<%BANNER%>

Using Antibody and Cell-Mediated Immune Response to Test Antigens in Periparturient Dairy Cows as a Measure of Disease R...

Permanent Link: http://ufdc.ufl.edu/UFE0022360/00001

Material Information

Title: Using Antibody and Cell-Mediated Immune Response to Test Antigens in Periparturient Dairy Cows as a Measure of Disease Resistance
Physical Description: 1 online resource (134 p.)
Language: english
Creator: De La Paz, Jason
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: antibody, categorization, cell, disease, immune, mediated, resistance, response
Veterinary Medicine -- Dissertations, Academic -- UF
Genre: Veterinary Medical Sciences thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Despite major advances in the dairy industry for sanitation, housing, milking strategies, and genetic trait selection; incidence of disease is still rising for Holstein dairy cows. This has sparked research aimed at identifying ways to incorporate genetic selection to improve broad-based immune responsiveness. For this to become a possibility, immune function must become a trait which can be quantified and correlated with risk of disease. For this study, both branches of adaptive immune function were considered due to the potential for an inverse relationship between the two. As a result, 774 cows were categorized based on their ability to mount an antibody-mediated immune response (AMIR), and 812 cows were categorized based on their ability to mount a cell-mediated immune response (CMIR). Immune response categorizations were high, medium, or low, such that the measured immune response for high > medium > low. Categorization status for AMIR as well as CMIR was found to be significantly associated with mastitis occurrence. Medium immune responders were 1.76 and 2.14 times more likely to have an occurrence of moderate or severe mastitis than high immune responders for AMIR and CMIR respectively. Low AMIR cows were 2.90 times more likely to have an occurrence of ketosis than high responders. This association with ketosis followed a low > medium > high pattern. For CMIR, low responders were 6.68 times more likely than high responders to have a retained fetal membrane (RFM). When only considering multiparous cows low responders for CMIR were 26.52 times more likely than high responders to have an occurrence of RFM. Although not statistically significant, medium CMIR status cows were 7.40 times more likely than high responders to have an occurrence of metritis. When considering the performance traits of fertility and milk yield, high AMIR status was associated with reduced fertility and reduced milk yield. However, high CMIR cows produced significantly greater milk yields than medium and low responders. Negative associations between higher levels of AMIR and reduced milk yield are likely attributed to neglecting immune function as a genetic selection trait. Associations between immune function and ketosis provide evidence for immune system involvement with energy-related metabolic conditions.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Jason De La Paz.
Thesis: Thesis (M.S.)--University of Florida, 2008.
Local: Adviser: Donovan, Gerald A.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2009-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022360:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022360/00001

Material Information

Title: Using Antibody and Cell-Mediated Immune Response to Test Antigens in Periparturient Dairy Cows as a Measure of Disease Resistance
Physical Description: 1 online resource (134 p.)
Language: english
Creator: De La Paz, Jason
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: antibody, categorization, cell, disease, immune, mediated, resistance, response
Veterinary Medicine -- Dissertations, Academic -- UF
Genre: Veterinary Medical Sciences thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Despite major advances in the dairy industry for sanitation, housing, milking strategies, and genetic trait selection; incidence of disease is still rising for Holstein dairy cows. This has sparked research aimed at identifying ways to incorporate genetic selection to improve broad-based immune responsiveness. For this to become a possibility, immune function must become a trait which can be quantified and correlated with risk of disease. For this study, both branches of adaptive immune function were considered due to the potential for an inverse relationship between the two. As a result, 774 cows were categorized based on their ability to mount an antibody-mediated immune response (AMIR), and 812 cows were categorized based on their ability to mount a cell-mediated immune response (CMIR). Immune response categorizations were high, medium, or low, such that the measured immune response for high > medium > low. Categorization status for AMIR as well as CMIR was found to be significantly associated with mastitis occurrence. Medium immune responders were 1.76 and 2.14 times more likely to have an occurrence of moderate or severe mastitis than high immune responders for AMIR and CMIR respectively. Low AMIR cows were 2.90 times more likely to have an occurrence of ketosis than high responders. This association with ketosis followed a low > medium > high pattern. For CMIR, low responders were 6.68 times more likely than high responders to have a retained fetal membrane (RFM). When only considering multiparous cows low responders for CMIR were 26.52 times more likely than high responders to have an occurrence of RFM. Although not statistically significant, medium CMIR status cows were 7.40 times more likely than high responders to have an occurrence of metritis. When considering the performance traits of fertility and milk yield, high AMIR status was associated with reduced fertility and reduced milk yield. However, high CMIR cows produced significantly greater milk yields than medium and low responders. Negative associations between higher levels of AMIR and reduced milk yield are likely attributed to neglecting immune function as a genetic selection trait. Associations between immune function and ketosis provide evidence for immune system involvement with energy-related metabolic conditions.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Jason De La Paz.
Thesis: Thesis (M.S.)--University of Florida, 2008.
Local: Adviser: Donovan, Gerald A.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2009-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022360:00001


This item has the following downloads:


Full Text

PAGE 1

USING ANTIBODY AND CELL-ME DIATED IMMUNE RESPONSE TO TEST ANTIGENS IN PERIPARTURIENT DAIRY COWS AS A MEASURE OF DISEASE RESISTANCE By JASON MICHAEL DE LA PAZ A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2008 1

PAGE 2

2008 Jason Michael De La Paz 2

PAGE 3

To my wife, the better half who inspires me to improve 3

PAGE 4

ACKNOWLEDGMENTS The road to completion of this degree program has certainly been educational for me and beneficial to the concept of disease resistance in Holstein dairy cattle. During this process there have been several individuals who served esse ntial roles and without their contribution, this venture would not have been as successful. Special recognition is extended toward the chai r of my supervisory committee, Dr. Arthur Donovan. This research was not possible without his efforts. I know I will always be grateful for this opportunity he chose to give me, which also provided a second chance to prove myself for veterinary school admission. His knowledge and practical approach to science is always very applicable and thought-provoking. I always en joy hearing his commentary on science publications; he has a way of uncovering details and revealing perspectives which are generally unconsidered. The initial study design was largely provided by the contributions of Dr. Bonnie Mallard and Dr. Armando Hernandez. Their efforts and experience with resear ch concerning immune function was critical to the succe ss of this project. I also must thank the owner of North Florida Holsteins, Mr. Don Bennink for his passion toward s the promotion of Holstein dairy cow wellbeing and willingness to allow his cows to be used for this study. Special thanks are given to the additional members of my supervisory committee, Dr. Pedro Melendez and Dr. Maureen Long. The insight into the statistical an alysis provided by Dr. Melendez certainly improved the research findi ngs. Dr. Longs knowledge of immune system helped determine the biological relevance of the data, which was essential to this study. All members of the FARMS department are well deserving of special appreciation for their contributions. I thank Dr. Louis Archibald fo r his knowledge and sincere enthusiasm for education and research. I was fortunate that Dr. Archibald appointed me as teaching assistant for 4

PAGE 5

veterinary theriogenology for one semester. I al ways enjoyed the conversa tions with Dr. Carlos Risco; whether referencing a reproductive con cern, a seminar topic, or the national champion football or basketball team, they were always very insightful. I am thankful for all the opportunities he provided for me working with Hols tein dairy cows or the ruthless water buffalo. I am very grateful to Dr Owen Rae for his know ledge, respect he shows others, and willingness to help answer any question that popped into my head. I also have to thank Deloris Foreman for her skills managing the department, problem-s olving ability, and her eagerness to organize various events for the department. I certainly en joyed our conversations about fishing. I also have to thank Dr. Pablo Pinedo and Dr. Maur icio Benzaquen. Both were fellow graduate students who befriended me; I lear ned a great deal from them bot h. They especially showed me how to bring a juicy red apple to their teacher for every class. During this degree program, I was also fortuna te enough to serve as teaching assistant for veterinary embryology and histology. Both opportunities woul d not have been possible or as successful without the efforts of Dr. Roger Reep. To prepare for these experiences, Dr. Reep gave me individual weekly training sessions. It is often said that te aching is the best way to learn, and this was exceptionally true for me. I will su rely have an advantag e in these courses for veterinary school. Being a father to my 2-year-old daughter Emily with a wife who must work, a typical concern would be for the wellbeing and teaching of my daughter in daycare. However, thanks to the eagerness of my parents, I have never ha d to worry about these matters. Their unrelenting willingness to care for Emily during the day has provided a great deal of relief knowing she is in good hands. This relief has allowed me the opportunity to focus more clearly on all my graduate 5

PAGE 6

studies. While on the subject of Emily, I should th ank her for being such a wonderful little girl and heavy sleeper; she began sleeping through the night at only 3 weeks old. I am sure I would not be anywhere close to he re if not for my wife. Without her, I would have settled for less a long time ago. By simply be ing herself, she makes me want to continue improving as a person. She values the important th ings in life; she is beautiful, strong, and always behind me for support. She gives more than her share to our marriage while working hard as a bank manager and terrific mother to our fi rst daughter Emily. For all these things and many more, she is the love of my life. 6

PAGE 7

TABLE OF CONTENTS page ACKNOWLEDGMENTs ................................................................................................................4 LIST OF TABLES .........................................................................................................................10 LIST OF FIGURES .......................................................................................................................11 ABSTRACT ...................................................................................................................................13 CHAPTER 1 INTRODUCTION................................................................................................................. .15 Disease Trend .........................................................................................................................15 Immune Suppression ..............................................................................................................15 Selection for Disease Resistance ............................................................................................15 Objectives ...............................................................................................................................18 2 REVIEW OF LITERATURE.................................................................................................19 Immune System Basics ...........................................................................................................19 Introduction .....................................................................................................................19 Innate Immunity ..............................................................................................................19 Acquired Immunity .........................................................................................................20 Intracellular immunity ..............................................................................................20 Extracellular immunity .............................................................................................21 Immunologic memory ..............................................................................................22 Relationship Between Intracellular and Extracellular Immunity ...........................................23 Delayed-Type Hypersensitivity ..............................................................................................24 Indirect Enzyme-Linked I mmunosorbent Assay (ELISA) .....................................................25 Periparturient Immune Suppression .......................................................................................26 Neuroendocrine Effect.....................................................................................................26 Effect of Negative Energy Balance .................................................................................27 Nonesterified fatty acids ...........................................................................................27 Effect of hyperketonemia .........................................................................................28 Effect of Hypocalcemia ...................................................................................................29 Lactogenesis Effect .........................................................................................................29 Dexamethasone ................................................................................................................30 Disease Trend .........................................................................................................................30 Selection for Disease Resistance ............................................................................................31 Direct Versus Indirect Selection ......................................................................................32 Disease Resistance Through Artificial Insemination ......................................................33 Somatic cell count ....................................................................................................33 Structural traits of th e udder and productive life ......................................................34 Disease Resistance Through Specific A ttributes of the Immune System .......................35 7

PAGE 8

Lymphocyte subsets .................................................................................................35 Major histocompatability complex ...........................................................................35 Broad-Based Immune Responsiveness ............................................................................36 Differences between high and low immune responders ...........................................37 Correlation with infectious disease risk ...................................................................38 Correlation with energy-related metabolic disease ..................................................39 Correlation with milk production .............................................................................41 Heritability of measures for immune responsiveness ..............................................42 Categorizing AMIR and CMIR .......................................................................................43 Introduction ..............................................................................................................43 Antibody-mediated immune response ......................................................................44 Cell-mediated immune response ..............................................................................45 3 METHODS UTILIZED FOR THE CA TEGORIZATION OF AMIR AFTER GENERATING ELISA OPTICAL DENSITY VALUES.....................................................53 Introduction .............................................................................................................................53 Study Population .....................................................................................................................53 Exclusion Criteria ............................................................................................................54 Removal of Cows Previously Exposed to Antigen .................................................................54 Methodology and Results ................................................................................................55 Effect of Parity on Antibody Response to Ovalbumin...........................................................55 Interval Variation Adjustment ................................................................................................55 Methods for OD Adjustment ...........................................................................................57 Results OD-3.............................................................................................................58 Results OD0..............................................................................................................58 Results OD+2.............................................................................................................59 Analysis of Classification Methods ........................................................................................60 Methods ...........................................................................................................................61 Results and Discussion ....................................................................................................62 Alternate Index Methods for AMIR Categorization .......................................................65 Results .............................................................................................................................67 4 CATEGORIZATION OF PERIPARTURIENT ANTIBODY RESPONSE TO OVALBUMIN AND ITS RELATIONSHIP WITH COMMON DISEASES AND PERFORMANCE MEASURES OF HOLSTEIN DAIRY CATTLE....................................82 Introduction .............................................................................................................................82 Materials and Methods ...........................................................................................................84 Research Sample Cows ...................................................................................................84 Animal Removal and Interval Criteria ............................................................................84 Body Condition Scoring ..................................................................................................85 Immunization...................................................................................................................86 Blood Collection and Processing .....................................................................................86 Enzyme-Linked Immunosorbent Assay ..........................................................................86 Preliminary Analysis .......................................................................................................88 Removal of non-naive ..............................................................................................88 8

PAGE 9

Parity effect ..............................................................................................................89 Interval analysis and opti cal density value adjustment ............................................89 Classification analysis ..............................................................................................90 Antibody-Mediated Immune Response Classification ....................................................93 Diseases ...........................................................................................................................94 Milk Yield, Somatic Cell Scor e, and Reproductive Efficiency .......................................95 Statistics ...........................................................................................................................96 Results .....................................................................................................................................97 Mastitis and Metritis ........................................................................................................97 Ketosis Displaced Abomasum and Retained Fetal Membrane .......................................97 Milk Yield and Somatic Cell Score .................................................................................98 Reproductive Efficiency ..................................................................................................98 Discussion ...............................................................................................................................98 5 CATEGORIZATION OF PERIPARTUR IENT CELL-MEDIATED IMMUNE RESPONSE TO A TEST ANTIGEN AND ITS RELATIONSHIP WITH COMMON DISEASES AND PERFORMANCE MEASUR ES OF HOLSTEIN DAIRY CATTLE....109 Introduction ...........................................................................................................................109 Materials and Methods .........................................................................................................110 Research Sample Cows .................................................................................................110 Animal Removal Criteria ..............................................................................................110 Body Condition Scoring ................................................................................................111 Delayed Type Hypersensitivity .....................................................................................111 Classification of Cell-Me diated Immune Response......................................................112 Diseases .........................................................................................................................113 Milk Yield, Somatic Cell Scor e, and Reproductive Efficiency .....................................114 Statistics .........................................................................................................................114 Results ...................................................................................................................................115 Mastitis ..........................................................................................................................115 Retained Fetal Membrane ..............................................................................................116 Metritis ..........................................................................................................................116 Ketosis and Displaced Abomasum ................................................................................116 Milk Yield, Somatic Cell Sc ore and Reproductive Efficiency ......................................117 Discussion .............................................................................................................................117 LIST OF REFERENCES .............................................................................................................124 BIOGRAPHICAL SKETCH .......................................................................................................134 9

PAGE 10

LIST OF TABLES Table page 3-1 Incidence of disease for each antibody response categorization method. .........................81 4-1 Odds ratios of disease inciden ce for antibody respons e categorizations. ........................105 5-1 Odds ratios of disease incidence for cell-mediated immune categorizations. .................121 10

PAGE 11

LIST OF FIGURES Figure page 2-1 General depiction of a primary a nd secondary response to antigen x ............................47 2-2 Depiction of the inverse relationship be tween AMIR and CMIR as a result of IL-10 ......48 2-3 Plasma NEFA in thin, medium and ov erconditioned cows during the peripartum period .................................................................................................................................49 2-4 Effects of NEFA on IgM secreti on in peripheral blood mononuclear cells ......................50 2-5 Effects of NEFA on interferonsecretion in periphera l blood mononuclear cells. .........51 2-6 Flow chart describing potential relations hip between energy-related metabolic condition (ketosis) and infectious condition (metritis) ......................................................52 3-1 Basic outline for treatment/ sampling fo r antibody mediated immune responsiveness .....70 3-2 Optical density values reflecting antibody response to OVA by sampling period............71 3-3 Scatter plot of OD-3 values by length of interval 1 for primiparous cows. ........................72 3-4 Scatter plot of OD-3 values by length of interval 1 for multiparous cows. ........................73 3-5 Scatter plot of OD0 values by length of interval 2 for primiparous cows. .........................74 3-6 Scatter plot of OD0 values by length of interval 2 for multiparous cows. .........................75 3-7 Scatter plot of OD+2 values by length of interval 3 for primiparous cows. .......................76 3-8 Scatter plot of OD+2 values by length of interval 3 for multiparous cows. ........................77 3-9 Graph reflecting the effect of maximal antibody response ................................................78 3-10 Graph reflecting the effect of maximal antibody response ................................................79 3-11 Depiction of a method used to categorize antibody mediated immune responsiveness.. ..................................................................................................................80 4-1 General outline of experimental design. ..........................................................................102 4-2 Polystyrene 96-well plate for en zyme linked immunosorbent assay. ..............................103 4-3 Diagram of the placement of test sera into 96 well polystyre ne plate for enzymelinked immunosorbent assay. ...........................................................................................104 4-4 Incidence ketosis by AMIR categorization within parity. ...............................................106 11

PAGE 12

4-5 Graph for the effect of antibody-mediat ed immune response (AMIR) categorization on milk yield. ...................................................................................................................107 4-6 Graph for the effect of antibody-mediat ed immune response (AMIR) categorization on pregnancy by 150 DIM. ..............................................................................................108 5-1 Graph showing increase in double skin -fold thickness respective of parity (multiparous or primiparous). ..........................................................................................119 5-2 Graph revealing parity differen ce for cell-mediated immune response ...........................120 5-3 Graph indicating the difference in risk of RFM between high and low cell-mediated immune response categorization. .....................................................................................122 5-4 Graph for the effect of cell-mediated immune response (CMIR) categorization on milk yield. ........................................................................................................................123 12

PAGE 13

Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science USING ANTIBODY AND CELL-ME DIATED IMMUNE RESPONSE TO TEST ANTIGENS IN PERIPARTURIENT DAIRY COWS AS A MEASURE OF DISEASE RESISTANCE By Jason Michael De La Paz August 2008 Chair: Arthur Donovan Major: Veterinary Medical Sciences Despite major advances in the dairy industr y for sanitation, housi ng, milking strategies, and genetic trait selection; incide nce of disease is still rising fo r Holstein dairy cows. This has sparked research aimed at identifying ways to incorporate genetic selection to improve broadbased immune responsiveness. For this to become a possibility, immune function must become a trait which can be quantified and correlated with risk of disease. For th is study, both branches of adaptive immune function were considered due to the potential for an inverse relationship between the two. As a result, 774 cows were categorized based on their ability to mount an antibody-mediated immune response (AMIR), an d 812 cows were categorized based on their ability to mount a cell-mediated immune res ponse (CMIR). Immune re sponse categorizations were high, medium, or low, such that the meas ured immune response for high > medium > low. Categorization status for AMIR as well as CM IR was found to be significantly associated with mastitis occurrence. Medium immune re sponders were 1.76 and 2.14 times more likely to have an occurrence of moderate or severe ma stitis than high immune responders for AMIR and CMIR respectively. Low AMIR cows were 2.90 ti mes more likely to have an occurrence of ketosis than high responders. Th is association with ketosis followed a low > medium > high 13

PAGE 14

pattern. For CMIR, low responders were 6.68 times more likely than high responders to have a retained fetal membrane (RFM). When only considering multiparous cows low responders for CMIR were 26.52 times more likely than high responders to have an occurrence of RFM. Although not statistic ally significant, medium CMIR status cows were 7.40 times more likely than high responders to have an occurrence of metritis. When considering the performance traits of fertility and milk yield, high AMIR status was associated with reduced fertility and reduced milk yield. However, high CMIR cows produced significantly greater milk yields than medium and low responders. Negative associations between higher levels of AMIR and reduced milk yield are likely attributed to neglecting immune function as a genetic selection trait. Associations between immune function and ketosis provide evidence for immune system involvement with energyrelated metabolic conditions. 14

PAGE 15

CHAPTER 1 INTRODUCTION Disease Trend The susceptibility of dairy cows to infectious disease is increasing. Genetic selection for increased milk yield without regard for disease resistance may be fueling this adverse effect (Harmon, 1994; Heringstad, 2000; Emanuelson, 1988). The rampant selection for increased milk production void of measures for resistance to ma stitis has been found to result in a genetic increase of 0.02 cases of mastitis per cow per ye ar (Strandberg and Shook, 1989). This translates into a genetic increase of 2 mastitis cases for every 100 dairy cows per year. Immune Suppression Immune suppression experienced around the time of calving has been well documented (Mallard et al., 1998; Lacetera et al., 2005; Kimura et al., 2006). This suppression is believed to be at least in part responsible for the increased risk of dis ease peripartum. The added stress associated with parturition and the abrupt change in lifestyle work together to suppress immune function (Mallard et al., 1997; Van Ka mpen and Mallard, 1997). Di fferent mediators of this immune depression include; endocrine hormones (Mallard et al., 1997), hypocalcemia (Kimura et al., 2006), and non-esterified fa tty acids (NEFA) (Lacetera et al., 2004, 2005). Selection for Disease Resistance Since a national database for Holstein health disorders in the United States does not exist, it is impossible to genetically select directly against specific health disorders. Artificial insemination has provided some opportunity for producer s to select for certain traits that vary in their degree of associat ion with disease resistance. Selecti on to improve somatic cell score (SCS) (a logarithmic transformation of somatic cell coun t), productive life and st ructural udder traits (udder cleft, udder depth, rear udder height, rear udder width, and fore udder attachment) have 15

PAGE 16

all been found to be significant pr edictors of susceptibil ity to clinical mastit is (Nash et al., 2000). Six studies estimating the genetic correlation be tween SCS and clinical mastitis all indicated a positive correlation averaging 0.71, and ranged from 0.37 to 0.98 (Nash et al., 2000). However, the relationship between SCC and intra-mammary inf ection is still unclear (Piccinini et al., 1999; Schukken et al., 1999). Also, with the exception of productive life, these traits are only associated with infections of the mammary gland. The increasing incidence of disease associat ed with selection for increasing milk production may be partially explained by the associa tion of increased severity and prevalence of immune suppression with elevated NEFA and deficiencies in calcium. These two postulated factors in immune suppression are also potentially correlated with increasing milk yield. However, previous work also indicates a st rong genetic influence on immune responsiveness (Wagter et al., 2000; Mallard et al., 1998, Biozzi et al., 1979). Application of measures to place genetic selection pressure on immune responsiveness could potentially be used to overcome increasing infectious disease tre nds. Concerns for a negative asso ciation between milk yield and genetic potential for immune responsiveness are not substantiated (Wagter et al., 2003; Detilleux et al., 1995). Breeding to improve immune function is a conc ept which has been examined and utilized in poultry (Soler et al., 2002; Cole, 1968), sw ine (Mallard et al., 1998), sheep (Woolaston and Baker, 1996), and mice (Biozzi et al., 1979). Ho wever, work aimed at categorizing general immune responsiveness in Holstein cattle is just recently gain ing attention (Wagter et al., 2000, 2003; Hernandez et al., 2003). The obvious motive be hind research aimed at identifying superior and inferior immune responsiveness is to serve as an indirect tr ait enabling selection for general broad-based disease resistance. This concept is appealing for several re asons. Since eradication 16

PAGE 17

of environmental infection-cau sing organisms is impossible, the domestic bovine must rely primarily on its immune system to fight pathogens Thus, cows are exposed to a wide variety of pathogens which are also proficie nt at altering their vi rulence mechanisms. As a result, selection should be for the cattle with the most robust repertoire of response against a variety of pathogens. Selection for increased immune re sponsiveness should reduce the strong dependency on vaccines and antibiotics. Increases in antibody titers to vaccinat ions should result in more efficient use of vaccine dosage and potentially a reduction in vaccine dosage (Wagter et al., 2000; Mallard et al. 1997). Selecting for increase d immune responsiveness should also take steps toward addressing the concerns consumers have for excessive use of antibiotics and animal welfare issues. Finally, through a reduction in ma stitis occurrence, SCC should also be reduced leading to increased cheese yiel d, dairy product quality and shelf life (National Mastitis Council, 1996). Research by Wagter et al. (2000) has found a significant vari ation in a cows ability to mount a humoral immune response. They also found that not all co ws experience immune suppression around calving. This research used ov albumin (OVA) as a novel antigen to elicit an antibody mediated immune response (AMIR) in 1 36 Holstein cows and heifers. They found the high responders had the lowest incidence of mastitis. They also found that antibody responsiveness to ovalbumin positively correlated with antibody response to the E. coli J5 vaccination (Rhne Mrieux, Lenexa, KS). Detil leux et al. (1995) found adequate genetic potential for immune responsiveness to exist among sires with top PTA for milk yield. An additional study found that selection for immune responsiveness does not predispose cows to reduced milk yield (Wagter et al., 2003). 17

PAGE 18

Objectives Using 136 cows the previous study by Wagter et al. (2000) had difficu lty finding statistical significance between AMIR and dis ease risk. Also, this study did not include a measure of cellmediated immune response (CMIR), which is the other branch of a cows adaptive immunity. In the present study we have further an alyzed the association between immune responsiveness and disease through th e use of a larger sample size. The objectives of this study were to categorize a cows humoral response to ovalbumin as a measure of AMIR as well as delayed-type hypersensitivity (DTH) response to Candida albicans as a measure of CMIR (Hernandez et al., 2003, 2005). Associations between these immune categorizations and disease risk, namely mastitis, metritis, retained fetal membrane, ketosis, and displaced abomasum were tested. The effect of these immune categori zations on milk production, somatic cell count, and fertility were also tested. 18

PAGE 19

CHAPTER 2 REVIEW OF LITERATURE Immune System Basics Introduction In order to better understand the mechanisms involved in this research study, a basic review of the immune system is required. The im mune system is comprised of several defense mechanisms. Most potential pathoge ns never elicit an immune re sponse because the bodys first line of defense, the epithelial surface, protects against the establishment of infection. The epithelial surface defense mechanisms can be divided into those mechanical, chemical, and microbiological. Mechanical is us ed to describe the defense provided through tight junctions between epithelial cells, or the movement of mucus by cilia in the respiratory tract. In the stomach, the low pH, enzymes (pepsin), antimicrobi al fatty acids, and peptides found in the stomach provide a chemical barrier to inf ection (Risso, 2000). The normal flora found on the skin and in the gut provides a microbiological defense mechanism. The initial response is characterized as an innate nonspecific immune response while extended and/or repeat exposure to pathogen leads to an acquired immune response acting only on specific antigen. Innate Immunity When foreign microbes are able to pass the bodys first line of defense, the immediate response is classified as innate immune re sponse. Innate immune function is adept at distinguishing self from non-self by the use of cell-surface receptors whic h react with general features that are common among microbes called pathogen associated molecular patterns (PAMPs). Therefore this res ponse is characterized as nonspe cific effector recognition of pathogen which involves the phagocytic and infl ammatory activity of the leukocytes, largely macrophages and neutrophils (Beutler and Rietschel, 2003). Macrophages and neutrophils 19

PAGE 20

release various cytokines and chemokines which can initiate the inflammatory response. The inflammatory response works by recruiting additiona l effector molecules to the infection site, reducing the spread of infecti on by microvascular coagulation, and by promoting tissue repair (van der Poll, 2001). Acquired Immunity Infectious disease occurs when a microor ganism succeeds in evading or overwhelming innate immune defenses. It is at this point where the acquired or adaptive immune response is required. Acquired immune function is characteri zed by specific recognition of antigen and has two main branches which provide means for th e elimination of both ex tra and intracellular pathogens. Elimination of extr acellular pathogen is performe d by antibody-mediated immune response (AMIR) pathways. Eradication of intra cellular pathogen is performed by cell-mediated immune response (CMIR) pathways. Intracellular immunity In certain instances, some forms of bacteria, parasites, and all viruses, replicate within cells and are not detected by extra cellular immunity. On these oc casions, the body employs methods to combat intracellular pathogens which largely involve the functi on of T-lymphocytes (T-cells). T-cells only respond to antigen that is accomp anied by a major histocompatibility complex (MHC), forming an MHC:antigen complex. This reac tion requires a specific T-cell receptor for a response to take place (Jensen, 2007). There are two classes of MHC molecules: MH C class I molecules are expressed by nearly every nucleated cell in vertebrates. MHC class II molecules are only ex pressed by professional antigen presenting cells (APC), which are a special group of phagocytic cells. Dendritic cells are the most common APC, but B-lymphocytes a nd macrophages can also function as APCs (Jensen, 2007). These APCs specialize in disp laying a small portion of processed antigen bound 20

PAGE 21

to an MHC class II molecule on the APCs surface (Savina and Amigorena, 2007). Antigen processing for presentation with an MHC class II molecule involves the endocytic pathway. This pathway begins when exogenous antigen has been endocytosed by an APC. There are two main classes of T-cells; CD4+ T-cells (T H ), and CD8+ T-cells (T C ). The T C cell is responsible for cytotoxic activity or cell killing. These cells initiate their cytotoxic activity on cells displaying antigen bound to an MHC class I molecule wh ich is expressed by nearly every nucleated cell in vertebrates. The T H cell is probably the most important of the T-cells. T H cells only respond to antigen bound to an MHC class II molecule, meaning they only respond to antigen presented by dendritic cells, macrophages or B-lymphocytes, also called APCs. There are two subsets of T H cell; T H 1 subset is responsible for cell-mediated functions such as delayedtype hypersensitivity (DTH) and act ivation of cytotoxic T-cells; T H 2 subset is largely responsible for B-cell activation. The functions of the two subsets of T H cell serve to activate cells in both branches of the adaptive imm une response, which is why T H cells are of special importance. Nave T-cell activation requires specific antigen presentation by an APC to a T H cell. The activation of a T H cell initiates the release of cytokines which can: activate the cytotoxic activity of T C cells, stimulate the chemotaxis of leukocyt es, activate B-lymphocytes, and also cause differentiation into memory T-cells. Activated T H and T C cells have a relatively short life-span while the memory T-cells being long-lived can last for the duration of the cows life. Extracellular immunity B-cells use immunoglobulins for antigen r ecognition and are c oncerned with the elimination of extracellular pathogens. Non-membranous imm unoglobulins, called antibodies, function by binding to the antigen which elicited the response. This binding or coating (opsonization) can neutralize the pathogen, and also flags the pathoge n for phagocytosis or complement activity. For a B-cell to differentiate into effector cells (activation), they require 21

PAGE 22

accessory signals from an activated T H 2 T-cell. B-cell activation occu rs when they bind antigen with its membranous immunoglobulin. The antigen is then internalized and degraded. A peptide fragment from the antigen is later displayed on the cells surface with an MHC class II molecule. A specific interaction between the MHC: antigen peptide complex and an armed T H 2 T-cell send activation signals to the B-cell. This allows th e differentiation into an tibody secreting plasma cells and memory B-cells (Parker, 1993). Immunologic memory Adaptive immunity is also associated with immunologic memory which results in a more rapid and effective immune response to pathogens that have been previously encountered. An antibody response profile indicates features which are common to every antibody response (Figure 2-1). The extent of these features or kinetics are differe nt for a primary immune response (first antigen encounter) and s econdary immune response (>1 antig en encounter). After antigen exposure, every antibody response begins with an initial lag phase to allow for somatic hypermutation and clonal differentiati on into effector cells. This pha se is followed by an increase in antibody concentration until a peak concentrati on is reached. After a peak response is reached, it is followed by a steady decline in antibody co ncentration. In a primary response the humoral response results from the activation of nave lymphocytes, whereas in a secondary immune response it is memory lymphocytes which are activated. These memory lymphocytes have greater affinity for their specific antigen, wh ich facilitates greate r immune response upon repeated exposure. Memory lymphocytes are also long-lived, and can prov ide life-long immunity to their specific pathogen. A secondary immune re sponse is characterized by a shorter lag phase, greater magnitude, longer duration and greater antibody affinity to antigen when compared to a primary immune response. A primary antibody response consists primarily of IgM isotype antibody, while a secondary response consists larg ely of IgG isotype. Ther e is also substantial 22

PAGE 23

variation in the kinetics of an antibody res ponse due to antigen type, administration route, presence of adjuvant, and species exposed to antigen. Relationship Between Intracellular and Extracellular Immunity Several studies have cited inverse relationships between intracellular immunity and extracellular immunity (de Vries, 1995; Biozzi et al., 1979; Rupp et al., 2007). Although the three referenced studies support the claim of an in verse relationship, they are all concerned with different aspects of the immune system. The work performed by Biozzi et al (1979) supported an inverse relationship between CMIR and AMIR on the basis of intracellular catabolism of antigen. Breeding mice for high antibody mediated immune responsiveness (AMIR) was associated with slower intracellular catabolism of antigen. The explanation for this finding was that slowed antigen processing also related to prolonged antigen presentation which re sults in greater stimulation for the production of antibodies. The mice selected for high AMIR we re more resistant to extracellular pathogens, however; the support for an inverse relationship en ters when these high AMIR mice were more susceptible to intracellular pathogens. The slow ed intracellular catabolism which was favorable for AMIR is believed to be unfavorably associat ed with relevant measures of cell-mediated immune function (CMIR). Research conducted by Rupp et al., (2007) found support for an inverse relationship on the basis of MHC alleles at the DRB3.2 locus. As previously discussed, MHC molecules are responsible for antigen presentation to T ly mphocytes. This experi ment found significant relationships between available alleles at the MHC locus DRB 3.2 and measures of AMIR and CMIR. However, these relationships were invers ely related. Alleles which confer high measures for CMIR associated with low measures of AMIR, and vise versa. 23

PAGE 24

Another paper by de Vries (1995) outlined a mediator for an inverse relationship on the basis of cytokine expression by T H 2 cells. As previously described, T H 2 lymphocytes are involved with activation of an AMIR while T H 1 lymphocytes are involved with activation of a CMIR. T H 1 lymphocytes secrete the cyt okine interleukin-10 (IL-10). Th is release of IL-10 works to block the function of T H 2 lymphocytes (Figure 2-2). Delayed-Type Hypersensitivity Hypersensitive immune responses are often termed, inappropriate immune responses to antigen. There are four types of hypersensitiv e reactions, three of which are mediated by antibodies (Type I-III), and one is mediated by T-lymphocytes (Type IV). A Type IV hypersensitivity response is called a delayed-ty pe hypersensitivity reaction (DTH) due to a characteristic delayed response by comparison to the acute phase reactions of immediate hypersensitivity (Type I). A DTH reaction gene rally peaks around 24-72 hrs post secondary exposure. The other three hypersensitive reac tions are termed immediate hypersensitivity because reaction peaks occur within mi nutes or hours of secondary exposure. A DTH immune response requi res the stimulation of T H 1 cells to form memory T H 1 cells. Therefore a DTH reaction requires previous expo sure to the specific antigen. Upon secondary antigen exposure, antigen presentation cells (A PC) take up the antigen and display it in conjunction with MHC class II molecule to the previously formed memory T H 1 cells. When memory T H 1 cells bind to APCs, cytokines are rele ased and chemotaxis of predominantly macrophages and neutrophils occur at the exposure site resulting in a granuloma. Clinically, a palpable lump occurs which is largely composed of these macrophages a nd neutrophils and to a far less degree, T-cells. DTH reactions are frequently used to detect exposure to large intracellular antigens. The most common of these antigens would be Mycobacterium tuberculosis DTH reactions have also 24

PAGE 25

been previously used as a means in which to qu antify cell-mediated immune function (Mallard et al., 1998; Hernandez et al., 2003, 20 05). There are no other known in vivo methods for quantifying CMIR. In order to mount a DTH i mmune response the subj ect must have been previously exposed to the antigen at least two weeks prior. Th is is required to provide an adequate period for T-cell clonal expansion and di fferentiation into memory T-cells. The antigen is then injected into the subj ect intradermally and after 24-72 hour s, detection of this responseis by examination for a palpable lump at the injection site. Indirect Enzyme-Linked Immunosorbent Assay (ELISA) The indirect ELISA is a method for detection and quantification of antibody that reacts against a specific antigen. These methods for indi rect ELISA have been used in the past as a means to quantify an individual cows antibody mediated immune responsiveness to a novel antigen (Burton et al., 1989; Ma llard et al., 1997; Wagt er et al., 2000). For the indirect ELISA, antigen is coated to the surface of a microtiter well and after su fficient time, excess free antigen is washed away. Nonspecific reactions are bl ocked by a non-reactive protein. Serum (or some other fluid) which potentially cont ains the anti-antigen primary an tibody of interest, is added to the microtiter well. This allows the primary antib ody to bind the antigen which is attached to the side of the well. After sufficient time has el apsed, the free unbound antibody is washed away. An enzyme-conjugated secondary antibody is added which binds to the constant region of the primary antibody which is adhered to the side of the well. After sufficient time, free secondary antibody-enzyme conjugate is washed away and a substrate for the remaining enzyme is added. The enzyme-substrate reaction is a color produci ng reaction which is qua ntified in terms of optical density using a specialized spectrophotometer. This application can be used to detect previous exposure to a particular infectious disease (Wild, 2001). 25

PAGE 26

Periparturient Immune Suppression The periparturient period for a dairy cow is accompanied by many abrupt events and changes in management which provide stress during this time frame. Various management strategies are employed during the transition period in an effort to ease this major adjustment. However, the act of parturition, lactation, changes in ration and en try into the milking herd have various implications which provide mediat ors for immune suppression. This immune suppression has been detected in several studi es (Lacetera et al., 2005; Saad et al., 1989; Detilleux et al., 1994; Park et al., 1992). This is believed to be at least partially responsible for the increased incidence of dis ease found soon after calving. Severa l studies have been conducted to identify the mediators responsible for immune suppression (Mallard et al., 1997; Lacetera et al., 2005; Kimura et al., 2006). Neuroendocrine Effect Neuroendocrine-immune modulation has been re searched due to the identification of neuroendocrine receptors on the surface of lymphoc ytes as well as their ability to release neurotransmitters and hormones such as growth hormone and insulin-l ike growth factor-1 (Badolato et al., 1994; Blalock, 1994). Findings such as these indicate th at metabolic changes induced by neuroendocrine mediators also have implications on the immune system (Besedovsky and Del Ray, 1996; Dardenne and Savino, 1996). To a certain degree imm une cells resemble small pituitary glands (Von Ruecker and Schmid t-Wolf, 2000). These discoveries have sparked research aimed at determining the effect stress hormones have on lymphocyte function in periparturient dairy cattle. Circulating levels of growth hormone (GH) has been found to have a positive correlation with antibody responsiveness to ovalbumin (OVA) measured in blood serum (r = 0.29, p 0.001) or milk whey (r = 0.31, p 0.0005) (Mallard et al., 19 97). Antibody produced in 26

PAGE 27

response to an E. coli J5 vaccination (Rhne Mrieux E. coli J5, Rhne Mrieux, Lenexa, KS) was also significantly correlated with GH (r = 0.18, p 0.04) (Mallard et al., 1997). Insulin-like growth f actor-1 (IGF-1) has been found to be negatively associated with antibody responsiveness measured in blood serum (r = -0.19, p 0.04) as well as milk whey (r = -0.22 p 0.01) (Mallard et al., 1997). Also, antibody responsiveness was significantly influenced by an interaction between week relative to calving and IGF-1 concentration (p 0.005) (Mallard et al., 1997). The correlation between cortisol concentration and antibody re sponsiveness has also been studied. Cortisol levels were f ound to be positively associated with antibody responsiveness (r = 0.17, p 0.06) (Mallard et al., 1997). The relationship these classical hormones have with each other was also examined. The relationship between GH and cortisol have shown to have a direct relationship with both having maximum concentrations at calving. On th e other hand, during this time frame, IGF-1 concentrations were at a minimum, yielding a negative correlation. GH and cortisol levels decreased in the weeks following calving, while IGF-1 concentrations increased until peak lactation (Mallard et al., 1997). Effect of Negative Energy Balance Nonesterified fatty acids Periparturient dairy cows at or near the onset of lacta tion, frequently undergo negative energy balance, which simply means, the energy requirements for milk production and final stages of calf development prepartum, exceed en ergy intake (Adewuyi et al., 2005). Dairy cows tend to amplify this condition by frequently exhi biting a reduction in dry matter intake beginning in the days prior to calving. To compensate fo r the deficit in energy, adipose tissue lipolysis occurs which produces free fatty acids in the bl ood called non esterified fatty acids (NEFA) 27

PAGE 28

(Adewuyi et al., 2005). Research has discovered that overconditioned cows not only lost significantly more body condition when compared to medium or thin conditioned cows, but also had significantly higher NEFA concentrations (F igure 2-3) (Lacetera et al., 2005). Periparturient dairy cow body condition and blood NEFA conc entration is negatively correlated with lymphocyte function as measured by reductions in peripheral blood mononuclear cell (PBMC); DNA synthesis, immunoglobulin M (IgM) secret ion (Figure 2-4), and interferon-gamma (IFN) secretion (Figure 2-5) (Lacetera et al., 2005, 2004). In this study body condition was a binary trait with overconditioned cows in one group wh ile medium and thin conditioned cows in the other. The significance of NEFA and/or body co ndition with lymphocyte function is generally only found in overconditioned cows. At the time of th is particular work it was speculated that for periparturient dairy cows, alterations in lymphoc yte function may proportionally relate to loss in body condition as assessed by changes in BCS. Effect of hyperketonemia The common state of negative energy balance for periparturient dairy cows also is a predisposing factor for development of hyperket onemia, a condition whereby levels of ketone bodies are elevated in the bl ood. The production of milk for t odays dairy cow requires large demands for glucose. To meet this demand duri ng a time of suppressed dry matter intake, dairy cows undergo intense gluconeogenesis. It is dur ing this time where a large portion of serum NEFA is directed to the liver which is then synthesized into ketone bodies. The serum ketones found in cattle are acetone, acetoacetate, and -hydroxybutyrate (Baird, 1982). Several studies have reported suppressed immune responsivenes s associated with the presence of elevated ketone bodies (Franklin et al ., 1991; Hoeben et al., 1997). However, other studies do not replicate this antagonistic re lationship with the immune sy stem (Nonnecke et al., 1992). 28

PAGE 29

Effect of Hypocalcemia Calcium plays a critical role in the activ ation of immune cells. Immune cell activation involves signal transduction pathways which involve inositol 1,4,5-tris phosphate binding to receptors on the endoplasmic reticulum (ER) which in turn stimulate the re lease of calcium ions (Ca 2+ ) into the cells cytoplasm (Grafton and Thwaite, 2001; Lewis, 2001). The level of the resulting rise in intracellular Ca 2+ has been used as a measure of immune cell responsiveness (Partiseti et al., 1994; Baus et al., 1996). Also, an in vivo study in rats showed that extracellular fluid calcium level is a primary indicator for intracellu lar calcium status (Mailhot et al., 2000). It is known that for a dairy cow in the peri parturient period, the de mands for calcium and risk of hypocalcemia greatly increase as producti on of colostrum and milk initiates. Because calcium is critical to immune cell activation, it has been hypothesized that the increased demands for calcium may unfavorably affect intracellula r calcium levels, which in turn could affect immune cell activation potential. Also, peripa rturient cows who have undergone mastectomy, do not develop hypocalcemia, and more importantly, do not encounter the same degree of immune suppression as lactating peripart urient cows (Goff and Kimura 2002; Nonnecke et al., 2003). Researchers have discovered th at calcium levels in the blood as well as calcium levels stored in the ER decline in the days up to cal ving. They also found serum calcium levels were significantly correlated w ith the intracellular Ca 2+ response as well as Ca 2+ stored in the ER (Kimura et al., 2006). Due to the tremendous pr oduction of milk and its demand for calcium, these research findings substantiate the claim that cows can be at least partially immune suppressed during the peripa rturient period due to de ficiencies in calcium. Lactogenesis Effect It has been hypothesized that immune suppression around calv ing is partly due to the sequestering of available syst emic immunoglobulins (Ig) into the mammary gland for colostrum 29

PAGE 30

and milk. This has also been theorized to expl ain the differences between high and low antibody responders to ovalbumin, where the low responders are low due to increased sequestering of Ig into the mammary gland. This theory was challenged when Wagter et al. (2000) compared the antibody response to OVA in serum to that in whey. For the theories to be supported by a correlation analysis, a negative or inverse rela tionship should be reveal ed, however, correlation analysis indicated a positive si gnificant relationship within ea ch test herd (Herd 1, r = 0.45, p <0.0001; Herd 2, r = 0.28, p <0.001; Herd 3, r = 0.44, p <0.001). Dexamethasone Dexamethasone is a synthetic glucocortico id which is commonly used to initiate parturition in cows within the last 30 days of gestation. This prepartum use provides consistent highly effective results; however, it has also been consistently associated with increased risk of retained fetal membranes (Beardsley et al., 1976; Peters and Poole, 1992). Also, several studies have identified a direct immunosupp ressive activity to measures of innate as well as adaptive immunity which are associated with the use of this glucocorticoid (Burton et al., 1995; Burton and Kehrli, 1995, 1996). Disease Trend Associated with breeding fo r increased milk production without regard for immune responsiveness, a concomitant rise in infec tious disease occurred (Emanuelson et al., 1988; Harmon, 1994; Wagter et al., 2003). For mastitis, Nord ic data reveal that genetic correlation with milk production ranges between 0.24 and 0.55 with an average of 0.43 when large field data sets are analyzed (Heringstad et al., 2000). After assuming a conservative gene tic correlation between mastitis and milk production of 0.30; Strandberg and Shook (1989) state that under traditional progeny testing programs without selection for mastitis, a genetic increase of 0.02 cases of 30

PAGE 31

mastitis per cow per year is the result. So for every 100 cows, there would be a genetic increase of 2 cases of mastitis per year. The mediators behind the correl ation between production and mas titis have not been fully determined. Detilleux et al. (1995) used 137 peripa rturient cows when they found selection for high milk yield did not produce genetic lines with unfavorable measures for innate and adaptive immune function. However, this study may not have been able to adequately reflect mediators for immune suppression which may bridge the gap between clinical mastitis and milk yield. The effect of NEFA (Lacetera et al., 2005), hyperketonemia (Suriyasathaporn et al., 2000) and calcium (Kimura et al., 2006) are three previously described me diators for immune suppression that can also correlate with milk production. As increased milk production per cow is achieved, demands for energy and calcium also increase This increase in demand for both energy and calcium could make the potential for a deficiency in both energy and calcium an ever increasing risk. Since a deficiency in energy can lead to the production of both NE FA and ketone bodies, and both can suppress the immune system. Also, a deficiency in calcium inhibits immune cell activation which also suppresses immune function (Kimura et al ., 2006); it can be hypothesized that selection for increased milk production without regard for immune responsiveness contribute to the mediators of immune suppres sion. The resulting increase in occurrence of immune suppression can be at least partly res ponsible for the increase d risk of disease. Selection for Disease Resistance The concept of breeding animals for disease resistance is not new. Selections against specific genes responsible for dise ase have provided means to pr event or reduce the risk of a given condition and have emphasized the effect gene tics plays in disease re sistance. Selection for disease resistance is most effective when the give n condition is associated with a single or small number of genes. In poultry, selections agains t specific genes responsib le for Mareks disease 31

PAGE 32

have made great strides (Cole, 1968). In dair y cattle, identification of a genetic defect responsible for bovine leukocyte adhesion deficiency (BLAD) has drastically reduced the risk of this condition (Kehrli et al., 1990). In sheep, the associa tion between a certain genotype and natural scrapie risk has been stud ied (Hunter et al., 1997). In swine, certain genes associated with risk of salmonella and E. coli diarrhea ha s been studied (Edfors-Lilia et al., 2000). Several different methods have been developed to select for disease resistance. Some of these methods are concerned with reducing the in cidence of a specific condition (Nash et al., 2000) or pathogen, while others ar e more broad-based in their a pproach and work to reduce the incidence of many related condi tions (Wagter et al., 2000). Generally speaking, more broadbased approaches tend to have slower genetic progress for a specific condition, but during that time frame the genetic progression is favorab le for a larger spect rum of conditions. Direct Versus Indirect Selection In certain instances, selection for disease re sistance is accomplished by directly selecting against the disease itself. Using this approach, one study compared mastitis frequencies of progeny from the best bulls for mastitis resistance to the progeny of the worst bulls (Steine, 1996). In this research, they found the three wo rst bulls for a mastitis resistance index had daughters with twice the mastitis frequency of daughters of the th ree bulls with the best index values. This type of selection requires an extensive database w ith records of disease occurrence in the pedigree of a given animal. Other methods of genetically selecting for dis ease resistance in dairy cows have involved the use of indirect traits (Nash et al., 2000; Heri ngstad et al., 2006; Wagter et al., 2000). Because there is not a national database for Holstein disease occurrence in the United States, it becomes impossible to directly select ag ainst particular inf ectious diseases. So identification of an 32

PAGE 33

effective indirect trait is required. For this to be useful, it must be correlated with the phenotype of interest (disease resistance), it must also be easy to measure a nd heritable (Kelm et al., 2001). Disease Resistance Through Artificial Insemination Somatic cell count In dairy cows, the widespread use of arti ficial insemination has enabled producers the ability to select sires based on predicted tr ansmitting abilities (PTA ) for phenotypic traits associated with disease resistance. Somatic cell sc ore (SCS) is a selection trait available for dairy producers using artificial insemination and is simp ly a logarithmic transformation of somatic cell count (SCC). Elevations in SCC are seen as a re sponse to microbial infestation in the mammary gland and are reported as the number of leukocytes present in 1 mL of milk. Thus, it can not only be used to identify the presence of microbes in the mammary gland but it can also to a certain degree be a potential measure of innate immune response to infection. Se lection for reduced SCS is considered because of the correlation betw een SCC and clinical mastitis. The association between SCC and clinical mastitis has been extens ively researched (Nash et al., 2000; Rogers et al., 1998; Heringstad et al., 2006). Several studie s cite strong correlations between SCC and clinical mastitis (Heringstad et al., 2006; Nash et al., 2000). Nash et al (2000) cites an average genetic correlation of 0.71 over six studies betwee n SCC and clinical mastitis. Heringstad et al. (2004) cites a range in genetic correlation de pendent on phase of lactation ranging from 0.37 to 0.73. The exact genetic relationship between selecti on for lower SCS and clinical mastitis is not completely understood. Because SCC reflects the amount of leukocytes in milk, and this can also be an indicator for innate immune responsiveness to microbial infestation; there is a belief that genetically selecting for reduced SCS in healthy cows may also be concurrently selecting for reduced immune responsiveness. Previous resear ch found that milk SCC prior to experimental 33

PAGE 34

challenge with S. aureus was actually higher in cows that re sisted infection compared to those who became infected (Piccinini et al., 1999; Sc hukken et al., 1999). However, there is substantial disagreement with this philosophy. Two studies found that high immune responders, determined by antibody responsiveness to ovalb umin, had significantly lower LS means for SCC than low responders (Wagter et al., 2000; Mallard et al., 1997). Additionally, Kelm et al. (1997) found a tendency for greater functional ability of neutr ophils from cows with lower estimated breeding values (EBV) for SCS during the periparturient period. In light of the concerns posed by Piccinini et al. (1999) and Schukken et al. (1999), recent work has concluded that SCS should be consider ed a heterogeneous trait with SCC of healthy cows separate from SCC of mastitic cows (Heringstad et al., 2006). This study found the heritability of SCS in healthy co ws to be 0.08, while the heritabil ity in mastitic cows equals 0.03. Structural traits of th e udder and productive life Udder conformational traits have also been used to help select for resistance to clinical mastitis infection. Udders with good attachment and cleft provide greater distance between the teat canal and the ground, or other potential fomites. Also, proper conformation can be associated with proper function and use of milk ing machines, which can also provide an avenue for microbial intramammary infiltration. PTAs fo r traits such as; udder cleft, udder depth, rear udder height, rear udder width, a nd fore udder attachment have b een found to be statistically significant predictors for clinical mastitis risk (Nash et al., 2000). The use of PTAs for SCC and udder conforma tional traits has provided some means to genetically select for disease resistance. However, these traits only associate with infections of the mammary gland, and for the case of selection fo r reduced SCS, we are still unsure if we are also unintentionally selecting for an unfavorable reduction in immune responsiveness. 34

PAGE 35

PTAs for productive life have also shown to be significant predictors of clinical mastitis infection. Productive life is a measure of the sta y ability or the ability of the daughters for a particular sire to resist being cu lled. Selection for this trait has a very crude application to disease resistance especially when you consider the arra y of potential reasons for being culled. However, productive life has been found to be significantly a ssociated with clinical mastitis (Nash et al., 2000). Disease Resistance Through Specific A ttributes of the Immune System Lymphocyte subsets Alterations in populations of T lymphocyte subsets have led to speculation about a potential role in predicting disease resistance. One study f ound a special correlation between features of T lymphocyte populations and the periparturient period. During this period of typical immune suppression, mammary gla nd secretions contained fewer numbers of T-lymphocytes and the ratio for subsets CD4:CD8+ was less than 1 (Park et al., 1992). Speci al attention was then placed on the ratio of T-lymphocyte subsets and their ability to predict disease. Additional studies have found that the CD4:CD8 ratio in ma mmary gland secretions was lowest in cows with Staphylococcus aureus mastitis (Park et al., 1993; Sordillo et al., 1991). Park et al. (1993) found that responsiveness to antigen by CD4+ T lymphocytes was incrementally reduced with increasing presence of CD8+ T lymphocytes. Park et al. (2004) found that mastitis susceptible cows had CD4:CD8 ratios of less than 1 in ma mmary gland secretions as well as peripheral blood. Major histocompatability complex Another arena for exploration involves the id entification of specific MHC haplotypes or gene alleles which are associated with favorab le measures of immune function. This could potentially serve as a genetic marker for disease resistance. The biological relevance stems from 35

PAGE 36

the previously discussed role of MHC in antigen presentation to T lymphocytes. The capacity to adequately present antigen is fundamental to mounting an effective immune response. Bovine MHC is also referred to as the bovine leukocyte antigen (BoL A) and is encoded by highly polymorphic genes. Several studies have found si gnificant differences in immune responsiveness and resistance to disease with alterations in class I and II MHC haplotypes and gene alleles (Rupp et al., 2007; Park et al., 2004; Aaerestrup et al ., 1995; Rupp and Didier 2003; Kelm et al., 1997). A recent study by Rupp et al (2007) looked at th e alleles for the DRB3.2 locus. This is the location encoding the MHC class II antig en binding site, making this region highly polymorphic. This study found several associations between different alleles and measures of immune function, disease resi stance, and performance. Broad-Based Immune Responsiveness Selection for improved immune responsiveness against disease has been studied in poultry (Soler et al., 2002; Heller et al., 1992; Kean et al., 1994), swine (Malla rd et al., 1998), sheep (Woolaston and Baker, 1996), mice (Biozzi et al., 1979) and Holstein cattl e (Mallard et al., 1997; Wagter et al., 2000; Hernandez et al., 2003). Selecting for increased immune responsiveness is to provide an indirect trait that potentially corre lates with broad-based disease resistance. Along with the previously mentioned mediators for immune suppression which affect immune responsiveness, there appears to be a significant genetic effect which determines the magnitude of a particular cows immune responsiveness. Significant variation exis ts in the ability of periparturient dairy cows to mount an immune response indicating that no t all cattle experience the same degree of immune suppression around calving (Mallard et al., 1997; Wagter et al., 2000). The concept of selecting for broad-based dise ase resistance as opposed to resistance to particular pathogens/diseases is a ppealing. It is true that selec tion against specific conditions will 36

PAGE 37

generally provide the quickest genetic progress for that particular condition. However, this often results in little or no genetic progress for other conditions. Also, due to the mechanics of the immune system, simply selecting for resistan ce to one condition without regard for other diseases, may introduce susceptibi lity to other conditions (Biozzi et al. 1979; de Vries, 1995; Rupp et al., 2007). The principles behind the decision to select for broad-based disease resistance are as follows: The cow relies on the immune system as the principle means to which pathogens are fought. Eradication of environmental infection-ca using organisms is impo ssible, so cows are going to be exposed to a wide variety of pa thogens. Through mutations, these organisms are capable of altering their virulence and defense mechanisms including resistance to antibiotics. Selection for increased immune responsiveness should reduce the dependency on vaccines and antibiotics. Increases in antibody t iters to vaccinations should resu lt in more efficient use of vaccine dosage and potentially a reduction in va ccine dosage (Wagter et al., 2000; Mallard et al. 1997). Selecting for increased immune responsiven ess should take steps toward addressing the concerns consumers have for excessive use of antibiotics and animal welfare issues. Through increased immune responsiveness and a reductio n in mastitis occurrence, SCC should also concurrently reduce (Mallard et al., 1997; Wagter et al., 2000) which is associated with increased cheese yield, shelf-life, and dairy product quality (National Mast itis Council, 1996). Differences between high and low immune responders Biozzi (1979) made some interesting discove ries while studying the differences between mice bred for high and low antibody responsiven ess. This study found that there was no difference in the amount of antibody released from individual plasma cells of the high and low responders. However, they found the high respon ders did multiply and differentiate at a significantly faster rate than lo w antibody responders. They found that antigen was catabolized at 37

PAGE 38

a significantly faster rate in the low responders when compared to high responders. The rationale explaining this finding is that slower catabolism of antigen is associated with slower antigen processing and prolonged antigen presentation. This extended antigen presentation is associated with greater lymphocyte s timulation and activation. Correlation with infectious disease risk Several studies have researched the corre lation between immune responsiveness and disease incidence. These studies can also be used to further unde rstand the individual aspects of the immune system and their particular role in the pr evention of a given condition. While breeding mice for high and low antibody responsiveness, Biozzi et al (1979) naturally found that his high antibody responder line was more resistant to extracellular pathogens. However, the high antibody responder s were more susceptib le to intracellular pathogen. This is believed to be due to the slow ed intracellular catabolism of antigen associated with the high antibody responders. When concerned with intracellular immunity, it is the speed at which the cell is able to break down the anti gen which positively reflects the potential of intracellular immunity. In swine, high immune responders as assesse d by measures for AMIR, CMIR, and innate immune function, were found to have considerab ly less peritonitis and pleuritis following a Mycoplasma hyorhinis infection (Mallard et al., 1998). However, it was also noted the high immune responders as assessed by estimated breeding values (EBV) for AMIR, CMIR, and innate immune function had more arthritis than the low responders. This is believed to be the result of selection for increas ed cell-mediated responsiveness and its association with the inflammatory response. Mastitis occurrence within antibody response categorization has been studied in Holstein dairy cows. Although not statistical ly significant, mastitis incidence for the high responders was 38

PAGE 39

lowest in 2 of the 3 study herds. In these two herds the high res ponders did not have an incidence of mastitis. There was also reason to question the va lidity of the herd having more mastitis in the high responders. In this herd, all cases of mastitis were in first parity heifers. The incidence of mastitis is generally higher for multiparous cows compared to primiparous. Relating to mastitis, two studies have also looked at the correlation between antib ody responsiveness to ovalbumin and antibody response to the E. coli J5 vaccination (Rhne Mrieux, Lenexa, KS) (Mallard et al., 1997; Wagter et al., 2000). In both instances they found that the correlation between antibody titers to the E. coli J5 vaccine and response to ovalbumin we re positive and significant. Wagter et al. (2000) reported a general correl ation of r = 0.56 (p < 0.0001). The E. coli J5 vaccination has been proven to be associated with re duced SCC, reduced time for clearance of E. coli in milk, and less milk production loss following intramammary challenge (Wilson et al., 2007). Because of this efficacy there is true biological relevance in selecting for increased immune responsiveness to ovalbumin. Mallard et al. (1997) also compared the incidence of disease between high and low AMIR cows. In this study their main focus was the effect of cortisol, GH, and IGF-1 on antibody response profiles. However, they also looked at disease occurren ce over the 3 categorizations for antibody response. They found that disease inciden ce was smallest for high responders (group 1). This finding also followed the pattern of high
PAGE 40

energy balance and the development of ketosis-fa tty liver complex. Another explanation for this association is provided through L acetera et al (2005) and Suri yasasathaporn et al. (2000). Because the presence of NEFA and ketone bodies is closely linked with negative energy balance and energy-related metabolic cond itions (ketosis); it is possible that the presence of NEFA and ketone bodies serve to suppress immune function wh ich increases the risk for infectious disease (Figure. 2-2). The discovery of elevated NEFA concentra tions in response to inflammatory agents (Steiger et al, 1999; Kushibiki et al., 2003) has sp arked research to determ ine if early lactation mastitis can cause ketosis-fatty liv er complex in dairy cows (Waldron et al, 2006). However, this work concluded that the results do not indicate mastitis to be ca usal for energy-related metabolic disorders. Instead, they did sugge st the possibility for a potential protective effect by the immune system on metabolism during early mammary infection. In some instances, research studying immune responsiveness reveals a relationship with a disease previously not underst ood to be correlated with th e immune system. Schukken et al. (1988) studied the relationship between an inf ectious disease (mastitis ) and retained fetal membranes. This study revealed th at cows having a retained fetal membrane were more likely to have a case of mastitis shortly after calving. Retained fetal memb rane is a condition which is now understood to be the result of a faulty immune response. The body must be able to identify the placenta as foreign and mount an appropria te immune response against the placenta soon after parturition. Research has demonstrated that neutrophil chemotaxis as well as killing ability is impaired in both pre and postpartum cows that will/had a retained placenta (Kimura et al., 2002). Because of this, it is possi ble that the prepartum use of de xamethasone and its association with increased risk of retained fetal me mbranes might be partially explained by the 40

PAGE 41

immunosuppressive action of dexamethasone (B urton et al., 1995; Burton and Kehrli, 1995, 1996). We do not exactly know the precise magnitude or the role the immune system components have on many metabolic activities. Th e complexity of the immune system in vivo often times provide unpredicted results during research study. The possibility of the immune system serving a substantial role in body meta bolism is considered plausible. Correlation with milk production Selection for improved immune responsiven ess should yield a tr end toward reduced incidence of disease. Although this strategy pr ovides many benefits for animal welfare, the extent to which this philosophy is adopted will be strongly influenced by economics. A decrease in disease occurrence will provide obvious fina ncial benefits through a reduction in treatment costs and also a reduction in milk loss from diseas e or treatment of disease. However, there is considerable speculation that a ny economic benefits observed may be counteracted by decreased performance. In other species, the energy and nutritive demands of a superior immune system have been shown to reduce performance. The en ergy and nutrients requir ed for maintenance and activation of a responsive immune system could otherwise be used for other phenotypic traits (Klasing et al., 1987, 1998; Soler et al., 2003). C oncerning dairy cows, the positive correlation between milk production and c linical mastitis (Emanuelson et al., 1988; Harmon, 1994) provides some evidence that financial benefits may be off-set by a reduction in milk yield. Along with the previously cited avian study, additional research has been conducted to study the ramifications of superi or immune responsiveness with pe rformance. In swine, Mallard et al. (1998) found that growth performance was consistently significantly greater for high immune responders when compared to both th e low immune responders and the control group. In this research, pigs were selectively mate d for high and low immune responsiveness over the 41

PAGE 42

course of eight generations using measures fo r innate and adaptive im munity (AMIR & CMIR). The trend of improved rate of gain for the high immune responder line compared to the rest was first identified in generation 0 and continued through generation 7. In Holstein cows, Wagter et al. (2003) looked at milk pr oduction within high, medium, and low AMIR categorizations. Because parity significantly contributed to variation in 305 day milk yield, they conducted their analys is within parity. For first pa rity cows the low responders produced significantly more milk than the medium and high responders. However, for second parity cows, there was no statistical differen ce between the high and low as well as high and medium response groups. For third parity cows, the high responders produ ced significantly more milk than the low and medium response groups (Wagter et al., 2003). Another study showed that dairy cows genetically selected for high milk yield over seven generations did not produce unfavorable measures for innate and adaptive immune function when compared to cows selected for average milk production (Detilleux et al., 1995). The findings of these studies indicate that although there is an association between selection for increased milk yield and increased risk of clinical mastitis, selection for improved immune responsiveness should not predispose cows to reduced milk yield. It also indicates adequate variation in immune responsiveness among sires with high PTA for milk yield to support selection for both increased milk yield as well as increased immune responsiveness Heritability of measures for immune responsiveness The level of heritability expected during selec tion for measures of immune responsiveness indicate the rate in which ge netic improvement can be made If the proposed measure for immune responsiveness has low leve ls of heritability, the effectiven ess of an indirect trait would be very limited. 42

PAGE 43

The heritability of measures for AMIR and CMIR has been tabulated in pigs (Mallard et al., 1998). These calculations were configured on over 1200 observations through 8 generations of selection for high and low immune responsivene ss. Heritability of AMIR was estimated to be 0.268 while CMIR heritability estimates equale d 0.163. The heritability of AMIR as measured by antibody responsiveness to ovalbumin has also been tabulated for Holstein periparturient cows (Wagter et al., 2000). These estimates of heritability ranged 0.32 to 0.64 dependent on week relative to calving. The lower value for this range (0.32) coincided with the heritability of antibody response to OVA measured at calving. Ha ving the lowest estimate occur around calving may be explained by the various stress factors and mediators for immune suppression which are occurring during this period. Categorizing AMIR and CMIR Introduction Studies which associate immune responsiven ess with disease risk usually involve correlative studies associating an iddices of immunity with risk for disease. The method chosen to stimulate the immune system by which to meas ure immune response is of primary importance. The technique used should represent a subjects ove rall potential to resist disease. Methods used to categorize immune responsiveness for the pur pose of selection for disease resistance should have a very general approach wh ich gives consideration to all aspects of immune function. When considering the separate roles in the immune system for AMIR and CMIR as well as the previous work which identifies a potential inverse rela tionship between the two (Biozzi et al., 1979; de Vries, 1995, Rupp et al. 2007), inclusion of bot h branches becomes essential. Due to the potential for an inverse relationship, a failure to cons ider both branches upon genetic selection may result in increased susceptibilit y in the branch not considered. 43

PAGE 44

Antibody-mediated immune response There are two different techniques that have previously been reported to categorize AMIR in Holstein cows. In both instances, cows were categorized during the peri parturient period with ovalbumin as the test antigen to elicit the humoral response. The study cows were injected with the antigen at week -8, week -3, and week 0 re lative to calving. Blood samples to measure the ensuing antibody response were co llected on week -8, week -3, w eek 0, week +3, and week +6. Antibody response was detected by ELISA and the re sulting OD values were used to categorize AMIR function (Wagter et al., 2000; Mallard et al., 1997). However, these studies differ in how AMIR categorizations were extrapolated from the OD values. Mallard et al. (1997) found that the sample si ze of 33 cows and heif ers partitioned into three groups. All cows responded well to the initia l antigen exposure at week -8, however it was the responses to the subsequent antigen exposures which determined their categorization. High responders had an above average response to a ll three antigen administrations. The medium responders responded well to both of the prep artum antigen administrations, yet responded poorly postpartum to the week 0 injection. The low responders mounted poor pre and postpartum responses to the week -3 and w eek 0 administrations of antigen. Wagter et al., (2000) categorized AMIR in 136 cows and heifers. In this study an index was generated which used the change in OD value over the intervals between the antigen injection/blood collection periods. This resultant value was then used to categorize cows as high, med, or low antibody responders. The formula fo r this index is as follows (Eq. 2-1): y total = I 1 + 1 I 2 + 2 I 3 + I 4 (2-1) Where: y total = total antibody I 1 = change in optical density (OD) between week -8 and week -3 44

PAGE 45

I 2 = change in OD between week -3 and week 0 I 3 = change in OD between week 0 and week +3 I 4 = change in OD between week +3 and week +6 1 & 2 = either 1.0 or 1.5 The coefficient 1 takes on a value of 1.0 if I 2 is positive, representing a positive antibody response around parturition. If I 2 is negative, representing a lack of response around calving, 1 takes on a value of 1.5 which serves to magnify or inflate the negative response. This rationale was also applied to 2 concerning I 3 To categorize immune responders, the mean a nd standard deviation for all the generated y total values was configured. Cows were classified as high responders if they had an y total value greater than the mean plus one standard deviation. A medium responder had an y total value within the mean plus one standard deviation and th e mean minus one standard deviation. A low responder had an y total value below the mean minus one standard deviation. Cell-mediated immune response Methods employed to categorize cell mediated immune function have all utilized delayedtype hypersensitivity (Hernandez et al., 2005, 2003; Mallard et al., 1998). Because this reaction is mediated by T H 1 cells it is an indicator of intracellula r immunity. This is the only in vivo method known which enables categorization of CMIR. Quantification and therefore categorization of the CMIR comes from measurements taken at the antigen injection site. As previously stated, DTH reactions require previous exposure to the antigen intended to elicit the DTH response. To in itiate the DTH response, the antigen is injected intradermally and double skin-fol d measurements are taken to serve as a baseline. All measurements should be taken w ith three repetitions while usi ng the average of the three for analysis. After 24 to 48 hours, measurements of the injection site are once again taken which 45

PAGE 46

should reflect the palpable lump indicative of a DTH response. The degree to which the measurements increased can be used as an indi cator of cell mediated immune responsiveness. 46

PAGE 47

Figure 2-1. General depiction of a primary and secondary response to an tigen x. There is substantial variation in the kinetics of an antibody response due to antigen type, administration route, presence of adj uvant, and species exposed to antigen. 47

PAGE 48

T-LYMPHOCYTES CD4+ CD8+ TH1 TH2 Secretes IL-10 Associated with humoral immunity IL-10 blocks cytokine secretion of TH2 cell Associated with cellular immunity Figure 2-2. Depiction of the inve rse relationship between AMIR a nd CMIR as a result of IL-10. 48

PAGE 49

Figure 2-3. Plasma NEFA in thin, medium a nd overconditioned cows during the peripartum period. Values with different letters differ significantly (P < 0.01). Values reported are LS means SEM. Reprinted with perm ission from: Lacetera, N., D. Scalia, U. Bernabucci, B. Ronchi, D. Pirazzi, and A. Nardone. 2005. Lymphocyte Functions in Overconditioned Cows Around Parturiti on. J. Dairy Sci. 88:2010-2016. Figure 2, page 2012. 49

PAGE 50

Figure 2-4. Effects of NEFA on IgM secretion in peripheral blood mononuclear cells stimulated with pokeweed mitogen. Values reported are mean SEM. Columns with different letters differ significantly (P < 0.05). Ns = Not stimulated. Reprinted with permission from: Lacetera, N., D. Scalia, O. Franci, U. Bernabucci, B. Ronchi, and A. Nardone. 2004. Short Comm: Effects of Nonesterifie d Fatty Acids on Lymphocyte Function in Dairy Heifers. J. Dairy Sci. 87:1012-1014. Figure 2, page 1014. 50

PAGE 51

Figure 2-5. Effects of NEFA on interferonsecretion in periphera l blood mononuclear cells stimulated with concanavalin A. Values reported are mean SEM. Columns with different letters differ significantly (P < 0.05). Ns: not stimulated; ND: not detectable. Reprinted with permission from: Lacetera, N., D. Scalia, O. Franci, U. Bernabucci, B. Ronchi, and A. Nardone. 2004. Short Comm: Ef fects of Nonesterified Fatty Acids on Lymphocyte Function in Dairy Heifers. J. Dairy Sci. 87:1012-1014. Figure 3, page 1014. 51

PAGE 52

Negative Energy Balance NEFA Ketosis Immune Suppression Metritis Figure 2-6. Flow chart describi ng potential relationship between energy-related metabolic condition (ketosis) and infectious condition (metritis). 52

PAGE 53

CHAPTER 3 METHODS UTILIZED FOR THE CATEGORIZ ATION OF AMIR AFTER GENERATING ELISA OPTICAL DENSITY VALUES Introduction Enzyme linked immunosorbent assay (ELISA) is commonly used to detect the presence of a particular antigen, or specifi c antibody in body fluids or tiss ues (Wild, 2001). Previous work used the generated optical density (OD) values from an indirect ELISA as a means to quantify antibody mediated immune responsiveness (AMIR) for individual dairy cows (Mallard et al., 1997; Wagter et al., 2000; Hernandez et al., 2003). In our research we utilized these previously described methods. However, after examination of the results, it was decided that further adjustments to the OD values were required. We al so found that not all cows were nave to the test antigen. This chapter discu sses the identification of cows previously exposed to the test antigen, and describes the reasoning behind the adjustments made to the OD values, and the methodology employed to make these adjustments. This chapter also discusses the alternate methods attempted to categorize AMIR and the justification for the method chosen. Study Population In total, 875 Holstein cows/heifers were enrolled into th e study population at approximately 8 weeks (wk-8) prior to expected calving. In cows, this wa s the initiation of the dry period. Animals were enrolled if the expected dry period length was less than 90 days, if reconfirmed pregnant at enroll ment and also if found in good health with no obvious signs of disease. All test animals were from a single herd in north central Florida which maintains exceptional record keeping. All cows and he ifers were enrolled between September 9 th and December 31 st 2004, calved between October 25 th 2004 and March 12 th 2005, and exited between November 9 th 2004 and March 28 th 2005. All cows and heifers received a routine dry off, prefresh, and fresh cow protocol. 53

PAGE 54

Exclusion Criteria A reduction in the study population was necessary to maintain the integrity of the study. Of the 875 cows and heifers, 13 were removed due to missing samples at one of the blood collection periods. An additional 88 were removed either b ecause they were not nave to test antigen or because they did not meet the interval or dry pe riod length exclusionary criteria which will be discussed later in the chapter. A total sample size of 774 with 433 cows and 341 heifers upon enrollment were analyzed. Because measurem ent of AMIR concluded 2 weeks after calving, heifers will be referred to as primiparous cows and the cows at enrollment will be referred to as multiparous cows (primiparous = first lactation; or multiparous = second or greater lactation). Removal of Cows Previous ly Exposed to Antigen Ovalbumin (OVA) was chosen as the antigen to stimulate AMIR. The rationale behind this decision came from the ability of OVA to stim ulate a humoral response, the low likelihood of prior exposure, and the previously successful use of this antigen as a tool to categorize AMIR in dairy cows (Mallard et al., 1997; Wagter et al., 2 000; Hernandez et al., 2003) For this trial, the periods of antigen exposure and blood collecti on occurred at enrollment (wk-8), entry into springer pen (wk-3), and calving (wk0). An additional blood sample was collected 2 weeks after calving (wk+2) (Figure 3-1). Bl ood samples were collected to determine antibody response to antigen. The OD value for wk-8 (OD -8 ) was to serve as a covariat e for the antibody response to OVA. Equal treatment for the measurement of AMIR requires that all anim als are nave to test antigen at this point. This is critically impor tant due to the differences between primary and secondary antibody responses (Figure 2-1). If a po rtion of cows mount a secondary response due to previous exposure and are be ing compared with cows mounting a primary response, this introduces unequal treatment into your study population. 54

PAGE 55

Methodology and Results After review of the OD -8 values it was apparent some cows were most likely not nave or had high nonspecific reactivity to the test antigen, so a na tural cut-off point of an OD -8 = 0.455 was used. This point corresponded with 1 sta ndard deviation above the mean for all OD -8 values. As a result of this analysis, 38 animals (23 primiparous and 15 multiparous) were termed not nave to test antigen and were removed from the study. Effect of Parity on Antibody Response to Ovalbumin A repeated measures analysis using OD -8 as a covariate with the mixed procedure of SAS revealed that multiparous cows responded significantly higher than primiparous cows at every antibody response measurement week (p<0.0001). Due to this effect, adjustments to OD values as well as AMIR categorizations were all made with respect to parity (F igure 3-2). Explanations for this finding may include higher levels of stress and therefore immune suppression for younger cows experiencing lactation and parturiti on for the first time compared to multiparous cows. Another explanation involves the presence of a more extens ive antibody repertoire in older cows. This could simply be due to the effect of time, allowing greater exposure to a broader array of pathogens. Another explanation Interval Variation Adjustment Defining Intervals: Antigen was injected and blood was collected at specified time frames (Figure 3-1). Variation in the number of days be tween these points of blood collection/antigen injection has a strong influence on the measured antibody response. These in terval lengths reflect the duration between previous antigen exposure and measurement of antibody response; which is relevant due to the kinetics of an antibody re sponse profile (Figure 2-1). In every antibody response there is a lag phase followed by a peri od of increasing antibody concentration up to a peak response which is followed by a steady de cline in antibody con centration. Introducing 55

PAGE 56

variation in the number of days between points of blood collection/antigen injection generates inconsistency in the phase of antibody response profile wher e antibody response was measured. For this research, interval 1 (Int 1 ) was defined as the duration in days between sampling periods identified as wk-8 and wk-3. Interval 2 (Int 2 ) was defined as the duration in days between sampling periods, wk-3 and wk0. Interval 3 (Int 3 ) was defined as the duration in days between sampling periods, wk0 and wk+2. Strictly for the purpose of monitoring the ch ange in body condition score (BCS), additional intervals were identified. Interval 4 (Int 4 ) was defined as period between wk-3 and wk0 which is an indication of the change in BCS ove r the transition period. Interval 5 (Int 5 ) was defined as period between wk-8 and w k0. Also, interval 6 (Int 6 ) was defined as period between wk-8 and wk+2. Interval Exclusionary Criteria: One exclusion criteria at study assignment was based on dry period length (Int 1 plus Int 2 ). If this period was greater th an 90 days, they were removed from the study. Although this re striction consists of Int 1 plus Int 2 it does not adequately put restrictions on individual lengths for either Int 1 or Int 2 However, it can eliminate cows with various metabolic problems from skewing results. A minimum interval length of 12 days was set for both Int 1 and Int 2 This does work to restrict specific interval lengths to a minimum, but it still leaves room for substantial variation to occur. Applying these exclusionary criteria an additional 50 animals were eliminated yielding a study population of 774. Due to the resources available and nature of a large-scale dairy farm, considerable variation in the lengths for Int 1 and Int 2 did occur. To correct the OD values for this effect, adjustments were made to OD values based on th e duration of the interval leading up to the 56

PAGE 57

respective OD value. For example, adjustments to the wk-3 OD (OD -3 ) were made based on the length of Int 1 for a respective cow. With Int 3 concluding at a sampling period occurri ng after calving (wk+2) with animals managed in the milking herd, this duration wa s under our control which greatly reduced the variation in Int 3 lengths. No adjustments were required for the OD -8 values because this was the initial point of exposure. Methods for OD Adjustment Statistical analysis to determine the effect of interval length on OD value was performed using the GLM procedure of SAS. The OD respons e to be adjusted was the outcome variable while potential explanatory variables for each cow included: Int Y = interval length in days for interval y BCS X = categorization of body condition score for sampling week x BCS Y = change in body condition score over interval y OD X = optical density value at sampling week x Dex = binary effect, whether cow recei ved dexamethasone prior to calving Sick = binary effect, incidence of either medium or severe case of mastitis, metritis, ketosis, or displaced abomasum within 16 DI M. Only used for analysis to correct OD +2 for reasons to be discussed later. Explanatory variables remained in the model if the effect showed a tendency (P < 0.10) to predict the desired OD value. With = 0.05, if statistical significance was found with the main effect (Int y ), the resultant parameter estimate was applied to Equation 3-1 to correct the corresponding OD values. If a given OD value was adjusted based on interval length, the model predicting the OD value for the subsequent samp ling period would have the adjusted OD from the previous sampling week. AOD ZPX = OD ZPX + PE YP (Mint YP Int ZPY ) (3-1) 57

PAGE 58

Where: AOD ZPX = the adjusted OD value for cow z in parity p for week x OD ZPX = OD value for cow z in parity p for week x PE YP = parameter estimate for the effect interval y has on OD x values for parity p. Also, interval y must always immediately precede week x Mint YP = median number of days for interval y for parity p Int ZPY = the actual interval length for cow z in parity p for interval y Results OD -3 The length in days for Int 1 was under considerable varia tion ranging from 12-55 days in primiparous cows (Figure 3-3).View of a scatter plot depicting the relationship between Int 1 and OD -3 values do not reveal any obvious pattern. Model effects were: Int 1 OD -8 BCS -8 and BCS 1 For primiparous cows, Int 1 was not found to be a significant predictor of OD -3 values (p = 0.34, = -0.0015). Because of this, no adjustments were made to primiparous cow OD -3 values. The Int 1 variation for multiparous cows ranged from 21-71 days (Figure 3-4). Inspection of the scatter plot depicti ng this relationship revealed an obvious association. The model effects were Int 1 and OD -8 There was a significant linear effect between Int 1 and OD -3 values (p < 0.0001, = -0.0081). This parameter estimate represents the slope of the fitted line for the linear model. The median Int 1 in multiparous cows equaled 36. This parameter estimate and median was then used to generate the adjusted OD -3 values for each respective multiparous cow (Eq. 31). Results OD 0 The range in days for Int 2 also had considerable varia tion. For primiparous cows, this range extended from 12 to 45 days (Figure 3-5). Model effects were Int 2 OD -3 BCS 2 and Dex. 58

PAGE 59

The variation in length of Int 2 had a significant effect on OD 0 values (p < 0.0001, = 0.0246).The median Int 2 duration of days was 21. This parameter estimate and median was then used on primiparous cows to adjust the OD 0 values (Eq. 3-1). For multiparous cows, the length of Int 2 ranged from 12-44 days (Figure 3-6). The scatter plot depiction of the relationship clearly indicates an association. This plot also clearly represents a characteristic secondary immune response w ith a longer duration of peak response. Model effects were Int 2 and OD -3 The effect of Int 2 length significantly effected OD 0 values (p < 0.0001, = -0.0211). The calculated median was 22 da ys. This parameter estimate and median was used to adjust OD 0 values for multiparous cows (Eq. 3-1). Results OD +2 For primiparous cows the variation in Int 3 ranged 12-21 days (Figure 3-7). For this analysis the model effects included Int 3 OD 0 and BCS 4 The length of Int 3 was a significant predictor for OD +2 in primiparous cows (p < 0.0032, = -0.0298). The calculated median number of days was 16. The parameter estimate and median was used to adjust OD +2 values in primiparous cows (Eq. 3-1). For multiparous cows the range for Int 3 was 12-20 days (Figure 3-8). This model included Dex,, OD 0 and Int 3 However, Int 3 .was not a significant predictor for OD +2 in multiparous cows (p = 0.21, = -0.0104). No adjustments were made to OD +2 in multiparous cows. An additional GLM model was run after the di scovery that parity was not a significant predictor for OD +2 as determined by a linear regression m odel (Eq. 3-2) which is later discussed. This new model analyzed associations with OD +2 irrespective of parit y. The remaining model effects were Sick, OD 0 and, Int 3 In this instance the model revealed Int 3 is a significant predictor for OD +2 (p < 0.0044, = -0.0182). After consider ation, this will not be used to adjust OD +2 values due to the added power of a repeated measure analysis in this application. 59

PAGE 60

Analysis of Classification Methods As previously discussed in chapter 2, there are two different techni ques that have been previously reported to categorize AMIR in peripa rturient Holstein cows. In both instances, cows were categorized with ovalbumin as the test antigen to elicit the humoral response. The study cows were injected with the antigen at wk8, wk-3, and wk0 relative to calving. The ensuing antibody response was detected by ELISA and the resulting OD values were used to categorize AMIR function (Wagter et al., 2000; Mallard et al., 1997). The response to antigen introduced at wk0 was detected on wk+3 and wk+6 postpartum. Use of interval changes in OD: The method for the actual AMIR categorization in the publication by Wagter et al. (2000) used an index based on the change in OD response over the intervals between blood sampling/ antigen injection periods (Eq. 2-1). As pr eviously discussed in chapter 2, this index we ights those intervals ( 1 & 2 ) around calving if they show a decline in OD value, which represents a decline in antibody concentration. y total = I 1 + 1 I 2 + 2 I 3 + I 4 (2-1) Where: y total = total antibody I 1 = change in optical density (OD) between week -8 and week -3 I 2 = change in OD between week -3 and week 0 I 3 = change in OD between week 0 and week +3 I 4 = change in OD between week +3 and week +6 1 & 2 = either 1.0 or 1.5 In this study, inspection of th e OD values for the respective weeks revealed a point where antibody response was at an appare nt maximum. If there was an improvement after subsequent antigen injection, these maximal poi nts in antibody concentration on ly slightly improved. If this peak response was achieve d early in the study (OD -3 ), which would indicate a high prepartum responder; there would be little if any room for an added re sponse. As a result an index 60

PAGE 61

concerned with weighting the change in OD dur ing the intervals adjacent to calving (Eq. 1-1), may negatively impact the categorization of hi gh immune responders who reached this peak response early because there was little or no ro om to further respond. Speculation for causation of this finding most likely involves the natural function of feedback inhibition. Early postpartum measures of immune function: The period immediately following parturition is a common occasion for increased in cidence of disease. As a result it may be hypothesized that substantial sickness could cont ribute to immune suppression. This effect of substantial sickness could be a confounding va riable for antibody responsiveness to OVA detected early postpartum. This would make it di fficult to study the effect measures of immune responsiveness have on disease ri sk if the association could al so be in the opposing direction. Objectives: The 2 objectives were as follows: 1) To analyze the potential for a maximal antibody response and its possible effect on an tibody response categorizat ion. 2) To study the potential for an effect of sickness on early postp artum measures of antibody responsiveness. This is performed by analyzing the relationship between OD +2 values and substantial sickness. Methods Maximal response: For this analysis, the relati onship between OD value and the subsequent interval change in OD (I) was studi ed. This was performed for the relationship between OD -3 and I 2 and also between OD 0 and I 3 The total sample size for this analysis was 774 cows. Cows were arranged according to their OD -3 value and grouped into one of 14 groups with 55 cows per group except for group 14 which had 59 cows. The top 55 cows for OD -3 became group 1 while the bottom 59 cows for OD -3 became group 14. This same process was also performed based on OD 0 values. A one-tailed two sample t-test was used to determine if the OD value group had a significant effect on the subsequent interv al change in OD. For this analysis, I 2 and I 3 had a 61

PAGE 62

normal distribution, however in both cases, onl y group 1 cows could be compared to group 2 cows due to statistical differences in variance between group 1 a nd the rest of the groups. The null hypothesis is: H o : Group 1 I 2 Group 2 I 2 0. Rejecting the null with statistical significance means that Group 1 I 2 is significantly smaller than Group 2 I 2 A linear regression model using the REG procedure of SAS was also used to see if OD -3 had a significant effect on I 2 or if OD 0 had a significant effect on I 3 Effect of sickness on OD +2 : For the statistical analysis a linear regression model with the REG procedure of SAS. OD +2 served as the outcome variable while potential explanatory variables included: BCS x = categorization of body condition score for sampling week x BCS y = change in body condition score over interval y OD 0 = optical density value at wk0 Dex = binary effect, whether cow recei ved dexamethasone prior to calving Sick = binary effect, incidence of either medium or severe case of mastitis, metritis, ketosis, or displaced abomasum within 16 DIM. Parity = binary effect, either primiparous or multiparous Explanatory variables remained in the model if the effect showed a tendency (P < 0.10) to predict the dependant OD value. For statistical significance, = 0.05. The Corr procedure of SAS was also used to test the correlation between OD 0 and OD +2 in healthy cows as well as those classified as sick within 16 DIM. Results and Discussion Maximal response: After sorting the OD -3 values from greatest to least, it was discovered that of the top 20 cows for OD -3 value, 15 had a smaller OD 0 value (75%), yielding a negative interval 2 change in OD. In th ese instances, 15 of the top 20 OD -3 responders would have an 62

PAGE 63

amplified ( = 1.5) negative I 2 value applied to their AMIR index if using an index which weights interval changes in OD. After grouping the OD -3 values from 1 to 14, with group 1 being the top 55 OD -3 values and group 14 the bottom 59 OD -3 values; this revealed a significantly smaller change in OD over interval 2 for group 1 cows compared to group 2 cows (p < 0.0001) (Figure 3-9). This value actually averaged below 0 (-0.05) for these top 55 OD -3 responders. Linear regression also revealed OD -3 is a significant predictor for I 2 (p < 0.0001; = -0.28). A negative indicates that as OD -3 values increase, I 2 values decrease. Repeating the same process by ranking cows based on OD 0 values in order to compare interval 3 change in OD, revealed similar results. Due to missing wk+2 blood samples the sample size for this analysis was 754. As a result group 14 had 39 cows while group 1 13 had 55. Of the top 20 cows for OD 0 values, 15 had negative interval 3 changes in OD (75%). Also, group 1 cows based on OD 0 values had a significantly smaller interval 3 change in OD compared to group 2 cows (p < 0.0001) (Figure 3-10). Linear regression also revealed OD 0 is a significant predictor for I 3 (p < 0.0001; = -0.50). A negative indicates that as OD 0 values increase, I 3 values decrease. In these instances of high OD values follo wed by a subsequent negative interval, the interval is not negative due to a poor subs equent OD value. Of the top 20 cows for OD -3 18 still had OD 0 values above the third quartile for the popul ation and the other 2 were still above the median for the population. For the top 20 cows for OD 0 all 20 still had OD +2 values above the third quartile for the population. The negative interval was simply the result of an inability to respond further. As a result, additional indexe s were generated and analyzed based on their correlation with disease incidence. This finding is likely due to feedback inhibition due to the 63

PAGE 64

difficulty in boosting a subject that already has a high anti body concentration. Although OD adjustments were made to accommodate for interv al length variation; a cow which experiences a secondary exposure during peak response from a previous exposure to th e specific antigen will not be boosted to the degree a cow is that rece ived the antigen after peak response due to feedback inhibition. Effect of sickness on OD+2: For this analysis, there were a total of 754 cows, 40 of these cows were classified as Sick as defined with in 16 DIM. Parity was not a significant predictor of OD +2 (p = 0.54); as a result, it was not included in the model and all cows were considered together (Eq. 3-2). OD +2 = Sick + Dex + OD 0 (3-2) The analysis revealed that sickness, as prev iously defined, was a si gnificant predictor of OD +2 (p = 0.0156). The difficulty in this analysis is proving the direction of the association. Did the occurrence of sickne ss within 16 DIM cause a suppression in immune responsiveness; or did inferior immune responsiveness cause sickness within 16 DIM. Because OD 0 occurs prior to the incidence of disease, and this was a significant predictor of OD +2 (p < 0.0001), you can be fairly certain the incidence of sickness had an effect on immune responsiveness. Correlation analysis was employed to study the relationship between OD 0 and OD +2 at fixed levels of sickness. Among cows considered healthy within 16 DIM, correlation analysis revealed OD 0 is positively correlated with OD +2 (r 2 = 0.64, p < 0.0001). Furthermore, within sick cows, the correlation between OD 0 and OD +2 has an even greater significantly positive correlation (r 2 = 0.73, p < 0.0001). For further understanding, an additional li near regression model was constructed. The rationale behind this analysis comes from OD 0 and OD +2 being positively correlated in sick as 64

PAGE 65

well as healthy cows. If it was the level of i mmune response that caused susceptibility to sickness, the strong positive correlation should allow OD 0 to be able to se rve as the outcome variable. It is known that OD 0 reflects an immune response occurring prior to the high risk period for sickness. For this analysis, parity was significant (p < 0.0001). As a result, this analysis was performed at fixed levels of parity. For the 333 primiparous cows the independent variables remaining in the model were, Sick, and OD +2 This analysis reveals sickne ss is not a predictor for OD 0 (p < 0.387). For the 422 multiparous cows the independent variables remaining in the model were Sick, and OD +2 This analysis also reveals sickness is not a predictor for OD 0 (p < 0.592). For further evidence, it will be discussed later how models which do not in clude the +2wk antibody response data tend to have a stronger association with disease risk. Alternate Index Methods for AMIR Categorization Although only similar to the AMIR index (Eq. 11) in Wagter et al. (2000), a comparable index was generated (Eq. 3-3). This was chosen based on the util ity of the previous index. In similar fashion, another index (Eq. 3-4) was derive d due to the previously described effect of early postpartum sickness on immune responsiveness. y total = I 1 + 1 I 2 + 2 I 3 (3-3) y total = I 1 + 1 I 2 (3-4) Where: y total = total antibody I 1 = change in optical density (OD) between week -8 and week -3 I 2 = change in OD between week -3 and week 0 I 3 = change in OD between week 0 and week +2 1 & 2 = either 1.0 or 1.5 65

PAGE 66

The rationale behind two additional indexes is based on the assumption that a favorable AMIR will have greater correla tion with the actual magnitude of the antibody concentration rather than changes in antibody c oncentration over inte rvals. In the case of a maximal antibody response, an index should reflect a cows ability to maintain a high concentration of specific antibody. This should all be accomplished while also using measures which give special attention to antibody respons es occurring peripartum. The first index (Eq. 3-5) includes the postpartum OD +2 while the second (Eq. 3-6) does not. In these indexes, the direct magnitudes of the OD values are considered. However, in the case of OD 0 and OD +2 they are still weighted, but this can be positively or negatively and only in proportion to the level of increase or decrease for I 2 and I 3 If I 2 is slightly negative, then OD 0 is multiplied by a number slightly under 1 yielding a smaller value. If I 2 is slightly positive, then OD 0 is multiplied by a number slightly over 1, yielding a larger value. y total = OD -3 + OD 0 *(1 + I 2 ) + OD +2 (1 + I 3 ) (3-5) y total = OD -3 + OD 0 *(1 + I 2 ) (3-6) Where: y total = total antibody OD -3 = optical density value at wk-3 OD 0 = optical density value at wk0 I 2 = change in OD between week -3 and week 0 I 3 = change in OD between week 0 and week +3 For each index (Eq. 3-3, Eq. 3-4, Eq. 3-5, Eq. 3-6), 2 different appro aches were taken to extrapolate AMIR categoriza tions from the generated y total values. In each approach the AMIR categorization was determined within parity due to the effect of parity on antibody responsiveness to OVA (Figure 3-2). The first approach involved calculating the mean and 66

PAGE 67

standard deviation of all y total values (Figure 3-11) (Wagter et al., 2000, Hernandez et al. 2003). This was performed separate for multiparous a nd primiparous cows. High responders were those cows with y total values greater than the mean plus one standard deviation. Low responders were those with y total values less than the mean minus one standard deviat ion. Medium responders had y total values within one standard deviation. If dealing with a normal distribution of data, roughly 68% of the data will fall within 1 stan dard deviation of the mean. This leaves roughly 16% for high AMIR responders and 16% for low AMIR responders. Because this method only categorizes the extr eme 32 % of the population into high or low immune responders, calculations of quartiles were used. This method allowed us to set the top 25% of data as high AMIR, while the bottom 25% as low AMIR. The resulting middle 50% was classified as medium responders. In this case, the extreme 50% of the population is categorized as high or low responders. The configuration of the quartiles was also calc ulated within parity. With 4 different possible equations (Eq. 33, 3-4, 3-5, 3-6) to generate antibody y total values and 2 different methods to extrapolate AMIR classifications from each equation, an analysis of each possibility is required to determine which me thod is chosen. For this determination a raw association between each method and incidence of disease was used (Table 3-1). Results The worst association with disease is clea rly Equation 3-5 using the SD classification method (Table 3-1). It appears E quation 3-3 has a closer associati on with resistance to mastitis and retained fetal membrane. However, Equa tion 3-4 and Equation 3-6 are more closely associated with resistance to metabolic conditions. These two equations do not include OD +2 data from early postpartum. Use of Equation 3-3, Eq uation 3-4, and Equation 3-6 appear to be the best options. Use of standard deviation ve rsus use of quartiles appears negligible. 67

PAGE 68

Using incidence of mastitis for an entire lact ation as a binary outcome variable, logistic regression using the LOGISTIC Procedure of SA S was used to identify the equation which can be best used to predict susceptibility to mastitis Independent variables remained in the model if they showed a tendency (p < 0.10) to predict mastitis inciden ce. The potential model effects were: BCS x = categorization of body condition score for sampling week x BCS y = change in body condition score over interval y Dex = binary effect, whether cow recei ved dexamethasone prior to calving Parity = binary effect, either primiparous or multiparous SCCavg = the average SCC collected monthly for the first 10 months ARC = antibody response categorization using e ither the standard deviation or quartile method for Equation 3-3, Equation 3-4, or Equation 3-6. The effects which remained in the model were: SCCavg, BCS 5 Parity, and ARC. Equation 3-3 was not a significant predictor for ma stitis using either the standard devation (p = 0.726) or the quartile method (p = 0.701). Equation 3-4 showed a tendency to predict mastitis incidence using the standard deviation (p = 0.116) and quartile method (p = 0.089). Equation 3-5 was not a significant predictor for mastitis using either method (p = 0.816, p = 0.893). Using Equation 3-6, the standard deviation method was not a significant pr edictor (p = 0.3646), however, using the quartile method, Equation 36 was a significant predictor for mastitis incidence (p= 0.028). The two equations revealing any ability to predict mastitis incidence did not include early postpartum (wk+2) measurements for antibody respons e. These methods appear to have a closer association with susceptibility to mastitis. E quation 3-6 was the only equa tion with a significant ability to predict mastitis. This is also the onl y equation which alleviates concerns about antibody 68

PAGE 69

saturation and the effect of early postpartum sickness. In light of these findings, in an effort to streamline the method in which AMIR categoriz ation is accomplished, Equation 3-6 using the quartile method will be used from here on out. 69

PAGE 70

Figure 3-1. Basic outline for treatment/ sampling for antibody mediated immune responsiveness. 70

PAGE 71

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8-8 -3 0 +2 Weeks Relative to CalvingOptical Density Figure 3-2. Optical density values reflec ting antibody response to OVA by sampling period separated by parity. A) Solid line is for multiparous cows. B) Dotted line is for primiparous cows. Bars indicate 95% confidence intervals. 71

PAGE 72

0.000 0.500 1.000 1.500 2.000 010203040506 Days Interval 1OD-3 Value 0 Figure 3-3. Scatter plot of OD -3 values by length of interval 1 for primiparous cows. Fitted line not significant (p = 0.34, = -0.0015). 72

PAGE 73

0.000 0.500 1.000 1.500 2.000 2.500 0102030405060708 Days Interval 1OD-3 Value 0 Figure 3-4. Scatter plot of OD -3 values by length of interval 1 for multiparous cows. Fitted line is significant (p < 0.0001, = -0.0081). 73

PAGE 74

0.000 0.500 1.000 1.500 2.000 2.500 01 02 03 04 05 Days Interval 2OD0 Value 0 Figure 3-5. Scatter plot of OD 0 values by length of interval 2 for primiparous cows. Fitted line is significant (p < 0.0001, = -0.0246). 74

PAGE 75

0.000 0.500 1.000 1.500 2.000 2.500 01 02 03 04 05 Days Interval 2OD0 Value 0 Figure 3-6. Scatter plot of OD 0 values by length of interval 2 for multiparous cows. Fitted line is significant (p < 0.0001, = -0.0211). 75

PAGE 76

0 0.5 1 1.5 2 2.5 10 12 14 16 18 20 22 Days Interval 3OD+2 Value Figure 3-7. Scatter plot of OD +2 values by length of interval 3 for primiparous cows. Fitted line is significant (p < 0.0032, = -0.0298). 76

PAGE 77

0 0.5 1 1.5 2 2.5 10121416182022 Days Interval 3OD+2 Value Figure 3-8. Scatter plot of OD +2 values by length of interval 3 for multiparous cows. Fitted line is not significant (p = 0.21, = -0.0104). 77

PAGE 78

-0.100 0.000 0.100 0.200 0.300 0.400 0.500 0.600 1234567891011121314 Ranking by OD-3 ValueChange in OD Over Interval 2 Figure 3-9. Graph reflecting the effect of maximal antibody response. Cows were ranked based on their OD -3 value (x-axis) with 55 cows pe r group ranking except group 14 which has 59 cows. Cows in group 1 have the top 55 OD -3 values. Cows in group 14 have the bottom 59 OD -3 values. The y axis represents OD 0 OD -3 Data points indicate the average change in OD over interv al 2 for each group. Bars reflect SEM 78

PAGE 79

-0.100 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 0.800 0.900 1.000 1234567891011121314 Ranking by OD0 ValueChange in OD Over Interval 3 Figure 3-10. Graph reflecting th e effect of maximal antibody response. Cows were ranked based on their OD 0 value (x-axis) with 55 cows pe r group ranking except group 14 which has 39 cows. Cows in group 1 have the top 55 OD 0 values. Cows in group 14 have the bottom 39 OD 0 values. The y axis represents the change in OD over Interval 3 (OD +2 OD 0 ). Data points indicate the average change in OD over interval 3 for each group. Bars reflect SEM. 79

PAGE 80

Figure 3-11. Depiction of a method used to categorize antib ody mediated immune responsiveness. Mean and standard deviation are calculated based on y total 1 values for the population. With a normal distribution roughly 16% of the population becomes high responders, 16% become low responders, and roughly 68% are medium responders. 1 y total total antibody response value ge nerated through the use of an immune response index (Equation 3-3, E quation 3-4, Equation 3-5, Equation 3-6). 80

PAGE 81

Table 3-1. Incidence of di sease for each antibody re sponse categorization method. Mastitis Equation 3-3 1 Equation 3-4 2 Equation 3-5 3 Equation 3-6 4 Categorization St.Dev 5 Quart 6 St.Dev Quart. St.Dev Quart. St.Dev Quart. Low 27.45 27.98 23.73 24.12 25.69 27.71 24.35 23.35 Medium 26.93 27.86 28.05 29.33 26.64 26.30 27.39 30.14 High 22.83 21.95 23.01 22.84 26.42 25.47 24.77 21.74 Metritis Equation 3-3 Equation 3-4 Equation 3-5 Equation 3-6 Categorization St.Dev Quart. St.Dev Quart. St.Dev Quart. St.Dev Quart. Low 4.50 5.32 7.69 7.45 6.56 5.32 7.69 6.38 Medium 5.08 4.79 4.08 3.46 4.17 4.26 4.06 3.72 High 7.41 6.42 7.63 6.95 8.80 7.49 7.81 7.49 Metritis in cows having retained fetal membrane. Equation 3-3 Equation 3-4 Equation 3-5 Equation 3-6 Categorization St.Dev Quart. St.Dev Quart. St.Dev Quart. St.Dev Quart. Low 22.22 37.50 43.75 42.11 45.45 33.33 43.75 36.84 Medium 32.50 21.43 20.69 19.05 22.86 20.83 21.43 20.00 High 0.00 27.27 20.00 20.00 22.22 30.77 18.18 25.00 Ketosis Equation 3-3 Equation 3-4 Equation 3-5 Equation 3-6 Categorization St.Dev Quart. St.Dev Quart. St.Dev Quart. St.Dev Quart. Low 4.50 4.79 6.92 7.45 4.92 5.85 9.23 7.45 Medium 4.68 5.03 4.67 4.50 4.74 4.76 3.84 4.23 High 2.75 2.66 0.76 1.06 2.38 2.13 1.55 1.60 Displaced Abomassum Equation 3-3 Equation 3-4 Equation 3-5 Equation 3-6 Categorization St.Dev Quart. St.Dev Quart. St.Dev Quart. St.Dev Quart. Low 4.50 4.26 3.08 3.72 4.92 4.26 3.85 3.19 Medium 2.06 1.85 2.24 2.12 1.78 1.85 2.02 2.38 High 3.67 2.66 3.79 2.66 3.97 2.66 3.88 2.66 Retained fetal membrane Equation 3-3 Equation 3-4 Equation 3-5 Equation 3-6 Categorization St.Dev Quart. St.Dev Quart. St.Dev Quart. St.Dev Quart. Low 8.11 8.51 12.31 10.11 9.02 9.57 12.31 10.11 Medium 7.49 7.41 5.89 5.56 6.92 6.35 5.66 5.29 High 5.50 5.85 7.58 7.98 7.14 6.91 8.53 8.51 1 Equation 3-3, refer to text. 2 Equation 3-4, refer to text. 3 Equation 3-5, refer to text. 4 Equation 36, refer to text. 5 St Dev.,standard deviation, refers to a method used to extrapolate antibody response categorizations (Figure 3-11). 6 Quart, quartiles, refers to a method used to extrapolate antibody response cat egorizations 81

PAGE 82

CHAPTER 4 CATEGORIZATION OF PERIPARTURIENT ANTIBODY RESPONSE TO OVALBUMIN AND ITS RELATIONSHIP WITH CO MMON DISEASES AND PERFORMANCE MEASURES OF HOLSTEIN DAIRY CATTLE Introduction The abrupt transition from a pregnant non-lactating cow to a non-pregnant lactating cow has a deleterious effect on immune function around the time of calving. This periparturient immune suppression is well documen ted (Lacetera et al., 2005; Saad et al., 1989; Detilleux et al., 1994; Park et al., 1992) and believed to be at least partially respons ible for the increased risk of disease during this period. This effect appears to have several contributing mediators such as, increased stress and hormones associated with st ress (Van Kampen and Mallard 1997; Mallard et al., 1997), hypocalcemia (Kimura et al., 2006), negative energy balan ce with the resultant effect of nonesterified fatty acids (Lacetera et al., 2005) and h yperketonemia (Franklin et al., 1991; Hoeben et al., 1997). Although these mediators have a significant e ffect on peripartum immune responsiveness, there also appears to be a gradual decline in immune competence associated with genetic selection focused on production traits which is made evident by the increasing risk of disease (Harmon, 1994; Heringstad, 2000; Emanuelson, 1988). Th is trend is occurring despite the efforts and major advances in sanitation and housing. This effect is believed to be fueled by placing the priority for genetic selection toward increased milk yield with little selection pressure for measures of disease resistance. Selection for increased milk production void of measures for resistance to mastitis has been found to result in a genetic increase of 0.02 cases of mastitis per cow per year (Strandberg and Shook, 1989). This transl ates into a genetic increase of 2 mastitis cases for every 100 dairy cows per year. 82

PAGE 83

To overcome this unfavorable tre nd, selection pressure is being applied to traits associated with disease resistance. Selection for decreased SCC, productive life, and structural traits of the udder have been studied (Nash et al., 2000; Heringstad et al ., 2006) and further implemented. However, associations between these traits and resistance to disease are rather crude and do not directly focus on the immune system which is primarily responsible for host defense. As a result, measures of the immune function have been eval uated as indicators of health (Wagter et al., 2000; Rupp et al., 2007; Park et al., 2004). Significant associations with mastitis risk between differing populations of T lymphocyte subsets have been reported (Park et al., 1993, 2004). Several studies have found significant differences in immune responsiveness and resistance to disease with alterati ons in class I and II MHC haplotypes and alleles (Rupp et al., 2007; Park et al., 2004; Rupp a nd Didier, 2003). Other studies have compared incidence of disease among cows categorized based on their antibodymediated immune responsiveness (AMIR) to test antigen (Wagter et al., 2000; Mallard et al., 1997; Hernandez et al., 2003). Research by Wagter et. al. (2000) reported substantial va riation among individual cows ability to mount an antibody response around calvi ng. In fact, not all cows experience peripartum immune suppression and cows c ould be categorized based on AM IR to ovalbumin (OVA). The high responders for AMIR had the lowest incidence of mastitis in two of the three study herds. Individual cows antibody responsiveness to OVA al so had a positive signifi cant correlation with antibody titers to E. coli J5 vaccination. This response to OVA was highly heritable, ranging from 0.32 to 0.62 depending on week relativ e to calving (Wagter et al., 2000). Although Wagter et al. (2000) di d find favorable associations between this measure of AMIR and disease risk; with 136 co ws and heifers spread over three herds, it was difficult to find 83

PAGE 84

statistical significance. In the current trial this problem was addressed by incorporating a much larger sample size in order to more adequately study the relationship between AMIR and disease risk. The hypothesis being that high AMIR will be associated with lower disease risk. The objectives were to categorize cows based on AM IR to OVA and then test for associations between AMIR and disease incidence, namely; mastitis, metritis, retained fetal membrane, ketosis, and displaced abomasum. Additiona lly, possible associations between AMIR categorization and reproductive efficiency, milk yield and somatic cell score were examined. Materials and Methods Research Sample Cows In total, 875 Holstein cows/heifers were enrolled into the study popul ation at 8 weeks (wk8) prior to calving. In cows, this was the initiation of the dry pe riod. Animals were enrolled if the expected dry period was less than 90 days, if reconfirmed pregnant and if found in good health with no obvious signs of disease. All test animals were from a singl e herd in north central Florida which maintains exceptional record keeping. Al l cows and heifers were: enrolled between September 9 th and December 31 st 2004; calved between October 25 th 2004 and March 12 th 2005; and sampling ended between November 9 th 2004 and March 28 th 2005. All cows and heifers received a routine dry off, prefresh, and fresh cow protocol. Animal Removal and Interval Criteria Of the originally enrolled 875 cows and he ifers, 13 were removed due to missing data, yielding 862. Interval 1 (Int 1 ) was defined as time from enrollment (wk-8) to entry into springer pen (wk-3). Interval 2 (Int 2 ) was defined as time from entry in to springer pen (wk-3) to calving (wk0) (Figure 4-1). Animals were removed from the study if the period from wk-8 to wk0 (dry period length for cows) was more than 90 days Additional animals were excluded if Int 1 or Int 2 was less than 12 days in length. These time frames are relevant because they coincided with 84

PAGE 85

antigen exposure and blood sampling to measur e antibody response to an tigen. Interval 3 (Int 3 ) was defined as the period between calving (wk0) and end of the sampling period (wk+2). No exclusionary criterion was needed for Int 3 because this interval was under investigator control. Of the 862 cows, 50 were removed because they did not meet one of these interval restrictions. A preliminary analysis was used to remove an additional 38 animals found not nave to test antigen. This resulted in a sample size of 774 with 433 cows and 341 heifers. The heifers upon enrollment will be referred to as primiparous cows while cows upon enrollment will be referred to as multiparous cows. Test animals were si red by 237 different sires, which should provide adequate variation in the gene pool for a sire effect on immune responsiveness. Body Condition Scoring Body condition was scored in point increments using the 5-point scale by Ferguson et al. (1994). Body condition scoring (BCS) was grouped into high, medium, and low categories. For cows during wk-8, wk-3, and wk0, a BCS between 3.0 3.75 was coded medium, those below 3.0 were coded low and those above 3.75 were coded high. Heifers at wk-8, wk-3, and wk0 were coded medium if BCS was between 3.0 3.5. A BCS above 3.5 was considered high, and a BCS below 3.0 was considered low. At wk+2, all anim als were coded normal if BCS fell in the range of 2.75 3.5. The loss of body condition, not just BCS alone was reported to be responsible for alterations in lymphocyte function (Lacetera et al., 2005; Wentink et al ., 1997; Kaneene et al., 1997). To account for this factor the interval chan ge in BCS for various intervals was calculated. Interval 4 was defined as the period between wk-3 and wk+2, which serves as an indicator of BCS lost over the transition period. Interval 5 wa s defined as the period between wk-8 and wk0. Interval 6 was defined as the period between wk-8 and wk+2. 85

PAGE 86

Immunization Ovalbumin was chosen as the test antigen due to its inert properties, its ability to stimulate antibody, a cows reduced likelihood of previous e xposure and its previous success as a tool to categorize AMIR (Wagter et al. 2000; Mallard et al., 1997; Hernandez et al., 2003). Animals received 1 mg ovalbumin (OVA; chicken albumi n, Type VII, Sigma-Aldrich, St. Louis, MO, USA) at three times; week -8 (w k-8), week -3 (wk-3), and week 0 (wk 0) relative to calving (Figure 4-1). The wk-8 and wk-3 OV A immunizations were dissolved in Escherichia coli J5 vaccine (J5 bacterin, Pfizer Inc., Kalamazoo, MI USA) with the manufacturers adjuvant which coincided with the farms routine vaccine protocol. At wk0, the 1 mg OVA was dissolved in 1 ml phosphate buffer saline (PBS, pH 7.4). All su spensions were then mixed with a type 1( C. albicans raw whole cell material, Greer Laboratories, Lenoir, NC, US A) antigen to stimulate a cell-mediated immune response (C MIR) and vortexed for at least one minute (Refer to chapter 5 for CMIR analysis). Blood Collection and Processing To determine antibody response to OVA, blood was collected by caudal venipuncture at; wk-8, wk-3, wk0, and wk+2 relative to calving (F igure 4-1). At calving (wk0), the blood sample was collected within 12 hours of pa rturition. Samples were collected into sterile 10 ml evacuated blood collection tubes with no additive, then put on ice during transport back to lab. Serum was collected after centrifugation at 4,000 rpm and stored at -70 C. Enzyme-Linked Immunosorbent Assay A cows specific antibody response to OVA was de tected using an indirect enzyme-linked immunosorbent assay (ELISA) method as previous ly described (Burton et al., 1989; Wagter et al., 2000; Mallard et al., 1997). For positive cont rol sample, 10 lactating cows received 1 mg OVA and 0.5 mg Quil A adjuvant suspended in 1 ml PBS on day 0 and 14. On day 21, serum 86

PAGE 87

from these cows was collected and pooled for positive control sample. Negative control samples consisted of a pool of serum from cows at wk-8. Polystyrene 96-well plates (Immulon II, Fisher Scientific Co. Ltd., Pittsburgh, PA, USA) (Figure 4-2) were stored for 48 h at 4 C after being coated with 100 L/ well of 1.4 mg OVA dissolved in 1 mL carbonate-bicar bonate coating buffer (pH 9.6). Plates were washed four times in a plate washer (ELX50 plate washer, Biotek Instruments, Inc., Winooski, VT, USA) with wash buffer solution containing PBS and 0.05 % Tween 20 (Sigma-A ldrich) (washing buffer, pH 7.4). Blocking solution (PBS pH 7.4, 3% Tween 20, 1% bovine serum albumin) was added (200 L / well) and plates incubated at room temperat ure for 1 h. Plates were washed 4 times before applying 100 L / well of control and test se ra. All samples were diluted to 1/50 and 1/200 dilutions using wash buffer solution. Positive and negative controls (1/50 and 1/200) were run in quadruplicates while test sera (1/ 50 and 1/200) were run in duplicat es placed in separate diagonal plate quadrants (Figure 4-3). Plates were then in cubated for 2 h at room temperature. After being washed 4 times, 100 L / well of alkaline phosp hatase-conjugated rabb it anti-bovine IgG (whole molecule; Sigma Chemical Co., St. Louis, MO) dissolved in Tris buffer solution (TBS, pH 7.4 and 0.05% Tween) at a 1/38000 dilution was added, and plates incubated for 1 h at room temperature. Plates were washed 4 times be fore 80 L / well of p-nitrophenyl phosphate disodium were added. Plates were incubated for 30 min at room temperature out of direct light. Optical density (OD) values were then determined for each sample with an ELISA plate reader (MRX Revelation, Dynes Technologies VA, USA) set at an absorbance of 405 m and 630 m (Revelation software, Dynes Technologies VA, USA). Coefficient of variation was calculated fo r the 1/50 and 1/200 positive controls to determine whether plates were accepted or rejected. The coefficient of variation for the 1/50 87

PAGE 88

dilution was calculated by dividing the standard deviation by the mean of the 1/50 positive control values. This was also performed for the 1/200 dilution. The maximum allowable variation for each plate was 20% for either the 1/50 or 1/200 positive control dilution. In order to correct for variation among plates a correction factor was determined for each plate. The correction factor was determined by comparing the positive control OD values of each plate to the mean of every plate. The mean of each plates 1/50 positive control dilution was summed with the mean of the 1/200 positive control dilution. This additive positive control value was then divided into the mean of every plates additive value. The resultant value functioned as the correction factor. Since each test sample was run in duplicate, each test samples mean 1/50 OD value was summed with the mean 1/200 OD value. This value was then multiplied by the plates correction factor to determine the samples co rrected OD value to be utilized in statistical analysis. Preliminary Analysis Removal of non-naive Due to the differences in the kinetics of a primary and sec ondary antibody response (Figure 2-1), it was important that all cows receive equal treatment which requires all cows to be nave to OVA at enrollment. This should result in low OD values for wk-8 (OD -8 ) due to the lack of antibody with affinity for OVA. Inspection of the OD -8 values revealed that some cows were not nave. Sorting the cows based on their OD -8 values revealed a natural cut-off where the values began increasing rapidly. This point also coincided with one sta ndard deviation above the mean for the OD -8 values. As a result 38 cows were excluded due to not being nave to test antigen. For further discussion refer to chapter 3. 88

PAGE 89

Parity effect A repeated measures analysis using OD -8 as a covariate with the mixed procedure of SAS revealed that multiparous cows responded significantly higher than primiparous cows at every antibody response measurement week (p<0.0001). Due to this effect, adjustments to OD values as well as AMIR categorizations were all made with respect to parity (F igure 3-2). Explanations for this finding may include the presence of a mo re extensive antibody repe rtoire in older cows. This could simply be due to the effect of time, allowing greater exposure to a broader array of pathogens. Interval analysis and optical density value adjustment The length of the interval between antigen administration and blood sampling has strong relevance due to the different phases of an an tibody response (Figure 21). As a result, an analysis was performed to determine this effect and subsequently adjust the OD values for interval length (Figure 3-3, 34, 3-5, 3-6, 3-7, 3-8). A generalize d linear model was constructed using the GLM procedure of SAS to determine if the duration in days of Int 1 Int 2 or Int 3 significantly influenced OD -3 OD 0 or OD +2 respectively. For this analysis, OD response served as the outcome variable while possible explanatory variables included: Int Y = interval length in days for interval y BCS X = categorization of body condition score for sampling week x BCS Y = change in body condition score over interval y OD X = optical density at previous sampling week x Dex = binary effect, whether cow recei ved dexamethasone prior to calving Sick = binary effect, incidence of either medium or severe case of mastitis, metritis, ketosis, or displaced abomasum within 16 DI M. Only used for analysis to correct OD +2 for reasons discussed in chapter 3. 89

PAGE 90

Explanatory variables remained in the model if the effect showed a tendency (P < 0.10) to predict the desired OD value. With = 0.05, if statistical significance was found with the main effect (Int y ), the resultant parameter estimate was applied to Equation 3-1 to correct the corresponding OD values. If a given OD value was adjusted based on interval length, the model predicting the OD value for the subsequent samp ling period would have the adjusted OD from the previous sampling period. AOD ZPX = OD ZPX + PE YP (Mint YP Int ZPY ) (3-1) Where: AOD ZPX = the adjusted OD value for cow z in parity p for week x OD ZPX = OD value for cow z in parity p for week x PE YP = parameter estimate for the effect interv al y has on ODx values for parity p. Also, interval y must always immediately precede week x. Mint YP = median number of days for interval y for parity p Int ZPY = the actual interval length for cow z in parity p for interval y As a result the OD -3 values were adjusted for multiparous cows only. The OD 0 values were adjusted for both multiparous and primiparous cows. The OD +2 values were adjusted for heifers only. For further reference refer to chapter 3. Classification analysis Maximal response: Previous work devised an index usi ng the interval change in OD value to generate total antibody values used to extr apolate AMIR classificati ons (Wagter et al. 2000). This index weighted those values around calving if they showed a decline in antibody response. In this study, inspection of the OD values for the respective weeks revealed what appeared to be a maximal antibody response. If there was an improvement after subsequent antigen injection, these maximal points in antibody concentration only slightly improved. If this peak response was 90

PAGE 91

achieved early in the study trial (OD -3 ), which would indicate a hi gh prepartum responder; there would be little if any room for an added response. As a result an index concerned with weighting the change in OD during the intervals adjacent to calving (Eq. 1-1), may negatively impact the categorization of great immune responders who reach ed this saturation poin t early because there was little or no room to further respond. After sorting the OD -3 values from greatest to least, it was discovered that of the top 20 cows for OD -3 value, 15 had a smaller OD 0 value (75%), yielding a nega tive interval 2 change in OD. In these instances, 15 of the top 20 OD -3 responders would have an amplified ( = 1.5) negative I 2 value applied to their AMIR index if usi ng an index which weights interval changes in OD. After grouping the OD -3 values from 1 to 14, with group 1 being the top 55 OD -3 values and group 14 the bottom 59 OD -3 values, a one-tailed two sample T-test was used to determine if the OD value level grouping had a significant effect on the subsequent inte rval change in OD. This revealed a significantly smaller change in OD over interval 2 for group 1 cows compared to group 2 cows (p < 0.0001) (Figure 3-9). This va lue actually averaged below 0 (-0.05) for these top 55 OD -3 responders. Linear regr ession also revealed OD -3 is a significant predictor for I 2 (p < 0.0001; = -0.28). A negative indicates that as OD -3 values increase, I 2 values decrease. Repeating the same process by ranking cows based on OD 0 values in order to compare interval 3 change in OD, revealed similar results. Due to missing wk+2 blood samples the sample size for this analysis was 754. As a result group 14 was comprised of 39 cows while group 1 13 comprised of 55. Of the top 20 cows for OD 0 values, 15 had negative interval 3 changes in OD (75%). Also, group 1 cows based on OD 0 values had a significantly smaller interval 3 change in OD compared to group 2 co ws (p < 0.0001) (Figure 3-10). Linear regression 91

PAGE 92

also revealed OD 0 is a significant predictor for I 3 (p < 0.0001; = -0.50). A negative indicates that as OD 0 values increase, I 3 values decrease. In these instances of high OD values follo wed by a subsequent negative interval, the interval is not negative due to a poor subs equent OD value. Of the top 20 cows for OD -3 18 still had OD 0 values above the third quartile for the popul ation and the other 2 were still above the median for the population. For the top 20 cows for OD 0 all 20 still had OD +2 values above the third quartile for the population. The negative interval was simply the result of an inability to respond further. As a result, additional indexe s were generated and analyzed based on their correlation with disease incidence. Early postpartum measures of immune function: Previous work utilized the antibody response measured early postpartum as a tool to categorize AMIR (Wagter et al., 2000; Mallard et al., 1997).This period immediat ely following parturition is a common occasion for increased incidence of disease. This effect of substa ntial sickness could be a confounding variable for antibody responsiveness to OVA detected early po stpartum. As a result it may be hypothesized that substantial sickness could contribute to immune suppression. This would make it difficult to study the effect measures of imm une responsiveness have on disease risk if the association could also be in the opposing direction. A linear regression model was used w ith the REG procedure of SAS. OD +2 served as the outcome variable while potential explanatory variables included: BCS x = categorization of body condition score for sampling week x. BCS y = change in body condition score over interval y. OD 0 = optical density value at wk0. Dex = binary effect, whether cow recei ved dexamethasone prior to calving. 92

PAGE 93

Sick = binary effect, incidence of either medium or severe case of mastitis, metritis, ketosis, or displaced abomasum within 16 DIM. Parity = binary effect, either primiparous or multiparous. Explanatory variables remained in the model if the effect showed a tendency (P < 0.10) to predict the dependant OD value. For statistical significance, = 0.05. The resulting model was as follows (Eq. 3-2): OD +2 = Sick + Dex + OD 0 (3-2) The analysis revealed that sickness, as prev iously defined, was a si gnificant predictor of OD +2 (p = 0.0156). The difficulty in this analysis is determination of causality, did the occurrence of sickness within 16 DIM cause a suppression in i mmune responsiveness; or did inferior immune responsiveness cause sickness within 16 DIM. Because OD 0 occurs prior to the incidence of disease, and this was a significant predictor of OD +2 (p < 0.0001), it appeared that the incidence of sickness had an effect on immune responsiveness. Correlation analysis using the Corr procedure of SAS was employed to study the relationship between OD 0 and OD +2 at fixed levels of sickness. Among cows considered healthy within 16 days in milk, corre lation analysis revealed OD 0 is positively correlated with OD +2 (r 2 = 0.64, p < 0.0001). Furthermore, within sick cows, the correlation between OD 0 and OD +2 has an even greater significantly positive correlation (r 2 = 0.73, p < 0.0001). For further reference see chapter 3. Antibody-Mediated Immune Response Classification To alleviate the concerns about the effect of antibody saturation and early postpartum sickness, a new index was generated for the ca tegorization of AMIR (Eq. 3-6). The rationale behind this index is that favorable AMIR will ha ve greater correlation with the actual magnitude of the antibody concentration rather than change s in antibody concentration over intervals. In the 93

PAGE 94

case of antibody saturation, an index should re flect a cows ability to maintain a high concentration of specific antibody. This should al l be accomplished while also using measures which give special attention to antib ody responses occurring peripartum. y total = OD -3 + OD 0 *(1 + I 2 ) (3-6) Where: y total = total antibody OD -3 = optical density value at wk-3 OD 0 = optical density value at wk0 I 2 = change in OD between week -3 and week 0 In this index, the direct magnitudes of the OD values are considered. However, in the case of OD 0 it is still weighted, but this can be positively or negatively and only in proportion to the level of increase or decrease for I 2 and I 3 Extrapolation of AMIR categorizations occurs by use of the y total values for individual cows and then configuring the quartiles respective of parity (primiparous vs. multiparous). Cows within the bottom 25% for a respective parity are categorized as low AMIR responders, while cows in the top 25% for a resp ective parity are termed high AM IR responders. The remaining middle 50% are categorized as medium AMIR responders. Diseases Identification of disease was performed by fa rm personnel who were blinded to immune response categorizations. The diseases of interest fo r this project were; mastitis, metritis, retained fetal membranes, ketosis, and displaced abomasum. All diseases were recorded as yes/no binary responses for the trial peri od of the current lactation. Mastitis was recorded through 365 DIM as light, medium, or severe. Severity was recorded as light if there were no systemic signs and milk was slightly watery with minimal clots (gargot). 94

PAGE 95

Severity was medium if there were no systemic signs and a substantial amount of clots were observed in the milk. Severity was severe if there were systemic signs, watery milk, and a substantial amount of clots in the milk. Metritis was recorded through 30 DIM as light, medium or severe. Severity was light if there was an abnormal vaginal discharge and a pa lpable uterine lumen. Severity was medium if there was a purulent vaginal discharge, with an enlarged, not-flaccid uterus. Severity was severe if there was a purulent foul-smelling vaginal di scharge, with an en larged flaccid uterus. Association between AMIR a nd energy related metabolic c onditions including ketosis and DA were also analyzed. Ketosis was recorded through 30 DIM as light, medium, or severe. Ketosis was coded as light if the urine cont ained 15 mg/dL of ketone bodies. Severity was medium if the urine contained 40 mg/dL of ketone bodies. The severity was severe if the urine contained > 80 mg/dL of ketone bodies. Displa ced abomasum was recorded through 50 DIM. Because retained fetal membranes are now understood to result from an inadequate immune response (Kimura et al., 2002), the re lationship between AMIR and RFM was also considered. Retained fetal membranes were id entified if there was placental retention 24 h postpartum. Milk Yield, Somatic Cell Score, and Reproductive Efficiency Milk yield was gathered from the Dairy Herd Information Association (DHIA) records for the current lactation us ing ME305, which is an estimate of the milk yield for 305 DIM. For categorical data, low milk producer s were identified if there 305 day milk was in the bottom 25% of the study group. High milk producers were th ose in the top 25% of the study group. The remaining middle 50% were classifi ed as medium milk producers. 95

PAGE 96

Somatic cell score (SCS) was gathered from DHIA records. The average SCS was determined for the first 3 test days, the first 6 te st days, and for the first 10 test days, which are an indication of the average SCS for the first 90, 180, and 300 DIM respectively. To analyze for associations between AMIR categorization and reproductive efficiency, a binary pregnancy term was used which simply indicated if a given cow was pregnant by 150 DIM. Addition quantitative variable measures of fe rtility included; number of days that a cow is not pregnant (days open), and number of times bred. Statistics To analyze the associations between risk of disease and reproductive efficiency with AMIR categorization, a logistic regression model was developed using the LOGISTIC procedure of SAS. Other than the main effect (AMIR categorization), possible explanatory variables include: BCS x = categorical effect of body condition score for sampling week x. BCS y = quantitative variable, change in body condition score over interval y. Dex = binary effect, whether cow recei ved dexamethasone prior to calving. Parity = binary effect, either primiparous or multiparous. RFM = binary effect, incidence of retained fetal membrane. CDiff = binary effect, if reported difficult calving. Ketosis = binary effect, incidence of ketosis. SCSavg = somatic cell score average over 10 monthly test days. For analyzing the relationship between AMIR categorization and the quantitative variables for SCS, 305 day milk yield, and reproductive efficiency, linear regression was used with the REG procedure of SAS. All categorical variables will be analyzed using logistic regression using the LOGISTIC procedure of SAS. All relevant effects were put in the model. A backwords elim ination procedure was used to determine the final model. Explanatory variable s remained in the model if they showed a 96

PAGE 97

tendency (p < 0.10) to predict the outcome variab le. Statistical significance was determined by setting = 0.05. Results Mastitis and Metritis No significant association was found between AMIR status and mastitis within 100 DIM. When considering the incidence within 365 days; there were 169 cases that were either medium or severe (22%). Forty (40) cases were reco rded in primiparous cows and 129 in multiparous cows. Effects remaining in the model were: SCS average, BCS 6 Parity. Antibody response categorization was a significant predictor of mo derate and severe mastitis risk (p = 0.0082) (Table 4-1). Although the low responders were no t statistically different than medium and high responders collectively (p = 0.12), the medium responders were 1.76 (CI = 1.08 2.89) times more likely to have an occurrence of moderate or severe mastitis than high responders. The recorded incidence for light medium or severe metritis during the first 30 DIM was remarkably low with only 41 cases; 27 cases in primiparous cows and 14 cases in older cows (5.3% overall). Remaining model effects were: Parity, BCS 0 Dex, and RFM. Antibody response categorization was not significantly associated with occurrence of metritis in this analysis (p = 0.29) (Table 4-1). Ketosis Displaced Abomasum and Retained Fetal Membrane Only 45 cases of ketosis occurred within 30 days of calving (5.8%). Primiparous cows accounted for 24 cases while multiparous accounted for 21 cases. The model included BCS 6 and the main effect (AMIR category). The low responders were 2.90 (CI = 1.10 7.62) times more likely to develop ketosis than hi gh responders (Table 4-1) (Figure 4-4). 97

PAGE 98

Displacement of the abomasum occurred in 21 cows (2.8%), with 12 from primiparous cows and 9 cases in multiparous cows. Remaining model effects were BCS 4 and ketosis. In this analysis, AMIR was not a significant predictor of DA incidence (Table 4-1). There were 57 cases of RFM for the study popul ation (7.4%). Primiparous cows accounted for 18 cases while multiparous cows accounted for 39 cases. Remaining model effects were: BCS 5 and BCS 0 The AMIR categorization was not a si gnificant predictor of RFM incidence (p = 0.17) (Table 4-1). Milk Yield and Somatic Cell Score For the analysis of the effect of AMIR cate gorization on milk yield, the contributing model effects were parity, number of days open, and the bi nary trait of mastitis. In this analysis, AMIR categorization had a tendency to predict milk yield (p = 0.06, = -347.34) (Figure 4-5). The analysis of the effect of AMIR on SCS included BCS 6 milk categorization, and the binary mastitis variable. The effect of AMIR on SCS wa s not a significant pred ictor of SCS (p = 0.40). Reproductive Efficiency The model analyzing the effect of AMIR cat egory on pregnancy at 150 DIM included the explanatory variables; BCS 4 and milk categorization. The e ffect of AMIR was a significant predictor for pregnancy by 150 DIM (p = 0.003). The low antibody-mediated responders were 2.32 (CI = 1.44 3.75) times more likely than high responders to become pregnant by 150 DIM. Also, the low antibody-mediated responders we re 1.57 (CI = 1.20 2.05) times more likely to become pregnant by 150 DIM than medium and low responders collectively (Figure 4-6). Discussion The significantly higher odds of mastitis for medium antibody-mediated responders compared to high responders indicate the va lidity of using antibody response to OVA as a measure of AMIR. These results coincide with the findings in Wagter et al. (2000) where low 98

PAGE 99

responders had the highest occurrence of mastitis, and antibody response to OVA was significantly correlated w ith antibody response to E. coli J5 vaccination. The lack of additional statistical findings for mastitis and metritis ar e likely due to low recorded disease frequency. Mechanisms linking energy-related metabolic disorders and measures of immune response are not completely understood. A ssociations between ketosis a nd suppressed immune response are largely believed to be the result of ketosis causing immune suppression. However in this study, AMIR categorization occurred pr epartum. Given that ketosis is a postpartum disorder, it is not possible that ketosis caused a suppressed immune response during categorization. With change in BCS remaining in the model, prepartu m BCS did not predispose cows to ketosis while influencing immune responsiveness. The odds for this effect followed a High < Medium < Low pattern. The discovery of inverse associations betw een AMIR categorization and milk yield and fertility was an unexpected result. With regard to milk yield, these findings do not agree with previous literature in dairy cows (Wagter et al., 2003; Detilleux et al., 1995) or with other performance measures in swine (Mallard et al., 1998). However, possible explanations may arise from studies in other metabolically active sp ecies. Poultry selected for greater immune responsiveness have reported a reduction in growth performa nce (Klasing et al., 1987, 1998; Soler et al., 2003). This was explained by bodily function competition for available nutrients. The energy and nutrients required for maintena nce and activation of a superior immune responder could otherwise be used for other phenotypic traits. There is no known biological rele vance for high immune responde rs to be predisposed to lower milk yields or decreased fertility. This finding may simply be the result of neglecting to 99

PAGE 100

select or having inadequate tools to select for immune responsiveness while putting direct selection pressure on the metabolica lly demanding trait of milk yield. There is another possible explanation for the unf avorable decline in fertility. This involves the maternal recognition of the conceptus by the immune system. During pregnancy the conceptus is a foreign body which otherwise would be subject to a maternal immune response followed by rejection. This reje ction, however, is blocked due to various immune suppressive activities which are initiated during maternal recognition of pregnancy (Hansen, 1997). This relationship between the immune system and pregna ncy brings the potential for an inappropriate immune response rejection of the conceptus. If it could be proven that cows who mount a superior antibody-mediated immune response are mo re likely to mount an inappropriate immune response against the conceptus, this could explain the present findings. In hindsight of this study, there are certain techniques and st rategies that could or should be implemented upon further research. Although it is difficult in a large sc ale dairy setting, and this aspect was accounted for in the present study, a tighter contro l on interval duration between antigen injection and blood co llection could be practiced. Another improvement would be in the use of antigen. A more represen tative categorization of AMIR could be obtained through two or more an tigens. This philosophy has been practiced in other species (Mallard et al.,1998). To maintain the integrity of a broad based approach to improving immune response when selection pressu re is applied for AMIR, it should be based on more than one antigen. Using the antibody response to OVA around calvi ng as a representative of the ability to mount an immune response during immune suppre ssion may be partially confounded due to the fact that OVA was previously injected at wk-8 and wk-3. This is because by this point memory 100

PAGE 101

lymphocytes have already formed for OVA and they are the cells being activated to mount the antibody response. These memory cells are much eas ier to activate and have greater affinity to antigen. For a true response during immune su ppression primary exposure to antigen should occur closer to parturition. As a result, if us ing multiple antigens, exposure to these antigens could be staggered or initially introduced at different time points with one occurring at calving. Due to the rising incidence of disease and the ever increasing necessity to produce milk as proficiently as possible, a more proactive and effective approach to disease resistan ce is required. Taking this initiative should also help alleviate concerns consumers have for animal welfare and usage of antibiotics. 101

PAGE 102

Figure 4-1. General outline of experimental design. 102

PAGE 103

Figure 4-2. Polystyrene 96well plate for enzyme linked immunosorbent assay. 103

PAGE 104

Figure 4-3. Diagram of the placement of test se ra into 96 well polysty rene plate for enzymelinked immunosorbent assay. Yellow wells identify 1/50 d ilution, while grey wells identify 1/200 dilution. All positive and negati ve controls were run in columns 1 and 7. Well identified as a is a 1/50 dilution of sample 1, which was run in column 2 of row A and duplicated in column 8 of row E. Sample identified as b is a 1/200 dilution of sample 1 run in column 2 of row B and duplicated in column 8 of row F. Matching background row colors indicate ha lf rows which were duplicated. Wells identified with B indicate blank wells for calibration of plate reader. 104

PAGE 105

Table 4-1. Odds ratios of disease inci dence for antibody res ponse categorizations. Mastitis Metritis Ketosis DA RFM 0.77 0.93 1.69 0.91 1.29 Low vs Med & High (0.55 1.07) (0.56 1.54) (1.06 2.69) (0.45 1.85) (0.86 1.94) 0.89 0.64 2.9 0.7 1.16 Low vs High (0.49 1.64) (0.27 1.56) (1.10 7.62) (0.19 2.6) (0.56 2.42) 1.76 0.521 1.75 0.65 0.63 Med vs High (1.08 2.89) (0.23 1.19) (0.68 4.45) (0.20 2.15) (0.31 1.27) Values indicate the estimated odds ratios for incidence of disease for the following comparisons among antibody-mediated immune categorizations. Values in parenthesis represent the 95% confidence intervals for the odds ratio estimate. 105

PAGE 106

0 2 4 6 8 10 12Incidence Ketosis % LowMediumHigh AMIR Categorization Primiparous Multiparous Figure 4-4. Incidence ketosis by AM IR categorization within parity. 106

PAGE 107

23000 23500 24000 24500 25000 25500 26000 LowMediumHigh AMIR CategorizationMilk Yield Primiparous Multiparous Figure 4-5. Graph for the effect of antibody-me diated immune response (AMIR) categorization on milk yield. Bars indicate SEM. 107

PAGE 108

0 10 20 30 40 50 60 70Pregnant by 150 DI M LowMediumHigh AMIR Categorization Primiparous Multiparous Figure 4-6. Graph for the effect of antibody-me diated immune response (AMIR) categorization on pregnancy by 150 DIM. 108

PAGE 109

CHAPTER 5 CATEGORIZATION OF PERIPARTURIENT CELL-MEDIATED IMMUNE RESPONSE TO A TEST ANTIGEN AND ITS RELATIONS HIP WITH COMMON DISEASES AND PERFORMANCE MEASURES OF HOLSTEIN DAIRY CATTLE Introduction The increasing risk of disease for Holstein dairy cows (Harmon, 1994; Heringstad, 2000; Emanuelson, 1988) has sparked interest in genetic selection for disease re sistance. Because the immune system is principally responsible for resi sting the array of potential pathogens; many of the methods studied have analyzed the relationshi p between disease risk and immune function or an aspect of a particular part of the immune system. Park et al. (2004) studied the ratio of CD4+ to CD8+ T lymphocyte subsets and its relationshi p with mastitis incidence. Other studies have placed special attention on characteristics of the major hist ocompatability complex (MHC) (Rupp et al., 2007; Park et al., 2004; Aaerestrup et al., 1995). So me studies have used test antigen to quantify an individuals antibody-mediat ed immune responsiveness (AMIR) as a tool to predict the risk of disease (Wagter et al., 2000; Mallard et al., 1997). An inverse relationship between cell-media ted immune response (CMIR) and AMIR has been documented (Rupp et al., 2007; Biozzi et al., 1979; de Vries, 1995). As a result, selection for increased AMIR without re gard for CMIR may confer su sceptibility to intracellular pathogens. Being mediated by T H 1 cells, delayed-type hypersensiti ve reactions (DTH) are largely concerned with elimination of intracellular an tigen. Therefore, DTH reactions have been previously used as a means to quantify CMIR (Mallard et al., 1998; Hernandez et al., 2003, 2005). However, few studies have analyzed the association between this measure of CMIR and disease susceptibility. 109

PAGE 110

The objectives of this research were to categ orize periparturient co ws using DTH reaction to a type 1 antigen and study the association it has with susceptibility to disease, namely; mastitis, metritis, retained fetal membrane (RFM), ketosis, and displaced abomasum (DA).The final objective was to examine for an effect of CMIR categorization on milk yield, somatic cell score (SCS), and fertility. Materials and Methods Research Sample Cows This research study was given IACUC approval. In total, 875 Holstein cows/heifers were enrolled into the study population at 8 weeks (wk-8) prior to exp ected calving. In cows, this was the initiation of the dry period. Animals were en rolled if the expected dry period was not longer than 90 days, if reconfirmed pregnant at enro llment and also if found in good health with no obvious signs of disease. All test animals were from a single herd in north central Florida which maintains exceptional record keeping. All cows and heifers were: enrolled between September 9 th and December 31 st 2004; calved between October 25 th 2004 and March 12 th 2005; and CMIR measurement occurred between November 2 nd 2004 and March 21 st 2005. All cows and heifers received a routine dry off, prefresh, and fresh cow protocol. Animal Removal Criteria Of the originally enrolled 875 cows and he ifers, 13 were removed due to missing data, yielding 862. Interval 1 (Int 1 ) was defined as the time from enro llment (wk-8) to entry into the springer pen (wk-3). Interval 2 (Int 2 ) was defined as the time from entry into the springer pen (wk-3) to calving (wk0) (Figure 41). Due to the restrictions of a parallel study involving AMIR, animals were removed from the study if the pe riod from wk-8 to wk0 (dry period length for cows) was more than 90 days. Add itional animals were excluded if Int 1 or Int 2 was less than 12 days in length. As a result 50 cows were rem oved leaving the study popula tion total at 812 with 110

PAGE 111

362 heifers and 450 cows. For the rest of this pape r heifers will be referred to as primiparous cows. Cows will be referred to as multipar ous cows. Test animals were sired by over 237 different sires, which should provi de adequate variation in the gene pool for a sire effect on disease resistance. Body Condition Scoring Body condition was scored in point increments using the 5-point scale by Ferguson et al. (1994). Body condition scoring (BCS) was grouped into high, medium, and low categories. For cows during wk-8, wk-3, and wk0, a BCS between 3.0 3.75 was coded medium, those below 3.0 were coded low and those above 3.75 were coded high. Heifers at wk-8, wk-3, and wk0 were coded medium if BCS was between 3.0 3.5. A BCS above 3.5 was considered high, and a BCS below 3.0 was considered low. At wk+2, all anim als were coded normal if BCS fell in the range of 2.75 3.5. It has been reported that fo r periparturient dairy cows, it was the loss of body condition, not just BCS alone that was res ponsible for alterations in lymphocyte function (Lacetera et al., 2005; Wentink et al., 1997; Kaneene et al., 1997). To account for this factor the interval change in BCS for various intervals was calculated. Interval 4 was defi ned as the period between wk-3 and wk+2, which serves as an indicator of BCS lost over the transiti on period. Interval 5 was defined as the period between wk-8 and wk0. Inte rval 6 was defined as the period between wk-8 and wk+2. Delayed Type Hypersensitivity To stimulate a DTH reaction, all cows received 0.5 mg C. albicans (raw whole cell material, Greer Laboratories, Le noir, NC, USA) at three times; wk-8, wk-3 and wk0. This was also included with a standard E. coli J5 vaccination program on wk-8 and wk-3 along with the OVA for AMIR. Within 12hours of calvi ng (wk0), animals received 0.5 mg C. albicans in 0.5 111

PAGE 112

mg Quil A (Accurate Chemicals and Scientific Corp., Westbury, NY, USA) adjuvant suspended in 1 mL PBS along with OVA for AMIR. At 1 week post-calving (wk+1), double skin-f old measurements were taken on the right and left skin folds under the base of the tail using a spring-loaded calip er (Harpenden skin-fold caliper, Creative Health Products Inc., Ann Arbor, Michigan, USA). This was performed after raising the tail 90 to a horizont al position and repeated measurements were taken 3 times. The average measurement was then recorded as ei ther the right (R0) or left (L0) time 0 h measurement. The locations of th e two measurements were cleaned with 70% isopropyl alcohol. The right side received an intradermal injection of 0.1 mg candin (C. albicans allergen extract, Greer Laboratories, Inc.) suspended in 0.1 mL PBS. The left tail fold (control side) received 0.1 ml PBS intradermally. All injections were give n with a 28 gauge needle. To identify the exact location of the measurement and injection, paper white-out solution marked the spot. Twentyfour hours later, injection sites were measured again to determ ine the increase in double skinfold thickness for the right side (R24) and left side (L24) as an indicator for the magnitude of the DTH (CMIR) response. Classification of Cell-Mediated Immune Response To obtain a normal distributi on, log transformations of th e DTH measurements were performed. The magnitude of the DTH response was determined by the following (Eq. 5-1): y = ln(R24) ln(R0) (5-1) A repeated measures preliminary analysis using the proc mixed procedure of SAS determined that multiparous cows tended to respond better than primiparous cows (p< 0.064) (Figure 5-1, Figure 5-2). Due to this effect of parity, determ ination of CMIR categorization occurred within parity. 112

PAGE 113

To extrapolate CMIR categorizations the mean and standard deviation for the y values were configured for all cows resp ective of parity. Cows with y values above the mean plus one standard deviation were classi fied as high CMIR responders. T hose cows below one standard deviation less than the mean were classified as low CMIR responders. All animals within one standard deviation of the mean were medium responders. Diseases Identification of disease was performed by fa rm personnel who were blinded to immune response categorizations. Disease information was co llected for; mastitis, metritis, retained fetal membrane, ketosis, and displaced abomasum. All diseases were recorded as yes/no binary responses for the trial peri od of the current lactation. Mastitis was recorded through 365 DIM as light, medium, or severe. Severity was recorded as light if there were no systemic signs and milk was slightly watery with minimal clots (gargot). Severity was medium if there were no systemic signs and a substantial amount of clots detected in the milk. Severity was severe if there were systemic signs, watery milk, and a substantial amount of clots in the milk. Metritis was recorded through 30 DIM as light, medium or severe. Severity was light if there was an abnormal vaginal discharge and a pa lpable uterine lumen. Severity was medium if there was a purulent vaginal discharge, with an enlarged, not-flaccid uterus. Severity was severe if there was a purulent foul-smelling vaginal di scharge, with an en larged flaccid uterus. Association between energy-re lated metabolic conditions an d CMIR categorization were also considered. Ketosis was recorded through 30 DIM as light, medium, or severe. Ketosis was coded as light if the urine contained 15 mg/dL of ketone bodies. Severity was medium if the urine contained 40 mg/dL of ketone bodies. The se verity was severe if the urine contained > 80 mg/dL of ketone bodies. Displaced abomas um was recorded through 16, and 50 DIM. 113

PAGE 114

Due to the involvement of the immune system in determining the expulsion of fetal membranes (Kimura et al., 2002), this study also analyzed the association between RFM and CMIR categorization. Retained feta l membranes were identified if still retained 24 h postpartum. Milk Yield, Somatic Cell Score, and Reproductive Efficiency Milk yield data was gathered from the Da iry Herd Information Association (DHIA) records for the current lactati on using ME305 which is an estim ate of the milk yield for 305 DIM. For categorical data, low milk producers were identified if there 305 day milk was in the bottom 25% of the study group. High milk producer s were those in the top 25% of the study group. The remaining middle 50% were clas sified as medium milk producers. Somatic cell score (SCS) was obtained from DHIA records. The average SCS was determined for the first 3 test days, the first 6 te st days, and for the first 10 test days, which are an indication of the average SCS for the first 90, 180, and 300 DIM respectively. To test for associations between CMIR categor ization and reproductive efficiency, a binary pregnancy term was used which simply indicated if a given cow was pregnant by 150 DIM. The number of number of days not pregnant (days open) and number of times bred were used as quantitative variables. Statistics To test for an association between CMIR categorization and the binary terms of specific disease risk and reproductive efficiency, a l ogistic regression model was used using the LOGISTIC procedure of SAS. Other than the main effect (CMIR categorization), possible explanatory variables include: BCS x = categorical effect of body condition score for sampling week x. BCS y = quantitative variable, change in body condition score over interval y. Dex = binary effect, whether cow recei ved dexamethasone prior to calving. Parity = binary effect, either primiparous or multiparous. 114

PAGE 115

RFM = binary effect, incidence of retained fetal membrane. CDiff = binary effect, if reported difficult calving. Ketosis = binary effect, incidence of ketosis. SCSavg = somatic cell score average over 10 monthly test days. For analyzing the relationship between CMIR categorization and the quantitative variables for SCS, 305 day milk yield, and reproductive efficiency, linear regression was used with the REG procedure of SAS. All categor ical variables were analyzed using logistic regression with the LOGISTIC procedure of SAS. All relevant effects were put in the model. A backword elim ination procedure was used to determine the final model. Explanatory variable s remained in the model if they showed a tendency (p < 0.10) to predict the outcome variab le. Statistical significance was determined by setting = 0.05. Results Mastitis Incidence of all types of mastitis within 100 DIM was not significantly associated with CMIR status. However, when only medium and severe cases of mastitis within 365 DIM were considered (138 cases in multiparous cows and 42 cases in primiparous cows), CMIR status was significantly associated with mastitis occurrence (Type 3 analysis p = 0.041). Parity and SCSavg were also significant variables in the model. Those cows categori zed as medium responders were 2.14 (CI = 1.13 4.08) times more likely to develop a medium or severe case of mastitis than high responders. When considering only multip arous cows, those categorized as low and medium responders were 2.80 (CI = 1.29 6.09) times more likely to develop a medium or severe case of mastitis than high responders. 115

PAGE 116

Retained Fetal Membrane There were 61 recorded cases of RFM (7.5 %). Of these 61, primiparous cows accounted for 22 while multiparous cows accounted fo r 39. Contributing model effects were:BCS 0 and BCS 5 The CMIR categorization was a significant predic tor for the risk of RFM (p = 0.0001) (Table 5-1). Low cell-mediated immune res ponders were 6.68 (CI = 1.87 23.84) times more likely to have a case of RFM than high i mmune responders. Also, if only considering multiparous cows, low cell-mediated immune responders were 26.52 (2.30 306.11) times more likely to have an RFM than high cell-mediated immune responders (Figure 5-3). Metritis For this analysis there were 43 cases of light, medium or se vere metritis within 30 DIM (5.3%). Primiparous cows accounted for 29 cases while multiparous cows accounted for 14. The remaining model contributing effects were: BCS 0 BCS 5 Dex, Parity, and RFM. Although not significantly different, the medium immune re sponders were 7.40 (CI = 0.91 60.25) times more likely to develop metritis than hi gh immune responders (Table 5-1). Ketosis and Displaced Abomasum There were 48 recorded incidences of light medium, or severe ketosis within 30 DIM (6.9%). Twenty-seven cases were observed in primiparous cows and 21 cases in multiparous cows. Remaining model effects were Dex and BCS 6 The categorization for CMIR was not a significant predictor for risk of ketosis (p = 0.97) (Table 5-1). There were only 21 recorded incidences of DA, 13 in primiparous and 8 in multiparous cows. Contributing model effects were: Ketosis, BCS 4 and Dex. The categorization for CMIR was not a significant predictor for incidence of DA (Table 5-1). 116

PAGE 117

Milk Yield, Somatic Cell Score and Reproductive Efficiency For the analysis of the effect of CMIR cate gorization on milk yield, the contributing model effects were parity, days open, and the binary variable mastitis. In this analysis, CMIR categorization was a significant pr edictor of milk yield (p = 0.049, = 508.08) (Figure 5-4). The analysis of the effect of CMIR on SCS included: BCS 6 milk categorization, and a binary mastitis variable. The e ffect of CMIR on SCS was not a si gnificant predictor of SCS (p = 0.83). The model analyzing the effect of AMIR category on pregnancy by 150 DIM, included BCS 4 and milk categorization. The effect of CMIR was not a signif icant predictor for pregnancy by 150 DIM (p = 0.77). Discussion The significant association be tween peripartum DTH response and mastitis identifies the role CMIR has on resisting infection in the ma mmary gland and the validity of using DTH to Candida albicans as a measure of CMIR. Although not sign ificant, the estimated odds of metritis show promise for the ability of DTH to predict metritis infection (Table 5-1). The results for RFM were in agreement with previous literature (Kimura et al., 2002). Although in the current study, because categorizat ion of CMIR occurred early postpartum, one could speculate that RFM served to suppress immune response. However, in Kimura et al (2002), neutrophil activity was suppressed 15 days prepartum in cows with RFM. The results for the associati on between milk yield and CMIR categorization prove that selection for improved CMIR does not predispose cows to lower milk production. In this study the significant effect of CMIR categorizati on on milk yield followed the pattern of High > Medium > Low. Because selection for improve d immune response should include both CMIR and AMIR, the potential negative association be tween AMIR and milk yield should balance out 117

PAGE 118

due to the positive effect of CMIR on milk yield. Previous studies in dairy cows have not studied the association between CMIR cat egorization and milk yield. Mallard et al. (1998) categorized CMIR in pigs and found greater growth rates in pigs with superior im mune responsiveness. Selection for decreased somatic cell score (SCS), increased productive life (PL), and improved structural traits of the udder, are all methods currently used to reduce incidence of disease. These methods, however, are largely based on fairly crude biologi cal associations with disease risk and due not addre ss the immune system which is pr incipally responsible for host disease resistance. These traits are also not c oncerned with broad-based resistance to disease. Selection for improved immune responsiveness as a means to reduce the risk of disease should take a broad based approach. This is primarily due to the vast array of potential pathogens and the various virulence mechanisms employed to in itiate disease. This broad based philosophy also becomes critically important due to th e complexity of the immune system in vivo Selection for specific improvements may prove inadequate, but may also confer unexpected disease susceptibility to additional varieties of pathoge ns. Inverse relationships between branches of adaptive immunity shed light on th e potential for this to occur (de Vries, 1995; Biozzi et al., 1979; Rupp et al., 2007). If there is one thing that should be le arned from the past 5 decades of applying selection pressure, it is that genetic selection should not focus or put too much pressure on specific traits. When this happens it is inev itable that unexpected unfavorable trends occur where selection pressure is neglected. 118

PAGE 119

5.0 5.5 6.0 6.5 7.0 7.5 0hrs 24hrs 0hrs 24hrs Primiparous Multiparous Double Skin-Fold (mm ) Figure 5-1. Graph showing in crease in double skin-fold thic kness respective of parity (multiparous or primiparous).Bars reflect SEM. 119

PAGE 120

0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 Primiparous Multiparous ParityDouble Skin-Fold Change (m m Figure 5-2. Graph revealing parity difference for cell-mediated immune response. Bars indicate SEM. 120

PAGE 121

Table 5-1. Odds ratios of disease inciden ce for cell-mediated immune categorizations. Mastitis Metritis Ketosis DA RFM 0.99 1.06 1.08 1.02 2.85 Low vs Med & High (0.64 1.53) (0.39 2.92) (0.56 2.06) (0.30 3.52) (1.67 4.86) 1.45 2.97 1.14 1.56 6.68 Low vs High (0.63 3.31) (0.29 30.51) (0.33 3.95) (0.12 20.21) (1.87 23.84) 2.14 7.4 1.05 2.28 1.94 Med vs High (1.13 4.08) (0.91 60.25) (0.40 2.78) (0.27 19.28) (0.58 6.45) Values indicate the estimated odds ratios for incidence of disease for the following comparisons among cell-mediated immune categorizations. Values in parenthesis represent the 95% confidence intervals for the odds ratio estimate. 121

PAGE 122

0 5 10 15 20Incidence RFM % LowMediumHigh CMIR Categorization Primiparous Multiparous Figure 5-3. Graph indicating the difference in risk of RFM betw een high and low cell-mediated immune response categorization. 122

PAGE 123

10200 10400 10600 10800 11000 11200 11400 11600 11800 12000 LowMediumHigh CMIR CategorizationMilk Yield (kg) Primiparous Multiparous Figure 5-4. Graph for the effect of cell-medi ated immune response (CMIR) categorization on milk yield. Bars indicate SEM. 123

PAGE 124

LIST OF REFERENCES Aarestrup, F. M., N. E. Jensen, and H. Osterg ard. 1995. Analysis of asso ciations between major histocompatibility complex (BoLA) class1 ha plotypes and subclinical mastitis of dairy cows. J. Dairy Sci. 78:1684-1692. Adewuyi, A. A., E. Gruys, and F. J. C. M. van Eerdenburg. 2005. Non esterified fatty acids (NEFA) in dairy cattle. A review Veterinary Quarterly 27(3):117-126. Badolato, R., H. M. Bond, G. Valerio, A. Petrella, G. Morrone, M. J. Waters, S. Venuta, and A. Tenore. 1994. Differential expression of su rface membrane growth hormone receptor on human peripheral blood lymphoc ytes detected by dual fluorochrome flow cytometry. J. Clin. Endocrinol. Metab. 79:984-990. Baird, G.D. 1982. Primary ketosis in the high-producing dairy cow: Clinical and subclinical disorders, treatment, prevention, an d outlook. J. Dairy Sci. 65(1):1-10. Baus, E., F. Andris, P. M. Dubois, J. Urbain, and O. Leo. 1996. Dexamethasone inhibits the early steps of antigen receptor signaling in ac tivated T lymphocytes. J. Immunol. 156:45554561. Beardsley, G. L., L. D. Muller, H. A. Garverick, F. C. Ludens, and W. L. Tucker. 1976. Initiation of parturition in dairy cows with dexame thasone. II. Response to dexamethasone in combination with estradiol benzoa te. J. Dairy Sci. 59(2):241-247. Besedovsky, H. O., and A. del Rey. 1996. Immune-n euro-endocrine interactions: Facts and hypothesis. Endocr. Rev. 17:64-102. Beutler, B., and E. T. Rietschel. 2003. Innate immune sensing and its roots: The story of endotoxin. Nat. Rev. Immunol. 3:169-176. Biozzi, G., D. Mouton, O. A. Sant Anna. H. C. Passos, M. Gennari M. H. Reis, V. C. Ferreira, A. M. Heumann, Y. Bouthillier, O. M. Ib anez, C. Stiffel, and M. Siqueira. 1979. Genetics of immunoresponsiveness to natura l antigens in the mouse. Curr. Top. Microbiol Immunol 85:31-98. Black, A. C. 1999. Delayed-type hypersensitivity: Cu rrent theories with a historic perspective. Dermatol. Online J. 5(1): 7-27. Blalock, J.E. 1994. The syntax of immune-neuroendocrine communication. Immunol. Today 15:504-511. Burton, J. L., and M. E. Kehrli, jr. 1995. Regul ation of neutrophil a dhesion molecules, and shedding of Staphylococcus aureus in milk of cortisoland dexamethasone-treated cows. Am. J. Vet. Res. 56:997-1006. 124

PAGE 125

Burton, J. L., and M. E. Kehrli, jr. 1996. Eff ects of dexamethasone on bovine circulating T lymphocyte populations. J. Leukocyte Biology 59:90-99. Burton, J. L., E. B. Burnside, B. W. Kenne dy, B. N. Wike, and J. H. Burton. 1989. Antibody responses to human erythrocytes and ovalbumin as marker tra its of disease resistance in dairy calves. J. Dairy Sci. 72:1252-1265. Burton, J. L., M. E. Kehrli, jr., S. Kapil, and R. L. Horst. 1995. Regulation of L-selectin and CD18 on bovine neutrophils by glucocorticoids: Effects of cortisol and dexamethasone. J. leukoc. Biol. 57:317-325. Castillo-Juarez, H., P. A. Oltenacu, R. W. Blak e, C. E. Mcculloch, and E. G. Cienfuegos-Rivas. 2000. Effect of herd environment on the ge netic and phenotypic relationships among milk yield, conception rate, and somatic cell sc ore in holstein cattle. J. Dairy Sci. 83:807814. Chertov, O., D. Yang, O. M. Howard, and J. J. Oppenheim. 2000. Leukocyte granule proteins mobilize innate host defenses and adaptiv e immune responses. Immunol. Rev. 177: 6878. Cole, R. K. 1968. Studies on gene tic resistance to marek's dis ease. Avian Diseases. 12:9-28. Corbeil, L. B., and R. H. BonDurant. 2001. Imm unity to bovine reproductive infections. Vet. Clin. of North Am. Food Anim. Pract. 17(3):567-83. Curtis, C. R., H. N. Erb, C. J. Sniffen, R. D. Smith, and D. S. Kronfeld. 1985. Path analysisof dry period nutrition, postpartum meta bolic and reproductive disorders, and mastitis in holstein cows. J. Dairy Sci. 68:2347-2360. Dardenne, M., and W. Savino 1996. Interdependen ce of the endocrine and immune systems. Adv. Neuroimmunol. 6:297-307. Davis, M. M., J. J. Boniface, Z. Reich, D. Lyons, J. Hampl, B. Arden, and Y. Chien. 1998. Ligand recognition by alphabeta T cell receptors. Annu. Rev. Immunol. 16:523-44. de Haas, Y., H. W. Barkema and R. F. V eerkamp. 2002. Genetic parameters for pathogenspecific incidence of clinical mastitis in dairy cows. Animal Sci. 74:233-242. de Vries, J. E. 1995. Immunosuppressive and an ti-inflammatory properties of interleukin 10. Ann. Med. 27(5):537-41. Detilleux, J. C., M. E. Kehrli, Jr., Stabel, A. E. Freeman, and D. H. Kelley. 1995. Study of immunological dysfunction in peri parturient holstein cattle selected for high and average milk production. Vet. Imm un. and Immunopathology 44:251-267. 125

PAGE 126

Edfors-Lilia, I., and P. Wallgreen. 2000. Escheric hia coli and Salmonella diarrhea in pigs. Breeding for disease resistance in farm animal s. 2nd Edn. Axford, R. F. E., S. C. Bishop, F. W. Nicholas and J. B. Owen. Walli ngford: CAB International. Pp 253-267. Emanuelson, U., B. Danell, and J. Phillipson. 19 88. Genetic parameters for clinical mastitis, somatic cell counts, and milk production estim ated by multiple-trait restricted maximum likelihood. J. Dairy Sci. 71:467-476. Ferguson, J. D., D. T. Galligan, N. Thomsen. 1994. Principle descriptors of body condition score in Holstein cows. J. Dairy Sci. 77(9):2695-2703 Franklin, S. T., J. W. Young, and B. J. Nonnecke. 1991. Effects of ketones, acetate, butyrate, and glucose on bovine lymphocyte prolif eration. J. Dairy Sci. 74:2507-2514. Gatti, E., and P. Pierre. 2003. Understanding the cell biology of antigen presentation: The dendritic cell contribution. Curr. Opin. in Cell Biol. 15:468-473. Goff, J. P., and K. Kimura. 2002. Effect of mastect omy on milk fever, energy, and vitamins A, E, and -carotene status at parturit ion. J. Dairy Sci. 85:1427-1436. Goff, J. P., and R. L. Horst. 1997. Physiological changes at parturiti on and relationship to metabolic disoders. J. Dairy Sci. 80:1260-1268. Grafton, G., and L. Thwaite. 2001. Calcium ch annels in lymphocytes. Immunology 104:119-126. Hansen, P. J. 1997. Interactions between the immune system and the bovine conceptus. Theriogenology 47:121-130 Harmon, R. J. 1994. Physiology of mastitis and f actors affecting milk somatic cell counts. J. Dairy Sci. 77:2103-2112. Heller, E. D., G. Leitner, A. Friedman, Z. Uni, M. Gutman, and A. Cahaner 1992. Immunological parameters in meat-type chic ken lines divergently selected by antibody response to Escherichia coli vaccinati on. Vet Immunol Immunopathol. 34:159-172. Heringstad, B., D. Gianola, Y. M. Chang, J. Odegard, and G. Klemetsdal. 2006. Genetic associations between clinical mastitis and so matic cell score in early first-lactation cows. J. Dairy Sci. 89: 2236-2244. Heringstad, B., G. Klemetsdal, and J. Ruane. 2000. Selection for mastitis resistance in dairy cattle: A review with focus on the situation in the nordic countries. Livestock Prod. Sci. 64: 95-106. Heringstad, B., Y. M. Chang, D. Gianola, and G. Klemetsdal. 2004. Multivariate threshold model analysis of clinical mastitis in multip arous norweigan dairy cattle. J. Dairy Sci. 87:3038-3046. 126

PAGE 127

Hernandez, A., J. A. Yager, B. N. Wilkie, K. E. Leslie, and B. A. Mallard. 2005. Evaluation of bovine cutaneous delayed-type hypersensitiv ity (DTH) to various test antigens and a mitogen using several adjuvants. Ve t Immun. And Immuno pathology 104:45-58. Hernandez, A., N. Karrow, and B. A. Mallard. 2003. Evaluation of immune responses of cattle as a means to identify high or low res ponders and use of a human microarray to differentiate gene expression. Genet. Sel. Evol. 35:S67-81. Hoeben, D., R. Heyneman, and C. Bu rvenich. 1997. Elevated levels of -hydroxybutyric acid in periparturient cows and in vi tro effect on respiratory burst activity of bovine neutrophils. Vet Immun. And Immunopathology 58:165-170. Hunter, N., L. Moore, B. D. Hosie, W. S. Di ngwall, and A. Greig. 1997. Association between natural scrapie and PrP genotype in a flock of suffolk sheep in Scotland. Veterinary Record 140:59-63. J. C. Detilleux. 2002. Genetic factors affecting sus ceptibility of dairy cows to udder pathogens. Vet. Immun. and Im munopathology 88:103-110. Jacysyn, J. F., I. H. Abrahamsohn, and M. S. Macedo. 2001. Modulation of delayed-type hypersensitivity during the time course of immune response to a protein antigen. Immunology 102: 373-379. Jensen, P. E. 2007. Recent advances in antige n processing and presentation. Nature Immunol. 8(10):1041-1048. Kadarmideen, H. N., R. Thompson, and G. Simm. 2000. Linear and threshold genetic parameters for disease fertility and milk production in dairy cattle. J. Anim. Sci. 71:411-419. Kaneene, J. B., R. A. Miller, T. H. Herdt, and J. C. Gardiner. 1997. The association of serum nonesterified fatty acids and cholesterol, management and feeding practices with peripartum disease in dairy cows. Prev. Vet. Med. 31:59-72. Kean, R. P., A. Cahaner, A. E. Freeman, S. J. Lamont. 1994. Direct and correlated responses to multitrait, divergent selection for im munocompetence. Poult. Sci. 73:18-32. Kehrli, M. E., and D. E. Shuster. 1994. Factors affecting milk somatic cells and their role in health of the bovine mammary gland. J. Dairy Sci. 77:619-627. Kehrli, M. E., and J. A. Harp. 2001. Immunity in the mammary gland. Vet. Clin. Of North Am Food Anim Pract 17(3):495-516. Kehrli, M.E. jr., F.C. Schmalstieg, D.C. Anderson, M. J. Van der Maaten, B. J. Hughes, M. R. Ackermann, C. L. Wilhelmse n, G. B. Brown, M. G. Steven s, and C. A. Whetstone 1990. Molecular definition of the bovine granul ocytopathy syndrome: Identification of deficiency of the Mac-1 (CD11b/CD18) glycoprotein. Am. J. Vet. Res. 51:1826-1836. 127

PAGE 128

Kelm, S. C., A. E. Freeman, and M. E. Kehrli, jr. 2001. Genetic control of disease resistance and immunoresponsiveness. Vet. Clin. Of North Am. Food Anim. Pract. 17(3):477-93. Kelm, S. C., J. C. Detilleux, and A. E. Freema n. 1997. Genetic association between parameters of innate immunity and measures of mastitis in periparturient holstein cattle. J. Dairy Sci. 80:1767-1775. Kimura, K., J. P. Goff, M. E. Kehrli, jr., and T. A. Reinhardt. 2002. Decreased neutrophil function as a cause of retained placenta in dairy cattle. J. Dairy Sci. 85:544-550. Kimura, K., T. A. Reinhardt, and J. P. Goff 2006. Parturition and hypocalcemia blunts calcium signals in immune cells of dair y cattle. J. Dairy Sci. 89:2588-2595. Klasing, K. C. 1998. Nutritional modulation of resi stance to infectious di sease. Poult. Sci. 77:1119-1125. Klasing, K. C., L. Lauring, R. Peng, and M. Fry. 1987. Immunologically mediated growth depression in chicks: Influence of feed intake corticosterone, and interleukin-1. J. Nutr. 117:1629-1637. Kremer, W. D. J., E. N. Noordhui zen-Stassen, F. J. Grommers, A. J. J. M. Daemen, P. A. J. Hendricks, and A. Brand. 1993. Preinfec tion chemotactic response of blood polymorphonuclear leukocytes to predict sever ity of Escherichia coli mastitis. J. Dairy Sci. 76:1568-1574. Kushibiki, S., K. Hodate, H. Shingu, Y. Obara, E. Touno, M. Shinoda, and Y. Yokomizo. 2003. Metabolic and lactational re sponses during recombinant bov ine tumor necrosis factortreatment in lactating co ws. J. Dairy Sci. 86:819-827. Lacetera, N., D. Scalia, O. Franci, U. Bern abucci, B. Ronchi, and A. Nardone. 2004. Short comm: Effects of nonesterified fatty acids on lymphocyte function in dairy heifers. J. Dairy Sci. 87:1012-1014. Lacetera, N., D. Scalia, U. Bernabucci, B. Ronchi, D. Pirazzi, and A. Nardone. 2005. Lymphocyte functions in overconditioned cows around parturition. J. Dairy Sci. 88:20102016. Lacetera, N., U. Bernabucci, B. Ronchi, and A. Nardone. 2001. Effects of subclinical pregnancy toxemia on immune responses in sheep. Am. J. Vet. Res. 62:1020-1024. Lewis, R. S. 2001. Calcium signaling mechan isms in T lymphocytes. Annu. Rev. Immunol. 19:497-521. Mailhot, G., J. L. Petit, C. Demers, and M. Ga scon-Barre. 2000. Influence of the in vivo calcium status on cellular calcium homeostasis and the level of the calcium-binding protein calreticulin in rat hepatocytes. Endocrinology 141:891-900. 128

PAGE 129

Mallard, B. A., B. N. Wilkie, B. W. Ke nnedy, J. Gibson, and M. Quinton. 1998. Immune responsiveness in swine: Eight generati ons of selection for high and low immune response in yorkshire pigs. Proceedings of the 6th World Congress on Genetics Applied to Livestock Production, 11-16 January Vol .27, University of New England, Armidale, pp. 257-262. Mallard, B. A., J. C. Dekkers, M. J. Ireland, K. E. Leslie, S. Sharif, C. Lacey Vankampen, L. Wagter, and B. N. Wilke. 1998. Alterati on in immune responsiveness during the peripartum period and its ramification on dair y cow and calf health. J. Dairy Sci. 81:585595. Mallard, B. A., L. C. Wagter, M. J. Ireland, and J. C. M. Dekkers. 1997. Effect of growth hormone, insulin-like growth factor-1, and cortisol on pe riparturient antibody response profiles of dairy cattle. Vet. Immunol. Immunopathol. 60:61-68. Meglia, G. E., A. Johannison, S. Agenas, K. Holtenius, and K. P. Waller. 2005. Effects of feeding intensity during the dry period on leukocyte and lymphocyte sub-populations, neutrophil function and health in peripartur ient dairy cows. The Vet. Journal 169:376384. Nagahata, H., A. Ogawa, Y. Sanada, H. Noda and S. Yamamoto. 1992. Peripartum changes in antibody producing capability of lymphocytes from dairy cows. Vet. Quart. 14:(1): 3940. Nash, D. L., G. W. Rogers, J. B. Cooper, G. L. Hargrove, J. F. Keown, and L. B. Hansen. 2000. Heritability of clinical mastitis incidence a nd relationships with sire transmitting abilities for somatic cell score, udder t ype traits, productive life, and protein yield. J. Dairy Sci. 83:2350-2360. National Mastitis Council. 1996. Cu rrent concepts of bovine mastitis. 4th ed. Natl. Mastitis Counc. Inc., Madison, WI. Neerhof, H. J., P. Madsen, V.P. Ducrocq, A. R. Vollema, J. Jensen, and I. R. Korsgaard. 2000. Relationships between mastitis and functional longevity in danish black and white dairy cattle estimated using survival an alysis. J. Dairy Sci. 83:1064-1071. Nonnecke, B. J., K. Kimura, J. P. Goff, J. Ma rcus, and M. E. Kehrli, jr. 2003. Effects of the mammary gland on functional capacities of blood mononuclear leukocyte populations from periparturient cows. J. Dairy Sci. 86:2359-2368. Nonnecke, B. J., S. T. Franklin, and J. W. Young. 1992. Effects of ketones, acetate, and glucose on In Vitro immunoglobulin se cretion by bovine lymphocytes J. Dairy Sci. 75:982-990. 129

PAGE 130

O'Neill, R. G., J. A. Woolliams, E. J. Gla ss, J. L. Williams, and J. L. Fitzpatrick. 2006. Quantitative evaluation of genetic and environmental parameters determining antibody response induced by vaccination against bovine respiratory syncytial virus. Vaccine 24: 4007-4016. Park, Y. H., L. K. Fox, M. J. Hamilton, and W. C. Davis. 1992. Bovine mononuclear leukocyte subpopulations in peripheral blood and mammary gland secret ions during lactation. J. Dairy Sci. 75:998-1006. Park, Y. H., L. K. Fox, M. J. Hamilton, and W. C. Davis. 1993. Suppression of proliferative response of BoCD4+ T lymphocytes by ac tivated BoCD8+ T lymphocytes in the mammary gland of cows with Staphylococcus aureus mastitis. Vet Immunol. Immunopathol. 36:137-151. Park, Y. H., Y. S. Joo, J. Y. Park, J. S. Moon, S. H. Kim, N. H. Kwon, J. S. Ahn, W. C. Davis, and C. J. Davies. 2004. Characterization of lymphocyte subpopulations and major histocompatibility complex haplotypes of mas titis-resistant and susceptible cows. J. Vet. Sci. 5:29-39. Parker, D. C. 1993. T cell-dependent B ce ll activation. Annu. Rev. Immunol. 11:331-60. Partiseti, M., F. L. Deist, C. Hivroz, A. Fisher, H. Korn, and D. Choquet. 1994. The calcium current activated by T cell receptor and store de pletion in human lymphocytes is absent in a primary immunodeficiency. J. Biol. Chem. 269:32327-32335. Peters, A.R., and D.A. Poole. 1992. Induction of pa rturition in dairy cows with dexamethasone. The Veterinary Record 131(25-26):576-578. Piccinini, R., C. Bronzo, P. Moroni, C. Luzza go, and A. Zecconi. 1999. Study on the relationship between milk immune factors and Staphyloco ccus aureus intramammary infections in dairy cows. J. Dairy Res. 66:501-510. Risso, A. 2000. Leukocyte antimicrobial peptides: multifunctional effector molecules of innate immunity. J. Leukocyte Biology 68:785-792. Rogers, G. W., G. Banos, U. S. Nielson, a nd J. Philipsson. 1998. Genetic correlations among somatic cell scores, productive life, and type traits from the United States and udder health measures from Denmark and Sweden. J. Dairy Sci. 81:1445-1453. Rupp, R., A. Hernandez, and B. A. Mallard. 2007. Association of bovi ne leukocyte antigen (BoLA) DRB3.2 with immune response mastit is, and production and type traits in Canadian holsteins. J. Dairy Sci. 90:1029-1038. Rupp, R., and D. Boichard. 2003. Genetics of resistan ce to mastitis in dairy cattle. Vet. Res. 34:671-688. 130

PAGE 131

Saad, A. M., C. Concha, and G. Astrom. 1989. Alterations in neutr ophil phagocytosis and lymphocyte blastogenesis in dairy cows around parturition. J. Vet. Med. 36(5): 337-45. Savina, A., and S. Amigorena. 2007. Phagocytosis and antigen presentation in dendritic cells. Immunol. Rev. 219: 143-56. Schukken, Y. H., H. N. Erb, and D. Smith. 1988. Th e relationship between mastitis and retained placenta in a commercial population of holst ein dairy cows. Preventive Veterinary Medicine 5:181-190. Schukken, Y. H., K. E. Leslie, D. A. Barnum, B. A. Mallard, J. H. Lumsden, P. C. Dick, G. H. Vessie, and M. E. Kehrli, Jr. 1999. Experime ntal Staphylococcus aureus intramammary challenge in late lactation dairy cows. Quarter and cow effects determining the probability of infection. J. Dairy Sci. 82:2393-2401. Shook, G. E., and M. M. Schutz. 1994. Selection on somatic cell score to improve resistance to mastitis in the United States. J. Dairy Sci. 77:648-658. Silvia, W. 2003. Addressing the decline in reprod uctive performance of lactating dairy cows: A researcher's perspective. Vet. Sci TommorrowVol. 3 May 2003. Soler, J. J., L. de Neve, T. Perez-Contreras, M. Soler, and G. Sorci. 2003. Trade-off between immunocompetence and growth in magpies: An experimental study. Proc. R. Soc. Lond. B 270:241-248. Sordillo, L. M., M. Campos, and L. A. Babi uk. 1991. Antibacterial activity of bovine mammary gland lymphocytes following treatment with interleukin-2. J. Dairy Sci. 74:3370-3375. Stear, M. J., S. C. Bishop, B. A. Mallard, and H. Raadsma. 2001. Rev: The sustainability, feasibility and desireability of breeding livesto ck for disease resistance. Research in Vet. Sci. 71:1-7. Steiger, M., M. Senn, G. Altreu ther, D. Werling, F. Sutter, M. Kreuzer, and W. Langhans. 1999. Effect of prolonged low-dose lipopolysaccharid e infusion on feed intake and metabolism in heifers. J. Anim. Sci. 77:2523-2532. Steine, T. 1996. Avlsarbeid og mastitt. Buskap 2: 8-11. (In Norweigian). Strandberg, E., and G. E. Shook. 1989. Genetic and economic responses to breeding programs that consider mastitis. J. Dairy Sci. 72:2136-2142. Suriyasasathaporn, W., C. Heuer, E. N. Noordhuizen-Stassen, and Y. H. Schukken. 2000. Hyperketonemia and the impairment of udde r defense: A review. Vet. Res. 31:397-412. Swanson, L. V., and A. G. Hunter. 1969. Egg yolk anti gens and their effect on fertility in rabbits. Biology of Reproduction 1: 324-329. 131

PAGE 132

Uribe, H. A., B. W. Kennedy, S. W. Martin, and D. F. Kelton. 1995. Genetic parameters for common health disorders of holstein cows. J. Dairy Sci. 78:421-430. van der Poll, T. 2001. Coagulation and infl ammation. J. Endotoxin Research 7:301-304. Van Kampen, C., and B. A Mallard. 1997. Effects of peripartum stress and health on circulating bovine lymphocyte subsets. Vet Imm unol Immunopathol. 59(1-2): 79-91. Van Werven, T., E. N. Noordhuizen-Stassen, A. J. J. M. Daemen, Y. H. Schukken, A. Brand, and C. Burvenich. 1997. Preinfection In Vitr o cemotaxis, phagocytosis, oxidative burst, and expression of CD11/CD18 receptors and th eir predictive capacity on the outcome of mastitis induced in dairy cows with Es cherichia coli. J. Dairy Sci. 80:67-74. von Ruecker, A., and I. G. H. Schmidt-Wolf. 2000. Strategies to evaluate metabolic stress and catabolism by means of immunological vari ables. Clinical Nutrition 19(3):147-156. Wagter, L. C., B. A. Mallard, B. N. Wilke, K. E. Leslie, P. J. Boettchart, and J. C. M. Dekkers. 2000. A quantitative approach to classi fying holstein cows based on antibody responsiveness and its relationship to peripartum mastitis occurrence. J. Dairy Sci. 83:488-498. Wagter, L. C., B. A. Mallard, B. N. Wilke, K. E. Leslie, P. J. Boettcher, and J. C. M. Dekkers. 2003. Relationship between milk production a nd antibody response to ovalbumin during the peripartum period. J. Dairy Sci. 86:169-173. Waldron, M. R., A. E. Kulick, A. W. Bell, and T. R. Overton. 2006. Acute experimental mastitis is not causal toward the development of energy-related metabolic disorders in early postpartum dairy cows. J. Dairy Sci. 89:596-610. Wedlock, D. N., F. E. Aldwell, D. M. Collins, G. W. de Lisle, T. Wilson, and B.M. Buddle. 1999. Immune responses induced in cattle by virulent and attenuated Mycobacterium bovis strains: correlation of de layed-type hypersensitivity with ability of strains to grow in macrophages. Infect. Immun. 67(5): 2172-2177. Weigel, K. A., A. E. Freeman, M. E. Kehrli, jr., J. R. Thurston, and D. H. Kelley. 1992. Relationship of In Vitro immune function with health and production in holstein cattle. J. Dairy Sci. 75:1672-1679. Weller, J. I., and E. Ezra. 1997. Genetic analysis of somatic cell score and female fertility of Israeli holsteins with an individual animal model. J. Dairy Sci. 80:586-593. Wentink, G. H., V. P. M. Rutten, T. S. van den Ingh, A. Hoek, K. E. M ller, and T. Wensing. 1997. Impaired specific immunoreactivity in cows with hepatic lipidos is. Vet. Immunol. Immunopathol. 56:77-83. Wild, D., ed. 2001. The immunoassay handbook. Nature Publishing Group, NY. 132

PAGE 133

Wilkie, B.N., B. A. Mallard, M. Quinton, and J. Gibson. 1998. Multi-trait selection for immune response: a possible alternativ e strategy for enhanced livesto ck health and productivity. Prog. Pig Sci. 29-38. Wilson, D. J., B. A. Mallard, J. L. Burton, Y. H. Schukken, and Y. T. Grohn. 2007. Milk and serum J5-specific antibody responses, milk production change, and clinical effects following intramammary E. coli challenge for J5 vaccinate and control cows. Clinical and Vaccine Immun. 14(6):693-699. Woolaston, R. R., and R. L. Baker. 1996. Prospects of breeding small ruminants for resistance to internal parasites. Internat J. of Parasitology 26:845-855. Yang, D., K. Nakada-Tsuki, M. Ohtani, R. Go to, T. Yoshimura, Y. Kobayashi, and N. Watanabe. 2001. Identification and cloning of genes associated with the guinea pig skin delayed-type hypersensitivity reaction. J. Biochem. 129:561-568. 133

PAGE 134

BIOGRAPHICAL SKETCH Jason De La Paz was born in Tampa, Florida in 1977. He realized his in terest in animals at an early age, and although never growing up on a farm he discovered a fascination for dairy cows while working toward an animal science degree at the University of Florida. During the four years of his undergraduate college education, he held a part-time position working as a veterinary technician. Jason rece ived his bachelors degree in 2001 and soon thereafter moved to Minnesota after accepting a position as a reprodu ctive specialist for ABS Global, which is a bovine genetics company. In this position, he ma naged the reproductive concerns for several large farms in north central Minnesota. With family, warmer climate and saltwater fishing awaiting him back in Florida, he took a posit ion with ABS Global which allowed him to move back. After working this position for a few year s, Jason began pursuing a Master of Science degree at the University of Florida college of Veterinary Medicine. With Dr. Arthur Donovan as chair of Jasons supervisory committee, Jason rece ived his Master of Science degree in August 2008 where he studied how to determine the immune response potential of individual Holstein dairy cows and how this is a ssociated with disease risk. For the years to come, Jason intends on con tinuing his focus on disease resistance through genetic selection for increased immune responsiveness. He has been married since 2002 to Amy, the woman he dated since high school. They now resi de in Ocala, Florida with their two-year-old daughter Emily. His interests include gator football, fishing, and tennis. 134



PAGE 1

USING ANTIBODY AND CELL-ME DIATED IMMUNE RESPONSE TO TEST ANTIGENS IN PERIPARTURIENT DAIRY COWS AS A MEASURE OF DISEASE RESISTANCE By JASON MICHAEL DE LA PAZ A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2008 1

PAGE 2

2008 Jason Michael De La Paz 2

PAGE 3

To my wife, the better half who inspires me to improve 3

PAGE 4

ACKNOWLEDGMENTS The road to completion of this degree program has certainly been educational for me and beneficial to the concept of disease resistance in Holstein dairy cattle. During this process there have been several individuals who served esse ntial roles and without their contribution, this venture would not have been as successful. Special recognition is extended toward the chai r of my supervisory committee, Dr. Arthur Donovan. This research was not possible without his efforts. I know I will always be grateful for this opportunity he chose to give me, which also provided a second chance to prove myself for veterinary school admission. His knowledge and practical approach to science is always very applicable and thought-provoking. I always en joy hearing his commentary on science publications; he has a way of uncovering details and revealing perspectives which are generally unconsidered. The initial study design was largely provided by the contributions of Dr. Bonnie Mallard and Dr. Armando Hernandez. Their efforts and experience with resear ch concerning immune function was critical to the succe ss of this project. I also must thank the owner of North Florida Holsteins, Mr. Don Bennink for his passion toward s the promotion of Holstein dairy cow wellbeing and willingness to allow his cows to be used for this study. Special thanks are given to the additional members of my supervisory committee, Dr. Pedro Melendez and Dr. Maureen Long. The insight into the statistical an alysis provided by Dr. Melendez certainly improved the research findi ngs. Dr. Longs knowledge of immune system helped determine the biological relevance of the data, which was essential to this study. All members of the FARMS department are well deserving of special appreciation for their contributions. I thank Dr. Louis Archibald fo r his knowledge and sincere enthusiasm for education and research. I was fortunate that Dr. Archibald appointed me as teaching assistant for 4

PAGE 5

veterinary theriogenology for one semester. I al ways enjoyed the conversa tions with Dr. Carlos Risco; whether referencing a reproductive con cern, a seminar topic, or the national champion football or basketball team, they were always very insightful. I am thankful for all the opportunities he provided for me working with Hols tein dairy cows or the ruthless water buffalo. I am very grateful to Dr Owen Rae for his know ledge, respect he shows others, and willingness to help answer any question that popped into my head. I also have to thank Deloris Foreman for her skills managing the department, problem-s olving ability, and her eagerness to organize various events for the department. I certainly en joyed our conversations about fishing. I also have to thank Dr. Pablo Pinedo and Dr. Maur icio Benzaquen. Both were fellow graduate students who befriended me; I lear ned a great deal from them bot h. They especially showed me how to bring a juicy red apple to their teacher for every class. During this degree program, I was also fortuna te enough to serve as teaching assistant for veterinary embryology and histology. Both opportunities woul d not have been possible or as successful without the efforts of Dr. Roger Reep. To prepare for these experiences, Dr. Reep gave me individual weekly training sessions. It is often said that te aching is the best way to learn, and this was exceptionally true for me. I will su rely have an advantag e in these courses for veterinary school. Being a father to my 2-year-old daughter Emily with a wife who must work, a typical concern would be for the wellbeing and teaching of my daughter in daycare. However, thanks to the eagerness of my parents, I have never ha d to worry about these matters. Their unrelenting willingness to care for Emily during the day has provided a great deal of relief knowing she is in good hands. This relief has allowed me the opportunity to focus more clearly on all my graduate 5

PAGE 6

studies. While on the subject of Emily, I should th ank her for being such a wonderful little girl and heavy sleeper; she began sleeping through the night at only 3 weeks old. I am sure I would not be anywhere close to he re if not for my wife. Without her, I would have settled for less a long time ago. By simply be ing herself, she makes me want to continue improving as a person. She values the important th ings in life; she is beautiful, strong, and always behind me for support. She gives more than her share to our marriage while working hard as a bank manager and terrific mother to our fi rst daughter Emily. For all these things and many more, she is the love of my life. 6

PAGE 7

TABLE OF CONTENTS page ACKNOWLEDGMENTs ................................................................................................................4 LIST OF TABLES .........................................................................................................................10 LIST OF FIGURES .......................................................................................................................11 ABSTRACT ...................................................................................................................................13 CHAPTER 1 INTRODUCTION................................................................................................................. .15 Disease Trend .........................................................................................................................15 Immune Suppression ..............................................................................................................15 Selection for Disease Resistance ............................................................................................15 Objectives ...............................................................................................................................18 2 REVIEW OF LITERATURE.................................................................................................19 Immune System Basics ...........................................................................................................19 Introduction .....................................................................................................................19 Innate Immunity ..............................................................................................................19 Acquired Immunity .........................................................................................................20 Intracellular immunity ..............................................................................................20 Extracellular immunity .............................................................................................21 Immunologic memory ..............................................................................................22 Relationship Between Intracellular and Extracellular Immunity ...........................................23 Delayed-Type Hypersensitivity ..............................................................................................24 Indirect Enzyme-Linked I mmunosorbent Assay (ELISA) .....................................................25 Periparturient Immune Suppression .......................................................................................26 Neuroendocrine Effect.....................................................................................................26 Effect of Negative Energy Balance .................................................................................27 Nonesterified fatty acids ...........................................................................................27 Effect of hyperketonemia .........................................................................................28 Effect of Hypocalcemia ...................................................................................................29 Lactogenesis Effect .........................................................................................................29 Dexamethasone ................................................................................................................30 Disease Trend .........................................................................................................................30 Selection for Disease Resistance ............................................................................................31 Direct Versus Indirect Selection ......................................................................................32 Disease Resistance Through Artificial Insemination ......................................................33 Somatic cell count ....................................................................................................33 Structural traits of th e udder and productive life ......................................................34 Disease Resistance Through Specific A ttributes of the Immune System .......................35 7

PAGE 8

Lymphocyte subsets .................................................................................................35 Major histocompatability complex ...........................................................................35 Broad-Based Immune Responsiveness ............................................................................36 Differences between high and low immune responders ...........................................37 Correlation with infectious disease risk ...................................................................38 Correlation with energy-related metabolic disease ..................................................39 Correlation with milk production .............................................................................41 Heritability of measures for immune responsiveness ..............................................42 Categorizing AMIR and CMIR .......................................................................................43 Introduction ..............................................................................................................43 Antibody-mediated immune response ......................................................................44 Cell-mediated immune response ..............................................................................45 3 METHODS UTILIZED FOR THE CA TEGORIZATION OF AMIR AFTER GENERATING ELISA OPTICAL DENSITY VALUES.....................................................53 Introduction .............................................................................................................................53 Study Population .....................................................................................................................53 Exclusion Criteria ............................................................................................................54 Removal of Cows Previously Exposed to Antigen .................................................................54 Methodology and Results ................................................................................................55 Effect of Parity on Antibody Response to Ovalbumin...........................................................55 Interval Variation Adjustment ................................................................................................55 Methods for OD Adjustment ...........................................................................................57 Results OD-3.............................................................................................................58 Results OD0..............................................................................................................58 Results OD+2.............................................................................................................59 Analysis of Classification Methods ........................................................................................60 Methods ...........................................................................................................................61 Results and Discussion ....................................................................................................62 Alternate Index Methods for AMIR Categorization .......................................................65 Results .............................................................................................................................67 4 CATEGORIZATION OF PERIPARTURIENT ANTIBODY RESPONSE TO OVALBUMIN AND ITS RELATIONSHIP WITH COMMON DISEASES AND PERFORMANCE MEASURES OF HOLSTEIN DAIRY CATTLE....................................82 Introduction .............................................................................................................................82 Materials and Methods ...........................................................................................................84 Research Sample Cows ...................................................................................................84 Animal Removal and Interval Criteria ............................................................................84 Body Condition Scoring ..................................................................................................85 Immunization...................................................................................................................86 Blood Collection and Processing .....................................................................................86 Enzyme-Linked Immunosorbent Assay ..........................................................................86 Preliminary Analysis .......................................................................................................88 Removal of non-naive ..............................................................................................88 8

PAGE 9

Parity effect ..............................................................................................................89 Interval analysis and opti cal density value adjustment ............................................89 Classification analysis ..............................................................................................90 Antibody-Mediated Immune Response Classification ....................................................93 Diseases ...........................................................................................................................94 Milk Yield, Somatic Cell Scor e, and Reproductive Efficiency .......................................95 Statistics ...........................................................................................................................96 Results .....................................................................................................................................97 Mastitis and Metritis ........................................................................................................97 Ketosis Displaced Abomasum and Retained Fetal Membrane .......................................97 Milk Yield and Somatic Cell Score .................................................................................98 Reproductive Efficiency ..................................................................................................98 Discussion ...............................................................................................................................98 5 CATEGORIZATION OF PERIPARTUR IENT CELL-MEDIATED IMMUNE RESPONSE TO A TEST ANTIGEN AND ITS RELATIONSHIP WITH COMMON DISEASES AND PERFORMANCE MEASUR ES OF HOLSTEIN DAIRY CATTLE....109 Introduction ...........................................................................................................................109 Materials and Methods .........................................................................................................110 Research Sample Cows .................................................................................................110 Animal Removal Criteria ..............................................................................................110 Body Condition Scoring ................................................................................................111 Delayed Type Hypersensitivity .....................................................................................111 Classification of Cell-Me diated Immune Response......................................................112 Diseases .........................................................................................................................113 Milk Yield, Somatic Cell Scor e, and Reproductive Efficiency .....................................114 Statistics .........................................................................................................................114 Results ...................................................................................................................................115 Mastitis ..........................................................................................................................115 Retained Fetal Membrane ..............................................................................................116 Metritis ..........................................................................................................................116 Ketosis and Displaced Abomasum ................................................................................116 Milk Yield, Somatic Cell Sc ore and Reproductive Efficiency ......................................117 Discussion .............................................................................................................................117 LIST OF REFERENCES .............................................................................................................124 BIOGRAPHICAL SKETCH .......................................................................................................134 9

PAGE 10

LIST OF TABLES Table page 3-1 Incidence of disease for each antibody response categorization method. .........................81 4-1 Odds ratios of disease inciden ce for antibody respons e categorizations. ........................105 5-1 Odds ratios of disease incidence for cell-mediated immune categorizations. .................121 10

PAGE 11

LIST OF FIGURES Figure page 2-1 General depiction of a primary a nd secondary response to antigen x ............................47 2-2 Depiction of the inverse relationship be tween AMIR and CMIR as a result of IL-10 ......48 2-3 Plasma NEFA in thin, medium and ov erconditioned cows during the peripartum period .................................................................................................................................49 2-4 Effects of NEFA on IgM secreti on in peripheral blood mononuclear cells ......................50 2-5 Effects of NEFA on interferonsecretion in periphera l blood mononuclear cells. .........51 2-6 Flow chart describing potential relations hip between energy-related metabolic condition (ketosis) and infectious condition (metritis) ......................................................52 3-1 Basic outline for treatment/ sampling fo r antibody mediated immune responsiveness .....70 3-2 Optical density values reflecting antibody response to OVA by sampling period............71 3-3 Scatter plot of OD-3 values by length of interval 1 for primiparous cows. ........................72 3-4 Scatter plot of OD-3 values by length of interval 1 for multiparous cows. ........................73 3-5 Scatter plot of OD0 values by length of interval 2 for primiparous cows. .........................74 3-6 Scatter plot of OD0 values by length of interval 2 for multiparous cows. .........................75 3-7 Scatter plot of OD+2 values by length of interval 3 for primiparous cows. .......................76 3-8 Scatter plot of OD+2 values by length of interval 3 for multiparous cows. ........................77 3-9 Graph reflecting the effect of maximal antibody response ................................................78 3-10 Graph reflecting the effect of maximal antibody response ................................................79 3-11 Depiction of a method used to categorize antibody mediated immune responsiveness.. ..................................................................................................................80 4-1 General outline of experimental design. ..........................................................................102 4-2 Polystyrene 96-well plate for en zyme linked immunosorbent assay. ..............................103 4-3 Diagram of the placement of test sera into 96 well polystyre ne plate for enzymelinked immunosorbent assay. ...........................................................................................104 4-4 Incidence ketosis by AMIR categorization within parity. ...............................................106 11

PAGE 12

4-5 Graph for the effect of antibody-mediat ed immune response (AMIR) categorization on milk yield. ...................................................................................................................107 4-6 Graph for the effect of antibody-mediat ed immune response (AMIR) categorization on pregnancy by 150 DIM. ..............................................................................................108 5-1 Graph showing increase in double skin -fold thickness respective of parity (multiparous or primiparous). ..........................................................................................119 5-2 Graph revealing parity differen ce for cell-mediated immune response ...........................120 5-3 Graph indicating the difference in risk of RFM between high and low cell-mediated immune response categorization. .....................................................................................122 5-4 Graph for the effect of cell-mediated immune response (CMIR) categorization on milk yield. ........................................................................................................................123 12

PAGE 13

Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science USING ANTIBODY AND CELL-ME DIATED IMMUNE RESPONSE TO TEST ANTIGENS IN PERIPARTURIENT DAIRY COWS AS A MEASURE OF DISEASE RESISTANCE By Jason Michael De La Paz August 2008 Chair: Arthur Donovan Major: Veterinary Medical Sciences Despite major advances in the dairy industr y for sanitation, housi ng, milking strategies, and genetic trait selection; incide nce of disease is still rising fo r Holstein dairy cows. This has sparked research aimed at identifying ways to incorporate genetic selection to improve broadbased immune responsiveness. For this to become a possibility, immune function must become a trait which can be quantified and correlated with risk of disease. For th is study, both branches of adaptive immune function were considered due to the potential for an inverse relationship between the two. As a result, 774 cows were categorized based on their ability to mount an antibody-mediated immune response (AMIR), an d 812 cows were categorized based on their ability to mount a cell-mediated immune res ponse (CMIR). Immune re sponse categorizations were high, medium, or low, such that the meas ured immune response for high > medium > low. Categorization status for AMIR as well as CM IR was found to be significantly associated with mastitis occurrence. Medium immune re sponders were 1.76 and 2.14 times more likely to have an occurrence of moderate or severe ma stitis than high immune responders for AMIR and CMIR respectively. Low AMIR cows were 2.90 ti mes more likely to have an occurrence of ketosis than high responders. Th is association with ketosis followed a low > medium > high 13

PAGE 14

pattern. For CMIR, low responders were 6.68 times more likely than high responders to have a retained fetal membrane (RFM). When only considering multiparous cows low responders for CMIR were 26.52 times more likely than high responders to have an occurrence of RFM. Although not statistic ally significant, medium CMIR status cows were 7.40 times more likely than high responders to have an occurrence of metritis. When considering the performance traits of fertility and milk yield, high AMIR status was associated with reduced fertility and reduced milk yield. However, high CMIR cows produced significantly greater milk yields than medium and low responders. Negative associations between higher levels of AMIR and reduced milk yield are likely attributed to neglecting immune function as a genetic selection trait. Associations between immune function and ketosis provide evidence for immune system involvement with energyrelated metabolic conditions. 14

PAGE 15

CHAPTER 1 INTRODUCTION Disease Trend The susceptibility of dairy cows to infectious disease is increasing. Genetic selection for increased milk yield without regard for disease resistance may be fueling this adverse effect (Harmon, 1994; Heringstad, 2000; Emanuelson, 1988). The rampant selection for increased milk production void of measures for resistance to ma stitis has been found to result in a genetic increase of 0.02 cases of mastitis per cow per ye ar (Strandberg and Shook, 1989). This translates into a genetic increase of 2 mastitis cases for every 100 dairy cows per year. Immune Suppression Immune suppression experienced around the time of calving has been well documented (Mallard et al., 1998; Lacetera et al., 2005; Kimura et al., 2006). This suppression is believed to be at least in part responsible for the increased risk of dis ease peripartum. The added stress associated with parturition and the abrupt change in lifestyle work together to suppress immune function (Mallard et al., 1997; Van Ka mpen and Mallard, 1997). Di fferent mediators of this immune depression include; endocrine hormones (Mallard et al., 1997), hypocalcemia (Kimura et al., 2006), and non-esterified fa tty acids (NEFA) (Lacetera et al., 2004, 2005). Selection for Disease Resistance Since a national database for Holstein health disorders in the United States does not exist, it is impossible to genetically select directly against specific health disorders. Artificial insemination has provided some opportunity for producer s to select for certain traits that vary in their degree of associat ion with disease resistance. Selecti on to improve somatic cell score (SCS) (a logarithmic transformation of somatic cell coun t), productive life and st ructural udder traits (udder cleft, udder depth, rear udder height, rear udder width, and fore udder attachment) have 15

PAGE 16

all been found to be significant pr edictors of susceptibil ity to clinical mastit is (Nash et al., 2000). Six studies estimating the genetic correlation be tween SCS and clinical mastitis all indicated a positive correlation averaging 0.71, and ranged from 0.37 to 0.98 (Nash et al., 2000). However, the relationship between SCC and intra-mammary inf ection is still unclear (Piccinini et al., 1999; Schukken et al., 1999). Also, with the exception of productive life, these traits are only associated with infections of the mammary gland. The increasing incidence of disease associat ed with selection for increasing milk production may be partially explained by the associa tion of increased severity and prevalence of immune suppression with elevated NEFA and deficiencies in calcium. These two postulated factors in immune suppression are also potentially correlated with increasing milk yield. However, previous work also indicates a st rong genetic influence on immune responsiveness (Wagter et al., 2000; Mallard et al., 1998, Biozzi et al., 1979). Application of measures to place genetic selection pressure on immune responsiveness could potentially be used to overcome increasing infectious disease tre nds. Concerns for a negative asso ciation between milk yield and genetic potential for immune responsiveness are not substantiated (Wagter et al., 2003; Detilleux et al., 1995). Breeding to improve immune function is a conc ept which has been examined and utilized in poultry (Soler et al., 2002; Cole, 1968), sw ine (Mallard et al., 1998), sheep (Woolaston and Baker, 1996), and mice (Biozzi et al., 1979). Ho wever, work aimed at categorizing general immune responsiveness in Holstein cattle is just recently gain ing attention (Wagter et al., 2000, 2003; Hernandez et al., 2003). The obvious motive be hind research aimed at identifying superior and inferior immune responsiveness is to serve as an indirect tr ait enabling selection for general broad-based disease resistance. This concept is appealing for several re asons. Since eradication 16

PAGE 17

of environmental infection-cau sing organisms is impossible, the domestic bovine must rely primarily on its immune system to fight pathogens Thus, cows are exposed to a wide variety of pathogens which are also proficie nt at altering their vi rulence mechanisms. As a result, selection should be for the cattle with the most robust repertoire of response against a variety of pathogens. Selection for increased immune re sponsiveness should reduce the strong dependency on vaccines and antibiotics. Increases in antibody titers to vaccinat ions should result in more efficient use of vaccine dosage and potentially a reduction in vaccine dosage (Wagter et al., 2000; Mallard et al. 1997). Selecting for increase d immune responsiveness should also take steps toward addressing the concerns consumers have for excessive use of antibiotics and animal welfare issues. Finally, through a reduction in ma stitis occurrence, SCC should also be reduced leading to increased cheese yiel d, dairy product quality and shelf life (National Mastitis Council, 1996). Research by Wagter et al. (2000) has found a significant vari ation in a cows ability to mount a humoral immune response. They also found that not all co ws experience immune suppression around calving. This research used ov albumin (OVA) as a novel antigen to elicit an antibody mediated immune response (AMIR) in 1 36 Holstein cows and heifers. They found the high responders had the lowest incidence of mastitis. They also found that antibody responsiveness to ovalbumin positively correlated with antibody response to the E. coli J5 vaccination (Rhne Mrieux, Lenexa, KS). Detil leux et al. (1995) found adequate genetic potential for immune responsiveness to exist among sires with top PTA for milk yield. An additional study found that selection for immune responsiveness does not predispose cows to reduced milk yield (Wagter et al., 2003). 17

PAGE 18

Objectives Using 136 cows the previous study by Wagter et al. (2000) had difficu lty finding statistical significance between AMIR and dis ease risk. Also, this study did not include a measure of cellmediated immune response (CMIR), which is the other branch of a cows adaptive immunity. In the present study we have further an alyzed the association between immune responsiveness and disease through th e use of a larger sample size. The objectives of this study were to categorize a cows humoral response to ovalbumin as a measure of AMIR as well as delayed-type hypersensitivity (DTH) response to Candida albicans as a measure of CMIR (Hernandez et al., 2003, 2005). Associations between these immune categorizations and disease risk, namely mastitis, metritis, retained fetal membrane, ketosis, and displaced abomasum were tested. The effect of these immune categori zations on milk production, somatic cell count, and fertility were also tested. 18

PAGE 19

CHAPTER 2 REVIEW OF LITERATURE Immune System Basics Introduction In order to better understand the mechanisms involved in this research study, a basic review of the immune system is required. The im mune system is comprised of several defense mechanisms. Most potential pathoge ns never elicit an immune re sponse because the bodys first line of defense, the epithelial surface, protects against the establishment of infection. The epithelial surface defense mechanisms can be divided into those mechanical, chemical, and microbiological. Mechanical is us ed to describe the defense provided through tight junctions between epithelial cells, or the movement of mucus by cilia in the respiratory tract. In the stomach, the low pH, enzymes (pepsin), antimicrobi al fatty acids, and peptides found in the stomach provide a chemical barrier to inf ection (Risso, 2000). The normal flora found on the skin and in the gut provides a microbiological defense mechanism. The initial response is characterized as an innate nonspecific immune response while extended and/or repeat exposure to pathogen leads to an acquired immune response acting only on specific antigen. Innate Immunity When foreign microbes are able to pass the bodys first line of defense, the immediate response is classified as innate immune re sponse. Innate immune function is adept at distinguishing self from non-self by the use of cell-surface receptors whic h react with general features that are common among microbes called pathogen associated molecular patterns (PAMPs). Therefore this res ponse is characterized as nonspe cific effector recognition of pathogen which involves the phagocytic and infl ammatory activity of the leukocytes, largely macrophages and neutrophils (Beutler and Rietschel, 2003). Macrophages and neutrophils 19

PAGE 20

release various cytokines and chemokines which can initiate the inflammatory response. The inflammatory response works by recruiting additiona l effector molecules to the infection site, reducing the spread of infecti on by microvascular coagulation, and by promoting tissue repair (van der Poll, 2001). Acquired Immunity Infectious disease occurs when a microor ganism succeeds in evading or overwhelming innate immune defenses. It is at this point where the acquired or adaptive immune response is required. Acquired immune function is characteri zed by specific recognition of antigen and has two main branches which provide means for th e elimination of both ex tra and intracellular pathogens. Elimination of extr acellular pathogen is performe d by antibody-mediated immune response (AMIR) pathways. Eradication of intra cellular pathogen is performed by cell-mediated immune response (CMIR) pathways. Intracellular immunity In certain instances, some forms of bacteria, parasites, and all viruses, replicate within cells and are not detected by extra cellular immunity. On these oc casions, the body employs methods to combat intracellular pathogens which largely involve the functi on of T-lymphocytes (T-cells). T-cells only respond to antigen that is accomp anied by a major histocompatibility complex (MHC), forming an MHC:antigen complex. This reac tion requires a specific T-cell receptor for a response to take place (Jensen, 2007). There are two classes of MHC molecules: MH C class I molecules are expressed by nearly every nucleated cell in vertebrates. MHC class II molecules are only ex pressed by professional antigen presenting cells (APC), which are a special group of phagocytic cells. Dendritic cells are the most common APC, but B-lymphocytes a nd macrophages can also function as APCs (Jensen, 2007). These APCs specialize in disp laying a small portion of processed antigen bound 20

PAGE 21

to an MHC class II molecule on the APCs surface (Savina and Amigorena, 2007). Antigen processing for presentation with an MHC class II molecule involves the endocytic pathway. This pathway begins when exogenous antigen has been endocytosed by an APC. There are two main classes of T-cells; CD4+ T-cells (T H ), and CD8+ T-cells (T C ). The T C cell is responsible for cytotoxic activity or cell killing. These cells initiate their cytotoxic activity on cells displaying antigen bound to an MHC class I molecule wh ich is expressed by nearly every nucleated cell in vertebrates. The T H cell is probably the most important of the T-cells. T H cells only respond to antigen bound to an MHC class II molecule, meaning they only respond to antigen presented by dendritic cells, macrophages or B-lymphocytes, also called APCs. There are two subsets of T H cell; T H 1 subset is responsible for cell-mediated functions such as delayedtype hypersensitivity (DTH) and act ivation of cytotoxic T-cells; T H 2 subset is largely responsible for B-cell activation. The functions of the two subsets of T H cell serve to activate cells in both branches of the adaptive imm une response, which is why T H cells are of special importance. Nave T-cell activation requires specific antigen presentation by an APC to a T H cell. The activation of a T H cell initiates the release of cytokines which can: activate the cytotoxic activity of T C cells, stimulate the chemotaxis of leukocyt es, activate B-lymphocytes, and also cause differentiation into memory T-cells. Activated T H and T C cells have a relatively short life-span while the memory T-cells being long-lived can last for the duration of the cows life. Extracellular immunity B-cells use immunoglobulins for antigen r ecognition and are c oncerned with the elimination of extracellular pathogens. Non-membranous imm unoglobulins, called antibodies, function by binding to the antigen which elicited the response. This binding or coating (opsonization) can neutralize the pathogen, and also flags the pathoge n for phagocytosis or complement activity. For a B-cell to differentiate into effector cells (activation), they require 21

PAGE 22

accessory signals from an activated T H 2 T-cell. B-cell activation occu rs when they bind antigen with its membranous immunoglobulin. The antigen is then internalized and degraded. A peptide fragment from the antigen is later displayed on the cells surface with an MHC class II molecule. A specific interaction between the MHC: antigen peptide complex and an armed T H 2 T-cell send activation signals to the B-cell. This allows th e differentiation into an tibody secreting plasma cells and memory B-cells (Parker, 1993). Immunologic memory Adaptive immunity is also associated with immunologic memory which results in a more rapid and effective immune response to pathogens that have been previously encountered. An antibody response profile indicates features which are common to every antibody response (Figure 2-1). The extent of these features or kinetics are differe nt for a primary immune response (first antigen encounter) and s econdary immune response (>1 antig en encounter). After antigen exposure, every antibody response begins with an initial lag phase to allow for somatic hypermutation and clonal differentiati on into effector cells. This pha se is followed by an increase in antibody concentration until a peak concentrati on is reached. After a peak response is reached, it is followed by a steady decline in antibody co ncentration. In a primary response the humoral response results from the activation of nave lymphocytes, whereas in a secondary immune response it is memory lymphocytes which are activated. These memory lymphocytes have greater affinity for their specific antigen, wh ich facilitates greate r immune response upon repeated exposure. Memory lymphocytes are also long-lived, and can prov ide life-long immunity to their specific pathogen. A secondary immune re sponse is characterized by a shorter lag phase, greater magnitude, longer duration and greater antibody affinity to antigen when compared to a primary immune response. A primary antibody response consists primarily of IgM isotype antibody, while a secondary response consists larg ely of IgG isotype. Ther e is also substantial 22

PAGE 23

variation in the kinetics of an antibody res ponse due to antigen type, administration route, presence of adjuvant, and species exposed to antigen. Relationship Between Intracellular and Extracellular Immunity Several studies have cited inverse relationships between intracellular immunity and extracellular immunity (de Vries, 1995; Biozzi et al., 1979; Rupp et al., 2007). Although the three referenced studies support the claim of an in verse relationship, they are all concerned with different aspects of the immune system. The work performed by Biozzi et al (1979) supported an inverse relationship between CMIR and AMIR on the basis of intracellular catabolism of antigen. Breeding mice for high antibody mediated immune responsiveness (AMIR) was associated with slower intracellular catabolism of antigen. The explanation for this finding was that slowed antigen processing also related to prolonged antigen presentation which re sults in greater stimulation for the production of antibodies. The mice selected for high AMIR we re more resistant to extracellular pathogens, however; the support for an inverse relationship en ters when these high AMIR mice were more susceptible to intracellular pathogens. The slow ed intracellular catabolism which was favorable for AMIR is believed to be unfavorably associat ed with relevant measures of cell-mediated immune function (CMIR). Research conducted by Rupp et al., (2007) found support for an inverse relationship on the basis of MHC alleles at the DRB3.2 locus. As previously discussed, MHC molecules are responsible for antigen presentation to T ly mphocytes. This experi ment found significant relationships between available alleles at the MHC locus DRB 3.2 and measures of AMIR and CMIR. However, these relationships were invers ely related. Alleles which confer high measures for CMIR associated with low measures of AMIR, and vise versa. 23

PAGE 24

Another paper by de Vries (1995) outlined a mediator for an inverse relationship on the basis of cytokine expression by T H 2 cells. As previously described, T H 2 lymphocytes are involved with activation of an AMIR while T H 1 lymphocytes are involved with activation of a CMIR. T H 1 lymphocytes secrete the cyt okine interleukin-10 (IL-10). Th is release of IL-10 works to block the function of T H 2 lymphocytes (Figure 2-2). Delayed-Type Hypersensitivity Hypersensitive immune responses are often termed, inappropriate immune responses to antigen. There are four types of hypersensitiv e reactions, three of which are mediated by antibodies (Type I-III), and one is mediated by T-lymphocytes (Type IV). A Type IV hypersensitivity response is called a delayed-ty pe hypersensitivity reaction (DTH) due to a characteristic delayed response by comparison to the acute phase reactions of immediate hypersensitivity (Type I). A DTH reaction gene rally peaks around 24-72 hrs post secondary exposure. The other three hypersensitive reac tions are termed immediate hypersensitivity because reaction peaks occur within mi nutes or hours of secondary exposure. A DTH immune response requi res the stimulation of T H 1 cells to form memory T H 1 cells. Therefore a DTH reaction requires previous expo sure to the specific antigen. Upon secondary antigen exposure, antigen presentation cells (A PC) take up the antigen and display it in conjunction with MHC class II molecule to the previously formed memory T H 1 cells. When memory T H 1 cells bind to APCs, cytokines are rele ased and chemotaxis of predominantly macrophages and neutrophils occur at the exposure site resulting in a granuloma. Clinically, a palpable lump occurs which is largely composed of these macrophages a nd neutrophils and to a far less degree, T-cells. DTH reactions are frequently used to detect exposure to large intracellular antigens. The most common of these antigens would be Mycobacterium tuberculosis DTH reactions have also 24

PAGE 25

been previously used as a means in which to qu antify cell-mediated immune function (Mallard et al., 1998; Hernandez et al., 2003, 20 05). There are no other known in vivo methods for quantifying CMIR. In order to mount a DTH i mmune response the subj ect must have been previously exposed to the antigen at least two weeks prior. Th is is required to provide an adequate period for T-cell clonal expansion and di fferentiation into memory T-cells. The antigen is then injected into the subj ect intradermally and after 24-72 hour s, detection of this responseis by examination for a palpable lump at the injection site. Indirect Enzyme-Linked Immunosorbent Assay (ELISA) The indirect ELISA is a method for detection and quantification of antibody that reacts against a specific antigen. These methods for indi rect ELISA have been used in the past as a means to quantify an individual cows antibody mediated immune responsiveness to a novel antigen (Burton et al., 1989; Ma llard et al., 1997; Wagt er et al., 2000). For the indirect ELISA, antigen is coated to the surface of a microtiter well and after su fficient time, excess free antigen is washed away. Nonspecific reactions are bl ocked by a non-reactive protein. Serum (or some other fluid) which potentially cont ains the anti-antigen primary an tibody of interest, is added to the microtiter well. This allows the primary antib ody to bind the antigen which is attached to the side of the well. After sufficient time has el apsed, the free unbound antibody is washed away. An enzyme-conjugated secondary antibody is added which binds to the constant region of the primary antibody which is adhered to the side of the well. After sufficient time, free secondary antibody-enzyme conjugate is washed away and a substrate for the remaining enzyme is added. The enzyme-substrate reaction is a color produci ng reaction which is qua ntified in terms of optical density using a specialized spectrophotometer. This application can be used to detect previous exposure to a particular infectious disease (Wild, 2001). 25

PAGE 26

Periparturient Immune Suppression The periparturient period for a dairy cow is accompanied by many abrupt events and changes in management which provide stress during this time frame. Various management strategies are employed during the transition period in an effort to ease this major adjustment. However, the act of parturition, lactation, changes in ration and en try into the milking herd have various implications which provide mediat ors for immune suppression. This immune suppression has been detected in several studi es (Lacetera et al., 2005; Saad et al., 1989; Detilleux et al., 1994; Park et al., 1992). This is believed to be at least partially responsible for the increased incidence of dis ease found soon after calving. Severa l studies have been conducted to identify the mediators responsible for immune suppression (Mallard et al., 1997; Lacetera et al., 2005; Kimura et al., 2006). Neuroendocrine Effect Neuroendocrine-immune modulation has been re searched due to the identification of neuroendocrine receptors on the surface of lymphoc ytes as well as their ability to release neurotransmitters and hormones such as growth hormone and insulin-l ike growth factor-1 (Badolato et al., 1994; Blalock, 1994). Findings such as these indicate th at metabolic changes induced by neuroendocrine mediators also have implications on the immune system (Besedovsky and Del Ray, 1996; Dardenne and Savino, 1996). To a certain degree imm une cells resemble small pituitary glands (Von Ruecker and Schmid t-Wolf, 2000). These discoveries have sparked research aimed at determining the effect stress hormones have on lymphocyte function in periparturient dairy cattle. Circulating levels of growth hormone (GH) has been found to have a positive correlation with antibody responsiveness to ovalbumin (OVA) measured in blood serum (r = 0.29, p 0.001) or milk whey (r = 0.31, p 0.0005) (Mallard et al., 19 97). Antibody produced in 26

PAGE 27

response to an E. coli J5 vaccination (Rhne Mrieux E. coli J5, Rhne Mrieux, Lenexa, KS) was also significantly correlated with GH (r = 0.18, p 0.04) (Mallard et al., 1997). Insulin-like growth f actor-1 (IGF-1) has been found to be negatively associated with antibody responsiveness measured in blood serum (r = -0.19, p 0.04) as well as milk whey (r = -0.22 p 0.01) (Mallard et al., 1997). Also, antibody responsiveness was significantly influenced by an interaction between week relative to calving and IGF-1 concentration (p 0.005) (Mallard et al., 1997). The correlation between cortisol concentration and antibody re sponsiveness has also been studied. Cortisol levels were f ound to be positively associated with antibody responsiveness (r = 0.17, p 0.06) (Mallard et al., 1997). The relationship these classical hormones have with each other was also examined. The relationship between GH and cortisol have shown to have a direct relationship with both having maximum concentrations at calving. On th e other hand, during this time frame, IGF-1 concentrations were at a minimum, yielding a negative correlation. GH and cortisol levels decreased in the weeks following calving, while IGF-1 concentrations increased until peak lactation (Mallard et al., 1997). Effect of Negative Energy Balance Nonesterified fatty acids Periparturient dairy cows at or near the onset of lacta tion, frequently undergo negative energy balance, which simply means, the energy requirements for milk production and final stages of calf development prepartum, exceed en ergy intake (Adewuyi et al., 2005). Dairy cows tend to amplify this condition by frequently exhi biting a reduction in dry matter intake beginning in the days prior to calving. To compensate fo r the deficit in energy, adipose tissue lipolysis occurs which produces free fatty acids in the bl ood called non esterified fatty acids (NEFA) 27

PAGE 28

(Adewuyi et al., 2005). Research has discovered that overconditioned cows not only lost significantly more body condition when compared to medium or thin conditioned cows, but also had significantly higher NEFA concentrations (F igure 2-3) (Lacetera et al., 2005). Periparturient dairy cow body condition and blood NEFA conc entration is negatively correlated with lymphocyte function as measured by reductions in peripheral blood mononuclear cell (PBMC); DNA synthesis, immunoglobulin M (IgM) secret ion (Figure 2-4), and interferon-gamma (IFN) secretion (Figure 2-5) (Lacetera et al., 2005, 2004). In this study body condition was a binary trait with overconditioned cows in one group wh ile medium and thin conditioned cows in the other. The significance of NEFA and/or body co ndition with lymphocyte function is generally only found in overconditioned cows. At the time of th is particular work it was speculated that for periparturient dairy cows, alterations in lymphoc yte function may proportionally relate to loss in body condition as assessed by changes in BCS. Effect of hyperketonemia The common state of negative energy balance for periparturient dairy cows also is a predisposing factor for development of hyperket onemia, a condition whereby levels of ketone bodies are elevated in the bl ood. The production of milk for t odays dairy cow requires large demands for glucose. To meet this demand duri ng a time of suppressed dry matter intake, dairy cows undergo intense gluconeogenesis. It is dur ing this time where a large portion of serum NEFA is directed to the liver which is then synthesized into ketone bodies. The serum ketones found in cattle are acetone, acetoacetate, and -hydroxybutyrate (Baird, 1982). Several studies have reported suppressed immune responsivenes s associated with the presence of elevated ketone bodies (Franklin et al ., 1991; Hoeben et al., 1997). However, other studies do not replicate this antagonistic re lationship with the immune sy stem (Nonnecke et al., 1992). 28

PAGE 29

Effect of Hypocalcemia Calcium plays a critical role in the activ ation of immune cells. Immune cell activation involves signal transduction pathways which involve inositol 1,4,5-tris phosphate binding to receptors on the endoplasmic reticulum (ER) which in turn stimulate the re lease of calcium ions (Ca 2+ ) into the cells cytoplasm (Grafton and Thwaite, 2001; Lewis, 2001). The level of the resulting rise in intracellular Ca 2+ has been used as a measure of immune cell responsiveness (Partiseti et al., 1994; Baus et al., 1996). Also, an in vivo study in rats showed that extracellular fluid calcium level is a primary indicator for intracellu lar calcium status (Mailhot et al., 2000). It is known that for a dairy cow in the peri parturient period, the de mands for calcium and risk of hypocalcemia greatly increase as producti on of colostrum and milk initiates. Because calcium is critical to immune cell activation, it has been hypothesized that the increased demands for calcium may unfavorably affect intracellula r calcium levels, which in turn could affect immune cell activation potential. Also, peripa rturient cows who have undergone mastectomy, do not develop hypocalcemia, and more importantly, do not encounter the same degree of immune suppression as lactating peripart urient cows (Goff and Kimura 2002; Nonnecke et al., 2003). Researchers have discovered th at calcium levels in the blood as well as calcium levels stored in the ER decline in the days up to cal ving. They also found serum calcium levels were significantly correlated w ith the intracellular Ca 2+ response as well as Ca 2+ stored in the ER (Kimura et al., 2006). Due to the tremendous pr oduction of milk and its demand for calcium, these research findings substantiate the claim that cows can be at least partially immune suppressed during the peripa rturient period due to de ficiencies in calcium. Lactogenesis Effect It has been hypothesized that immune suppression around calv ing is partly due to the sequestering of available syst emic immunoglobulins (Ig) into the mammary gland for colostrum 29

PAGE 30

and milk. This has also been theorized to expl ain the differences between high and low antibody responders to ovalbumin, where the low responders are low due to increased sequestering of Ig into the mammary gland. This theory was challenged when Wagter et al. (2000) compared the antibody response to OVA in serum to that in whey. For the theories to be supported by a correlation analysis, a negative or inverse rela tionship should be reveal ed, however, correlation analysis indicated a positive si gnificant relationship within ea ch test herd (Herd 1, r = 0.45, p <0.0001; Herd 2, r = 0.28, p <0.001; Herd 3, r = 0.44, p <0.001). Dexamethasone Dexamethasone is a synthetic glucocortico id which is commonly used to initiate parturition in cows within the last 30 days of gestation. This prepartum use provides consistent highly effective results; however, it has also been consistently associated with increased risk of retained fetal membranes (Beardsley et al., 1976; Peters and Poole, 1992). Also, several studies have identified a direct immunosupp ressive activity to measures of innate as well as adaptive immunity which are associated with the use of this glucocorticoid (Burton et al., 1995; Burton and Kehrli, 1995, 1996). Disease Trend Associated with breeding fo r increased milk production without regard for immune responsiveness, a concomitant rise in infec tious disease occurred (Emanuelson et al., 1988; Harmon, 1994; Wagter et al., 2003). For mastitis, Nord ic data reveal that genetic correlation with milk production ranges between 0.24 and 0.55 with an average of 0.43 when large field data sets are analyzed (Heringstad et al., 2000). After assuming a conservative gene tic correlation between mastitis and milk production of 0.30; Strandberg and Shook (1989) state that under traditional progeny testing programs without selection for mastitis, a genetic increase of 0.02 cases of 30

PAGE 31

mastitis per cow per year is the result. So for every 100 cows, there would be a genetic increase of 2 cases of mastitis per year. The mediators behind the correl ation between production and mas titis have not been fully determined. Detilleux et al. (1995) used 137 peripa rturient cows when they found selection for high milk yield did not produce genetic lines with unfavorable measures for innate and adaptive immune function. However, this study may not have been able to adequately reflect mediators for immune suppression which may bridge the gap between clinical mastitis and milk yield. The effect of NEFA (Lacetera et al., 2005), hyperketonemia (Suriyasathaporn et al., 2000) and calcium (Kimura et al., 2006) are three previously described me diators for immune suppression that can also correlate with milk production. As increased milk production per cow is achieved, demands for energy and calcium also increase This increase in demand for both energy and calcium could make the potential for a deficiency in both energy and calcium an ever increasing risk. Since a deficiency in energy can lead to the production of both NE FA and ketone bodies, and both can suppress the immune system. Also, a deficiency in calcium inhibits immune cell activation which also suppresses immune function (Kimura et al ., 2006); it can be hypothesized that selection for increased milk production without regard for immune responsiveness contribute to the mediators of immune suppres sion. The resulting increase in occurrence of immune suppression can be at least partly res ponsible for the increase d risk of disease. Selection for Disease Resistance The concept of breeding animals for disease resistance is not new. Selections against specific genes responsible for dise ase have provided means to pr event or reduce the risk of a given condition and have emphasized the effect gene tics plays in disease re sistance. Selection for disease resistance is most effective when the give n condition is associated with a single or small number of genes. In poultry, selections agains t specific genes responsib le for Mareks disease 31

PAGE 32

have made great strides (Cole, 1968). In dair y cattle, identification of a genetic defect responsible for bovine leukocyte adhesion deficiency (BLAD) has drastically reduced the risk of this condition (Kehrli et al., 1990). In sheep, the associa tion between a certain genotype and natural scrapie risk has been stud ied (Hunter et al., 1997). In swine, certain genes associated with risk of salmonella and E. coli diarrhea ha s been studied (Edfors-Lilia et al., 2000). Several different methods have been developed to select for disease resistance. Some of these methods are concerned with reducing the in cidence of a specific condition (Nash et al., 2000) or pathogen, while others ar e more broad-based in their a pproach and work to reduce the incidence of many related condi tions (Wagter et al., 2000). Generally speaking, more broadbased approaches tend to have slower genetic progress for a specific condition, but during that time frame the genetic progression is favorab le for a larger spect rum of conditions. Direct Versus Indirect Selection In certain instances, selection for disease re sistance is accomplished by directly selecting against the disease itself. Using this approach, one study compared mastitis frequencies of progeny from the best bulls for mastitis resistance to the progeny of the worst bulls (Steine, 1996). In this research, they found the three wo rst bulls for a mastitis resistance index had daughters with twice the mastitis frequency of daughters of the th ree bulls with the best index values. This type of selection requires an extensive database w ith records of disease occurrence in the pedigree of a given animal. Other methods of genetically selecting for dis ease resistance in dairy cows have involved the use of indirect traits (Nash et al., 2000; Heri ngstad et al., 2006; Wagter et al., 2000). Because there is not a national database for Holstein disease occurrence in the United States, it becomes impossible to directly select ag ainst particular inf ectious diseases. So identification of an 32

PAGE 33

effective indirect trait is required. For this to be useful, it must be correlated with the phenotype of interest (disease resistance), it must also be easy to measure a nd heritable (Kelm et al., 2001). Disease Resistance Through Artificial Insemination Somatic cell count In dairy cows, the widespread use of arti ficial insemination has enabled producers the ability to select sires based on predicted tr ansmitting abilities (PTA ) for phenotypic traits associated with disease resistance. Somatic cell sc ore (SCS) is a selection trait available for dairy producers using artificial insemination and is simp ly a logarithmic transformation of somatic cell count (SCC). Elevations in SCC are seen as a re sponse to microbial infestation in the mammary gland and are reported as the number of leukocytes present in 1 mL of milk. Thus, it can not only be used to identify the presence of microbes in the mammary gland but it can also to a certain degree be a potential measure of innate immune response to infection. Se lection for reduced SCS is considered because of the correlation betw een SCC and clinical mastitis. The association between SCC and clinical mastitis has been extens ively researched (Nash et al., 2000; Rogers et al., 1998; Heringstad et al., 2006). Several studie s cite strong correlations between SCC and clinical mastitis (Heringstad et al., 2006; Nash et al., 2000). Nash et al (2000) cites an average genetic correlation of 0.71 over six studies betwee n SCC and clinical mastitis. Heringstad et al. (2004) cites a range in genetic correlation de pendent on phase of lactation ranging from 0.37 to 0.73. The exact genetic relationship between selecti on for lower SCS and clinical mastitis is not completely understood. Because SCC reflects the amount of leukocytes in milk, and this can also be an indicator for innate immune responsiveness to microbial infestation; there is a belief that genetically selecting for reduced SCS in healthy cows may also be concurrently selecting for reduced immune responsiveness. Previous resear ch found that milk SCC prior to experimental 33

PAGE 34

challenge with S. aureus was actually higher in cows that re sisted infection compared to those who became infected (Piccinini et al., 1999; Sc hukken et al., 1999). However, there is substantial disagreement with this philosophy. Two studies found that high immune responders, determined by antibody responsiveness to ovalb umin, had significantly lower LS means for SCC than low responders (Wagter et al., 2000; Mallard et al., 1997). Additionally, Kelm et al. (1997) found a tendency for greater functional ability of neutr ophils from cows with lower estimated breeding values (EBV) for SCS during the periparturient period. In light of the concerns posed by Piccinini et al. (1999) and Schukken et al. (1999), recent work has concluded that SCS should be consider ed a heterogeneous trait with SCC of healthy cows separate from SCC of mastitic cows (Heringstad et al., 2006). This study found the heritability of SCS in healthy co ws to be 0.08, while the heritabil ity in mastitic cows equals 0.03. Structural traits of th e udder and productive life Udder conformational traits have also been used to help select for resistance to clinical mastitis infection. Udders with good attachment and cleft provide greater distance between the teat canal and the ground, or other potential fomites. Also, proper conformation can be associated with proper function and use of milk ing machines, which can also provide an avenue for microbial intramammary infiltration. PTAs fo r traits such as; udder cleft, udder depth, rear udder height, rear udder width, a nd fore udder attachment have b een found to be statistically significant predictors for clinical mastitis risk (Nash et al., 2000). The use of PTAs for SCC and udder conforma tional traits has provided some means to genetically select for disease resistance. However, these traits only associate with infections of the mammary gland, and for the case of selection fo r reduced SCS, we are still unsure if we are also unintentionally selecting for an unfavorable reduction in immune responsiveness. 34

PAGE 35

PTAs for productive life have also shown to be significant predictors of clinical mastitis infection. Productive life is a measure of the sta y ability or the ability of the daughters for a particular sire to resist being cu lled. Selection for this trait has a very crude application to disease resistance especially when you consider the arra y of potential reasons for being culled. However, productive life has been found to be significantly a ssociated with clinical mastitis (Nash et al., 2000). Disease Resistance Through Specific A ttributes of the Immune System Lymphocyte subsets Alterations in populations of T lymphocyte subsets have led to speculation about a potential role in predicting disease resistance. One study f ound a special correlation between features of T lymphocyte populations and the periparturient period. During this period of typical immune suppression, mammary gla nd secretions contained fewer numbers of T-lymphocytes and the ratio for subsets CD4:CD8+ was less than 1 (Park et al., 1992). Speci al attention was then placed on the ratio of T-lymphocyte subsets and their ability to predict disease. Additional studies have found that the CD4:CD8 ratio in ma mmary gland secretions was lowest in cows with Staphylococcus aureus mastitis (Park et al., 1993; Sordillo et al., 1991). Park et al. (1993) found that responsiveness to antigen by CD4+ T lymphocytes was incrementally reduced with increasing presence of CD8+ T lymphocytes. Park et al. (2004) found that mastitis susceptible cows had CD4:CD8 ratios of less than 1 in ma mmary gland secretions as well as peripheral blood. Major histocompatability complex Another arena for exploration involves the id entification of specific MHC haplotypes or gene alleles which are associated with favorab le measures of immune function. This could potentially serve as a genetic marker for disease resistance. The biological relevance stems from 35

PAGE 36

the previously discussed role of MHC in antigen presentation to T lymphocytes. The capacity to adequately present antigen is fundamental to mounting an effective immune response. Bovine MHC is also referred to as the bovine leukocyte antigen (BoL A) and is encoded by highly polymorphic genes. Several studies have found si gnificant differences in immune responsiveness and resistance to disease with alterations in class I and II MHC haplotypes and gene alleles (Rupp et al., 2007; Park et al., 2004; Aaerestrup et al ., 1995; Rupp and Didier 2003; Kelm et al., 1997). A recent study by Rupp et al (2007) looked at th e alleles for the DRB3.2 locus. This is the location encoding the MHC class II antig en binding site, making this region highly polymorphic. This study found several associations between different alleles and measures of immune function, disease resi stance, and performance. Broad-Based Immune Responsiveness Selection for improved immune responsiveness against disease has been studied in poultry (Soler et al., 2002; Heller et al., 1992; Kean et al., 1994), swine (Malla rd et al., 1998), sheep (Woolaston and Baker, 1996), mice (Biozzi et al., 1979) and Holstein cattl e (Mallard et al., 1997; Wagter et al., 2000; Hernandez et al., 2003). Selecting for increased immune responsiveness is to provide an indirect trait that potentially corre lates with broad-based disease resistance. Along with the previously mentioned mediators for immune suppression which affect immune responsiveness, there appears to be a significant genetic effect which determines the magnitude of a particular cows immune responsiveness. Significant variation exis ts in the ability of periparturient dairy cows to mount an immune response indicating that no t all cattle experience the same degree of immune suppression around calving (Mallard et al., 1997; Wagter et al., 2000). The concept of selecting for broad-based dise ase resistance as opposed to resistance to particular pathogens/diseases is a ppealing. It is true that selec tion against specific conditions will 36

PAGE 37

generally provide the quickest genetic progress for that particular condition. However, this often results in little or no genetic progress for other conditions. Also, due to the mechanics of the immune system, simply selecting for resistan ce to one condition without regard for other diseases, may introduce susceptibi lity to other conditions (Biozzi et al. 1979; de Vries, 1995; Rupp et al., 2007). The principles behind the decision to select for broad-based disease resistance are as follows: The cow relies on the immune system as the principle means to which pathogens are fought. Eradication of environmental infection-ca using organisms is impo ssible, so cows are going to be exposed to a wide variety of pa thogens. Through mutations, these organisms are capable of altering their virulence and defense mechanisms including resistance to antibiotics. Selection for increased immune responsiveness should reduce the dependency on vaccines and antibiotics. Increases in antibody t iters to vaccinations should resu lt in more efficient use of vaccine dosage and potentially a reduction in va ccine dosage (Wagter et al., 2000; Mallard et al. 1997). Selecting for increased immune responsiven ess should take steps toward addressing the concerns consumers have for excessive use of antibiotics and animal welfare issues. Through increased immune responsiveness and a reductio n in mastitis occurrence, SCC should also concurrently reduce (Mallard et al., 1997; Wagter et al., 2000) which is associated with increased cheese yield, shelf-life, and dairy product quality (National Mast itis Council, 1996). Differences between high and low immune responders Biozzi (1979) made some interesting discove ries while studying the differences between mice bred for high and low antibody responsiven ess. This study found that there was no difference in the amount of antibody released from individual plasma cells of the high and low responders. However, they found the high respon ders did multiply and differentiate at a significantly faster rate than lo w antibody responders. They found that antigen was catabolized at 37

PAGE 38

a significantly faster rate in the low responders when compared to high responders. The rationale explaining this finding is that slower catabolism of antigen is associated with slower antigen processing and prolonged antigen presentation. This extended antigen presentation is associated with greater lymphocyte s timulation and activation. Correlation with infectious disease risk Several studies have researched the corre lation between immune responsiveness and disease incidence. These studies can also be used to further unde rstand the individual aspects of the immune system and their particular role in the pr evention of a given condition. While breeding mice for high and low antibody responsiveness, Biozzi et al (1979) naturally found that his high antibody responder line was more resistant to extracellular pathogens. However, the high antibody responder s were more susceptib le to intracellular pathogen. This is believed to be due to the slow ed intracellular catabolism of antigen associated with the high antibody responders. When concerned with intracellular immunity, it is the speed at which the cell is able to break down the anti gen which positively reflects the potential of intracellular immunity. In swine, high immune responders as assesse d by measures for AMIR, CMIR, and innate immune function, were found to have considerab ly less peritonitis and pleuritis following a Mycoplasma hyorhinis infection (Mallard et al., 1998). However, it was also noted the high immune responders as assessed by estimated breeding values (EBV) for AMIR, CMIR, and innate immune function had more arthritis than the low responders. This is believed to be the result of selection for increas ed cell-mediated responsiveness and its association with the inflammatory response. Mastitis occurrence within antibody response categorization has been studied in Holstein dairy cows. Although not statistical ly significant, mastitis incidence for the high responders was 38

PAGE 39

lowest in 2 of the 3 study herds. In these two herds the high res ponders did not have an incidence of mastitis. There was also reason to question the va lidity of the herd having more mastitis in the high responders. In this herd, all cases of mastitis were in first parity heifers. The incidence of mastitis is generally higher for multiparous cows compared to primiparous. Relating to mastitis, two studies have also looked at the correlation between antib ody responsiveness to ovalbumin and antibody response to the E. coli J5 vaccination (Rhne Mrieux, Lenexa, KS) (Mallard et al., 1997; Wagter et al., 2000). In both instances they found that the correlation between antibody titers to the E. coli J5 vaccine and response to ovalbumin we re positive and significant. Wagter et al. (2000) reported a general correl ation of r = 0.56 (p < 0.0001). The E. coli J5 vaccination has been proven to be associated with re duced SCC, reduced time for clearance of E. coli in milk, and less milk production loss following intramammary challenge (Wilson et al., 2007). Because of this efficacy there is true biological relevance in selecting for increased immune responsiveness to ovalbumin. Mallard et al. (1997) also compared the incidence of disease between high and low AMIR cows. In this study their main focus was the effect of cortisol, GH, and IGF-1 on antibody response profiles. However, they also looked at disease occurren ce over the 3 categorizations for antibody response. They found that disease inciden ce was smallest for high responders (group 1). This finding also followed the pattern of high
PAGE 40

energy balance and the development of ketosis-fa tty liver complex. Another explanation for this association is provided through L acetera et al (2005) and Suri yasasathaporn et al. (2000). Because the presence of NEFA and ketone bodies is closely linked with negative energy balance and energy-related metabolic cond itions (ketosis); it is possible that the presence of NEFA and ketone bodies serve to suppress immune function wh ich increases the risk for infectious disease (Figure. 2-2). The discovery of elevated NEFA concentra tions in response to inflammatory agents (Steiger et al, 1999; Kushibiki et al., 2003) has sp arked research to determ ine if early lactation mastitis can cause ketosis-fatty liv er complex in dairy cows (Waldron et al, 2006). However, this work concluded that the results do not indicate mastitis to be ca usal for energy-related metabolic disorders. Instead, they did sugge st the possibility for a potential protective effect by the immune system on metabolism during early mammary infection. In some instances, research studying immune responsiveness reveals a relationship with a disease previously not underst ood to be correlated with th e immune system. Schukken et al. (1988) studied the relationship between an inf ectious disease (mastitis ) and retained fetal membranes. This study revealed th at cows having a retained fetal membrane were more likely to have a case of mastitis shortly after calving. Retained fetal memb rane is a condition which is now understood to be the result of a faulty immune response. The body must be able to identify the placenta as foreign and mount an appropria te immune response against the placenta soon after parturition. Research has demonstrated that neutrophil chemotaxis as well as killing ability is impaired in both pre and postpartum cows that will/had a retained placenta (Kimura et al., 2002). Because of this, it is possi ble that the prepartum use of de xamethasone and its association with increased risk of retained fetal me mbranes might be partially explained by the 40

PAGE 41

immunosuppressive action of dexamethasone (B urton et al., 1995; Burton and Kehrli, 1995, 1996). We do not exactly know the precise magnitude or the role the immune system components have on many metabolic activities. Th e complexity of the immune system in vivo often times provide unpredicted results during research study. The possibility of the immune system serving a substantial role in body meta bolism is considered plausible. Correlation with milk production Selection for improved immune responsiven ess should yield a tr end toward reduced incidence of disease. Although this strategy pr ovides many benefits for animal welfare, the extent to which this philosophy is adopted will be strongly influenced by economics. A decrease in disease occurrence will provide obvious fina ncial benefits through a reduction in treatment costs and also a reduction in milk loss from diseas e or treatment of disease. However, there is considerable speculation that a ny economic benefits observed may be counteracted by decreased performance. In other species, the energy and nutritive demands of a superior immune system have been shown to reduce performance. The en ergy and nutrients requir ed for maintenance and activation of a responsive immune system could otherwise be used for other phenotypic traits (Klasing et al., 1987, 1998; Soler et al., 2003). C oncerning dairy cows, the positive correlation between milk production and c linical mastitis (Emanuelson et al., 1988; Harmon, 1994) provides some evidence that financial benefits may be off-set by a reduction in milk yield. Along with the previously cited avian study, additional research has been conducted to study the ramifications of superi or immune responsiveness with pe rformance. In swine, Mallard et al. (1998) found that growth performance was consistently significantly greater for high immune responders when compared to both th e low immune responders and the control group. In this research, pigs were selectively mate d for high and low immune responsiveness over the 41

PAGE 42

course of eight generations using measures fo r innate and adaptive im munity (AMIR & CMIR). The trend of improved rate of gain for the high immune responder line compared to the rest was first identified in generation 0 and continued through generation 7. In Holstein cows, Wagter et al. (2003) looked at milk pr oduction within high, medium, and low AMIR categorizations. Because parity significantly contributed to variation in 305 day milk yield, they conducted their analys is within parity. For first pa rity cows the low responders produced significantly more milk than the medium and high responders. However, for second parity cows, there was no statistical differen ce between the high and low as well as high and medium response groups. For third parity cows, the high responders produ ced significantly more milk than the low and medium response groups (Wagter et al., 2003). Another study showed that dairy cows genetically selected for high milk yield over seven generations did not produce unfavorable measures for innate and adaptive immune function when compared to cows selected for average milk production (Detilleux et al., 1995). The findings of these studies indicate that although there is an association between selection for increased milk yield and increased risk of clinical mastitis, selection for improved immune responsiveness should not predispose cows to reduced milk yield. It also indicates adequate variation in immune responsiveness among sires with high PTA for milk yield to support selection for both increased milk yield as well as increased immune responsiveness Heritability of measures for immune responsiveness The level of heritability expected during selec tion for measures of immune responsiveness indicate the rate in which ge netic improvement can be made If the proposed measure for immune responsiveness has low leve ls of heritability, the effectiven ess of an indirect trait would be very limited. 42

PAGE 43

The heritability of measures for AMIR and CMIR has been tabulated in pigs (Mallard et al., 1998). These calculations were configured on over 1200 observations through 8 generations of selection for high and low immune responsivene ss. Heritability of AMIR was estimated to be 0.268 while CMIR heritability estimates equale d 0.163. The heritability of AMIR as measured by antibody responsiveness to ovalbumin has also been tabulated for Holstein periparturient cows (Wagter et al., 2000). These estimates of heritability ranged 0.32 to 0.64 dependent on week relative to calving. The lower value for this range (0.32) coincided with the heritability of antibody response to OVA measured at calving. Ha ving the lowest estimate occur around calving may be explained by the various stress factors and mediators for immune suppression which are occurring during this period. Categorizing AMIR and CMIR Introduction Studies which associate immune responsiven ess with disease risk usually involve correlative studies associating an iddices of immunity with risk for disease. The method chosen to stimulate the immune system by which to meas ure immune response is of primary importance. The technique used should represent a subjects ove rall potential to resist disease. Methods used to categorize immune responsiveness for the pur pose of selection for disease resistance should have a very general approach wh ich gives consideration to all aspects of immune function. When considering the separate roles in the immune system for AMIR and CMIR as well as the previous work which identifies a potential inverse rela tionship between the two (Biozzi et al., 1979; de Vries, 1995, Rupp et al. 2007), inclusion of bot h branches becomes essential. Due to the potential for an inverse relationship, a failure to cons ider both branches upon genetic selection may result in increased susceptibilit y in the branch not considered. 43

PAGE 44

Antibody-mediated immune response There are two different techniques that have previously been reported to categorize AMIR in Holstein cows. In both instances, cows were categorized during the peri parturient period with ovalbumin as the test antigen to elicit the humoral response. The study cows were injected with the antigen at week -8, week -3, and week 0 re lative to calving. Blood samples to measure the ensuing antibody response were co llected on week -8, week -3, w eek 0, week +3, and week +6. Antibody response was detected by ELISA and the re sulting OD values were used to categorize AMIR function (Wagter et al., 2000; Mallard et al., 1997). However, these studies differ in how AMIR categorizations were extrapolated from the OD values. Mallard et al. (1997) found that the sample si ze of 33 cows and heif ers partitioned into three groups. All cows responded well to the initia l antigen exposure at week -8, however it was the responses to the subsequent antigen exposures which determined their categorization. High responders had an above average response to a ll three antigen administrations. The medium responders responded well to both of the prep artum antigen administrations, yet responded poorly postpartum to the week 0 injection. The low responders mounted poor pre and postpartum responses to the week -3 and w eek 0 administrations of antigen. Wagter et al., (2000) categorized AMIR in 136 cows and heifers. In this study an index was generated which used the change in OD value over the intervals between the antigen injection/blood collection periods. This resultant value was then used to categorize cows as high, med, or low antibody responders. The formula fo r this index is as follows (Eq. 2-1): y total = I 1 + 1 I 2 + 2 I 3 + I 4 (2-1) Where: y total = total antibody I 1 = change in optical density (OD) between week -8 and week -3 44

PAGE 45

I 2 = change in OD between week -3 and week 0 I 3 = change in OD between week 0 and week +3 I 4 = change in OD between week +3 and week +6 1 & 2 = either 1.0 or 1.5 The coefficient 1 takes on a value of 1.0 if I 2 is positive, representing a positive antibody response around parturition. If I 2 is negative, representing a lack of response around calving, 1 takes on a value of 1.5 which serves to magnify or inflate the negative response. This rationale was also applied to 2 concerning I 3 To categorize immune responders, the mean a nd standard deviation for all the generated y total values was configured. Cows were classified as high responders if they had an y total value greater than the mean plus one standard deviation. A medium responder had an y total value within the mean plus one standard deviation and th e mean minus one standard deviation. A low responder had an y total value below the mean minus one standard deviation. Cell-mediated immune response Methods employed to categorize cell mediated immune function have all utilized delayedtype hypersensitivity (Hernandez et al., 2005, 2003; Mallard et al., 1998). Because this reaction is mediated by T H 1 cells it is an indicator of intracellula r immunity. This is the only in vivo method known which enables categorization of CMIR. Quantification and therefore categorization of the CMIR comes from measurements taken at the antigen injection site. As previously stated, DTH reactions require previous exposure to the antigen intended to elicit the DTH response. To in itiate the DTH response, the antigen is injected intradermally and double skin-fol d measurements are taken to serve as a baseline. All measurements should be taken w ith three repetitions while usi ng the average of the three for analysis. After 24 to 48 hours, measurements of the injection site are once again taken which 45

PAGE 46

should reflect the palpable lump indicative of a DTH response. The degree to which the measurements increased can be used as an indi cator of cell mediated immune responsiveness. 46

PAGE 47

Figure 2-1. General depiction of a primary and secondary response to an tigen x. There is substantial variation in the kinetics of an antibody response due to antigen type, administration route, presence of adj uvant, and species exposed to antigen. 47

PAGE 48

T-LYMPHOCYTES CD4+ CD8+ TH1 TH2 Secretes IL-10 Associated with humoral immunity IL-10 blocks cytokine secretion of TH2 cell Associated with cellular immunity Figure 2-2. Depiction of the inve rse relationship between AMIR a nd CMIR as a result of IL-10. 48

PAGE 49

Figure 2-3. Plasma NEFA in thin, medium a nd overconditioned cows during the peripartum period. Values with different letters differ significantly (P < 0.01). Values reported are LS means SEM. Reprinted with perm ission from: Lacetera, N., D. Scalia, U. Bernabucci, B. Ronchi, D. Pirazzi, and A. Nardone. 2005. Lymphocyte Functions in Overconditioned Cows Around Parturiti on. J. Dairy Sci. 88:2010-2016. Figure 2, page 2012. 49

PAGE 50

Figure 2-4. Effects of NEFA on IgM secretion in peripheral blood mononuclear cells stimulated with pokeweed mitogen. Values reported are mean SEM. Columns with different letters differ significantly (P < 0.05). Ns = Not stimulated. Reprinted with permission from: Lacetera, N., D. Scalia, O. Franci, U. Bernabucci, B. Ronchi, and A. Nardone. 2004. Short Comm: Effects of Nonesterifie d Fatty Acids on Lymphocyte Function in Dairy Heifers. J. Dairy Sci. 87:1012-1014. Figure 2, page 1014. 50

PAGE 51

Figure 2-5. Effects of NEFA on interferonsecretion in periphera l blood mononuclear cells stimulated with concanavalin A. Values reported are mean SEM. Columns with different letters differ significantly (P < 0.05). Ns: not stimulated; ND: not detectable. Reprinted with permission from: Lacetera, N., D. Scalia, O. Franci, U. Bernabucci, B. Ronchi, and A. Nardone. 2004. Short Comm: Ef fects of Nonesterified Fatty Acids on Lymphocyte Function in Dairy Heifers. J. Dairy Sci. 87:1012-1014. Figure 3, page 1014. 51

PAGE 52

Negative Energy Balance NEFA Ketosis Immune Suppression Metritis Figure 2-6. Flow chart describi ng potential relationship between energy-related metabolic condition (ketosis) and infectious condition (metritis). 52

PAGE 53

CHAPTER 3 METHODS UTILIZED FOR THE CATEGORIZ ATION OF AMIR AFTER GENERATING ELISA OPTICAL DENSITY VALUES Introduction Enzyme linked immunosorbent assay (ELISA) is commonly used to detect the presence of a particular antigen, or specifi c antibody in body fluids or tiss ues (Wild, 2001). Previous work used the generated optical density (OD) values from an indirect ELISA as a means to quantify antibody mediated immune responsiveness (AMIR) for individual dairy cows (Mallard et al., 1997; Wagter et al., 2000; Hernandez et al., 2003). In our research we utilized these previously described methods. However, after examination of the results, it was decided that further adjustments to the OD values were required. We al so found that not all cows were nave to the test antigen. This chapter discu sses the identification of cows previously exposed to the test antigen, and describes the reasoning behind the adjustments made to the OD values, and the methodology employed to make these adjustments. This chapter also discusses the alternate methods attempted to categorize AMIR and the justification for the method chosen. Study Population In total, 875 Holstein cows/heifers were enrolled into th e study population at approximately 8 weeks (wk-8) prior to expected calving. In cows, this wa s the initiation of the dry period. Animals were enrolled if the expected dry period length was less than 90 days, if reconfirmed pregnant at enroll ment and also if found in good health with no obvious signs of disease. All test animals were from a single herd in north central Florida which maintains exceptional record keeping. All cows and he ifers were enrolled between September 9 th and December 31 st 2004, calved between October 25 th 2004 and March 12 th 2005, and exited between November 9 th 2004 and March 28 th 2005. All cows and heifers received a routine dry off, prefresh, and fresh cow protocol. 53

PAGE 54

Exclusion Criteria A reduction in the study population was necessary to maintain the integrity of the study. Of the 875 cows and heifers, 13 were removed due to missing samples at one of the blood collection periods. An additional 88 were removed either b ecause they were not nave to test antigen or because they did not meet the interval or dry pe riod length exclusionary criteria which will be discussed later in the chapter. A total sample size of 774 with 433 cows and 341 heifers upon enrollment were analyzed. Because measurem ent of AMIR concluded 2 weeks after calving, heifers will be referred to as primiparous cows and the cows at enrollment will be referred to as multiparous cows (primiparous = first lactation; or multiparous = second or greater lactation). Removal of Cows Previous ly Exposed to Antigen Ovalbumin (OVA) was chosen as the antigen to stimulate AMIR. The rationale behind this decision came from the ability of OVA to stim ulate a humoral response, the low likelihood of prior exposure, and the previously successful use of this antigen as a tool to categorize AMIR in dairy cows (Mallard et al., 1997; Wagter et al., 2 000; Hernandez et al., 2003) For this trial, the periods of antigen exposure and blood collecti on occurred at enrollment (wk-8), entry into springer pen (wk-3), and calving (wk0). An additional blood sample was collected 2 weeks after calving (wk+2) (Figure 3-1). Bl ood samples were collected to determine antibody response to antigen. The OD value for wk-8 (OD -8 ) was to serve as a covariat e for the antibody response to OVA. Equal treatment for the measurement of AMIR requires that all anim als are nave to test antigen at this point. This is critically impor tant due to the differences between primary and secondary antibody responses (Figure 2-1). If a po rtion of cows mount a secondary response due to previous exposure and are be ing compared with cows mounting a primary response, this introduces unequal treatment into your study population. 54

PAGE 55

Methodology and Results After review of the OD -8 values it was apparent some cows were most likely not nave or had high nonspecific reactivity to the test antigen, so a na tural cut-off point of an OD -8 = 0.455 was used. This point corresponded with 1 sta ndard deviation above the mean for all OD -8 values. As a result of this analysis, 38 animals (23 primiparous and 15 multiparous) were termed not nave to test antigen and were removed from the study. Effect of Parity on Antibody Response to Ovalbumin A repeated measures analysis using OD -8 as a covariate with the mixed procedure of SAS revealed that multiparous cows responded significantly higher than primiparous cows at every antibody response measurement week (p<0.0001). Due to this effect, adjustments to OD values as well as AMIR categorizations were all made with respect to parity (F igure 3-2). Explanations for this finding may include higher levels of stress and therefore immune suppression for younger cows experiencing lactation and parturiti on for the first time compared to multiparous cows. Another explanation involves the presence of a more extens ive antibody repertoire in older cows. This could simply be due to the effect of time, allowing greater exposure to a broader array of pathogens. Another explanation Interval Variation Adjustment Defining Intervals: Antigen was injected and blood was collected at specified time frames (Figure 3-1). Variation in the number of days be tween these points of blood collection/antigen injection has a strong influence on the measured antibody response. These in terval lengths reflect the duration between previous antigen exposure and measurement of antibody response; which is relevant due to the kinetics of an antibody re sponse profile (Figure 2-1). In every antibody response there is a lag phase followed by a peri od of increasing antibody concentration up to a peak response which is followed by a steady de cline in antibody con centration. Introducing 55

PAGE 56

variation in the number of days between points of blood collection/antigen injection generates inconsistency in the phase of antibody response profile wher e antibody response was measured. For this research, interval 1 (Int 1 ) was defined as the duration in days between sampling periods identified as wk-8 and wk-3. Interval 2 (Int 2 ) was defined as the duration in days between sampling periods, wk-3 and wk0. Interval 3 (Int 3 ) was defined as the duration in days between sampling periods, wk0 and wk+2. Strictly for the purpose of monitoring the ch ange in body condition score (BCS), additional intervals were identified. Interval 4 (Int 4 ) was defined as period between wk-3 and wk0 which is an indication of the change in BCS ove r the transition period. Interval 5 (Int 5 ) was defined as period between wk-8 and w k0. Also, interval 6 (Int 6 ) was defined as period between wk-8 and wk+2. Interval Exclusionary Criteria: One exclusion criteria at study assignment was based on dry period length (Int 1 plus Int 2 ). If this period was greater th an 90 days, they were removed from the study. Although this re striction consists of Int 1 plus Int 2 it does not adequately put restrictions on individual lengths for either Int 1 or Int 2 However, it can eliminate cows with various metabolic problems from skewing results. A minimum interval length of 12 days was set for both Int 1 and Int 2 This does work to restrict specific interval lengths to a minimum, but it still leaves room for substantial variation to occur. Applying these exclusionary criteria an additional 50 animals were eliminated yielding a study population of 774. Due to the resources available and nature of a large-scale dairy farm, considerable variation in the lengths for Int 1 and Int 2 did occur. To correct the OD values for this effect, adjustments were made to OD values based on th e duration of the interval leading up to the 56

PAGE 57

respective OD value. For example, adjustments to the wk-3 OD (OD -3 ) were made based on the length of Int 1 for a respective cow. With Int 3 concluding at a sampling period occurri ng after calving (wk+2) with animals managed in the milking herd, this duration wa s under our control which greatly reduced the variation in Int 3 lengths. No adjustments were required for the OD -8 values because this was the initial point of exposure. Methods for OD Adjustment Statistical analysis to determine the effect of interval length on OD value was performed using the GLM procedure of SAS. The OD respons e to be adjusted was the outcome variable while potential explanatory variables for each cow included: Int Y = interval length in days for interval y BCS X = categorization of body condition score for sampling week x BCS Y = change in body condition score over interval y OD X = optical density value at sampling week x Dex = binary effect, whether cow recei ved dexamethasone prior to calving Sick = binary effect, incidence of either medium or severe case of mastitis, metritis, ketosis, or displaced abomasum within 16 DI M. Only used for analysis to correct OD +2 for reasons to be discussed later. Explanatory variables remained in the model if the effect showed a tendency (P < 0.10) to predict the desired OD value. With = 0.05, if statistical significance was found with the main effect (Int y ), the resultant parameter estimate was applied to Equation 3-1 to correct the corresponding OD values. If a given OD value was adjusted based on interval length, the model predicting the OD value for the subsequent samp ling period would have the adjusted OD from the previous sampling week. AOD ZPX = OD ZPX + PE YP (Mint YP Int ZPY ) (3-1) 57

PAGE 58

Where: AOD ZPX = the adjusted OD value for cow z in parity p for week x OD ZPX = OD value for cow z in parity p for week x PE YP = parameter estimate for the effect interval y has on OD x values for parity p. Also, interval y must always immediately precede week x Mint YP = median number of days for interval y for parity p Int ZPY = the actual interval length for cow z in parity p for interval y Results OD -3 The length in days for Int 1 was under considerable varia tion ranging from 12-55 days in primiparous cows (Figure 3-3).View of a scatter plot depicting the relationship between Int 1 and OD -3 values do not reveal any obvious pattern. Model effects were: Int 1 OD -8 BCS -8 and BCS 1 For primiparous cows, Int 1 was not found to be a significant predictor of OD -3 values (p = 0.34, = -0.0015). Because of this, no adjustments were made to primiparous cow OD -3 values. The Int 1 variation for multiparous cows ranged from 21-71 days (Figure 3-4). Inspection of the scatter plot depicti ng this relationship revealed an obvious association. The model effects were Int 1 and OD -8 There was a significant linear effect between Int 1 and OD -3 values (p < 0.0001, = -0.0081). This parameter estimate represents the slope of the fitted line for the linear model. The median Int 1 in multiparous cows equaled 36. This parameter estimate and median was then used to generate the adjusted OD -3 values for each respective multiparous cow (Eq. 31). Results OD 0 The range in days for Int 2 also had considerable varia tion. For primiparous cows, this range extended from 12 to 45 days (Figure 3-5). Model effects were Int 2 OD -3 BCS 2 and Dex. 58

PAGE 59

The variation in length of Int 2 had a significant effect on OD 0 values (p < 0.0001, = 0.0246).The median Int 2 duration of days was 21. This parameter estimate and median was then used on primiparous cows to adjust the OD 0 values (Eq. 3-1). For multiparous cows, the length of Int 2 ranged from 12-44 days (Figure 3-6). The scatter plot depiction of the relationship clearly indicates an association. This plot also clearly represents a characteristic secondary immune response w ith a longer duration of peak response. Model effects were Int 2 and OD -3 The effect of Int 2 length significantly effected OD 0 values (p < 0.0001, = -0.0211). The calculated median was 22 da ys. This parameter estimate and median was used to adjust OD 0 values for multiparous cows (Eq. 3-1). Results OD +2 For primiparous cows the variation in Int 3 ranged 12-21 days (Figure 3-7). For this analysis the model effects included Int 3 OD 0 and BCS 4 The length of Int 3 was a significant predictor for OD +2 in primiparous cows (p < 0.0032, = -0.0298). The calculated median number of days was 16. The parameter estimate and median was used to adjust OD +2 values in primiparous cows (Eq. 3-1). For multiparous cows the range for Int 3 was 12-20 days (Figure 3-8). This model included Dex,, OD 0 and Int 3 However, Int 3 .was not a significant predictor for OD +2 in multiparous cows (p = 0.21, = -0.0104). No adjustments were made to OD +2 in multiparous cows. An additional GLM model was run after the di scovery that parity was not a significant predictor for OD +2 as determined by a linear regression m odel (Eq. 3-2) which is later discussed. This new model analyzed associations with OD +2 irrespective of parit y. The remaining model effects were Sick, OD 0 and, Int 3 In this instance the model revealed Int 3 is a significant predictor for OD +2 (p < 0.0044, = -0.0182). After consider ation, this will not be used to adjust OD +2 values due to the added power of a repeated measure analysis in this application. 59

PAGE 60

Analysis of Classification Methods As previously discussed in chapter 2, there are two different techni ques that have been previously reported to categorize AMIR in peripa rturient Holstein cows. In both instances, cows were categorized with ovalbumin as the test antigen to elicit the humoral response. The study cows were injected with the antigen at wk8, wk-3, and wk0 relative to calving. The ensuing antibody response was detected by ELISA and the resulting OD values were used to categorize AMIR function (Wagter et al., 2000; Mallard et al., 1997). The response to antigen introduced at wk0 was detected on wk+3 and wk+6 postpartum. Use of interval changes in OD: The method for the actual AMIR categorization in the publication by Wagter et al. (2000) used an index based on the change in OD response over the intervals between blood sampling/ antigen injection periods (Eq. 2-1). As pr eviously discussed in chapter 2, this index we ights those intervals ( 1 & 2 ) around calving if they show a decline in OD value, which represents a decline in antibody concentration. y total = I 1 + 1 I 2 + 2 I 3 + I 4 (2-1) Where: y total = total antibody I 1 = change in optical density (OD) between week -8 and week -3 I 2 = change in OD between week -3 and week 0 I 3 = change in OD between week 0 and week +3 I 4 = change in OD between week +3 and week +6 1 & 2 = either 1.0 or 1.5 In this study, inspection of th e OD values for the respective weeks revealed a point where antibody response was at an appare nt maximum. If there was an improvement after subsequent antigen injection, these maximal poi nts in antibody concentration on ly slightly improved. If this peak response was achieve d early in the study (OD -3 ), which would indicate a high prepartum responder; there would be little if any room for an added re sponse. As a result an index 60

PAGE 61

concerned with weighting the change in OD dur ing the intervals adjacent to calving (Eq. 1-1), may negatively impact the categorization of hi gh immune responders who reached this peak response early because there was little or no ro om to further respond. Speculation for causation of this finding most likely involves the natural function of feedback inhibition. Early postpartum measures of immune function: The period immediately following parturition is a common occasion for increased in cidence of disease. As a result it may be hypothesized that substantial sickness could cont ribute to immune suppression. This effect of substantial sickness could be a confounding va riable for antibody responsiveness to OVA detected early postpartum. This would make it di fficult to study the effect measures of immune responsiveness have on disease ri sk if the association could al so be in the opposing direction. Objectives: The 2 objectives were as follows: 1) To analyze the potential for a maximal antibody response and its possible effect on an tibody response categorizat ion. 2) To study the potential for an effect of sickness on early postp artum measures of antibody responsiveness. This is performed by analyzing the relationship between OD +2 values and substantial sickness. Methods Maximal response: For this analysis, the relati onship between OD value and the subsequent interval change in OD (I) was studi ed. This was performed for the relationship between OD -3 and I 2 and also between OD 0 and I 3 The total sample size for this analysis was 774 cows. Cows were arranged according to their OD -3 value and grouped into one of 14 groups with 55 cows per group except for group 14 which had 59 cows. The top 55 cows for OD -3 became group 1 while the bottom 59 cows for OD -3 became group 14. This same process was also performed based on OD 0 values. A one-tailed two sample t-test was used to determine if the OD value group had a significant effect on the subsequent interv al change in OD. For this analysis, I 2 and I 3 had a 61

PAGE 62

normal distribution, however in both cases, onl y group 1 cows could be compared to group 2 cows due to statistical differences in variance between group 1 a nd the rest of the groups. The null hypothesis is: H o : Group 1 I 2 Group 2 I 2 0. Rejecting the null with statistical significance means that Group 1 I 2 is significantly smaller than Group 2 I 2 A linear regression model using the REG procedure of SAS was also used to see if OD -3 had a significant effect on I 2 or if OD 0 had a significant effect on I 3 Effect of sickness on OD +2 : For the statistical analysis a linear regression model with the REG procedure of SAS. OD +2 served as the outcome variable while potential explanatory variables included: BCS x = categorization of body condition score for sampling week x BCS y = change in body condition score over interval y OD 0 = optical density value at wk0 Dex = binary effect, whether cow recei ved dexamethasone prior to calving Sick = binary effect, incidence of either medium or severe case of mastitis, metritis, ketosis, or displaced abomasum within 16 DIM. Parity = binary effect, either primiparous or multiparous Explanatory variables remained in the model if the effect showed a tendency (P < 0.10) to predict the dependant OD value. For statistical significance, = 0.05. The Corr procedure of SAS was also used to test the correlation between OD 0 and OD +2 in healthy cows as well as those classified as sick within 16 DIM. Results and Discussion Maximal response: After sorting the OD -3 values from greatest to least, it was discovered that of the top 20 cows for OD -3 value, 15 had a smaller OD 0 value (75%), yielding a negative interval 2 change in OD. In th ese instances, 15 of the top 20 OD -3 responders would have an 62

PAGE 63

amplified ( = 1.5) negative I 2 value applied to their AMIR index if using an index which weights interval changes in OD. After grouping the OD -3 values from 1 to 14, with group 1 being the top 55 OD -3 values and group 14 the bottom 59 OD -3 values; this revealed a significantly smaller change in OD over interval 2 for group 1 cows compared to group 2 cows (p < 0.0001) (Figure 3-9). This value actually averaged below 0 (-0.05) for these top 55 OD -3 responders. Linear regression also revealed OD -3 is a significant predictor for I 2 (p < 0.0001; = -0.28). A negative indicates that as OD -3 values increase, I 2 values decrease. Repeating the same process by ranking cows based on OD 0 values in order to compare interval 3 change in OD, revealed similar results. Due to missing wk+2 blood samples the sample size for this analysis was 754. As a result group 14 had 39 cows while group 1 13 had 55. Of the top 20 cows for OD 0 values, 15 had negative interval 3 changes in OD (75%). Also, group 1 cows based on OD 0 values had a significantly smaller interval 3 change in OD compared to group 2 cows (p < 0.0001) (Figure 3-10). Linear regression also revealed OD 0 is a significant predictor for I 3 (p < 0.0001; = -0.50). A negative indicates that as OD 0 values increase, I 3 values decrease. In these instances of high OD values follo wed by a subsequent negative interval, the interval is not negative due to a poor subs equent OD value. Of the top 20 cows for OD -3 18 still had OD 0 values above the third quartile for the popul ation and the other 2 were still above the median for the population. For the top 20 cows for OD 0 all 20 still had OD +2 values above the third quartile for the population. The negative interval was simply the result of an inability to respond further. As a result, additional indexe s were generated and analyzed based on their correlation with disease incidence. This finding is likely due to feedback inhibition due to the 63

PAGE 64

difficulty in boosting a subject that already has a high anti body concentration. Although OD adjustments were made to accommodate for interv al length variation; a cow which experiences a secondary exposure during peak response from a previous exposure to th e specific antigen will not be boosted to the degree a cow is that rece ived the antigen after peak response due to feedback inhibition. Effect of sickness on OD+2: For this analysis, there were a total of 754 cows, 40 of these cows were classified as Sick as defined with in 16 DIM. Parity was not a significant predictor of OD +2 (p = 0.54); as a result, it was not included in the model and all cows were considered together (Eq. 3-2). OD +2 = Sick + Dex + OD 0 (3-2) The analysis revealed that sickness, as prev iously defined, was a si gnificant predictor of OD +2 (p = 0.0156). The difficulty in this analysis is proving the direction of the association. Did the occurrence of sickne ss within 16 DIM cause a suppression in immune responsiveness; or did inferior immune responsiveness cause sickness within 16 DIM. Because OD 0 occurs prior to the incidence of disease, and this was a significant predictor of OD +2 (p < 0.0001), you can be fairly certain the incidence of sickness had an effect on immune responsiveness. Correlation analysis was employed to study the relationship between OD 0 and OD +2 at fixed levels of sickness. Among cows considered healthy within 16 DIM, correlation analysis revealed OD 0 is positively correlated with OD +2 (r 2 = 0.64, p < 0.0001). Furthermore, within sick cows, the correlation between OD 0 and OD +2 has an even greater significantly positive correlation (r 2 = 0.73, p < 0.0001). For further understanding, an additional li near regression model was constructed. The rationale behind this analysis comes from OD 0 and OD +2 being positively correlated in sick as 64

PAGE 65

well as healthy cows. If it was the level of i mmune response that caused susceptibility to sickness, the strong positive correlation should allow OD 0 to be able to se rve as the outcome variable. It is known that OD 0 reflects an immune response occurring prior to the high risk period for sickness. For this analysis, parity was significant (p < 0.0001). As a result, this analysis was performed at fixed levels of parity. For the 333 primiparous cows the independent variables remaining in the model were, Sick, and OD +2 This analysis reveals sickne ss is not a predictor for OD 0 (p < 0.387). For the 422 multiparous cows the independent variables remaining in the model were Sick, and OD +2 This analysis also reveals sickness is not a predictor for OD 0 (p < 0.592). For further evidence, it will be discussed later how models which do not in clude the +2wk antibody response data tend to have a stronger association with disease risk. Alternate Index Methods for AMIR Categorization Although only similar to the AMIR index (Eq. 11) in Wagter et al. (2000), a comparable index was generated (Eq. 3-3). This was chosen based on the util ity of the previous index. In similar fashion, another index (Eq. 3-4) was derive d due to the previously described effect of early postpartum sickness on immune responsiveness. y total = I 1 + 1 I 2 + 2 I 3 (3-3) y total = I 1 + 1 I 2 (3-4) Where: y total = total antibody I 1 = change in optical density (OD) between week -8 and week -3 I 2 = change in OD between week -3 and week 0 I 3 = change in OD between week 0 and week +2 1 & 2 = either 1.0 or 1.5 65

PAGE 66

The rationale behind two additional indexes is based on the assumption that a favorable AMIR will have greater correla tion with the actual magnitude of the antibody concentration rather than changes in antibody c oncentration over inte rvals. In the case of a maximal antibody response, an index should reflect a cows ability to maintain a high concentration of specific antibody. This should all be accomplished while also using measures which give special attention to antibody respons es occurring peripartum. The first index (Eq. 3-5) includes the postpartum OD +2 while the second (Eq. 3-6) does not. In these indexes, the direct magnitudes of the OD values are considered. However, in the case of OD 0 and OD +2 they are still weighted, but this can be positively or negatively and only in proportion to the level of increase or decrease for I 2 and I 3 If I 2 is slightly negative, then OD 0 is multiplied by a number slightly under 1 yielding a smaller value. If I 2 is slightly positive, then OD 0 is multiplied by a number slightly over 1, yielding a larger value. y total = OD -3 + OD 0 *(1 + I 2 ) + OD +2 (1 + I 3 ) (3-5) y total = OD -3 + OD 0 *(1 + I 2 ) (3-6) Where: y total = total antibody OD -3 = optical density value at wk-3 OD 0 = optical density value at wk0 I 2 = change in OD between week -3 and week 0 I 3 = change in OD between week 0 and week +3 For each index (Eq. 3-3, Eq. 3-4, Eq. 3-5, Eq. 3-6), 2 different appro aches were taken to extrapolate AMIR categoriza tions from the generated y total values. In each approach the AMIR categorization was determined within parity due to the effect of parity on antibody responsiveness to OVA (Figure 3-2). The first approach involved calculating the mean and 66

PAGE 67

standard deviation of all y total values (Figure 3-11) (Wagter et al., 2000, Hernandez et al. 2003). This was performed separate for multiparous a nd primiparous cows. High responders were those cows with y total values greater than the mean plus one standard deviation. Low responders were those with y total values less than the mean minus one standard deviat ion. Medium responders had y total values within one standard deviation. If dealing with a normal distribution of data, roughly 68% of the data will fall within 1 stan dard deviation of the mean. This leaves roughly 16% for high AMIR responders and 16% for low AMIR responders. Because this method only categorizes the extr eme 32 % of the population into high or low immune responders, calculations of quartiles were used. This method allowed us to set the top 25% of data as high AMIR, while the bottom 25% as low AMIR. The resulting middle 50% was classified as medium responders. In this case, the extreme 50% of the population is categorized as high or low responders. The configuration of the quartiles was also calc ulated within parity. With 4 different possible equations (Eq. 33, 3-4, 3-5, 3-6) to generate antibody y total values and 2 different methods to extrapolate AMIR classifications from each equation, an analysis of each possibility is required to determine which me thod is chosen. For this determination a raw association between each method and incidence of disease was used (Table 3-1). Results The worst association with disease is clea rly Equation 3-5 using the SD classification method (Table 3-1). It appears E quation 3-3 has a closer associati on with resistance to mastitis and retained fetal membrane. However, Equa tion 3-4 and Equation 3-6 are more closely associated with resistance to metabolic conditions. These two equations do not include OD +2 data from early postpartum. Use of Equation 3-3, Eq uation 3-4, and Equation 3-6 appear to be the best options. Use of standard deviation ve rsus use of quartiles appears negligible. 67

PAGE 68

Using incidence of mastitis for an entire lact ation as a binary outcome variable, logistic regression using the LOGISTIC Procedure of SA S was used to identify the equation which can be best used to predict susceptibility to mastitis Independent variables remained in the model if they showed a tendency (p < 0.10) to predict mastitis inciden ce. The potential model effects were: BCS x = categorization of body condition score for sampling week x BCS y = change in body condition score over interval y Dex = binary effect, whether cow recei ved dexamethasone prior to calving Parity = binary effect, either primiparous or multiparous SCCavg = the average SCC collected monthly for the first 10 months ARC = antibody response categorization using e ither the standard deviation or quartile method for Equation 3-3, Equation 3-4, or Equation 3-6. The effects which remained in the model were: SCCavg, BCS 5 Parity, and ARC. Equation 3-3 was not a significant predictor for ma stitis using either the standard devation (p = 0.726) or the quartile method (p = 0.701). Equation 3-4 showed a tendency to predict mastitis incidence using the standard deviation (p = 0.116) and quartile method (p = 0.089). Equation 3-5 was not a significant predictor for mastitis using either method (p = 0.816, p = 0.893). Using Equation 3-6, the standard deviation method was not a significant pr edictor (p = 0.3646), however, using the quartile method, Equation 36 was a significant predictor for mastitis incidence (p= 0.028). The two equations revealing any ability to predict mastitis incidence did not include early postpartum (wk+2) measurements for antibody respons e. These methods appear to have a closer association with susceptibility to mastitis. E quation 3-6 was the only equa tion with a significant ability to predict mastitis. This is also the onl y equation which alleviates concerns about antibody 68

PAGE 69

saturation and the effect of early postpartum sickness. In light of these findings, in an effort to streamline the method in which AMIR categoriz ation is accomplished, Equation 3-6 using the quartile method will be used from here on out. 69

PAGE 70

Figure 3-1. Basic outline for treatment/ sampling for antibody mediated immune responsiveness. 70

PAGE 71

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8-8 -3 0 +2 Weeks Relative to CalvingOptical Density Figure 3-2. Optical density values reflec ting antibody response to OVA by sampling period separated by parity. A) Solid line is for multiparous cows. B) Dotted line is for primiparous cows. Bars indicate 95% confidence intervals. 71

PAGE 72

0.000 0.500 1.000 1.500 2.000 010203040506 Days Interval 1OD-3 Value 0 Figure 3-3. Scatter plot of OD -3 values by length of interval 1 for primiparous cows. Fitted line not significant (p = 0.34, = -0.0015). 72

PAGE 73

0.000 0.500 1.000 1.500 2.000 2.500 0102030405060708 Days Interval 1OD-3 Value 0 Figure 3-4. Scatter plot of OD -3 values by length of interval 1 for multiparous cows. Fitted line is significant (p < 0.0001, = -0.0081). 73

PAGE 74

0.000 0.500 1.000 1.500 2.000 2.500 01 02 03 04 05 Days Interval 2OD0 Value 0 Figure 3-5. Scatter plot of OD 0 values by length of interval 2 for primiparous cows. Fitted line is significant (p < 0.0001, = -0.0246). 74

PAGE 75

0.000 0.500 1.000 1.500 2.000 2.500 01 02 03 04 05 Days Interval 2OD0 Value 0 Figure 3-6. Scatter plot of OD 0 values by length of interval 2 for multiparous cows. Fitted line is significant (p < 0.0001, = -0.0211). 75

PAGE 76

0 0.5 1 1.5 2 2.5 10 12 14 16 18 20 22 Days Interval 3OD+2 Value Figure 3-7. Scatter plot of OD +2 values by length of interval 3 for primiparous cows. Fitted line is significant (p < 0.0032, = -0.0298). 76

PAGE 77

0 0.5 1 1.5 2 2.5 10121416182022 Days Interval 3OD+2 Value Figure 3-8. Scatter plot of OD +2 values by length of interval 3 for multiparous cows. Fitted line is not significant (p = 0.21, = -0.0104). 77

PAGE 78

-0.100 0.000 0.100 0.200 0.300 0.400 0.500 0.600 1234567891011121314 Ranking by OD-3 ValueChange in OD Over Interval 2 Figure 3-9. Graph reflecting the effect of maximal antibody response. Cows were ranked based on their OD -3 value (x-axis) with 55 cows pe r group ranking except group 14 which has 59 cows. Cows in group 1 have the top 55 OD -3 values. Cows in group 14 have the bottom 59 OD -3 values. The y axis represents OD 0 OD -3 Data points indicate the average change in OD over interv al 2 for each group. Bars reflect SEM 78

PAGE 79

-0.100 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 0.800 0.900 1.000 1234567891011121314 Ranking by OD0 ValueChange in OD Over Interval 3 Figure 3-10. Graph reflecting th e effect of maximal antibody response. Cows were ranked based on their OD 0 value (x-axis) with 55 cows pe r group ranking except group 14 which has 39 cows. Cows in group 1 have the top 55 OD 0 values. Cows in group 14 have the bottom 39 OD 0 values. The y axis represents the change in OD over Interval 3 (OD +2 OD 0 ). Data points indicate the average change in OD over interval 3 for each group. Bars reflect SEM. 79

PAGE 80

Figure 3-11. Depiction of a method used to categorize antib ody mediated immune responsiveness. Mean and standard deviation are calculated based on y total 1 values for the population. With a normal distribution roughly 16% of the population becomes high responders, 16% become low responders, and roughly 68% are medium responders. 1 y total total antibody response value ge nerated through the use of an immune response index (Equation 3-3, E quation 3-4, Equation 3-5, Equation 3-6). 80

PAGE 81

Table 3-1. Incidence of di sease for each antibody re sponse categorization method. Mastitis Equation 3-3 1 Equation 3-4 2 Equation 3-5 3 Equation 3-6 4 Categorization St.Dev 5 Quart 6 St.Dev Quart. St.Dev Quart. St.Dev Quart. Low 27.45 27.98 23.73 24.12 25.69 27.71 24.35 23.35 Medium 26.93 27.86 28.05 29.33 26.64 26.30 27.39 30.14 High 22.83 21.95 23.01 22.84 26.42 25.47 24.77 21.74 Metritis Equation 3-3 Equation 3-4 Equation 3-5 Equation 3-6 Categorization St.Dev Quart. St.Dev Quart. St.Dev Quart. St.Dev Quart. Low 4.50 5.32 7.69 7.45 6.56 5.32 7.69 6.38 Medium 5.08 4.79 4.08 3.46 4.17 4.26 4.06 3.72 High 7.41 6.42 7.63 6.95 8.80 7.49 7.81 7.49 Metritis in cows having retained fetal membrane. Equation 3-3 Equation 3-4 Equation 3-5 Equation 3-6 Categorization St.Dev Quart. St.Dev Quart. St.Dev Quart. St.Dev Quart. Low 22.22 37.50 43.75 42.11 45.45 33.33 43.75 36.84 Medium 32.50 21.43 20.69 19.05 22.86 20.83 21.43 20.00 High 0.00 27.27 20.00 20.00 22.22 30.77 18.18 25.00 Ketosis Equation 3-3 Equation 3-4 Equation 3-5 Equation 3-6 Categorization St.Dev Quart. St.Dev Quart. St.Dev Quart. St.Dev Quart. Low 4.50 4.79 6.92 7.45 4.92 5.85 9.23 7.45 Medium 4.68 5.03 4.67 4.50 4.74 4.76 3.84 4.23 High 2.75 2.66 0.76 1.06 2.38 2.13 1.55 1.60 Displaced Abomassum Equation 3-3 Equation 3-4 Equation 3-5 Equation 3-6 Categorization St.Dev Quart. St.Dev Quart. St.Dev Quart. St.Dev Quart. Low 4.50 4.26 3.08 3.72 4.92 4.26 3.85 3.19 Medium 2.06 1.85 2.24 2.12 1.78 1.85 2.02 2.38 High 3.67 2.66 3.79 2.66 3.97 2.66 3.88 2.66 Retained fetal membrane Equation 3-3 Equation 3-4 Equation 3-5 Equation 3-6 Categorization St.Dev Quart. St.Dev Quart. St.Dev Quart. St.Dev Quart. Low 8.11 8.51 12.31 10.11 9.02 9.57 12.31 10.11 Medium 7.49 7.41 5.89 5.56 6.92 6.35 5.66 5.29 High 5.50 5.85 7.58 7.98 7.14 6.91 8.53 8.51 1 Equation 3-3, refer to text. 2 Equation 3-4, refer to text. 3 Equation 3-5, refer to text. 4 Equation 36, refer to text. 5 St Dev.,standard deviation, refers to a method used to extrapolate antibody response categorizations (Figure 3-11). 6 Quart, quartiles, refers to a method used to extrapolate antibody response cat egorizations 81

PAGE 82

CHAPTER 4 CATEGORIZATION OF PERIPARTURIENT ANTIBODY RESPONSE TO OVALBUMIN AND ITS RELATIONSHIP WITH CO MMON DISEASES AND PERFORMANCE MEASURES OF HOLSTEIN DAIRY CATTLE Introduction The abrupt transition from a pregnant non-lactating cow to a non-pregnant lactating cow has a deleterious effect on immune function around the time of calving. This periparturient immune suppression is well documen ted (Lacetera et al., 2005; Saad et al., 1989; Detilleux et al., 1994; Park et al., 1992) and believed to be at least partially respons ible for the increased risk of disease during this period. This effect appears to have several contributing mediators such as, increased stress and hormones associated with st ress (Van Kampen and Mallard 1997; Mallard et al., 1997), hypocalcemia (Kimura et al., 2006), negative energy balan ce with the resultant effect of nonesterified fatty acids (Lacetera et al., 2005) and h yperketonemia (Franklin et al., 1991; Hoeben et al., 1997). Although these mediators have a significant e ffect on peripartum immune responsiveness, there also appears to be a gradual decline in immune competence associated with genetic selection focused on production traits which is made evident by the increasing risk of disease (Harmon, 1994; Heringstad, 2000; Emanuelson, 1988). Th is trend is occurring despite the efforts and major advances in sanitation and housing. This effect is believed to be fueled by placing the priority for genetic selection toward increased milk yield with little selection pressure for measures of disease resistance. Selection for increased milk production void of measures for resistance to mastitis has been found to result in a genetic increase of 0.02 cases of mastitis per cow per year (Strandberg and Shook, 1989). This transl ates into a genetic increase of 2 mastitis cases for every 100 dairy cows per year. 82

PAGE 83

To overcome this unfavorable tre nd, selection pressure is being applied to traits associated with disease resistance. Selection for decreased SCC, productive life, and structural traits of the udder have been studied (Nash et al., 2000; Heringstad et al ., 2006) and further implemented. However, associations between these traits and resistance to disease are rather crude and do not directly focus on the immune system which is primarily responsible for host defense. As a result, measures of the immune function have been eval uated as indicators of health (Wagter et al., 2000; Rupp et al., 2007; Park et al., 2004). Significant associations with mastitis risk between differing populations of T lymphocyte subsets have been reported (Park et al., 1993, 2004). Several studies have found significant differences in immune responsiveness and resistance to disease with alterati ons in class I and II MHC haplotypes and alleles (Rupp et al., 2007; Park et al., 2004; Rupp a nd Didier, 2003). Other studies have compared incidence of disease among cows categorized based on their antibodymediated immune responsiveness (AMIR) to test antigen (Wagter et al., 2000; Mallard et al., 1997; Hernandez et al., 2003). Research by Wagter et. al. (2000) reported substantial va riation among individual cows ability to mount an antibody response around calvi ng. In fact, not all cows experience peripartum immune suppression and cows c ould be categorized based on AM IR to ovalbumin (OVA). The high responders for AMIR had the lowest incidence of mastitis in two of the three study herds. Individual cows antibody responsiveness to OVA al so had a positive signifi cant correlation with antibody titers to E. coli J5 vaccination. This response to OVA was highly heritable, ranging from 0.32 to 0.62 depending on week relativ e to calving (Wagter et al., 2000). Although Wagter et al. (2000) di d find favorable associations between this measure of AMIR and disease risk; with 136 co ws and heifers spread over three herds, it was difficult to find 83

PAGE 84

statistical significance. In the current trial this problem was addressed by incorporating a much larger sample size in order to more adequately study the relationship between AMIR and disease risk. The hypothesis being that high AMIR will be associated with lower disease risk. The objectives were to categorize cows based on AM IR to OVA and then test for associations between AMIR and disease incidence, namely; mastitis, metritis, retained fetal membrane, ketosis, and displaced abomasum. Additiona lly, possible associations between AMIR categorization and reproductive efficiency, milk yield and somatic cell score were examined. Materials and Methods Research Sample Cows In total, 875 Holstein cows/heifers were enrolled into the study popul ation at 8 weeks (wk8) prior to calving. In cows, this was the initiation of the dry pe riod. Animals were enrolled if the expected dry period was less than 90 days, if reconfirmed pregnant and if found in good health with no obvious signs of disease. All test animals were from a singl e herd in north central Florida which maintains exceptional record keeping. Al l cows and heifers were: enrolled between September 9 th and December 31 st 2004; calved between October 25 th 2004 and March 12 th 2005; and sampling ended between November 9 th 2004 and March 28 th 2005. All cows and heifers received a routine dry off, prefresh, and fresh cow protocol. Animal Removal and Interval Criteria Of the originally enrolled 875 cows and he ifers, 13 were removed due to missing data, yielding 862. Interval 1 (Int 1 ) was defined as time from enrollment (wk-8) to entry into springer pen (wk-3). Interval 2 (Int 2 ) was defined as time from entry in to springer pen (wk-3) to calving (wk0) (Figure 4-1). Animals were removed from the study if the period from wk-8 to wk0 (dry period length for cows) was more than 90 days Additional animals were excluded if Int 1 or Int 2 was less than 12 days in length. These time frames are relevant because they coincided with 84

PAGE 85

antigen exposure and blood sampling to measur e antibody response to an tigen. Interval 3 (Int 3 ) was defined as the period between calving (wk0) and end of the sampling period (wk+2). No exclusionary criterion was needed for Int 3 because this interval was under investigator control. Of the 862 cows, 50 were removed because they did not meet one of these interval restrictions. A preliminary analysis was used to remove an additional 38 animals found not nave to test antigen. This resulted in a sample size of 774 with 433 cows and 341 heifers. The heifers upon enrollment will be referred to as primiparous cows while cows upon enrollment will be referred to as multiparous cows. Test animals were si red by 237 different sires, which should provide adequate variation in the gene pool for a sire effect on immune responsiveness. Body Condition Scoring Body condition was scored in point increments using the 5-point scale by Ferguson et al. (1994). Body condition scoring (BCS) was grouped into high, medium, and low categories. For cows during wk-8, wk-3, and wk0, a BCS between 3.0 3.75 was coded medium, those below 3.0 were coded low and those above 3.75 were coded high. Heifers at wk-8, wk-3, and wk0 were coded medium if BCS was between 3.0 3.5. A BCS above 3.5 was considered high, and a BCS below 3.0 was considered low. At wk+2, all anim als were coded normal if BCS fell in the range of 2.75 3.5. The loss of body condition, not just BCS alone was reported to be responsible for alterations in lymphocyte function (Lacetera et al., 2005; Wentink et al ., 1997; Kaneene et al., 1997). To account for this factor the interval chan ge in BCS for various intervals was calculated. Interval 4 was defined as the period between wk-3 and wk+2, which serves as an indicator of BCS lost over the transition period. Interval 5 wa s defined as the period between wk-8 and wk0. Interval 6 was defined as the period between wk-8 and wk+2. 85

PAGE 86

Immunization Ovalbumin was chosen as the test antigen due to its inert properties, its ability to stimulate antibody, a cows reduced likelihood of previous e xposure and its previous success as a tool to categorize AMIR (Wagter et al. 2000; Mallard et al., 1997; Hernandez et al., 2003). Animals received 1 mg ovalbumin (OVA; chicken albumi n, Type VII, Sigma-Aldrich, St. Louis, MO, USA) at three times; week -8 (w k-8), week -3 (wk-3), and week 0 (wk 0) relative to calving (Figure 4-1). The wk-8 and wk-3 OV A immunizations were dissolved in Escherichia coli J5 vaccine (J5 bacterin, Pfizer Inc., Kalamazoo, MI USA) with the manufacturers adjuvant which coincided with the farms routine vaccine protocol. At wk0, the 1 mg OVA was dissolved in 1 ml phosphate buffer saline (PBS, pH 7.4). All su spensions were then mixed with a type 1( C. albicans raw whole cell material, Greer Laboratories, Lenoir, NC, US A) antigen to stimulate a cell-mediated immune response (C MIR) and vortexed for at least one minute (Refer to chapter 5 for CMIR analysis). Blood Collection and Processing To determine antibody response to OVA, blood was collected by caudal venipuncture at; wk-8, wk-3, wk0, and wk+2 relative to calving (F igure 4-1). At calving (wk0), the blood sample was collected within 12 hours of pa rturition. Samples were collected into sterile 10 ml evacuated blood collection tubes with no additive, then put on ice during transport back to lab. Serum was collected after centrifugation at 4,000 rpm and stored at -70 C. Enzyme-Linked Immunosorbent Assay A cows specific antibody response to OVA was de tected using an indirect enzyme-linked immunosorbent assay (ELISA) method as previous ly described (Burton et al., 1989; Wagter et al., 2000; Mallard et al., 1997). For positive cont rol sample, 10 lactating cows received 1 mg OVA and 0.5 mg Quil A adjuvant suspended in 1 ml PBS on day 0 and 14. On day 21, serum 86

PAGE 87

from these cows was collected and pooled for positive control sample. Negative control samples consisted of a pool of serum from cows at wk-8. Polystyrene 96-well plates (Immulon II, Fisher Scientific Co. Ltd., Pittsburgh, PA, USA) (Figure 4-2) were stored for 48 h at 4 C after being coated with 100 L/ well of 1.4 mg OVA dissolved in 1 mL carbonate-bicar bonate coating buffer (pH 9.6). Plates were washed four times in a plate washer (ELX50 plate washer, Biotek Instruments, Inc., Winooski, VT, USA) with wash buffer solution containing PBS and 0.05 % Tween 20 (Sigma-A ldrich) (washing buffer, pH 7.4). Blocking solution (PBS pH 7.4, 3% Tween 20, 1% bovine serum albumin) was added (200 L / well) and plates incubated at room temperat ure for 1 h. Plates were washed 4 times before applying 100 L / well of control and test se ra. All samples were diluted to 1/50 and 1/200 dilutions using wash buffer solution. Positive and negative controls (1/50 and 1/200) were run in quadruplicates while test sera (1/ 50 and 1/200) were run in duplicat es placed in separate diagonal plate quadrants (Figure 4-3). Plates were then in cubated for 2 h at room temperature. After being washed 4 times, 100 L / well of alkaline phosp hatase-conjugated rabb it anti-bovine IgG (whole molecule; Sigma Chemical Co., St. Louis, MO) dissolved in Tris buffer solution (TBS, pH 7.4 and 0.05% Tween) at a 1/38000 dilution was added, and plates incubated for 1 h at room temperature. Plates were washed 4 times be fore 80 L / well of p-nitrophenyl phosphate disodium were added. Plates were incubated for 30 min at room temperature out of direct light. Optical density (OD) values were then determined for each sample with an ELISA plate reader (MRX Revelation, Dynes Technologies VA, USA) set at an absorbance of 405 m and 630 m (Revelation software, Dynes Technologies VA, USA). Coefficient of variation was calculated fo r the 1/50 and 1/200 positive controls to determine whether plates were accepted or rejected. The coefficient of variation for the 1/50 87

PAGE 88

dilution was calculated by dividing the standard deviation by the mean of the 1/50 positive control values. This was also performed for the 1/200 dilution. The maximum allowable variation for each plate was 20% for either the 1/50 or 1/200 positive control dilution. In order to correct for variation among plates a correction factor was determined for each plate. The correction factor was determined by comparing the positive control OD values of each plate to the mean of every plate. The mean of each plates 1/50 positive control dilution was summed with the mean of the 1/200 positive control dilution. This additive positive control value was then divided into the mean of every plates additive value. The resultant value functioned as the correction factor. Since each test sample was run in duplicate, each test samples mean 1/50 OD value was summed with the mean 1/200 OD value. This value was then multiplied by the plates correction factor to determine the samples co rrected OD value to be utilized in statistical analysis. Preliminary Analysis Removal of non-naive Due to the differences in the kinetics of a primary and sec ondary antibody response (Figure 2-1), it was important that all cows receive equal treatment which requires all cows to be nave to OVA at enrollment. This should result in low OD values for wk-8 (OD -8 ) due to the lack of antibody with affinity for OVA. Inspection of the OD -8 values revealed that some cows were not nave. Sorting the cows based on their OD -8 values revealed a natural cut-off where the values began increasing rapidly. This point also coincided with one sta ndard deviation above the mean for the OD -8 values. As a result 38 cows were excluded due to not being nave to test antigen. For further discussion refer to chapter 3. 88

PAGE 89

Parity effect A repeated measures analysis using OD -8 as a covariate with the mixed procedure of SAS revealed that multiparous cows responded significantly higher than primiparous cows at every antibody response measurement week (p<0.0001). Due to this effect, adjustments to OD values as well as AMIR categorizations were all made with respect to parity (F igure 3-2). Explanations for this finding may include the presence of a mo re extensive antibody repe rtoire in older cows. This could simply be due to the effect of time, allowing greater exposure to a broader array of pathogens. Interval analysis and optical density value adjustment The length of the interval between antigen administration and blood sampling has strong relevance due to the different phases of an an tibody response (Figure 21). As a result, an analysis was performed to determine this effect and subsequently adjust the OD values for interval length (Figure 3-3, 34, 3-5, 3-6, 3-7, 3-8). A generalize d linear model was constructed using the GLM procedure of SAS to determine if the duration in days of Int 1 Int 2 or Int 3 significantly influenced OD -3 OD 0 or OD +2 respectively. For this analysis, OD response served as the outcome variable while possible explanatory variables included: Int Y = interval length in days for interval y BCS X = categorization of body condition score for sampling week x BCS Y = change in body condition score over interval y OD X = optical density at previous sampling week x Dex = binary effect, whether cow recei ved dexamethasone prior to calving Sick = binary effect, incidence of either medium or severe case of mastitis, metritis, ketosis, or displaced abomasum within 16 DI M. Only used for analysis to correct OD +2 for reasons discussed in chapter 3. 89

PAGE 90

Explanatory variables remained in the model if the effect showed a tendency (P < 0.10) to predict the desired OD value. With = 0.05, if statistical significance was found with the main effect (Int y ), the resultant parameter estimate was applied to Equation 3-1 to correct the corresponding OD values. If a given OD value was adjusted based on interval length, the model predicting the OD value for the subsequent samp ling period would have the adjusted OD from the previous sampling period. AOD ZPX = OD ZPX + PE YP (Mint YP Int ZPY ) (3-1) Where: AOD ZPX = the adjusted OD value for cow z in parity p for week x OD ZPX = OD value for cow z in parity p for week x PE YP = parameter estimate for the effect interv al y has on ODx values for parity p. Also, interval y must always immediately precede week x. Mint YP = median number of days for interval y for parity p Int ZPY = the actual interval length for cow z in parity p for interval y As a result the OD -3 values were adjusted for multiparous cows only. The OD 0 values were adjusted for both multiparous and primiparous cows. The OD +2 values were adjusted for heifers only. For further reference refer to chapter 3. Classification analysis Maximal response: Previous work devised an index usi ng the interval change in OD value to generate total antibody values used to extr apolate AMIR classificati ons (Wagter et al. 2000). This index weighted those values around calving if they showed a decline in antibody response. In this study, inspection of the OD values for the respective weeks revealed what appeared to be a maximal antibody response. If there was an improvement after subsequent antigen injection, these maximal points in antibody concentration only slightly improved. If this peak response was 90

PAGE 91

achieved early in the study trial (OD -3 ), which would indicate a hi gh prepartum responder; there would be little if any room for an added response. As a result an index concerned with weighting the change in OD during the intervals adjacent to calving (Eq. 1-1), may negatively impact the categorization of great immune responders who reach ed this saturation poin t early because there was little or no room to further respond. After sorting the OD -3 values from greatest to least, it was discovered that of the top 20 cows for OD -3 value, 15 had a smaller OD 0 value (75%), yielding a nega tive interval 2 change in OD. In these instances, 15 of the top 20 OD -3 responders would have an amplified ( = 1.5) negative I 2 value applied to their AMIR index if usi ng an index which weights interval changes in OD. After grouping the OD -3 values from 1 to 14, with group 1 being the top 55 OD -3 values and group 14 the bottom 59 OD -3 values, a one-tailed two sample T-test was used to determine if the OD value level grouping had a significant effect on the subsequent inte rval change in OD. This revealed a significantly smaller change in OD over interval 2 for group 1 cows compared to group 2 cows (p < 0.0001) (Figure 3-9). This va lue actually averaged below 0 (-0.05) for these top 55 OD -3 responders. Linear regr ession also revealed OD -3 is a significant predictor for I 2 (p < 0.0001; = -0.28). A negative indicates that as OD -3 values increase, I 2 values decrease. Repeating the same process by ranking cows based on OD 0 values in order to compare interval 3 change in OD, revealed similar results. Due to missing wk+2 blood samples the sample size for this analysis was 754. As a result group 14 was comprised of 39 cows while group 1 13 comprised of 55. Of the top 20 cows for OD 0 values, 15 had negative interval 3 changes in OD (75%). Also, group 1 cows based on OD 0 values had a significantly smaller interval 3 change in OD compared to group 2 co ws (p < 0.0001) (Figure 3-10). Linear regression 91

PAGE 92

also revealed OD 0 is a significant predictor for I 3 (p < 0.0001; = -0.50). A negative indicates that as OD 0 values increase, I 3 values decrease. In these instances of high OD values follo wed by a subsequent negative interval, the interval is not negative due to a poor subs equent OD value. Of the top 20 cows for OD -3 18 still had OD 0 values above the third quartile for the popul ation and the other 2 were still above the median for the population. For the top 20 cows for OD 0 all 20 still had OD +2 values above the third quartile for the population. The negative interval was simply the result of an inability to respond further. As a result, additional indexe s were generated and analyzed based on their correlation with disease incidence. Early postpartum measures of immune function: Previous work utilized the antibody response measured early postpartum as a tool to categorize AMIR (Wagter et al., 2000; Mallard et al., 1997).This period immediat ely following parturition is a common occasion for increased incidence of disease. This effect of substa ntial sickness could be a confounding variable for antibody responsiveness to OVA detected early po stpartum. As a result it may be hypothesized that substantial sickness could contribute to immune suppression. This would make it difficult to study the effect measures of imm une responsiveness have on disease risk if the association could also be in the opposing direction. A linear regression model was used w ith the REG procedure of SAS. OD +2 served as the outcome variable while potential explanatory variables included: BCS x = categorization of body condition score for sampling week x. BCS y = change in body condition score over interval y. OD 0 = optical density value at wk0. Dex = binary effect, whether cow recei ved dexamethasone prior to calving. 92

PAGE 93

Sick = binary effect, incidence of either medium or severe case of mastitis, metritis, ketosis, or displaced abomasum within 16 DIM. Parity = binary effect, either primiparous or multiparous. Explanatory variables remained in the model if the effect showed a tendency (P < 0.10) to predict the dependant OD value. For statistical significance, = 0.05. The resulting model was as follows (Eq. 3-2): OD +2 = Sick + Dex + OD 0 (3-2) The analysis revealed that sickness, as prev iously defined, was a si gnificant predictor of OD +2 (p = 0.0156). The difficulty in this analysis is determination of causality, did the occurrence of sickness within 16 DIM cause a suppression in i mmune responsiveness; or did inferior immune responsiveness cause sickness within 16 DIM. Because OD 0 occurs prior to the incidence of disease, and this was a significant predictor of OD +2 (p < 0.0001), it appeared that the incidence of sickness had an effect on immune responsiveness. Correlation analysis using the Corr procedure of SAS was employed to study the relationship between OD 0 and OD +2 at fixed levels of sickness. Among cows considered healthy within 16 days in milk, corre lation analysis revealed OD 0 is positively correlated with OD +2 (r 2 = 0.64, p < 0.0001). Furthermore, within sick cows, the correlation between OD 0 and OD +2 has an even greater significantly positive correlation (r 2 = 0.73, p < 0.0001). For further reference see chapter 3. Antibody-Mediated Immune Response Classification To alleviate the concerns about the effect of antibody saturation and early postpartum sickness, a new index was generated for the ca tegorization of AMIR (Eq. 3-6). The rationale behind this index is that favorable AMIR will ha ve greater correlation with the actual magnitude of the antibody concentration rather than change s in antibody concentration over intervals. In the 93

PAGE 94

case of antibody saturation, an index should re flect a cows ability to maintain a high concentration of specific antibody. This should al l be accomplished while also using measures which give special attention to antib ody responses occurring peripartum. y total = OD -3 + OD 0 *(1 + I 2 ) (3-6) Where: y total = total antibody OD -3 = optical density value at wk-3 OD 0 = optical density value at wk0 I 2 = change in OD between week -3 and week 0 In this index, the direct magnitudes of the OD values are considered. However, in the case of OD 0 it is still weighted, but this can be positively or negatively and only in proportion to the level of increase or decrease for I 2 and I 3 Extrapolation of AMIR categorizations occurs by use of the y total values for individual cows and then configuring the quartiles respective of parity (primiparous vs. multiparous). Cows within the bottom 25% for a respective parity are categorized as low AMIR responders, while cows in the top 25% for a resp ective parity are termed high AM IR responders. The remaining middle 50% are categorized as medium AMIR responders. Diseases Identification of disease was performed by fa rm personnel who were blinded to immune response categorizations. The diseases of interest fo r this project were; mastitis, metritis, retained fetal membranes, ketosis, and displaced abomasum. All diseases were recorded as yes/no binary responses for the trial peri od of the current lactation. Mastitis was recorded through 365 DIM as light, medium, or severe. Severity was recorded as light if there were no systemic signs and milk was slightly watery with minimal clots (gargot). 94

PAGE 95

Severity was medium if there were no systemic signs and a substantial amount of clots were observed in the milk. Severity was severe if there were systemic signs, watery milk, and a substantial amount of clots in the milk. Metritis was recorded through 30 DIM as light, medium or severe. Severity was light if there was an abnormal vaginal discharge and a pa lpable uterine lumen. Severity was medium if there was a purulent vaginal discharge, with an enlarged, not-flaccid uterus. Severity was severe if there was a purulent foul-smelling vaginal di scharge, with an en larged flaccid uterus. Association between AMIR a nd energy related metabolic c onditions including ketosis and DA were also analyzed. Ketosis was recorded through 30 DIM as light, medium, or severe. Ketosis was coded as light if the urine cont ained 15 mg/dL of ketone bodies. Severity was medium if the urine contained 40 mg/dL of ketone bodies. The severity was severe if the urine contained > 80 mg/dL of ketone bodies. Displa ced abomasum was recorded through 50 DIM. Because retained fetal membranes are now understood to result from an inadequate immune response (Kimura et al., 2002), the re lationship between AMIR and RFM was also considered. Retained fetal membranes were id entified if there was placental retention 24 h postpartum. Milk Yield, Somatic Cell Score, and Reproductive Efficiency Milk yield was gathered from the Dairy Herd Information Association (DHIA) records for the current lactation us ing ME305, which is an estimate of the milk yield for 305 DIM. For categorical data, low milk producer s were identified if there 305 day milk was in the bottom 25% of the study group. High milk producers were th ose in the top 25% of the study group. The remaining middle 50% were classifi ed as medium milk producers. 95

PAGE 96

Somatic cell score (SCS) was gathered from DHIA records. The average SCS was determined for the first 3 test days, the first 6 te st days, and for the first 10 test days, which are an indication of the average SCS for the first 90, 180, and 300 DIM respectively. To analyze for associations between AMIR categorization and reproductive efficiency, a binary pregnancy term was used which simply indicated if a given cow was pregnant by 150 DIM. Addition quantitative variable measures of fe rtility included; number of days that a cow is not pregnant (days open), and number of times bred. Statistics To analyze the associations between risk of disease and reproductive efficiency with AMIR categorization, a logistic regression model was developed using the LOGISTIC procedure of SAS. Other than the main effect (AMIR categorization), possible explanatory variables include: BCS x = categorical effect of body condition score for sampling week x. BCS y = quantitative variable, change in body condition score over interval y. Dex = binary effect, whether cow recei ved dexamethasone prior to calving. Parity = binary effect, either primiparous or multiparous. RFM = binary effect, incidence of retained fetal membrane. CDiff = binary effect, if reported difficult calving. Ketosis = binary effect, incidence of ketosis. SCSavg = somatic cell score average over 10 monthly test days. For analyzing the relationship between AMIR categorization and the quantitative variables for SCS, 305 day milk yield, and reproductive efficiency, linear regression was used with the REG procedure of SAS. All categorical variables will be analyzed using logistic regression using the LOGISTIC procedure of SAS. All relevant effects were put in the model. A backwords elim ination procedure was used to determine the final model. Explanatory variable s remained in the model if they showed a 96

PAGE 97

tendency (p < 0.10) to predict the outcome variab le. Statistical significance was determined by setting = 0.05. Results Mastitis and Metritis No significant association was found between AMIR status and mastitis within 100 DIM. When considering the incidence within 365 days; there were 169 cases that were either medium or severe (22%). Forty (40) cases were reco rded in primiparous cows and 129 in multiparous cows. Effects remaining in the model were: SCS average, BCS 6 Parity. Antibody response categorization was a significant predictor of mo derate and severe mastitis risk (p = 0.0082) (Table 4-1). Although the low responders were no t statistically different than medium and high responders collectively (p = 0.12), the medium responders were 1.76 (CI = 1.08 2.89) times more likely to have an occurrence of moderate or severe mastitis than high responders. The recorded incidence for light medium or severe metritis during the first 30 DIM was remarkably low with only 41 cases; 27 cases in primiparous cows and 14 cases in older cows (5.3% overall). Remaining model effects were: Parity, BCS 0 Dex, and RFM. Antibody response categorization was not significantly associated with occurrence of metritis in this analysis (p = 0.29) (Table 4-1). Ketosis Displaced Abomasum and Retained Fetal Membrane Only 45 cases of ketosis occurred within 30 days of calving (5.8%). Primiparous cows accounted for 24 cases while multiparous accounted for 21 cases. The model included BCS 6 and the main effect (AMIR category). The low responders were 2.90 (CI = 1.10 7.62) times more likely to develop ketosis than hi gh responders (Table 4-1) (Figure 4-4). 97

PAGE 98

Displacement of the abomasum occurred in 21 cows (2.8%), with 12 from primiparous cows and 9 cases in multiparous cows. Remaining model effects were BCS 4 and ketosis. In this analysis, AMIR was not a significant predictor of DA incidence (Table 4-1). There were 57 cases of RFM for the study popul ation (7.4%). Primiparous cows accounted for 18 cases while multiparous cows accounted for 39 cases. Remaining model effects were: BCS 5 and BCS 0 The AMIR categorization was not a si gnificant predictor of RFM incidence (p = 0.17) (Table 4-1). Milk Yield and Somatic Cell Score For the analysis of the effect of AMIR cate gorization on milk yield, the contributing model effects were parity, number of days open, and the bi nary trait of mastitis. In this analysis, AMIR categorization had a tendency to predict milk yield (p = 0.06, = -347.34) (Figure 4-5). The analysis of the effect of AMIR on SCS included BCS 6 milk categorization, and the binary mastitis variable. The effect of AMIR on SCS wa s not a significant pred ictor of SCS (p = 0.40). Reproductive Efficiency The model analyzing the effect of AMIR cat egory on pregnancy at 150 DIM included the explanatory variables; BCS 4 and milk categorization. The e ffect of AMIR was a significant predictor for pregnancy by 150 DIM (p = 0.003). The low antibody-mediated responders were 2.32 (CI = 1.44 3.75) times more likely than high responders to become pregnant by 150 DIM. Also, the low antibody-mediated responders we re 1.57 (CI = 1.20 2.05) times more likely to become pregnant by 150 DIM than medium and low responders collectively (Figure 4-6). Discussion The significantly higher odds of mastitis for medium antibody-mediated responders compared to high responders indicate the va lidity of using antibody response to OVA as a measure of AMIR. These results coincide with the findings in Wagter et al. (2000) where low 98

PAGE 99

responders had the highest occurrence of mastitis, and antibody response to OVA was significantly correlated w ith antibody response to E. coli J5 vaccination. The lack of additional statistical findings for mastitis and metritis ar e likely due to low recorded disease frequency. Mechanisms linking energy-related metabolic disorders and measures of immune response are not completely understood. A ssociations between ketosis a nd suppressed immune response are largely believed to be the result of ketosis causing immune suppression. However in this study, AMIR categorization occurred pr epartum. Given that ketosis is a postpartum disorder, it is not possible that ketosis caused a suppressed immune response during categorization. With change in BCS remaining in the model, prepartu m BCS did not predispose cows to ketosis while influencing immune responsiveness. The odds for this effect followed a High < Medium < Low pattern. The discovery of inverse associations betw een AMIR categorization and milk yield and fertility was an unexpected result. With regard to milk yield, these findings do not agree with previous literature in dairy cows (Wagter et al., 2003; Detilleux et al., 1995) or with other performance measures in swine (Mallard et al., 1998). However, possible explanations may arise from studies in other metabolically active sp ecies. Poultry selected for greater immune responsiveness have reported a reduction in growth performa nce (Klasing et al., 1987, 1998; Soler et al., 2003). This was explained by bodily function competition for available nutrients. The energy and nutrients required for maintena nce and activation of a superior immune responder could otherwise be used for other phenotypic traits. There is no known biological rele vance for high immune responde rs to be predisposed to lower milk yields or decreased fertility. This finding may simply be the result of neglecting to 99

PAGE 100

select or having inadequate tools to select for immune responsiveness while putting direct selection pressure on the metabolica lly demanding trait of milk yield. There is another possible explanation for the unf avorable decline in fertility. This involves the maternal recognition of the conceptus by the immune system. During pregnancy the conceptus is a foreign body which otherwise would be subject to a maternal immune response followed by rejection. This reje ction, however, is blocked due to various immune suppressive activities which are initiated during maternal recognition of pregnancy (Hansen, 1997). This relationship between the immune system and pregna ncy brings the potential for an inappropriate immune response rejection of the conceptus. If it could be proven that cows who mount a superior antibody-mediated immune response are mo re likely to mount an inappropriate immune response against the conceptus, this could explain the present findings. In hindsight of this study, there are certain techniques and st rategies that could or should be implemented upon further research. Although it is difficult in a large sc ale dairy setting, and this aspect was accounted for in the present study, a tighter contro l on interval duration between antigen injection and blood co llection could be practiced. Another improvement would be in the use of antigen. A more represen tative categorization of AMIR could be obtained through two or more an tigens. This philosophy has been practiced in other species (Mallard et al.,1998). To maintain the integrity of a broad based approach to improving immune response when selection pressu re is applied for AMIR, it should be based on more than one antigen. Using the antibody response to OVA around calvi ng as a representative of the ability to mount an immune response during immune suppre ssion may be partially confounded due to the fact that OVA was previously injected at wk-8 and wk-3. This is because by this point memory 100

PAGE 101

lymphocytes have already formed for OVA and they are the cells being activated to mount the antibody response. These memory cells are much eas ier to activate and have greater affinity to antigen. For a true response during immune su ppression primary exposure to antigen should occur closer to parturition. As a result, if us ing multiple antigens, exposure to these antigens could be staggered or initially introduced at different time points with one occurring at calving. Due to the rising incidence of disease and the ever increasing necessity to produce milk as proficiently as possible, a more proactive and effective approach to disease resistan ce is required. Taking this initiative should also help alleviate concerns consumers have for animal welfare and usage of antibiotics. 101

PAGE 102

Figure 4-1. General outline of experimental design. 102

PAGE 103

Figure 4-2. Polystyrene 96well plate for enzyme linked immunosorbent assay. 103

PAGE 104

Figure 4-3. Diagram of the placement of test se ra into 96 well polysty rene plate for enzymelinked immunosorbent assay. Yellow wells identify 1/50 d ilution, while grey wells identify 1/200 dilution. All positive and negati ve controls were run in columns 1 and 7. Well identified as a is a 1/50 dilution of sample 1, which was run in column 2 of row A and duplicated in column 8 of row E. Sample identified as b is a 1/200 dilution of sample 1 run in column 2 of row B and duplicated in column 8 of row F. Matching background row colors indicate ha lf rows which were duplicated. Wells identified with B indicate blank wells for calibration of plate reader. 104

PAGE 105

Table 4-1. Odds ratios of disease inci dence for antibody res ponse categorizations. Mastitis Metritis Ketosis DA RFM 0.77 0.93 1.69 0.91 1.29 Low vs Med & High (0.55 1.07) (0.56 1.54) (1.06 2.69) (0.45 1.85) (0.86 1.94) 0.89 0.64 2.9 0.7 1.16 Low vs High (0.49 1.64) (0.27 1.56) (1.10 7.62) (0.19 2.6) (0.56 2.42) 1.76 0.521 1.75 0.65 0.63 Med vs High (1.08 2.89) (0.23 1.19) (0.68 4.45) (0.20 2.15) (0.31 1.27) Values indicate the estimated odds ratios for incidence of disease for the following comparisons among antibody-mediated immune categorizations. Values in parenthesis represent the 95% confidence intervals for the odds ratio estimate. 105

PAGE 106

0 2 4 6 8 10 12Incidence Ketosis % LowMediumHigh AMIR Categorization Primiparous Multiparous Figure 4-4. Incidence ketosis by AM IR categorization within parity. 106

PAGE 107

23000 23500 24000 24500 25000 25500 26000 LowMediumHigh AMIR CategorizationMilk Yield Primiparous Multiparous Figure 4-5. Graph for the effect of antibody-me diated immune response (AMIR) categorization on milk yield. Bars indicate SEM. 107

PAGE 108

0 10 20 30 40 50 60 70Pregnant by 150 DI M LowMediumHigh AMIR Categorization Primiparous Multiparous Figure 4-6. Graph for the effect of antibody-me diated immune response (AMIR) categorization on pregnancy by 150 DIM. 108

PAGE 109

CHAPTER 5 CATEGORIZATION OF PERIPARTURIENT CELL-MEDIATED IMMUNE RESPONSE TO A TEST ANTIGEN AND ITS RELATIONS HIP WITH COMMON DISEASES AND PERFORMANCE MEASURES OF HOLSTEIN DAIRY CATTLE Introduction The increasing risk of disease for Holstein dairy cows (Harmon, 1994; Heringstad, 2000; Emanuelson, 1988) has sparked interest in genetic selection for disease re sistance. Because the immune system is principally responsible for resi sting the array of potential pathogens; many of the methods studied have analyzed the relationshi p between disease risk and immune function or an aspect of a particular part of the immune system. Park et al. (2004) studied the ratio of CD4+ to CD8+ T lymphocyte subsets and its relationshi p with mastitis incidence. Other studies have placed special attention on characteristics of the major hist ocompatability complex (MHC) (Rupp et al., 2007; Park et al., 2004; Aaerestrup et al., 1995). So me studies have used test antigen to quantify an individuals antibody-mediat ed immune responsiveness (AMIR) as a tool to predict the risk of disease (Wagter et al., 2000; Mallard et al., 1997). An inverse relationship between cell-media ted immune response (CMIR) and AMIR has been documented (Rupp et al., 2007; Biozzi et al., 1979; de Vries, 1995). As a result, selection for increased AMIR without re gard for CMIR may confer su sceptibility to intracellular pathogens. Being mediated by T H 1 cells, delayed-type hypersensiti ve reactions (DTH) are largely concerned with elimination of intracellular an tigen. Therefore, DTH reactions have been previously used as a means to quantify CMIR (Mallard et al., 1998; Hernandez et al., 2003, 2005). However, few studies have analyzed the association between this measure of CMIR and disease susceptibility. 109

PAGE 110

The objectives of this research were to categ orize periparturient co ws using DTH reaction to a type 1 antigen and study the association it has with susceptibility to disease, namely; mastitis, metritis, retained fetal membrane (RFM), ketosis, and displaced abomasum (DA).The final objective was to examine for an effect of CMIR categorization on milk yield, somatic cell score (SCS), and fertility. Materials and Methods Research Sample Cows This research study was given IACUC approval. In total, 875 Holstein cows/heifers were enrolled into the study population at 8 weeks (wk-8) prior to exp ected calving. In cows, this was the initiation of the dry period. Animals were en rolled if the expected dry period was not longer than 90 days, if reconfirmed pregnant at enro llment and also if found in good health with no obvious signs of disease. All test animals were from a single herd in north central Florida which maintains exceptional record keeping. All cows and heifers were: enrolled between September 9 th and December 31 st 2004; calved between October 25 th 2004 and March 12 th 2005; and CMIR measurement occurred between November 2 nd 2004 and March 21 st 2005. All cows and heifers received a routine dry off, prefresh, and fresh cow protocol. Animal Removal Criteria Of the originally enrolled 875 cows and he ifers, 13 were removed due to missing data, yielding 862. Interval 1 (Int 1 ) was defined as the time from enro llment (wk-8) to entry into the springer pen (wk-3). Interval 2 (Int 2 ) was defined as the time from entry into the springer pen (wk-3) to calving (wk0) (Figure 41). Due to the restrictions of a parallel study involving AMIR, animals were removed from the study if the pe riod from wk-8 to wk0 (dry period length for cows) was more than 90 days. Add itional animals were excluded if Int 1 or Int 2 was less than 12 days in length. As a result 50 cows were rem oved leaving the study popula tion total at 812 with 110

PAGE 111

362 heifers and 450 cows. For the rest of this pape r heifers will be referred to as primiparous cows. Cows will be referred to as multipar ous cows. Test animals were sired by over 237 different sires, which should provi de adequate variation in the gene pool for a sire effect on disease resistance. Body Condition Scoring Body condition was scored in point increments using the 5-point scale by Ferguson et al. (1994). Body condition scoring (BCS) was grouped into high, medium, and low categories. For cows during wk-8, wk-3, and wk0, a BCS between 3.0 3.75 was coded medium, those below 3.0 were coded low and those above 3.75 were coded high. Heifers at wk-8, wk-3, and wk0 were coded medium if BCS was between 3.0 3.5. A BCS above 3.5 was considered high, and a BCS below 3.0 was considered low. At wk+2, all anim als were coded normal if BCS fell in the range of 2.75 3.5. It has been reported that fo r periparturient dairy cows, it was the loss of body condition, not just BCS alone that was res ponsible for alterations in lymphocyte function (Lacetera et al., 2005; Wentink et al., 1997; Kaneene et al., 1997). To account for this factor the interval change in BCS for various intervals was calculated. Interval 4 was defi ned as the period between wk-3 and wk+2, which serves as an indicator of BCS lost over the transiti on period. Interval 5 was defined as the period between wk-8 and wk0. Inte rval 6 was defined as the period between wk-8 and wk+2. Delayed Type Hypersensitivity To stimulate a DTH reaction, all cows received 0.5 mg C. albicans (raw whole cell material, Greer Laboratories, Le noir, NC, USA) at three times; wk-8, wk-3 and wk0. This was also included with a standard E. coli J5 vaccination program on wk-8 and wk-3 along with the OVA for AMIR. Within 12hours of calvi ng (wk0), animals received 0.5 mg C. albicans in 0.5 111

PAGE 112

mg Quil A (Accurate Chemicals and Scientific Corp., Westbury, NY, USA) adjuvant suspended in 1 mL PBS along with OVA for AMIR. At 1 week post-calving (wk+1), double skin-f old measurements were taken on the right and left skin folds under the base of the tail using a spring-loaded calip er (Harpenden skin-fold caliper, Creative Health Products Inc., Ann Arbor, Michigan, USA). This was performed after raising the tail 90 to a horizont al position and repeated measurements were taken 3 times. The average measurement was then recorded as ei ther the right (R0) or left (L0) time 0 h measurement. The locations of th e two measurements were cleaned with 70% isopropyl alcohol. The right side received an intradermal injection of 0.1 mg candin (C. albicans allergen extract, Greer Laboratories, Inc.) suspended in 0.1 mL PBS. The left tail fold (control side) received 0.1 ml PBS intradermally. All injections were give n with a 28 gauge needle. To identify the exact location of the measurement and injection, paper white-out solution marked the spot. Twentyfour hours later, injection sites were measured again to determ ine the increase in double skinfold thickness for the right side (R24) and left side (L24) as an indicator for the magnitude of the DTH (CMIR) response. Classification of Cell-Mediated Immune Response To obtain a normal distributi on, log transformations of th e DTH measurements were performed. The magnitude of the DTH response was determined by the following (Eq. 5-1): y = ln(R24) ln(R0) (5-1) A repeated measures preliminary analysis using the proc mixed procedure of SAS determined that multiparous cows tended to respond better than primiparous cows (p< 0.064) (Figure 5-1, Figure 5-2). Due to this effect of parity, determ ination of CMIR categorization occurred within parity. 112

PAGE 113

To extrapolate CMIR categorizations the mean and standard deviation for the y values were configured for all cows resp ective of parity. Cows with y values above the mean plus one standard deviation were classi fied as high CMIR responders. T hose cows below one standard deviation less than the mean were classified as low CMIR responders. All animals within one standard deviation of the mean were medium responders. Diseases Identification of disease was performed by fa rm personnel who were blinded to immune response categorizations. Disease information was co llected for; mastitis, metritis, retained fetal membrane, ketosis, and displaced abomasum. All diseases were recorded as yes/no binary responses for the trial peri od of the current lactation. Mastitis was recorded through 365 DIM as light, medium, or severe. Severity was recorded as light if there were no systemic signs and milk was slightly watery with minimal clots (gargot). Severity was medium if there were no systemic signs and a substantial amount of clots detected in the milk. Severity was severe if there were systemic signs, watery milk, and a substantial amount of clots in the milk. Metritis was recorded through 30 DIM as light, medium or severe. Severity was light if there was an abnormal vaginal discharge and a pa lpable uterine lumen. Severity was medium if there was a purulent vaginal discharge, with an enlarged, not-flaccid uterus. Severity was severe if there was a purulent foul-smelling vaginal di scharge, with an en larged flaccid uterus. Association between energy-re lated metabolic conditions an d CMIR categorization were also considered. Ketosis was recorded through 30 DIM as light, medium, or severe. Ketosis was coded as light if the urine contained 15 mg/dL of ketone bodies. Severity was medium if the urine contained 40 mg/dL of ketone bodies. The se verity was severe if the urine contained > 80 mg/dL of ketone bodies. Displaced abomas um was recorded through 16, and 50 DIM. 113

PAGE 114

Due to the involvement of the immune system in determining the expulsion of fetal membranes (Kimura et al., 2002), this study also analyzed the association between RFM and CMIR categorization. Retained feta l membranes were identified if still retained 24 h postpartum. Milk Yield, Somatic Cell Score, and Reproductive Efficiency Milk yield data was gathered from the Da iry Herd Information Association (DHIA) records for the current lactati on using ME305 which is an estim ate of the milk yield for 305 DIM. For categorical data, low milk producers were identified if there 305 day milk was in the bottom 25% of the study group. High milk producer s were those in the top 25% of the study group. The remaining middle 50% were clas sified as medium milk producers. Somatic cell score (SCS) was obtained from DHIA records. The average SCS was determined for the first 3 test days, the first 6 te st days, and for the first 10 test days, which are an indication of the average SCS for the first 90, 180, and 300 DIM respectively. To test for associations between CMIR categor ization and reproductive efficiency, a binary pregnancy term was used which simply indicated if a given cow was pregnant by 150 DIM. The number of number of days not pregnant (days open) and number of times bred were used as quantitative variables. Statistics To test for an association between CMIR categorization and the binary terms of specific disease risk and reproductive efficiency, a l ogistic regression model was used using the LOGISTIC procedure of SAS. Other than the main effect (CMIR categorization), possible explanatory variables include: BCS x = categorical effect of body condition score for sampling week x. BCS y = quantitative variable, change in body condition score over interval y. Dex = binary effect, whether cow recei ved dexamethasone prior to calving. Parity = binary effect, either primiparous or multiparous. 114

PAGE 115

RFM = binary effect, incidence of retained fetal membrane. CDiff = binary effect, if reported difficult calving. Ketosis = binary effect, incidence of ketosis. SCSavg = somatic cell score average over 10 monthly test days. For analyzing the relationship between CMIR categorization and the quantitative variables for SCS, 305 day milk yield, and reproductive efficiency, linear regression was used with the REG procedure of SAS. All categor ical variables were analyzed using logistic regression with the LOGISTIC procedure of SAS. All relevant effects were put in the model. A backword elim ination procedure was used to determine the final model. Explanatory variable s remained in the model if they showed a tendency (p < 0.10) to predict the outcome variab le. Statistical significance was determined by setting = 0.05. Results Mastitis Incidence of all types of mastitis within 100 DIM was not significantly associated with CMIR status. However, when only medium and severe cases of mastitis within 365 DIM were considered (138 cases in multiparous cows and 42 cases in primiparous cows), CMIR status was significantly associated with mastitis occurrence (Type 3 analysis p = 0.041). Parity and SCSavg were also significant variables in the model. Those cows categori zed as medium responders were 2.14 (CI = 1.13 4.08) times more likely to develop a medium or severe case of mastitis than high responders. When considering only multip arous cows, those categorized as low and medium responders were 2.80 (CI = 1.29 6.09) times more likely to develop a medium or severe case of mastitis than high responders. 115

PAGE 116

Retained Fetal Membrane There were 61 recorded cases of RFM (7.5 %). Of these 61, primiparous cows accounted for 22 while multiparous cows accounted fo r 39. Contributing model effects were:BCS 0 and BCS 5 The CMIR categorization was a significant predic tor for the risk of RFM (p = 0.0001) (Table 5-1). Low cell-mediated immune res ponders were 6.68 (CI = 1.87 23.84) times more likely to have a case of RFM than high i mmune responders. Also, if only considering multiparous cows, low cell-mediated immune responders were 26.52 (2.30 306.11) times more likely to have an RFM than high cell-mediated immune responders (Figure 5-3). Metritis For this analysis there were 43 cases of light, medium or se vere metritis within 30 DIM (5.3%). Primiparous cows accounted for 29 cases while multiparous cows accounted for 14. The remaining model contributing effects were: BCS 0 BCS 5 Dex, Parity, and RFM. Although not significantly different, the medium immune re sponders were 7.40 (CI = 0.91 60.25) times more likely to develop metritis than hi gh immune responders (Table 5-1). Ketosis and Displaced Abomasum There were 48 recorded incidences of light medium, or severe ketosis within 30 DIM (6.9%). Twenty-seven cases were observed in primiparous cows and 21 cases in multiparous cows. Remaining model effects were Dex and BCS 6 The categorization for CMIR was not a significant predictor for risk of ketosis (p = 0.97) (Table 5-1). There were only 21 recorded incidences of DA, 13 in primiparous and 8 in multiparous cows. Contributing model effects were: Ketosis, BCS 4 and Dex. The categorization for CMIR was not a significant predictor for incidence of DA (Table 5-1). 116

PAGE 117

Milk Yield, Somatic Cell Score and Reproductive Efficiency For the analysis of the effect of CMIR cate gorization on milk yield, the contributing model effects were parity, days open, and the binary variable mastitis. In this analysis, CMIR categorization was a significant pr edictor of milk yield (p = 0.049, = 508.08) (Figure 5-4). The analysis of the effect of CMIR on SCS included: BCS 6 milk categorization, and a binary mastitis variable. The e ffect of CMIR on SCS was not a si gnificant predictor of SCS (p = 0.83). The model analyzing the effect of AMIR category on pregnancy by 150 DIM, included BCS 4 and milk categorization. The effect of CMIR was not a signif icant predictor for pregnancy by 150 DIM (p = 0.77). Discussion The significant association be tween peripartum DTH response and mastitis identifies the role CMIR has on resisting infection in the ma mmary gland and the validity of using DTH to Candida albicans as a measure of CMIR. Although not sign ificant, the estimated odds of metritis show promise for the ability of DTH to predict metritis infection (Table 5-1). The results for RFM were in agreement with previous literature (Kimura et al., 2002). Although in the current study, because categorizat ion of CMIR occurred early postpartum, one could speculate that RFM served to suppress immune response. However, in Kimura et al (2002), neutrophil activity was suppressed 15 days prepartum in cows with RFM. The results for the associati on between milk yield and CMIR categorization prove that selection for improved CMIR does not predispose cows to lower milk production. In this study the significant effect of CMIR categorizati on on milk yield followed the pattern of High > Medium > Low. Because selection for improve d immune response should include both CMIR and AMIR, the potential negative association be tween AMIR and milk yield should balance out 117

PAGE 118

due to the positive effect of CMIR on milk yield. Previous studies in dairy cows have not studied the association between CMIR cat egorization and milk yield. Mallard et al. (1998) categorized CMIR in pigs and found greater growth rates in pigs with superior im mune responsiveness. Selection for decreased somatic cell score (SCS), increased productive life (PL), and improved structural traits of the udder, are all methods currently used to reduce incidence of disease. These methods, however, are largely based on fairly crude biologi cal associations with disease risk and due not addre ss the immune system which is pr incipally responsible for host disease resistance. These traits are also not c oncerned with broad-based resistance to disease. Selection for improved immune responsiveness as a means to reduce the risk of disease should take a broad based approach. This is primarily due to the vast array of potential pathogens and the various virulence mechanisms employed to in itiate disease. This broad based philosophy also becomes critically important due to th e complexity of the immune system in vivo Selection for specific improvements may prove inadequate, but may also confer unexpected disease susceptibility to additional varieties of pathoge ns. Inverse relationships between branches of adaptive immunity shed light on th e potential for this to occur (de Vries, 1995; Biozzi et al., 1979; Rupp et al., 2007). If there is one thing that should be le arned from the past 5 decades of applying selection pressure, it is that genetic selection should not focus or put too much pressure on specific traits. When this happens it is inev itable that unexpected unfavorable trends occur where selection pressure is neglected. 118

PAGE 119

5.0 5.5 6.0 6.5 7.0 7.5 0hrs 24hrs 0hrs 24hrs Primiparous Multiparous Double Skin-Fold (mm ) Figure 5-1. Graph showing in crease in double skin-fold thic kness respective of parity (multiparous or primiparous).Bars reflect SEM. 119

PAGE 120

0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 Primiparous Multiparous ParityDouble Skin-Fold Change (m m Figure 5-2. Graph revealing parity difference for cell-mediated immune response. Bars indicate SEM. 120

PAGE 121

Table 5-1. Odds ratios of disease inciden ce for cell-mediated immune categorizations. Mastitis Metritis Ketosis DA RFM 0.99 1.06 1.08 1.02 2.85 Low vs Med & High (0.64 1.53) (0.39 2.92) (0.56 2.06) (0.30 3.52) (1.67 4.86) 1.45 2.97 1.14 1.56 6.68 Low vs High (0.63 3.31) (0.29 30.51) (0.33 3.95) (0.12 20.21) (1.87 23.84) 2.14 7.4 1.05 2.28 1.94 Med vs High (1.13 4.08) (0.91 60.25) (0.40 2.78) (0.27 19.28) (0.58 6.45) Values indicate the estimated odds ratios for incidence of disease for the following comparisons among cell-mediated immune categorizations. Values in parenthesis represent the 95% confidence intervals for the odds ratio estimate. 121

PAGE 122

0 5 10 15 20Incidence RFM % LowMediumHigh CMIR Categorization Primiparous Multiparous Figure 5-3. Graph indicating the difference in risk of RFM betw een high and low cell-mediated immune response categorization. 122

PAGE 123

10200 10400 10600 10800 11000 11200 11400 11600 11800 12000 LowMediumHigh CMIR CategorizationMilk Yield (kg) Primiparous Multiparous Figure 5-4. Graph for the effect of cell-medi ated immune response (CMIR) categorization on milk yield. Bars indicate SEM. 123

PAGE 124

LIST OF REFERENCES Aarestrup, F. M., N. E. Jensen, and H. Osterg ard. 1995. Analysis of asso ciations between major histocompatibility complex (BoLA) class1 ha plotypes and subclinical mastitis of dairy cows. J. Dairy Sci. 78:1684-1692. Adewuyi, A. A., E. Gruys, and F. J. C. M. van Eerdenburg. 2005. Non esterified fatty acids (NEFA) in dairy cattle. A review Veterinary Quarterly 27(3):117-126. Badolato, R., H. M. Bond, G. Valerio, A. Petrella, G. Morrone, M. J. Waters, S. Venuta, and A. Tenore. 1994. Differential expression of su rface membrane growth hormone receptor on human peripheral blood lymphoc ytes detected by dual fluorochrome flow cytometry. J. Clin. Endocrinol. Metab. 79:984-990. Baird, G.D. 1982. Primary ketosis in the high-producing dairy cow: Clinical and subclinical disorders, treatment, prevention, an d outlook. J. Dairy Sci. 65(1):1-10. Baus, E., F. Andris, P. M. Dubois, J. Urbain, and O. Leo. 1996. Dexamethasone inhibits the early steps of antigen receptor signaling in ac tivated T lymphocytes. J. Immunol. 156:45554561. Beardsley, G. L., L. D. Muller, H. A. Garverick, F. C. Ludens, and W. L. Tucker. 1976. Initiation of parturition in dairy cows with dexame thasone. II. Response to dexamethasone in combination with estradiol benzoa te. J. Dairy Sci. 59(2):241-247. Besedovsky, H. O., and A. del Rey. 1996. Immune-n euro-endocrine interactions: Facts and hypothesis. Endocr. Rev. 17:64-102. Beutler, B., and E. T. Rietschel. 2003. Innate immune sensing and its roots: The story of endotoxin. Nat. Rev. Immunol. 3:169-176. Biozzi, G., D. Mouton, O. A. Sant Anna. H. C. Passos, M. Gennari M. H. Reis, V. C. Ferreira, A. M. Heumann, Y. Bouthillier, O. M. Ib anez, C. Stiffel, and M. Siqueira. 1979. Genetics of immunoresponsiveness to natura l antigens in the mouse. Curr. Top. Microbiol Immunol 85:31-98. Black, A. C. 1999. Delayed-type hypersensitivity: Cu rrent theories with a historic perspective. Dermatol. Online J. 5(1): 7-27. Blalock, J.E. 1994. The syntax of immune-neuroendocrine communication. Immunol. Today 15:504-511. Burton, J. L., and M. E. Kehrli, jr. 1995. Regul ation of neutrophil a dhesion molecules, and shedding of Staphylococcus aureus in milk of cortisoland dexamethasone-treated cows. Am. J. Vet. Res. 56:997-1006. 124

PAGE 125

Burton, J. L., and M. E. Kehrli, jr. 1996. Eff ects of dexamethasone on bovine circulating T lymphocyte populations. J. Leukocyte Biology 59:90-99. Burton, J. L., E. B. Burnside, B. W. Kenne dy, B. N. Wike, and J. H. Burton. 1989. Antibody responses to human erythrocytes and ovalbumin as marker tra its of disease resistance in dairy calves. J. Dairy Sci. 72:1252-1265. Burton, J. L., M. E. Kehrli, jr., S. Kapil, and R. L. Horst. 1995. Regulation of L-selectin and CD18 on bovine neutrophils by glucocorticoids: Effects of cortisol and dexamethasone. J. leukoc. Biol. 57:317-325. Castillo-Juarez, H., P. A. Oltenacu, R. W. Blak e, C. E. Mcculloch, and E. G. Cienfuegos-Rivas. 2000. Effect of herd environment on the ge netic and phenotypic relationships among milk yield, conception rate, and somatic cell sc ore in holstein cattle. J. Dairy Sci. 83:807814. Chertov, O., D. Yang, O. M. Howard, and J. J. Oppenheim. 2000. Leukocyte granule proteins mobilize innate host defenses and adaptiv e immune responses. Immunol. Rev. 177: 6878. Cole, R. K. 1968. Studies on gene tic resistance to marek's dis ease. Avian Diseases. 12:9-28. Corbeil, L. B., and R. H. BonDurant. 2001. Imm unity to bovine reproductive infections. Vet. Clin. of North Am. Food Anim. Pract. 17(3):567-83. Curtis, C. R., H. N. Erb, C. J. Sniffen, R. D. Smith, and D. S. Kronfeld. 1985. Path analysisof dry period nutrition, postpartum meta bolic and reproductive disorders, and mastitis in holstein cows. J. Dairy Sci. 68:2347-2360. Dardenne, M., and W. Savino 1996. Interdependen ce of the endocrine and immune systems. Adv. Neuroimmunol. 6:297-307. Davis, M. M., J. J. Boniface, Z. Reich, D. Lyons, J. Hampl, B. Arden, and Y. Chien. 1998. Ligand recognition by alphabeta T cell receptors. Annu. Rev. Immunol. 16:523-44. de Haas, Y., H. W. Barkema and R. F. V eerkamp. 2002. Genetic parameters for pathogenspecific incidence of clinical mastitis in dairy cows. Animal Sci. 74:233-242. de Vries, J. E. 1995. Immunosuppressive and an ti-inflammatory properties of interleukin 10. Ann. Med. 27(5):537-41. Detilleux, J. C., M. E. Kehrli, Jr., Stabel, A. E. Freeman, and D. H. Kelley. 1995. Study of immunological dysfunction in peri parturient holstein cattle selected for high and average milk production. Vet. Imm un. and Immunopathology 44:251-267. 125

PAGE 126

Edfors-Lilia, I., and P. Wallgreen. 2000. Escheric hia coli and Salmonella diarrhea in pigs. Breeding for disease resistance in farm animal s. 2nd Edn. Axford, R. F. E., S. C. Bishop, F. W. Nicholas and J. B. Owen. Walli ngford: CAB International. Pp 253-267. Emanuelson, U., B. Danell, and J. Phillipson. 19 88. Genetic parameters for clinical mastitis, somatic cell counts, and milk production estim ated by multiple-trait restricted maximum likelihood. J. Dairy Sci. 71:467-476. Ferguson, J. D., D. T. Galligan, N. Thomsen. 1994. Principle descriptors of body condition score in Holstein cows. J. Dairy Sci. 77(9):2695-2703 Franklin, S. T., J. W. Young, and B. J. Nonnecke. 1991. Effects of ketones, acetate, butyrate, and glucose on bovine lymphocyte prolif eration. J. Dairy Sci. 74:2507-2514. Gatti, E., and P. Pierre. 2003. Understanding the cell biology of antigen presentation: The dendritic cell contribution. Curr. Opin. in Cell Biol. 15:468-473. Goff, J. P., and K. Kimura. 2002. Effect of mastect omy on milk fever, energy, and vitamins A, E, and -carotene status at parturit ion. J. Dairy Sci. 85:1427-1436. Goff, J. P., and R. L. Horst. 1997. Physiological changes at parturiti on and relationship to metabolic disoders. J. Dairy Sci. 80:1260-1268. Grafton, G., and L. Thwaite. 2001. Calcium ch annels in lymphocytes. Immunology 104:119-126. Hansen, P. J. 1997. Interactions between the immune system and the bovine conceptus. Theriogenology 47:121-130 Harmon, R. J. 1994. Physiology of mastitis and f actors affecting milk somatic cell counts. J. Dairy Sci. 77:2103-2112. Heller, E. D., G. Leitner, A. Friedman, Z. Uni, M. Gutman, and A. Cahaner 1992. Immunological parameters in meat-type chic ken lines divergently selected by antibody response to Escherichia coli vaccinati on. Vet Immunol Immunopathol. 34:159-172. Heringstad, B., D. Gianola, Y. M. Chang, J. Odegard, and G. Klemetsdal. 2006. Genetic associations between clinical mastitis and so matic cell score in early first-lactation cows. J. Dairy Sci. 89: 2236-2244. Heringstad, B., G. Klemetsdal, and J. Ruane. 2000. Selection for mastitis resistance in dairy cattle: A review with focus on the situation in the nordic countries. Livestock Prod. Sci. 64: 95-106. Heringstad, B., Y. M. Chang, D. Gianola, and G. Klemetsdal. 2004. Multivariate threshold model analysis of clinical mastitis in multip arous norweigan dairy cattle. J. Dairy Sci. 87:3038-3046. 126

PAGE 127

Hernandez, A., J. A. Yager, B. N. Wilkie, K. E. Leslie, and B. A. Mallard. 2005. Evaluation of bovine cutaneous delayed-type hypersensitiv ity (DTH) to various test antigens and a mitogen using several adjuvants. Ve t Immun. And Immuno pathology 104:45-58. Hernandez, A., N. Karrow, and B. A. Mallard. 2003. Evaluation of immune responses of cattle as a means to identify high or low res ponders and use of a human microarray to differentiate gene expression. Genet. Sel. Evol. 35:S67-81. Hoeben, D., R. Heyneman, and C. Bu rvenich. 1997. Elevated levels of -hydroxybutyric acid in periparturient cows and in vi tro effect on respiratory burst activity of bovine neutrophils. Vet Immun. And Immunopathology 58:165-170. Hunter, N., L. Moore, B. D. Hosie, W. S. Di ngwall, and A. Greig. 1997. Association between natural scrapie and PrP genotype in a flock of suffolk sheep in Scotland. Veterinary Record 140:59-63. J. C. Detilleux. 2002. Genetic factors affecting sus ceptibility of dairy cows to udder pathogens. Vet. Immun. and Im munopathology 88:103-110. Jacysyn, J. F., I. H. Abrahamsohn, and M. S. Macedo. 2001. Modulation of delayed-type hypersensitivity during the time course of immune response to a protein antigen. Immunology 102: 373-379. Jensen, P. E. 2007. Recent advances in antige n processing and presentation. Nature Immunol. 8(10):1041-1048. Kadarmideen, H. N., R. Thompson, and G. Simm. 2000. Linear and threshold genetic parameters for disease fertility and milk production in dairy cattle. J. Anim. Sci. 71:411-419. Kaneene, J. B., R. A. Miller, T. H. Herdt, and J. C. Gardiner. 1997. The association of serum nonesterified fatty acids and cholesterol, management and feeding practices with peripartum disease in dairy cows. Prev. Vet. Med. 31:59-72. Kean, R. P., A. Cahaner, A. E. Freeman, S. J. Lamont. 1994. Direct and correlated responses to multitrait, divergent selection for im munocompetence. Poult. Sci. 73:18-32. Kehrli, M. E., and D. E. Shuster. 1994. Factors affecting milk somatic cells and their role in health of the bovine mammary gland. J. Dairy Sci. 77:619-627. Kehrli, M. E., and J. A. Harp. 2001. Immunity in the mammary gland. Vet. Clin. Of North Am Food Anim Pract 17(3):495-516. Kehrli, M.E. jr., F.C. Schmalstieg, D.C. Anderson, M. J. Van der Maaten, B. J. Hughes, M. R. Ackermann, C. L. Wilhelmse n, G. B. Brown, M. G. Steven s, and C. A. Whetstone 1990. Molecular definition of the bovine granul ocytopathy syndrome: Identification of deficiency of the Mac-1 (CD11b/CD18) glycoprotein. Am. J. Vet. Res. 51:1826-1836. 127

PAGE 128

Kelm, S. C., A. E. Freeman, and M. E. Kehrli, jr. 2001. Genetic control of disease resistance and immunoresponsiveness. Vet. Clin. Of North Am. Food Anim. Pract. 17(3):477-93. Kelm, S. C., J. C. Detilleux, and A. E. Freema n. 1997. Genetic association between parameters of innate immunity and measures of mastitis in periparturient holstein cattle. J. Dairy Sci. 80:1767-1775. Kimura, K., J. P. Goff, M. E. Kehrli, jr., and T. A. Reinhardt. 2002. Decreased neutrophil function as a cause of retained placenta in dairy cattle. J. Dairy Sci. 85:544-550. Kimura, K., T. A. Reinhardt, and J. P. Goff 2006. Parturition and hypocalcemia blunts calcium signals in immune cells of dair y cattle. J. Dairy Sci. 89:2588-2595. Klasing, K. C. 1998. Nutritional modulation of resi stance to infectious di sease. Poult. Sci. 77:1119-1125. Klasing, K. C., L. Lauring, R. Peng, and M. Fry. 1987. Immunologically mediated growth depression in chicks: Influence of feed intake corticosterone, and interleukin-1. J. Nutr. 117:1629-1637. Kremer, W. D. J., E. N. Noordhui zen-Stassen, F. J. Grommers, A. J. J. M. Daemen, P. A. J. Hendricks, and A. Brand. 1993. Preinfec tion chemotactic response of blood polymorphonuclear leukocytes to predict sever ity of Escherichia coli mastitis. J. Dairy Sci. 76:1568-1574. Kushibiki, S., K. Hodate, H. Shingu, Y. Obara, E. Touno, M. Shinoda, and Y. Yokomizo. 2003. Metabolic and lactational re sponses during recombinant bov ine tumor necrosis factortreatment in lactating co ws. J. Dairy Sci. 86:819-827. Lacetera, N., D. Scalia, O. Franci, U. Bern abucci, B. Ronchi, and A. Nardone. 2004. Short comm: Effects of nonesterified fatty acids on lymphocyte function in dairy heifers. J. Dairy Sci. 87:1012-1014. Lacetera, N., D. Scalia, U. Bernabucci, B. Ronchi, D. Pirazzi, and A. Nardone. 2005. Lymphocyte functions in overconditioned cows around parturition. J. Dairy Sci. 88:20102016. Lacetera, N., U. Bernabucci, B. Ronchi, and A. Nardone. 2001. Effects of subclinical pregnancy toxemia on immune responses in sheep. Am. J. Vet. Res. 62:1020-1024. Lewis, R. S. 2001. Calcium signaling mechan isms in T lymphocytes. Annu. Rev. Immunol. 19:497-521. Mailhot, G., J. L. Petit, C. Demers, and M. Ga scon-Barre. 2000. Influence of the in vivo calcium status on cellular calcium homeostasis and the level of the calcium-binding protein calreticulin in rat hepatocytes. Endocrinology 141:891-900. 128

PAGE 129

Mallard, B. A., B. N. Wilkie, B. W. Ke nnedy, J. Gibson, and M. Quinton. 1998. Immune responsiveness in swine: Eight generati ons of selection for high and low immune response in yorkshire pigs. Proceedings of the 6th World Congress on Genetics Applied to Livestock Production, 11-16 January Vol .27, University of New England, Armidale, pp. 257-262. Mallard, B. A., J. C. Dekkers, M. J. Ireland, K. E. Leslie, S. Sharif, C. Lacey Vankampen, L. Wagter, and B. N. Wilke. 1998. Alterati on in immune responsiveness during the peripartum period and its ramification on dair y cow and calf health. J. Dairy Sci. 81:585595. Mallard, B. A., L. C. Wagter, M. J. Ireland, and J. C. M. Dekkers. 1997. Effect of growth hormone, insulin-like growth factor-1, and cortisol on pe riparturient antibody response profiles of dairy cattle. Vet. Immunol. Immunopathol. 60:61-68. Meglia, G. E., A. Johannison, S. Agenas, K. Holtenius, and K. P. Waller. 2005. Effects of feeding intensity during the dry period on leukocyte and lymphocyte sub-populations, neutrophil function and health in peripartur ient dairy cows. The Vet. Journal 169:376384. Nagahata, H., A. Ogawa, Y. Sanada, H. Noda and S. Yamamoto. 1992. Peripartum changes in antibody producing capability of lymphocytes from dairy cows. Vet. Quart. 14:(1): 3940. Nash, D. L., G. W. Rogers, J. B. Cooper, G. L. Hargrove, J. F. Keown, and L. B. Hansen. 2000. Heritability of clinical mastitis incidence a nd relationships with sire transmitting abilities for somatic cell score, udder t ype traits, productive life, and protein yield. J. Dairy Sci. 83:2350-2360. National Mastitis Council. 1996. Cu rrent concepts of bovine mastitis. 4th ed. Natl. Mastitis Counc. Inc., Madison, WI. Neerhof, H. J., P. Madsen, V.P. Ducrocq, A. R. Vollema, J. Jensen, and I. R. Korsgaard. 2000. Relationships between mastitis and functional longevity in danish black and white dairy cattle estimated using survival an alysis. J. Dairy Sci. 83:1064-1071. Nonnecke, B. J., K. Kimura, J. P. Goff, J. Ma rcus, and M. E. Kehrli, jr. 2003. Effects of the mammary gland on functional capacities of blood mononuclear leukocyte populations from periparturient cows. J. Dairy Sci. 86:2359-2368. Nonnecke, B. J., S. T. Franklin, and J. W. Young. 1992. Effects of ketones, acetate, and glucose on In Vitro immunoglobulin se cretion by bovine lymphocytes J. Dairy Sci. 75:982-990. 129

PAGE 130

O'Neill, R. G., J. A. Woolliams, E. J. Gla ss, J. L. Williams, and J. L. Fitzpatrick. 2006. Quantitative evaluation of genetic and environmental parameters determining antibody response induced by vaccination against bovine respiratory syncytial virus. Vaccine 24: 4007-4016. Park, Y. H., L. K. Fox, M. J. Hamilton, and W. C. Davis. 1992. Bovine mononuclear leukocyte subpopulations in peripheral blood and mammary gland secret ions during lactation. J. Dairy Sci. 75:998-1006. Park, Y. H., L. K. Fox, M. J. Hamilton, and W. C. Davis. 1993. Suppression of proliferative response of BoCD4+ T lymphocytes by ac tivated BoCD8+ T lymphocytes in the mammary gland of cows with Staphylococcus aureus mastitis. Vet Immunol. Immunopathol. 36:137-151. Park, Y. H., Y. S. Joo, J. Y. Park, J. S. Moon, S. H. Kim, N. H. Kwon, J. S. Ahn, W. C. Davis, and C. J. Davies. 2004. Characterization of lymphocyte subpopulations and major histocompatibility complex haplotypes of mas titis-resistant and susceptible cows. J. Vet. Sci. 5:29-39. Parker, D. C. 1993. T cell-dependent B ce ll activation. Annu. Rev. Immunol. 11:331-60. Partiseti, M., F. L. Deist, C. Hivroz, A. Fisher, H. Korn, and D. Choquet. 1994. The calcium current activated by T cell receptor and store de pletion in human lymphocytes is absent in a primary immunodeficiency. J. Biol. Chem. 269:32327-32335. Peters, A.R., and D.A. Poole. 1992. Induction of pa rturition in dairy cows with dexamethasone. The Veterinary Record 131(25-26):576-578. Piccinini, R., C. Bronzo, P. Moroni, C. Luzza go, and A. Zecconi. 1999. Study on the relationship between milk immune factors and Staphyloco ccus aureus intramammary infections in dairy cows. J. Dairy Res. 66:501-510. Risso, A. 2000. Leukocyte antimicrobial peptides: multifunctional effector molecules of innate immunity. J. Leukocyte Biology 68:785-792. Rogers, G. W., G. Banos, U. S. Nielson, a nd J. Philipsson. 1998. Genetic correlations among somatic cell scores, productive life, and type traits from the United States and udder health measures from Denmark and Sweden. J. Dairy Sci. 81:1445-1453. Rupp, R., A. Hernandez, and B. A. Mallard. 2007. Association of bovi ne leukocyte antigen (BoLA) DRB3.2 with immune response mastit is, and production and type traits in Canadian holsteins. J. Dairy Sci. 90:1029-1038. Rupp, R., and D. Boichard. 2003. Genetics of resistan ce to mastitis in dairy cattle. Vet. Res. 34:671-688. 130

PAGE 131

Saad, A. M., C. Concha, and G. Astrom. 1989. Alterations in neutr ophil phagocytosis and lymphocyte blastogenesis in dairy cows around parturition. J. Vet. Med. 36(5): 337-45. Savina, A., and S. Amigorena. 2007. Phagocytosis and antigen presentation in dendritic cells. Immunol. Rev. 219: 143-56. Schukken, Y. H., H. N. Erb, and D. Smith. 1988. Th e relationship between mastitis and retained placenta in a commercial population of holst ein dairy cows. Preventive Veterinary Medicine 5:181-190. Schukken, Y. H., K. E. Leslie, D. A. Barnum, B. A. Mallard, J. H. Lumsden, P. C. Dick, G. H. Vessie, and M. E. Kehrli, Jr. 1999. Experime ntal Staphylococcus aureus intramammary challenge in late lactation dairy cows. Quarter and cow effects determining the probability of infection. J. Dairy Sci. 82:2393-2401. Shook, G. E., and M. M. Schutz. 1994. Selection on somatic cell score to improve resistance to mastitis in the United States. J. Dairy Sci. 77:648-658. Silvia, W. 2003. Addressing the decline in reprod uctive performance of lactating dairy cows: A researcher's perspective. Vet. Sci TommorrowVol. 3 May 2003. Soler, J. J., L. de Neve, T. Perez-Contreras, M. Soler, and G. Sorci. 2003. Trade-off between immunocompetence and growth in magpies: An experimental study. Proc. R. Soc. Lond. B 270:241-248. Sordillo, L. M., M. Campos, and L. A. Babi uk. 1991. Antibacterial activity of bovine mammary gland lymphocytes following treatment with interleukin-2. J. Dairy Sci. 74:3370-3375. Stear, M. J., S. C. Bishop, B. A. Mallard, and H. Raadsma. 2001. Rev: The sustainability, feasibility and desireability of breeding livesto ck for disease resistance. Research in Vet. Sci. 71:1-7. Steiger, M., M. Senn, G. Altreu ther, D. Werling, F. Sutter, M. Kreuzer, and W. Langhans. 1999. Effect of prolonged low-dose lipopolysaccharid e infusion on feed intake and metabolism in heifers. J. Anim. Sci. 77:2523-2532. Steine, T. 1996. Avlsarbeid og mastitt. Buskap 2: 8-11. (In Norweigian). Strandberg, E., and G. E. Shook. 1989. Genetic and economic responses to breeding programs that consider mastitis. J. Dairy Sci. 72:2136-2142. Suriyasasathaporn, W., C. Heuer, E. N. Noordhuizen-Stassen, and Y. H. Schukken. 2000. Hyperketonemia and the impairment of udde r defense: A review. Vet. Res. 31:397-412. Swanson, L. V., and A. G. Hunter. 1969. Egg yolk anti gens and their effect on fertility in rabbits. Biology of Reproduction 1: 324-329. 131

PAGE 132

Uribe, H. A., B. W. Kennedy, S. W. Martin, and D. F. Kelton. 1995. Genetic parameters for common health disorders of holstein cows. J. Dairy Sci. 78:421-430. van der Poll, T. 2001. Coagulation and infl ammation. J. Endotoxin Research 7:301-304. Van Kampen, C., and B. A Mallard. 1997. Effects of peripartum stress and health on circulating bovine lymphocyte subsets. Vet Imm unol Immunopathol. 59(1-2): 79-91. Van Werven, T., E. N. Noordhuizen-Stassen, A. J. J. M. Daemen, Y. H. Schukken, A. Brand, and C. Burvenich. 1997. Preinfection In Vitr o cemotaxis, phagocytosis, oxidative burst, and expression of CD11/CD18 receptors and th eir predictive capacity on the outcome of mastitis induced in dairy cows with Es cherichia coli. J. Dairy Sci. 80:67-74. von Ruecker, A., and I. G. H. Schmidt-Wolf. 2000. Strategies to evaluate metabolic stress and catabolism by means of immunological vari ables. Clinical Nutrition 19(3):147-156. Wagter, L. C., B. A. Mallard, B. N. Wilke, K. E. Leslie, P. J. Boettchart, and J. C. M. Dekkers. 2000. A quantitative approach to classi fying holstein cows based on antibody responsiveness and its relationship to peripartum mastitis occurrence. J. Dairy Sci. 83:488-498. Wagter, L. C., B. A. Mallard, B. N. Wilke, K. E. Leslie, P. J. Boettcher, and J. C. M. Dekkers. 2003. Relationship between milk production a nd antibody response to ovalbumin during the peripartum period. J. Dairy Sci. 86:169-173. Waldron, M. R., A. E. Kulick, A. W. Bell, and T. R. Overton. 2006. Acute experimental mastitis is not causal toward the development of energy-related metabolic disorders in early postpartum dairy cows. J. Dairy Sci. 89:596-610. Wedlock, D. N., F. E. Aldwell, D. M. Collins, G. W. de Lisle, T. Wilson, and B.M. Buddle. 1999. Immune responses induced in cattle by virulent and attenuated Mycobacterium bovis strains: correlation of de layed-type hypersensitivity with ability of strains to grow in macrophages. Infect. Immun. 67(5): 2172-2177. Weigel, K. A., A. E. Freeman, M. E. Kehrli, jr., J. R. Thurston, and D. H. Kelley. 1992. Relationship of In Vitro immune function with health and production in holstein cattle. J. Dairy Sci. 75:1672-1679. Weller, J. I., and E. Ezra. 1997. Genetic analysis of somatic cell score and female fertility of Israeli holsteins with an individual animal model. J. Dairy Sci. 80:586-593. Wentink, G. H., V. P. M. Rutten, T. S. van den Ingh, A. Hoek, K. E. M ller, and T. Wensing. 1997. Impaired specific immunoreactivity in cows with hepatic lipidos is. Vet. Immunol. Immunopathol. 56:77-83. Wild, D., ed. 2001. The immunoassay handbook. Nature Publishing Group, NY. 132

PAGE 133

Wilkie, B.N., B. A. Mallard, M. Quinton, and J. Gibson. 1998. Multi-trait selection for immune response: a possible alternativ e strategy for enhanced livesto ck health and productivity. Prog. Pig Sci. 29-38. Wilson, D. J., B. A. Mallard, J. L. Burton, Y. H. Schukken, and Y. T. Grohn. 2007. Milk and serum J5-specific antibody responses, milk production change, and clinical effects following intramammary E. coli challenge for J5 vaccinate and control cows. Clinical and Vaccine Immun. 14(6):693-699. Woolaston, R. R., and R. L. Baker. 1996. Prospects of breeding small ruminants for resistance to internal parasites. Internat J. of Parasitology 26:845-855. Yang, D., K. Nakada-Tsuki, M. Ohtani, R. Go to, T. Yoshimura, Y. Kobayashi, and N. Watanabe. 2001. Identification and cloning of genes associated with the guinea pig skin delayed-type hypersensitivity reaction. J. Biochem. 129:561-568. 133

PAGE 134

BIOGRAPHICAL SKETCH Jason De La Paz was born in Tampa, Florida in 1977. He realized his in terest in animals at an early age, and although never growing up on a farm he discovered a fascination for dairy cows while working toward an animal science degree at the University of Florida. During the four years of his undergraduate college education, he held a part-time position working as a veterinary technician. Jason rece ived his bachelors degree in 2001 and soon thereafter moved to Minnesota after accepting a position as a reprodu ctive specialist for ABS Global, which is a bovine genetics company. In this position, he ma naged the reproductive concerns for several large farms in north central Minnesota. With family, warmer climate and saltwater fishing awaiting him back in Florida, he took a posit ion with ABS Global which allowed him to move back. After working this position for a few year s, Jason began pursuing a Master of Science degree at the University of Florida college of Veterinary Medicine. With Dr. Arthur Donovan as chair of Jasons supervisory committee, Jason rece ived his Master of Science degree in August 2008 where he studied how to determine the immune response potential of individual Holstein dairy cows and how this is a ssociated with disease risk. For the years to come, Jason intends on con tinuing his focus on disease resistance through genetic selection for increased immune responsiveness. He has been married since 2002 to Amy, the woman he dated since high school. They now resi de in Ocala, Florida with their two-year-old daughter Emily. His interests include gator football, fishing, and tennis. 134