<%BANNER%>

Comparison of a Silicone Membrane Blood-Feeding System to Human Skin in Repellency Bioassays with Aedes aegypti Mosquitoes

Permanent Link: http://ufdc.ufl.edu/UFE0045130/00001

Material Information

Title: Comparison of a Silicone Membrane Blood-Feeding System to Human Skin in Repellency Bioassays with Aedes aegypti Mosquitoes
Physical Description: 1 online resource (77 p.)
Language: english
Creator: Agramonte, Natasha M
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

Subjects / Keywords: aedes -- aegypti -- behavior -- bioassay -- cloth -- deet -- dose -- ed50 -- ed95 -- ed99 -- effective -- feeding -- in -- ir3535 -- kbr3023 -- med -- membrane -- module -- pmd -- repellent -- response -- skin -- vitro -- vivo
Entomology and Nematology -- Dissertations, Academic -- UF
Genre: Entomology and Nematology thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: It is expected that laboratory-based repellent bioassays should reliably evaluate the efficacy of compounds that deter mosquito feeding behavior. The variety of repellent bioassays available allows for flexibility in design, but makes it difficult to compare any two methods, including in vitro and in vivo comparisons. The most reliable data come from skin assays; however, this exposes volunteers to chemicals and mosquito bites. In this study, four repellent active ingredients were used: DEET, IR3535, picaridin and para-menthane-3,8-diol. Results from bioassays with a module-based method were operated in vitro with a membrane and in vivo on the skin of the leg and were then compared to a standard in vitro method where repellent-treated cloth is placed over an arm that is inserted into a cage of mosquitoes. Pooled data from six volunteers were used to create dose response curves and estimate effective doses for four repellents at the 50, 95, and 99% levels (ED50,ED95, ED99) using dose response curves with a probit model from the module tests. The ED99 was estimated with repellent-treated cloth as the concentration that prevented 99% of the mosquitoes from feeding.  The baseline attraction did not differ statistically for the membrane and skin module tests. The membrane and skin module results also did not differ significantly as repellent treatments below 31 nmol/cm2 dose level. However, since the chemical dose at which the skin and membrane curves diverge is lower than standard thresholds used in screening these chemicals, the use of the membrane module system will require further alterations before it can substitute current screening methods using human volunteers. Interestingly, all four repellents tested on the membrane required much higher doses of repellent chemical to reach full protection. Based on the results of this study, cage-based tests appear to be a more reliable estimate of repellent activity on skin compared to module-based tests on membrane. However, with the knowledge of the effective dose ratios, module-based tests can be utilized for testing with infected vectors with previously evaluated repellents.
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 Natasha M Agramonte.
Thesis: Thesis (M.S.)--University of Florida, 2012.
Local: Adviser: Bernier, Ulrich R.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0045130:00001

Permanent Link: http://ufdc.ufl.edu/UFE0045130/00001

Material Information

Title: Comparison of a Silicone Membrane Blood-Feeding System to Human Skin in Repellency Bioassays with Aedes aegypti Mosquitoes
Physical Description: 1 online resource (77 p.)
Language: english
Creator: Agramonte, Natasha M
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

Subjects / Keywords: aedes -- aegypti -- behavior -- bioassay -- cloth -- deet -- dose -- ed50 -- ed95 -- ed99 -- effective -- feeding -- in -- ir3535 -- kbr3023 -- med -- membrane -- module -- pmd -- repellent -- response -- skin -- vitro -- vivo
Entomology and Nematology -- Dissertations, Academic -- UF
Genre: Entomology and Nematology thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: It is expected that laboratory-based repellent bioassays should reliably evaluate the efficacy of compounds that deter mosquito feeding behavior. The variety of repellent bioassays available allows for flexibility in design, but makes it difficult to compare any two methods, including in vitro and in vivo comparisons. The most reliable data come from skin assays; however, this exposes volunteers to chemicals and mosquito bites. In this study, four repellent active ingredients were used: DEET, IR3535, picaridin and para-menthane-3,8-diol. Results from bioassays with a module-based method were operated in vitro with a membrane and in vivo on the skin of the leg and were then compared to a standard in vitro method where repellent-treated cloth is placed over an arm that is inserted into a cage of mosquitoes. Pooled data from six volunteers were used to create dose response curves and estimate effective doses for four repellents at the 50, 95, and 99% levels (ED50,ED95, ED99) using dose response curves with a probit model from the module tests. The ED99 was estimated with repellent-treated cloth as the concentration that prevented 99% of the mosquitoes from feeding.  The baseline attraction did not differ statistically for the membrane and skin module tests. The membrane and skin module results also did not differ significantly as repellent treatments below 31 nmol/cm2 dose level. However, since the chemical dose at which the skin and membrane curves diverge is lower than standard thresholds used in screening these chemicals, the use of the membrane module system will require further alterations before it can substitute current screening methods using human volunteers. Interestingly, all four repellents tested on the membrane required much higher doses of repellent chemical to reach full protection. Based on the results of this study, cage-based tests appear to be a more reliable estimate of repellent activity on skin compared to module-based tests on membrane. However, with the knowledge of the effective dose ratios, module-based tests can be utilized for testing with infected vectors with previously evaluated repellents.
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 Natasha M Agramonte.
Thesis: Thesis (M.S.)--University of Florida, 2012.
Local: Adviser: Bernier, Ulrich R.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0045130:00001


This item has the following downloads:


Full Text

PAGE 1

1 COMPARISON OF A SILICONE MEMBRANE BLOOD FEEDING SYSTEM TO HUMAN SKIN IN REPELLENCY BIOASSAYS WITH AEDES AEGYPTI MOSQUITOES By NATASHA MARIE AGRAMONTE A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2012

PAGE 2

2 2012 Natasha Marie Agramonte

PAGE 3

3 To my husband and my family for their unending love and support

PAGE 4

4 ACKNOWLEDGMENTS First and foremost, I want to express my great appreciation for Dr. Uli Bernier, my major advisor, for his patience, knowledge and assistance throughout my m research project. He has aided and pushed me, encouraging high achievements as well as believing in me often mo re than I did myself. I also wish to thank him for allowing me the opportunity to pursue my degree the financial support to continue my education and the flexibility to do all this while also working as his full time technician. I also wish to thank Dr. J ames Maruniak, the other half of my graduate committee. Dr. Maruniak provided encouragement and guidance with my presentation skills as well as with my writing, suggesting I write early and often. I only regret that I could not put more of the molecular b iology techniques that I learned from him to good use for this project. An unanticipated benefit t o having Dr. Maruniak on my committee was his wife, Dr. Alejandra Garcia Maruniak who was always ready to listen and helped put things in perspective when I w as discouraged by the slow progress of a large research project. I would also like to thank Drs. Gary Clark, Jerry Hogsette, Dan Kline, Phil Kaufman, Sandra Allan, Jeff Bloomquist, and Paul Weldon and for their kind words, their invaluable knowledge regard ing experimental design, data analysis, and writing, and their availab i l ity to answer many of my questions regarding the minutiae of the thesis process. I also have great appreciation for the help of Dr. Salvador Gezan for his comprehension, patience and e xpertise in statistical analysis. He was tremendously helpful in adding to the statistical and scientific soundness of my research I am indebted to the Florida Mosquito Control Association for awarding me the T. Wainwright Miller, Jr. Scholarship to suppo rt my research and academic efforts.

PAGE 5

5 Thanks are also extended to my friends and laboratory mates Greg Allen, Nathan Newlon Maia Tsikolia, Katelyn Chalaire and Neil Sanscrainte for their time and assistance with several aspects of my project. I would also like to thank Maria Theresa Cuevas, the undergraduate student who took a special interest in my project and worked meticulously for hours counting mosquitoes, noting feeding behavior and administering bi oassays on my human volunteers. Tremendous thanks are extended to all of my family members, my sisters and friends, and especially my parents, Regina and Tony, my grandmothers, Maria and Norma, and my mother in law, Rosa, for their love and support. I also wish to remember my late grandfather, Dirk Smit who did not live to see me achieve all that he firmly believed that I would. It is my greatest hope that my work here makes them proud. Last but not least, I owe the world of gratitude to my husband, Kent for all his love, support, encourag ement and the hours of his time that he devoted to listening to the d etails of my project even w he n he falling asleep on the floor beside me as I stayed up all night writing and the hours of painful research that I subjected h im to as a participant in my project Few people would give their time, much less their bodies for the advancement of science. He is my rock and I am lucky to have him

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ............... 4 LIST OF TABLES ................................ ................................ ................................ ........................... 8 LIST OF FIGURES ................................ ................................ ................................ ......................... 9 ABSTRACT ................................ ................................ ................................ ................................ ... 11 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .................. 13 Mosquitoes and their Medical Relevance ................................ ................................ ............... 13 A Brief History of Repe llent Research ................................ ................................ ................... 14 2 A REVIEW OF MOSQUITO HOST LOCATION ................................ ................................ 16 Chemical Cues and Olfaction ................................ ................................ ................................ 16 Semiochemicals and their Role in Olfaction ................................ ................................ ... 16 Carbon Dioxide as an Attractant ................................ ................................ ..................... 16 Skin Emanations and th eir Metabolites as Attractants ................................ .................... 17 L Lactic Acid as an Attractant ................................ ................................ ........................ 19 Repellent Chemicals and Associated Mosquito Behavior ................................ ............... 20 Mosquito Biology and Behavior Used in Host Location ................................ ........................ 21 3 A REVIEW OF MOSQUITO REPELLENT TESTING METHODS ................................ ... 24 Current Standards for Repellent Testing Methods ................................ ................................ 24 Measurements of Repellent Efficacy ................................ ................................ ...................... 25 A Survey of in vitro Test Methods ................................ ................................ ......................... 27 A Survey of in vivo Test Methods ................................ ................................ .......................... 30 Previous Reviews of Mosquito Repellency Methods ................................ ............................. 32 4 COMPARISON OF DOSE RESPONSE CURVES FOR MOSQUITO REPELLENTS TESTED IN VIVO AND IN VITRO USING A MODULE SYSTEM ................................ ... 33 Background ................................ ................................ ................................ ............................. 33 Materials and Methods ................................ ................................ ................................ ........... 34 Mosquito Rearing and Selection ................................ ................................ ..................... 34 Chemical Treatments and Control ................................ ................................ ................... 35 Alterations Made to Module from Previous Designs ................................ ...................... 35 in vivo Module Bioassays on Skin ................................ ................................ ................... 36 in vitro Module Bioassays on Silicone Membranes Treated with Skin Odors ................ 38 Statistical Analysis ................................ ................................ ................................ .......... 39

PAGE 7

7 Results ................................ ................................ ................................ ................................ ..... 39 Discussion ................................ ................................ ................................ ............................... 40 5 ASSESSMENT OF THREE MEDIA FOR THE ESTIMATION OF EFFECTIVE DOSE IN MOSQUITO REPEL LENT BIOASSAYS ................................ ............................ 51 Background ................................ ................................ ................................ ............................. 51 Materials and Methods ................................ ................................ ................................ ........... 52 Mosquit o Rearing and Selection ................................ ................................ ..................... 52 Repellent Chemical Treatments and Control ................................ ................................ .. 53 Skin Bioassays for the Estimation of Effective Dose ................................ ...................... 53 Membrane Bioassays for Estimation of Effective Dose ................................ .................. 54 Cloth Bioassays for Estimation of Minimum Effective Dose ................................ ......... 55 Statistical Analysis ................................ ................................ ................................ .......... 56 Results ................................ ................................ ................................ ................................ ..... 57 Discussion ................................ ................................ ................................ ............................... 58 6 SUMMARY OF FINDINGS AND IMPLICATIONS FOR FUTURE RESEARCH ............ 68 LIST OF REFERENCES ................................ ................................ ................................ ............... 71 BIOGRAPHI CAL SKETCH ................................ ................................ ................................ ......... 77

PAGE 8

8 LIST OF TABLES Table page 4 1 Chemical structures, names and other properties of the four repellent treatments. ........... 44 4 2 Results of F Tests in SAS 9.2 for the fitted model using pooled data from six subjects examining the effects of the chemicals, the testing media, the dose of repellent as well as their interactions. ................................ ................................ ................................ ... 44 5 1 Total number of mosquitoes bioassayed per treatment for six human volunteers. ............ 61 5 2 Mean ED 50 (95% CI) and ED 50 ratios fo r four chemicals estimated using repellency bioassay data (nmol/cm 2 ) pooled from six human volunteers on membrane and skin. ..... 61 5 3 Mean ED 95 (95% CI) and ED 95 ratios for four chemicals estim ated using repellency bioassay data (nmol/cm 2 ) pooled from six human volunteers on membrane and skin. ..... 61 5 4 Mean ED 99 (95% CI) and ED 99 ratios to skin and cloth for four chemicals estimated using repellency bioassay data (nmol/cm 2 ) pooled from six human volunteers on membrane, skin and cloth. ................................ ................................ ................................ 62 5 5 Densities calculated for all three bioassay types using the internal volume of each test chamber or cage (cm 3 ) and the number of mosquitoes used for each test (N). ........... 62

PAGE 9

9 LIST OF FIGURES Figure page 4 1 Pictures of in vitro load ing and feeding modules. Photos courtesy of Natasha M. Agramonte. ................................ ................................ ................................ ......................... 45 4 2 Assembled mosquito loading and feeding modules with arrow indicating the CO 2 tubing connected in parallel to the fron t of the six testing chambers via the cork hole. Photo courtesy of Natasha M. Agramonte. ................................ ................................ ........ 45 4 3 Confirmation of mosquito blood feeding from module tests. Photos courtesy of Natasha M. Agramon te. ................................ ................................ ................................ ..... 45 4 4 Assembled mosquito loading and feeding modules connected to heated water circulator via tubing for in vitro bioassay, showing rubber elastic bands placed around each of the six te sting chambers. Photo courtesy of Natasha M. Agramonte. ....... 46 4 5 Stacked column chart comparing the mean attraction for the volunteers individually and pooled comparing in vivo vs. in vitro mo squito feeding on the control with aste ....... 46 4 6 Line graph of dose response data indicating the proportion of mosquitoes repelled by DEET on skin ( in vivo ) and on membrane ( in vitro ) with asterisk indicating where ................................ ...... 47 4 7 Line graph of dose response data indicating the propo rtion of mosquitoes repelled by KBR3023 on skin ( in vivo ) and on membrane ( in vitro ) with asterisk indicating ........................... 47 4 8 Li ne graph of dose response data indicating the proportion of mosquitoes repelled by IR3535 on skin ( in vivo ) and on membrane ( in vitro ) with asterisk indicating where ................................ ...... 48 4 9 Line graph of dose response data indicating the proportion of mosquitoes repelled by PMD on skin ( in vivo ) and on membrane ( in vitro ) with asterisk indicating where the two media differed from each other sign ................................ ............ 48 4 10 Dose response curves produced by PoloPlus displaying the percentage of mosquitoes repelled by DEET on skin (left) and on the membrane (right) over a range of ch emical doses on a log scale. ................................ ................................ ........................... 49 4 11 Dose response curves produced by PoloPlus displaying the percentage of mosquitoes repelled by KBR3023 on skin (left) and on the membrane (right) over a r ange of chemical doses on a log scale. ................................ ................................ ........................... 49

PAGE 10

10 4 12 Dose response curves produced by PoloPlus displaying the percentage of mosquitoes repelled by IR3535 on skin (left) and on the membrane (right) over a range of chemical doses on a log scale. ................................ ................................ ........................... 50 4 13 Dose response curves produced by PoloPlus displaying the percentage of mosquitoes repelled by PMD on skin (left) and on the membrane ( right) over a range of chemical doses on a log scale. ................................ ................................ ................................ ........... 50 5 1 Pictures of in vitro loading and feeding modules. Photos courtesy of Natasha M. Agramonte. ................................ ................................ ................................ ......................... 63 5 2 Assembled mosquito loading and feeding modules with arrow indicating the CO 2 tubing connected in parallel to the front of the six testing chambers via the cork hole. Photo courtesy of Natasha M. Agramonte. ................................ ................................ ........ 63 5 3 Confirmation of mosquito blood feeding from module tests. Photos courtesy of Natasha M. Agramonte. ................................ ................................ ................................ ..... 63 5 4 Assembled mosquito loading an d feeding modules connected to heated water circulator via tubing for in vitro bioassay, showing rubber elastic bands placed around each of the six testing chambers. Photo courtesy of Natasha M. Agramonte. ....... 64 5 6 Ae. aegypti feeding behavior on hands without (left) and with (right) repellent DEET. Photo courtesy of Greg Allen. ................................ ................................ ............... 65 5 8 Cloth patch test showing confirmation of mosq uito blood feeding. Photos courtesy of Greg Allen. ................................ ................................ ................................ ......................... 66 5 9 Bar graph of the mean ED 99 (nmol/cm 2 ) estimated for each of the four repellent treatments evaluated on three media. ................................ ................................ ................. 66 5 10 Bar graph highlighting the mean ED 99 (nmol/cm 2 ) estimates on skin and cloth media for the four repellent treatments rescaled for direct comparison. ................................ ...... 67

PAGE 11

11 Abstract of T hesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science COMPARISON OF A SILICONE MEMBRANE BLOOD FEEDING SYSTEM TO HUMAN SKIN IN REPELLENCY BIOASSAYS WITH AEDES AEGYPTI MOSQUITOES By Natasha Marie Agramonte December 2012 Chair: Ulrich Bernier Major: Entomology and Nematology It is expected that laboratory based repellent bioassays should reliably evaluate the efficacy of compounds that deter mosqui to feeding behavior. The variety of repellent bioassays available allows for flexibility in design, but makes it difficult to compare any two methods, including in vitro and in vivo comparisons. The most reliable data come from skin assays; however, this e xposes volunteers to chemicals and mosquito bites. In this study, four repellent active ingredients were used: DEET, IR3535, picaridin and para menthane 3,8 diol. Results from bioassays with a module based method were operated in vitro with a membrane and in vivo on the skin of the leg and were then compared to a standard in vitro method where repellent treated cloth is placed over an arm that is inserted into a cage of mosquitoes. Pooled data from six volunteers were used to create dose response curves and estimate effective doses for four r epellents at the 50, 95, and 99 % levels (ED 50 ED 95 ED 99 ) using dose response curve s with a probit model from the module tests. The ED 99 was estimated with repellent treated cloth as the concentration that prevented 99 % of the mosquitoes from feeding. The baseline attraction did not differ statistically for the membrane and skin module tests. The membrane and skin module results also did not differ significantly as repellent treatments below 31 nmol/cm 2 dose level. Howe ver, since the chemical dose at which the skin and

PAGE 12

12 membrane curves diverge is lower than standard thresholds used in screening these chemicals, the use of the membrane module system will require further alterations before it can substitute current screenin g methods using human volunteers. Interestingly, all four repellents tested on the membrane required much hig her doses of repellent chemical to reach full protection. Based on the results of this study, cage based tests appear to be a more reliable estimat e of repellent activity on skin compared to module based tests on membrane. However, with the knowledge of the effective dose ratios, module based tests can be utilized for testing with infected vectors with previously evaluated repellent s

PAGE 13

13 CHAPTER 1 INTR ODUCTION Mosquitoes and their Medical Relevance There are over 3,500 described species of mosquitoes (Diptera: Culicidae), of which 156 species are of significant medical importance (Harbach 2012, Gaffigan et al. 2012, WRBU 2012). While both male and femal e mosquitoes feed on plant sugars, only adult females rely on haematophagy as a source of protein for the production of eggs. As humans have evolved with mosquitoes, our experience with these organisms has been less than pleasant especially since humans a re the preferred host for anthropophilic species, and serve as suitable alternatives for other opportunistic species. The severity of this experience ranges from a nuisance during the biting process and the allergic response that itches and irritates the s kin, to the transmission of pathogens which cause debilitating disease and death. Over the last few decades, global climate change urbanization and increased human travel have contributed to the growing threat and spread of vector borne diseases ( Weaver a nd Reisen 2010, Gubler et al 2001 ) Countries that had previously eradicated or controlled mosquito borne illnesses, such as yellow fever, dengue, and malaria are experiencing a marked resurgence in cases. In Brazil, where a malaria eradication program ha d previously controlled the local vector population, the incidence of malaria increased 76% in the Amazon region primarily due to migratory workers and deforestation related to urbanization efforts ( Gratz 1999, Sawyer 1986 ) Additionally, arthropod borne viruses or arboviruses have emerged and spread into areas where they formerly could not have been sustained. Increased temperature and rainfall in areas with a previously temperate climate have provided new breeding grounds for tropical species. The vector for dengue and yellow fever, the Aedes aegypti (Linnaeus) mosquito originally from Northern Africa, has established itself in much of the western hemisphere including the United

PAGE 14

14 States. However, unlike many Central and South American countries, dengue fe ver is not endemic to the U.S. (Gratz 1999). Some dengue transmission had previously been reported along the Texas Mexico border since 1980, and more recently several locally acquired cases of d engue in Key West, FL were confirmed in 2009 2010 (CDC 1999, C DC 2010). Knowledge of these cases and of juxtaposed Ae. aegypti mosquito populations highlight the need to explore preventative measures in anticipation of dengue outbreaks. The most recent major arbovirus outbreak of note was the West Nile virus epidemic in which most cases in the United States occurred from 1999 to 2002 (Huhn et al. 2003). This event was particularly significant for developmental repellent research because it promoted the use of insect repellents to the public for the prevention of mosq uito borne diseases. A Brief History of Repellent Research L aboratory based mosquito repellent testing methods can be classified broadly into those that involve application of compounds to skin ( in vivo ) and those that involve some surrogate, such as appli cation of the compound to a membrane or cloth ( in vitro ) The in vivo methods simulate closely the actual use of insect repellents because the repellent is applied directly on the skin surface. In vitro methods for mosquito repellents refer to the testing in an indirect fashion, such as testing a candidate repellent soaked into a piece of cloth, followed by placement of the treated cloth over the skin surface. A common assumption in this area of research is that in vitro testing methods provide a fairly acc urate means to screen, predict and extrapolate in vivo testing results and that these assays can be conducted in a convenient environment that exposes research staff and study volunteers to less overall risk. While the use of plant derived compounds to red uce the nuisance from insects has been documented since the time of the Egyptians by Herodotus and found across many native cultures, the extensive research and development effort to discover modern insect repellents

PAGE 15

15 began during World War II with a goal t o reduce the impact of disease to U.S. troops fighting overseas (Moore and Debboun 2007). An overwhelming breakthrough for the field of repellent research came in 1952 when N,N diethyl 3 methylbenzamide (DEET) was invented by USDA in Beltsville, MD and tes ted by the USDA laboratory in Orlando, FL (Bernier and Tsikolia 2011). Since its discovery, DEET has been considered the gold standard for repellent testing against mosquitoes and many other arthropod species due to its effectiveness at low doses, broad sp ectrum repellency and low toxicity to humans. Even DEET, however, is not an ideal repellent. DEET is not recommended for use on children or pregnant women, has an oily texture when applied on the skin, and is known to have plasticizing properties (Moore an d Debboun 2007). Since the discovery of DEET, there are a number of effective repellents have been developed from two main areas, plant derived ex tracts and synthetic compounds.

PAGE 16

16 CHAPTER 2 A REVIEW OF MOSQUITO HOST LOCATION Chemical Cues and Olfaction Semi oche micals and their Role in Olfacti on Semiochemicals are chemical s that serve as mess engers emitted by an organism that can modify the behavior emitting or receiving the chemical. Semiochemicals can be subdivided into three major classes, kairomones, allo mones, and synomones. A kairomone is a chemical message that is released by one organism but which benefits the organism that receives the chemical message often to the detriment of the producer. Kairomones are contained within the skin emanations of human s and other animals. These kairomones have the undesired effect of providing the insect pest with olfactory cues to find its host. Conversely, an allomone is a chemical message that benefits the organism producing it rather than the organism that receives the chemical message. Allomonal compounds frequently associated with mosquitoes are those that deter a mosquito from finding a host. A synomone is a chemical message that benefits both the producer and the receiver such as a pheromone use d to attract a s ex of one species to the other sex Of these, kairomones and allomones are the most significant in odor mediated host vector relationships. Carbon Dioxide as an Attractant Several chemicals have been historically implicated as kairomones for the benefit of mosquito attraction to human hosts. Some of these chemicals include carbon dioxide, lactic acid, acetone, ammonia, 1 octen 3 ol, among others. The first of these, carbon dioxide is the oldest known attractant and is one highly prevalent both in the litera ture and in commercially available mosquito trapping equipment. Carbon dioxide has been debated as to whether it acts only as an activator to flight towards a potential host or whether it also has a synergistic response in

PAGE 17

17 combination with host odors ( Dekk er et al. 2005, Roachell 1997, Gillies 1980). Some believe that carbon dioxide exhaled by a vertebrate is utilized by mosquitoes as well as other haematopha gous arthropods as a cue for foraging behavior, helping them orient towards a host organism (Lehane 2005). Khan et al. examined the role of carbon dioxide and its role at various concentrations in 71 ). Addition of carbon dioxide did not increase blood feeding behavior via probing, but was found to illicit an increase in flight activity towards the host although an increase in the amount of carbon dioxide did not increase flight activity further. Attraction to a human hand was found to be much preferred than a combination of heat, moisture and carbon dioxide, and the study was pessimistic that this sort of combination would not be sufficient to redirect mosquitoes from a host, stating that carbon dioxide was believed to act primarily as a flight stimulant in mosquitoes (Khan et al. 1967). Sk in Emanations and their Metabolites as Attractants Hundreds of semiochemicals have been identified from compounds dissolved in sweat collected on human skin. Some of these are from sweat glands, from the breakdown of the foods we eat, and from bacteria tha t live on our skin. There were more than 270 chemicals identified on human skin (Bernier et al. 2000, 2002 ). Many volatile compounds found to be attractive to mosquitoes may occur as result of perspiration, which are produced by different types of sweat gl ands. Eccrine sudoriferous glands are most abundant on the hands and feet, but also on other parts of the body. Apocrine sudoriferous glands are more common in the armpits and the groin area. Sebaceous glands are found on the face and the scalp (Roachell 1997, Takken 1991). Studies using fresh vs. incubated sweat with Anopheles gambiae s.s. show that fresh sweat secretions are less attractive to mosquitoes than incubated sweat, indicating that bacterial decomposition of sweat by the microflora of the skin play a greater role in attracting mosquitoes

PAGE 18

18 than do volatiles derived directly from human sweat glands and incubated sweat in the absence of bacteria diminished the attractiveness of the sample (Braks and Takken 1999, Braks et al. 2000). Khan et al. belie ved that convective heat, rather than simply being attractive in and of itself, aided in mosquito host location by means of odor transport of skin chemicals (1967), which has some merit as well because volatile chemicals are stimulated by heat and are able to stimulate mosquito at greater distances. Schreck et al. also found this to be true, noting that more Ae. aegypti mosquitoes were attracted to heated rather than unheated residues on the surfaces of glass that had been handled by human hands (1981). Exa mination of skin exudates individually or in aggregates has increased the compounds known to attract mosquitoes. Brown suggested that the amino acids and estrogens have an additive effect to the attractiveness of a host to a mosquito, particularly when the se are built upon that chemicals other than carbon dioxide must be sufficiently attractive in its absence since only a negligible amount of carbon dioxide is r espired via the skin (Frame et al. 1972). Ammonia was found to be an important attractant for An. gambiae s.s. mosquitoes, unlike lactic acid, and this was likely cause by the breakdown of urea and amino acids by bacterial microorganisms on the skin surfa ce, increasing the pH of the incubated sweat (Braks et al. 2001) Incubated sweat was also found to be attractive to Ae. aegypti thus ammonia may also be an attractant for this species (Geier et al. 1999). 1 octen 3 ol was originally identified as an attra ctant for tsetse flies from oxen was also found to be attractive to mosquitoes (Kline 1998). Human foot odor was attractive to Ae aegypti with 66% of mosquitoes responding; however this was still less than the 80% which found the human hand to be most att ractive in dual port olfactomet er tests (Kline 1998).

PAGE 19

19 Some other potential attractants have been largely underexplored Limburger cheese was identified as attractive to Anopheles gambiae sensu strict o (Knols and De Jong 1996), however despite its similarit y to human foot odor, limburger cheese was not found to be attractive to Ae aegypti (Kline 1998). Implications have been made about a correlation of attractiveness and human blood type but these have been largely refuted (Curtis 1986). Ahmadi and McClell and have suggested from experiments on Aedes sierrensis that an additional factor in host seeking behavior might be an aggregation pheromone from mosquitoes; however, this theory appears to have garnered little support from a lack of subsequent testing or mention by other scientists (Ahmadi and McClelland 1985). Mehr et al. found that at concentrations of 10 6 to 10 4 DEET was attractive to Ae. aegypti and Aedes taeniorhynchus but not to Anopheles albimanus (Mehr et al. 1990), however this result was foun d to occur only over a short range of very low concentrations and the primary use of DEET currently is as a mosquito repellent. L Lactic Acid as an Attractant L lactic acid is another chemical that has been widely studied and debated for its attractive abi lities (Reference, Acree et al. 1968 Science paper). Smith et al. reported that lactic acid was an attractant to Aedes aegypti at low concentrations, similar to those found on the skin surface or from breath, but repellent at much higher doses (19 63 ). Chem ical analyses of human skin emanations have shown that carboxylic acids and lactic acid are present in highest concentrations among many other volatile and moderately volatile skin compounds (Bernier et al. 2000, 2002). Bernier found evidence of attractant chemical synergism when he tested binary blends using L lactic acid mixed with either dichloromethane, dimethyl disulfide or acetone and found that these blends produced a greater attraction of about 80% in Ae aegypti mosquitoes than the mean attraction of about 50% for the two individual chemicals (Bernier et al. 2003). Later, a three compound mixture using lactic acid, acetone and dimethyl disulfide was found to provide even

PAGE 20

20 greater attraction with almost 85% of mosquitoes responding to the blend (Berni er et al. 2007 a ). Conversely to the results found for Ae aegypti tests done with Culex quinquefasciatus found lactic acid to be attractive at concentrations that would normally be repellent to other species (Braks et al. 1999). Lactic acid was found to be attractive to Ae aegypti as well as surfaces touched by hands, but it has been contested that this is only true in the presence of carbon dioxide (Schreck et al. 1981). Repellent Chemicals and Associated Mosquito Behavior Repellents belong to a family of substances called allomones, i.e. substances that benefit the sender of chemical signals rather than receiver. chemical which causes insects to make oriented movement away from its source (1960) Usually, this is accomplished by a mosquito or other arthropod coming into contact with the repellent applied to the skin surface of a human or other animal. There is some disagreement with what is strictly considered a repellent vs. other classifications such as a pest i cide or a deterrent (Cantrell et al. 2005) Although the US EPA definition of pesti cides is any substance or mixture of substances intended for preventing, destroying, repelling, or mitigating any pest, the true mechanism of repellency does not necessaril y involve insecticidal action ( 20 12 ) While insecticides often irritate an insect and cause it to move away from the target source, sufficient exposure to the insecticide will ultimately kill the insect. There are many examples of repellents that are relat ively non toxic to target organisms. Dethier also noted that substances often classified as repellents may sometimes be more correctly defined as errent s n situations where the inhibition of feeding or oviposition is being measured rather than the mo vement away from a source However, Dethier et al. also noted that the same compound may have multiple (Dethier et al. 1960).

PAGE 21

21 A second type of allomone involved in the vector host relationship is an attraction inhibitor This is a compound produced by a host organism which induces anosmia or hyposmia in arthropods thereby masking host kairomones from the arthropod vector (Bernier et al. 2007 b ). This is the least understood class of compounds wi th the greatest potential for future researc h. Mosquito Biology and Behavior Used in Host Location Mosquito host location is controlled by both inherited behaviors and environmental factors whereby a mosquito is able to sense a potential host from a distan ce, locate a host in space at a short range, confirm the identity of a potential host, and initiate feeding behavior. While seemingly simple, these interactions rely on a number of cues that a mosquito must interpret in the central nervous system, weigh ag ainst internal cues and process to determine the appropriate action to be taken. Initiation of host location behavior can be passive or active. Some species actively seek out hosts once their internal physiological state has been assessed as adequate for t he development of eggs. Other species feed opportunistically when a potential host wanders into a certain radius, which varies by species or by the strength of the odor stimulus, where a mosquito is resting (Day 2005). Once an organism has entered into thi s proximity, mosquitoes follow an odor plume produced by both skin chemicals and carbon dioxide towards a potential host organism (Dekker et al. 2001). Mosquitoes usually fly upwind into a downwind plume. This anemotaxis towards human odors has been repor ted to occur from up to 30 ft away (Brown 1966). Mosquitoes may also orient upwind under continuous odor stimulation (Geier et al. 1999). Evidence of anemotaxis to host odors was provided by an experiment by Happold where individuals stood on a 20 ft platf orm above a forest (Happold 1965). The individuals were undisturbed by mosquitoes until a fan was used to direct a wind over the individual and into the forest, after which mosquitoes flew upwards 20 ft to the platform (Maibach 1969). In locating a host, a mosquito moves along a

PAGE 22

22 route which provides an increasing concentration of stimulus and turns away from areas where the stimulus gradient declines (Brown 1966). This mechanism relating to the turning frequency, independent of orientation (which may also r esult), is known as klinokinesis. Environmental conditions also play an important role in long range host location, particularly ambient between 50 90% is necessary fo r a mosquito to adequately locate a host (Brown 1966). The blood seeking behavior in a female mosquito relies on receptors found on several anatomical structures to sense its external environment. Receptors vary in type and in number ody such as the ommatidia in the compound eyes, grooved pegs, small sensilla coeloconica and ampullacea, capitate palpal pegs, and for female Ae. aegypti the blunt tipped type I trichodea and short, pointed trichodea (McIver 1982). Receptors located in se nsilla transduction studies to be used in the detection of odors from potential host organisms (Zwiebel and Takken 2004). There are also receptors for odor, heat tarsi and proboscis, but these play a larger role in close range and tactile recognition, whereas the sensory receptors in the antennae are used in long range detection (McIver 1982 ). The exact function and binding abi lities of many receptors are still an area of active research, particularly in relation to specific attractant or repellent chemical substances. Sensory receptors are used for the detection of heat (thermoreceptors), the detection of specific classes of ch emicals (chemoreceptors), the detection of humidity (hygroreceptors) and the detection of movement (mechanoreceptors) (McIver 1982). Also, visual receptors in ommatidia are used for light/dark detection and shape recognition in long range host location (M cIver 1982). Visual cues are also utilized in long range detection of a host and Ae. aegypti has been found to be attracted to darker

PAGE 23

23 colors in clothing (Brett 1938). Although the mosquito uses both visual and olfactory cues to orient toward a host, the ol factory system plays a prominent role in the response and its subsequent action (Bowen 1991). After landing, tactile and thermal stimuli play an important role in confirming the host and the decision making process to f eed on the host (Maibach 1969). Mosqu itoes also react to their own physiological state when determining whether or not they require a blood meal. Adult female mosquitoes reproduce and sugar feed soon after emerging and then a few days later, they seek a host for a bloodmeal. Internal cues suc h as a changes the tendency of a mosquito to seek out a host (Bowen 1991). In a 1984 study by Davis, Ae aegypti mosquitoes were found to not display host seeking be havior within two days of adult emergence, but after 5 days more than 90% of female mosquitoes display host seeking behavior and this continued consistently up to 15 days post emergence, showing that host seeking behavior is age dependent (Roachell 1997). Other studies have also found that the frequency of bloodmeal seeking increases in mosquitoes a s they age (Khan et al. 1971). Host seeking behavior studied in Culex pipiens stimulated with lactic acid has been shown to be related to sensitivity of sensilla type A3 receptors for females that have terminated diapause and resumed reproductive development vs. females that were still in reproductive diapause (Roachell 1997, Bowen 1991).

PAGE 24

24 CHAPTER 3 A REVIEW OF MOSQUITO REPELLENT TESTING ME THODS Current Standards for Repellent Testing Methods The major guidelines that govern the testing of mosquito repellents are contained in the three main documents that delineate the recommended materials and methods for testing. Two of the se guidelines are produced by the Americ an Society for Testing and Materials (ASTM) for the laboratory testing of mosquito repellent formulations on the skin (2000a) and for the field testing of topical repellent compounds (2000b) The third is guidelines fo r efficacy testing of mosquito repellents for use on human skin (20 09). Although the methods are meant for application to a range of vector species, the WHO recommends the use of Aedes aegypti for use in laboratory tests. There are several reasons for sele cting this particular species. Not only does Ae aegypti respond well to laboratory rearing and can produce a large amount of insects for testing at regular intervals, but it is also an avid biter. Additionally, Ae aegypti is a diurnal feeder which is pre ferred over other crepuscular mosquito species. This single species of mosquito is often used for large scale screening of potential repellent compounds, but other medically important genera should also be considered. Mosquito species of the Anopheles and Culex genera are frequently used in addition to Ae aegypti mosquitoes in the most accurate evaluations (WHO 2009). The benefit of using all three genera in laboratory testing is to provide for a range of repellent responses. Data derived from experiments using all three genera can then more accurately predict the response of the mosquito population that the researcher is attempting to characterize, since several spe cies inhabit the same regions. The three previous guidelines cited stress safety and standar dized processes which are aimed at providing a baseline for comparative analysis. This baseline is important as the

PAGE 25

25 techniques used for repellent testing of mosquitoes can vary widely from region to region and across different target species of mosquitoes. While these methods allow for customization, the basic recommended procedures stated by the guidelines are noted below alongside examples of individual in vitro and in vivo methods. First, to provide some context for the individual examples, the three ma jor systems of repellency measurement will be described to establish the basics involved in all testing methods. Measurements of Repellent Efficacy With the increased need for repellent research came the necessity to quantify the efficacy of potential repe llent chemicals. The most widely used measure for determining effectiveness was one of the earliest developed, the time to first bite (Granett 1940) or complete protection time (CPT) which measures the time that the repellent offers 100% protection from bi tes. For minutes to a biting insect until a bite occurs. The complete protection time is then recorded when a confirmatory bite occurs either during the same th ree minute period or during the subsequent three minute expo sure period. Another metric to assess repellency that is frequently used is percent protection, which calculates the percentage of biting mosquitoes prevented from taking a blood meal during a spe cified amount of time. This measurement is common for laboratory rather than field use as it requires an accurate count of the total female mosquitoes exposed to the test subject as well as the total number of blood fed mosquitoes (Barnard et al. 2007). Fe males mosquitoes are almost exclusively used in the testing of mosquito repellents as they and not males are haematopha gous or blood feeding, requiring a blood meal in order to lay eggs. Blood feeding can be determined by direct observation of mosquito pro bing behavior on the test subject, by counting the number

PAGE 26

26 of visibly engorged mosquitoes, or by sacrificing and examining the mosquitoes directly by individually crushing them on white paper and noting blood spots. A third, more involved measure of repelle ncy is the ED 50 measurement or average effective dose. While the time to first bite measurement offers a practical and intuitive frame of reference for the typical repellent user, the ED 50 or sometimes the ED 90 or ED 9 9 provide the researcher with a standar dized measurement for comparative analysis over a range of testing mosquitoes for a thirty second period first without repellent, as a control exposure, and then with increasingly higher doses of the experimental repellent compound until complete repellency is attained. Mosquito biting is measured through observation of probing behavior during the exposure period. After the complete repellency dose is determined, the arm is re introduced hourly until the number of bites received during the exposure period equals 50% of the bites received during the control exp osure (Moore and Debboun 2007). A fourth measurement conducted to measure repellent efficacy in the laborat ory is the measurement of the minimum effective dosage (MED) (USDA 1977). The MED is an estimate of the surface concentration at which a repellent compound fails to prevent bites. Depending on the number of mosquitoes in the trial and the defined failure p oint, e.g. for 500 mosquitoes and a failure point of 5 bites, the MED is essentially an estimate of the ED 99 The MED also has a bearing on the measurement of CPT. As a compound evaporates from skin, the ability of the compound to repel will decrease and i n theory, the surface concentration at which the repellent fails is the MED. However, the CPT measurement is influence d by many other factors such as evaporative loss of the compound, abrasive loss from volunteers usually by accidental contact with other s urfaces, and dermal absorption and migration. Measurement of the MED removes all

PAGE 27

27 these other sources of variability and focuses solely on the amount of repellent needed to deter mosquitoes (or other arthropods) from biting. A Survey of in vitro Test Method s The use of cloth sleeves as an in vitro test method is perhaps the simplest experimental design of this type. The idea behind the use of a cloth sleeve stems from the precautionary need to avoid direct contact with the skin. Direct contact with the skin is avoided in compounds where toxicological studies have not been performed and the safety of the chemical for use on skin is questionable. Cloth sleeves typically are fabricated of a well ventilated stretchable material to allow for the emanations from hu man skin to pass through and illicit the host seeking behavior in the mosquitoes. These sleeves are often constructed for testing on the forearms, spanning the area from the wrist to elbow. A sleeve can also be used for testing on the lower leg, spanning t he area from the ankle to the knee. In order to prevent direct contact of the treated sleeve with the exposed skin of the participant, a second untreated sleeve is worn underneath the first to provide a protective barrier since the chemical treated in the sleeve may produce harmf ul effect to humans (WHO 2009). Experimental repellent chemicals for use on cloth are diluted in 95% ethanol or acetone and the concentration applied is between 1.5 mg/cm 2 or 1 g of compound for a 650 cm 2 area. Ethanol and acetone a re used as solvents in these experiments due to their high volatility. After the application of the chemical mixture to the cloth, the sleeve is allowed to dry for a specific amount of time. Typically fifteen minutes is sufficient time to allow for all of the volatile solvent to evaporate from the cloth sleeve, leaving only the desired test chemical. Once dry, the sleeve can be placed over the arm or leg of the volunteer. When testing the forearm, a latex glove is worn over the hand to protect the volunteer from painful bites on that sensitive region of the body. The arm or leg is then placed into a screened cage containing a minimum of 200 adult

PAGE 28

28 female mosquitoes of a certain species, aged 5 10 d and preferably starved for one hour pr ior to testing (Barnard 2005). A modification of the cloth sleeve experiment is the use of a cloth patch. In this experiment the area of skin on the arm or leg exposed to bites from mosquitoes is much smaller, and can vary according the needs of the experimenter. One example of this test uses a 50 cm 2 area of cloth soaked in the test chemical/solvent mixture as used for the cloth sleeve ( Carroll et al. 2011, Katritzky et al. 2008 2010 Rosa et al. 2012 ). The concentration of test chemical is also similar, often 1.5 mg/cm 2 For testing, however, the volunteer requires not only a latex glove to cover the hand, but also an impenetrable material to cover the rest of the arm as well. One model for this covering is the use of a Velcro lined vinyl sleeve fashioned to the shape of the v arm with area cut out to allow for mosquitoes to bite through. The cloth patch can then be placed over this window during the testing procedure. The cloth patch modification is beneficial for large scale screening of chemicals because less cloth is used per test chemical/concentration and cloth patches can easily be replaced over the vinyl sleeve window for rapi d trials (Bernier et al. 2005). When human volunteers are scarce, the use of animals can provide a skin surface on which repellent chemi cals can be tested. Rats, rabbits, gerbils, or guinea pigs are among the animals most often chosen for this type of assay. The desired area of skin surface is exposed by first shaving the animal and then the repellent chemical is applied to the skin surfac e. Sedation is recommended for the animals in these tests so that the chemical is not removed by rubbing and so that the mosquitoes may feed freely (Robert et al. 1991). Another benefit of this type of experiment is that it removes the potential risk to hu man volunteers with the use of wild caught mosquitoe s in disease endemic areas.

PAGE 29

29 The use of synthetic membranes that mimic the skin surface can be utilized as well. Several types of synthetic membranes have been explored as cost effective means to avoid th e use of human volunteers. Some of these membranes include collagen film, cow mesentery or baudruche film, and lamb skin membranes. These synthetic membranes can be stretched over receptacles of any size filled with a blood source. It is preferable that th e blood source be maintained at a temperature of 28 to seeking out for their host blood meal will be consistent with the membrane/blood reservoir system. Although the concept of using an artificial m embrane can be applied to most in vitro methods, this particular modification was developed in conjunction with the Klun & Debboun (K & D) in vitro module (Klun et al. 2005). The K & D module itself is a modification of an ASTM module for the testing multi ple repellents on the skin of a single research participant at once (ASTM 2000b). The K & D module consists of a Plexiglas rectangle divided into six individual cells measuring 3 cm by 4 cm by 4 cm each. Each cell was designed with a small, pluggable hole at the top of the cell for the loading of 5 mosquitoes into each compartment. At the bottom of each cell is a Plexiglas sliding door to introduce the mosquitoes to the desired repellent testing surface, in this case the membrane/blood reservoir system. Clo th patches cut to the dimensions of then placed in between the K & D module containing the mosquitoes and the blood reservoir system. An uncommon and often crit icized method for determining repellency involves the use of an olfactometer. Olfactometers are elongated chambers which are normally used for determining attraction in mosquitoes. Mosquitoes are placed at one end of the chamber and an odor stimulus is int roduced at the other end of the chamber. The circulation of air uniformly through the

PAGE 30

30 chamber allows an odor plume produced by the chemical stimulus to guide mosquitoes. Movement of the mosquitoes in the chamber toward the source of the odor stimulus provi des a basis as to whether a chemical is attractant or not. If a chemical is not attractant, the mosquito will not be stimulated to fly and will remain in the area of the chamber where it settled before the test. Some researchers have modified this design t o allow for the testing of attraction inhibitors and other repellents in the olfactometer ( Bernier et al. 2005, Kline et al. 2003 ). One such modification involves the use of a three chamber olfactometer. In this design, mosquitoes are places in a median ch amber prior to the test and allowed to acclimate for an hour. At the start of the test a repellent chemical is placed in one end, referred to as the proximal chamber. At the other end of the olfactometer is the distal chamber which has a mesh netting funne l which mosquitoes must pass through into a collection tube. Barriers separating the proximal and distal chambers from the median chamber are opened once the chemical stimulus has been placed in the proximal end. Movement of the mosquitoes away from the so urce of the chemical stimulus is indicative of a repellent and the repellency can be quantified by calculating the percentage of mosquitoes in each of the three chambers compared to the total number of mosquitoes (Dogan and Rossignol 1999). For this design male mosquitoes may be used as well as female mosquitoes since the focus is not on mosquitoes seeking a blood source, but simply being repelled from their p osition in the median chamber. A Survey of in vivo Test Methods In contrast to the great variation of in vitro methods, there are only a few in vivo methods. In vivo methods are often more rigid than in vitro methods since only a few chemicals that have previously passed toxicological safety tests can be administered and screened on a volunteer at a ti me. However, in vivo methods provide fairly consistent results, with high correlation to other in vivo tests and are realistically comparable to the end use of the repellent.

PAGE 31

31 Topical skin repellent tests are very closely related to the cloth sleeve tests, and are very similar for both laboratory and field tests. Experimental repellent chemicals are diluted to the desired concentration, often in 25% ethanol or acetone. The concentration typically applied is 1.5 mL of compound for a 650 cm 2 area, the average skin surface area from the wrist to the elbow. Typically fifteen minutes is sufficient time to allow for all of the volatile solvent to evaporate from the skin surface, leaving only the desired test chemical. If the forearm is being tested, the volunteer s hould place a latex glove over the hand to protect from painful bites. The arm is then placed into a screened cage containing a minimum of 200 adult female mosquitoes of a certain species, aged 5 10 d and starved for one hour prior to t esting (Barnard and Xue 2004). Human skin topical repellent tests in large outdoor cage or field tests follow the same basic procedure as the laboratory tests but are subject to the weather and environmental effects of the region in which the test is performed (Debboun et al. 2000, ASTM 2000a). Recommendations from the WHO guidelines indicate that repellent tests of this kind should be performed in the vicinity of human homes, have test subjects spaced 10 m apart and rotated in a randomized fashion to minimiz e positional error s (WHO 2009). The K & D module for in vivo testing systems, like the module used for in vitro testing also consists of a Plexiglas rectangle divided into six individual cells measuring 3 cm x 4 cm x 4 cm each. Each cell was designed with a small, pluggabl e hole at the top of the cell for the loading of 5 mosquitoes into each compartment. At the bottom of each cell, which is curved for in vivo methods to lay closer to the skin surface of the rounded leg or arm, is a Plexiglas sliding door to introduce the m corresponding rectangular treatments have been delineated and dried (Klun et al. 2006). This design is a marked improvement over the ASTM module, although is it also still used (Rutledge

PAGE 32

32 and Gu pta 2004). The ASTM module consists of a long, rectangular loading cell where 10 to 15 mosquitoes are loaded from a pluggable hole in one side. The bottom of this module is fashioned with five 29 mm diameter openings which are placed on the arm or leg skin surface of the volunteer where corresponding treatments have been deli neated and dried (ASTM 2000b). Previous Reviews of Mosquito Repellency Methods While there is great diversity in the techniques employed in repellent testing methods, and several scient ists have utilized these various methods to compare the efficacy of repellent compounds across mosquitoes species or even provide comparative analysis across several medically relevant insect orders, in the past forty years only two researchers have attemp ted to provide a direct comparative analysis and review of the many of the methods involved. Earlier reviews of this topic have had a broader focus, providing critiques of insect repellents in general, as in the 1977 article by Schreck. While mosquitoes pr repellent testing measures for tick, fleas, and other biting flies are also mentioned in detail. While experimental designs can overlap to other insects from mosquito research, results from tick, flea, or other biti ng fly research are not usually applicable to mosquitoes and vice versa. One alteration Schreck advocates is the use of a large amount of mosquitoes for laboratory cage testing, suggesting that a higher biting pressure reduces the variability (Schreck 1977 ). W hile a thorough comparison of all relevant in vivo methods are presented Barnard largely disregards in vitro methods, citing that the comparability of in vitro methods is poor and should be avoided in favor of in vivo methods (Barn ard 2005).

PAGE 33

33 CHAPTER 4 C OMPARISON OF DOSE RESPONSE CURVES FOR MOSQUITO REPELLE NTS TESTED IN VIVO AND IN VITRO USING A MODULE SYSTE M Background Though most in vitro methods for mosquito repellent tests provide a fast, safe, inexpensive way to test chemica l compounds regardless of whether toxicity has been established or not, there are limitations that should be recognized when results from these methods are compared or even more importantly, used to estimate performance in the field. One limitation is the difficultly in correlation of results from in vitro studies to in vivo stud ies Since by its nature an in vitro test is not directly tested on the target host out of convenience or necessity, extrapolation of these results to in vivo systems without a bas is for comparison limits their usefulness F ew studies have been conducted to compare these methods and validate that they provide results that are an accurate prediction of how repellents will perform on skin. In vivo methods, however, are not without the ir disadvantages. These studies require h uman volunteers that are willing to subject themselves to the accumulation of bites and the potential for allergic reactions makes it difficult to perform statistically significant experiments with large sample size s. Although field testing is most similar to the real use of repellent chemicals, the risk of exposure to pathogens carried by wild mosquitoes in disease endemic areas makes it impractical and unethical to perform these outdoor experiments in some location s. Despite the drawbacks of current in vitro systems and their unpopularity with some researchers, in vitro systems are still valuable tools for the future of mosquito repellent research because they can serve to predict repellency when in vivo testing is not feasible. The desire to produce an accurate, rapid screening method has produced a myriad of in vitro testing methods in comparison to the handful of in vivo methods. Modification to some of the more promising in vitro methods could result in a mechani sm by which in vivo test results could be reliably

PAGE 34

34 predicted without the need for human volunteers. With the ability to narrow down potential repellents through screening the costly safety screening process for promising novel repellent compounds could al so be minimized. Furthermore the US EPA H uman S tudies R eview B oard (HSRB) has recently expressed concerns over use of humans in repellent studies, and these results would be of interest to US EPA for their new rules on repellent registrations The first o bjective of this research project wa s to examine the dose response curves of four mosquito repelle nt active ingredients tested in vivo on human skin and in vitro on a silicone membrane using a module system for significant differences. Materials and Method s Mosquito Rearing and Selection The mosquitoes used in all bioassays were female Ae. aegypti (Orlando strain, 1952) from the colony maintained at the Center for Medical, Agricultural, and Veterinary Entomology location of the United States Department of A griculture Agricultural Research Service (USDA ARS CMAVE) in Gainesville, FL. Pupa e were obtained from the colony and kept in laboratory cages where newly emerged mosquitoes were maintained ad libitum on a 10% sucrose solution at 25 28 o C ambient temperatu re, 60 80% relative humidity and a 14:10 (light:dark) photoperiod. Nulliparous female mosquitoes aged six to eleven days displaying host seeking behavior were pre selected from stock cages using a hand draw box and trapped in a collection trap (Posey et al 1981). Female mosquitoes were then transferred to a smaller cage ( 28,316 cm 3 ) from which the mosquitoes were sorted into groups of 10 by mechanically aspirating them into acrylic holding tubes (15 cm long, 1.25 cm wide) Each tube contained approximately 10 mosquitoes, contained by screen gauze on one end and s top pered with a small cork (Fisher Scientific, Catalog No. 07781D Size 1 ) at the other end. Mosquitoes held in the tubes were allowed to acclimatize

PAGE 35

35 for 15 20 min (Barnard et al. 2007) before being transfer red into a n empty chamber of the module for testing. Chemical Treatments and Control Chemical compounds used as repellent treatments included technical N,N diethyl 3 methylbenzamide 97% (Aldrich, CAS# 134 62 3 ) hereafter referred to as DEET tech nical 1 piperidinecarboxylic acid 2 (2 hydroxyethyl) 1 methylpropylester, 98% (CAS# 119515 38 7 ) hereafter referred to as KBR3023 tech nical ethyl 3 [acetyl(butyl)amino]propanoate 98% (CAS# 52304 36 6 ) hereafter referred to as IR3535 and technical p ar a menthane 3,8 diol 98% (Bedoukian, CAS# 42822 86 6) hereafter referred to as P MD The repellents used in this study are summarized in Table 4 1. Serial dilutions of these four chemical s were obtained by adding 1mL of completely denatured ethyl alcohol (Mallinckrodt) into 1mL of the previous concentration of chemical to produce a range of concentrations (7840 nmol/cm 2 3920 nmol/cm 2 1960 nmol/cm 2 980 nmol/cm 2 490 nmol/cm 2 245 nmol/cm 2 123 nmol/cm 2 61 nmol/cm 2 31 nmol/cm 2 15 nmol/cm 2 8 nmol/cm 2 ) when applied to the 14.19 cm 2 treatment area The control treatment consisted of only completely denatured ethanol. Alterations Made to Module from Previous Designs Several modifications were made to the module and protocols from those used by Klun and De bboun ( 2000 ) and Weldon et al. ( 2003 ) Modification s made to the original Klun and Debboun (K & D) module by Weldon included an increase in the i nte rnal volume of each chamber from 100 cm 3 to 125 cm 3 an increase in the spacing between the chambers from 0. 25 cm to 1.25 cm, and the opening under each chamber was reduced to a circular aperture with a diameter of 4.25 cm (Weldon et al. 2003 ). Modifications made the Weldon module system for this project include an increase in th e number of mosqui toes placed in each chamber from 5 to 10, to allow for more precise feeding proportions to be recorded at each concentration level. A

PAGE 36

36 glass plate with drilled holes in line with the blood wells and apertures of the feeding module was added. The glass plate was placed und er the feeding module to act as a n inert barrier between the feeding module and the loading module or between the feeding module and the skin of the volunteers (Fig. 4 1) This glass plate also prevent ed absorption of chemicals into the Plexiglass bottom o f the module from which contract with the chemicals was most likely to occur. Elastic r ubber bands were added to the sliding doors to prevent accidental opening of the chamber doors by the volunteer s during testing. A procedural modification to this experi ment was the use of l ayering of chemical doses on the skin or membrane so that each dose can be evaluated in all 6 chambers simultaneously This differed from other protocols that used different doses in each of the 6 chambers. The testing of the same con centration across all chambers, in combination with the increased spac ing between the chambers is an improvement over the original K & D design which assured that there wa s minimal effect of the repellent dose in one chamber affecting an adjacent chamber. An early obstacle encountered during the trials was the lengthy amount of time needed to test volunteers by running each of the six chamber s individually in a series To overcome this, a multi port connector was fashioned from nylon tubing and nylon barbed T ee connect ors, arranged in parallel that allow ed for CO 2 to be simultaneously pumped into several chambers (Fig. 4 2) This allowed three or six chambers to be tested concurrently and cut down testing time at least three fold fr om the previous testing sc heme. in vivo Module Bioassays on Skin For this part of the study the proportion of Ae. aegypti mosquitoes blood feeding on six human volunteers (4 male, 2 female) was determined for a range of concen trations to the repellent chemical s DEET IR3535, KBR30 23, and PMD T en nulliparous female Ae. aegypti mosquitoes (6 1 1 d) w ere loaded into each of the six testing chambers (5 cm x 5 cm x 5 cm) in

PAGE 37

37 the previously described loading module (Klun and Debboun 2000, Klun et al. 2005, Weldon et al. 2003, Rutledge and Gupta 2004). Each of the six chamber s in the loading module had two apertures, a 4.25 cm diameter circular aperture at the bottom of each chamber covered by a sliding door, as well as a small 1.5 cm diameter circular aperture on the front of each chamber closed by a small cork For the in vivo test, the loading module was placed over the thighs of each human volunteer The areas where test solution was to be delivered on the thighs were demarcated with 4.25 cm diameter circles corresponding to the six cham diluted dose of DEET IR3535, KBR3023, or PMD ( 7840 nmol/cm 2 3920 nmol/cm 2 1960 nmol/cm 2 980 nmol/cm 2 490 nmol/cm 2 245 nmol/cm 2 123 nmol/cm 2 61 nmol/cm 2 31 nmol/cm 2 15 nmol/cm 2 8 nmol/cm 2 ) wa s applied to the six circles in s uccessive layered doses from the lowest concentration to the highest concentration The first treatment applied in any set of tests was the ethanol control to establish the baseline for mosquito feeding behavior. The application wa s allowed to dry for 3 5 min to allow for the ethanol to evaporate. The loading module wa s then placed between the module and the skin to prevent direct contact of the module with the chemicals. A sliding door wa s opened under three chambers at a time to expose mosquitoes in those chambers to the repellent for a 3 min period At the end of the exposure period, the mosquitoes were knocked down with CO 2 gas via the corked hole, removed from the chamber with an aspirator and crushe d to record the proportion of mosquitoes blood feeding out of ten (Fig. 4 3) This process wa s repeated with the rest of the six chambers and with all concentrations of the repellent until all feeding in all chambers ceased. Volunteers for all repellent te sts signed informed consent forms and were enrolled in an IRB study (Project # 636 2005).

PAGE 38

38 in vitro Module Bioassays on Silicone Membranes Treated with Skin Odors In the second part of this study the loading module wa s placed over the feeding module with s ix 4.25 cm diameter wells, corresponding to the six circular openings in the loading chamber (Klun and Debboun 2000, Klun et al. 2005, Weldon et al. 2003, Rutledge and Gupta 2004). Each of the wells in the feeding module was filled with 7 mL of citrated bo vine blood, continuously pumping hot water through the feeding chamber with a Cole Parmer Polystat circulating water bath (Fig. 4 4) Prior to testing, strips of silicone membranes we re worn against the upper thigh by the volu nteers for 3 4 h, held in place with an Ace elastic bandage. The worn silicone membranes we re then placed across the six wells of the feeding module to come into contact with the blood, and the glass spacer was placed over the membranes leaving only the m embrane covered diluted dose of DEET IR3535, KBR3023, or PMD ( 7840 nmol/cm 2 3920 nmol/cm 2 1960 nmol/cm 2 980 nmol/cm 2 490 nmol/cm 2 245 nmol/cm 2 123 nmol/cm 2 61 nmol/cm 2 31 nmol/cm 2 15 nmol/cm 2 8 nmol/cm 2 ) wa s ap plied to the six circles on the silicone membrane above each well in successive layered doses from the lowest concentration to the highest. The first treatment applied in any set of tests was the ethanol control to establish the baseline for mosquito feedi ng behavior. The dose wa s allowed to dry for 3 5 min to allow for the ethanol to evaporate. The loading module wa s placed onto the feeding module and lined up to correspond to each well. A sliding door wa s opened under all six chambers to expose the mosqui toes to the repellent treatment for a 3 min period At the end of the exposure period, the mosquitoes we re knocked down with CO 2 gas via the corked hole, removed from the chamber with an aspirator and crushed to record the proportion of mosquitoes blood fe eding out of ten. This process was repeated with the rest of the six chambers and with all concentrations of the repellent until all feeding in all chambers ceased.

PAGE 39

39 Statistical Analysis Data from the six volunteers were pooled for each of the four repellen t chemical s and at each of the concentration level s to reduce random effects that were the result of biological testing as well as to minimize person to person variability in attraction of the mosquitoes. The p ooled data from the repellency bioassays were analyzed using a generalized linear mixed model using a binomial distribution with a probit link. The equation for the fitted model wa s: probit(y/n) = mu + Nchem + Medium + Nchem*Medium + x + Nchem*x + Medium*x Nchem*Medium*x + e where mu was the overall m ean, Nchem wa s the fixed effect of the chemical, Medium wa s the fixed effect of the me dium, Nchem*Medium wa s the interacti on between Nchem and Medium, x wa s a fixed covariate corr esponding to x = log(Dose+100), and the other terms correspond to interaction s wit h this variate. In addition, e wa s the random error with e ~N(0, 2 ). Also an overdispersion parameter was considered for this model. The above model was fitted using the procedure GLIMMIX as implemented in SAS 9.2 ( SAS 20 12 ) The s ignifican ce of the model term effects we re evaluated using an approximated F test. PoloP lus v2.0 was also used to produce dose response curves with the pooled data for visual comparison of treatments tested in vivo and in vitro ( PoloPlus 201 2 ). Results The results from the approximated F test analysis, summarized in Table 4 2, indicate that t here are significant differences for all of the effects accounted for in the fitted model. Hence, chemical and medium differ among its levels and there are significant interactions among them. Also, the continuous variate (logDose) is significant, and the slope associated with each of the factors Nchem and Medium and their interaction, are also highly significant.

PAGE 40

40 A comparison of baseline attraction revealed that the mean blood feeding on the control treatment was 59.16% for in vivo and 56.17% for in vitro which is not a statistically significant difference for all volunteers pooled together (Fig. 4 5). Data from preliminary tests conducted with unworn membranes using both blood and a 10% sucrose solution treated with red food coloring failed to adequately attract mosquitoes, 38.3% and 12.8% bloodfeeding respectively, in the module tests at a level comparable to that of human skin. When examined individually, volunteer M5 had a significantly different proportion of mosquitoes blood feeding on skin vs membra ne without repellent (Fig. 4 5), and because of the high variability in individual data only test for paired means at the p=0.05 level. Line graphs of the pooled dose response data for DEET, KBR3023, IR3535, and PMD illustrated that the proportion of mosquitoes repelled on skin vs. membrane differed significantly from each other beginning at either the 31 nmol/cm 2 (Fig. 4 6), the 61 nmol/cm 2 (Fig. 4 7) or at the 123 nmo l/cm 2 (Fig. 4 8, Fig. 4 9) dose level, according to a test of means with a confidence level of p=0.05 Probit linked dose response curves were produced in PoloPlus for visual comparison of the paired treatments for each of the chemicals tested (Fig. 4 10, Fig. 4 11, Fig. 4 12, Fig. 4 13). These illustrated the sigmoidal curve expected and provided visual cues as to the precision of the six data points at each dose level. Discussion T wo sources of kairomones, human skin and silicone membranes tr eated with human odors, were used to evaluate the blood feeding behavior of female mosquitoes when exposed to four repellent chemicals in a laboratory setting. Since the baseline attraction was found to be statistically the same for both of these attractan t sources the attractiveness of the membrane to mosquitoes supports the hypothesis that this could be a good surrogate for humans in laboratory tests. The membranes were worn on the skin of the volunteers for 3 4 h prior to testing in an

PAGE 41

41 effort to transfe r some of the attractive skin chemicals on the membranes. This appears to have been successful, since the attractiveness of the membrane tested without human odors using both blood and a 10% sugar solution, did differ significantly from the control treatme nts on human skin. The membrane served as a suitable alternative for the skin in the testing of all four repellents at lower concentrations, since the repellent results tested on skin and those tested on the membrane did not differ significantly below the 31 nmol dose level. Above this level, DEET differed significantly beginning at 31 nmol/cm 2 (Fig. 4 6), KBR3023 differed at 61 nmol/cm 2 (Fig. 4 7), and both IR3535 and PMD differed at 123 nmol/cm 2 (Fig. 4 8, Fig. 4 9). This indicates that near the point whe re half of the mosquitoes are deterred from feeding, the two models diverge and above these doses the membrane fails to provide a good estimate of the repellent dose required on the skin. Interestingly, all four repellents tested on the membrane required much higher doses of repellent chemicals to reach full protection from mosquito blood feeding, i.e. 100% of the mosquitoes deterred from feeding. The mean membrane dose level f or full protection was 2 3 dose levels higher or 4 6 times as much repellent th an was needed for the skin. Given that the chemical doses were doubled at each interval, much more of the repellent chemical was necessary to achieve comparable feeding deterrence in the membrane vs. the skin. This is interesting because it is unclear whet her the membrane became more desirable to the mosquitoes than the skin at these doses or if perhaps the skin became less desirable to the mosquitoes. Natural repellents as well as attraction inhibitors have been documented in the exudate of humans (Bernier et al. 2002, 2005, 2007b). These natural allomones have also been found to vary from individual to individual. Perhaps the mixture of these naturally repellent or inhibiting

PAGE 42

42 chemicals produced by the skin when mixed with the higher doses of repellents app lied during the laboratory test, reduced the total amount of repellent needed to prevent blood feeding by the female mosquitoes. Although, the membranes worn on the skin of the volunteers appears to have successfully transferred attractive skin chemicals o nto the membranes, it is unclear whether any of these allomonal compounds were transferred or if they are produced at higher rates when humans undergo stress, i.e. chemicals for feeding dete rrence by insects. The human skin chemical interactions with the repellent compounds were not explicitly explored in this study but could be further examined in additional studies with the membrane module system. By applying a standardized human derived complement of chemicals to the membrane along with repellent treatments, the membrane system could be examined to see if it produces results more similar to the skin at dose applications higher than 31 nmol/cm 2 Preliminary data on the individual voluntee rs showed evidence of an increase in feeding behavior that was inconsistent with the increased application of repellent chemical. This occurred mostly at the dose level just prior to the statistically significant divergence of the skin and membrane curves from each other. However, this effect was reduced by the pooling of the data from all volunteers, but is still evident on visual examination of the curves and line graphs. These unexpected results may be due to potentially attractive properties of repellen t chemicals at very low doses, which have been documented with DEET (Mehr et al. 1990, Dogan and Rossignol 1999, Bernier et al. 2005). Further studies with lower doses of KBR3023, IR3535, and PMD should be conducted to test whether this phenomenon is occur ring with these compounds as well.

PAGE 43

43 There are several benefits of using an in vitro testing method in the laboratory instead of performing testing directly on human volunteers, which are not limited to exposing humans to less risk by not subjecting them to being bitten by mosquitoes, testing compounds of unknown toxicity, screening many successive compounds quickly, not needing to schedule human volunteers for testing or even needing to register volunteers for an IRB study. This last consideration encompasse s significant savings in time and cost for scientists that wish to conduct repellent testing However, since the chemical dose at which the skin and membrane curves diverge is lower than standard thresholds used in screening these chemicals, the use of the membrane module system would require some further modifications before it can be fully utilized as a replacement for screening methods using human volunteers.

PAGE 44

44 Table 4 1 Chemical structures, names and other properties of the four repellent treatments. Repellent Name Structure MW (g/mol) CAS # DEET N,N diethyl 3 methylbenzamide 191.27 134 62 3 KBR3023 2 (2 hydroxyethyl) 1 methylpropylester 229.32 119515 38 7 IR3535 ethyl 3 [acetyl(butyl)amino] propanoate 215.29 52304 36 6 PMD p menthane 3,8 di ol 172.26 42822 86 6 Table 4 2. Results of F Tests in SAS 9.2 for the fitted model using pooled data from six subjects examining the effects of the chemicals, the testing media, the dose of repellent as well as their interactions. Effect Numerator DF Denominator DF F Value Pr > F Nchem 3 71 6.48 0.0006 Medium 1 71 83.23 <.0001 Nchem*Medium 3 71 5.30 0.0023 logDose 1 71 646.71 <.0001 logDose*Nchem 3 71 7.05 0.0003 logDose*Medium 1 71 101.93 <.0001 logDose*Nchem*Medium 3 71 5.91 0.0012

PAGE 45

45 Figure 4 1. Pictures of in vitro loading and feeding modules. a) M osquito loading module with six testing chambers corked and containing mosquitoes. b) M o squito feeding module with 6 well receptacles for blood feeding Photos courtesy of Natasha M. Agramon te Figure 4 2. Assembled mosquito loading and feeding modules with arrow indicating the CO 2 tubing connected in parallel to the front of the six testing chambers via the cork hole. Photo courtesy of Natasha M. Agramonte Figure 4 3. C onfirmation of m osquito blood feeding from module tests. a) E vidence from module test indicating 0 mosquitoes feeding. b) E vidence from module test indicating 1 0 mosquitoes feeding. Photos courtesy of Natasha M. Agramonte

PAGE 46

46 Figure 4 4. Assembled mosquito loading and fe eding modules connected to heated water circulator via tubing for in vitro bioassay, showing rubber elastic bands placed around each of the six testing chambers. Photo courtesy of Natasha M. Agramonte Figure 4 5 Stacked column chart comparing the me an attraction for the volunteers individually and pooled comparing in vivo vs. in vitro mosquito feeding on the control with asterisk indicating where controls differed from each other significantly (p

PAGE 47

47 Figure 4 6 Line graph of dose response data indicating the proportion of mosquitoes repelled by DEET on skin ( in vivo ) and on membrane ( in vitro ) with asterisk indicating where Figure 4 7 Line graph of dose response data indicating the proportion of mosquitoes repelled by KBR3023 on skin ( in vivo ) and on membrane ( in vitro ) with asterisk indicating

PAGE 48

48 Figure 4 8 Line graph of dose response data indicating the proportion of mosquitoes repelled by IR3535 on skin ( in vivo ) and on membrane ( in vitro ) with asterisk indicating Figure 4 9 Line graph of dose response data indicating the proportion of mosquitoes repelled by PMD on skin ( in vivo ) and on membrane ( in vitro ) with asterisk indicating where

PAGE 49

49 Figure 4 10 Dose response curve s produced by PoloPlus displaying the percentage of mosquitoes repelled by DEET on skin (left) and on the membrane (right) over a range of chemical doses on a log scale Figure 4 1 1 Dose response curves produced by PoloPlus displaying the percentage of mosquitoes repelled by KBR3023 on skin (left) and on the membrane (right) over a range of chemical doses on a log scale.

PAGE 50

50 Figure 4 1 2 Dose response curves produced by PoloPlus displaying the percentage of mosquitoes repelled by I R3535 on skin (left) and on the membrane (right) over a range of chemical doses on a log scale. Figure 4 1 3 Dose response curves produced by PoloPlus displaying the percentage of mosquitoes repelled by PMD on skin (left) and on the membrane (right) over a range of chemical doses on a log scale.

PAGE 51

51 CHAPTER 5 ASSESSMENT OF THREE MEDIA FOR THE ESTIMA TION OF EFFECTIVE DO SE IN MOSQUITO REPELLENT B IOASSAYS Background An ideal repellent screening test method sh ould provide a fast, safe, inexpensive way to te st chemical compounds, whether toxicity has been established or not, and without requiring the use of human volunteers. Currently, no such testing method exists but advances in mosquito repellent testing have allowed for indirect methods of testing mosquit o repellents. Alternative testing strategies have included testing on cloth, membranes, animals of other species, as well as the use of complex human surrogate systems. While many clever testing methods have been fashioned, many of these novel methods have not undergone comparative analysis with skin or with other standardized methods of repellent screening to examine how well the results using the alternative methods accurately predict how well the repellent will perform on the skin. One method used for th e rapid screening of repellent compounds of unknown toxicity is the cloth patch bioassay ( Carroll et al. 2011, Katritzky et al. 2008, 2010, Rosa et al. 2012 ) Although the cloth patch bioassay allow s for the presence of human skin volatiles to affect the b ehavior of the mosquitoes in combination with repellent treatments, complex chemical interactions are not fully explored due to the protective barriers used in these tests. Because of this, potentially synergistic interactions of repellent chemicals with k airomonal skin emanations such as CO 2 lactic acid, acetone, and dimethyl disulfide cannot be explored until the chemicals are deemed safe. Several areas of arboviral research would benefit f rom a well defined correlation of indirect and direct methodologi es particularly in the testing of virus infected mosquitoes. Since it is not ethical to perform assays with virus infected mosquitoes on human volunteers using direct testing methods a well developed indirect met hod would be beneficial, especially one th at

PAGE 52

52 precludes the use of human volunteers. Testing methods that preclude the use of human volunteers would also expand the ability of scientists to test mosquito repellents in their laboratories without the hassle of first obtaining approval from a human us e review board Materials and Methods Mosquito Rearing and Selection The mosquitoes used in all bioassays were female Ae. aegypti (Orlando strain, 1952) from the colony maintained at the Center for Medical, Agricultural, and Veterinary Entomology location of the United States Department of Agriculture, Agricultural Research Service (USDA ARS CMAVE) in Gainesville, FL. Pupae were obtained from the colony and kept in laboratory cages where newly emerged mosquitoes were maintained ad libitum on a 10% sucrose s olution at 25 28 o C ambient temperature, 60 80% relative humidity and a 14:10 (light:dark) photoperiod. Nulliparous female mosquitoes aged six to eleven days displaying host seeking behavior were pre selected from stock cages using a hand draw box and trapp ed in a collection trap (Posey et al. 1981). For use in the module bioassays on skin and membrane, female mosquitoes were then transferred to a smaller cage (28,316 cm 3 ) from which the mosquitoes were sorted into groups of 10 by mechanically aspirating the m into acrylic holding tubes (15 cm long, 1.25 cm wide). Each tube contained approximately 10 mosquitoes, contained by screen gauze on one end and stoppered with a small cork (Fisher Scientific, Catalog No. 07781D Size 1) at the other end. Mosquitoes held in the tubes were allowed to acclimatize for 15 20 min (Barnard et al. 2007) before being transferred into an empty cha mber of the module for testing. For use in the cage tests on cloth, 500 ( 10%) females were preselected and collected in the trap and tr ansferred to a test cage (approximately 59,000 cm 3 with dimensions 45 x 37.5 x 35 cm) and allowed to acclimatize for 15 20 min prior to testing (Barnard et al. 2007).

PAGE 53

53 Repellent Chemical Treatments and Control Chemical compounds used as repellent treatments included technical N,N diethyl 3 methylbenzamide 97% (Aldrich, CAS#134 62 1 piperidinecarboxylic acid 2 (2 hydroxyethyl) 1 methylpropylester, 98% (CAS#119515 38 7) al ethyl 3 [acetyl(butyl)amino]propanoate 98% (CAS#52304 36 menthane 3,8 diol, 98% (Bedoukian, CAS# 42822 86 chemicals were obtained by adding 1mL of completely denatured ethyl alcohol (Mallinckrodt) into 1mL of the previous concentration of chemical to produce a range of concentrations (7840 nmol/cm 2 3920 nmol/cm 2 1960 nmol/cm 2 980 nmol/cm 2 490 nmol/cm 2 245 nmol/cm 2 123 nmol/cm 2 61 nmol/cm 2 31 nmol/cm 2 15 nmol/cm 2 8 nmol/cm 2 ) when applied to the 14.19 cm 2 treatment area. The control treat ment consisted of only ethanol. Skin Bioassays for the Estimation of Effective Dose In the first portion of this study, the proportion of Ae aegypti mosquitoes blood feeding on the skin of six human volunteers (4 male, 2 female) was evaluated for a range of concentrations to the repellent chemicals DEET, IR3535, KBR3023, and PMD. Ten nulliparous female Ae. aegypti mosquitoes ( aged 6 11d) were loaded into each of the six testing chambers (5 cm x 5 cm x 5 cm) in the previously described loading module (Klun and Debboun 2000, Klun et al. 2005, Weldon et al. 2003, Rutledge and Gupta 2004). Each of the six chambers in the loading module had two ape rtures, a 4.25 cm diameter circular aperture at the bottom of each chamber, covered by a sliding door, as well as a small 1.5 cm diameter circular aperture on the front of each chamber closed by a small cork (Fig. 5 1a) For the skin test, the loading modu le was placed over the thighs of each human volunteer On the thighs, areas for treatment were demarcated with 4.25 cm diameter circles corresponding

PAGE 54

54 diluted dose of DEET, IR3535, KBR3023, or PMD ( 7840 nmol/cm 2 3920 nm ol/cm 2 1960 nmol/cm 2 980 nmol/cm 2 490 nmol/cm 2 245 nmol/cm 2 123 nmol/cm 2 61 nmol/cm 2 31 nmol/cm 2 15 nmol/cm 2 8 nmol/cm 2 ) was then applied to the six circles in successive layered doses from the lowest concentration to the highest concentration. Th e first treatment applied in any set of tests was the ethanol control to establish the baseline for mosquito feeding behavior. The dose was allowed to dry for 3 5 min to allow for the ethanol to thigh with a glass spacer placed between the module and the skin to prevent direct contact of the module with the chemicals. A sliding door was opened under three chambers at a time to expose mosquitoes in those chambers to the repellent for a 3 min perio d. At the end of the exposure period, the mosquitoes were knocked down with CO 2 gas via the corked hole (Fig. 5 2) removed from the chamber with an aspirator and crushed to record the proportion of mosquitoes blood feeding out of ten (Fig. 5 3) This proc ess was repeated with the rest of the six chambers and with all concentrations of the repellent until all feeding in all chambers ceased. Volunteers for all repellent tests signed informed consent forms and were enrolled in an IRB study (Project # 636 2005 ). Membrane Bioassays for Estimation of Effective Dose In the second portion of this study which examined the membrane medium, the loading module was placed over the feeding module with six 4.25 cm diameter wells, corresponding to the six circular openings in the loading chamber (Klun and Debboun 2000, Klun et al. 2005, Weldon et al. 2003, Rutledge and Gupta 2004). Each of the wells in the feeding module were filled with 7 mL of citrated bovine blood (Fig. 5 1b) continuously p umping hot water through the feeding chamber with a Cole Parmer Polystat circulating water bath (Fig. 5 4) Prior to testing, strips of silicone membranes were worn against the upper thigh by the volunteers for 3 4 h, held in place with an Ace elastic ban dage.

PAGE 55

55 The worn silicone membranes were placed across the six wells of the feeding module to come into contact with the blood, and the glass spacer was placed over the membranes leaving only the membrane diluted do se of DEET, IR3535, KBR3023, or PMD ( 7840 nmol/cm 2 3920 nmol/cm 2 1960 nmol/cm 2 980 nmol/cm 2 490 nmol/cm 2 245 nmol/cm 2 123 nmol/cm 2 61 nmol/cm 2 31 nmol/cm 2 15 nmol/cm 2 8 nmol/cm 2 ) was then applied to the six circles on the silicone membrane above each well in successive layered doses from the lowest concentration to the highest. The first treatment applied in any set of tests was the ethanol control to establish the baseline for mosquito feeding behavior. The dose was allowed to dry for 3 5 min to allow for the ethanol to evaporate. The loading module was placed onto the feeding module and lined up to correspond to each well. A sliding door was opened under all six chambers to expose the mosquitoes to the repellent treatment for a 3 min period. At t he end of the exposure period, the mosquitoes were knocked down with CO 2 gas via the corked hole, removed from the chamber with an aspirator and crushed to record the proportion of mosquitoes blood feeding out of ten. This process was repeated with the res t of the six chambers and with all concentrations of the repellent until all feeding in all chambers ceased. Cloth Bioassays for Estimation of Minimum Effective Dose The third portion of this study examined the estimation of minimum effective dose for repe llency on cloth. In preparation for these tests 2 dram screw top glass vials containing the 1mL ethanol diluted dose s of DEET, IR3535, KBR3023, or PMD ( 7840 nmol/cm 2 3920 nmol/cm 2 1960 nmol/cm 2 980 nmol/cm 2 490 nmol/cm 2 245 nmol/cm 2 123 nmol/cm 2 6 1 nmol/cm 2 31 nmol/cm 2 15 nmol/cm 2 8 nmol/cm 2 ) were arranged and a 50 cm 2 (5 cm x 10 cm) piece of muslin cloth was rolled up and placed inside each vial to soak up the chemical treatment. Just prior to the experiment, the pieces of treated cloth we re re moved from the vials and card stock tabs (5 cm x 3 cm) were stapled onto them lengthwise. These cloth patches wer e

PAGE 56

56 hung on a drying rack using masking tape for 5 min to allow the ethanol to volatilize off, leaving only the chemical treatment on the dry clo th Volunteers in the study used latex gloves to pull a n ylon stocking over their arm and a Velcro sealed vinyl sleeve was then placed over the ir forearm (Fig. 5 5) The sleeve had a 32 cm 2 (4 cm x 8cm) window to allow attractive skin odors to dra w mosqui toes to that open area. The purpose of the nylon stocking was to produce a barrier between the dried cloth and the skin, thereby avoiding dire ct contact of chemical to skin. Gloves were worn to protect the sensitive hands of the volunteers which would be b itten by the mosquitoes if not covered (Fig. 5 6). The dried cloth patch was then attached with masking tape over the opening in the sleeve Participants then inserted their arm with the sleeve and patch into a screened cage that contained 500 female Ae. a egypti mosquitoes (Fig. 5 7) Tests were conducted on each control or treated patch for 1 min periods A control patch treated only with ethanol was tested prior to the start of experiments (Fig. 5 8a) When testing a treated patch, if approximately 1% or 5 mosquito bites were received within 1 min, th e chemical dose was considered to have failed, i.e. was not repellent at that concentration (Fig. 5 8b) If a treated cloth patch received 0 4 bites within a minute, then it was considered as passed, i.e., re pellent at that concentration. A median concentration treated patch was tested in the first round and treated patches were then tested successively at higher or lower concentrations depending upon whether the previous patch failed or passed, respectively. The estimate of the minimum effective dose, or E D 99 was the lowest concentration that passed for each repellent chemical. This process of estimating an ED 99 was replicated five times f or each volunteer in the study. Statistical Analysis Pooled data from re pellency analysis was used to estimate the various effective doses for each treatment for the three media examined. Effective doses at 50, 95, and 99 were estimated

PAGE 57

57 for the skin and membrane bioassays using a dose response curve model with a probit link. T he equation used for the model is: probit(y/n) = ln (Dose_nmol +100) where y is the pooled number of mosquitoes repelled at each treatment group across all replicates and all volunteers, n is the pooled total number of mosquitoes exposed at each treatment group across all replicates and all volunteers (Table 5 1), and Dose_nmol is the concentration of repellent chemical to which mosquitoes were exposed at each concentration interval. Estimations of the effective dose at 50, 95, and 99 (ED 50 ED 95 and ED 99 ) were made using SAS 9.2 (SAS 2012). Estimation of 95% confidence intervals (CI 95) for each treatment and effective dose from the skin and membrane bioassays were calculated using trial and error in SAS 9.2. The significance at a confidence of p=0.05 for the model term effects, as well as their interactions were evaluated using an approximated F test. PoloPlus v2.0 was also used to produce dose response curves for secondary confirmation of effective dose estimations at 50, 95, and 99 levels for the skin an d membrane media (PoloPlus 2012). Pooled data was used to estimate the MED on the cloth medium for each repellent compound tested. Effective doses at 50 and 95 could not be determined for this test since only the point estimate at ED 99 was estimated. The E D 99 values for each repellent chemical were then averaged for the five replicates and then across the six study volunteers, and reported as the mean ED 99 The 95% confidence intervals were estimated using twice the value of the standard deviation the mea n ED 99 Results Mean values for the ED 50 95 (nmol/cm 2 ) for the four repellent chemicals tested were estimated for the skin and membrane media using data pooled from all six volunteers (Table 5 2, 5 3). Ratios comparing the ED 50 and ED 95 results o n membrane to the skin for the

PAGE 58

58 four repellent chemicals were also included (Table 5 2, 5 3). Mean values for the ED 99 (nmol/cm 2 ) for the four repellent chemicals tested were estimated for the skin, membrane and cloth media using pooled data as well (Tab le 5 4). Ratios comparing the ED 50 and ED 95 results from membrane and cloth to the skin as well as from membrane and skin to the cloth for the four repellent chemicals were also included (Table 5 4). Secondary confirmation of the estimated ED 50 ED 95 and ED 99 values on skin and membrane were supported by estimates provided by PoloPlus using the dose response curves data described in the previous chapter. Bar graphs were used to illustrate the differences in the mean ED 99 nt treatments on skin, membrane and cloth (Figure 5 9). The ED 99 results for the skin and cloth were more similar than either skin or cloth to the membrane, and these fine differences were highlighted in a second bar graph which excluded the membrane value s (Figure 5 10). Discussion The objective of this portion of the study was to compare the skin and membrane bioassay results with the results of a standard method to assess the usefulness of this alternative method in estimating repellent dose. This study allowed for a comparison of three methods using different media to estimate minimum effective dose (ED 99 ). Only two media, the skin and the membrane could be compared at the median effective dose (ED 50 ) and the 95% effective dose (ED 95 ). Effective doses a t 50 and 95 could not be determined for the cloth patch bioassay since this type of assay is used only to estimate the ED 99 dose These three dose levels are those that are most commonly reported in repellency and toxicity studies, although toxicity studie s report lethal dose levels (LD). A comparison of the ED 99 s that the minimum effective doses estimated on membrane were much higher than those of the same chemical repellents applied to skin or cloth.

PAGE 59

59 There are potentially for two reasons for this discrepancy. One reason that the results of the membrane ED 99 are not comparable to the cloth results, as was alluded to in the previous chapter previously since the membrane results were statistically different from the results on skin possibly above at least 31 nmol/cm 2 It is possible that allomonal chemicals produced by the skin adversely effected the concentration of repellent needed to achieve full protection, both in the skin as well as in the cloth bioassays. While the bioassays using cloth did not have repellent directly applied to the skin surface, the skin of the volunteers was available to the mosquitoes to bite, as well as to provide odors for attraction or possibly repulsion via allomonal chemicals. The presence of these human skin volatil es could have affected the behavior of the mosquitoes in combination with repellent treatments. A second reason for the discrepancy may lie in the validity of the ED 99 estimation. The ED 99 results for the membrane contain a high level of inaccuracy due to estimation of a point in an area of the dose response curve where small changes in response (repellency) occur over a relatively larger range of concentrations. According to Robertson et al., the sample size necessary for an accurate estimation of the ED 9 5 from a dose response curve require s a sample size of 600 1000 to determine significance (1984, 2007). This means that for the estimation of an ED 99 more than 1000 samples, likely between 2000 5000 samples would be necessary for accurate estimation of an ED 99 using a dose response cu rve. This sacrifice in accuracy is accepted since the number of mosquitoes and replicates required to produce an estimate with a high level of confidence was beyond the time and mosquito resources available for this study. For an accurate estimation of ED 50 however, a sample size of only 300 500 is necessary to determine significance (Robertson et al. 1984, 2007), which was easily achieved in all the bioassay runs since the average number of mosquitoes tested per volunteer per treatment was

PAGE 60

60 between 400 700 (Table 5 1). For this reason, the ED 50 estimates on the skin and membrane are more likely to predict the median effective dose and the ED 99 is likely a less precise indicator since these were estimated from dose response curv es, which is further supported by the wide confidence intervals calculated for both skin and membrane bioassays at the ED 95 and ED 99 levels (Table 5 3, 5 4) While the estimated ED 99 the membrane results, the wider confidence intervals for the skin indicate a lower degree of precision when tested by this method. Thus, the cloth module bioassay appears to produce the best estimation of ED 99 of these three bioassays for mosquito repellent sc reening. Interestingly, the ED 99 values estimated on cloth in this study were roughly two times higher than estimates for these same four repellent chemicals tested by the same method, but using acetone as a solvent (Agramonte and Bernier, unpublished data ). The density of mosquitoes within each testing chamber for all three bioassays was 0.08 mosquitoes per cm 3 or 1 mosquito per 12.5 cm 3 for the modules tests using skin and membrane, and 0.008 mosquitoes per cm 3 or 1 mosquito per 118 cm 3 for the cage tes ts using cloth (Table 5 5). This means that the mosquito density for module bioassays is approximately 10 times greater than the density of mosquitoes used in the cage tests with cloth. Further testing should examine ED 99 ch as those in the cloth patch bioassay, to rule out the effect of higher mosquito density or limited ventilation in the module.

PAGE 61

61 Table 5 1. Total number of mosquito es bioassayed per treatment for six human volunteers. Treatment 1 Total Mosquitoes Tested 2 (N) DEET skin 2570 KBR3023 skin 2574 IR3535 skin 3298 PMD skin 2800 DEET membrane 3900 KBR3023 membrane 3840 IR3535 membrane 4018 PMD membrane 3714 DEET cloth 2500 3 KBR3023 cloth 2500 3 IR3535 cloth 2500 3 PMD cloth 2500 3 1 Treatm ent defined as repellent chemical applied and medium on which it was applied 2 Average number of mosquitoes tested per volunteer per treatment is ~400 700 3 Number of mosquitoes per test cage in cloth patch bioassays was 500 (25) Table 5 2. Mean ED 50 ( 95% CI) and ED 50 ratios for four chemicals estimated using repellency bioassay data (nmol/cm 2 ) pooled from six human volunteers on membrane and skin. Medium DEET KBR3023 IR3535 PMD Skin 11.55 (1.45 20.28) 4.05 ( 14.20 20.46) 17.04 (4.41 28.74) 20.28 (8.53 31.14) Membrane 43.05 (17.68 69.02) 35.07 (7.39 63.37) 60.72 (30.26 92.87) 23.73 ( 3.91 51.41) Membrane: Skin Ratio 3.73 8.66 3.56 1.17 Table 5 3. Mean ED 95 (95% CI) and ED 95 ratios for four chemicals estimated using repellency bioassay data (nmol/ cm 2 ) pooled from six human volunteers on membrane and skin. Medium DEET KBR3023 IR3535 PMD Skin 218.14 (169.89 307.89) 751.64 (519.55 1278.84) 385.25 (299.41 537.78) 341.19 (261.04 491.70) Membrane 2480.87 (1681.12 4185.53) 3448.89 (2182.44 6520.99 ) 4190.35 (2721.43 7477.97) 3498.19 (2137.24 7050.95) Membrane: Skin Ratio 11.37 4.59 10.88 10.25

PAGE 62

62 Table 5 4. Mean ED 99 (95% CI) and ED 99 ratios to skin and cloth for four chemicals estimated using repellency bioassay data (nmol/cm 2 ) pooled from s ix human volunteers on membrane, skin and cloth. Medium DEET KBR3023 IR3535 PMD Skin 391.13 (287.61 606.27) 1934.65 (1174.11 4075.54) 774.74 (561.82 1202.45) 655.94 (467.93 1052.86) Membrane 8456.34 (4914.05 17754.31) 13648.87 (7159.02 34100.65) 16 629.62 (9082.00 38077.44) 14436.30 (7159.02 40034.84) Cloth 239.00 (173.97 304.03) 258.67 ( 211.24 441.28) 1494.67 (130.81 3865.76) 381.33 (38.13 567.88) Membrane: Skin Ratio 21.63 7.06 21.48 22.04 Membrane: Cloth Ratio 35.38 52.90 11.13 37.89 Skin : Cloth Ratio 1.64 7.50 0.52 1.72 Table 5 5 Densities calculated for all three bioassay types using the internal volume of each test chamber or cage (cm 3 ) and the number of mosquitoes used for each test (N). Bioassay Type N (mosquitoes) Volume (cm 3 ) Density (N/cm 3 ) Module Test on Skin 10 125 0.080 Module Test on Membrane 10 125 0.080 Cage Test on Cloth 500 59000 0.008

PAGE 63

63 Figure 5 1. Picture s of in vitro loading and feeding modules. a) M osquito loading module with six testing chambers corked and containing mosquitoes. b) M osquito feeding module with 6 well receptacles for blood feeding. Photos courtesy of Natasha M. Agramonte. Figure 5 2. Assembled mosquito loading and feeding modules with arrow indicating the CO 2 tubing connected in paralle l to the front of the six testing chambers via the cork hole. Photo courtesy of Natasha M. Agramonte. Figure 5 3. C onfirmation of mosquito blood feeding from module tests. a) E vidence from module test indicating 0 mosquitoes feeding. b) E vidence from mod ule test indicating 10 mosquitoes feeding. Photos courtesy of Natasha M. Agramonte.

PAGE 64

64 Figure 5 4. Assembled mosquito loading and feeding modules connected to heated water circulator via tubing for in vitro bioassay, showing rubber elastic bands placed ar ound each of the six testing chambers. Photo courtesy of Natasha M. Agramonte. Figure 5 5 Materials used for the c loth patch bioassay. Photo courtesy of Greg Allen.

PAGE 65

65 Figure 5 6 Ae. aegypti feeding behavior on hands without (left) and with (right) repellent DEET. Photo courtesy of Greg Allen. Figure 5 7 Example of c loth patch bioassay in progress Photo courtesy of Greg Allen.

PAGE 66

66 Figure 5 8 Cloth patch test showing confirmation of mosquito blood feeding. a) C loth patch with 0 mosquitoes blo od feeding. b) C loth patch with > 5 mosquitoes blood feeding. Photos courtesy of Greg Allen. Figure 5 9 Bar graph of the mean ED 99 (nmol/cm 2 ) estimated for each of the four repellent treatments evaluated on three media. a b

PAGE 67

67 Figure 5 1 0 Bar graph h ighlighting the mean ED 99 (nmol/cm 2 ) estimates on skin and cloth media for the four repellent treatments rescaled for direct comparison.

PAGE 68

68 CHAPTER 6 SUMMARY OF FINDINGS AND IMPLICA TIONS FOR FUTURE RES EARCH The purpose of this research was to examine a silic one membrane blood feeding system and its effectiveness as an alternative to application of repellents directly on human skin. A f e w standardized in vivo repellency assay method s exist which have directly utilized humans in testing repellents on the skin; how ever these had not been shown to be sufficiently comparable to in vitro repellency assay methods. The goal of this project was to examine the results obtained from an in vitro assay method to determine the level at which results from this method compare to in vivo testing results in an effort to preclude the use of human volunteers in these kinds of studies Based on literature on mosquito host seeking behavior, it was logical to hypothesize that use of a silicone membrane treated with human odors and place d over a warmed blood source would produce similar repellent responses in the Aedes aegypti mosquito when compared to human skin The efforts to transfer some of the attractive skin chemicals on to the membranes appear ed to have been successful, since the at tractiveness of the membrane tested without human odors using both blood and a 10% sugar solution, did differ significantly from the tests on human skin and from membranes treated with attractive skin chemicals. Additionally the baseline attraction for the membrane s worn by the volunteers was found not to d iffer statistically from the attractiveness for tests performed on the skin of these same volunteers. This project also sought to create dose response curves for four major repellent chemicals and examine if these could simulate the results of testing on skin. A comparison of the ED 99 indicated that the minimum effective doses estimated on membrane were much higher than those of the same chemical applied to skin or cloth. Based on the comparative assessment of thre e repellent screening media, namely cloth, membrane and skin, the cloth bioassay appears to

PAGE 69

69 produce the best estimation of the results on skin as well as the most accurate method for estimation of ED 99 values from these bioassays for mosquito repellent scr eening. All four repellents tested on the membrane required much higher doses of repellent chemicals to reach full protection from mosquito blood feeding, i.e. 100% of the mosquitoes deterred from feeding. This is interesting because it is unclear whether the membrane became more desirable to the mosquitoes than the skin at these doses or if perhaps the skin became less desirable to the mosquitoes. If these are the result of allomonal compounds produced from human skin it was unclear whether any of these al lomonal compounds were transferred on the membranes or if they are produced at higher rates when humans undergo stress, i.e. being bitten Also, early data fr om the individual volunteers showed evidence of an increase in feeding behavior that was inconsistent with the increased application of repellent chemical. This occurred mostly at the dose level just prior to the statistically significant divergence of the skin and membrane curves from each other. These unexpected results may be due to potentially attractive properties of repellent chemicals at very low doses. A third question of interest was whether the solvent the repellent in dissolved in plays a role in the effect of a repellent, since the ED 99 values estimated on cloth in this study were roughly two times higher than estimates for these same four repellent chemicals tested by the same method, but using acetone as a solvent. The human skin chemical int eraction s with the repellent compounds could be further examined in additional studies with the membrane module system by applying a standard human derived chemical treatment to the membrane along with repellents to examine if it produces results more simi lar to the skin at dose applications higher than 31 nmol/cm 2 Further studies with lower doses of KBR3023, IR3535, and PMD should be conducted to test whether

PAGE 70

70 these compounds are also attractive to mosquitoes at low concentrations. Finally, further testing should examine ED 99 rule out the effect of higher mosquito density or limited ventilation in the module and the effect of solvent on the minimum effective dose of a repellent ch emical. A dose response study could be undertaken with the cloth patch assay to better establish how well this method compares to in vivo module assays. Finally, a dose response study could be conducted in vivo with direct topical application of these comp ounds since all four repellents are registered for use on human skin.

PAGE 71

71 LIST OF REFERENCES Ahmadi, A., and G. McClelland. 1985. Mosquito mediated attraction of female mosquitoes to a host. Physiol. Entomol. 10: 251 255. American Society for Testing and Mat erials [ASTM] 2000a. ASTM E939 94, Standard test method of field testing topical applications of compounds as repellents for m edically important and pest arthropods (including insects, ticks, and mites): 1 m osquitoes ASTM International, West Conshohocke n, PA, 2003, DOI: 10.1520/E0939 94R12, www.astm.org American Society for Testing and Materials [ASTM] 2000b. ASTM E951 94, Standard test methods for laboratory testing of non commercial mosquito repellent formulation s on the skin. ASTM International, West Conshohocken, PA, 2003, DOI: 10.1520/E0951 94R06, www.astm.org Barnard, D. R. 2005. Biological assay methods for mosquito repellents. J. Am. Mosq. Control Assoc. 21: S12 S16. Barn ard, D. R., U. R. Bernier, R. Xue, and M. Debboun. 2007. Standard methods for testing mosquito repellents, pp. 103 110. In M. Debboun, S. P. Frances and D. Strickman (eds.), Insect repellents: principles, methods, and uses. CRC Press, Boca Raton, FL. Barna rd, D. R., and R. Xue. 2004. Laboratory evaluation of mosquito repellents against Aedes albopictus Culex nigripalpus and Ochlerotatus triseriatus (Diptera: Culicidae). J. Med. Entomol. 41: 726 730. Bernier, U. R., K. D. Furman, D. L. Kline, S. A. Allan, and D. R. Barnard. 2005. Comparison of contact and spatial repellency of catnip oil and N,N diethyl 3 methylbenzamide (DEET) against mosquitoes. J. Med. Entomol. 42: 306 311. Bernier, U., D. Kline, S. Allan, and D. Barnard. 2007a. Laboratory comparison of Aedes aegypti attraction to human odors and to synthetic human odor compounds and blends. J. Am. Mosquito Contr. 23: 288 293. Bernier, U. R., D. L. Kline, D. R. Barnard, C. E. Schreck, and R. A. Yost. 2000. Analysis of human skin emanations by gas chromato graphy/mass spectrometry. 2. Identification of volatile compounds that are candidate attractants for the Yellow Fever mosquito ( Aedes aegypti ). Anal. Chem. 7 2: 747 756. Bernier, U. R., D. L. Kline, and K. H. Posey. 2007b. Human emanations and related natur al compounds that inhibit mosquito host finding abilities, pp. 77 102. In M. Debboun, S. P. Frances and D. Strickman (eds.), Insect repellents: principles, methods, and uses. CRC Press, Boca Raton, FL. Bernier, U., D. Kline, K. Posey, M. Booth, R. Yost, an d D. Barnard. 2003. Synergistic attraction of Aedes aegypti (L.) to binary blends of L lactic acid and acetone, dichloromethane, or dimethyl disulfide. J. Med. Entomol. 40: 653 656.

PAGE 72

72 Bernier, U., D. Kline, C. Schreck, R. Yost, and D. Barnard. 2002. Chemical analysis of human skin emanations: comparison of volatiles from humans that differ in attraction of Aedes aegypti (Diptera: Culicidae). J. A m. Mosquito Contr. 18: 186 195. Bernier, U. R., and M. Tsikolia. 2011. Development of novel repellents using struct ure activity modeling of compounds in the USDA archival database, pp. 21 46. In J. R. Coats, and G. E. Paluch (eds.), Recent developments in invertebrate repellents. American Chemical Society: Washington, D.C. Bowen, M. F. 1991. The sensory physiology of h ost seeking behavior in mosquitoes. Annu. Rev. Entomol. 36: 139 158. Braks, M., C. Koenraadt, O. Briet, and W. Takken. 1999. Behavioral responses of Culex quinquefasciatus (Diptera: Culicidae) to human odour, carbon dioxide and lactic Acid. Exp. Appl. Ento mol. 10: 91 97. Braks, M., J. Meijerink, and W. Takken. 2001. The response of the malaria mosquito, Anopheles gambiae to two components of human sweat, ammonia and L lactic Acid, in an olfactometer. Physiol. Entomol. 26: 142 148. Braks, M., E. Scholte, W. Takken, and T. Dekker. 2000. Microbial growth enhances the attractiveness of human sweat for the malaria mosquito, Anopheles gambiae sensu stricto (Diptera: Culicidae). Chemoecology. 10: 129 134. Braks, M., and W. Takken. 1999. Incubated human sweat but n ot fresh sweat attracts the malaria mosquito Anopheles gambiae sensu stricto J. Chem. Ecol. 25: 663 672. Brett, G. A. 1938. On the relative attractiveness to Aedes aegypti of certain coloured cloths. T. Roy. Soc. Trop. Med. H. 32: 113 124. Brown, A. 1966. The attraction of mosquitoes to hosts. J. Am. Med. Assoc. 196: 159 162. Cantrell, C. L., J. A. Klun, C. T. Bryson, M. Kobaisy, and S. O. Duke. 2005. Isolation and identification of mosquito bite deterrent terpenoids from leaves of American ( Callicarpa am ericana ) and Japanese ( Callicarpa japonica ) beautyberry. J. Agric. Food Chem. 53: 5948 5953. Carroll, J. F., N. Tabanca, M. H. Kramer, N. M. Elejalde, D. E. Wedge, U. R. Bernier, M. R. Coy, J. J. Becnel, B. Demirci, K. H. Can Baser, J. Zhang and S. Zhang. 2011. Activity of Cupresus funebris Juniperis communis and J. chinensis (Cupressaceae) Essential Oils as Repellents Against Ticks (Acari: Ixodidae) and as Repellents and Toxicants Against Mosquitoes (Diptera: Culicidae). J. Vector Ecol. 36: 258 268. Cent ers for Disease Control and Prevention [CDC] 1999. Imported Dengue Florida, 1997 1998. Morbidity and Mortality Weekly Report. 48: 1150 1152. Centers for Disease Control and Prevention [CDC] 2010. Locally acquired Dengue Key West, Florida, 2009 2010. Morbidity and Mortality Weekly Report. 59: 577 581.

PAGE 73

73 Curtis, C. 1986. Fact and fiction in mosquito attraction and repulsion. Parasitol. Today. 2: 316 318. Day, J. F. 2005. Host seeking strategies of mosquito disease vectors. J. Am. Med. Assoc. 21: S17 S22. Debboun, M., D. Strickman, V. B. Solberg, R. C. Wilkerson, K. R. McPherson, C. Golenda, L. Keep, R. A. Wirtz, R. Burge, and T. A. Klein. 2000. Field evaluation of DEET and a piperidine repellent against Aedes communis (Diptera: Culicidae) and Simulium venu stum (Diptera: Simuliidae) in the Adirondack mountains of New York. J. Med. Entomol. 37: 919 923. Dekker, T., M. Geier, and R. T. Carde. 2005. Carbon dioxide instantly sensitizes female yellow fever mosquitoes to human skin odours. J. Exp. Biol. 208: 2963 2972. Dekker, T., W. Takken, and R. T. Carde. 2001. Structure of host odour plumes influences catch of Anopheles gambiae s.s. and Aedes aegypti in a dual choice olfactometer. Physiol. Entomol. 26: 124 134. Dethier, V. G., L. B. Browne, and C. N. Smith. 196 0. The designation of chemicals in terms of the responses they elicit from insects. J. Econ. Entomol. 53: 134 136. Dogan, E. B., and P. A. Rossignol. 1999. An olfactometer for discriminating between attraction, inhibition, and repellency in mosquitoes (Dip tera: Culicidae). J. Med. Entomol. 36: 788 793. Frame, G. W., W. G. Strauss, and H. I. Maibach. 1972. Carbon dioxide emission of the human arm and hand. J. Invest. Dermatol. 59: 155 159. Gaffigan, T. V., R. C. Wilkerson, J. E. Pecor, J. A. Stoffer, and T. Anderson. 2012. Systematic catalog of Culicidae. http://www.mosquitocatalog.org/ Geier, M., O. Bosch, and J. Boeckh. 1999. Influence of odour plume structure on upwind flight of mosquitoes towards hosts. J. Exp. Biol. 202: 1639 1648. Gillies, M. T. 1980. The role of carbon dioxide in host finding by mosquitoes (Diptera: Culicidae): a review. Bull. Ent. Res. 70: 525 532. Granett, P. 1940. Studies of mosquito repellents, I. test procedure and method of evaluat ing test data. J. Econ. Entomol. 33: 563 565. Gratz, N. G. 1999. Emerging and resurging vector borne diseases. Annu. Rev. Entomol. 44: 51 75. Gubler, D. J., P. Reiter, K. L. Ebi, W. Yap, R. Nasci, and J. A. Patz. 2001. Climate variability and change in the United States: potential impacts on vector and rodent borne diseases. Environ. Health Persp. 109: 223 233.

PAGE 74

74 Harbach, R.E. 2012. Mosquito taxonomic inventory. http://mosquito taxonomic inventory.info/ Happold, D. C. D. 1965. Mosquito ecology in Central Alberta, II. Adult population and activities. Canad. J. Zool. 43: 821 846. Huhn, G. D., J. J. Sejvar, S. P. Montgomery, and M. S. Dworkin. 2003. West Nile virus in the United States: an update on an emerging infectious disease. Am. Fam. Physican 68: 653 660. Katritzky, A. R., Z. Wang, S. Slavov, D. A. Dobchev, C. D. Hall, M. Tsikolia, U. R. Bernier, N. M. Elejalde, G. G. Clark, and K. J. Linthicum. 2010. N ovel carboxamides as potential mosquito repellents. J. Med. Entomol. 47: 924 938. Katritzky, A. R., Z. Wang, S. Slavov, M. Tsikolia, D. Dobchev, N. G. Akhmedov, C. D. Hall, U. R. Bernier, G. G. Clark, and K. J. Linthicum. 2008. Synthesis and bioassay of im provement mosquito repellents predicted from chemical structure. Proc. Natl. Acad. Sci. 105: 7359 7364. Khan, A., H. Maibach, and W. Strauss. 1971. A quantitative study of variation in mosquito response and host attractiveness. J. Med. Entomol. 8: 41 43. K han, A., H. Maibach, W. Strauss, and J. Fisher. 1970. Differential attraction of the yellow fever mosquito to vertebrate hosts. Mosq. News. 30: 43 47. Khan, A., W. Strauss, H. Maibach, and W. Fenley. 1967. Comparison of the attractiveness of the human palm and other stimuli to the yellow fever mosquito. J. Econ. Entomol. 60: 318 320. Kline, D. 1998. Olfactory responses and field attraction of mosquitoes to volatiles from limburger cheese and human foot odor. J. Vector Ecol. 23: 186 194. Kline, D. L., Bernie r, U. R., Posey, K. H., and D. R. Barnard. 2003. Olfactometric evaluation of spatial repellents for Aedes aegypti J. Med. Entomol. 40: 463 467. Klun, J., and M. Debboun. 2000. A new module for quantitative evaluation of repellent efficacy using human subj ects J. Med. Entomol. 37: 177 181. Klun, J. A., A. Khrimian, and M. Debboun. 2006. Repellent and deterrent effects of SS220, Picaridin, and DEET suppress human blood feeding by Aedes aegypti Anopheles stephensi and Phlebotomus papatasi J. Med. Entomol. 43: 34 39. Klun, J. A., M. Kramer, and M. Debboun. 2005. A new in vitro bioassay system for discovery of novel human use mosquito repellents. J. Am. Mosq. Control Assoc. 21: 64 70. Knols, B. G. J., and R. De Jong. 1996. Limburger cheese as an attractant for the malaria mosquito Anopheles gambiae s.s. Parasitol. Today. 12: 159 161.

PAGE 75

75 Lehane, M. J. 2005. Location of the host, pp. 27 55 In The biology of blood sucking in insects, Cambridge U niversity Press, Cambridge, UK. Mai bach, H., A. Khan, W. Strauss, and W. Skinner. 1969. Human skin in relationship to mosquito attraction and r epulsion. Conn. Med. 33: 23 28. McCabe, E. T., W. F. Barthel, S. I. Gertler, and S. A. Hall. 1954. Insect repellents. III. N,N diethylamides. J. Org Chem. 19: 493 498. McIver, S. B. 1982. Sensilla of mosquitoes (Diptera: Culicidae). J. Med. Ent. 19: 489 535. Mehr, Z., L. Rutledge, M. Buescher, R. Gupta, and M. Zakaria. 1990. Attraction of mosquitoes to diethyl methylbenzamide and ethyl hexanediol. J. Am. Mosquito Contr. 6: 469 476. Moore, S. J. and M. Debboun. 2007. History of insect repellents, pp. 3 30 In M. Debboun, S. P. Frances and D. Strickman (eds.), Insect repellents: principles, methods, and uses. CRC Press, Boca Raton, FL. PoloPlus. 2012. P robit and logit statistical analysis computer software program, version 2.0. LeOra Software, Petaluma, CA. Posey K. H., and C. E. Schreck. 1981. An airflow apparatus for selecting female mosquitoes for use in repellent and attraction studies. Mosq. News 41 : 566 568. Roachell, W. 1997. Olfactory cues in mosquito host location. Insect Behavior Review Articles. http://www.colostate.edu/Depts/Entomology/courses/en50 7/papers_1997/roachell.html Robert, L. L., J. A. Hallam, D. C. Seeley, L. W. Roberts, and R. A. Wirtz. 1991. Comparative sensitivity of four Anopheles (Diptera: Culicidae) to five repellents. J. Med. Entomol. 28: 417 420. Robertson, J. L., R. M. Russell H. K. Preisler, and N. E. Savin. 2007. Binary quantal response: dose number, dose selection, and sample size. pp. 55 70. In Bioassays with arthropods. CRC Press, Boca Raton, FL. Robertson, J. L., K. C. Smith, N. E. Savin and R. J. Lavigne 1984. Effects of dose selection and sample size on the precision of lethal dose estimates in dose mortality regression. J. Econ. Entomol. 77: 833 837. Rosa, L. H., N. Tabanca, N. Techen, D. E. Wedge, Z. Pan, U. R. Bernier, J. J. Becnel, N. M. Agramonte, L. A. Walker, a nd R. M. Moraes. 2012. Fungal diversity associated with micropropogated medicinal plant of Echinacea purpurea (L.) Moench Amer. J. Plant Sci. 3: 1105 1164. Rutledge, L. C. and R. K. Gupta. 2004. Evaluation of an in vitro bloodfeeding system for testing mo squito repellents. J. Am. Mosq. Control Assoc. 20: 150 154.

PAGE 76

76 Sawyer, D. R. 1986. Malaria on the Amazon frontier: economic and social aspects of transmission and control. Southeast Asia J. Trop. Med. Public Health. 17(3): 342 345. Statistic al Analysis Softwa re Institute [SAS] 2012. Statistical analysis computer software program, version 9.2. SAS Institute, Cary, NC. Schreck, C. E. 1977. Techniques for the evaluation of insect repellents: a critical review. Ann. Rev. Entomol. 22: 101 119. Schreck, C., N. Smit h, D. Carlson, G. Price, D. Haile, and D. Godwin. 1981. A material isolated from human hands that attracts female mosqui toes. J. Chem. Ecol. 8: 429 38. Smith, C., I. Gilbert, H. Gouck, M. Bowman, F. Acree Jr., and C. Schmidt. 1963. Factors affecting the pr otection period of mosquito repellents. USDA Technical Bulletin 1285: 1 36. Takken, W. 1991. The role of olfaction in host seeking of mosquitoes: a review. Insect. Sci. Appl. 12: 287 295. U.S. Department of Agriculture [USDA] 1977. Repellent activity of compounds submitted by Walter Reed Army Institute of Research. I. Protection time and minimum effective dosage against Aedes aegypti mosquitoes. USDA Technical Bulletin No. 1549. U.S. Environmental Protection Agency [US EPA] 20 12 Federal Insecticide, Fun gicide, and Rodenticide Act (FIFRA) P.L. 112 177, effective Sept. 28, 2012, 7 U.S.C. § 136 et seq. http://www.ag.senate.gov/download/fifra Weaver, S. C. and W. K. Reisen. 2010. Present and future ar boviral threats. Antiviral Research 85(2): 328 345. Weldon, P., J. Aldrich, J. Klun, J. Oliver, and M. Debboun. 2003. Benzoquinones from millipedes deter mosquitoes and elicit self anointing in capuchin monkeys ( Cebus spp.). N aturwissenschaften 90: 301 304 Walter Reed Biosystematics Unit [WRBU] 2012. Medically important mosquito species. http://wrbu.si.edu/command_aors_MQ.html World Health Organization [WHO] 2009. Guidelines for efficacy testing of mosquito repellents for human skin. Ref:WHO/CDS/NTD/WHOPES/2009.4, http://www.who.int/whopes/guidelines/en/ Zwiebel, L. J., and W. Takken. 2004. Olfactory regulation of mosquito host interactions. Insec. Biochem. Molec. 34: 645 652.

PAGE 77

77 BIOGRAPHICAL SKETCH Natasha Marie Elejalde was born in Miami, FL, the first of three daughters by Regina Maria Smit and Jose Antonio Elejalde, III. She grew up in Miami where she attended St. Theresa Catholic School for much of her primary education and Our Lady of Lourdes Academy for high school from 1998 to 2002. At 17, she enrolled at the University of Florida in Gainesville, FL and took courses in the basic sciences, psychology, genetics and entomology. Driven by her interest in molecular biology and genetics, she volunteered in evolutionary biology laboratory for two years of her undergraduate career and was selected as an undergraduate teaching assistant for genetics course. S he received her Bachelor of Science in Zoology in 2006 from the University of Florida The semester prior to graduation, Natasha left genetics laboratory to begin working as a student assistant for the USDA in mosquito r epellents laboratory at the Center for Medical, Veterinary and Agricultural Entomology in Gainesville, FL. She worked up to a full time position as a research technician testing experimental chemicals as insect repellents as well as assisting with other pr ojects testing military uniforms for efficacy in insecticide treatment. Her passion for public health issues specifically those related to the transmission and spread of insect borne diseases led her to enroll at the University of Florida again in 2009 to pursue a graduate education in medical entomology under Dr. Bernier She maintained her full time position at the USDA while pursuing her graduate degree part time and developed a research project examining alternative method s for testing mosquito repelle nts which precluded direct testing on human volunteers. In 2012, she married her boyfriend of five years Kent Josep h Agramonte, and took his name. She received her Master of Science in Entomology and Nematology in 2012 from the University of Florida.