Anti-exocrine autoimmunity in the NOD mouse model of Sjogren's Syndrome


Material Information

Anti-exocrine autoimmunity in the NOD mouse model of Sjogren's Syndrome
Physical Description:
x, 150 leaves : ill. ; 29 cm.
Robinson, Christopher Paul, 1970-
Publication Date:


Subjects / Keywords:
Research   ( mesh )
Sjogren's Syndrome -- etiology   ( mesh )
Sjogren's Syndrome -- physiopathology   ( mesh )
Autoimmunity   ( mesh )
Exocrine Glands -- physiopathology   ( mesh )
Exocrine Glands -- chemistry   ( mesh )
Mice, Inbred, NOD   ( mesh )
Mice, SCID   ( mesh )
Models, Immunological   ( mesh )
Cytokines   ( mesh )
Gene Expression   ( mesh )
Salivary Proteins   ( mesh )
Salivary Glands   ( mesh )
Lacrimal Apparatus   ( mesh )
Department of Pathology and Laboratory Medicine thesis Ph.D   ( mesh )
Dissertations, Academic -- College of Medicine -- Department of Pathology and Laboratory Medicine -- UF   ( mesh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph.D.)--University of Florida, 1997.
Bibliography: leaves 136-149.
Statement of Responsibility:
by Christopher Paul Robinson.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 002287013
oclc - 49346360
notis - ALP0164
System ID:

This item is only available as the following downloads:

Full Text






No usual thanks are sufficient to acknowledge my debt and sincere appreciation to

both of my co-mentors, Dr. Michael G. Humphreys-Beher and Dr. Ammon B. Peck.

Their insight has influenced both my personal and professional progression through this

graduate program. In addition, I would like to thank Dr. Linda Brinkley, Dr. Michael

Clare-Salzler, and Dr. Joel Schiffenbauer for their advice and contributions toward this


1 am deeply grateful to Janet Cornelius, Micah Kerr, Jeff Anderson, Elizabeth

Bowen, Jason Brayer, and Kim Nguyen for providing technical assistance on this project.

Special thanks to all of my friends in the Peck, Humphreys-Beher, Clare-Salzler, and

Hillman laboratories who have made these past four years a genuine pleasure.

I would like express my sincere gratitude to my parents, Paul and Wesley

Robinson, without whose love, encouragement and support I would not be the person that

I am today. Lastly, I wish to thank my wife, Meryl, for her love, understanding, and

confidence in me.



ACKNOWLEDGEMENTS...................................................... ii

LIST OF TABLES... ............... .. .... ................... ........................ v

LIST OF FIGURES ....................................... ... .................... .. vi

ABSTRACT......................... ................... ...................... ix


1 INTRODUCTION ............................ ........ ...... ........ 1

NOD M ICE ................. .......... ........ ... ................... 30

Introduction ............ ............ .......... .. .......... 30
Materials and Methods.............. ..... ......... ... .. 32
Results .................. ............................. 37
Discussion .................................... .. ..... .. 50

OADENITIS OF NOD MICE ................ ... .......... 56

Introduction ................. .. .............. ......... 56
M materials and M ethods ........................................... 57
R results .................. ..... .................................. 6 1
Discussion ............ ... ............. ......... ......... .... 74


Introduction .............. .......................... .......... 80
Materials and Methods ............................... ....... 81
Results............... .. .................... ....... 87
D discussion ............. .. ...................... .............. 10 1


Introduction ............... .. ... ..... ......... 105
M materials and M ethods ................. .......... ............ 107
Results ...................... .......... .......... ........... 116
Discussion ................. .......... ................. 124

6 CONCLUSION AND FUTURE DIRECTION ...................... 129

RE FEREN CE S ............................ .... .. ................................. .. 136

BIOGRAPHICAL SKETCH .................................................. ......... 150


Table Page

1. Murine Primer and Probe Sequences ..... ....... ..... ............. 34

2. Percentage of CD3 and B220 in the Lymphocytic Infiltrates of
Submandibular (SMG) and Lacrimal (LAC) Glands of NOD
M ice ................................................... .............. ........ 43

3. CD45RB Expression in Lymphocytic Infiltrating Cells of
Submandibular (SMG) and Lacrimal (LAC) Glands of NOD
Mice...................... ...................... .......... 44

4. Analysis of NOD-scid Saliva................. ................... ....... ............ 62

5. Cysteine Protease Activity in Saliva...................... ............ 115


Figure Page

1. Histological Profile of Tissues Showing Lymphocytic Infiltrates of the
Exocrine Tissues and Insulitis in the NOD Mouse............................ 38

2. Flow Cytometric Analysis ofCD4* and CD8* T Cell Populations in
Lymphocytic Infiltrates in Tissues From 12 wk NOD Mice................ 39

3. Histogram of the Temporal Expression ofCD4' and CD8' T cells in
Lymphocytic Infiltrates of Submandibular (SMG), Lacrimal (LAC),
Parotid Glands (PAR) and Spleen (SPL) from NOD Mice................... 40

4. Histogram of Selective TCR Vp Distribution in Infiltrating Lymphocytes
Isolated From Spleen, Pancreatic Islet, Submandibular, and Lacrimal
Glands of NOD M ice................................................ .......... 45

5. Interleukin mRNA Expression of Lacrimal and Submandibular Glands as
Determined By RT-PCR and Southern Blotting.................................. 48

6. Proinflammatory mRNA Expression of Lacrimal and Submandibular
Glands as Determined by RT-PCR and Southern Blotting................... 49

7. Cytoplasmic Amylase Activity of Parotid and Submandibular Glands............ 63

8. Temporal Changes in the Protein Profiles of Saliva From NOD-scid
M ice..................................... .... ........................... 64

9. N-Terminal Amino Acid Residue Sequences of the 32 kDa, 27 kDa,
and 20 kDa Protein Bands (Shown on Fig 8) and Their Alignment
With the Published Sequence of Murine Parotid Secretory
Protein (PSP)....................... ...... ........ ..... ................... 65

10. Identification of Parotid Secretory Protein in Saliva Using Polyclonal
Anti-PSP Antibody on Western Blot.................... ................ 69

11. Western Blot Analysis of Cytoplasmic Fractions of Submandibular and
Parotid Cell Lysates Using Anti-PSP Polyclonal Antibody..................... 70

12. Western Blot Analysis of Saliva and Cytoplasmic Fractions of Parotid
and Submandibular Cell Lysates Using Anti-Proline-Rich Protein
Polyclonal Antibody...... ........ ................. ............ .............. 71

13. Morphological Changes in the Salivary Glands of NOD-scid Mice................ 72

14. RT-PCR and Southern Blot Detection ofPSP mRNA Isolated From
M urine Lacrimal Glands............. ..... ..... ...... .......... .......... 88

15. Western Blot Detection of Parotid Secretory Protein in Murine Lacrimal
Glands.................................... .................. ... .. ........... 89

16. Purification of Parotid Secretory Protein ............................................. 90

17. Amylase Enzyme Activity in the Presence of Saliva Proteins ....................... 91

18. Bacterial Binding of PSP................ ................................ ............ 93

19. Autoradiograph of PSP Binding to Bacterial Membrane Proteins ................. 94

20. RT-PCR and Southern Blot Detection of PSP Transcripts in NOD
and C3H/HeJ Tissues .......................................................... 96

21. Western Blot Detection of Parotid Secretory Protein in 8 wk NOD
and C3H/HeJ Tissue Lysates........................... ......... ............ 97

22. Sequence of Parotid Secretory Protein cDNA Derived From the Parotid
Gland mRNA From the NOD Mouse Strain ............... .................. 98

23. Autoradiogram of Differential PSP Migration Following Incubation
With Saliva or Salivary Gland Lysates........................................ Ill

24. Western Blot of Differential PSP Migration Following Incubation of
NOD and BALB/c Saliva.................................... ................... 112

25. Western Blot Depicting EDTA Inhibition of the Proteolytic Cleavage
ofPSP.................. ....... .. ...... ................. 113

26. Western Blot Depicting Proteolytic Cleavage of PSP in the Presence
ofEGTA..................... ..... ..... ...... .................... ..... ..... ......... .. 114

27. Histogram of Cysteine Protease Activity in Salivary Gland Lysates................ 119

28. Western Blot Analysis of Saliva for the Presence of Apoptotic
Proteases.. ............................ ... ............... ............... 120

29. Western Blot Analysis of Saliva for the Presence of the Cysteine
Protease Inhibitor; Cystatin.................. ... ............................. 121

30. Zymogram Gel Showing Distinct MMP Activities in NOD and BALB/c
Saliva... ........................... ..... ................ .. 122

Abstract of Dissertation Presented to the Graduate School of the University of Florida in
Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy



Christopher P. Robinson

May, 1997

Chairman: Dr. Ammon B. Peck
Cochairman: Dr. Michael G. Humphreys-Beher
Major Department: Pathology and Laboratory Medicine

The primary goal of this dissertation is to elucidate key features of the

pathogenesis of anti-exocrine autoimmunity in the NOD (non-obese diabetic) mouse. In

addition to autoimmune islet cell destruction, the NOD mouse develops chronic

lymphocytic infiltration of the salivary and lacrimal tissues, leading to dramatic declines in

exocrine gland secretion. The finding that secretary dysfunction in the NOD mouse

correlates to the presence of leukocyte infiltration of the exocrine tissues prompted the

study of this strain as a novel model of secondary Sjbgren's Syndrome. To further

develop this model, the studies presented in this dissertation investigate the anti-exocrine

infiltrating cell populations, physiological and biochemical alterations in NOD exocrine

tissues, and the contribution of the NOD genetic background to the development of anti-

exocrine autoimmunity. By flow cytometric analysis, the infiltrating lymphocyte

populations are shown to be quantitatively and qualitatively similar to that observed in

exocrine biopsies from Sjogren's patients and other animal models of this disease.

Similarly, RT-PCR detection of cytokine gene expression demonstrates similar cytokine

profiles in the NOD exocrine tissues as that seen in human patients and other disease

models. The investigation of the NOD-scid strain shows that morphological and

biochemical alterations of the salivary glands of NOD mice occur in the absence of

lymphocytic infiltration, suggesting that intrinsic alterations of the salivary glands may

underlie autoimmune invasion. Furthermore, aberrant gene expression and proteolytic

processing of the salivary protein, PSP, were detected in both NOD and NOD-scid

animals. This protein, previously reported as salivary gland specific, is further found to be

expressed in the lacrimal glands of mice. The novel cleavage of PSP is the result of a

proteolytic activity uniquely present in the saliva of aging NOD and NOD-scid animals.

Both cysteine protease and matrix metalloproteinase activities were detected in older

NOD and NOD-scid samples, potentially indicating increased apoptotic activity and

glandular restructuring, respectively. Novel proteolytic activity may explain the loss of

submandibular acini in NOD-scid mice, as well as potential candidate enzymes for PSP

cleavage. As such, the findings presented in this dissertation lay the foundation for future

genetic and molecular studies investigating the underlying pathogenesis of autoimmune

Sicca syndrome.


Sjogren's Syndrome is a progressive, debilitating disorder in which the body's

immune system destroys the mucinous secreting exocrine tissues resulting in the hallmark

features of dry mouth and dry eyes (sicca syndrome) (1). While classified as an orphaned

autoimmune disease, the estimated number of cases of Sjogren's Syndrome in the United

States is believed to range between 200,000 to 4 million (2). This wide range in

estimated case numbers is attributed to the difficulty in properly diagnosing the seemingly

ubiquitous nature of many of the patient complaints. In Europe, where the criteria for

diagnosis is less stringent than in the United States, the prevalence is estimated at 0.1 to

1.0% of the population. As typical of many autoimmune diseases, over 90% of Sjogren's

Syndrome patients are women, most of which are post-menopausal (2). Common

complaints include "trouble chewing or parched mouth," "burning throat," "grit or glass in

the eyes," "blurred vision," "itching skin," or "debilitating fatigue"(3). Indeed, Sjogren's

has been termed the "great mimicker" since many symptoms, often considered minor or

vague, resemble those seen in other disease states such as hepatitis C infection, autonomic

neuropathy, or drug treatment (1). Since symptoms are often reported to medical

specialists such as dentists or optometrists, Sjbgren's Syndrome is often misdiagnosed and

symptoms treated as individual entities instead of systemic disorders. These complications

in diagnosis have led to the consensus described by the National Organization for Rare

Disorders (NORD) president Abbey Meyers that "Sjogren's Syndrome isn't a rare disease;

it's just massively under-diagnosed". The etiology of Sjogren's is unknown, and there are

no known cures. It was not until 1933, however, that sufficient interest led to the

landmark studies presented to the medical community by Swedish physician, Henrik

Sjogren (4).

Henrik Sjogren

Individual case studies of dry mouth or dry eye patients were presented as early as

the late 1800's (5). By 1927, Gougerot made the connection between salivary and

lacrimal sicca symptoms, and the linkage of these symptoms to arthritic disease was first

presented (6,7). Tying these findings together, Henrik Sjogren studied a patient group

predominately consisting of women over 40 years of age and displaying

keratoconjunctivitis sicca (4). Sjogren noted the appearance of lymphocytic infiltrates in

both the salivary and lacrimal tissues, as well as salivary gland swelling. Over 50% of the

patients had a history of arthritis. For the next 20 years, little advancement was made

beyond these descriptive observations. In 1951, after publishing numerous supplemental

studies, Sjogren concluded that the major criteria for diagnosis were keratoconjunctivitis

sicca, xerostomia, and polyarthritis (8). While the etiology of disease remained unknown,

Sjogren's microscopic and descriptive findings laid the foundation for the diagnostic

criteria which is still in use today.

Clinical Presentation and Complications of Siogren's Syndrome

The protein and mucin-rich secretions derived from the salivary, lacrimal, and

other minor exocrine tissues, i.e. labial and hardarian glands, are essential for maintaining

the health and integrity of the oral and ocular surfaces. These secretions provide not only

the fluid and electrolytes necessary for tissue homeostasis, but also contain several

additional classes of protein constituents (9). These constituents include important anti-

microbial defense mechanisms against pathogens such as immunoglobulins, iron chelators,

and proteases; growth factors important for mucosal tissue maintenance and regeneration,

such as epidermal growth factor (EGF), nerve growth factor (NGF), and transforming

growth factors (TGF-a and TGF-0); and mucinous lubricating agents. For the most part,

both tear and saliva secretions serve similar functions and contain many of the same

protein constituents, e.g., EGF, NGF, TGF-a, lactoferrin, lysozyme, and immunoglobulins

(9,10). At the same time, however, saliva and/or tear specific secretary proteins, as

evidenced by salivary amylase and digestive enzymes, provide for specialized physiological

functions of the individual secretary fluids (9).

The clinical presentation of sicca syndrome is associated with the loss of both the

fluid and proteinacious phases of saliva (1). This is demonstrated by the fact that

supplemental tear and saliva substitutes lacking the protein constituents of naturally-

occurring exocrine fluids are not complete remedies for patient discomfort associated with

dryness (11). Similarly, complaints of chronic dryness are often displayed in a portion of

autoimmune diabetic patients with normal flow rates (12,13). This indicates that the

protective protein and mucous constituents in addition to fluid saliva are essential for the

maintenance of oral and ocular health.

Disruption of exocrine secretion has severe clinical implications. In addition to

patient discomfort, corneal damage due to lack of lubricating fluids and chronic ocular

infections can lead to blindness in severely affected patients (14. Similarly, loss of

protective saliva and hydration provided by mucinous A*oatings leads to rampant

periodontal disease, caries, candidal infection, and backingg and loss of teeth despite

rigorous dental treatment regimens (15). Drypess and severe cracking of the tongue in

many patients leads to difficulties in speaking, chewing and swallowing most foods. As in

many autoimmune diseases, Sj6gren',patients often develop a waxing and waning in the

severity of symptoms, leading t periods of moderate health followed by chronic episodes

(16). To combat dryness, ~tificial saliva, tears, and ointments can provide symptomatic

relief for many patient; however, severe cases are also treated with anti-inflammatory

drugs (16).

Numerous systemic complications frequently appear in Sjogren's patients. The

most severe of these is a marked increase in the risk (>40 fold) of developing non-

Hodgkin's lymphomas in the salivary glands or cervical lymph nodes (17). In 1991, Fox

et al. demonstrated that salivary gland lymphomas predominately involve a t(14:18)

translocation which allows for increased synthesis of the proto-oncogene bcl-2 (17).

Over-expression ofbcl-4 is known to rescue activated lymphocytes from apoptotic death


Other symptoms include chronic fatigue, itchy skin attributed to both dryness and

vasculitis, and digestive disorders. Many patients appear to have central nervous system

disorders, which, in addition to chronic fatigue, further adds to the systemic nature of this

autoimmune disease (19).

Modem Diagnosis of Sjgiren's Syndrome

New advances in our immunologic and molecular understandings have led to an

expansion of the criteria determining the diagnosis of SjOgren's Syndrome. Today,

Sjogren's syndrome is subdivided into primary and secondary classifications. Primary

Sjogren's syndrome is defined as autoimmune sicca syndrome in the absence of other

autoimmune conditions (1). Secondary Sjogren's syndrome occurs in the presence of

other autoimmune connective tissue diseases, such as rheumatoid arthritis, systemic lupus

erythematosis, or scleroderma. The presence of dry eyes is measured through use of the

Schirmer's test of tear production, and Rose Bengal corneal-staining dye, once used by

Henrik Sjogren, is still used to visualize corneal damage. Salivary gland scintigraphy is

used to evaluate saliva flow rates. In 1960, Bloch et al. detected the presence of anti-

nuclear antibodies in the sera of approximately 70% of Sjogren's patients (20). This

finding was further expanded to identify two specific ribonucleoprotein antigens named

SS-A/Ro (60 and 52 kDa isoforms) and SS-B/La (47 kDa) (21). Antibodies directed

against SS-A and SS-B are present in approximately 90%/ and 70%, respectively, of sera

of Sjogren's syndrome patients (22). Rheumatoid factor is also present in 70%/ of

patients, including those that do not display joint involvement (23). Therefore, presence of

these autoantibodies has been used as a critical criteria for diagnostic purposes.

In the United States, the San Francisco criteria is used to diagnose candidate

patients displaying sicca syndrome (1). Under this criteria a positive diagnosis includes the

detection of lymphocytic infiltration in labial lip biopsies. Unfortunately, different

standards for diagnosis in the international community have led to great controversy in

regards to the patient populations used in clinical and basic research studies (24,25). In

the European (EEC) criteria, for example, examination of labial lip biopsies for presence

of lymphocytic infiltration is not essential for diagnosis. Thus, only an estimated 15% of

EEC patients would be diagnosed with Sjogren's Syndrome under the San Francisco

criteria (25). Since the patient populations are more restrictive in the San Francisco

criteria, comparison of international studies remains controversial.

Genetic Factors in Susceptibility

While the initiating environmental triggers of Sjogren's syndrome are unclear,

intrinsic genes contributing to disease susceptibility are thought to be critical potentiators

in the development of autoimmune disease. Approximately 12% of SS Sj6gren's

syndrome patients have a relative with this disease, demonstrating a familial aggregation

(26). The strongest disease associations appear related to specific alleles of the human

leukocyte antigen (HLA) complex or the corresponding major histocompatibility complex

(MHC) in murine models (27). These highly polymorphic gene products serve to bind and

present peptide epitopes to lymphocytes. A specific allele has unique peptide binding

specificities which are dependent upon the polymorphic amino acid composition in the

binding/anchoring clefts of the molecule. The subsequent presentation of peptides serves

as a foundation for the development of both tolerance to self tissue antigens and the ability

to bind and present novel exogenous peptides to the immune system (28). Therefore, it is

theorized that individuals with specific HLA alleles may be predisposed to selected tissue

specific autoimmunity based on the capacity of tissue autoantigens to bind in the HLA

binding clefts. In support of this theory is the finding that different autoimmune diseases

often correlate to the presence of specific HLA alleles.

In Caucasians with primary Sjogren's syndrome, an increased frequency of HLA-

B8 is found in -60% of Sjogren's patients and only 20% in controls (29). Additionally,

83% of primary Sjogren's syndrome patients displayed HLA-DW3 as compared with 24%

in control populations (30). Racial distinction in patient populations is also detected, as

Japanese patients display higher frequencies of HLA-DRW52 (31). While these findings

were generally true for primary Sjogren's patients, secondary Sjogren's syndrome

associated with rheumatoid arthritis did not display an association with either HLA-DW3

or HLA-B8, indicating genetic distinction between patient groups (29,30). Similarly,

Sjogren's syndrome associated with systemic lupus erythematosis appears related to DR3,

DQw2, and C4AQ0 (complement gene mapped to the HLA region) (31). Specific HLA

haplotypes are further associated with differences in the levels of anti-Ro antibody

production or high autoantibody levels in general. HLA-DW3, for instance, is associated

with decreased salivary flow rates and larger focal infiltration among patients (32).

Therefore, the presence of particular HLA alleles may not only contribute to disease risk,

but also to specific characteristics of an individual patient's disease.

The importance of non-HLA related genes appears to be important as well. A

review of familial studies by Goldstein et al. suggested the presence of a dominant, non-

HLA linked gene (33). In addition, discordance of Sjogren's Syndrome among identical

twins suggests that an environmental trigger(s) plays a critical role as well. The complex

polygenic nature of this disease may explain the widely varying complaints and disease

severity among patients afflicted with Sjogren's syndrome.

Infiltrating Lymphocytes

One advantage to the clinical study of Sj6gren's Syndrome as compared to many

other autoimmune diseases is the routine accessibility of tissue biopsies. Focal infiltrates

of the salivary glands are predominately CD4' T lymphocytes, with a smaller percentage

of CD8 T cells than detected in peripheral tissues (11). Helper T-cells are predominately

of a memory phenotype. Although polyclonal activation of B-lymphocytes is a hallmark

feature of this disease, only 10-15% of salivary leukocytic infiltrates are comprised of B-

lymphocytes. Interestingly, lacrimal gland infiltrates contain nearly double (-30%) the

numbers of B-cells as seen in the salivary glands (34). Whether the increased percentage

of B lymphocytes within the lacrimal glands is due to increased B-cell proliferation or

active recruitment remains unknown. However, detection of CD23 (blast-2 antigen) in the

lacrimal infiltrates of Sjogren's Syndrome patients support the possibility of


Diversity of the T-cell receptor (TCR) repertoire is achieved through the random

arrangements of the genes encoding the a and P chains. While numerous TCR a chains

are available to generate unique VDJ rearrangements, less than thirty P chains are

available to compliment the heterodimeric formation of the TCR. Therefore, it was

questioned whether the generation of autoimmune activity against selected tissue antigens

involved the preferential usage of specific variable P (Vp) chains which had a biased

affinity for tissue antigen epitopes (35). This idea is supported by the expansion of

dominant VB and Vo T-cells specific for defined experimental antigens, such as myelin

basic protein (36). In theory, if a biased TCR VP gene predominated a given autoimmune

disease, specific TCR depletion therapies could be used to slow disease progression

without completely immuno-compromising the patient.

In primary Sjogren's patients, a preferential usage of several TCR Vp genes has

been demonstrated in salivary gland biopsies (35,37,38). However, a similar study of

lacrimal tissues showed a highly diverse TCR repertoire, providing a potential indication

of the involvement of multiple tissue antigens. Unfortunately, the study of T-cell

phenotypes in human patients is hampered by the fact that tissue samples are from the late

stages of autoimmune progression. Therefore, the identification of the lymphocyte subsets

critical for the initiation of disease necessitates the use of animal models.

Cvtokines in Siogren's Syndrome

The immunoregulatory effects of cytokines have stimulated great interest among

researchers. Of note, several studies using immunohistochemistry, in situ RT-PCR, or

RT-PCR methods have shown that production of inflammatory cytokines is not restricted

to infiltrating immunological cells. Indeed, the exocrine tissues appear to produce IL-1

and IL-6 which may contribute to inflammation (39,40). Constitutive production of

transforming growth factor beta (TGFI), a strong immunosuppressive factor, is also

found in these tissues (41). Furthermore, production of IFNy by infiltrating immune cells

has been shown to induce the expression of HLA-DR on the surface of salivary gland

epithelium (42). Cytokine production is also important in upregulating expression of

ICAM-1 and E-selectin on the surface of salivary endothelial cells in Sjogren's patients

(43). Together, these findings have led to the current hypothesis that exocrine cells were

not merely passive targets of an aggressive immune system, but may be active participants

in the autoimmune process. Whether aberrant tissue expression of HLA protein is

involved in stimulating or propagating the autoimmune response has yet to be resolved.

In addition to IL-1 and IL-6, high levels of IL-2, IL-10, IL-12, TNFa and IFNy

are detected in Sjogren's biopsies (11). Levels of IL-4 are typically beyond detection.

While much emphasis has been placed upon delineating whether Thl (cytotoxic response)

or Th2 humorall) T-helper responses predominate in Sjogren's Syndrome, a clear picture

of the complex cytokine interactions is still beyond reach. The presence of IL-10,

described to direct the immune response to Th2 phenotype, in addition to IL-12 and IFNy,

which mediates the opposing Thl phenotype, suggests that the distinction of a Thl or Th2

response may not be applicable for this disease (44). Indeed, the cytokine profiles are

highly suggestive of a generalized pro-inflammatory process. Nonetheless, potential

clinical usage of cytokine-blocking antibodies or cytokine therapy to direct the immune

response to a beneficial state is of extreme interest among researchers.

In addition to immunoregulation, cytokines may play a more direct role in tissue

destruction. In vitro studies have now shown that IFNy is toxic to a human salivary gland

ductal cell line (HSG) (45). Similarly, TNFa, whose receptor bears homology to the

death domain associated with FAS-induced apoptosis, may play a critical role in mediating

both inflammation and apoptotic events (46). Wu et al. demonstrated that TNFa, while

not able to induce significant death of the HSG line, was able to act synergistically with

IFNy to induce rapid cell death (45). Whether these cytokines directly serve to induce

exocrine cell death in the autoimmune state has yet to be determined.

Implications of Viral Expression

Despite extensive research, the initiating triggers for this disease remain a mystery.

Discordance between identical twins and the late disease onset suggests that an

exogenous agent, such as glandular trauma or viral infection, may potentiate

autoaggression in susceptible individuals. Much interest has focused on the potential role

of virus in the initiation of Sjogren's Syndrome. Hepatitis C and HIV infection often

result in dry eye and mouth syndromes, but do not have the same histological and

autoimmune phenotypes of Sjogren's patients (47,48). Transgenic expression of the tax

gene from the HTLV-1, a human leukemia virus, in mice results in lymphocytic infiltration

of the exocrine glands (49). Detectable lax gene sequences were evident in 25% of labial

salivary gland biopsies of Sjogren's patients (50). Epstein-Barr Virus (EBV) antigens and

genomic DNA has been detected in the lacrimal tissue of Sjogren's patients, however high

levels in normal biopsies complicates these findings (51). Of particular interest is the

finding that the SS-B/La nuclear autoantigen is associated with viral translation and may

also be redistributed to the cytoplasm and cell surface following infection (52). Whether

the generation of autoantibodies against the Ro and La ribonucleoproteins is potentially

due to their association with viral RNAs or molecular mimicry of viral epitopes is

currently a hot topic among researchers. Thus, the potential involvement of viral infection

in the initiation or exacerbation of this disease is intriguing; however, researchers have yet

to demonstrate a primary etiologic role of virus in Sj6gren's Syndrome.

Autoimmune Disruption of Exocrine Secretion

Stimulation of exocrine secretion is under the control of both sympathetic and

parasympathetic innervation. Parasympathetic nerves, primarily allowing for the release of

watery phase of saliva, are responsible for basal levels of saliva production as well as

mechanically stimulated saliva flow, such as when eating (53). Parasympathetic

stimulation leads to the release of electrolytes which form the osmotic gradient needed for

salivation. Mucin rich saliva secretion is mediated through sympathetic stimulation.

Pilocarpine, a muscarinic receptor agonist, is a drug currently used to treat patients

through parasympathetic stimulation (54). In our studies, pilocarpine was used to

stimulate watery saliva, while the B-adrenergic receptor agonist, isoproterenol, was used

to stimulate proteinacious saliva flow.

Disruption of the naturally occurring secretary process is a frequent consequence

of pharmaceutical drugs or the effects of irradiation in anti-cancer therapies (1). While the

pathogenic mechanism of exocrine dysfunction in Sjogren's Syndrome is unclear,

autoimmune disruption of the secretary process may occur through degeneration of

glandular innervation, blockage of receptor stimulation, cellular destruction, or

degeneration of the secreting cells. It is generally believed that immune destruction of the

exocrine cells is not the primary mechanism for the loss of exocrine secretion (11). This is

in stark contrast to Type 1 insulin dependent diabetes, for instance, where disease onset

occurs when over 90% of the insulin secreting 3 cell mass is eliminated (55). Histological

evaluation of exocrine biopsies reveals that only approximately 10% of the tissue area is

replaced by lymphocytic infiltration (56). In areas not in direct contact with lymphocytic

foci, large areas of seemingly normal tissues are present. Furthermore, the severity of

sialadenitis does not appear correlated with measured loss of glandular secretions. Equally

perplexing is the fact that loss of glandular secretion appears inversely related to the

presence ofanti-ductal cell antibodies in patient sera (57). These findings have led to the

current theory that the exocrine glands of Sjogren's patients are "loafing"(l 1). Chronic

destruction of the normal glandular architecture could lead to tissue dedifferentiation and

loss of secretary function.

The classical waxing and waning of symptoms in Sjogren's patients suggests that

lymphocytes play a direct role in effecting the loss of salivary function. For instance, if

glandular innervation was effectively destroyed, abrogation of the immune response would

not be expected to restore function. Therefore, it is possible that cytokines and antibodies

released by immune components are playing an active role in disrupting neuro-glandular

stimulation. Autoantibodies against the Ro and La antigens, while prevalent in Sjogren's

patients, have yet to be directly implicated in the loss of exocrine function; although high

levels of maternal autoantibody correlates to an increased risk of severe heart defects in

children (1).

Animal Models of Sjogren's Syndrome

The inherent scientific and ethical limitations in the study of human subjects has

forced basic researchers to turn to the study of animal models which display specific

disease traits. Several murine models for Sjogren's Syndrome have been suggested based

on the presence oflymphocytic infiltration in the salivary glands, including the NZB/NZW

F1, MRLApr, NFS/sld, and TGFp knockout congenic mice (58-61). The TGF3 knockout

mouse develops extensive systemic autoimmunity including anti-exocrine infiltration and

typically dies within one month of birth (61). In 1994, Haneji et al. detailed anti-exocrine

autoaggression in NFS/sld mice thymectomized at birth, however exocrine dysfunction

was not evaluated in this study (60). The NFS/sld mouse was shown to contain a single

recessive gene defect in sublingual gland development, indicating that abnormalities in

glandular development may contribute to autoimmune tissue targeting.

The NZB/NZW Fi and the MRL/lpr mice display lymphocytic infiltration of both

the lacrimal and salivary glands in addition to lupus-like disorders, and have therefore been

studied as models for secondary Sjogren's Syndrome (58,59). Interestingly, the presence

of lymphocytes in these tissues and apparent histological changes in the exocrine glands of

these mice does not correlate to severe loss of glandular function. Only NZB/NZW mice

>4 months of age displayed slightly abnormal Schirmer tests, and none of these strains

showed a loss of salivary flow (54). Walcott et al. recently described progressive

degeneration of glandular innervation in the lacrimal glands of NZB/NZW mice over 6

months of age (62). Similar to that of human disease, focal infiltration of the exocrine

tissues in NZB/W and MRIJIpr mice was more prominent in female than male mice,

occupied no more than 30% of tissue sections, and intensity of the infiltration did not

correlate to glandular dysfunction (54).

Lymphocytes infiltrating the exocrine tissues of MRL/Ipr mice, which have a

genetic disruption of Fas-FasLigand-mediated apoptosis, are predominately CD4' T-cells

(63). Analysis of TCR Vp expression revealed a diverse T-cell repertoire, however, a

skewing of the population to V18.1.2 and V36 was detected (64). Purified CD4', but not

CD8' lymphocytes isolated from MRL/lpr salivary glands are able to transfer sialadenitis

to immunocompromised CB17-scid mice (65). Interestingly, specific depletion of donor

VP38 or VP6* lymphocytes prior to transfer significantly inhibited sialadenitis in recipient


Both the NZB/NZW Fl and MRL/lpr mice display similar cytokine profiles as

those detected in human tissue biopsies (66). Therefore, since these mice do not develop

severe functional deficiency, it is unlikely that the cytotoxic effects of hallmark cytokines

such as IFNy are primary effectors of exocrine dysfunction in murine models. Together

these findings demonstrate that the mere homing of leukocytes to the exocrine tissues is

not sufficient to cause the loss of exocrine function. In 1992, an advance in the search for

an animal model for Sjogren's was made by Hu el al. in their description of secretary

dysfunction in autoimmune NOD (nonobese diabetic) mice (67).

NOD Mouse

First introduced in 1980, the inbred non-obese diabetic (NOD) mouse exhibits a

strikingly similar pathology to that of human Type 1 insulin dependent diabetes mellitus

(IDDM) (55). Lymphocytic destruction of the p-cells of the islets of Langerhans results in

a loss of blood glucose regulation due to the loss of insulin secretion. Immune infiltration

of the pancreas can begin in NOD mice as early as two weeks of age and begins with the

appearance of Class 11 monocytes and CD8 T lymphocytes (68). Overt diabetes in NOD

mice generally begins between 8-12 wks of age and occurs when >90% of the B-cell mass

is destroyed. By 30 weeks of age roughly 80% of female and 20% of male mice become

diabetic (55). Interestingly, NOD mice kept in specific pathogen free (SPF) colonies

develop diabetes more frequently and at an earlier age that those kept in traditional

colonies (69). This indicates that, similar to the human condition, environmental factors

play a role in the development of autoimmunity in these mice.

The observation of the lymphocytic infiltration in the NOD mouse is not confined

to the pancreas, however, but is also observed in the salivary and lacrimal glands of this

strain (70). Interestingly, the parotid gland does not develop extensive lymphocytic

infiltration. The first appearance of periductal and perivascular lymphocytic infiltration in

the submandibular and lacrimal glands begins at 8-10 wks and 10-12 wks of age,

respectively (71). By 18 week of age, the exocrine glands display focal lymphocytic

lesions with an observable disorganization of normal acinar structure. Despite the early

appearance of lymphocytic infiltrates within the submandibular and lacrimal glands, loss

of saliva flow and tear production does not typically occur until 14-16 wks of age (67).

This result was expanded to show that NOD mice lose approximately 90%/ of their saliva

flow and 30% of tear flow between 8 and 20 weeks of age (71). Unlike diabetes in the

NOD, both male and female mice develop infiltration of the exocrine tissues at similar

rates (70). These findings reveal that the NOD mouse represents the first-described animal

model for the spontaneous autoimmune-induced loss of both saliva flow and tear

production, and, as such, is emerging as an excellent model for the study of Sjogren's

Syndrome in humans.

Immunogenetics of the NOD Mouse

The NOD mouse arose during the selective breeding of the inbred cataract

Shionogi (CTS) strain from outbred ICR mice (55). In 1974, a single female mouse

exhibiting polyuria, severe glycosuria, and weight loss was discovered, and, after extensive

inbreeding, the NOD strain was introduced in 1980. Since this time, at least 13 genetic

loci have been discovered which contribute to the development of diabetes in NOD mice

(72). An additional two genetic loci, Idd-7 and Idd-8 found in the B10 background, are

also able to increase diabetes progression when bred onto the NOD background (69).

Genetic influence contributes to both the incidence of diabetes, as in Idd-l, as well the

timing of disease onset, as in Idd-2 and Idd-4.

The MHC-linked locus, Idd-1, is essential for the development of diabetes, as

replacement of the of the NOD MHC loci with the B10-derived allele completely inhibits

both diabetes and insulitis, but not sialadenitis (73). The NOD MHC, H-287 haplotype

contains several unique features, including a structurally distinct I-A molecule and a

deletion of I-E expression (74). Conserved serine-aspartic acid or proline-aspartic acid at

the A3 amino acid positions 56 and 57 are replaced with histidine and serine in the NOD

allele (75). In humans, replacement of the aspartic acid with uncharged amino acids at

position 57 of the human DQP has been implicated in increased diabetes risk (76).

Transgenic expression of I-Ak in NOD mice inhibits insulitis and diabetes, however

sialadenitis was not affected (77). The lack of I-E expression is not unique to the NOD

strain, but is also a feature of inbred C57BL/6, SJL, ACA, and DBA/I mice (78). Since

the B10 mice used in generating the NOD.B0I-H-2b congenic are also I-E negative, the

impact of 1-E expression on the NOD strains was not investigated in this model.

Generation of transgenic NOD mice which express the I-E molecule has determined that I-

E expression is protective against diabetes and insulitis (79). Sialadenitis was not

investigated in these mice. Interestingly, Faustman found occasional lymphocytic

infiltration in the submandibular glands in the majority of the I-E negative strains

mentioned above (78). The I-E molecule is responsible for the deletion of TCR vI35 T-

lymphocytes, suggesting that vp5' lymphocytes may play a role in the development of

sialadenitis. However, NOD mice backcrossed with SWR mice, which contain a deletion

in v35, vp8 and vp11 gene segments, still develop typical frequencies of insulitis and

diabetes (80). Together these findings demonstrate that both the unique NOD I-Ag7

molecule and the lack of I-E expression are important for diabetes, however exocrine

gland infiltration is only associated with the lack of I-E expression.

The contributions of non-MHC genes to the development of diabetes and

sialadenitis is an area of intense interest. Garchon et al. linked a centromeric locus on

chromosome 1 with diabetes susceptibility (81). Interestingly, a second susceptibility

locus on the same chromosome was linked to sialadenitis, peri-insulitis, and hyper-IgG

production. The development of sialadenitis was found to be a dominant trait in these

studies (81). To evaluate immune and non-immune components of autoimmunity in the

NOD mouse, Leiter et al. bred the scid mutation onto the NOD genetic background,

creating the NOD-scid strain which was extensively used in the current studies (82).

Homozygosity at the scid locus leads to the loss of functional T and B lymphocytes, and,

therefore, these animals do not develop insulitis, sialadenitis, or diabetes. This study also

demonstrated that, similar to parental NOD mice, NOD-scid mice have complement

deficiency, functionally immature macrophage populations, and decreased natural killer

cell activity (82). The sum results of these immunogenetic findings are promising in that

they strongly suggest that the ultimate goal of genetically separating diabetes from

exocrine gland autoimmunity to develop a primary marine model for Sjogren's Syndrome

will be attainable.

Role of Helper T-Lymphocvtes

Prior to this study, the lymphocyte phenotypes infiltrating the exocrine glands were

largely unexplored. Therefore, our knowledge of autoimmunity in the NOD mouse comes

almost exclusively from diabetes research. While autoimmune infiltration includes several

immune cell types, researchers have predominately focused on the role of CD4' T

lymphocytes which represent the largest component of the infiltrate. Neonatal

thymectomy of NOD mice generally inhibits insulitis and diabetes, and athymic nude NOD

mice do not develop either phenotype (83,84). Similarly, injection of anti-Thyl.2, anti-

CD4, or anti-CD3 antibodies suppresses diabetes (55). A direct role of lymphocytic

involvement is demonstrated by the ability of splenic or tissue infiltrating lymphocytes

from diabetic donors to rapidly transfer both diabetes and sialadenitis to non-diabetic

NOD or NOD-scid recipients (85). As opposed to rapid transfer of disease, splenic

lymphocyte preparations from prediabetic mice suppressed disease onset. Therefore, the

ability to successfully transfer disease is not only dependent upon the presence of CD4* T

lymphocytes, but also their activation state.

When transferred to immunocompromised recipients, purified naive CD4+ T cells,

represented by the CD4'CD45RBhi phenotype, induce a wasting disease associated with

intestinal inflammation (86). Interestingly, co-transfer ofCD4+CD45RBo memory T cells

suppresses the wasting disease (87). Two distinct populations of CD4' memory T cells

with opposing diabetogenic potential have been separated from prediabetic and diabetic

NOD mice based upon their cytokine expressions (87). Memory cells producing high

levels of IFNy (Thl) and low levels of IL-4 transfer diabetes rapidly to recipient mice.

However, memory cells producing low amounts of IFNy and high levels of IL-4 (Th2)

confer active suppression (87). This indicates that the ability to progress from a protective

to a pathogenic state may lie in the relative levels of cytokines present in the autoimmune

lesion. This is supported by similar findings using diabetogenic CD4' T-cell clones.

Haskins et al. demonstrated that shifting high IFNy producing T cell clones to IL-4

producers in long term culture suppressed their ability to transfer disease (88).

Furthermore, successful transfer of diabetes with CD4' T cell clones into NOD-scid

recipients demonstrates the potential of CD4' cells to transfer diabetes in the absence of

either CD8' T cells or B lymphocytes.

Analysis of TCR V3 usage in the pancreas of NOD mice shows heterogeneous

TCR populations. However, much like the MRL/lpr mouse, a preferential usage of V08

and Vp6 genes is present in diabetogenic T cells (89). Edouard et al. showed that

selective deletion of Vp6' or Vp38 T cells prior to adoptive transfer significantly lowered

the ability to transfer diabetes to irradiated NOD recipients (90). Although suggestive of a

pathogenic role for Vp6' and VB8* T cells, NOD X SWR mice develop insulitis and

diabetes despite deletions in Vp5, V38, and Vl 31 gene segments (80). In addition, TCR

usage from islet specific T cell clones of diabetic NOD mice include Vp4, Vp6, V18.2,

Vp112, Vp16, and V319 gene usage, further indicating that glandular specific TCR gene

usage is not restricted (55). Recent findings by Anderson (personal communication) have

indicated that TCR V1 restriction may be present in a CD3*CD4"CD8 "double negative"

T cell population which is present in early islet infiltrates; however, this has yet to be


CD8' T Lvmphocvtes and B Lymphocytes

While the necessity of CD4' lymphocytes is readily apparent, the role of CD8' T

cells and B cells in the initiation of autoimmunity is less clear. Transfer of purified CD8* T

cells is typically insufficient to cause diabetes, however a recent report demonstrated that

diabetes can be transferred by a CD8+ T cell clone (85,91). Using a gene knockout

strategy, Wicker et al. determined that disruption of the 0-2 microglobulin gene, which

results in a specific loss of CD8 T-cells, inhibits both insulitis and diabetes in NOD mice

(92). However, in prediabetic mice, co-transfer of both CD4+ and CD8' populations is

necessary to initiate diabetes (85). Therefore, while an aggressive CD4* or CD8' T cell

clone may be able to cause diabetes, both cell types appear essential for the initiation of


Similarly, despite the fact that autoantibodies are detected on the surface of islet

cells prior to leukocytic infiltration, the role of B-cells has been largely downplayed in the

literature (93). Renewed interest in the B cell component is being generated with the

discovery that NOD mice containing a p-heavy chain knockout which halts the generation

of mature B lymphocytes do not develop insulitis or diabetes (94). While exocrine

dysfunction has yet to be evaluated in these mice, several studies have indicated that the B

cell component may be important for exocrine dysfunction in NOD mice. In addition to

the presence of antibodies directed against islet cell components, NOD sera contains both

anti-nuclear antibodies and antibodies directed against salivary gland cells (95,96). In an

analysis of the signal transduction pathways responsible for secretion from salivary and

lacrimal glands, Humphreys-Beher et al. identified a subclass of autoantibodies directed

against the muscarinic and p-adrenergic cell surface receptors that initiate this process

(97,98). The density of these cell surface receptors is reduced on both the parotid and

submandibular glands, along with secretary response to 0-adrenergic,

muscarinic/cholinergic and neuropeptide agonists during disease. Thus, it is possible that

soluble factors such as autoantibodies may ultimately be responsible for the loss of

secretary function in NOD mice. This may also explain how parotid gland function is

inhibited despite the lack of focal lymphocytic infiltration. Together, these findings

suggest that the initiation of autoimmunity in the NOD mouse involves multiple immune

components and numerous cell types.

Potential Dichotomy of Diabetes and Sialoadenitis

Approximately 30% of autoimmune diabetic patients suffer concomitantly from

xerostomia and xerophthalmia due to insufficient glandular secretions (13). These clinical

symptoms are thought to result from poorly maintained blood glucose regulation, lack of

insulin secretion, or neuropathy (99). Expansion of the autoimmune processes into the

exocrine glands of diabetic patients is rarely encountered and generally not considered to

be the causative factor for exocrine dysfunction in diabetics (12). In support of this, tight

glucose control generally alleviates the feeling of dryness in most patients. Therefore, is it

possible that the loss of exocrine secretion in NOD mice is due to the loss of blood

glucose regulation, and not anti-exocrine autoaggression?

In addition to the immunogenetic studies discussed above, several studies have

illustrated a dichotomy between autoimmune diabetes and autoimmune sialadenitis in the

NOD mouse. Loss of secretary function in NOD mice, while more severe after diabetes

onset, is present in both male and female prediabetic mice, which lose 50-70% of

stimulated salivary flow between 8 and 20 weeks of age (67). Similarly, injection of

diabetic mice with daily insulin injections does not restore secretary function. Recent

findings using NOD.B10-H-2b mice have shown that the loss of secretary function occurs

in the absence of either insulitis or diabetes (unpublished observations). Treatment of

NOD mice with antibodies against alpha 4-integrin and L-selectin was able to significantly

inhibit both insulitis and diabetes, however, sialadenitis was unaffected (100). In humans,

increased expression of ICAM-1 and E-selectin is detected on salivary epithelial cells (43).

From the standpoint of tissue tolerance, intrathymic injection of islet cell homogenates or

transgenic expression of proinsulin II in NOD mice prevents diabetes but fails to protect

against sialadenitis (101,102). These studies suggest that the antigens involved in the anti-

exocrine response are distinct from those in the islet, and that loss of exocrine gland

tolerance is not secondary to p cell autoimmunity.

Changes in Saliva Proteins in the NOD mouse

In addition to loss of secretary function in NOD mice, the protein constituents of

NOD saliva change over time. This is reflected by reductions in amylase activity (>50%)

and loss of ductal cell secretion of EGF (67). In newly diabetic male NOD mice, over

97% of EGF production is lost. In this dissertation, specific changes in both parotid

secretary protein (PSP) and proline rich proteins (PRP) are observed in both NOD and

NOD-scid animals which parallel changes in EGF and amylase. The proline rich proteins

are latent in normal mice; however, they comprise >70% of the protein in human saliva,

where they serve to bind both calcium and hydroxyapatite (103). Expression of the PRPs

can be induced in both the marine parotid and submandibular glands by chronic B-

adrenergic stimulation or the introduction of high levels of tannic acid in the diet (104). In

these studies we demonstrate the abnormal presence of PRPs in both submandibular and

parotid glands of aging NOD mice.

In normal mice, PSP is described as a 20 kDa, leucine-rich glycoprotein (-23%

leucine; 235 amino acids) of unknown function that is secreted predominately by amylase

producing acinar cells of the parotid gland (105). Studies have shown a developmental

coordination of murine PSP and salivary amylase expression in the parotid gland of adult

mice where the two proteins appear in constant ratios (106). Coordinate expression,

however, is not determined by the rate of gene expression, since developmental expression

of the PSP gene occurs before amylase expression (107). Although PSP is specific to the

parotid and sublingual glands of adult mice, PSP is expressed in the developing

submandibular gland up to 5 days of age (108). This is important in that abnormal re-

expression of PSP in the submandibular gland is detected in aging NOD mice and may

indicate dedifferentiation of the glandular acini. While the function of PSP is unknown,

our findings suggest that it may play a role in anti-microbial binding. To date, PSP has not

been detected in human saliva (our anti-mouse PSP antibody does not crossreact),

however a single copy of the PSP gene is present in human cells and PSP mRNA has been

detected in human parotid tumors (109).

Results in this dissertation show that PSP is enzymatically cleaved in older NOD

saliva. The cleavage occurs between leucine and asparagine at the 26th and 27"' amino

acids of the protein. Database searches uncovered no known enzymes which may be

responsible for this cleavage. However, since PSP cleavage corresponds to the time of

dramatic acinar cell loss in the NOD mouse, we investigated the potential of both matrix

metalloproteinases and apoptotic cysteine proteases to cleave PSP through bystander

activity. These proteins, involved in glandular restructuring and apoptotic cell death,

respectively, may play a key role in the loss of submandibular acini in NOD and NOD-scid


Matrix Metalloproteinases and Cysteine Proteases

Cellular homeostasis depends on regulated cell proliferation coupled to cell death

(110). The matrix metalloproteinases (MMPs) are a class of zinc-dependent enzymes

which, in conjunction with their specific inhibitors (TIMPS), are responsible for

restructuring and maintenance of the extracellular matrix. Over 12 MMPs have been

described and include the collagenases, gelatinases (collagen type IV), stromalysins

(laminases), and elastins which are able to cleave virtually all components of the

extracellular matrix (111). Expression of specific MMPs during the developmental

process is responsible for sculpturing the extracellular environment and thereby dictating

cellular turnover and differentiation. While much research has focused on their association

with tissue homeostasis and tumor metastasis, a prominent role of MMPs in autoimmune

processes has recently been described (111). Monocyte production of MMPs is involved

in dictating the extent of tissue damage caused by inflammation and may additionally be

critical for matrix degradation necessary for lymphocyte chemotaxis. Inhibitors of MMPs

have been shown to significantly reduce the appearance and severity of EAE in animal

models (112). This effect was attributed to the inhibition of monocyte-derived gelatinases

responsible for the degradation of the blood-brain barrier basement membrane. Similarly,

MMP inhibitors significantly reduce joint inflammation and tissue destruction in collagen-

induced murine model of rheumatoid arthritis (113). Studies have further demonstrated

that MMP activity is connected to the apoptotic pathway through the ability to cleave cell

surface FAS on apoptotic target cells (111). In humans, SjOgren's patients have elevated

levels of collagenase (MMP-1) in their saliva, and supematants from excised biopsies

contain elevated MMP activities (114). Whether elevated or aberrant MMP production

precedes or is a result of tissue inflammation in autoimmune target tissues is still unknown,

however the extensive changes in glandular architecture seen in tissue biopsies may, in

part, be attributed to MMP activity.

Extracellular signaling molecules are capable of regulating glandular cell

populations through a series of intracellular events termed programmed cell death or

apoptosis (110). Apoptosis is distinct from cell lysis in that the target cell actively and

efficiently mediates its own death. This process involves intracellular activation, reduction

of cell volume, chromatin condensation, and endonuclease cleavage of DNA within the

cell membrane (115). By this process, the apoptotic cell contents are kept safely inside the

cell membrane until phagocytized by macrophages or neighboring cells. Immune system

molecules capable of mediating apoptosis include the cytokine tumor necrosis factor

(TNF) and the related protein CD95 (FAS) (115). Mice that lack functional FAS surface

protein develop a lymphoproliferative disorder and lupus-like pathology associated with

the inability to delete lymphocytes during the immunological education process (116). In

Sjogren's patients, it is hypothesized that dysfunction in apoptosis results in abnormal

longevity of both T and B lymphocytes and may underlie polyclonal B-cell activation.

Elevated levels of bcl-2, which is able to rescue target cells from apoptotic death, are

detected in salivary gland biopsies (117). Interestingly, 2-fold increases in surface FAS

expression are detected in peripheral blood T cells from both Sjogren's Syndrome and

SLE patients than controls which correlated with accelerated apoptosis of these cells in

vitro (118). Therefore, it is unlikely that defective FAS molecule is directly responsible

for Sj6gren's Syndrome or SLE in most diseased patients.

Activation of the FAS or TNF receptors leads to a proteolytic cascade involving

members of the cysteine protease family (110). These proteases cleave important cellular

proteins including pro-enzymes of other members of this class of proteases and

endonucleases responsible for DNA cleavage. A prototype protein of the cysteine

proteases is interleukin-l -converting enzyme (ICE) which has been characterized as the

activator of the cytokine interleukin-lp through cleavage of its precursor at

Aspl 16/Alal 17 (119). All members of the ICE-related cysteine protease family, including

Nedd-2 and apopain/cpp32, cleave their substrates after an aspartate residue followed by a

small amino acid residue which is important for substrate consensus recognition (120).

Expression of ICE in neuroblastoma or fibroblast cell lines leads to apoptotic death (110).

Therefore, high levels of cysteine protease activity may be indicative of both the

activation of apoptotic mechanisms as well as the processing of proinflammatory cytokine

precursors. Since both MMPs and cysteine proteases appear involved in regulated cell

proliferation and cell death, it is possible that upregulation of these enzymes in the NOD

mouse are directly involved in acinar cell death and the aberrant cleavage of additional

salivary proteins.

Further Development of the NOD Mouse Model of Soigren's Syndrome

The physiological loss of secretary function makes the NOD mouse an ideal animal

model for the study of human Sjogren's Syndrome. Despite the limited amount of

exocrine gland research, the most exciting aspect of the NOD mouse is the ability to apply

the findings of over 15 years of diabetes research to the understanding of autoimmune

sicca syndrome. Additionally, the seeming dichotomy between sialadenitis and diabetes

suggests that the study of congenic NOD mice, such as the NOD.BIOH-2b, will lead to the

development of a primary model to study anti-exocrine autoimmunity. To begin these

studies, however, it is necessary to determine the immunological and physiological

processes underlying exocrine dysfunction in the NOD mice. Therefore, the studies of this

dissertation explore the following specific aims:

1. To determine the temporal changes in exocrine gland histology, infiltrating lymphocyte

populations, and cytokine production spanning the initiation of the autoimmune

process to the final stages of glandular destruction.

2. To investigate exocrine gland abnormalities in immunodeficient NOD-scid mice in

order to separate immune and non-immune components of the disease process.

3. To investigate the tissue specificity and function of parotid secretary protein, which is

found to be aberrantly expressed and processed in aging NOD mice.

4. To evaluate the proteolytic activity in NOD saliva responsible for the aberrant

cleavage of PSP and investigate candidate enzymes for PSP cleavage.



During the past several years significant interest has developed in detailing the

autoimmune destruction of the salivary and lacrimal tissues in the NOD mouse.

Pioneering studies have demonstrated that the lymphocytic infiltration of the salivary and

lacrimal glands correlates with a functional decline in saliva flow and tear production

independent of the loss of blood glucose regulation observed in NOD mice (67,71).

Lymphocyte transfer studies have now shown that the induction of thymic tolerance to the

pancreatic islets does not confer immunologic tolerance to the salivary tissues, suggesting

a potential dichotomy of disease states between these target tissues (101). This is

supported, as well, by immunogenetic studies linking sialoadenitis, hyper-IgG production,

and peri-insulitis to a centromeric loci on Chromosome 1, while insulitis is linked to a

telomeric loci on the same chromosome (81). In addition to anti-pancreatic and anti-

insulin autoantibodies, NOD sera has been shown to contain autoantibodies targeting the

acinar and ductal cells of the submandibular and parotid glands, including those directed

against the muscarinic and P-adrenergic receptors responsible for the generation of saliva

flow (97,98). The decline in saliva output is accompanied by changes in salivary protein

content as well. Both alterations appear to be the result of down-regulation of signal

transduction components of the salivary glands (97). Abnormalities in gene transcription

and protein processing in the salivary glands of immunodeficient NOD-scid mice suggest

that glandular dysfunction in NOD mice may precede lymphocytic infiltration and provide

a potential trigger for the autoimmune targeting of the salivary glands (121).

The NOD mouse represents the first-described animal model for the spontaneous

autoimmune-induced loss of both saliva flow and tear production, and, as such, is

emerging as an excellent model for the study of Sjogren's Syndrome in humans. Sjogren's

Syndrome is an autoimmune disease characterized by dry eye and dry mouth syndromes

due to the destruction of exocrine tissue. Clinical patients develop chronic lymphocytic

infiltration of the salivary and lacrimal glands as well as a cell-mediated and autoantibody

response against the exocrine tissue (11). Diagnosis of Sj6gren's Syndrome often relies

on the detection of lymphocytic infiltration in labial lip biopsies excised from patients,

which have revealed a predominance of CD4' T lymphocytes with an oligoclonality of the

T-cell receptor repertoire (38). In a majority of patients, autoantibodies to ribonuclear

protein antigens are also present (20). In addition, increases in IFNy, TNFa, IL-2, IL-6,

and IL-10 cytokine production have been described in biopsy tissues (11). However, since

patient samples are from a late stage, the initial events surrounding the development of

autoimmune infiltration and destruction of tissue remain unknown.

In this study I have characterized the infiltrating lymphocyte repertoire and

cytokine mRNA profile of the lacrimal, parotid, and submandibular glands of NOD mice

during the course ofimmunopathogenesis. Spleen and islet-infiltrating cells were analyzed

as control lymphocyte populations throughout the time course of disease progression.

This has allowed for a qualitative comparison of lymphocyte populations infiltrating the

pancreas, submandibular and lacrimal tissues of NOD mice, as well as provide a detailed

account of the anti-exocrine autoimmune response which can be compared to that seen in

Sjogren's Syndrome.

Materials and Methods


Female NOD mice (7 wks of age) were purchased from Jackson Laboratories (Bar

Harbor, ME) and maintained throughout the course of the study in the animal facility at

the Health Science Center at the University of Florida (Gainesville, FL). Diabetes was

diagnosed by elevated blood glucose levels using Chemstrip bG reagent strips (Boehringer

Mannheim, Indianapolis, IN). Mice with blood glucose >200 mg/dL were given insulin

injections (lU/mouse/day ip) (67). Evidence of diabetes was first noted at 14 weeks of

age. At 16 weeks, 3 of 5 mice were diabetic; at 18 weeks, 1 of 5, and at 20 weeks, 2 of 5

mice were diabetic.


Monoclonal antibodies used in this study were purchased from PharMingen (San

Diego, CA) and are as follows; CD3e (clone 145-2cll), CD4 (RM4-5), CD8a (53-6.7),

CD45RB/B220 (RA3-6B2), CD45RB (23G2), TCR Vp3 (KJ25), TCR VI6 (RR4-7),

TCR V08.1, 8.2 (MR5-2), TCR V39 (MR10-2), TCR V31l (RR3-15), and TCR Vpl7a


Tissue Preparation

Spleen, pancreas, lacrimal, parotid and submandibular glands were removed at

each harvest. A small piece was cut from each tissue, placed in 10% buffered formalin

and submitted to the Diagnostic Referral Laboratory at the University of Florida

(Gainesville, FL) for histologic sectioning and staining. The remainder was processed as a

pool from the 5 mice per group. A small aliquot of each pool was removed for RNA

isolation and all remaining tissue prepared for flow cytometric analyses.

Single cell suspensions of splenic leukocytes were obtained by gently pressing

spleens through wire mesh screens and washing in PBS (68). Red blood cells were lysed

with 0.84% ammonium chloride. After washing, the remaining leukocytes were aliquoted

at 1 x 106/tube and washed with FACS buffer (PBS with 0.1% NaN3 (FISHER Scientific,

Orlando, FL) and 0.5% BSA (SIGMA Chemical Co, St. Louis, MO) prior to antibody


Tissue Digestion

Except for spleen, all pooled tissues were dissociated by gentle mincing followed

by digestion at 37C in a shaking water bath for 15 minutes in a mixture of 4 mg/ml

collagenase Type V (SIGMA) + 100 U/ml DNase Type II (SIGMA). Digested tissue was

further dissociated with vigorous pipetting, removed (after allowing large, undigested

pieces to settle) and placed in ice-cold HBSS with 2% FBS (GIBCO/BRL, Grand Island,

NY). Digestion was continued using a mixture of 2 mg/ml collagenase + 100 U/ml DNase

in 5 min incubations at 37C in a shaking water bath and repeated until completion.


U- U F- H U U0
g ^s^ y s

Ys B u o-

U. &
| 8: | i o ^| g I

F- c Y

.1 0
o u C o^C uS

St U 00U0O U U0U
o u iS c ^ c E- ouu^

U US < o

3 U t O u ou y -t U F-

E 2222 2222
Is 111 e gih

i .pUS I G : s
C uip u u ^ ^ U
- y c u^ Y

3 *< S s Z B 4 5
*g u uo b : U ^i ^U

h, U U 0 0 0

e.EeEe 5

Islets were picked manually from digested pancreatic tissue under a dissecting

microscope and collected in ice-cold IMDM (GIBCO/BRL) with 1% NMS (68). Islets

were collected by centrifugation, supernatant discarded and the cells dispersed by

continuous pipetting in 10 ml of a solution containing trypsin/EDTA (GIBCO/BRL) +

2000 U DNase (SIGMA) at 37"C for 10 min. Digestion was stopped by adding ice-cold

IMDM with 1% NMS.

Digested lacrimal, parotid and submandibular tissue and dispersed islets were

washed and then separated by centrifugation through a 55% Percoll (SIGMA) gradient.

The infiltrating cells were collected in the pellet and contaminating red blood cells lysed

with 0.84% ammonium chloride. After washing, cells were divided into aliquots

containing approximately 1 x 106 cells and washed in FACS buffer prior to antibody


Flow Cvtometry

Aliquots of all cell populations were resuspended in 100 il FACS buffer and

stained with antibody at 1 .tg/106 cells. Cells were stained first with anti-CD3 for 40 min

at 4C, washed with FACS buffer and then stained with the appropriate second antibody in

a 40-min incubation at 4C. After a final wash, cells were suspended in FACS buffer for


Flow cytometric analyses were performed using a FACScan flow cytometer

(Becton Dickinson, Mountain View, CA) equipped with a 15 milliwatt, 488nm air-cooled

argon-ion laser and using LYSYSTM II software (68). Ten thousand events were collected

per sample from a population gated on a window encompassing the splenic lymphocyte


RNA Isolation and RT-PCR Detection of Cvtokine mRNA

Pooled tissues were minced in PBS, placed in lysis buffer and mRNA isolated

using a Micro-FastTrack" Kit (Invitrogen). Isolated mRNA was stored at -70C in

ethanol until all samples were collected. mRNA was pelleted by centrifugation and cDNA

prepared by reverse transcription using Superscript IIT Reverse Transcriptase

(GIBCO/BRL). cDNA was quantified using a DNA DipstickT Kit (Invitrogen, San

Diego CA). Equal quantities of cDNA from each sample (50 ng per reaction) were

amplified by PCR for 40 cycles at 60C annealing (1 min) and 72'C elongation (2 min)

using cytokine primer pairs shown in Table 1. PCR products were separated on 1.2%

agarose gels and transferred to positively charged nylon membranes (Boehringer

Mannheim) by Southern blotting (122).

Detection of Cvtokine mRNA

Specific PCR products were identified using the Genius" system of

nonradioactive DNA labeling and detection (Boehringer Mannheim) according to the

manufacturer's protocols. Briefly, internal oligonucleotide probes specific for each

cytokine (Table 1) were labeled by random primed incorporation of digoxigenin-labeled

deoxyuridine-triphosphate. After overnight hybridization at 65"C, the PCR products were

detected colorimetrically using an anti-digoxigenin alkaline phosphatase conjugate in an

enzyme-linked immunoassay. Concanavalin A (SIGMA)-stimulated mouse splenocytes

were used as a positive control for the primers and probes and G3PDH was used as a

positive control for the isolation of nRNA (123). All nucleotide primers and probes were

synthesized in the Interdisciplinary Center for Biotechnology Research DNA Synthesis

Core Laboratory at the University of Florida (Gainesville, FL).

Densitometric Analysis

Semi-quantitative analyses of cytokine PCR products was done using densitometry

and One-Dscan software. Sample calculations were standardized using G3PDH values to

ensure that equal amounts of nRNA were present in each sample thus allowing for semi-

quantitative comparison between sample bands (G3PDH values for 8 wk lacrimal and 10

wk submandibular samples are reported as raw data since they did not appear to be

consistent with experimental samples). Lanes which did not exceed background were

given a value of zero and subsequent sample values were determined using auto-

background calculations.


Tissue Histology

Lacrimal, submandibular, parotid and pancreatic tissues were surgically removed

from groups of 5 NOD mice at 2 wk intervals from 8 through 20 weeks of age. A small

piece of each tissue was stained with hematoxylin/eosin(H&E) and examined for leukocyte

infiltration. As expected, leukocytic infiltration was observed in the pancreatic islets of 8

wk old mice and increased in severity over time. By 18 wks, few islets remained due to

Figure 1 Histological profile of tissues showing lymphocytic infiltrates of the exocrine
tissues and insulitis in the NOD mouse. Tissue sections were stained with
hematoxylin/eosin. (A) 10 wk pancreas; (B) 12 wk submandibular gland; (C) 12 wk
lacrimal gland; (D) 16 wk parotid gland.

,, s

4 G 0

0 U

" _
03 0 0

Figure 3 Histogram of the temporal expression of CD4' and CD8' T-cells in lymphocytic
infiltrates of submandibular (SMG), lacrimal (LAC), parotid (PAR) glands and spleen
(SPL) from NOD mice. Data was obtained from flow cytometric analyses of infiltrating
cells using FITC-conjugated CD3 and PE-conjugated CD4 or CD8. All values are
expressed as a percent of CD3' cells.




autoimmune destruction. At 8 wks of age the submandibular glands of 2 of the 5 mice

showed small focal areas of infiltration while the lacrimal and parotid glands remained

normal. By 12 wks, the lacrimal glands showed infiltration, but their glands had fewer and

smaller lymphocytic foci than the submandibular glands which were heavily infiltrated in

all mice by 14 wks. In contrast, the parotid glands showed no focal lymphocytic

infiltration up to 20 wks of age, although a few mononuclear cells were occasionally seen

in some mice from 14 to 20 wks of age. Representative histological profiles of glandular

infiltrates are shown in Fig 1.

Flow cytometric analysis of glandular infiltrating lymphocytes

Monoclonal antibodies to cell surface molecules have been widely used as

phenotypic markers corresponding to functionally distinct subsets of lymphocyte

populations. We have used flow cytometric analysis to determine the cellular phenotypes

of lymphocytes infiltrating the pancreas, lacrimal, parotid and submandibular glands of the

NOD mouse from the first appearance of lymphocytes at 8 wks through 20 wks, which is

4 wk beyond the detection of exocrine gland dysfunction. Gates set on the NOD spleen

lymphocyte populations were used to select the infiltrating populations of the other tissue


The percentage of the cell populations falling in this gated window remained fairly

constant throughout the time course of the study averaging approximately 73% for spleen,

47% for islets, 59% for submandibular gland, 24% for lacrimal gland and 17% for parotid

gland. Within these gated populations the percentage of CD3' cells also remained

relatively constant over the time course of the study at 48% for spleens, 41% for islets,

37% for lacrimal glands and 39% for parotid glands. However, in the submandibular

gland, there was an increase in CD3* cells from 41% at 8 wks to 70% at 20 wks.

Double labeling of the infiltrating populations with CD3/CD4 or CD3/CD8 is

shown in Fig 2 for the 12 wk time point. Similar profiles were obtained for each time

period to determine the percent of CD4 or CD8 cells (Fig 3). The CD4* population in the

submandibular gland approximated that in the spleen: 68% and 62% respectively. In the

lacrimal gland, the CD4* population was about half that of the spleen at 8 wks (-30%) and

increased over time to the level of the spleen. The parotid gland showed a decrease from

84% CD4' cells at 8 wks to -50% at 20 wks. However, the low total number of

lymphocytes detected in the parotid glands as well as their similarity to splenic profiles

may be indicative of contamination by lymphatic vessels which are interspersed throughout

the gland. There was greater variability in the infiltrating CD8' populations of the

exocrine glands than seen in CD4' cells (Fig 3). Generally, fewer than 10% of the CD3*

population were CD8' in these glands at 8 wks -- considerably lower than the 27% found

in the spleen. As the disease progressed, the percentage of CD8* cells increased in all 3

glands, but it is unclear if the apparent decrease at 20 wks is significant.

As indicated by the spleen population, the ratio of peripheral CD4' to CD8* cells

remained relatively constant at ~2.5:1. In the submandibular gland, the CD4:CD8 ratio

was 6:1 at 8 wks, and decreased to an average of 3.5:1 from 12 wks on. Both the lacrimal

and parotid glands also had higher ratios in early infiltrates, but decreased to spleen levels

at 12 wks.

The CD4* and CD8' populations did not account for all of the CD3' cells. The

CD3', CD4CD8 population accounted for 5% of the spleen cells, 8% of the

0 c c_ _
S 0, fl ^
'- >/] -

r- 0
N r-N
N en 0

Lr 't 0 'C, en en .

3 01 N

.5- N N 'P N
'C N' 00 N 0^ 't' en 0
- 0 "' Ci N r3

I ? 0 N' 00 6 DO In
^- 'C 'C N- N- M *

vI n In In
p 0< ^ r"; C^ \6 I
n v N, -t (

- (t 1 -<). e en
2 re o

N "jN

0 0 o o0 6 0

5 0 L

c 0 5
F- a

_ I I

~ -a- '"' \ 0
a o a o

S- N -

00 0 '0 c o


LLJ ---- -
I I O 00 ''-0

n 0 i 0 N06

'O r N I -C

0 N n -1 rni q 00o
Nl C' 00 00 r'. N 0 00
CI ^- N 't r _

0 U U

+ +
I In

u A

w C, \o


S9 I11 17

Figure 4 Histogram of selective TCR Vp3 distribution in infiltrating lymphocytes isolated
from spleen, pancreatic islet, submandibular, and lacrimal glands of NOD mice. Data from
flow cytometric analyses of 8 through 20 wk samples were averaged for each tissue.
Spleen-solid bar; Pancreatic islet-horizontal striped bar, Submandibular gland-clear bar;
and Lacrimal gland-diagonal striped bar. Data are presented as percentage of CD3' cells
+ SE of the mean.

submandibular cells, 15% of the lacrimal cells and 9% of the parotid cells. This indicates

an increase in the double negative population in the exocrine tissues over that of the


The infiltrating populations of the submandibular and lacrimal glands were further

characterized using antibodies against the B-cell marker B220 (63) and the

activation/memory marker CD45RB (123). As shown in Table 2, the CD3-B220' B-cell

population in the submandibular gland increased from -1% at 8 wks to -15% at 10 wks

and maintained that level throughout the rest of the study. In the lacrimal gland, CD3

B220* cells also started at 1% at 8 wks and increased to -33% at 12 weeks. Whether the

fluctuation seen between 10 and 14 weeks is meaningful is unknown at this time. In

addition, the presence of a CD3'B220* cell population was detected in both of these

tissues. This phenotype did not appear in significant numbers until 16 wks of age in either

gland and never exceeded 10% in the submandibular gland or 5% in the lacrimal gland.

The CD3'B220' phenotype has been reported as a double negative population in MRL/lpr

mice (63) and as a lymphokine activated killer phenotype (LAK cell) in other studies


CD45RB staining presented a much more complex picture (Table 3). In

submandibular glands at 8 wks, -3% of the CD3' cells were CD45RB"' (naive T-cell)

and from 10 wk on, this CD45RB'" population remained relatively stable at -16%. The

percentage of CD3'CD45RB"+ (memory T-cells) population was 15% at 8 wks, showed a

1.8-fold increase to 27% at 12 wks and remained at that level through 20 wks. The CD3

CD45RB"' population in the submandibular gland remained constant at -25% from 10 -

20 wks. In the lacrimal gland, the picture was less clear due to a smaller sample size

(Table 3). In general, the CD45RBhil population (both CD3' and CD3) increased to a

maximum of -55% at 12 wks, plateaued and then decreased at 18 and 20 wks to -24%.

The CD3'CD45RB'* population remained fairly constant at ~20% from 10 wks on.

Analysis of TCR VB usage was done using the monoclonal antibodies listed in the

Materials section. It has been previously reported that the majority of the infiltrating

populations of the submandibular and lacrimal glands are TCRoap (64). In the VPs tested,

there was no appreciable time-related variation in response, therefore, the data from all

time points were averaged for VB distribution (Fig 4). The submandibular and lacrimal

glands showed a similar distribution pattern to that seen in the spleen. Both Vp6 and V38

were significantly increased over background as represented by VB3, while, to a lesser

extent, V39 and Vp17 were also increased. These increases were much more dramatic in

the submandibular gland than in the lacrimal gland which paralleled the response in the

periphery as represented by the spleen.

Cvtokine mRNA expression in salivary and lacrimal tissues

Temporal expression of mRNA transcripts for selected cytokines expressed within

the infiltrates of lacrimal and submandibular tissues from each experimental age group was

determined through the use of semi-quantitative RT-PCR. PCR bands were quantified

using densitometry and compared against PCR bands of G3PDH to provide a

measurement of temporal changes in the production of mRNA transcripts and provide a

detailed analysis of cytokine expression throughout the progression of the autoimmune

activity. As presented in Fig 5, mRNA transcripts for a number of interleukins were

detected generally at both an earlier age and with greater intensity in the lacrimal glands




I-S Ij*




Figure 5 Interleukin mRNA expression of lacrimal and submandibular glands as
determined by RT-PCR and Southern blotting. Blots were scanned and analyzed by
densitometric comparison using G3PDH as a standard. Values were plotted by arbitrary
density units and analyzed using a best-fit polynomial or exponential trendline.



~y"-~ ~1IIII



T .




Figure 6 Proinflammatory mRNA expression of lacrimal and submandibular glands as
determined by RT-PCR and Southern blotting. Blots were scanned and analyzed by
densitometric comparison using G3PDH as a standard. Values were plotted by arbitrary
density units and analyzed using a best-fit polynomial or exponential trendline.

than submandibular glands. In the lacrimal tissue, IL-Il, IL-5, IL-6, IL-7, and IL-10

mRNA transcripts were detected as early as the 8 wk as well as at most later time points.

In the submandibular gland infiltrates, detectable interleukin mRNA transcripts were

usually not observed before 14 wk of age. Of note, mRNA transcripts for IL-4 were

absent from lacrimal glands, while IL-4 and IL-5 were absent from the submandibular

glands. Both interleukin transcripts were detected in control mRNA obtained from ConA

stimulated NOD splenocytes, while IL-4 is commonly seen in islet infiltrates (122). mRNA

transcripts for three effector cytokines, IFNy, TNFa and iNOS were detected in both

lacrimal and submandibular gland infiltrates of mice aged 8 and 12 wks, respectively (Fig

6). All three cytokines exhibited increased levels of mRNA transcription through 16-18

wks. Cytokine mRNA transcripts were rarely detected in control parotid tissues.

Considered in lolo, these results indicate that major temporal changes in the cytokine

profiles occur in the lacrimal and submandibular glands of NOD mice between 12 and 16

wks of age. In the case of the submandibular glands, this is several weeks following the

first appearance oflymphocytic infiltration.


Overt diabetes in our NOD mouse colony generally begins at 12 wks of age, while

the first appearance of focal lymphocytic lesions in the submandibular and lacrimal glands

begins at 8-10 wks and 10-12 wks of age, respectively. Despite the early appearance of

lymphocytic infiltrates within the submandibular and lacrimal glands, loss of saliva flow

and tear production does not occur until 14-16 wks of age (71). For this reason, the

argument has been advanced that the autoimmune attack against the salivary and lacrimal

tissues is merely secondary to the loss of immunological tolerance for a pancreatic 0 cell

antigen also expressed on the salivary and lacrimal glands. A second argument suggests

that the decreased saliva and tear flow is a consequence of the loss of blood glucose


Several recent studies have raised serious questions with both arguments. First,

Leiter et al. (101) have reported that development of immunological tolerance against the

P cell following intrathymic injection of islet cell homogenates into neonatal NOD mice

prevented the development of diabetes, but not the autoimmune attack against the salivary

glands. Second, Garchon et al (81) have linked sialoadenitis, hyper-IgG production and

peri-insulitis to a telometric locus, but insulitis and diabetes to a centromeric locus on

chromosome 1. Third, I have observed (unpublished data) that loss of secretary function

in the salivary glands proceeds the onset of changes in blood glucose levels and diabetes in

NOR/chr 9 mice, a strain which has delayed onset diabetes (85). These studies indicate

diabetes and salivary/lacrimal gland complications in NOD mice are most likely distinct


In the present study, I have attempted to define the temporal development of the

autoimmune attack against the salivary and lacrimal glands of NOD mice. A unique

feature of this pathogenesis, and comparable to Sjogren's Syndrome in humans, is the loss

of secretary function resulting in clinical presentations of xerostomia and xerophthalmia

(71,11). As such, I believe these data will allow for a comparison of the autoimmune

processes, e.g., lymphocyte phenotypes, activation states and cytokine profiles, within

several organs of the same host as well as a comparison between this animal model and

Sjogren's Syndrome.

Analyses of the lymphocytic infiltrates of the salivary and lacrimal glands

uncovered a relatively high percentage of CD4' T-cells, as might be expected from

previous studies on the islet infiltrating lymphocyte populations (68). In contrast, only

10% of the T-cells detected were CD8', and even fewer were CD3'/CD4"CD8" double-

negative T-cells. While generally accepted that CD4* T-lymphocytes are necessary for the

immunopathogenesis of diabetes in the NOD mouse, the role of the CD8' population,

particularly in the initiation of the disease, has remained more elusive (92). T-cell transfer

studies in NOD mice have confirmed the necessity of CD4* T-cell populations to transfer

autoimmunity, whereas neither purified CD8' T nor B-cells were capable (85).

It is becoming increasingly apparent that the activation state of transferred

lymphocyte populations is critical. Using the CD45RB as a marker of lymphocyte

activation, 1 observed that the majority of T-cells present in the infiltrates of both

submandibular and lacrimal glands were either of a memory (CD45RB') phenotype or a

phenotype (CD45RBh") suggesting transition to an activated stage. The presence of a

distinct CD45RB" population throughout the time course of this study suggests that a

population of naive T-cells are being actively recruited to the salivary and lacrimal glands.

Interestingly, a recent study has shown that the ability to transfer diabetes in NOD mice

is associated with a CD45RB'" memory T-cell population and a concomitant increase in

the 1FNy to IL-4 ratio (87). T-cells that produce L-4 have consistently failed to transfer

disease. Markedly increased levels of IFNy mRNA transcripts were observed in both the

submandibular and lacrimal glands, possibly indicating a switch in the cytokine production

by activated CD45RBI T-cells.

In addition to exploring the lymphocytic activation states, I have also investigated

selected Vo TCR phenotypes of the infiltrating cells. While most VD TCR populations

were detected in the salivary and lacrimal gland infiltrates (as expected in the NOD

mouse), an increased percentage of Vp6 and Vp8 expression was present, consistent with

previous observations of preferential TCR usage observed immunohistochemically (89).

The importance of V38 and Vp6 lymphocytes in the development of autoimmunity in both

NOD and MRLlpr mice has been demonstrated through the use of both T-cell transfer

and specific anti-Vp therapy (64,65,125).

Differences in the numbers of B lymphocytes infiltrating the salivary and lacrimal

glands were also observed: approximately 15% in the submandibular gland, but 33% in

the lacrimal tissue. Whether the increased percentage of B lymphocytes within the lacrimal

glands is due to increased B-cell proliferation or active recruitment remains unknown.

However, detection of CD23 (blast-2 antigen) in the lacrimal infiltrates of Sjogren's

Syndrome patients support the possibility of lymphoproliferation (34).

In this study, I have also documented several age-dependent increases in cytokine

mRNA expression in the salivary and lacrimal glands. Of particular interest is the

increased expression of IFNy, TNFa, and iNOS in both salivary and lacrimal tissues

starting at 12-14 wks of age and increasing to maximally detected levels by 20 wks of age.

By 16 wks, high levels of L-10 mRNA production was detected as well. The dramatic

rise of each of these transcripts coincides with the time of the first evidence of the loss of

saliva flow as reported in previous studies (67). Whether these cytokines are involved in

effecting loss of exocrine function directly via a cytotoxic pathway or their presence is

necessary to initiate lymphocyte-mediated cell killing is not yet known.

Many recent studies have focused on the potential roles of cytokines as cytotoxic

effectors of exocrine tissue destruction. Evidence indicates that IL-1, IFNy, TNFt and

nitric oxide may be key mediators in the pathogenic process of islet cell destruction in both

humans and NOD mice (126-129). Of particular interest, production of high levels of

IFNy by islet infiltrating cells has been repeatedly demonstrated in NOD mice and appears

tightly linked to P cell destruction (127,130). Recent studies by Mathis et al.

demonstrated that an islet-specific Thl T-cell clone producing high levels of IFNy is able

to rapidly transfer diabetes to young NOD recipients (88). However, when shifted in vitro

to an IL-4 producing phenotype, the same T-cell clone was unable to transfer diabetes.

Furthermore, IFNy has been noted to induce cell death of both islet and salivary gland

cells in vitro, and may represent a non-specific mediator of exocrine cell destruction in

vivo (45). However, a caveat to this argument is the detection of high levels of IFNy and

TNFa in the salivary glands of MRL/lpr mice which do not lose secretary function


Unexpectedly, cytokine mRNA expression in the lacrimal glands typically appeared

at an earlier age and in larger quantities than in the submandibular gland. This occurred

despite the fact that both fewer and smaller focal lymphocytic lesions were present in the

lacrimal glands in the earlier age groups, thus suggesting that lymphocytes infiltrating the

lacrimal tissue are activated at an earlier stage than those in the submandibular gland.

Alternatively, a lag time between the detection of focal lymphocytic lesions in the

submandibular gland (8-10 wks) and the detection of increased cytokine mRNA

expression (12-16 wks) may indicate that the submandibular gland-infiltrates remain

functionally quiescent and require a signal of activation.

In conclusion, this data demonstrate numerous similarities in both lymphocyte

phenotypes (CD4/CD8 ratios, V3 TCR restriction, and B220 populations) and cytokine

expression (increased IL-Il, IL-2, IL-10, TNFa and IFNy) between the NOD mouse and

other animal models for autoimmune sialoadenitis as well as Sjogren's Syndrome.

Although the initiating agent for Sjogren's Syndrome remains unknown, it is often

believed that extrinsic factors, i.e., viral agents, may be responsible for the breakdown of

tolerance in immunologically susceptible individuals. In NOD mice, however, the

necessary intrinsic elements for the breakdown of salivary gland tolerance exist in their

genetic background. Using the NOD-scid mice, I have recently described multiple salivary

gland abnormalities of NOD-scid mice that do not appear to be immunologically related

(121). These alterations are detectable starting at 8-10 wks of age and include

morphological abnormalities, aberrant gene expression, and increased proteolytic activity.

While any number of immunological changes can result in aggregation of lymphocytes in

exocrine tissues, such as the Ipr/ gld mutation of MRL mice, graft vs. host models (131),

TGF-P knockouts (61), and bcl-2 overexpression (18), only with the appearance of

developmental defects or extrinsic destruction of the exocrine tissue leads to the

development of secretary dysfunction through loss of immune regulation.



The recent development of the NOD-scid congenic strain provides a unique model

to investigate the role of the immune response in the pathogenesis of Sjogren's Syndrome

(82). The NOD-scid mouse is homozygous at the scid (severe combined

immunodeficiency) locus and thus lacks functional T and B lymphocytes. The scid

mutation prevents the spontaneous development of sialoadenitis, insulitis, and diabetes in

these mice; however, the transfer ofT lymphocytes from diabetic NOD mice to recipient

NOD-scid mice can restore an autoimmune phenotype (85). In addition, because of its

NOD background, the NOD-scid has impaired NK (natural killer) cell and reduced

complement activity (82). Since NOD-scid mice share the same NOD genetic

background, this model is ideal for studying the non-immune genetic factors that

contribute to the development of autoimmunity. Previously, the dramatic changes in

exocrine gland histology, protein synthesis, and secretary dysfunction are thought to be a

direct result of the autoimmune lymphocytic component (67). By investigating exocrine

gland function in the absence of functional lymphocytes, this specific aim demonstrates

that temporal changes in salivary gland function occur in the absence of overt

autoimmunity. This indicates that the resulting immune response may actually be

triggered by aberrant physiological changes in exocrine gland homeostasis and function.

Materials and Methods


BALB/cJ, C3H, CBA, and NOD/Uf mice were bred and maintained in the

Department of Pathology's mouse facility (University of Florida, Gainesville, FL).

NOD-scid mice were bred and housed in the Department of Pathology's transgenic mouse

breeding colony. Both male and female mice ranging in age from 3 to 30 weeks were

used. Maintenance of the scid mutation was assessed in experimental animals by flow

cytometry of spleen cells and RT-PCR analysis of CD4, CD8, and TCR VP repertoire

from isolated submandibular gland total RNA. NOD mice were routinely tested for blood

glucose levels using Chemstrip bG reagent strips (Boehringer Mannheim, Indianapolis,

IN). Consecutive elevated fasting blood glucose levels >240 mg/dl were considered onset

of diabetes, after which time, the mice were maintained on daily insulin injections.

Saliva Collection and Flow Rate

Saliva was collected from control and experimental groups of male and female

mice following stimulation of secretion using isoproterenol (0.20 mg/100 g body weight)

and pilocarpine (0.05 mg/100 g body weight) (Sigma Chemicals, St. Louis, MO) dissolved

in saline. The secretogogue cocktail was injected (0.1 ml volume) intraperitoneally and

saliva was collected, starting 1 min post-injection, for 10 min from the oral cavity by

micropipette and placed into chilled 1.5 ml microcentrifuge tubes (67). Volume was

determined by measurement with 200 pl micropipettes. Saliva samples were collected

from groups of 7 male or 5 female mice then frozen at -700C until analyzed for temporal

protein changes by enzyme assay, radio-receptor assay, SDS polyacrylamide gels and

Western blotting

Total Salivary Protein and a-Amvlase Analysis

Saliva samples were analyzed for total protein using bovine serum albumin as the

standard (132). Amylase was determined by its ability to hydrolyze starch according to

published protocols (133). In brief, 500-1000 fold dilutions of saliva in

phosphate-buffered saline (PBS) were added to a solution containing 0.4 g soluble starch

in 60 mM tris(hydroxymethyl)aminomethane (Tris)-HCI, 0.15 M NaCI and 3 mM CaCI2.

The reaction was stopped after 5 or 10 min by the addition of 0.045% 12, 0.045% KI, and

0.03 N HC1. Absorbance was measured at wavelength 620 nm. One unit of amylase was

defined as the amount that hydrolyzed 1 mg starch/min/mg protein at 370C.

EGF Analysis

Salivary epidermal growth factor (EGF) was estimated by the procedure of Booth

et al. (134). Saliva was diluted 10 fold in PBS containing 0.2 mg/ml bovine serum

albumin (BSA). Reactions consisting of 100 pl diluted saliva, 100 pl human placental

microvilli membranes and 100 pl 1251-labeled human EGF were incubated for 24 hr at

4C (67). Each reaction was then diluted with 3.5 ml ice-cold 0.1% BSA in PBS solution,

centrifuged for 20 min at 7,000 g in an RC-3B Sorvall centrifuge and the quantity of

radiolabel associated with the membrane pellet determined using a Beckman gamma

counter. The concentration of EGF was compared to a standard curve generated with

dilutions of known quantities of human recombinant EGF. Membrane binding competition

is independent of species origin for the source of EGF (135).

Isolation of Salivary Gland Tissues.

Parotid and submandibular glands were excised from mice killed by cervical

dislocation. Each gland was freed of connective tissue, fat, and any lymph nodes, then

homogenized in 10 mM Tris buffer (Ph 7.4) containing 100 pM phenylmethylsulfonyl

fluoride, 1 liM leupeptin, and 100 lM benzamidine. The slurry was then centrifuged at

100,000 g for 30 min to recover total membrane (67). The supernatant was saved and

frozen until analysis for a-amylase activity.

Polvacrylamide gel electrophoresis and Western blot analysis

Total saliva proteins (5 pg of total protein or 5 pl of total saliva volume) were

subjected to electrophoretic separation on a 1.5 mm thick 10% SDS-polyacrylamide gel

(12% SDS for Western blots) using a modified Tris-Glycine system of Pugsley and

Schnaitman (136). Western blots of gland lysates contained 20 lg protein per lane. Gels

were fixed and stained using Coomassie Brilliant Blue R-250 or transferred to

Immobilon-P membranes (Millipore, Boston, MA) for 2 hr at 70v for Western blotting

(137,138). The blocking buffer consisted of 3% nonfat dry milk and 3% BSA in

Tris-buffered saline. Polyclonal rabbit anti-mouse parotid secretary protein (mPSP) IgG

antibody (139), kindly provided by Dr. William Ball (Dept. of Anatomy, Howard

University) or polyclonal rabbit anti-rat proline rich protein (140), kindly provided by Dr.

David Castle (Dept. of Anatomy and Cell Biology, University of Virginia), was incubated

with each membrane for 2 hr at 250C. Following three 10 min washes, the membranes

were incubated in alkaline phosphatase conjugated goat anti-rabbit immunoglobulin

(Sigma Chemical Co.) and exposed to substrate as previously described (95).

Protein Sequencing

Salivary proteins were subjected to electrophoresis on 10% SDS-polyacrylamide

gels as described above and transferred to Immobilon-P membranes. Selected protein

bands were carefully cut from the membranes and subjected to N-terminal sequencing

using Applied Biosystems Model 470A Gas Phase Protein Sequencer with Model 120A

on-line PTH analyzer (University of Florida ICBR Protein Sequencing Core Laboratory).

Protein sequences for the first 12 amino acids of each protein were entered into a protein

database for comparison with known protein sequences.


Freshly excised submandibular glands were placed immediately into 10% PBS-

buffered formaldehyde (pH 7.2). Each tissue was embedded in paraffin, sectioned, then

stained with hematoxylin/eosin dyes (UF Diagnostics Referral Laboratory, Gainesville,

FL). The stained sections were viewed using light microscopy.

Statistical Analyses

All measures of variance are given as standard deviations of the mean. Tests of

significance for differences between independent means were performed with the unpaired

Student I-test. Results in which p < 0.05 were considered significant.


Analysis of NOD-scid Saliva

To determine the impact on salivary function of the scid mutation in the NOD

mouse, I analyzed whole saliva samples from individual NOD-scid mice for total volume,

total protein, amylase activity and EGF concentration. Saliva from 10-12 wk old

NOD-scid mice, an age at which salivary glands in NOD mice are devoid of detectable

lymphocytic infiltration, was compared to that of >20 wk old NOD-scid mice, an age

when lymphocytic infiltration is present in the salivary glands of NOD mice (Table 4).

While there was no significant age or sex differences observed for salivary flow rates

(P>0.05), significant sex-related differences were observed for protein content (P<0.05).

Saliva from female NOD-scid mice contained approximately 40% less total protein than

that of male mice in both experimental groups.

Saliva from NOD-scid mice were analyzed for two proteins, EGF (a product of

submandibular gland ductal cells) and amylase (a product of parotid gland acinar cells),

normally secreted in high quantities. As presented in Table 4, the quantity of EGF in the

saliva from >20 wk old male NOD-scid mice showed a 21% decline compared to the

amount present in 10-12 wk old male mice (P<0.002). A comparison of amylase activity

+1 +1 +1 2 C14
> o o" ? o_
1= *
+1+1 +1 + Z

SE 0
E o


Wa m a
^~ *-


8s *
*d a r x

po C V) 00 "T
E o6 -; 6a r -' m- .
0 -6b +I i 1 + + + 13
--Z N 00 g

+1 +1 +1 +1 +1 +1

1y 0 0

+1 +1 +1 +1 +1 +1
F cI 0 QC

(-'I CC> d M B

a E- *
-s S

.h ^^^ .6 |i
Ig _^ (M 8 cn ro3mSc

(^ rc > EE
00 2 s 2 ?| ;

3 I I
'-4 A

2 ]- 10 wkNOD-scid
,> Q] 20wk NOD-scid
S E -U 20 wk BALB/c

E 50-

E -

Parotid Submandibular

Figure 7 Cytoplasmic amylase activity of parotid and submandibular glands.
Homogenates of parotid and submandibular glands were centrifuged at 100,000 x g.
Lysates prepared from the glands of 10-12 wk old NOD-scid (open bars), >20 wk old
NOD-scid (hatched bars), and 20 wk old BALB/cJ mice (solid bars) were tested for
amylase activity in a starch hydrolysis assay. Values are expressed as means of 4
experimental animals performed in triplicate : the standard deviations.







STD 6 wks 10 wks 15 wks 20 wks

Figure 8 Temporal changes in the protein profiles of saliva from NOD-scid mice. Saliva
was collected from 6 wk old NOD-scid (Lanes 1-3), 10 wk old NOD-scid (Lanes 4 and
5), 15 wk old NOD-scid (Lanes 6 and 7), and 20 wk old NOD-scid mice (Lanes 8 and 9)
following isoproterenol/pilocarpine stimulation. Salivary protein (5 p-g protein in 35 pl
SDS-PAGE sample buffer containing p-mercaptoethanol) was loaded on each lane and
separated by electrophoresis through a 10% polyacrylamide gel. The gel was stained with
Coomassie Brilliant Blue R-250. Prestained molecular weight markers (STD) were: 96
kDa, Phosphorylase B; 68 kDa, bovine serum albumin; 45 kDa, ovalbumin; 32 kDa,
carbonic anhydrase; 24 kDa, soybean trypsin inhibitor.



E 9s
uj u


r- e

5 NO

z z z
> > >





E N C\


in the saliva of old versus young male and female NOD-scid mice (Table 4) also revealed

significant decreases over time. Again, no sex-related differences in amylase activity were

observed, with both male and female NOD-scid mice showing approximately 50%

declines in activity between the two age groups (P<0.0002).

Analysis of Ctoplasmic Amylase

The temporal decline in amylase activity present in saliva of NOD-scid mice was

also reflected in the amylase activity present in parotid and submandibular gland lysates

(Fig 7). Cytoplasmic fractions from parotid gland lysates showed a 50% reduction in

amylase activity in the >20 wk old NOD-scid mice when compared to those of 10-12 wk

NOD-scid mice and a 90% reduction when compared to levels obtained from parotid

gland lysates of BALB/c mice. As expected, submandibular gland lysates exhibited low

levels of amylase activity in all experimental groups.

Temporal Changes in NOD-scid Salivary Proteins by SDS-PAGE

To evaluate temporal changes of salivary proteins in NOD-scid mice, we

compared electrophoretic separations in SDS-polyacrylamide gels of three individual

saliva samples collected from 6 wk old female NOD-scid mice, plus two samples collected

from each of 10 wk, 15 wk, and 20 wk old female animals (Fig 8). Although the total

protein concentration in each saliva sample is constant, the gel profiles revealed substantial

differences in protein composition between animals of the various age groups. These

protein profiles were similar whether constant protein (5 gg) or constant volumes (5 il)

were applied to the SDS-PAGE. The temporal decline (then precipitous disappearance

from the saliva samples of 20 wk old NOD-scid mice) of the protein observed at

approximately 55 kDa, shown to be amylase (105), is consistent with reported changes

observed for salivary amylase in NOD mice when analyzed on SDS-PAGE gels.

Two additional protein bands, one at 32 kDa and one at 20 kDa (indicated by

arrows in Fig 8), present in relatively high amounts in saliva of NOD-scid mice of 10 wks

of age or less, virtually disappear from saliva of NOD-scid mice 15 wks of age and older.

Interestingly, the decline of these two proteins appeared to correlate with the emergence

of a third protein band at approximately 27 kDa (indicated by an asterisk in Fig 8). By 20

wks of age, this 27 kDa protein band appears to be one of the most abundant proteins in

the salivary samples.

Sequence Homology of the 20 kDa. 27 kDa and 32 kDa Proteins with Parotid Secretory


N-terminal amino acid sequence analyses of the 20 kDa, 27 kDa and 32 kDa

protein bands indicated that all three were homologous to parotid secretary protein (PSP)

(Fig 9). Interestingly, the N-terminal amino acid residues of both the 32 kDa and 20 kDa

proteins, the two isoforms observed prior to the time NOD mice present with

sialoadenitis, were not only identical to each other, but were also identical to the published

murine PSP sequence beginning at the signal cleavage site of the secretary form (108;141-

142). In contrast, the N-terminal sequence of the 27 kDa protein, appearing at a time

NOD mice begin exhibiting sialoadenitis, matched an internal portion of the PSP starting

at the 27th amino acid after the signal cleavage site. The secretary form of PSP has been

reported to be a 20 kDa, leucine-rich glycoprotein that is abundant in the secretary fluids

of the parotid gland (105).

Antigenic Cross-reactivity of Salivary PSP from NOD-scid. NOD and BALB/c Mice

To confirm the identity of the 20 kDa, 27 kDa and 32 kDa protein bands as PSP,

Western blots of SDS-PAGE separated saliva from 10 wk old NOD-scid, 20 wk old

NOD-scid, diabetic NOD and normal BALB/c mice were treated with anti-PSP polyclonal

antibody. As shown in Fig 10, anti-PSP antibody detected each of the three isoforms.

Saliva from BALB/c and 10 wk old NOD-scid mice contained predominately the 20 and

32 kDa isoforms, while saliva from diabetic NOD contained the 27 and 20 kDa isoforms

and lacked the 32 kDa isoform. Saliva from 20 wk old NOD-scid mice contained each of

the isoforms, but had greatly reduced levels of both the 20 and 32 kDa isoforms compared

to the BALB/c and 10 wk NOD-scid mice.

PSP Detected in the Cytoplasmic Fractions of Salivary Glands of NOD-scid Mice

Cytoplasmic lysates of the submandibular and parotid glands were analyzed on

Western blots for the presence of PSP using anti-PSP antibody (Fig 11). Lysates of

submandibular glands from 10 wk old NOD-scid mice revealed the presence of all three

PSP isoforms, and each isoform increased in quantity by 20 wks of age (Lane 2 versus

Lane 3). No PSP was present in the lysates of BALB/c submandibular glands (Lane 1).

Lysates of parotid glands from 10 wk old NOD-scid mice contained the 32 kDa

protein band plus a comparatively low level of the 20 kDa protein (lane 5). In contrast, by

26 wks of age, parotid gland lysates from NOD-scid mice exhibited all three isoforms

3 3

0o o o

S 0 < 0 0
Z z z





Figure 10 Identification of parotid secretary protein in saliva using polyclonal anti-PSP
antibody on Western blots. Protein (5 gg/lane) in whole saliva from 20 wk old diabetic
NOD (Lane 1), 20 wk old BALB/c (Lane 2), 10 wk old NOD-scid (Lane 3, and 20 wk old
NOD-scid mice (Lane 4) was separated using SDS-PAGE, electroblotted to Immobilon-P
membranes, and reacted with rabbit anti-mouse PSP antibody. The Western blots were
then developed using alkaline phosphatase-conjugated anti-rabbit IgG plus substrate. The
32 kDa and 20 kDa bands are indicated with arrows and the 27 kDa protein with a star.
Prestained molecular weight markers are 45 kDa, ovalbumin; 32 kDa, carbonic anhydrase;
and 24 kDa, soybean trypsin inhibitor (Lane S).

Submandibular Parotid

s o o o o

z Z E Z


32 O Q


Figure 11 Representative Western blot analysis of cytoplasmic fractions of submandibular
and parotid cell lysates using anti-PSP polyclonal antibody. The cytoplasmic proteins (20
gg/lane) of either submandibular gland lysates prepared from 26 wk old BALB/c (Lane 1),
10 wk old NOD-scid (Lane 2), and 26 wk old NOD-scid mice (Lane 3) or parotid gland
lysates prepared from 26 wk old BALB/c (Lane 4), 10 wk old NOD-scid (Lane 5), and 26
wk old NOD-scid mice (Lane 6) were separated on a 12% polyacrylamide gel and
transferred to Immobilon P membranes. Membranes were treated with alkaline
phosphatase-conjugated rabbit anti-mouse PSP and substrate. The 32 kDa and 20 kDa
protein bands are indicated with arrows, while the 27 kDa protein band is shown with a
star. Prestained molecular weight markers are the same as for Fig 9.

AI a s ,t

53.2 -

34.9 -- .

28.7 -

20.5 -


I 2 c O 2
< < ) < 2 oo
IL U a (0 a M z 0

0. Q 0 0 1 z a
l ZZ 0

53.2 -
34.9 -- i

20.5 -
Figure 12 Western blot analysis of saliva and cytoplasmic fractions of parotid and
submandibular cell lysates using anti-Proline-Rich Protein polyclonal antibody. Panel A,
10% SDS-PAGE of constant saliva protein of 5 pg from mouse strains C3H, BALB/c,
CBA, NOD, and NOD-scid. P-DM, prediabetic NOD; DM, diabetic NOD. Panel B,
represents the antibody staining of PRP in parotid (PAR) and submandibular (SM) gland
lysates from BALB/c, P-DM and DM NOD, and NOD-scid mice. Each lane contained 20
pg of protein from total cell lysate.

Figure 13 Morphological changes in the salivary glands of NOD-scid mice.
Hematoxylin/eosin-stained tissue sections of submandibular glands from 10 wk old NOD-
scid (A: 100X, B: 200X), 30 wk old NOD-scid (C: 100X, D: 200X) and 20 wk old NOD
mice (E: 100X, F: 200X).

(lane 6). Again, lysates ofBALB/c parotid glands contained only the 32 kDa protein (lane

4), comparable with the profile of young NOD-scid mice. Interestingly, the anti-PSP

antibody bound to a protein band of approximately 65 kDa present in the parotid gland

lysates of both BALB/c and 10 wk old NOD-scid mice; however, this protein band has

not been studied. These results show the temporal nature of the appearance of the 27 kDa

isoform of PSP in the salivary glands and the abnormal expression of PSP by

submandibular glands of NOD-scid mice.

Detection of Proline-Rich Protein (PRP) in the Saliva and Salivary Glands of NOD and
NOD-scid mice

The PRPs are a set of proteins whose synthesis are induced in the salivary glands

of mice following chronic p-adrenergic agonist treatment or by the introduction of

deleterious dietary changes (104). Saliva from NOD-scid, diabetic or prediabetic NOD

mice, as well as BALB/c, CBA, and C3H strains were evaluated by Western blot for the

presence of PRP. As indicated in Fig 12A, antibody to rat PRP was able to cross-react

with a protein of approximately 33 kDa in saliva of diabetic NOD, 10 wk and 25 wk

NOD-scid mice. In contrast, very little of this protein was detectable in the saliva from

control BALB/c and prediabetic NOD mice. No reactivity was detected in the saliva of

C3H or CBA mice. N-terminal amino acid sequence analysis confirmed the identity of this

protein as PRP (data not shown). An examination of parotid and submandibular gland

lysates revealed that PRP expression was occurring in both glands of the NOD-scid,

prediabetic, and diabetic NOD mice, but not in control BALB/c animals (Fig 12B).

Histological Changes in the Submandibular Glands of NOD-scid mice

Histological sections of submandibular glands from 15 individual NOD-scid mice

ranging in age from 3 to 30 wks were evaluated for structural changes. Figure 13 shows

hematoxylin/eosin stained sections of submandibular glands from 10 wk (Fig 13A & 13B)

and 30 wk (Fig 13C & 13D) old female NOD-scid mice. A comparison of these two time

points reveal a marked, progressive loss of acinar tissue and a decline in the acinar to

ductal cell ratio. The histology of the parotid gland was also examined. However, there

were no observable changes in the cellular or glandular morphology between 10 wk and

30 wk NOD-scid mice (data not shown) as was seen in the submandibular gland. As

expected, the NOD-scid tissue displayed no signs of immune cell infiltration. In contrast,

submandibular gland tissue from diabetic NOD mice, which appear to have only slight loss

ofacinar tissue, is highly infiltrated with mononuclear cells (Fig 13E & 13F).


The clinical symptom of xerostomia, or oral dryness, associated with human

Sjogren's Syndrome is commonly attributed to two problems: the loss of the fluid phase

of saliva and changes in salivary protein composition (1). Both problems are considered a

result of a progressive autoimmune response against the salivary glands. This hypothesis

is supported by studies using the NOD mouse model in which changes in both salivary

flow rates and protein composition are associated with the appearance of immune cells in

the salivary glands (67). Unfortunately, these studies in man and the NOD mouse fail to

determine if the loss of salivary function is a direct result of an autoimmune attack or if the

autoimmunity is in response to specific changes in the exocrine glands.

In the present study, I have attempted to dissociate immune and non-immune

factors which may contribute to the loss of salivary gland function by using the NOD-scid

mouse. As expected, histological analyses of salivary glands from NOD-scid mice did not

reveal the presence oflymphocytic infiltrates. Furthermore, total salivary flow and protein

concentration measured in NOD-scid mice appeared similar to the published values of

BALB/c mice and prediabetic NOD mice (67). The composition of saliva proteins, as

shown by EGF and amylase concentration in NOD-scid saliva, however, showed

significant changes with age, whereas the levels detected in BALB/c saliva remained

relatively constant over the same time period (67,71). This was further reflected by

greatly reduced amylase activity (90% reduction) in parotid gland lysates of NOD-scid

mice as compared to BALB/c. In addition, amylase activity detected in NOD-scid parotid

lysates declined nearly 50% between 10 and 25 wks of age. Temporal analyses of

NOD-scid saliva by SDS-PAGE again revealed the age related decline of amylase as well

as several additional proteins. Age-related changes in marine saliva volume and

composition in non-NOD strains does not typically occur until after 12 months of age


The most striking findings, however, were the discovery of a novel expression of

PRP and an internally cleaved PSP isoform (27 kDa) which was prominent in 15 wk

NOD-scid saliva but was not detected in saliva from BALB/c mice or younger NOD-scid.

The abundant appearance of these proteins may explain why total saliva protein

concentration remained constant despite the decline in other major proteins. Proline-rich

protein is a latent constituent of murine saliva which is induced in response to glandular

trauma (104). In normal mice, PSP is a 20 kDa, leucine-rich glycoprotein of unknown

function that is secreted predominately by amylase producing acinar cells of the parotid

gland (105). Studies have shown a developmental coordination of murine PSP and

salivary amylase expression in the parotid gland of adult mice where the two proteins

appear in constant ratios (107,109). With the onset of diabetes in NOD mice (>14 wks)

and the aging of NOD-scid mice (>15 wks), both the 32 kDa and 20 kDa isoforms of PSP

are replaced with a 27 kDa protein band detected with the anti-PSP antibody. Sequence

analysis of the N-terminal amino acid residues of the 27 kDa isoform revealed that the

protein started at an internal region (+27 aa) of the PSP protein. Interestingly, all three

isoforms of PSP appeared in the lysates of the submandibular glands of older NOD-scid

mice but not of normal BALB/c mice. These findings suggest that, first, the glandular

specificity of PSP gene expression is lost over time in NOD and NOD-scid mice, and

second, the transition to a new isoform of PSP might involve differential splicing and/or

alteration ofpost-translational modifications.

Taken together, these findings suggest that both submandibular and parotid gland

function, as shown by alterations in gland specific PSP and PRP gene regulation and

salivary EGF and amylase concentrations are altered in NOD-scid animals between 10 and

25 wks of age. Therefore, changes in saliva protein content in the NOD-scid mice are

probably not due to the insufficiencies of an individual gland but are multiglandular and

potentially affecting other exocrine gland function. The low level of detection of amylase

activity in parotid gland lysates, ectopic expression of PRP, and the appearance of a novel

isoform of PSP in both the parotid and submandibular glands further suggest that these

changes take place at the intracellular level.

Histological examination of the submandibular gland, but not the parotid gland, of

aging NOD-scid mice revealed a remarkable decline in the acinar to ductal cell ratio. This

may be due to acinar cell loss, the hyperproliferation of ductal cells, or a combination of

both. Acinar cell loss in the submandibular gland may be triggered by the observed

defects in protein synthesis or processing which could potentially lead to cell death. Since

whole salivary flow rates in the older NOD-scid animals do not decline despite this

apparent loss of submandibular acini, it is possible that the parotid, sublingual, and minor

salivary glands in the oral mucosa have increased salivary output in order to compensate

for this loss of acinar cells (143,144). An interesting alternate hypothesis to acinar cell

loss is the hyperproliferation of the submandibular ductal cells, which are known to have

self-renewing capacity. Hyperplasia and metaplasia of the salivary ductal epithelium is a

hallmark feature seen in labial salivary biopsies of human Sjogren's patients (15). It should

be noted too that abnormal proliferation of differentiated tissue is often accompanied by

the loss of differentiated cell protein synthesis and function (145). This hypothesis is

especially attractive in light of studies showing that the ductal cells of the NOD exocrine

pancreas can undergo hyperproliferation (146,147). Therefore, it is altogether possible

that similar developmental abnormalities are occurring in the submandibular gland and

pancreas of NOD mice that may precede the development of sialoadenitis and insulitis

respectively, and be inherently involved in the pathogenesis of these autoimmune lesions.

The de novo production of the 27 kDa PSP isoform in aging NOD-scid mice

suggests that aberrant proteolytic processing may play a role in the generation of

otherwise hidden cryptic antigens, priming the immune system for an autoimmune

response. The time at which this new isoform of PSP appears in the saliva and

submandibular glands of NOD-scid mice coincides with the time of appearance of

lymphocytic infiltrates in the salivary glands of NOD mice. In addition, morphological

changes in the salivary glands of NOD-scid mice involving loss of acinar tissue are

detectable by histology in the absence of lymphocytic infiltrates. These observations

suggest that changes in the salivary glands of NOD mice occur independently of

lymphocytic infiltration and that development of the associated autoimmune activity may

actually occur in response to these changes. This is consistent with the proposed model

for the induction of human Sjogren's Syndrome (1), except that the initial "injury" to the

gland is intrinsic i.e. genetically programmed in the NOD mouse rather than extrinsic i.e. a

consequence of viral infection or other insult.

A newly emerging model of pathogenesis of autoimmune sialoadenitis in the NOD

mouse suggests that the initial trigger may reside in a defect in salivary gland homeostasis,

leading to the production of new or altered proteins. As antigenic forms of cellular and

secretary proteins are released, the immune system responds by initiating the homing of

immune cells to the exocrine tissue. Production of cytotoxic cytokines and direct cell

killing by activated T cells may play a role in furthering glandular damage. However, the

presence of a subclass of autoantibodies recognizing the j-adrenergic and muscarinic-

cholinergic receptors alludes to a potential mechanism for the decline in saliva production

and protein output which involves antibody-mediated impairment of neuro-glandular

stimulation. Thus, the presence of autoantibodies may be responsible for secondary

effects on gland function following a primary immunological disturbance generated by a

genetic defect in the salivary glands. This model is useful in explaining why first, the

presence of activated lymphocytes is necessary to develop the loss of the fluid phase of

saliva in the NOD mouse while NOD-scid salivary flow rates remain normal, and second,

how an autoimmune response against relatively small portions of the total area of the

submandibular gland may dramatically affect the total salivary output of the parotid,

sublingual and submandibular glands combined.

In conclusion, by using the NOD-scid mouse model, I have begun to dissociate

elements of salivary dysfunction in the NOD mouse attributable to the progression of the

autoimmune disease from normal/abnormal events dictated by the genetic background of

the animal. These findings are striking in that they suggest an underlying defect in salivary

gland homeostasis in NOD mice which results in a histopathological and functional

phenotype similar to human Sj6gren's syndrome.



The protein and mucin-rich secretions derived from the salivary, lacrimal, and

other minor exocrine tissues, i.e.labial and hardarian glands, are essential for maintaining

the health and integrity of the oral and ocular surfaces (9). For the most part, both tear and

saliva secretions serve similar functions and contain many of the same protein constituents,

e.g., EGF, NGF, TGF-a, lactoferrin, lysozyme, and immunoglobulins (9,10). At the same

time, however, saliva and/or tear specific secretary proteins, as evidenced by salivary

amylase and digestive enzymes, provide for specialized physiological functions of the

individual secretary fluids (9).

Chapter 3 of this dissertation documented unique changes in the synthesis of

several proteins noted specifically in saliva, including de novo synthesis of PSP as well as

decreased concentrations of amylase and EGF in NOD mice (121). Similarly, protein

expression abnormalities were detected in NOD-scid mice. These findings suggest that

changes in saliva protein constituents in the NOD strain are independent of the

autoimmune destruction of the glands. The aberrant synthesis of PSP in the

submandibular glands of NOD mice led us to examine the question of whether PSP may

be abnormally synthesized in other exocrine tissue of NOD mice, and whether

transcriptional splicing is likely to account for the novel PSP isofonn described in Aim 3.

In this specific aim, I show that while PSP is synthesized in the lacrimal glands of NOD

mice, it is detected in the lacrimal glands in several other laboratory mouse strains as well.

As a constituent in saliva and tears, I provide evidence of a potential anti-microbial

function for PSP which may explain its normal synthesis by both the salivary and lacrimal

glands. In addition, NOD mice were found to have lost the normal regulation of PSP gene

transcription in a number of organs/tissues. No alternatively spliced mRNA was detected,

however, the NOD PSP gene contains numerous base pair changes indicative of strain

specific differences between mice.

Materials and Methods


BALB/c, CBA/J, and NOD/Uf mice were bred and maintained under SPF

conditions in the mouse colony of the Department of Pathology and Laboratory Medicine

at the University of Florida, Gainesville, FL. C3H/HeJ and NOD-scid mice were

purchased from The Jackson Laboratories (Bar Harbor, ME). Both male and female mice

ranging in age from 8 to 25 weeks were used. NOD mice were routinely tested for blood

glucose levels using Chemstrip bG reagent strips (Boehringer Mannheim, Indianapolis,

IN). Consecutive elevated fasting blood glucose levels >240 mb/dl were considered onset

of diabetes, after which time the mice were maintained on daily insulin injections (67).

NOD mice of 15 weeks of age were separated into diabetic and prediabetic sample groups

for protein studies.


Lysteria monocytogenes and Streplococus mulans were gifts from Dr. F.

Southwick, Department of Infectious Diseases, University of Florida, and Dr. A. Bleiweis,

Department of Oral Biology, University of Florida. Actinobacillus

actinomycelenmcomitans was generously provided by Dr. P. Fives-Taylor, Department of

Microbiology, University of Vermont.

Isolation of Tissues for Protein Studies

Tissues were excised from mice killed by cervical dislocation. Glands were pooled

from a minimum of two age- and sex- matched mice of each strain, freed of connective

tissue, fat, and any lymph nodes, then homogenized in 10 mM Tris buffer (pH 7.4). The

slurry was centrifuged at 500 g for 15 min to pellet cellular debris and protein

concentrations of the resulting supernatant were determined by the method of Bradford

(132) using bovine serum albumin as the standard. Supernatants were frozen at -700 C for

Western blot analysis.

Isolation of total RNA

Excised tissues were immediately homogenized in 2 ml of 4M Guanidinium

Isothiocynate (GITC) as described by Chirgwin et al. (148). Briefly, 100 pl of 10%

Sarkosyl and 14.2 il of 2-Mercaptoethanol were added to homogenates and total RNA

was isolated by centrifugation at 35,000 g over a 2 M CsCI gradient. RNA pellets were

washed in 200 pl of 80/20% ethanol/Diethyl Pyrocarbonate (DEPC) treated water and

then resuspended in 200 1l of DEPC-treated water. RNA was precipitated through the

addition of 500 pl of 100% ethanol and 8 p. of 5M NaCI. Quantitation of total RNA was

determined by spectrophotometric analysis at 260 nm wavelength.

RT-PCR and Southern Blot Detection ofPSP PCR Products

Total tissue RNA (2 pg) was pelleted by centrifugation and reverse transcribed

using Superscript II Reverse Transcriptase (Gibco BRL) (122). The resulting cDNA was

amplified by PCR for 40 cycles using a 94C denaturation (1 min), 600C primer annealing

(1 min) and 72C elongation (2 min) using 5'GCAGAGAAACAAGGATCTCG and 3'

CACTGGAGAGTAGCCAGCAGG PSP-specific primers. These primers spanned a

region preceding the start codon and extending beyond the translation termination site

(149). PCR products were separated on 1.2% agarose gels and transferred to nylon

membrane by Southern blotting. To confirm the identity of specific PCR products,

hybridization of digoxigenin-labeled oligonucleotide internal probe (5'

AATGCGACCGTTCTTGCC) specific for PSP cDNA was carried out using the Genius

system ofBoehringer-Mannheim (Indianapolis, IN) (150). Briefly, oligonucleotide probes

were labeled with digoxigenin using terminal transferase. Southern blot membranes were

baked at 80C for I hour, blocked with pre-hybridization buffer (Genius Kit), and

hybridized with labeled probes overnight at 650C. Colorimetric detection of PSP product

was assayed using an anti-digoxigenin alkaline phosphatase conjugated antibody according

to manufacturer's instructions. Primers and probes specific for G3PDH were used as

positive controls for all PCR reactions. All nucleotide primers and probes were

synthesized in the University of Florida's ICBR DNA Synthesis Laboratory (Gainesville,


Polvacrylamide eel electrophoresis and Western blot analysis

Gland lysates (30 pg of total protein per lane) were subjected to electrophoretic

separation on a 12% SDS-polyacrylamide gel using a modified Tris-Glycine system of

Pugsley and Schnaitman (136). Gels were stained by using Coomassie Brilliant Blue R-

250 (137) or transferred to Immobilon-P membranes (Millipore, Boston, MA) for 2 hr at

70v for Western blotting (138). The blocking buffer consisted of 3% nonfat dry milk and

3% BSA in Tris-buffered saline. Polyclonal rabbit anti-mouse parotid secretary protein

(mPSP) IgG antibody, kindly provided by Dr. William Ball (Dept. of Anatomy, Howard

University) (139) was incubated with each membrane for 2 hr at 25"C. Following three

10 min washes, the membranes were incubated in alkaline phosphatase conjugated goat

anti-rabbit immunoglobulin (Sigma Chemical Co.) and exposed to substrate as previously

described (95).

Purification and Radiolabeling of PSP

Parotid secretary protein was purified using a one-step procedure by

electrophoretic separation of saliva proteins on a 3 mm 10% polyacrylamide prep gel.

Four hundred Il of whole saliva from C3H/HeJ or BALB/c mice was separated using a

single large well in the stacking gel. Molecular weight standards (Bio-Rad) and 0.5 cm of

the sample well were cut and transferred to PVDF membrane for Western blotting.

Following detection of PSP migration, the similar region on the remaining unfixed gel was

removed, macerated by an electric homogenizer in 3 ml PBS containing 0.02% NaN3 and

0.2% SDS, and placed on an orbital shaker overnight at 4C. The extracted protein was

dialyzed against ddH20 and lyophilized to concentrate the purified protein. Typically, 1.5-

2 .ig of pure PSP was recovered for 400 pl of saliva protein. Purity of PSP was

determined by electrophoresis and staining of polyacrylamide gels with Coomasie Brilliant

Blue R-250, autoradiography and Western blot. Pure PSP, M, 25,000 was radiolabeled

using chloramine-T and Na1251 obtained through Amersham (Arlington Heights, IL).

Radiolabeled PSP was purified from free [1251], potentially other contaminating proteins

and radiolabeled BSA in the incubation buffer by molecular sieve chromatography on

Sephadex G-75 obtained from Pharmacia (Uppsala, Sweden).

PSP Binding to Bacteria

Four strains of bacteria, L. monocytogenes, E. coli, S. mutans, and A.

actinomycetemcomitans were grown in overnight cultures at 370C. The bacteria were

pelleted by centrifugation at 40C for 5 min at 15,000 x g and washed in PBS buffer

containing 0.5% BSA, 0.02% NaN3, 1 mM CaCI2, 1 mM MgC2, and 1 mM ZnCl2. The

bacteria were resuspended in 1.0 ml of the above buffer at 108 cells/ml. [1251]- Labeled

PSP (104 cpm) was added to the cells and incubated at 370C or 230C for 2 hr on an orbital

shaker. The cells were pelleted by centrifugation, washed twice, and the radiolabel bound

to the bacteria quantitated by a Beckman gamma counter. Specificity of PSP binding was

determined by pre-incubation of the cells with 10 pg/ml of unlabeled PSP. Cation-

dependent binding was determined by incubation where one or all of the salts were altered

in the incubation buffer.

Ligand blot assays were conducted as described previously (151). In brief,

bacteria were lysed by sonication (4C in PBS), and the membrane fraction collected

following centrifugation. The membranes were sonicated a second time (in buffer

containing 0.5% SDS), followed by centrifugation and the soluble material heated to 1000

C for 5 min. Sample (the optical density adjusted to A2s0= 2.6 units/ml: 10pil (-0.1 units)

was mixed with SDS containing sample buffer, separated on 10% polyacrylamide gel, and

transferred to PVDF. The membrane was blocked using a modification of the method of

Hossenlopp et al. (151). Briefly, the blot was incubated for 30 min in 10 mM Tris-HCI,

pH 7.4 containing 150 mM NaCI, 0.02% NaN3, and 3% NP-40, 30 min in Tris buffer

containing 1% BSA in place of detergent, and 30 min in Tris buffer containing 0.1%

Tween-20. The membrane was incubated overnight in TBS containing 1% BSA, 0.1%

Tween-20, and 10' cpm ['25I]-PSP. The blot was washed 3 times in TBS and exposed to

X-ray film for 12 hr at -800 C using Kodak XAR-5 film. Specificity of PSP binding to

bacterial proteins was determined by pre-incubation of the filter with unlabeled protein.

Ligand binding blots were run on 3 separate occasions for reproducibility using 2 separate

preparations of ['25I]-PSP from C3H/HeJ and BALB/c saliva.

Amylase Assay

The activity of human salivary amylase (SIGMA Chemical) was determined in the

presence of varying concentrations of pure PSP, proline-rich protein (PRP), and BSA.

Amylase activity was determined by the method of Bernfeld using starch as a standard

substrate (133). Human amylase (lgg/ml) was resuspended in PBS. The incubation

solution was comprised of 0.4 g soluble starch in 60 mM Tris-HCI containing 0.15M NaCI

and 3 mM CaC12 and ZnCI2. The stop solution consisted of 0.45% 12, 0.045% KI and

0.03 N HCI. After termination of the reaction at 5, 10, 15, and 20 min, the enzyme

activity was defined as the amount that hydrolyzed 1 mg starch/min/mg of protein. All

values are expressed as mean + S.E. for 3 separate determinations.


Detection ofPSP RNA Transcripts in Murine Lacrimal Gland

Total RNA derived from lacrimal tissue isolated from NOD, BALB/c and

C3H/HeJ mice was assayed for the presence of PSP RNA transcripts by RT-PCR and

Southern-blotting. As shown in Fig 14A, a strong PCR band was observed at 785 bp in

lacrimal RNA of NOD but not C3H/HeJ or BALB/c mice. When hybridized with a PSP-

specific internal oligonucleotide-probe, both NOD and BALB/c developed a band at the

expected base pair size (Fig 14B) while the C3H/HeJ remained negative. Interestingly, the

lack of PSP expression in C3H/HeJ mice is in concordance with previous reports by

Hjorth et al (105) describing the parotid specificity of PSP gene expression using

C3H/HeJ mice. Detection of the housekeeping gene, G3PDH, was used as a positive

control to indicate that similar amounts of RNA was utilized in each PCR reaction (Fig

14C). G3PDH bands of 983 base pairs were present in all samples.


800 bp
700 bp


800 bp
700 bp


1,000 bp
900 bp



a n r- G3PDH

Figure 14 RT-PCR and Southern blot detection of PSP mRNA isolated from murine
lacrimal glands. Total tissue RNA (2gg) was reverse-transcribed and resulting cDNA
amplified by PCR using PSP-specific primers spanning the entire of length of the PSP
transcript (785 bp). Panel A; Ethidium bromide-stained agarose gel containing PSP-
amplified PCR product (10 l/lane) from 8 week C3H/HeJ (Lane 1), NOD (Lane 2), and
BALB/c (Lane 3) lacrimal tissues. Panel B; Agarose gel (from Panel A) was transferred
to a nylon membrane by Southern blotting and hybridized with a digoxigenin-labeled
oligonucleotide probe specific for an internal PSP sequence. Blots were developed using
an alkaline phosphatase labeled anti-digoxigenin antibody. Panel C; Positive control of
RT-PCR and Southern Blot procedure using G3PDH-specific primers (983 bp product)
and probe for C3H/HeJ (Lane 1), NOD (Lane 2), and BALB/c (Lane 3) RNA samples.

I 'b ol
0 0
. 0
Q:~~ Z s


84.0 -
53.2 0

34.9 h1
28.7 Z. *
kDa "

Figure 15 Western blot detection of parotid secretary protein in murine lacrimal glands.
Total protein in pooled tissue lysates (n=3 mice/sample) from C3H/HeJ parotid gland
(Lane 2; 1 pg/lane), was compared with pooled lacrimal gland lysates (30 jg/lane) from
C3H/HeJ (Lane 3), CBA/J (Lane 4), BALB/c (Lanes 5 and 6), NOD (Lanes 7-10) and
NOD-scid (Lane 11) mouse strains. Age (in weeks) and sex of sample groups are labeled
above, with 15 week NOD mice being diabetic. Briefly, proteins were separated on 12%
SDS-PAGE gels under reducing conditions, transferred to Immobilon P membranes, and
incubated with rabbit anti-PSP antibody. Blots were developed with alkaline phosphatase-
conjugated goat anti-rabbit antibody and substrate as per methods section. Prestained
molecular weight markers are as follows; Bovine serum albumin, 84.0 kDa; Ovalbumin,
53.2 kDa; Carbonic Anhydrase, 34.9 kDa; and Soybean trypsin inhibitor, 28.7 kDa.







Figure 16 Purification of Parotid Secretory Protein. Four hundred I. of total saliva was
separated in a prep gel well (Panel A). Parotid secretary protein was identified by
Western blot using antibody to PSP (Panel B) and cut from the gel. The purified protein
was dialyzed, radiolabeled, and reanalyzed for purity by autoradiography and Western blot
reactivity (Panel C lanes 2 and 3, respectively). Molecular weight standards (Bio-Rad) are
Phosphorylase B, 116 kDa; Bovine Serum Albumin, 84 kDa; Ovalbumin, 53 kDa;
Carbonic Anhydrase, 35 kDa; and Soybean Trypsin Inhibitor, 28 kDa.

116 l
84 -db

53 -i


Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd