Journal of Autoimmunity 26 (2006) 90e103www.elsevier.com/locate/issn/08968411

Sjogren’s syndrome in the NOD mouse model is an interleukin-4time-dependent, antibody isotype-specific autoimmune disease

Juehua Gao a, Smruti Killedar b,1, Janet G. Cornelius a, Cuong Nguyen a,Seunghee Cha b, Ammon B. Peck a,b,c,*

a Department of Pathology, Immunology and Laboratory Medicine, University of Florida College of Medicine, Gainesville, FL 32610, USAb Department of Oral Biology, University of Florida College of Dentistry, P.O. Box 100424, Gainesville, FL 32610, USA

c Center for Orphan Autoimmune Diseases, University of Florida, Gainesville, FL 32610, USA

Received 13 July 2005; revised 15 November 2005; accepted 21 November 2005

Abstract

NOD.B10-H2b and NOD/LtJ mice manifest many features of primary and secondary Sjogren’s syndrome (SjS), respectively, an autoimmunedisease affecting primarily the salivary and lacrimal glands leading to xerostomia (dry mouth) and xerophthalmia (dry eyes). A previous studysuggested that the TH2 cytokine, interleukin (IL)-4, plays an integral role in the development and onset of SjS-like disease in the NOD mousemodel. To define further the role of IL-4 in onset of murine SjS-like disease, we have examined two IL4 gene knockout (KO) mouse strains,NOD.IL4�/� and NOD.B10-H2b.IL4�/�. Unlike NOD.IL4�/� mice, NOD.B10-H2b.IL4�/� mice are resistant to development of diabetes.The presence of a dysfunctional IL4 gene did not impede leukocyte infiltration of the salivary glands, yet prevented development of secretorydysfunction. Whereas NOD.B10-H2b.IL4�/� mice exhibited many pathophysiological manifestations of SjS-like disease common to the parentalstrains, these mice failed to produce anti-muscarinic acetylcholine type-3 receptor (M3R) autoantibodies of the IgG1 isotype. Cytokine mRNAexpression profiles and adoptive transfers of T lymphocytes from NOD.B10-H2b.Gfp mice into NOD.B10-H2b.IL4�/� mice at different agessuggest IL-4 is required during the pre-clinical disease stage (around 12 weeks of age) to initiate clinical xerostomia. The results of this studyindicate that the failure of NOD.IL4�/� and NOD.B10-H2b.IL4�/� mice to synthesize anti-M3R autoantibodies of the IgG1 isotype apparentlyexplains why these mice fail to develop exocrine gland dysfunction, despite exhibiting pre-clinical manifestations of SjS-like disease.� 2005 Elsevier Ltd. All rights reserved.

Keywords: NOD mouse; Sjogren’s syndrome; Anti-muscarinic acetylcholine type-3 receptor autoantibodies; Interleukin-4

1. Introduction

Sjogren’s syndrome (SjS) is a human autoimmune diseasecharacterized by the loss of exocrine function as a result of

Abbreviations: ANA, anti-nuclear antibodies; KO, knockout; M3R,

muscarinic acetylcholine type-3 receptor; PSP, parotid secretory protein;

SjS, Sjogren’s syndrome.

* Corresponding author. Department of Oral Biology, University of Florida,

1600 SW Archer Road, P.O. Box 100424, Gainesville, FL 32610, USA. Tel.:

þ1 352 392 3046; fax: þ1 352 392 3474.

E-mail address: peck@pathology.ufl.edu (A.B. Peck).1 Present address: The Johns Hopkins Asthma and Allergy Center, 5501

Hopkins Bayview Circle, Rm 2A 58, Baltimore, MD 21224, USA.

0896-8411/$ - see front matter � 2005 Elsevier Ltd. All rights reserved.

doi:10.1016/j.jaut.2005.11.004

a chronic immune attack directed primarily against the sali-vary and lacrimal glands [1e3]. Although the underlyingcause of SjS remains elusive, a number of studies using theNOD mouse model and its congenic strains have led us to pro-pose the concept that autoimmune excrinopathy progresses inthree consecutive phases [4e8]. In phase 1, a number of aber-rant genetic, physiological and biochemical activities associatedwith retarded salivary gland organogenesis occur sequentiallyprior to and independent of initiation of an autoimmuneattack. In phase 2, leukocytes infiltrate the exocrine glandswith a concomitant increase in the expression of inflammatorycytokines. In phase 3, secretory dysfunction of the salivaryand lacrimal glands occurs, most likely the result of

91J. Gao et al. / Journal of Autoimmunity 26 (2006) 90e103

production of anti-muscarinic acetylcholine type-3 receptor(M3R) autoantibodies [6,9e12]. An interruption within anyone of these three phases prevents onset of clinical SjS-likedisease.

Autoantibodies play an important role in the pathogenesis ofSjS. Patients with SjS commonly have severe hypergammaglo-bulinemia that can lead to complications of both liver and kid-neys [13]. SjS patients, as well as NOD mice, develop anextensive array of autoantibodies reactive with nuclearantigens, intracellular components, membrane proteins andsecreted products of exocrine tissues [9,14e18]. Althoughanti-nuclear autoantibodies (ANAs), specifically anti-SS-A/Ro and anti-SS-B/La, have been found in 40e70% of patientsera and are considered one of the better markers of SjSdisease [19], recent attention has focused on anti-M3R autoanti-bodies. Preliminary studies suggest that anti-M3R autoantibod-ies may be present in 80e100% of sera from both SjS patientsand NOD mice [4,6,20] and may be an important effector ofglandular dysfunction by blocking normal signal transductionpathways or desensitizing acinar cells to normal neural stim-ulations [12,21,22]. This concept is supported, in part, bystudies showing that the IgG fractions of sera obtainedfrom SjS patients or NOD mice with disease can suppressstimulated salivary flow rates when infused into healthy, nor-mal mice [11].

The possibility that anti-M3R autoantibodies may be in-volved in development of xerostomia and xerophthalmia inSjS has led us to examine this issue in greater detail [4,6,20].Preliminary studies utilizing a number of congenic partnerstrains of NOD with non-functional cytokine genes revealedthat the interleukin (IL)-4 gene knockout (KO) mouse,NOD.IL4�/�, fails to develop salivary gland dysfunction de-spite severe leukocytic infiltration of the salivary glands,accompanied by detectable increases in the expression of pro-inflammatory cytokines. IL-4 is a pleiotropic cytokine involvedin cell proliferation, activation and differentiation. This TH2 cy-tokine appears to promote cell survival and proliferation, espe-cially of B lymphocytes [23]. Although it is not a growth factorper se for resting lymphocytes, it can substantially prolong thelives of both T and B lymphocytes in culture [24], and functionas a co-mitogen for B cell growth [25]. IL-4 is also well-recog-nized as a regulatory molecule for isotype switching, specificallystimulating germline g1 and 3 immunoglobulin gene transcrip-tion. Thus, in combination with CD40L, and possibly otherfactors, IL-4 promotes immunoglobulin isotype switching toeither IgG1 or IgE in vivo [26,27]. A precise role for IL-4in autoimmune exocrinopathy, or SjS-like disease, of NODmice is still unknown. IL-4 might act as a promoter of lym-phocyte activation and proliferation that results in the survivaland expansion of autoimmune T and/or B cells, or alternativelyas a regulator molecule for the production of specific isotypicautoantibodies. In the present study, we have expanded ourinvestigations to gain insight into how the different biologicalroles of IL-4 impact development and onset of clinical SjS-like disease through comparing the autoimmune processesof NOD.IL4�/� and NOD.B10-H2b.IL4�/� mice with paren-tal NOD/LtJ and NOD.B10-H2b mice.

2. Materials and methods

2.1. Animals

NOD/LtJ, NOD.IL4�/� and NOD.B10-H2b mice were pur-chased from Jackson Laboratories (Bar Harbor, ME) and main-tained under specific pathogen-free conditions in the mousefacility of the Department of Pathology, University of Florida,Gainesville, FL. NOD.B10-H2b.IL4�/� mice were generatedby crossing a NOD.IL4�/� male with NOD.B10-H2b femalemice. The F1 heterozygotes were intercrossed to produce F2generations that were screened for the presence and/or absenceof the disrupted IL-4 gene by PCR and the MHC H-2b locus bymicrosatellite marker genotyping. PCR primers for IL-4 werepurchased from Invitrogen Life Technologies (Carlsbad, CA).Microsatellite marker primers (Dmit17e21 MapPair) were ob-tained from The Jackson Laboratory (Bar Harbor, ME). Micehomozygous for both the neomycin-disrupted IL-4 gene andthe H-2IAb gene were bred to obtain the founder mice. TheNOD.B10-H2b.IL4�/� line is being carried via single descentthrough brotheresister matings. NOD.B10-H2b.Gfp micewere constructed by crossing a NOD.B10-H2b male withNOD.Gfp female mice. The F1 heterozygotes were intercrossedto produce F2 generations that were screened for the presenceand/or absence of the H-2IAb locus and a Gfp-positive pheno-type. Mice homozygous for both markers were bred to obtainthe founder mice. The NOD.B10-H2b.Gfp line is being carriedvia single line of descent. NOD.Gfp mice were generously pro-vided to the University of Florida by Ixion Biotechnology Inc.(Alachua, FL). Studies described herein were approved by theUniversity of Florida IACUC.

2.2. Histology

Submandibular and lacrimal glands were surgically removedfrom euthanized mice, placed in 10% phosphate buffered for-malin for 24 h, then embedded in paraffin. The embedded tis-sues were sectioned at 5 mm thickness and mounted ontomicroscope slides. Slides were de-paraffinized by immersingin xylene then dehydrated in ethanol. Tissue sections werestained with Mayer’s hematoxylin and eosin (H/E) dye, and ob-served at 40� magnification. Alternatively, the tissue sectionswere washed 5 min with PBS at 25 �C, then incubated 1 hwith blocking solution containing normal rabbit serum diluted1:50 in PBS. Rat anti-mouse B220 (BD Biosciences Pharmin-gen, San Diego, CA) diluted 1:10 and goat anti-mouse CD3(Santa Cruz Biotechnology, Santa Cruz, CA) diluted 1:50were added to individual sections for 1 h at 25 �C. The slideswere washed three times with PBS for 5 min per wash followedby a 1 h incubation at 25 �C in a mixture of secondary antibod-ies (Texas Red-conjugated rabbit anti-rat IgG (Biomedia,Foster City, CA) diluted 1:25 and FITC-conjugated rabbitanti-goat IgG (Sigma Chemicals, St. Louis, MO) diluted1:100. The slides were thoroughly washed with PBS, treatedwith Vectashield DAPI-mounting medium (Vector Laboratory,Burlingame, CA) and overlaid with glass coverslips. Stainedsections were visualized at 400� magnification.

92 J. Gao et al. / Journal of Autoimmunity 26 (2006) 90e103

2.3. Measurement of salivary flow rates

To measure stimulated flow rates of saliva, individual micewere given an intraperitoneal injection of isoproterenol(0.2 mg/100 g body weight) plus pilocarpine (0.05 mg/100 gbody weight) dissolved in Ca2þ- and Mg2þ-free phosphatebuffered saline (PBS). Saliva samples were collected fromeach mouse for 10 min starting 1 min after injection of the se-cretagogue. The volume of each saliva sample was measuredand normalized based on body weight. The saliva sampleswere then frozen at �80 �C until analysis for protein concen-trations and proteolytic activities.

2.4. Analysis of saliva amylase activity

Amylase activity in saliva was determined using theInfinity� Liquid Amylase Kit (Thermo Trace ElectronCorp., Waltham, MA) in which ethylidene-pNP-G7 was thesubstrate. Saliva samples were diluted 250-fold with deionizedwater and added to1 ml of the Infinity Amylase Liquid StableReagent. Following 1 min and 2 min incubations at 37 �C, ab-sorbance was measured at a wavelength of 405 nm. Amylaseactivity was calculated according to the manufacture’s instruc-tions using the formula: Amylase activity (U/L) ¼ 6A/2 � 5140 � Sample dilution.

2.5. Proteolysis of parotid secretory protein (PSP)

Detection of PSP proteolysis was carried out by incubatingwhole saliva specimens with a synthesized oligopeptide corre-sponding to amino acids 20 through 34 of the published se-quence for mouse PSP. This oligopeptide contains theproteolytic site (NLNL) for a serine kinase present in salivaryglands following onset of SjS-like disease in the NOD mouse[35]. Eight microliters of saliva collected from individual micewere mixed with 42 ml of the PSP oligopeptide (2.5 mg/ml)and incubated at 42 �C for 2 h. Following incubation, 50 mlTriseHCl buffer (50 mM, pH 8.0) was added and the mixturecentrifuged through Micro-spin filter tubes at 14,000 rpm for10 min. The filtrates were analyzed by HPLC (Dionex Sys-tems, Sunnyvale, CA) for proteolytic products. Control sam-ples consisted of 50 ml of the PSP oligopeptide.

2.6. Detection of cytokine and transcriptionfactor mRNA expressions

Total RNA was prepared from freshly explanted submandib-ular glands using the RNeasy Mini Kit (Qiagen, Valencia, CA)according to the manufacturer’s protocol. cDNA was synthe-sized using 4 mg of total RNA, Superscript II reverse transcrip-tase (Invitrogen Life Technologies), and pd(T)12e18oligomeric DNA (Amersham Pharmacia, Piscataway, NJ). ThecDNAwas quantified by spectrophotometry. PCRs were carriedout using 1 mg of cDNA as template. Following an initial dena-turation at 94 �C for 4 min, each PCR was carried out for 40cycles consisting of 94 �C for 1 min, 60 �C for 45 s and72 �C for 2 min. PCR products were size separated by

electrophoresis using a 0.9% agarose gels and visualized withethidium bromide staining. The primer sequences used are asfollows:

Interferon (IFN)-g 5#: TGAACGCTACACACTGCATCTTGG; IFN-g 3#: CGACTCCTTTTCCGCTTCCTGAG;IL-1 5#: ATGGCCAAAGTTCCTGACTTGTTT; IL-1 3#:CCTTCAGCAACACGGGCTGGTC;IL-2 5#: ATGTACAGCATGCAGCTCGCATC; IL-2 3#:GGCTTGTTGAGATGATGCTTTGACA;IL-4 5#: ATGGGTCTCAACCCCCAGCTAGT; IL-4 3#:GCT CTTTAGGCTTTCCAGGAAGTC;IL-6 5#: ATGAAGTTCCTCTCTGCAAGAGACT; IL-6 3#:CACTAGGTTTGCCGAGTAGATCTC;IL-10 5#: GCAGGGGCCAGTACAGCCGGGAA; IL-103#: GCTTTTCATTTTGATCATCATGT;IL-12 5#: ATGACATGGTGAAGACGGCCAGAG; IL-123#: TCACGACGCGGGTGCTGAAGGCGTG;G3PDH 5#: TGAAGGTCGGTGTGAACGGATTTGGC;G3PDH 3#: CATGTAGGCCATGAGGTCCACCAC;NFkB 5#: TCCACGCTGGCTGAAAATCC; NFkB 3#:CACGGGAGACACAGACGAACAGTT;STAT6 5#: TCTGTCCTTGGTGGTCATCGTG; STAT6 3#:GGAGATGGGGTTCTTTGGGTTT;SOS 5#: TCCTCATCCTATTGATAAATGGGC; SOS 3#:CCAAGGGCACATAGTGACAACC;PI3Kp85 5#: ACAGACTGGTCCCTGAGTGACTTG; PI3Kp85 3#: GGAGTCCTTTCCGCCCTTGTTGTT;RAF 5#: AGCACCACCTTCTTTCCCAATGC; RAF 3#:CCGCTAACACTGAGTCACCACACT

2.7. Detection of anti-nuclear autoantibodies (ANAs)

Serum ANAs were detected by indirect immunofluorescentstaining of the human cell line, HEp-2G, using the AntinuclearAntibody Kit (Sigma Diagnostics, St. Louis, MO). Each serumsample was tested after being diluted 50-fold with PBS. TheHEp-2G cells were incubated with test sera for 3 h in a humid-ified chamber at 37 �C, then washed three times with PBS.The HEp-2G cells were then treated with FITC-conjugatedgoat anti-mouse Ig for 45 min, followed by two washes withPBS. Fluorescence was visualized at 100� magnification us-ing a fluorescent microscope.

2.8. Detection of anti-mM3R autoantibody in mouse sera

Blood samples were collected from the tail veins of exper-imental and control mice at 4 and 20 weeks of age and allowedto clot at room temperature. Sera were removed and stored at�80 �C until testing for the presence of anti-M3R autoanti-body using a mouse M3R-transfected CHO cell line. Genera-tion of the M3R-transfected CHO cell line assay was carriedout as detailed for the construction of the human and mouseM3R-transfected CHO cell lines [20,30].

Prior to testing for the presence of autoantibody using themM3R-transfected Flp-In CHO cells, each test sera was firstpre-absorbed with non-transfected CHO cells (5 � 106 CHO

93J. Gao et al. / Journal of Autoimmunity 26 (2006) 90e103

cells per 10 ml sera at 25 �C for 2 h). At time of testing, thetransfected and non-transfected CHO cells were washedonce with PBS and resuspended in FACS buffer to a densityof 1 � 106 cells/ml. Aliquots of cells (0.1 ml) were incubated2 h at 4 �C with 10 ml sera. The cells were washed once withFACS buffer, then incubated with either FITC-conjugated goatanti-mouse IgG antibody (BD Biosciences Pharmingen) orFITC-conjugated goat anti-mouse IgG1, IgG2a, IgG2b,IgG3, IgA, IgE and IgM antibodies (Accurate ChemicalCorp.) or goat anti-mouse IgG2a and IgG2c antibodies (South-ern Biotechnology Associates Inc., Birmingham, AL) for45 min at 4 �C. After a final wash with FACS buffer, the cellswere resuspended in 102 ml FACS buffer and analyzed usinga FACScan cytometer (Becton Dickinson).

2.9. Adoptive transfer of T cells intoNOD.B10-H2b.IL4�/� mice

Single cell suspensions of splenic leukocytes, prepared fromeither 4-, 10- or 12-week-old NOD.B10-H2b.Gfp mice, wereincubated with PE-conjugated goat anti-mouse CD4 antibody(1 mg/106 cells) for 1 h at 4 �C, washed once, and resuspendedin FACS buffer. The labeled cells were subjected to FACS sort-ing to isolate CD4þ Gfpþ T cells. Sorted T cells (1 � 106 cellsin 0.1 ml volume) were adoptively transferred into sex- andage-matched 4-, 10- or 12-week-old NOD.B10-H2b.IL4�/�

mice (n ¼ 4 per group) via tail vein injections once a weekfor 2 weeks. Two weeks after transfer, IL-4 production inGFPþ donor cells was confirmed by intracellular staining withPE-labeled goat anti-mouse IL-4 and analyzed by FACScan.Animals were then monitored for salivary secretory functionto 26 or 36 weeks of age. The animals were euthanized andtheir organs examined histologically for disease.

3. Results

3.1. Leukocytic infiltration of the submandibularand lacrimal glands of NOD.B10-H2b.IL4�/� mice

Over the past several years, NOD/Lt mice have been shown toexhibit many feature of SjS, including loss of stimulated fluid se-cretion concomitant with the appearance of leukocyte infiltratesin the salivary and lacrimal glands. To determine the impact ofIL4 gene expression on SjS-like disease in NOD mice in the ab-sence of potentially confounding factors associated with onsetof type 1 diabetes, we generated the NOD.B10-H2b.IL4�/�

mouse, a line carrying a disrupted IL4 gene and exhibiting no di-abetes (due to the replacement of the Idd1 s locus with an Idd1r

locus derived from the C57BL/10 mouse). As presented inFig. 1A, histological analyses of the exocrine glands of thisNOD.B10-H2b.IL4�/� mouse revealed that both the subman-dibular and lacrimal glands show significant focal leukocytic in-filtrations, indicating that a dysfunctional IL4 gene does notprevent lymphocytes from entering the exocrine glands. In gen-eral, the levels of infiltration were visibly similar to or increasedfrom levels observed in either parental NOD/LtJ or NOD.B10-H2b mice. This was true for both male and female mice.

Previous flow cytometric analyses of the lymphocyte popula-tions infiltrating the submandibular and lacrimal glands of theparental NOD/Lt mouse have indicated that the majority ofthe infiltrating cells are CD4þ T cells with lesser numbers ofCD8þ T and B cells [28]. In the current analysis, a comparisonof direct immunofluorescent staining of the submandibularand lacrimal glands from both NOD.B10-H2b and NOD.B10-H2b.IL4�/� mice revealed a much higher level of B cellsthan T cells (Fig. 1B), suggesting a loss of B lymphocyteswhen isolating lymphocytes from the exocrine gland tissue. Inaddition, foci of NOD.B10-H2b mice appear to contain relativelyfewer numbers of T cells than foci of NOD.B10-H2b.IL4�/�

mice and, at the same time, give the impression that the infiltrat-ing T cells are more randomly scattered throughout the foci.

3.2. Lack of immune-mediated secretorydysfunction in the NOD.B10-H2b.IL4�/� mouse

To examine whether NOD.B10-H2b.IL4�/� mice retain thepredisposition for secretory dysfunction that develops in pa-rental NOD/LtJ or NOD.B10-H2b mice, temporal changes inthe saliva flow rates were compared among NOD.B10-H2b,NOD.IL4�/� and NOD.B10-H2b.IL4�/� mice. Saliva secre-tions were stimulated by intra-peritoneal injections of an iso-proterenol/pilocarpine solution. At 20 weeks of age, a timepoint at which secretory dysfunction is exhibited in NOD/LtJ and NOD.B10-H2b mice, only NOD.B10-H2b miceshowed a statistically significant decrease ( p < 0.05) whencompared to salivary flow rates at 4 weeks of age, an arbitrarytime point prior to onset of disease (Fig. 2A). NeitherNOD.IL4�/� nor NOD.B10-H2b.IL4�/� mice exhibited a sta-tistically significant loss of secretory activity. Furthermore,NOD.B10-H2b.IL4�/� mice retained normal salivary flowrates to 36 weeks of age (data not shown).

Reduced secretory function in NOD/LtJ and NOD.B10-H2b

mice results in temporal increases in protein concentrations insaliva and tears [29] (Fig. 2B). As expected, saliva collectedfrom both NOD.IL4�/� and NOD.B10-H2b.IL4�/� micemaintained consistently normal salivary protein concentrationsover time (Fig. 2B). Thus, the lack of a functional IL4 generesulted in retention of normal secretory flow rates and nomeasurable temporal changes in the overall levels of salivaryproteins, despite having little impact on leukocytic infiltrationof the submandibular and lacrimal glands.

3.3. Amylase activity in the saliva ofNOD.B10-H2b.IL4�/� mice

Concomitant with the development of xerostomia, parentalNOD/LtJ and NOD.B10-H2b mice also exhibit a number of al-tered physiological properties associated with the loss of sali-vary gland homeostasis. These include measurable decreasesin epidermal growth factor and amylase, as well as increasesin MMP activity, caspases and apoptosis as the acinar cellmass is disrupted and declines [8,29,30]. To determine if am-ylase activity in the two IL4 KO mouse strains, NOD.IL4�/�

and NOD.B10-H2b.IL4�/�, also decreases, amylase activity

94 J. Gao et al. / Journal of Autoimmunity 26 (2006) 90e103

Fig. 1. Lymphocytic infiltration of the salivary and lacrimal glands in the NOD.B10-H2b and NOD.B10-H2b.IL4�/� mouse. (A) Submandibular (SMX) and lac-

rimal glands were surgically explanted from euthanized NOD.B10-H2b and NOD.B10-H2b.IL4�/� mice at 20 weeks of age. The glands were fixed in 10% for-

malin, embedded in paraffin, and sections of 5 mm thickness cut. Each section was stained with Mayer’s hematoxylin and eosin (H/E) dye. Stained sections were

observed at 40� magnification. (B) Tissue sections (5 mm) of submandibular and lacrimal glands were mounted on electrostatically-treated slides and incubated 1 h

with blocking solution containing normal rabbit serum. After a wash in PBS, the sections were stained with a goat anti-mouse CD3 or a rat anti-mouse CD45/B220

antibody, followed by a mixture of Texas Red-conjugated rabbit anti-rat IgG and FITC-conjugated rabbit anti-goat IgG antibodies. The slides were washed with

PBS, treated with Vectashield DAPI-mounting medium and visualized at 400� magnification.

95J. Gao et al. / Journal of Autoimmunity 26 (2006) 90e103

Fig. 2. Saliva volumes, salivary protein concentrations and salivary amylase activities of NOD.B10.H2b, NOD.IL4�/�, NOD.B10-H2b.IL4�/� mice. (A) Saliva was

collected from NOD.B10.H2b, NOD.B10-H2b.IL4�/� and NOD.IL4�/� mice (n ¼ 6) at 4 and 20 weeks of age, following stimulation with an i.p. injection of iso-

proterenol / pilocarpine mixture. Saliva samples were collected from each mouse for 10 min starting 1 min after injection. The volume of each saliva sample was

measured and adjusted for body weight. (B) The protein concentration for each saliva sample was determined using the Pierce Coomassie Plus Protein Assay Kit

and a microplate reader. (C) Individual saliva specimens, collected from NOD.B10.H2b, NOD.B10-H2b.IL4�/� and NOD.IL4�/� mice (n ¼ 5) at 4 and 20 weeks of

age following stimulation with an i.p. injection of isoproterenol / pilocarpine mixture, were tested for amylase activity using the Infinity� Liquid Amylase Kit

(ThermoTrace). Activities (U/L) were calculated as 6 Absorbance/2 � 5140 � Sample dilution. Values are expressed as the means (U/mL) � SD. Inter-group

comparisons were calculated using the unpaired Student t-test based on the mean � SE (*p < 0.05).

96 J. Gao et al. / Journal of Autoimmunity 26 (2006) 90e103

of whole saliva was measured by the ability to degrade thesubstrate, ethylidene-pNP-G7. As shown in Fig. 2C,NOD.B10.H2b mice exhibited a decline in amylase activityfrom 386.9 � 49.5 to 288.9 � 33.7 U/L between 4 and20 weeks of age ( p < 0.05). In contrast, NOD.IL4�/� andNOD.B10-H2b.IL4�/� mice exhibited slight increases in amy-lase activities, from 234.9 � 26.7 to 291.6 � 59.3 U/L and181.4 � 25.2 to 250.7 � 26.9 U/L, respectively, between 4and 20 weeks of age. The lower levels of amylase activity inthe two IL4 KO strains compared to the parental NOD.B10-H2b strain are believed to result from significant alterationsin the IRS1/2 signal transduction pathways controlled in partby the IL-4 receptor pathway (unpublished observation). Nev-ertheless, NOD.IL4�/� and NOD.B10-H2b.IL4�/� mice do notexhibit temporal declines in amylase activity, suggesting re-tention of functional salivary glands.

3.4. Detection of ANAs in the sera of NOD.B10-H2b

and NOD.B10-H2b.IL4�/� mice

NOD/Lt and NOD.B10-H2b mice produce a variety of auto-antibodies reactive with nuclear, cellular and secreted proteinsof the exocrine glands, mimicking the humoral response ob-served in the human disease [31]. More specifically, the pres-ence of ANAs and/or rheumatoid factor (RF) in the sera of SjSpatients is a common feature of SjS and used extensively indisease diagnoses [32e34]. To determine if ANAs are presentin the sera of mice with a dysfunctional IL4 gene, sera collectedfrom 20 wk old NOD.IL4�/� and NOD.B10-H2b.IL4�/� mice,as well as NOD.B10-H2b and BALB/c mice, were diluted 1:50and incubated with HEp-2G cells fixed to glass slides. Thepresence of ANAs was detected by indirect immunofluo-rescence. As presented in Fig. 3, NOD.IL4�/�, NOD.B10-H2b.IL4�/� and NOD.B10-H2b mice all showed the presenceof ANAs in their sera, while non-autoimmune BALB/c micewere negative. Thus, ANAs are synthesized in the IL4 geneKO mice, but appear to have little or no direct role in exocrinedysfunction in this model.

3.5. Proteolytic activity in the salivaof NOD.B10-H2b.IL4�/� mice

In addition to the appearance of sialoadenitis in the salivaryglands, parental NOD/LtJ and NOD.B10-H2b mice also exhibita number of altered biochemical and physiological properties as-sociated with the exocrine glands. The most interesting of theseis the appearance of aberrant proteolytic activities, one of whichis the proteolysis of the anti-microbial protein [35], parotid secre-tory protein (PSP), at an unique NLNL sequence site within theN-terminal region of the protein molecule [8,29,30,36] by an ac-tivated serine protease. We have long used the appearance of thisserine protease as a marker for the eventual development of SjS-like disease in the NOD mouse model.

To determine if NOD.B10-H2b.IL4�/� mice retain this earlyphenotypic marker of SjS-like disease, saliva collected fromindividual mice were tested for the presence of PSP proteolyticactivity using HPLC to detect the appearance of the cleaved

products from a synthetic oligopeptide carrying the NLNLamino acid sequence. As presented in Fig. 4, the synthetic oli-gopeptide elutes at 13.9 min. When incubated with saliva fromnon-disease animals, e.g., BALB/c mice, the oligonucleotidepeptide is not degraded and elutes reproducibly at 13.9 min.In contrast, when incubated with saliva from NOD.B10-H2b,NOD.IL4�/� or NOD.B10-H2b.IL4�/� mice, the oligonucleo-tide peptide is partially, or completely, degraded, showing thetwo degradation products that elute at 9.2 and 12.8 min. Thus,the NOD.IL4�/� and NOD.B10-H2b.IL4�/� mice continue toexpress the PSP proteolytic serine protease.

3.6. Absence of anti-M3R autoantibodies in thesera of NOD.B10-H2b.IL4�/� mice

A number of recent studies have provided evidence support-ing the concept that autoantibodies reactive with the M3R arethe primary effector molecules leading to exocrine gland acinarcell dysfunction [12,21,22], and such autoantibodies are presentin both NOD/Lt and NOD.B10-H2b mice [10,30]. To determineif NOD.B10-H2b.IL4�/� mice synthesize anti-M3R autoanti-bodies, a transgenic CHO cell line metastably expressing themouse M3R protein was employed to detect the levels of differ-ent Ig subclasses of anti-M3R antibodies in the mice at either 4or 20 weeks of age. Sera, prepared from blood of femaleNOD.B10-H2b and NOD.B10-H2b.IL4�/� mice (n ¼ 5 perstrain) bled first at 4 weeks of age, then subsequently at20 weeks of age, were pooled for each group and age, incubatedwith the M3R-transfected CHO cells, followed by FITC-conju-gated goat anti-mouse IgG, IgA, IgG1, IgG2a, IgG2b, IgG3,IgE or IgM, then analyzed using flow cytometry. As shown inFig. 5A, the overall profiles for the major subclasses (IgM,IgA, IgE and IgG) were similar for both NOD.B10-H2b andNOD.B10-H2b.IL4�/� mice. On the other hand, the anti-M3Rantibody profiles for the different IgG isotypes proved distinct(Fig. 5B). Sera of NOD.B10-H2b mice at 4 weeks of age con-tained IgG1 and IgG2b isotypic anti-M3R antibodies, whilethe sera of these mice at 20 weeks of age contained IgG1,IgG2a and IgG3 isotypic anti-M3R antibodies. In contrast,sera of NOD.B10-H2b.IL4�/� mice at 4 weeks contained lowlevels of Ig2b and IgG3 isotypic antibodies, whereas thesera of these mice at 20 weeks of age contained Ig2a/cand IgG3 isotypic antibodies. As expected, there was an ab-sence of IgG1 antibodies in the IL4 gene KO mice, and this ap-peared to result in a concomitant elevation in IgG3 isotypicantibodies.

Although the NOD mouse expresses the Igh1b gene encod-ing IgG2c, virtually all IgG subclass studies carried out withthis strain use IgG2a reagents that are reported to exhibit 15e20% cross-reactivity with IgG2c. As a result, expression levelsof anti-M3R antibodies of the IgG2c isotype in NOD.B10-H2b

and NOD.B10-H2b.IL4�/� mice were also examined. As pre-sented in Fig. 5C, the profiles for IgG2c showed approximately7-fold higher detection levels over those for IgG2a, at the sametime indicated the presence of IgG2c isotype anti-M3R antibod-ies in both NOD.B10-H2b and NOD.B10-H2b.IL4�/� mice.Most importantly, the level of IgG2c isotype anti-M3R

97J. Gao et al. / Journal of Autoimmunity 26 (2006) 90e103

Fig. 3. Detection of antinuclear antibodies (ANAs) in sera of NOD.IL4�/�, NOD.B10.H2b and NOD.B10-H2b.IL4�/� mice. ANAs were detected by indirect im-

munofluorescence staining using the Sigma Diagnostics Antinuclear Antibody Kit, which utilizes human epithelial (HEp-2G) cells. Sera collected from BALB/cJ,

NOD.B10.H2b, NOD.B10-H2b.IL4�/� or NOD.IL4�/� mice were diluted 1:50 and incubated on the HEp-2 cells for 3 h in a humidity chamber. The cells were

rinsed and incubated with FITC-conjugated goat anti-mouse whole Ig for 45 min. The slides were washed and examined at 100� magnification. Negative and

positive control sera were supplied with the kit, while sera from BALB/c was used as the comparative negative sera.

antibodies in NOD.B10-H2b.IL4�/� mice increased over age,despite lack of onset of clinical disease.

3.7. Altered temporal expressions of cytokine mRNAtranscripts in the submandibular glands ofNOD.B10-H2b.IL4�/� mice

A comparison of cytokine mRNA expression profiles in thesubmandibular glands from NOD.B10-H2b and NOD.B10-H2b.IL4�/� mice from 4 to 20 weeks of age was carried outusing RT-PCR analysis. As shown in Fig. 6, the cytokinemRNA profiles proved temporally quite different. In the SjS-like disease-prone NOD.B10-H2b mice at 4 weeks of age,when little or no leukocytic infiltration is present in the

submandibular glands, only IL-10 mRNA could be detectedof the relatively small number of cytokines tested. By 8 weeksof age, the time at which leukocytes first begin invading thesubmandibular glands, IL-6 mRNA was detected, as well aslow levels of IFN-g and IL-12. By 12 weeks of age, when leu-kocytic infiltrations are clearly present in the submandibularglands, mRNA for each cytokine tested, including IFN-g,IL-1, IL-2, IL-4, IL-6, IL-10 and IL-12, were detected. By20 weeks of age, many of the cytokines appeared to be ex-pressed at lower levels. By contrast, in NOD.B10-H2b.IL4�/�

mice, cytokine mRNA profiles of the submandibular glandsindicated that IFN-g, IL-1, IL-2, IL-6, IL-10 and IL-12 wereexpressed at a much earlier age and for a much longer time(Fig. 6), consistent with the histological findings that

98 J. Gao et al. / Journal of Autoimmunity 26 (2006) 90e103

Fig. 4. Detection of aberrant enzymatic activity in the submandibular glands of 20 wk old NOD.IL4�/�, NOD.B10.H2b, and NOD.B10-H2b.IL4�/� mice. Appear-

ance of an activated serine protease in saliva capable of proteolytic digestion of parotid secretory protein (PSP) was carried out by incubating saliva pooled from

20-week-old BALB/cJ, NOD.B10.H2b, NOD.B10-H2b.IL4�/� or NOD.IL4�/� mice (n ¼ 6 per strain) with a synthesized oligopeptide carrying the NLNL pro-

teolytic site. After a 2 h incubation at 42 �C, each reaction was diluted with TriseHCl buffer (50 mM, pH 8.0), centrifuged through Micro-spin filter tubes,

and the filtrates analyzed by HPLC. The PSP oligopeptide was run as a control (top panel) to ensure proper time of elution.

leukocyte infiltration occurs earlier in this mouse and tends tobecome more extensive with age when compared toNOD.B10-H2b mice. Focal leukocyte infiltration has been ob-served as early as 6 weeks of age in NOD.B10-H2b.IL4�/�

mice (unpublished observation). Interestingly, IL-2 and IL-12 expression is lost by 12 weeks of age, and as expected,IL-4 was not detected at any time point.

3.8. Induction of salivary gland dysfunction inNOD.B10-H2b.IL4�/� mice following adoptivetransfer of T cells from NOD.B10-H2b

mice capable of IL-4 synthesis

As presented above, NOD.B10-H2b.IL4�/� mice exhibita pre-clinical disease profile quite similar to that ofNOD.B10-H2b mice, including leukocyte infiltration of thesubmandibular gland, production of autoantibodies and ap-pearance of aberrant physiological activities. Nevertheless,NOD.B10-H2b.IL4�/� mice fail to develop clinical disease,maintaining normal saliva flow rates, an observation correlat-ing with the lack of anti-M3R autoantibodies of the IgG1isotype. The cytokine mRNA profile for IL-4 exhibited byNOD.B10-H2b mice indicates high levels of expression at

a well-defined time point in the development of disease,i.e., a time encompassing 12 weeks of age. To determinethe importance of IL-4 expression, NOD.B10-H2b.IL4�/�

mice at 4, 10 and 16 weeks of age were infused withCD4þ T cells isolated from the spleens of similarly agedNOD.B10-H2b.Gfp mice and examined for subsequent onsetof salivary gland dysfunction evaluated by stimulated salivaproduction. As presented in Fig. 7A, saliva volumes ofNOD.B10-H2b.IL4 �/� recipient mice that were adoptivelytransferred with NOD.B10-H2b.Gfp T cells at either 4 or10 weeks of age decreased markedly starting at 16 weeksof age so that by 26 weeks of age, these animals had salivaryflow rates reflective of parental NOD.B10.H2b mice. Serafrom these two sets of adoptively transferred mice containeddetectable IgM and IgG1 anti-M3R antibodies, but no IgG3anti-M3R antibodies (Fig. 7B). In contrast, saliva volumesof NOD.B10-H2b.IL4�/� recipient mice that were adoptivelytransferred with NOD.B10-H2b.Gfp T cells at 16 weeks ofage exhibited reduced, but not significant, loss of saliva se-cretions, even when measured to 36 weeks of age. Most im-portantly, sera from these adoptively transferred micecontained only IgM, and possibly low levels of IgG3, anti-M3R antibodies. IgG1 anti-M3R antibodies were not

99J. Gao et al. / Journal of Autoimmunity 26 (2006) 90e103

Fig. 5. Absence of IgG1 isotypic anti-M3R autoantibody in sera of NOD.B10-H2b.IL4�/� mice. Sera, collected from NOD.B10.H2b and NOD.B10-H2b.IL4�/�

mice at 4 weeks and again at 20 weeks of age, were pre-absorbed on non-transfected Flp-In CHO cells for 2 h at 25 �C. Each serum was then incubated with

mM3R-transfected Flp-In CHO cells for 2 h at 4 �C, the cells washed once with FACS buffer, then incubated with either FITC-conjugated goat anti-mouse

IgG antibody or FITC-conjugated goat anti-mouse IgM, IgA, IgE or IgG (A), IgG1, IgG2a, IgG2b or IgG3 (B) and IgG2a and IgG2c (C) antibodies for

45 min at 4 �C. After a final wash with FACS buffer, the cells were resuspended in FACS buffer and analyzed using a FACScan cytometer. To detect the appearance

of anti-M3R autoantibodies in a serum, the profiles obtained with sera from animals at 20 weeks of age (solid gray histograms) were overlaid onto profiles obtained

from the same animals at 4 weeks of age (open histograms). Isotype-specific controls for each antibody are shown (solid black histograms).

detected (Fig. 7B). These profiles, where IgM anti-M3R anti-bodies are the more dominant subclass, appear to mimic theearly disease in the NOD.B10-H2b or NOD.B10-H2b.IL4�/�

mice (see Fig. 5).

4. Discussion

In the present studies, we have investigated the role of IL-4in the development of SjS-like disease in the NOD mouse

100 J. Gao et al. / Journal of Autoimmunity 26 (2006) 90e103

A. NOD.B10-H2b

B. NOD.B10-H2b.IL4

-/-

4 weeks

IFN-γ IL-1 IL-2 IL-4 IL-6 IL-10 IL-12 G3PDH

8 weeks

12 weeks

20 weeks

4 weeks

8 weeks

12 weeks

20 weeks

Fig. 6. Temporal expressions of major cytokine mRNA transcripts in submandibular glands of NOD.B10.H2b and NOD.B10-H2b.IL4�/�. Total RNA was prepared

from the submandibular glands of NOD.B10-H2b and NOD.B10-H2b.IL4�/� mice at 4, 8, 12 or 20 weeks of age. Total RNA was reverse-transcribed and the re-

sulting cDNA used as template in PCR using cytokine-specific primers. Each PCR product was size separated by electrophoresis and visualized using ethidium

bromide staining. G3PDH was used as a positive control, while lack of IL-4 products using cDNA from NOD.B10-H2b.IL4�/� tissue acted as the negative control.

using our recently constructed IL4-gene KO mouse, NOD.B10-H2b.IL4�/�. Although earlier studies using theNOD.IL4�/� mouse indicated that secretory function of thesalivary gland was not lost if the IL4 gene was non-functional,it was not possible to determine how development of type 1diabetes influenced the outcome. For this reason, we generatedthe NOD.B10-H2b.IL4�/� mouse by crossing NOD.B10-H2b(a NOD mouse line that does not develop type 1 diabetes) withNOD.IL4�/� (a NOD mouse line that carries a disrupted IL4gene, but retains its susceptibility for type 1 diabetes)[4,31,37], then selecting for expression of both the MHC H-2b and dysfunctional IL-4 genes. This newly constructedmouse allows direct comparisons of disease phenotypesamong 4 congenic strains: NOD/Lt, NOD.IL4�/�,NOD.B10-H2b and NOD.B10-H2b.IL4�/�. As anticipated,the NOD.B10-H2b.IL4�/� mouse exhibited a number of char-acteristics observed in disease-prone NOD/Lt and NOD.B10-H2b mice. These included (a) the presence of leukocyticinfiltration of the submandibular and lacrimal glands, (b) theproduction of anti-nuclear antibodies and autoantibodies reac-tive with tissues of the salivary glands, (c) the activation ofa PSP-specific serine protease, and (d) the expression of majorpro-inflammatory cytokines within the salivary glands associatedwith the leukocytic infiltrates. In contrast, two characteristicsof SjS-like disease were not observed in the NOD.B10-H2b.IL4�/� mice. First, there was no temporal loss ofpilocarpine/isoproterenol-stimulated saliva flow rates. As aconsequence, the protein concentrations of the saliva remainedwithin normal ranges, as did the measured amylase activities.Second, there was no temporal appearance of IgG1 isotypicanti-M3R autoantibodies at the age when onset of exocrine

gland dysfunction would be occurring. Taken together, thesephenotypic characteristics suggest that both NOD.IL4�/�

and NOD.B10-H2b.IL4�/� mice express a similar disease phe-notype, solely dependent on the role of IL-4 in development ofthe SjS-like autoimmune exocrinopathy.

Our previous studies using the B lymphocyte-deficientNOD.Igmnull mouse revealed that B cells are essential for theonset of salivary gland dysfunction that leads to a loss ofstimulated secretion of saliva in parental NOD/Lt mice [4,11].Even more revealing, however, have been the observations thatNOD/Lt mice with a dysfunctional IL-4 gene also fail toexhibit loss of stimulated secretion of saliva despite havingnormal numbers of B lymphocytes [4]. IL-4 has been shown tohave at least two major effects on B lymphocytes [38e41]:first, IL-4 regulates B cell proliferation and survival duringB cell ontogeny and activation in part via the IRS signal trans-duction pathways, and second, IL-4 participates in the IgM toIgG1 isotypic switch via the JAK-STAT6 signal transductionpathway. As a cytokine directly involved in B lymphocytegrowth, proliferation, survival and differentiation, as well asregulation of IgG isotype switching [23,37,42e46], IL-4 clearlyplays many important roles in immune responses, one or moreof which appears to be critical to the pathogenesis of SjS-likedisease of the NOD mouse, as shown with the present studies.Results also point to the role IL-4 plays in regulating IgG1 iso-type switching, but studies to identify the precise mechanismare ongoing.

The concept that the effector molecule of clinical diseasemay be restricted to IgG1 isotypic anti-M3R autoantibodiesis supported by the observation that sera of both NOD.B10-H2b and NOD.B10-H2b.IL4�/� mice temporally contain

101J. Gao et al. / Journal of Autoimmunity 26 (2006) 90e103

Fig. 7. Loss of saliva secretory function in NOD.B10-H2b.IL4�/� mice following infusions of CD4þ T cells from NOD.B10-H2b.Gfp mice. (A) Single cell sus-

pensions of spleens freshly explanted from 4, 10 and 16 week old NOD.B10-H2b.Gfp mice were incubated with PE-conjugated goat anti-mouse CD4 antibody for

1 h at 4 �C, washed once, and resuspended in FACS buffer. The labeled cells were subjected to FACS sorting to isolate CD4þ Gfpþ T cells. Sorted T cells (1 � 106

cells in 0.1 ml volume) were adoptively transferred into sex- and age-matched 4 (A), 10 (-) or 16 (:) week-old NOD.B10-H2b.IL4�/� mice (n ¼ 4 per group)

via tail vein injections once a week for 2 weeks. Animals were monitored for salivary secretory function to 26 or 36 weeks of age. Saliva secretory function was

monitored monthly by measuring secreted saliva volumes following intraperitoneal injections of an isoproterenol/pilocarpine secretagogue. (B) Sera, collected

from each of the adoptively transferred NOD.B10-H2b.IL4�/� mice at time of euthanization, were pooled and pre-absorbed on non-transfected Flp-In CHO cells

for 2 h at 25 �C. Each pooled serum was then incubated with either non-transfected or mM3R-transfected Flp-In CHO cells for 2 h at 4 �C, followed by an in-

cubation with either FITC-conjugated goat anti-mouse IgG antibody or FITC-conjugated goat anti-mouse IgG1, IgG2a or IgM antibodies for 45 min at 4 �C. The

cell suspensions were analyzed using a FACScan cytometer. Profiles obtained with sera from NOD.B10-H2b.IL4�/� mice at time of euthanasia (solid histograms)

were overlaid onto profiles obtained from prior to being adoptively transplanted (4 weeks of age) (open histograms).

102 J. Gao et al. / Journal of Autoimmunity 26 (2006) 90e103

a number of different subclasses of anti-mM3R autoantibodies(detected on transfected CHO cells expressing the mouseM3R), but only the sera of NOD.B10-H2b mice contain anti-M3R autoantibodies of IgG1 isotype. In addition, adoptivetransfer into NOD.B10-H2b.IL4�/� mice of T cells capableof IL-4 synthesis at either 4 or 10 weeks of age, but not at16 weeks of age, resulted in production of IgG1 isotypicanti-M3R autoantibodies and subsequent decline in salivarysecretory activity. Whether other isotypic anti-M3R autoanti-bodies, especially those of the IgG3 isotype, can contributeto disease still needs to be determined. Furthermore, whyNOD mice develop a humoral response against exocrine glandacinar cells and autoantibodies reactive with the M3R remainsunknown; however, we have speculated that delayed organo-genesis of the submandibular glands in mice of the NOD ge-netic background may permit the retention of anti-self Blymphocytes reactive with membrane-associated proteins,like M3R, due to delayed expression [5]. Delayed expressionof M3Rs would result in circumvention of clonal deletionand/or self-tolerance [47,48]. Alternatively, activation of theautoimmune response may occur through delayed expressionof antigens such as endogenous neo-antigens or viral proteins[49,50].

An additional interesting observation is the apparent time-dependent expression of IL-4 for development and onset ofthe SjS-like disease in the NOD mouse model. First, detectionof IL-4 mRNA expression in parental NOD.B10-H2b mice isrestricted to a time around 12 weeks of age. Second, onsetof secretory dysfunction in the NOD.B10-H2b.IL4�/� mousesubsequent to adoptive transfer of T cells from NOD.B10-H2b.Gfp mice occurred if the adoptive transfers were carriedout at either 4 or 10 weeks of age (before 12 weeks of ageand/or disease activation), but not if carried out at 16 weeksof age (after 12 weeks of age and/or disease activation).Once again, the timing of IL-4 synthesis correlates with theappearance of anti-M3R autoantibodies.

While immunoglobulins appear to play a major role in xe-rostomia, it is interesting that the degree of leukocyte infiltra-tion within the submandibular and lacrimal glands ofNOD.B10-H2b.IL4�/� mice occurs at an earlier age and be-comes more intense than that observed in NOD.B10-H2b

mice, as measured by the number of foci and the area of infil-tration. Despite the fact that the submandibular gland can benearly over-run by infiltrating leukocytes, the capacity to se-crete normal levels of saliva persists. This is different thanwhat is observed in parental NOD/Lt or NOD.B10-H2b

mice, where onset of xerostomia occurs concomitantly withleukocyte infiltration. These differences strongly suggest a ma-jor role for a systemic soluble factor, like IgG, in the onset ofxerostomia rather than a leukocyte-acinar cell contact. Thismay explain why other mouse models of SjS have severe dis-connects between loss of salivary flow rates and leukocyte in-filtration of the salivary glands. Although IL-4 may control theexpression or synthesis of a specific soluble factor, it is alsopossible that it permits survival and persistence of a uniquecell subpopulation with the capacity to synthesize an inhibitoryfactor. Alternatively, IL-4 may merely alter the physiological

environment leading to injury to the exocrine glands that stim-ulates an autoimmune response. In such a scenario, IL-4 wouldmimic the effects of hormonal changes, especially thoseknown for androgen/estrogen [51e53].

In summary, the use of the IL4 gene knockout mouse,NOD.B10-H2b.IL4�/�, provides an opportunity to study therole of IL-4 in development of the SjS-like autoimmune exo-crinopathy present in the NOD mouse. While one facet of IL-4activity may regulate the survival of autoreactive B cells and,indirectly, degree of lymphocyte infiltration into the salivaryand lacrimal glands, a second facet seems to permit the pro-duction of IgG1 anti-M3R autoantibodies thought to effectexocrine gland dysfunction and promote onset of xerostomiaand/or xerophthalmia. Whether the function of IL-4 in thisSjS-like disease of NOD mice is restricted to the productionof IgG1 isotypic autoantibodies or is involved in multiple fac-ets of the autoimmune response against the submandibularglands will require further studies.

Acknowledgments

This work was supported by National Institutes of HealthGrants DE55304, DE014344 and AI47483 (to A.B.P.). S.C.was supported by a post-doctoral fellowship from the Sjog-ren’s Syndrome Foundation.

References

[1] Fox RI, Kang HI. Pathogenesis of Sjogren’s syndrome. Rheum Dis Clin

North Am 1992;18:517e38.

[2] Fox RI, Michelson P. Approaches to the treatment of Sjogren’s syn-

drome. J Rheumatol Suppl 2000;61:15e21.

[3] Johnson EO, Skopouli FN, Moutsopoulos HM. Neuroendocrine manifes-

tations in Sjogren’s syndrome. Rheum Dis Clin North Am 2000;26:

927e49.

[4] Brayer JB, Cha S, Nagashima H, Yasunari U, Lindberg A, Diggs S, et al.

IL-4-dependent effector phase in autoimmune exocrinopathy as defined

by the NOD.IL-4-gene knockout mouse model of Sjogren’s syndrome.

Scand J Immunol 2001;54:133e40.

[5] Cha S, van Blockland SC, Versnel MA, Homo-Delarche F, Nagashima H,

Brayer J, et al. Abnormal organogenesis in salivary gland development

may initiate adult onset of autoimmune exocrinopathy. Exp Clin Immu-

nogenet 2001;18:143e60.

[6] Nguyen KH, Brayer J, Cha S, Diggs S, Yasunari U, Hilal G, et al.

Evidence for antimuscarinic acetylcholine receptor antibody-mediated

secretory dysfunction in nod mice. Arthritis Rheum 2000;43:2297e306.

[7] Brayer JB, Humphreys-Beher MG, Peck AB. Sjogren’s syndrome: immu-

nological response underlying the disease. Arch Immunol Ther Exp

2001;49:353e60.

[8] Cha S, Peck AB, Humphreys-Beher MG. Progress in understanding au-

toimmune exocrinopathy using the non-obese diabetic mouse: an update.

Crit Rev Oral Biol Med 2002;13:5e16.

[9] Bacman S, Sterin-Borda L, Camusso JJ, Arana R, Hubscher O, Borda E.

Circulating antibodies against rat parotid gland M3 muscarinic receptors

in primary Sjogren’s syndrome. Clin Exp Immunol 1996;104:454e9.

[10] Yamamoto H, Sims NE, Macauley SP, Nguyen KH, Nakagawa Y,

Humphreys-Beher MG. Alterations in the secretory response of non-obese

diabetic (NOD) mice to muscarinic receptor stimulation. Clin Immunol

Immunopathol 1996;78:245e55.

[11] Robinson CP, Brayer J, Yamachika S, Esch TR, Peck AB, Stewart CA,

et al. Transfer of human serum IgG to nonobese diabetic Igmu null

mice reveals a role for autoantibodies in the loss of secretory function

103J. Gao et al. / Journal of Autoimmunity 26 (2006) 90e103

of exocrine tissues in Sjogren’s syndrome. Proc Natl Acad Sci USA

1998;95:7538e43.

[12] Waterman SA, Gordon TP, Rischmueller M. Inhibitory effects of musca-

rinic receptor autoantibodies on parasympathetic neurotransmission in

Sjogren’s syndrome. Arthritis Rheum 2000;43:1647e54.

[13] Jonsson R, Haga H, Gordon T. Sjogren’s syndrome. In: Koopman WJ,

editor. Arthritis and allied conditions: a textbook of rheumatology. Phil-

adelphia: Lippincott Williams and Wilkins; 2001. p. 1736e59.

[14] Kino-Ohsaki J, Nishimori I, Morita M, Okazaki K, Yamamoto Y,

Onishi S, et al. Serum antibodies to carbonic anhydrase I and II in

patients with idiopathic chronic pancreatitis and Sjogren’s syndrome.

Gastroenterology 1996;110:1579e86.

[15] Feist E, Kuckelkorn U, Dorner T, Donitz H, Scheffler S, Hiepe F, et al.

Autoantibodies in primary Sjogren’s syndrome are directed against pro-

teasomal subunits of the alpha and beta type. Arthritis Rheum 1999;

42:697e702.

[16] Billaut-Mulot O, Cocude C, Kolesnitchenko V, Truong MJ, Chan EK,

Hachula E, et al. SS-56, a novel cellular target of autoantibody responses

in Sjogren syndrome and systemic lupus erythematosus. J Clin Invest

2001;108:861e9.

[17] Harley JB, Alexander EL, Bias WB, Fox OF, Provost TT, Reichlin M,

et al. Anti-Ro (SS-A) and anti-La (SS-B) in patients with Sjogren’s syn-

drome. Arthritis Rheum 1986;29:196e206.

[18] Haneji N, Nakamura T, Takio K, Yanagi K, Higashiyama H, Saito I, et al.

Identification of alpha-fodrin as a candidate autoantigen in primary Sjog-

ren’s syndrome. Science 1997;276:604e7.

[19] Reichlin M. Antibodies to Ro and La. Ann Med Interne (Paris) 1998;149:

34e41.

[20] Gao J, Cha S, Jonsson R, Opalko J, Peck AB. Detection of anti-type 3

muscarinic acetylcholine receptor autoantibodies in the sera of Sjogren’s

syndrome patients by use of a transfected cell line assay. Arthritis Rheum

2004;50:2615e21.

[21] Cavill D, Waterman SA, Gordon TP. Antiidiotypic antibodies neutralize

autoantibodies that inhibit cholinergic neurotransmission. Arthritis

Rheum 2003;48:3597e602.

[22] Li J, Ha YM, Ku NY, Choi SY, Lee SJ, Oh SB, et al. Inhibitory effects of

autoantibodies on the muscarinic receptors in Sjogren’s syndrome. Lab

Invest 2004;27:27.

[23] Singh RR. IL-4 and many roads to lupuslike autoimmunity. Clin

Immunol 2003;108:73e9.

[24] Hu-Li J, Shevach EM, Mizuguchi J, Ohara J, Mosmann T, Paul WE.

B cell stimulatory factor 1 (interleukin 4) is a potent costimulant for normal

resting T lymphocytes. J Exp Med 1987;165:157e72.

[25] Howard M, Farrar J, Hilfiker M, Johnson B, Takatsu K, Hamaoka T, et al.

Identification of a T cell-derived b cell growth factor distinct from inter-

leukin 2. J Exp Med 1982;155:914e23.

[26] Defrance T, Aubry JP, Rousset F, Vanbervliet B, Bonnefoy JY, Arai N,

et al. Human recombinant interleukin 4 induces Fc epsilon receptors

(CD23) on normal human B lymphocytes. J Exp Med 1987;165:1459e67.

[27] Noelle R, Krammer PH, Ohara J, Uhr JW, Vitetta ES. Increased expres-

sion of Ia antigens on resting B cells: an additional role for B-cell growth

factor. Proc Natl Acad Sci USA 1984;81:6149e53.

[28] Robinson CP, Cornelius J, Bounous DE, Yamamoto H, Humphreys-

Beher MG, Peck AB. Characterization of the changing lymphocyte pop-

ulations and cytokine expression in the exocrine tissues of autoimmune

NOD mice. Autoimmunity 1998;27:29e44.

[29] Robinson CP, Yamamoto H, Peck AB, Humphreys-Beher MG. Genetical-

ly programmed development of salivary gland abnormalities in the NOD

(nonobese diabetic)-scid mouse in the absence of detectable lymphocytic

infiltration: a potential trigger for sialoadenitis of NOD mice. Clin

Immunol Immunopathol 1996;79:50e9.

[30] Cha S, Brayer J, Gao J, Brown V, Killedar S, Yasunari U, et al. A dual role

for IFN-gamma in the pathogenesis of Sjogren’s syndrome-like autoimmune

exocrinopathy in the NOD mouse. Scand J Immunol 2004;60:552e65.

[31] Robinson CP, Yamachika S, Bounous DI, Brayer J, Jonsson R,

Holmdahl R, et al. A novel NOD-derived murine model of primary Sjog-

ren’s syndrome. Arthritis Rheum 1998;41:150e6.

[32] Tensing EK, Nordstrom DC, Solovieva S, Schauman KO,

Sippo-Tujunen I, Helve T, et al. Salivary gland scintigraphy in Sjogren’s

syndrome and patients with sicca symptoms but without Sjogren’s

syndrome: the psychological profiles and predictors for salivary gland

dysfunction. Ann Rheum Dis 2003;62:964e8.

[33] Diaz-Lopez C, Geli C, Corominas H, Malat N, Diaz-Torner C,

Llobet JM, et al. Are there clinical or serological differences between

male and female patients with primary Sjogren’s syndrome? J Rheumatol

2004;31:1352e5.

[34] Newkirk MM. Rheumatoid factors: what do they tell us? J Rheumatol

2002;29:2034e40.

[35] Robinson CP, Bounous DI, Alford CE, Nguyen KH, Nanni JM,

Peck AB, et al. PSP expression in murine lacrimal glands and function

as a bacteria binding protein in exocrine secretions. Am J Physiol

1997;272:G863e71.

[36] Humphreys-Beher MG, Peck AB. New concepts for the development of

autoimmune exocrinopathy derived from studies with the NOD mouse

model. Arch Oral Biol 1999;44(Suppl 1):S21e5.

[37] Wang B, Gonzalez A, Hoglund P, Katz JD, Benoist C, Mathis D. Inter-

leukin-4 deficiency does not exacerbate disease in NOD mice. Diabetes

1998;47:1207e11.

[38] Jiang H, Harris MB, Rothman P. IL-4/IL-13 signaling beyond JAK/STAT.

J Allergy Clin Immunol 2000;105:1063e70.

[39] Boothby M, Mora AL, Aronica MA, Youn J, Sheller JR, Goenka S, et al.

IL-4 signaling, gene transcription regulation, and the control of effector T

cells. Immunol Res 2001;23:179e91.

[40] Agnello D, Lankford CS, Bream J, Morinobu A, Gadina M, O’Shea JJ,

et al. Cytokines and transcription factors that regulate T helper cell

differentiation: new players and new insights. J Clin Immunol 2003;23:

147e61.

[41] Keegan AD, Nelms K, Wang LM, Pierce JH, Paul WE. Interleukin 4 re-

ceptor: signaling mechanisms. Immunol Today 1994;15:423e32.

[42] Keegan AD, Johnston JA, Tortolani PJ, McReynolds LJ, Kinzer C,

O’Shea JJ, et al. Similarities and differences in signal transduction by in-

terleukin 4 and interleukin 13: analysis of Janus kinase activation. Proc

Natl Acad Sci USA 1995;92:7681e5.

[43] Nakajima A, Hirose S, Yagita H, Okumura K. Roles of IL-4 and IL-12 in

the development of lupus in NZB/W F1 mice. J Immunol 1997;158:

1466e72.

[44] Peng SL, Moslehi J, Craft J. Roles of interferon-gamma and interleukin-4

in murine lupus. J Clin Invest 1997;99:1936e46.

[45] Ruger BM, Erb KJ, He Y, Lane JM, Davis PF, Hasan Q. Interleukin-4

transgenic mice develop glomerulosclerosis independent of immunoglob-

ulin deposition. Eur J Immunol 2000;30:2698e703.

[46] Hill N, Sarvetnick N. Cytokines: promoters and dampeners of autoimmu-

nity. Curr Opin Immunol 2002;14:791e7.

[47] Kyewski B, Derbinski J. Self-representation in the thymus: an extended

view. Nat Rev Immunol 2004;4:688e98.

[48] Mathis D, Benoist C. Back to central tolerance. Immunity 2004;20:

509e16.

[49] Hunziker L, Recher M, Macpherson AJ, Ciurea A, Freigang S,

Hengartner H, et al. Hypergammaglobulinemia and autoantibody

induction mechanisms in viral infections. Nat Immunol 2003;4:

343e9.

[50] Yarilin DA, Valiando J, Posnett DN. A mouse herpesvirus induces re-

lapse of experimental autoimmune arthritis by infection of the inflamma-

tory target tissue. J Immunol 2004;173:5238e46.

[51] Bao M, Yang Y, Jun HS, Yoon JW. Molecular mechanisms for gender dif-

ferences in susceptibility to T cell-mediated autoimmune diabetes in non-

obese diabetic mice. J Immunol 2002;168:5369e75.

[52] Pearce RB, Formby B, Healy K, Peterson CM. Association of an

androgen-responsive T cell phenotype with murine diabetes and Idd2.

Autoimmunity 1995;20:247e58.

[53] Shim GJ, Warner M, Kim HJ, Andersson S, Liu L, Ekman J, et al.

Aromatase-deficient mice spontaneously develop a lymphoproliferative

autoimmune disease resembling Sjogren’s syndrome. Proc Natl Acad

Sci USA 2004;101:12628e33. Epub 12004 Aug 12616.