Sjögren’s syndrome (SS) is an autoimmune disease in which exocrine tissues are affected by cellular and humoral immunity. As a result, the salivary and lacrimal glands of patients with SS are damaged, leading to xerostomia (dry mouth) and keratoconjunctivitis sicca (dry eyes). Because experimental approaches to investigate SS pathogenesis in human patients are limited, development of a mouse model is indispensable for understanding the disease. In this study, we show that special AT-rich sequence binding protein-1 conditional knockout (SATB1cKO) mice, in which the SATB1 gene is specifically deleted from hematopoietic cells, develop SS by 4 wk of age, soon after weaning. Female mice presented an earlier onset of the disease than males, suggesting that female SATB1cKO mice are more susceptible to SS. T cell–dominant immune cell infiltration was observed in the salivary glands of 4 wk old SATB1cKO mice, and the frequency of B cells gradually increased as the mice aged. Consistently, levels of anti-SSA and anti-SSB Abs were increased around 8 wk of age, after salivary production reached its lowest level in SATB1cKO mice. These results suggest that SATB1cKO mice can be a novel SS model, in which the progression and characteristics of the disease resemble those of human SS.
Autoimmune diseases occur when immune tolerance is broken and the immune system attacks the organism’s own cells and tissues (1). In healthy conditions, autoreactive T cells are eliminated by negative selection during T cell ontogeny in the thymus (2). However, because some developing T cells escape from negative selection, autoreactive T cells are present in the periphery even in healthy individuals (3). Therefore, peripheral tolerance, a mechanism suppressing the activation of autoreactive T cells, is important for the maintenance of immune tolerance together with central tolerance.
Various humoral and cellular factors are involved in the establishment of peripheral tolerance. Among them, regulatory T (Treg) cells play an important role in the suppression of autoreactive T cells because Treg deficiency causes aggressive and fatal systemic autoimmune manifestations in both humans and mice (4). This loss of immune tolerance results in various autoimmune diseases (3). One such systemic autoimmune disorder is Sjögren’s syndrome (SS), with clinical hallmarks of keratoconjunctivitis sicca (dry eyes) and xerostomia (dry mouth) (5).
The etiology of SS in humans remains unclear, although several different pathophysiological models have been proposed (6). SS pathogenesis appears to strongly depend on humoral immunity because autoantibodies such as anti-SSA/Ro and anti-SSB/La are observed in patients with SS (5, 7). However, T cells might play a role in SS onset because mice with T cell–specific knockout of Id3 or PI3K develop SS-like diseases (8, 9). Although SS model mice share some characteristics with human SS patients, not all diagnostic criteria are fulfilled in any known mouse model of SS, even if keratoconjunctivitis sicca or xerostomia are observed (6, 10). Therefore, an SS mouse model that better recapitulates the course of human SS progression and presents all SS diagnostic criteria would facilitate the understanding of SS pathogenesis in humans.
Special AT-rich sequence binding protein-1 (SATB1) is a genome organizer that has been identified in a human Jurkat T cell line and binds AT-rich sequences (also known as base-unpairing regions) (11, 12). SATB1 forms complexes with SWI/SNF factors and regulates the expression of numerous genes (12–14). SATB1 expression is abundant in thymocytes among hematopoietic cells, although it is even observed in hematopoietic stem cells (15, 16). SATB1 is also expressed in nonhematopoietic cells, resulting in the death of SATB1-null mice by 3 wk of age (14). Therefore, to determine the role of SATB1 in hematopoietic cells, we generated SATB1 floxed mice and crossed them with Vav-Cre mice (SATB1fl/fl Vav-Cre+ mice) (17). In these SATB1 conditional knockout (cKO) mice, T cell numbers are severely reduced because positive selection is impaired. Negative selection is also impaired in SATB1cKO mice, in which levels of anti-dsDNA Abs are increased at adulthood (17). Therefore, SATB1cKO mice are autoimmune-prone. However, the details of autoimmune manifestations in SATB1cKO mice remain unclear.
In this study, we demonstrated that SATB1cKO mice develop an SS-like disease that fulfills almost all criteria necessary for diagnosis of SS in humans. Infiltration of T and B lymphocytes was observed in the salivary glands of SATB1cKO mice as early as 4 wk of age. Treg cells were absent in SATB1cKO mice for a week after birth, although the number increased to near normal levels with age. Dramatic improvement of saliva production was observed when Treg cells from adult wild-type (WT) mice or SATB1cKO mice were injected into 3 d old SATB1cKO mice. Therefore, the lack of Treg cells in the first week after birth as well as defective negative selection in the thymus might cause SS in SATB1cKO mice. In addition, SATB1cKO mice would be a useful mouse model for clarification of SS etiology.
Materials and Methods
SATB1fl/fl Vav-Cre+ mice were generated as previously described (17), and used in this study as SATB1cKO mice. Vav-Cre mice (18) were purchased from the Jackson Laboratory (Bar Harbor, ME). RAG2−/− mice were maintained in our laboratory. C57BL/6 mice were obtained from CLEA Japan (Tokyo, Japan). All mice presented the C57BL/6 background and were maintained under specific pathogen-free conditions at the animal facility of Toho University School of Medicine. We used both male and female mice with no distinction in this study except for the examination shown in Fig. 1, as stated in the figure legends. All experiments using mice received approval from the Toho University Administrative Panel for Animal Care (16-55-118) and Recombinant DNA (16-52-117).
The following Abs were used for cell-surface and intracellular staining: B220 (RA3-6B2) Pacific Blue and CD25 (PC61) PE were purchased from BioLegend (San Diego, CA). CD4 (GK1.5) PE-Cy7 and CD8 (53-6.7) Pacific Blue were purchased from TONBO Biosciences (San Diego, CA). Thy1.2 (30-H12) PE and Foxp3 (FJK-16a) FITC were purchased from eBioscience (San Diego, CA). Cells from the thymus, spleen, or lymph nodes were depleted of erythrocytes by hypotonic lysis and stained with fluorophore-conjugated Abs. Intracellular staining for Foxp3 was performed after fixation and permeabilization according to the manufacturer’s protocol (eBioscience). Samples were analyzed or sorted with FACSCanto II or FACSAria III flow cytometers (BD Biosciences), and the data were analyzed with FlowJo software (Tree Star, Ashland, OR).
Saliva secretion test
Mice were anesthetized by i.p. injection of 0.75 mg/kg medetomidine (Nippon Zenyaku Kogyo, Koriyama, Japan), 4 mg/kg midazolam (Sandoz, Yamagata, Japan), and 5 mg/kg butorphanol tartrate (Meiji Seika Pharma, Tokyo, Japan), and then injected i.p. with 0.5 mg/kg pilocarpine hydrochloride (Sigma-Aldrich, St. Louis, MO) to stimulate saliva production. Saliva was collected with a 20 μl micropipette after pilocarpine injection for 15 min. The volume of saliva was normalized relative to body weight.
Histopathology and immunohistochemistry
Mouse tissues were fixed in 10% formalin solution (Wako, Osaka, Japan) and embedded in paraffin. Sections were stained with H&E and observed using a BX63 microscope (Olympus, Tokyo, Japan). Immunofluorescence microscopic analysis was performed with fluorophore-conjugated Abs on frozen tissue sections. Salivary glands were frozen in OCT compound (Sakura Finetek, Tokyo, Japan) and fixed after cryostat sectioning with 4% PFA for 10 min at room temperature (RT). Fixed sections were washed twice with PBS and blocked with 5% BSA and 5% rat serum in PBS for 1 h at RT. The sections were then incubated with CD4 (GK1.5) Alexa Fluor 594, CD8 (53-6.7) Alexa Fluor 647, and B220 (RA3-6B2) Alexa Fluor 488 (all from BioLegend) overnight at 4°C. Kidneys were frozen in OCT compound and fixed after cryostat sectioning with −30°C acetone for 30 s. Fixed sections were washed three times with PBS and blocked with 10% goat serum and 5% rat serum in PBS for 1 h at RT. The sections were then stained with goat-anti mouse IgG (H+L) Fab-Alexa Fluor 488 (Jackson ImmunoResearch Laboratories, West Grove, PA), goat-anti mouse IgM, human ads-TXRD (SouthernBiotech, Birmingham, AL), and rat anti-mouse C3 (11H9; Abcam, Cambridge, U.K.) overnight at 4°C. The slides were then blocked using an avidin/biotin blocking kit according to the manufacturer’s protocol (Vector Laboratories, Burlingame, CA). Anti-rat IgG-biotin was then added and incubated for 30 min at RT. Following a PBS wash, Alexa Fluor 647–streptavidin (Molecular Probes, Eugene, OR) was added for 30 min at RT. After staining, slides were mounted with Dako fluorescence mounting medium (Dako, Santa Clara, CA). Images were observed with an A1R confocal laser-scanning microscope (Nikon, Tokyo, Japan).
The extent of inflammatory lesions in the salivary glands was scored by counting the number of lymphocyte aggregates composed of >50 lymphocytes per 4 mm2 of tissue (19, 20). Focus score 1 indicates a single focus composed of >50 lymphocytes per 4 mm2. Whole salivary gland section images were generated by stitching together multiple images using a BX63 microscope and cellSens software (Olympus), and all scoring was performed using cellSens software.
ELISA using mouse serum anti–SS-A, anti–SS-B, and antinuclear Abs (ANA) was performed using Mouse Anti-SSA, Anti-SSB, and ANA Total Ig ELISA kits, respectively (α Diagnostic, San Antonio, TX).
Isolation of salivary gland–infiltrating cells
Salivary glands were cut into small pieces and digested using 0.5 mg/ml collagenase A (Roche, Basel, Switzerland) in the presence of 0.2 mg/ml DNAse I (Sigma-Aldrich) for 35 min at 37°C. After incubation, EDTA (final concentration: 20 mM) was added to stop the enzyme reaction. These cells were passed through a 70 mm filter (BD Biosciences) and washed. Then the cell pellet was resuspended in 10 ml of 40% (v/v) Percoll (GE Healthcare Bioscience, Uppsala, Sweden), overlaid with 2 ml of 80% (v/v) Percoll, and centrifuged at 2400 rpm for 20 min at RT without accelerator and brake. Salivary gland infiltrated cells were collected from the interphase and washed with PBS.
Urine protein test
Urinary proteins were measured using Pretest 3aII (Wako Pure Chemical, Osaka, Japan).
Adoptive transfer of T cells
Thy1.2+ B220− cells were sorted from cervical lymph nodes or spleens of SATB1cKO donor mice (12–15 wk old) using a FACSAria III cell sorter. Purified donor cells (5 × 106) were transferred into Rag2−/− recipient mice (8–12 wk old) by i.v. injection. Saliva secretion tests were performed at 8 wk post-transfer.
Adoptive transfer of Treg cells
CD4+CD25+ Treg cells were sorted from the spleen and lymph nodes of WT or SATB1cKO donor mice using a FACSAria III cell sorter. Donor Treg cells (3 × 105) were adoptively transferred into 3 d old SATB1cKO mice by i.p. injection. Saliva secretion tests were performed at 15 wk post-transfer.
Statistical analysis was performed using the Student t test to compare means or the Mann–Whitney U test with assumption of differing variance and a confidence level of 95%.
SATB1cKO mice develop SS
We previously reported that SATB1cKO mice are autoimmune-prone, as levels of autoantibodies such as anti-dsDNA in the serum increase with age (17). The presence of anti-dsDNA Abs is highly correlated with systemic lupus erythematosus (SLE) in humans (21), suggesting that systemic autoimmune disorders occur in SATB1cKO mice. However, because SATB1cKO mice start to die at ∼24 wk of age, it is possible that some other autoimmune diseases might occur earlier. To uncover the pathogenicity associated with the lack of immune tolerance in the absence of SATB1, we first identified organs functionally affected by autoimmune manifestation in young SATB1cKO mice.
Saliva production was significantly decreased in SATB1cKO mice, but not in WT littermates, by 4 wk of age (Fig. 1A). At this age, the saliva production of female SATB1cKO mice but not male SATB1cKO mice was reduced (Fig. 1B), implying that female mice are more susceptible to SS as observed in human SS cases (22). No difference in saliva production levels between male and female SATB1cKO mice was observed after 8 wk of age. Impairment of saliva production in SATB1cKO mice was accompanied by immune cell infiltration and the destruction of the gland structure (Fig. 1C, 1D). Saliva production reached the lowest level in SATB1cKO mice at ∼8 wk of age (Fig. 1A), whereas destruction of salivary gland structures continued even after 8 wk of age (Fig. 1C), suggesting that the loss of gland function may occur prior to the progression of gland destruction. The magnitude of immune cell infiltration also progressively increased as the mice aged (Fig. 1D). Therefore, the symptom severity may not reflect the degree of organ destruction in autoimmune diseases. Consistently, although immune cell infiltration was observed in various organs of SATB1-deficient mice such as the pancreas (17, 23), impairment of glucose tolerance was not observed in SATB1cKO mice (data not shown).
Next, we examined the types of lymphocytes that infiltrated the salivary glands in SATB1cKO mice. In 4 wk old SATB1cKO mice, half of the infiltrated cells were CD4+ T cells (Fig. 2A, 2B). CD8+ T cells were clearly observed as well. Of note, because the expression of CD4 and CD8 is deregulated in T cells in the absence of SATB1, CD4+CD8+ T cells, which are not normally observed in the periphery, were also present in the salivary glands of SATB1cKO mice (Fig. 2B). The number of B cells gradually increased in SATB1cKO mice and these cells dominated in older mice (Fig. 2). This B cell–dominant infiltration is also observed in patients with SS (24), suggesting that the course of SS in SATB1cKO mice resembles that observed in human SS cases. SATB1cKO mice fulfilled another criterion for SS diagnosis: the presence of anti-SSA and anti-SSB Abs (5). However, the levels of anti-SSA and anti-SSB Abs in SATB1cKO mice were significantly higher than those in WT littermates at 10 wk and 7 wk of age, respectively (Fig. 3), suggesting that levels of anti-SSA and anti-SSB Abs might increase not when gland functions are impaired, but at the chronic phase of SS after damage progression in salivary tissues.
Lupus nephritis occurs in SATB1cKO mice after failure of salivary gland function
SATB1cKO mice start to die at ∼24 wk of age as we previously reported (17). Increased levels of anti-dsDNA Abs imply that SLE occurs in aged SATB1cKO mice as the presence of anti-dsDNA Abs is one of the factors for diagnosis of SLE (25). Other SLE-associated autoantibodies, such as ANA, were also detected at high levels in SATB1cKO mice after 15 wk of age (Fig. 4A). Because lupus nephritis is one of the phenotypes of SLE, we tested urine from SATB1cKO mice at various ages. We found that some of these SATB1cKO mice, but none of the age-matched WT littermates, clearly had proteinuria after 15 wk of age (Fig. 4B). Additionally, IgM and IgG deposition, most likely a part of immune complexes, was observed around the glomerulus in the kidneys of SATB1cKO mice at 15 wk of age (Fig. 4C). A component of complement C3 was also present in association with Ig deposition (Fig. 4C), suggesting that kidney lesions resembling lupus nephritis developed in aged SATB1cKO mice. In contrast, no kidney lesions were observed in SATB1cKO mice at 4–5 wk of age (Fig. 4B, 4C), in which saliva production was already decreased (Fig. 1A). These data indicate that renal failure due to kidney lesions is a cause of death of SATB1cKO mice. In humans, primary SS is an exocrine-tissue specific autoimmune disease and is not associated with other autoimmune disorders such as SLE, whereas secondary SS occurs in addition to other autoimmune diseases (5). Therefore, SS in SATB1cKO mice can be considered as secondary.
T cells play a primary role in SS pathogenicity in SATB1cKO mice
SATB1 is expressed specifically in T cells in mature hematopoietic cells, although it has been recently reported that dendritic cells also express SATB1 (26). Therefore, we examined whether SS pathogenicity in SATB1cKO mice would be solely present in T cells. To this end, we purified T cells from cervical lymph nodes or spleens of WT and SATB1cKO mice at 12–15 wk of age and injected the cells into RAG2−/− mice (8–12 wk of age). Although too few T cells were isolated from cervical lymph nodes of WT mice to be applied to this assay, the cervical lymph nodes of SATB1cKO mice bulged, and sufficient T cells for this experiment could be collected. We measured saliva production after 8 wk of injection and found that the saliva volume from RAG2−/− mice receiving SATB1-deficient T cells from cervical lymph nodes was significantly lower than that from RAG2−/− mice without any treatment (Fig. 5A). In contrast, no change in saliva production was observed in RAG2−/− mice when splenic T cells from either WT or SATB1cKO mice were injected (Fig. 5A). Acinar cell destruction was observed in the salivary glands of RAG2−/− mice receiving T cells from cervical lymph nodes of SATB1cKO mice (Fig. 5B). These results demonstrate that T cells play a significant role in the pathogenesis of SS in SATB1cKO mice, at least during the early phase of the disease.
If peripheral tolerance is perfectly established, autoimmune diseases should not occur even if autoreactive T cells are present in the periphery. The importance of Treg cells in neonates has been emphasized for suppression of autoimmune diseases in mice (4). Because the number of Treg cells in the periphery is not significantly different between SATB1cKO and WT mice at adulthood (above 5 wk of age) (17), we next examined Treg cell numbers in younger SATB1cKO mice and compared them to Treg cell numbers of age-matched WT mice. Foxp3+CD25+ Treg cells were observed as early as 3 d after birth in the spleen of WT littermates (Fig. 6A) as previously reported (27). In contrast, splenic Treg cells were absent in SATB1cKO mice from birth until 1 wk of age (Fig. 6A). No significant difference in the number of Treg cells was observed after 2 wk of age between SATB1cKO and WT littermates (Fig. 6B).
To investigate the relationship between the lack of Treg cells in the first week after birth and the development of SS in SATB1cKO mice, we transferred Treg cells derived from the spleen of matured WT mice into 3 d old SATB1cKO mice and examined the appearance of SS symptoms by measuring saliva production. Although saliva production was still reduced in SATB1cKO mice injected with WT Treg cells, xerostomia was significantly improved in SATB1cKO mice by WT Treg cell transfer (Fig. 6C). In accordance with this observation, infiltration of lymphocytes in salivary glands was detected in SATB1cKO mice with WT Treg cell transfer, but at a much lesser degree compared with control SATB1cKO mice without cell transfer (Fig. 6D). In addition, Treg cells from adult SATB1cKO mice protected against the development of SS at a similar or higher better level compared with WT Treg cells (Fig. 6C). These results strongly imply that the absence of Treg cells in the first week after birth is one of the etiologic factors for SS occurrence in SATB1cKO mice.
SS is an autoimmune disease, in which external glands are targets of autoreactive lymphocytes, leading to impaired production of saliva and tears (28). One factor in the diagnosis of SS is the high level of autoantibodies such as anti-SSA and anti-SSB Abs, which is uniquely observed in patients with SS (5). These anti-SSA and anti-SSB Abs and other autoantibodies are suggested to play a role in SS progression (7, 29). In addition, accumulation of B cells that produce autoantibodies reacting to external glands is observed in patients with SS (30). There are various mouse models of SS with different etiologies and disease phenotypes. For example, in NZB/W F1 mice, infiltration of immune cells, most of which are T cells, is observed in the salivary and lacrimal glands by 4 mo of age when symptoms of lupus nephritis are observed (31, 32). Therefore, SS-like symptoms in NZB/W F1 mice resemble secondary SS as in the case of SATB1cKO mice. In another model, BAFF transgenic mice develop SS after 1 y of age with infiltration of B cells in salivary glands without an increase of anti-SSA or anti-SSB Abs (33, 34), which is different from the typical SS phenotype observed in humans. Therefore, damage to exocrine glands by autoimmunity can be induced by various causes, although the importance of IFN-γ in the etiology of human SS has been emphasized recently (6).
In SATB1cKO mice, CD4+ T cells were the major cell infiltrates around 4 wk after birth and are presumably responsible for the onset of the disease and the destruction of salivary and lacrimal glands before the levels of anti-SSA and anti-SSB Abs increase. Although we cannot rule out the possibility that autoantibodies other than anti-SSA and anti-SSB Abs are involved in the destruction of exocrine glands in SATB1cKO mice at the early phase (younger than 5 wk of age), it is possible that tissue damage is first caused by cellular immunity, followed by humoral immunity, because transfer of T cells from SATB1cKO mice is sufficient to induce the development of SS in RAG2−/− mice (Fig. 5). We should note that transfer of SATB1cKO T cells from the cervical lymph nodes, but not from the spleen, caused SS in RAG2−/− mice, suggesting that pathogenic T cells may be enriched in regional lymph nodes of exocrine glands. In addition, although B cell malignancy is occasionally accompanied by SS in humans (7), we have not observed development of B cell leukemia or lymphoma in SATB1cKO mice thus far (data not shown). Further investigation is necessary to clarify the pathophysiology of SS in SATB1cKO mice.
In the current study, Treg cell deficiency was observed in SATB1cKO mice until 1 wk after birth. Similar results have been obtained with a different SATB1cKO mouse line (SATB1fl/fl CD4Cre+ mice), although Treg cells are present at 1 wk of age in this line (23). Treg cells play an important role in the establishment of peripheral tolerance. Because autoreactive T cells are present in the periphery, the lack of Treg cells results in the development of autoimmune diseases (4). In particular, the presence of Treg cells soon after birth is critical for prevention of autoimmune diseases thereafter in mice (27). Accordingly, if Treg cells are transferred into SATB1cKO mice at 3 d after birth, the SS phenotype becomes less severe (Fig. 6C). Therefore, the presence of Treg cells in the 1 wk window after birth, at least in the case of SATB1cKO mice, is critical for suppression of autoreactive T cells and the development of autoimmune diseases. Interestingly, although the suppression of SS by injection of Treg cells at day 3 after birth in SATB1cKO mice is partial (Fig. 6C, 6D), the production of anti-dsDNA Abs is almost completely blocked (data not shown), suggesting that distinct pathogenic T cells might be responsible for SS and lupus nephritis development in SATB1cKO mice. This also suggests that pathogenic T cells in SS are already activated before or on day 3 after birth in SATB1cKO mice.
We recently demonstrated that SATB1 is indispensable for the development of thymic Treg cells, but not for the differentiation of peripheral Treg (pTreg) cells (23). Therefore, the Treg cells in the periphery of SATB1cKO mice are mostly pTreg cells. Because Treg cells derived from SATB1cKO mice suppressed the development of SS in SATB1cKO mice to a degree similar to that of Treg cells from WT mice (Fig. 6C), it is highly possible that both thymic Treg and pTreg cells present a similar suppressive function against activation of T cells involved in the pathogenesis of autoimmune diseases. We should note that SS symptoms can be transferred in RAG2−/− mice by injecting SATB1-deficient T cells that contain Treg cells (Fig. 5). Therefore, once pathogenic T cells are activated, the Treg cells may not be able to suppress activity of such T cells.
In conclusion, we found in this study that SS develops in SATB1cKO mice because of the lack of pTreg cells for a week after birth in addition to defective negative selection. These results indicate that SATB1cKO mice provide an ideal SS model to investigate the etiology of SS and for the development of new treatments for SS. Further studies with SATB1cKO mice will help to elucidate the molecular mechanisms underlying the establishment of immune tolerance as well.
We thank Dr. Toshihiro Nanki for discussion and Yutaka Yazaki and Toyonobu Eguchi for technical assistance.
This work was supported by a Japan Society for the Promotion of Science KAKENHI (JP24390121 to M.K. and Y.T., JP26670240 to M.K.); a Strategic Research Foundation Grant-aided Project for Private Schools at Heisei 26th (S1411015) from the Ministry of Education, Culture, Sports, Science and Technology to M.K.; a Research Promotion Grant from the Toho University Graduate School of Medicine (14-02 to M.K.); Project Research Grants from the Toho University School of Medicine (23-4, 26-22 and 27-10 to Y.T., 27-14 to A.I.); the Public Foundation of the Vaccination Research Center to M.K.; Initiative for Realizing Diversity in the Research Environment from the Japan Science and Technology Agency to Y.T.; a National Institutes of Health Grant (R37CA39681) to T.K.-S.; and a Grant-in Aid for Private University Research Branding Project from the Ministry of Education, Culture, Sports, Science and Technology to M.K.
Abbreviations used in this article:
peripheral regulatory T
special AT-rich sequence binding protein-1
SATB1 conditional knockout
systemic lupus erythematosus
The authors have no financial conflicts of interest.