Abstract
Spontaneous autoimmune thyroiditis (SAT) is an organ-specific autoimmune disease characterized by chronic inflammation of the thyroid by T and B lymphocytes. To investigate the roles of Th1 and Th2 cytokines in the pathogenesis of SAT, IFN-γ−/− and IL-4−/− NOD.H-2h4 mice were generated. IL-4−/− mice developed lymphocytic SAT (L-SAT) comparable to that of wild-type (WT) mice. They produced little anti-mouse thyroglobulin (MTg) IgG1, but had levels of anti-MTg IgG2b comparable to WT mice. Compared with WT mice, IFN-γ−/− mice produced significantly less anti-MTg IgG1 and IgG2b. Absence of IFN-γ resulted in abnormal proliferation of thyroid epithelial cells with minimal lymphocyte infiltration. Thyroids of IFN-γ−/− mice had markedly reduced B lymphocyte chemoattractant expression, B cell and plasma cell infiltration, and decreased MHC class II expression on thyrocytes compared with WT mice. Adoptive transfer of WT splenocytes to IFN-γ−/− mice restored the capacity to develop typical L-SAT, enhanced anti-MTg IgG1 and IgG2b production, up-regulated MHC class II expression on thyrocytes and decreased thyrocyte proliferation. These results suggest that IFN-γ plays a dual role in the development of SAT. IFN-γ is required for development of L-SAT, and it also functions to inhibit thyroid epithelial cell proliferation.
Spontaneous autoimmune thyroiditis (SAT)3 in NOD.H-2h4 mice is a chronic organ-specific autoimmune disease characterized by infiltration of thyroids with CD4+ and CD8+ T cells and B cells (1, 2, 3, 4, 5). Levels of mouse thyroglobulin (MTg)-specific autoantibodies generally correlate well with SAT severity scores (2, 6). B cells and CD8+ T cells are required for the initiation of SAT (2, 6), whereas CD4+ T cells are required not only for the initiation of SAT, but also for the maintenance of SAT lesions (2, 3). It is an excellent model to study the mechanisms of autoreactive T and B cell interaction and the roles of cytokines in autoimmune diseases.
IFN-γ and IL-4 are prototypic Th1- and Th2-type cytokines, respectively (7). Functions of IFN-γ include activation of macrophages, induction of MHC class I and II expression for Ag presentation, up-regulation of adhesion molecules, and recruitment of Th1 cells to sites of inflammation (8). IFN-γ can either enhance or suppress autoimmune thyroiditis depending on the experimental conditions (9, 10, 11, 12, 13, 14). IL-4, on the other hand, stimulates Th2 T cell development, activates B cells, induces MHC class II expression on B cells, and induces Ig class switching to IgG1 (15, 16). The role of IL-4 in autoimmune pathogenesis is still unclear. In some studies, IL-4 suppressed organ-specific autoimmune diseases (17, 18). However, IL-4 could exacerbate experimental autoimmune uveoretinitis (19) and adjuvant-induced arthritis (20), and IL-4 gene-disrupted mice develop most autoimmune diseases including experimental autoimmune thyroiditis (EAT) (21).
Most animal models of autoimmune disease are induced by Ags plus adjuvants (22). NOD.H-2h4 mice spontaneously develop SAT (1, 2, 3, 4, 5, 6). Increased dietary iodine accelerates development of SAT in NOD.H-2h4 mice (1, 2, 3), humans, obese strain (OS) chickens, and BB/Wor rats (23, 24, 25). Recent work showed that enhanced iodination of thyroglobulin facilitates processing and presentation of a cryptic pathogenic peptide and this might partly account for the association of high iodine intake and autoimmune thyroiditis (26). The goal of this study was to investigate the role of IFN-γ and IL-4 in the development of SAT. Using IFN-γ and IL-4 gene KO mice eliminates potential problems due to incomplete cytokine neutralization with Ab treatment and allows a more detailed analysis of autoimmune parameters. The results presented here show that IFN-γ is required for the development of typical lymphocytic SAT (L-SAT). Most IFN-γ−/− mice developed thyroid lesions characterized by thyroid follicular cell proliferation with relatively few lymphocytes, while lack of IL-4 had no effect on SAT.
Materials and Methods
Mice
NOD.H-2h4 mice were originally provided by Dr. L. Wicker (Merck, Rahway, NJ) (27) and subsequently bred under specific pathogen-free conditions in our animal facility (2, 5, 6). NOD.H-2h4 mice are I-E negative and express H-2Kk, I-Ak, and Dd on the NOD background (27). IFN-γ−/− and IL-4−/− male NOD mice were provided by Dr. D. Serreze (The Jackson Laboratory, Bar Harbor, ME) and were crossed with NOD.H-2h4 females to generate F1 mice. F1 mice were used to generate F2 mice that were selected for the expression of Kk and absence of Kd by flow cytometry of peripheral blood. The progeny were further screened by PCR amplification of tail DNA for the IFN-γ and IL-4 gene deletions to select IFN-γ−/− and IL-4−/− NOD.H-2h4 mice for further breeding. Mice homozygous for the disrupted IFN-γ and IL-4 genes as well as wild-type (WT) NOD.H-2h4 and IFN-γ × NOD.H-2h4 heterozygous (IFN-γ−/+) mice received 0.05% NaI in their drinking water for 4–12 wk beginning at 7–8 wk of age (2). NOD.Thy1.1 males, obtained from The Jackson Laboratory, were crossed with NOD.H-2h4 females (Thy1.2) to generate NOD.H-2h4.Thy1.1 mice as described above. Mice expressing Thy1.1 but not Thy1.2 as determined by flow cytometry of peripheral blood were selected for further breeding and used as donors for some cell transfer experiments to distinguish donor (Thy1.1+) T cells and recipient (Thy1.2+) T cells.
Evaluation of thyroiditis
At various intervals after receiving NaI water, one thyroid lobe from each mouse was collected, fixed in formaldehyde, sectioned, and stained with H&E as previously described (2, 28). All slides were coded before being scored by two individuals, one of whom had no knowledge of the experimental groups. The other thyroid lobe from each animal was snap frozen in liquid nitrogen and stored at −70°C for use in immunohistochemical staining or for isolation of RNA for RT-PCR. As shown previously (2, 5), thyroid lesions reach maximal severity 8 wk after NOD.H-2h4 mice are given NaI in the drinking water. Thyroids were scored for the extent of normal thyroid follicle replacement or destruction using a scale of 0–5+ as previously described (2, 5, 28). Briefly, a score of 0 indicates a normal thyroid, while 0+ indicates there are mild follicular changes and/or a few inflammatory cells infiltrating the thyroids. The other scores are as follows: 1+ thyroiditis is defined as an inflammatory infiltrate of at least 125 cells in one or several foci or proliferative thyroid epithelial cell changes sufficient to cause replacement of several follicles; 2+ represents 10–20 foci of cellular infiltration, each the size of several follicles, or epithelial cell changes causing replacement or destruction of up to one-fourth of the gland; 3+ indicates that one-fourth to one-half of the gland is destroyed by infiltrating inflammatory cells or proliferative epithelial cell changes; and 4+ indicates that more than one-half of the gland is destroyed. Thyroids given a score of 5+ had few or no remaining intact follicles. Qualitatively, the thyroid inflammatory cell infiltrate seen in WT NOD.H-2h4 mice was typical of that seen in conventional lymphocytic EAT, with large clusters of lymphocytes in thyroids with 2+ or greater severity scores (5). Thyroids of most IFN-γ−/− mice had variable degrees of thyroid epithelial cell enlargement or proliferation, fewer lymphocytes, and no evident lymphocyte clusters. Thyroid lesions in IFN-γ−/− mice graded 0+ to 2+ were characterized by areas containing groups of very small follicles (microfollicles) closely juxtaposed with compression of the interstitial areas. The follicular epithelial cells were enlarged and became cuboidal or columnar, and small numbers of lymphocytes were scattered through these areas. Thyroid follicles were often devoid of colloid. In some areas there was mild epithelial cell proliferation and only mild infiltration of lymphocytes. The more severe lesions in IFN-γ−/− mice (graded 3–5+ based on the percentage of normal thyroid follicular structure remaining) had widespread clusters of proliferating follicular epithelial cells and histiocytes surrounded by lymphocytic infiltration and, in a few cases, collagen deposition evident by H&E staining.
Measurement of serum anti-MTg autoantibody
MTg-specific IgG autoantibodies in serum from individual mice were measured by ELISA as previously described (2). The contribution of various IgG subclasses to the total IgG autoantibody response was assessed using alkaline phosphatase-conjugated Abs specific for IgG1 and IgG2b. Dilutions of the subclass-specific Abs (1/6000–1/8000) were determined in preliminary titrations to detect optimal Ab activity of serum on MTg-coated plates while giving minimal activity (OD < 0.05) on plates coated with an irrelevant protein (OVA) or of normal mouse serum (1/100 dilution) on MTg-coated plates. Results are expressed as A410 of 1/50 dilutions of serum (2, 6). All assays were repeated at least once and sera from several experiments were run at the same time. In the experiment shown in Table II⇓, WT and IFN-γ−/− mice were immunized with MTg (150 μg) plus LPS (15 μg) twice at 10-day intervals. Seven days after the second immunization, serial dilutions of sera were analyzed for MTg-specific Abs as described above. The end point of the titration was considered to be the dilution at which the average OD for a particular group was <0.1.
Immunohistochemical staining
Immunohistochemical staining was done as previously described (55), Thy1.2 and Thy1.1 staining of the same thyroid frozen sections was done according to the Vector Laboratories illustration. The primary staining for Thy1.2 was done as above using VIP as substrate, which gives a purple color. After primary staining, the avidin-biotin blocking kit (Vector Laboratories) was used for 15 min, then biotinylated anti-Thy1.1 and the Vectastain Elite avidin-biotin complex were used for the second color staining. The secondary staining was developed using the SG substrate (Vector Laboratories), which gives a gray color (5).
Cytokeratin staining was done using paraffin sections of thyroids. Slides were deparaffinized in xylene and dehydrated in graded alcohol. The slides were treated by microwave for 10 min twice at 15-min intervals in PBS to abolish endogenous mouse Ig. The remaining steps were the same as for staining of frozen sections. Mouse anti-mouse cytokeratin (1/400, PCK-26; Sigma-Aldrich) was used as primary Ab, and biotinylated goat anti-mouse F(ab′)2 (1/4000; Kirkegaard & Perry Laboratories, Gaithersburg, MD) was used as secondary Ab.
RT-PCR amplification
RT-PCR was done as previously described (2, 5, 21, 29). To determine the relative initial amounts of target cDNA, each cDNA sample was serially diluted 1/5 and 1/25, and each dilution was amplified with cytokine- or target-specific primers as previously described (29). Hypoxanthine phosphoribosyltransferase (HPRT) was used as a housekeeping gene to verify that the same amount of RNA was amplified (29). The primer sequences for HPRT, IFN-γ, IL-4, IL-12, and IL-13 have been previously described (29). PCR products were separated by electrophoresis in 2% agarose gels and visualized by UV light following ethidium bromide staining. Densitometry analysis was performed using an IS-1000 Digital Imaging System (Life Sciences, St. Louis, MO). Samples within the linear relationship between input cDNA and final PCR products (usually 1/25 cDNA dilution) were collected, and the densitometric units for each cytokine band were normalized to that for the corresponding HPRT band (29). A ratio of 100 indicates a 1:1 ratio between a particular cytokine and HPRT.
Adoptive transfer
Splenocytes from 6- to 8-wk-old WT NOD.H-2h4 mice that had not received NaI water were injected i.v. (3.5 × 107 cells) into lightly irradiated (300 rad) IFN-γ−/− recipient mice. WT mice and IFN-γ−/− mice that did not receive WT splenocytes were used as controls and were also irradiated (300 rad). In some experiments, to distinguish the thyroid-infiltrating lymphocytes that came from the WT donors vs the IFN-γ−/− recipient mice, splenocytes from WT Thy1.1-positive NOD.H-2h4 mice were used as donors and lightly irradiated (300 rad) Thy1.2-positive NOD.H-2h4 IFN-γ−/− mice were used as recipients. Beginning on the day of cell transfer, recipient mice received 0.05% NaI water, and thyroids were removed 8 wk later. SAT severity and anti-MTg autoantibodies were assessed as described above.
Student’s t test
Statistical analysis of data was performed using an unpaired two-tailed Student’s t test as indicated in the tables and figure legends. A value of p < 0.05 was considered to be statistically significant.
Results
Impaired L-SAT development in IFN-γ−/− mice but not in IL-4−/− mice
To determine the effect of IL-4 or IFN-γ gene deletions on the development of SAT, mice were given 0.05% NaI in their water and thyroids were removed 8 wk later for histologic evaluation. Representative results from two separate experiments are shown in Table I⇓. As previously reported (2), L-SAT developed in most WT NOD.H-2h4 mice after 8 wk on NaI water. The incidence and severity of L-SAT in IL-4−/− mice was similar to that of WT mice and the histology was comparable (Fig. 1⇓, A–F). In contrast, no IFN-γ−/− mice developed typical lymphocytic thyroiditis, and IFN-γ−/− mice maintained on NaI water for up to 16 wk also did not have L-SAT (data not shown). Although IFN-γ−/− mice did not develop conventional L-SAT, nearly all of them had abnormal thyroids consisting of relatively mild (score, 1–2+) or more severe (score, 3–5+) lesions. These lesions varied from a few sites of thyroid epithelial cell enlargement with microfollicles and a few scattered lymphocytes to more severe changes in which the normal follicular architecture was replaced by large clusters of proliferating thyroid epithelial cells with a relatively modest lymphocytic infiltration (Fig. 1⇓, G–I). The large clusters of lymphocytes typically observed in WT and IL-4−/− mice were not observed in IFN-γ−/− mice (Fig. 1⇓, G and H). IFN-γ heterozygous (IFN-γ−+/−) NOD.H-2h4 mice developed L-SAT similar to that of WT mice (data not shown).
H&E staining of thyroid sections demonstrating L-SAT in WT (A and B) and IL-4−/− mice (D and E). There are many infiltrating lymphocytes forming aggregates in thyroids with 2–3+ L-SAT. These lymphocyte aggregates were not observed in thyroids of IFN-γ−/− mice; they developed lesions predominated by thyroid follicular cell proliferation (G and H). Transfer of WT splenocytes (Sp) into IFN-γ−/− mice (see Materials and Methods) restored development of L-SAT (J and K). Cytokeratin staining of thyroid follicular cells (C, F, I, and L; arrows). Note the masses of proliferating follicular cells in thyroids of IFN-γ−/− mice in I (arrowheads). Original magnification: A, D, G, and J, ×100; B, C, E, F, H, I, K, and L, ×400. Severity scores for each of the thyroids shown was 2–3+ and representative areas of the thyroid are shown.
SAT severity and autoantibody production in WT, IL-4−/−, and IFN-γ−/− NOD.H-2h4 mice
Cytokeratin staining was used to identify thyroid epithelium (30). Some enlarged thyrocytes in WT and IL-4−/− mice and the proliferating thyrocytes in IFN-γ−/− mice were positive for cytokeratin (Fig. 1⇑, C, F, and I). Enlarged thyroid epithelial cells surrounding microfollicles and areas of proliferating epithelial cells resulted in loss of normal thyroid follicles (Fig. 1⇑I, short arrows). Antiproliferating cell nuclear Ag staining indicated that proliferating thyrocytes were proliferating cell nuclear Ag-positive (data not shown). The abnormal proliferation was observed only in the thyroids of mice given NaI water; thyroids of mice not given NaI water and other organs such as spleen, lymph nodes, and salivary glands were normal. Thyroid epithelial cell proliferative lesions with characteristics like those described for the IFN-γ−/− NOD.H-2h4 mice with 3–4+ severity scores were also observed in WT NOD.H-2h4 mice given anti-IFN-γ mAb (data not shown).
The more severe (score, 3–5+) proliferative SAT lesions in IFN-γ−/− mice occurred in 23 of 125 IFN-γ−/− mice examined in these experiments. The majority (n = 84) had scores of 1+ to 2+, whereas the remaining 18 mice had scores of 0+, with obvious follicular abnormalities, but insufficient replacement of normal thyroid follicles for a score of 1+ (data not shown). However, in contrast to thyroids of IFN-γ−/− mice with granulomatous EAT (14), there were no eosinophils and fewer inflammatory cells in thyroids of IFN-γ−/− NOD.H-2h4 mice.
These results indicate that IFN-γ, but not IL-4, is critical for the development of L-SAT and suggest that the absence of IFN-γ can promote proliferation of thyroid epithelial cells.
Autoantibody production in WT, IL-4−/−, and IFN-γ−/− mice
IFN-γ is a key cytokine to promote IgG2a production by B cells (31) and IL-4 is important for the production of IgG1 (32). We determined the levels of MTg-specific autoantibodies in sera of WT, IL-4−/−, and IFN-γ−/− mice after 8 wk on NaI water. As shown previously, WT mice produced MTg-specific IgG1 and IgG2b autoantibodies. Although the incidence and severity of SAT in IL-4−/− mice was comparable to that of WT mice, they produced essentially no detectable MTg-specific IgG1 autoantibody (Table I⇑). However, IL-4−/− mice produced IgG2b autoantibody comparable to that of WT mice. Compared with WT mice, lower levels of MTg-specific IgG1 and IgG2b autoantibodies were produced in IFN-γ−/− mice (Table I⇑). However, total serum IgG levels were similar for both WT and IFN-γ−/− mice (data not shown), and IFN-γ−/− mice had anti-MTg autoantibody responses comparable to those of WT mice when they were immunized with MTg and the adjuvant LPS (Table II⇓). NOD.H-2h4 mice do not produce MTg-specific IgG2a autoantibody (2, 3), and this was also true for IFN-γ−/− (Table II⇓) and IL-4−/− NOD.H-2h4 mice (data not shown).
Anti-MTg autoantibody production in WT and IFN-γ−/− NOD.H-2h4 mice immunized with MTg and LPS
Inflammatory cell distribution and BLC expression in thyroids
Using immunohistochemical staining, the cellular composition and organization of thyroid inflammatory infiltrates of WT, IL-4−/−, and IFN-γ−/− mice after 8 wk on NaI water was examined. As in previous studies, CD4+ T cells and B220+ B cells predominated in WT thyroids, forming lymphoid aggregates, whereas CD8+ T cells were scattered throughout the lesions (5) (Fig. 2⇓, A–C). The phenotype and histologic distribution of inflammatory cells in thyroids of IL-4−/− mice with a 2–3+ SAT was similar to the those of WT mice (data not shown). The number of inflammatory cells was markedly reduced in the thyroids of IFN-γ−/− mice with 2–3+ thyroid epithelial cell proliferation (Fig. 2⇓, H–J). RT-PCR results for CD4, CD8, and Ig-β in thyroids of IFN-γ−/−, IL-4−/−, and WT mice with similar SAT severity scores were consistent with the histologic results (data not shown). As in WT mice, CD8+ T cells in thyroids of IFN-γ−/− mice were scattered throughout the thyroid (Fig. 2⇓I). CD4+ T cells outnumbered CD8+ T cells. Most of the CD4+ T cells were distributed between areas of proliferating thyroid epithelial cells (Fig. 2⇓H, asterisk) and did not form the lymphoid aggregates typically observed in thyroids of WT and IL-4−/− mice with L-SAT. Very few B cells were detected in the thyroids of IFN-γ−/− mice (Fig. 2⇓J), whereas thyroids of IL-4−/− and WT mice had numerous B cells (Fig. 2⇓C). BLC is important for recruiting B cells to inflammatory sites (33). BLC expression was also markedly reduced in thyroids of IFN-γ−/− mice (Fig. 2⇓K) compared with WT mice (Fig. 2⇓D), and this may account for the minimal B cell infiltration of thyroids in the absence of IFN-γ.
Representative sections of thyroid-infiltrating cells (2–3+ severity scores) in WT mice (A–G), IFN-γ−/− mice (H–N), and IFN-γ−/− mice given WT splenocytes (O–U) after 8 wk on NaI water. Thyroids of WT mice with 3+ L-SAT showed many CD4+ T cells (A) and B cells (C) forming aggregates (data not shown). Infiltration of CD4+ T cells (H) and B cells (J) was markedly reduced and there were no lymphocyte aggregates in thyroids of IFN-γ−/− mice with 3+ thyroid follicular cell proliferative lesions. CD8+ T cells were scattered throughout the lesions in both WT and IFN-γ−/− thyroids (B and I). Many BLC-expressing cells were clustered in the thyroids of WT mice with 2–3+ L-SAT (D), while only a few BLC-expressing cells were scattered in the thyroids of IFN-γ−/− mice with 3+ follicular cell proliferation (K). Clustered sydencan-1-positive plasma cells were located near the thyroid follicles in thyroids of WT mice with 2–3+ L-SAT (E), whereas very few plasma cells were observed in thyroids of IFN-γ−/− mice with 2–3+ follicular cell proliferation (L). MHC class II was strongly expressed on thyrocytes (arrows) and infiltrating mononuclear cells in thyroids of WT mice with 2–3+ L-SAT (F). MHC class II expression was considerably reduced on thyrocytes (arrows) of IFN-γ−/− mice with 3–4+ follicular cell proliferation (M), but was strongly expressed on infiltrating mononuclear cells (M). IFN-γ−/− mice that received WT splenocytes had L-SAT comparable to that of WT mice, with aggregates of CD4+ T cells (O) and B cells (Q) and scattered CD8+ T cells (P). BLC-positive cells (R), syndecan-1-positive plasma cells, and MHC class II expression on thymocytes were also similar to those in WT mice (D–F). Thy-1.2-positive lymphocytes (purple) were observed in thyroids of WT (G) and IFN-γ−/− mice (N). After reconstitution of IFN-γ−/− (Thy 1.2+) mice with cells from Thy1.1+ WT mice, both Thy1.1-positive donor WT T cells (gray) and Thy1.2-positive recipient IFN-γ−/− lymphocytes (purple) infiltrated the thyroids of IFN-γ−/− recipients (U). Original magnification, ×400.
Plasma cells in thyroids and spleens of WT, IL-4−/−, and IFN-γ−/− mice
Syndecan-1 (CD138) is a conventional marker to identify plasma cells (34, 35). Anti-CD138 mAb staining showed many plasma cell clusters in the thyroids of IL-4−/− mice (data not shown) and WT mice with L-SAT (Fig. 2⇑E). When viewed at high power, these cells showed the characteristic morphology of mature plasma cells, with a small eccentric nucleus and abundant cytoplasm. The clustered plasma cells were usually located outside of the T and B cell aggregates (data not shown) and near the thyroid follicles (Fig. 2⇑E). Very few plasma cells were detected in the thyroids of IFN-γ−/− mice with follicular proliferative SAT (Fig. 2⇑L), and no plasma cells were observed in normal NOD.H-2h4 thyroids (data not shown). Since few plasma cells were observed in thyroids of IFN-γ−/− mice, it was of interest to examine plasma cells in the spleen. The architecture of the spleen appeared to be normal except there were few or no peanut agglutinin-positive germinal centers in spleens of IL-4−/− and IFN-γ−/− mice compared with WT mice (data not shown). Flow cytometry analysis indicated that there were no significant differences in the relative proportions of CD4+ or CD8+ T cell or B cell populations in spleens of WT, IL-4−/−, and IFN-γ−/− mice (data not shown). Few plasma cells were found in the spleens of NOD.H-2h4 mice that did not receive NaI water. Many clusters of plasma cells were present outside the B follicles in spleens of IL-4−/− and WT mice after 8 wk on NaI water, whereas very few plasma cells were observed scattered between the B cell follicles in the spleens of IFN-γ−/− mice after 8 wk on NaI water (data not shown). These results are consistent with the results in Table I⇑ that showed very low MTg autoantibody production in the sera of IFN-γ−/− mice. Thus, a lack of IFN-γ inhibits intrathyroidal accumulation of B cells and plasma cells and the spontaneous production of anti-MTg autoantibodies.
MHC molecule expression on thyrocytes
IFN-γ is important for induction of MHC class I and II expression on epithelial cells (36). Because IFN-γ−/− mice did not develop L-SAT, this might be associated with a low level of MHC expression on thyrocytes or inflammatory cells. Thyrocytes of IFN-γ−/−, IL-4−/−, and WT mice without SAT expressed MHC class I but did not express MHC class II (data not shown). Thyrocytes of IL-4−/− (data not shown) and WT mice with L-SAT expressed both MHC class I (data not shown) and MHC class II (Fig. 2⇑F). Although thyrocytes of IFN-γ−/− mice with proliferative SAT lesions had comparable MHC class I expression (data not shown), class II expression by their thyrocytes was markedly reduced (Fig. 2⇑M, arrows). In contrast, expression of MHC class II on thyroid-infiltrating mononuclear cells of IFN-γ−/− mice (Fig. 2⇑M) was comparable to that of IL-4−/− (data not shown) and WT mice (Fig. 2⇑F). Similar numbers of CD11b+ inflammatory cells were observed in the thyroids of WT and IL-4−/− mice with L-SAT and IFN-γ−/− mice with follicular proliferative SAT (data not shown). The observation that the MHC class II staining pattern was similar to that of CD11b+ cell staining in the thyroids of IFN-γ−/− mice suggests that most MHC class II-positive mononuclear cells in thyroids of these IFN-γ−/− mice were CD11b+. Thus, MHC class II expression on thyrocytes appears to be more dependent on IFN-γ than is MHC class II expression on mononuclear cells.
Adoptive transfer of WT splenocytes into IFN-γ−/− mice results in development of L-SAT
To begin to understand the basis of SAT pathogenesis in IFN-γ−/− mice, adoptive transfer studies were done to determine whether IFN-γ−/− mice would develop L-SAT after receiving WT splenocytes. WT splenocytes were injected i.v. into IFN-γ−/− recipient mice as described in Materials and Methods. As shown in Table III⇓, IFN-γ−/− recipient mice not receiving WT cells all developed epithelial cell proliferative changes in their thyroids. In contrast, IFN-γ−/− mice receiving WT splenocytes developed L-SAT and had increased levels of MTg-specific IgG1 and IgG2b autoantibodies comparable to those of WT mice (Table III⇓). Further histologic analysis of the thyroids showed intense infiltration of inflammatory cells forming lymphoid aggregates (Figs. 1⇑, J and K, and 2, O–Q). BLC expression was comparable to that of WT mice (Fig. 2⇑R), many B cells and plasma cells infiltrated the thyroids (Fig. 2⇑, Q and S), and thyrocytes strongly expressed MHC class II molecules (Fig. 2⇑T). Most important, thyrocyte epithelial cell proliferation was inhibited in the presence of WT lymphoid cells (Fig. 1⇑L). To determine whether the infiltrating T cells in thyroids of IFN-γ−/− recipients of WT donor cells were all of WT origin or whether T cells from the IFN-γ−/− recipients were also present in the thyroid infiltrates, Thy1.1-positive WT splenocytes were adoptively transferred into Thy1.2 IFN-γ−/− mice. IFN-γ−/− recipients given Thy1.1-positive WT splenocytes also developed L-SAT (data not shown). The results of immunohistochemical staining suggested that both donor (Thy1.1) and recipient T lymphocytes (Thy1.2) infiltrated the recipient thyroids (Fig. 2⇑U), and as shown below, some of the donor T cells expressed IFN-γ mRNA. These results demonstrate that IFN-γ derived from WT lymphocytes is necessary and sufficient for development of L-SAT and that IFN-γ inhibits proliferation of thyroid follicular cells.
IFN-γ−/− NOD.H-2h4 mice develop L-SAT after receiving WT spleen cells
Cytokine expression in spleens and thyroids
Cytokine mRNA expression in spleens and thyroids of some mice was determined by RT-PCR. HPRT was used as a housekeeping gene to normalize cytokine gene expression in spleens and thyroids of different mice. IL-4 and IFN-γ mRNA was undetectable in the spleens and thyroids of IL-4−/− and IFN-γ−/− mice, respectively, confirming their genotype (Fig. 3⇓). Importantly, IFN-γ mRNA was detected in thyroids of IFN-γ−/− mice given WT splenocytes, indicating that IFN-γ-producing cells from WT donors migrated to the thyroids. Because Th1 and Th2 cytokines can cross-regulate each other, we examined whether elimination of IFN-γ or IL-4 resulted in altered expression of other cytokines. Although IL-13 mRNA expression was reduced in spleens of IL-4−/− mice, expression of IL-13 and other cytokines in thyroids of IL-4−/− mice was comparable to that of WT mice except for the absence of IL-4 (data not shown). With the exception of TGF-β, which is constitutively expressed in thyroids of NOD.H-2h4 mice (2), expression of most cytokine mRNAs was markedly reduced in thyroids of IFN-γ−/− mice with 0–1+ severity scores (Fig. 3⇓B). IL-1β (data not shown), IL-12, and IL-13 expression (Fig. 3⇓B) in thyroids of IFN-γ−/− mice with 2–4+ severity scores was comparable to that of WT mice, but expression of IL-4 (Fig. 3⇓B) and TNF-α (data not shown) was low. Cytokine gene expression in thyroids of IFN-γ−/− mice reconstituted with WT spleen cells was comparable to that of WT mice. Expression of several chemokines, including IFN-γ-inducible protein 10 (IP-10), RANTES, BLC, and CXCR-4, was also reduced in the thyroids of IFN-γ−/− mice (data not shown).
Expression of cytokine mRNA in spleens and thyroids. Spleens (A) and thyroid lobes (B) were obtained from five to six individual WT mice, IL-4−/− mice, or IFN-γ−/− mice that received NaI water for 8 wk. Thyroids and spleens from 2- to 3-mo-old normal WT NOD.H-2 h4 mice that did not receive NaI water were used as controls. B, The group designated IFN-γ−/− low-D were thyroids from IFN-γ−/− mice with 0+ to 1+ severity scores, the IFN-γ−/− Hi-D group were from IFN-γ−/− mice with 2+ to 4+ severity scores, and the IFN-γ−/− plus WT splenocytes (Sp) were from IFN-γ−/− mice reconstituted with WT splenocytes. The disease severity scores of the WT and IFN-γ−/− plus WT Sp groups were 2–3+. Cytokine gene expression was determined by RT-PCR as described in Materials and Methods. Data are expressed as the mean ± SEM of the ratio of cytokine:HPRT mRNA of five to six mice per group. Results are representative of two separate experiments. ∗, Compared with WT mice, p < 0.01; ∗∗, p < 0.05.
Discussion
To elucidate more clearly the roles of prototypic Th1 and Th2 cytokines in the pathogenesis of L-SAT, we generated NOD.H-2h4 mice genetically deficient in IFN-γ or IL-4 and demonstrated the critical role of IFN-γ in the pathogenesis of L-SAT. IFN-γ deletion markedly reduced the production of MTg-specific IgG1 and IgG2b autoantibodies, and IFN-γ−/− mice did not develop typical L-SAT. In contrast, IL-4 deletion resulted in diminished MTg-specific IgG1 autoantibody production, but IgG2b autoantibody production was normal, and there was no effect on the development of L-SAT.
An interesting observation in these studies was the development of a different type of thyroid lesion, characterized by thyroid epithelial cell proliferative changes with few lymphocytes, in IFN-γ−/− NOD.H-2h4 mice. In some cases, these lesions progressed to widespread clusters of proliferating thyrocytes and histiocytes surrounded by collagen and lymphocytes, with almost total destruction of the normal thyroid architecture. The more severe pathologic changes observed in some IFN-γ−/− mice were very similar to thyroid lesions that develop in most WT NOD.H-2h4 mice given anti-IFN-γ mAb beginning 2–3 wk after NaI water, when SAT lesions are beginning to develop (data not shown). Future studies will determine whether the timing of IFN-γ neutralization is critical for generating these severe proliferative lesions. As shown here (Table III⇑), adoptive transfer of WT splenocytes into IFN-γ−/− mice inhibited the development of thyrocyte proliferation and resulted in the development of typical L-SAT lesions. Thus, IFN-γ deficiency allows severe thyroid epithelial cell proliferation to develop in NOD.H-2h4 mice. This is in agreement with results of others suggesting that IFN-γ can act as an antiproliferative factor on hemopoietic stem cells (37), human bronchial epithelial cells (38), and Mac-1+ (CD11b+) mononuclear cells (39). IFN-γ has also been shown to promote apoptosis in some models (37, 40, 41), and decreased apoptosis could possibly result in proliferation of thyroid epithelial cells. Clearly, further studies, currently in progress, are needed to determine how IFN-γ inhibits thyrocyte proliferation in this model.
The reasons why epithelial cell proliferative changes were relatively mild in most IFN-γ−/− mice but very marked in some mice is unknown. It is unlikely that lesions had simply not fully developed in the mice with milder lesions, since the incidence of severe proliferative changes was not increased when IFN-γ−/− mice received NaI water for 16 rather than 8 wk (data not shown). IFN-γ is apparently important for the spontaneous production of autoantibodies, because the production of MTg-specific IgG1 and IgG2b autoantibodies was markedly reduced in IFN-γ−/− mice. This was not due to an inability of IFN-γ−/− mice to produce anti-MTg autoantibodies, since they developed autoantibody responses comparable to those of WT mice after immunization with MTg and LPS (Table II⇑). These results suggest that T cells and B cells in IFN-γ−/− mice are not intrinsically dysfunctional or unresponsive to MTg, since immunization with Ag and adjuvant could overcome the deficiency of IFN-γ mice to produce anti-MTg Abs. This supports previous findings that adjuvants could overcome some changes related to the lack of IFN-γ (22). However, in nonobese diabetic mice, IFN-γ−/− mice develop diabetes normally, while adjuvants such as CFA can inhibit diabetes development by an IFN-γ-dependent mechanism (42).
B cells are required for the development of SAT (6). All mice that develop SAT produce anti-MTg autoantibodies and autoantibody levels generally correlate with SAT severity scores (2, 6). It is interesting that large numbers of plasma cells and B cells were detected in the thyroids of WT and IL-4−/− mice with L-SAT, while thyroids of IFN-γ−/− mice had very few B cells or plasma cells. Plasma cells but not B cells were also reduced in the spleens of IFN-γ−/− compared with spleens of IL-4−/− and WT mice (data not shown). It is possible that the lack of IFN-γ affected B cell infiltration in the thyroid due to reduction of BLC (Fig. 2⇑K) or other B cell-attracting chemokines in thyroids of IFN-γ−/− mice (data not shown). Such low numbers of B cells and plasma cells infiltrating IFN-γ−/− thyroids may further affect T cell and B cell interaction locally, resulting in few T cell and B cell aggregates in IFN-γ−/− thyroids. Our observation that there was reduced infiltration of T cells and B cells in thyroids of IFN-γ−/− mice may be due to deficient trafficking or homing of cells as has been reported by others for IFN-γ−/− mice and mice with low levels of IFN-γ (43, 44). The clusters of plasma cells in thyroids of WT NOD.H-2h4 mice were localized outside the CD4+ T cell/B cell clusters and close to the thyroid follicles (Fig. 2⇑E). Whether the plasma cells mature locally in the thyroid or whether they first mature in secondary lymphoid organs and then migrate to thyroids is unknown. Recent work in autoimmune NZB/W mice suggested that plasma cells were initially generated in secondary lymphoid organs, then accumulated and persisted in inflamed kidneys (45). However, other studies demonstrated that plasma cells matured locally at the site of inflammation in rheumatoid arthritis patients (46). Additional studies are needed to determine which of these mechanisms is operative in this SAT model.
The reasons for the failure of L-SAT to develop in the absence of IFN-γ may include both immunologic and nonimmunologic mechanisms. Thyrocytes have been reported to express increased MHC class I and II after lymphocytic infiltration (4, 47). In the current studies, lack of IFN-γ severely impaired MHC class II expression on thyrocytes and this may have affected local immune responses within thyroids. Such a mechanism has been suggested in other studies for the absence of diabetes in IFN-γ−/− mice (48) and for the requirement for IFN-γ in murine lupus (49). IP-10 has been shown to play a critical role in recruiting inflammatory cells to sites of inflammation in some autoimmune diseases (43, 50). Our preliminary results showed that expression of BLC (Fig. 2⇑), IP-10, RANTES, and CXCR-4 mRNA was markedly reduced in the thyroids of IFN-γ−/− mice (data not shown). Reduced expression of BLC, IP-10, or other chemokines or adhesion molecules may impair the recruitment of inflammatory cells to the thyroid. This in turn could affect the local environment, resulting in reduced MHC class II expression on thyrocytes and reduced expression of cytokines in thyroids (Fig. 3⇑). However, MHC class II expression on thyroid-infiltrating inflammatory cells and MHC class I on thyrocytes was comparable in both IFN-γ−/− and WT mice. The reason why MHC class I expression on thyrocytes was apparently not reduced is unclear, although MHC class I is expressed constitutively by NOD.H-2h4 thyrocytes, but MHC class II is not (data not shown). Furthermore, in the MHC class II transactivator gene promoter pIV KO mice, there was selective loss of IFN-γ-induced MHC class II expression on non-bone marrow-derived cells, but the deletion of PIV had no effect on MHC class I expression (51).
Adoptive transfer of WT splenocytes to IFN-γ−/− mice induced up-regulation of MHC class II expression on their thyrocytes (Fig. 2⇑T), enhanced MTg autoantibody production, inhibited thyrocyte proliferation, and restored development of L-SAT (Table III⇑ and Fig. 1⇑). These mice expressed IFN-γ (Fig. 3⇑) as well as B cell/CD4+ T cell aggregates and plasma cells in thyroids (Fig. 2⇑). By adoptive transfer of Thy1.1-positive WT splenocytes into Thy1.2 IFN-γ−/− mice, the thyroid-infiltrating T cells that derived from the WT donors were distinguished from those of the IFN-γ−/− recipients. These results showed that both the donor IFN-γ+ WT splenocytes (Thy1.1) and recipient IFN-γ−/− lymphocytes (Thy1.2) infiltrated the recipient thyroids (Fig. 2⇑U). Because many of the infiltrating T cells were from the IFN-γ−/− recipients, these results suggest that in the presence of IFN-γ-producing WT cells, cells from IFN-γ−/− mice can migrate to the thyroid and contribute to the development of typical L-SAT. Because many of the cells in the thyroid infiltrates are B cells which obviously do not express Thy1, it would also be important to know whether the B cells derive from the WT or the IFN-γ−/− mice. Studies to address this issue are currently in progress.
Although IL-4 can suppress some autoimmune diseases (17, 18), IL-4-deficient NOD.H-2h4 mice developed L-SAT and had relative proportions of CD4+ and CD8+ T cells, B cells, and plasma cells in their thyroids comparable to those of WT mice. The fact that IL-4-deficient mice did not have accelerated or increased severity of SAT suggests that IL-4 does not play a protective role in SAT. Because IL-4−/− mice did not produce MTg-specific IgG1 autoantibody, our results also indicate that IgG1 is not a critical isotype for the development of L-SAT.
These results stress the importance of Th1 cytokines in the pathogenesis of L-SAT and are consistent with our previous results indicating that the expression of Th1 cytokines was predominant before expression of Th2 cytokines in thyroids of NOD.H-2h4 mice with L-SAT (2). Because IL-4 expression in L-SAT thyroids was maximal at or after the time of maximal SAT severity, we suggested IL-4 might be important for maintaining chronic inflammation and for clustering of CD4+ T cells and B cells in thyroids (2). The present results clearly indicate this is not the case, since IL-4−/− mice had typical CD4+/B cell clusters in their thyroids, and SAT severity was maintained when IL-4−/− mice received NaI water for up to 4 mo (data not shown).
In conclusion, this study demonstrated that IFN-γ was essential for the development of L-SAT and that the absence of IFN-γ was associated with development of a proliferative epithelial cell histopathologic process that was inhibited by transfer of WT splenocytes. These results are consistent with other reports demonstrating a role for IFN-γ in inhibiting some inflammatory processes and potentiating others (8, 9, 10, 11, 12, 13, 14, 22, 42, 48, 52, 53). Our results may also be relevant for understanding the possible role of IFN-γ in other autoimmune diseases such as granulomatous vasculitis described in MRL/Mp-lpr/lpr mice (54). These studies described pathologic findings similar to those described here for the hypertrophied and proliferating thyroid epithelial cells with interspersed lymphocytes in IFN-γ−/− mice, although the relationship of the pathology to IFN-γ was not examined in those studies. Further studies, currently in progress, are needed to elucidate the nature of the cytokines and/or chemokines involved in the development of proliferative thyroid epithelial cell lesions in IFN-γ−/− mice and to determine how IFN-γ protects against this proliferative pathologic process.
Acknowledgments
We thank Patti Mierzwa and Jennifer Campbell for skilled technical assistance and Louise Barnett for performing the flow cytometry analyses. We also thank Dr. David Serreze (The Jackson Laboratory) for providing the breeding stock of NOD.IL-4−/− and NOD.IFN-γ−/− mice and Dr. Edward Leiter (The Jackson Laboratory) for providing the breeding stock of NOD.Thy1.1 mice.
Footnotes
-
↵1 This work was supported by a Merit Review Grant from the Veterans Affairs and by the Children’s Miracle Network, the Missouri Chapter of the Arthritis Foundation, and the A. P. Green Foundation.
-
↵2 Address correspondence and reprint requests to Dr. Helen Braley-Mullen, Department of Medicine, Division of Immunology and Rheumatology, University of Missouri, M450 Medical Sciences, Columbia, MO 65212. E-mail address: mullenh{at}health.missouri.edu
-
↵3 Abbreviations used in this paper: SAT, spontaneous autoimmune thyroiditis; L-SAT, lymphocytic SAT; MTg, mouse thyroglobulin; BLC, B lymphocyte chemoattractant; WT, wild type; HPRT, hypoxanthine phosphoribosyltransferase; IP-10, IFN-γ-inducible protein 10.
- Received April 22, 2002.
- Accepted July 19, 2002.
- Copyright © 2002 by The American Association of Immunologists