Kemin Chen

Decreasing TNF-α results in less fibrosis and earlier resolution of granulomatous experimental autoimmune thyroiditis

Granulomatous experimetal autoimmune thyroiditis (G‐EAT) is induced in DBA/1 mice by adoptive transfer of mouse thyroglobulin (MTg)‐primed spleen cells. TNF‐α is an important proinflammatory cytokine and apoptotic molecule involved in many autoimmune diseases. To study its role in G‐EAT, anti‐TNF‐α mAb was given to recipient mice. Disease severity was comparable between mice with or without anti‐TNF‐α treatment at days 19–21, the time of maximal severity of G‐EAT, suggesting TNF‐α is not essential for development of thyroid inflammation. However, thyroid lesions resolved at day 48 in anti‐TNF‐α‐treated mice, while thyroids of rat Ig‐treated controls had fibrosis. These results suggested that reducing TNF‐α contributed to resolution of inflammation and inhibited fibrosis. Gene and protein expression of inflammatory molecules was examined by RT‐PCR and immunostaining, and apoptosis was detected using TUNEL staining and an apoptosis kit. Thyroids of anti‐TNF‐α‐treated controls had reduced proinflammatory and profibrotic molecules, e.g., IFN‐γ, IL‐1β, IL‐17, inducible NOS and MCP‐1, at day 19 compared with thyroids of rat Ig‐treated mice. There were more apoptotic thyrocytes in rat Ig‐treated controls than in anti‐TNF‐α‐treated mice. The site of expression of the anti‐apoptotic molecule FLIP also differed between rat Ig‐treated and anti‐TNF‐α‐treated mice. FLIP was predominantly expressed by inflammatory cells of rat Ig‐treated mice and by thyrocytes of anti‐TNF‐α‐treated mice. These results suggest that anti‐TNF‐α may regulate expression of proinflammatory cytokines and apoptosis in thyroids, resulting in less inflammation, earlier resolution, and reduced fibrosis.

Inhibition of TGFβ1 by Anti-TGFβ1 Antibody or Lisinopril Reduces Thyroid Fibrosis in Granulomatous Experimental Autoimmune Thyroiditis

In this study, a murine model of granulomatous experimental autoimmune thyroiditis (G-EAT) was used to determine the role of TGFβ1 in fibrosis initiated by an autoimmune inflammatory response. The fibrotic process was evaluated by staining thyroid tissue for collagen, α-smooth muscle actin, TGFβ1, and angiotensin-converting enzyme (ACE), and measuring serum thyroxine in mice given anti-TGFβ1 or the ACE inhibitor lisinopril. The role of particular inflammatory cells in fibrosis was tested by depletion experiments, and the cytokine profile in thyroids was examined by RT-PCR. Neutralization of TGFβ1 by anti-TGFβ1 or lisinopril resulted in less collagen deposition and less accumulation of myofibroblasts, and levels of active TGFβ1 and ACE were reduced in thyroids of treated mice compared with those of untreated controls. Other profibrotic molecules, such as platelet-derived growth factor, monocyte chemotactic protein-1, and IL-13, were also reduced in thyroids of anti-TGFβ1- and lisinopril-treated mice compared with those of controls. Confocal microscopy showed that CD4+ T cells and macrophages expressed TGFβ1. Fibrosis was reduced by injection of anti-CD4 mAb on day 12, when G-EAT was very severe (4–5+). Together, these results suggest a critical role for TGFβ1 in fibrosis initiated by autoimmune-induced inflammation. Autoreactive CD4+ T cells may contribute to thyroid fibrosis through production of TGFβ1. This G-EAT model provides a new model to study how fibrosis associated with autoimmune damage can be inhibited.

Mechanisms of Spontaneous Resolution versus Fibrosis in Granulomatous Experimental Autoimmune Thyroiditis

When granulomatous experimental autoimmune thyroiditis (G-EAT) was induced in CBA/J or DBA/1 mice, thyroid lesions resolved in less severe (3+) G-EAT in wild-type mice or severe (5+) G-EAT in IFN-γ−/− mice, but progressed to fibrosis in 5+ G-EAT in wild-type mice. To define the mechanisms leading to these distinct outcomes, the expression of inflammatory and apoptotic molecules and infiltrating cells was evaluated using immunohistochemistry, RT-PCR, and confocal microscopy. The ratio of CD4+/CD8+ T cells in thyroid infiltrates was one factor that predicted G-EAT outcome. CD4+ T cells outnumbered CD8+ T cells when lesions progressed to fibrosis, while CD8+ T cells outnumbered CD4+ T cells in thyroids that resolved. Fas, Fas ligand, FLIP, TNF-α, inducible NO synthase, TGF-β, and IFN-γ were highly expressed by infiltrating cells when G-EAT progressed to fibrosis. The expression of active caspase-3 was low, possibly contributing to the persistence of CD4+ T cells in fibrosis. In contrast, FLIP was mainly expressed by thyrocytes in resolving G-EAT, the expression of active caspase-3 was high, and resolution correlated with apoptosis of infiltrating cells. There was also relatively less expression of TGF-β, IFN-γ, TNF-α, and inducible NO synthase and higher expression of IL-10 in resolving G-EAT than in G-EAT that progressed to fibrosis. These differences were particularly striking when comparing IFN-γ−/− vs wild-type mice. These results suggest that several opposing biological mechanisms contribute to the outcome of an ongoing autoimmune response. These include differential expression of pro- and antiapoptotic molecules, cytokines, and the ratio of CD4+ vs CD8+ T cells.

Expression and Regulation of Fas and Fas Ligand on Thyrocytes and Infiltrating Cells During Induction and Resolution of Granulomatous Experimental Autoimmune Thyroiditis

Granulomatous experimental autoimmune thyroiditis (G-EAT) is induced by mouse thyroglobulin-sensitized spleen cells activated in vitro with mouse thyroglobulin, anti-IL-2R, and IL-12. G-EAT lesions reach maximal severity 19–21 days after cell transfer, and lesions almost completely resolve by day 35. Depletion of CD8+ cells delays resolution and reduces Fas ligand (FasL) mRNA expression in thyroids. This study was undertaken to analyze Fas and FasL protein expression in the thyroid during induction and resolution of G-EAT and to determine whether CD8+ cells might regulate Fas or FasL expression in the thyroid. Fas and FasL expression was analyzed by immunohistochemical staining or in situ hybridization in thyroids of mice with or without depletion of CD8+ cells. Fas and FasL proteins were not detectable in normal thyroids, but expression of both proteins increased during development of G-EAT. Fas was expressed primarily by inflammatory cells; some enlarged thyrocytes were also Fas+. Thyrocytes had intense FasL immunoreactvity, and many CD8+ cells were also FasL positive. Depletion of CD8+ cells resulted in decreased FasL expression by thyrocytes and inflammatory cells, but had no effect on Fas expression. TUNEL assay detected many apoptotic inflammatory cells in proximity to thyrocytes. CD8-depleted thyroids had ongoing inflammation with fewer apoptotic infiltrating cells at day 35. Administration of a neutralizing anti-FasL mAb had no apparent effects on development of G-EAT, but anti-FasL was as effective as anti-CD8 in preventing G-EAT resolution. These results suggested that CD8+ T cells and thyrocytes may kill inflammatory cells through the Fas pathway, contributing to G-EAT resolution.

Role of TGFβ in Development of Spontaneous Autoimmune Thyroiditis in NOD.H-2h4 Mice

Nearly 100% of NOD.H-2h4 mice develop spontaneous autoimmune thyroiditis (SAT) and produce anti-mouse thyroglobulin autoantibodies when they receive 0.05% NaI in their drinking water beginning at 8 wk of age. Our previous studies showed that TGFβ1 mRNA was constitutively expressed in thyroids and spleens of normal NOD.H-2h4 mice but not other strains of mice. To determine whether TGFβ might have a role in SAT, mice were given anti-TGFβ mAb at various times during development of SAT. Anti-TGFβ markedly inhibited development of SAT and production of anti-mouse thyroglobulin IgG1 autoantibodies. Anti-TGFβ was most effective in inhibiting SAT when given during the time thyroid lesions were developing, i.e., starting 4 wk after administration of NaI water. The active form of the TGFβ1 protein was present in thyroids of mice with SAT but not in normal NOD.H-2h4 thyroids. However, thyrocytes of normal NOD.H-2h4 thyroids did express latent TGFβ1. TGFβ1 protein expression in the thyroid correlated with SAT severity scores, and administration of anti-TGFβ inhibited TGFβ1 protein expression in both the thyroid and spleen. TGFβ1 was produced primarily by inflammatory cells and was primarily localized in areas of the thyroid containing clusters of CD4+ T and B cells. Depletion of CD8+ T cells had no effect on TGFβ1 protein expression. Activation of splenic T cells was apparently not inhibited by anti-TGFβ, because up-regulation of mRNA for cytokines and other T cell activation markers was similar for control and anti-TGFβ-treated mice. TGFβ1 may function by promoting migration to, or retention of, inflammatory cells in the thyroid.

Characterization of thyroid fibrosis in a murine model of granulomatous experimental autoimmune thyroiditis.

This study was initiated to identify and characterize thyroid fibrosis in a murine model of granulomatous experimental autoimmune thyroiditis (G‐EAT) and determine if TGF‐β1 might be involved in fibrosis. G‐EAT was induced by transfer of mouse thyroglobulin‐sensitized spleen cells activated in vitro with thyroglobulin, anti‐IL‐2R, and IL‐12. There was almost complete destruction of thyroid follicles, leading to fibrosis of the gland and reduced serum T4 levels. Fibrosis was confirmed by staining for collagen and α smooth‐muscle actin, a marker of myofibroblasts. Kinetic studies characterized the onset and development of thyroid fibrosis. TGF‐β1 was increased at mRNA and protein levels, and expression of TGF‐β1 protein paralleled G‐EAT severity. Comparison of staining patterns showed that TGF‐β1 was expressed in areas of myofibroblast and collagen accumulation, implying that TGF‐β1 may play a role in fibrosis in G‐EAT. Further studies demonstrated that myofibroblasts, macrophages, and thyrocytes contributed to TGF‐β1 production. This provides an excellent model to study the mechanisms of fibrosis associated with autoimmune damage.

Chemokine expression during development of fibrosis versus resolution in a murine model of granulomatous experimental autoimmune thyroiditis

Severe granulomatous eperimental autoimmune thyroiditis (G‐EAT) in DBA/1 or CBA/J wild type (WT) mice at day 19 progresses to fibrosis by day 35, but severe G‐EAT in DBA/1 interferon (IFN)‐γ−/− mice or less‐severe G‐EAT at day 19 in WT mice resolves by day 35. To study the role of chemokines in autoimmune diseases and fibrosis, profiles of chemokines and chemokine receptors were analyzed in DBA/1 WT versus IFN‐γ−/− and CBA/J thyroids, which have distinct outcomes of autoimmune inflammation. Gene expression of CXC chemokine ligand 1 (CXCL1) and CXC chemokine receptor 2 (CXCR2) paralleled neutrophil infiltration and thyrocyte destruction in DBA/1 WT or CBA/J thyroids, and gene expression of CC chemokine ligand 11 (CCL11), CCL8, and CC chemokine receptor 3 paralleled eosinophil infiltration in IFN‐γ−/− thyroids. Gene and protein expression of CXCL10, CXCL9, and CXCR3 was significantly lower in IFN‐γ−/− compared with DBA/1 WT thyroids. Moreover, immunostaining showed that CXCL10 was expressed by thyrocytes and inflammatory cells, and strong expression of CXCL10 by thyrocytes was as early as day 7. High expression of CCL2 was only observed in severely destroyed DBA/1 WT or CBA/J thyroids, which would develop fibrosis. Thus, the differential expression of chemokines may direct distinct cellular populations in DBA/1 WT versus IFN‐γ−/− thyroids. Up‐regulation of CXCL10 by thyrocytes suggests its role in regulating the recruitment of specific subsets of activated lymphocytes to the thyroid during autoimmune inflammation. The early expression of CXCL1, CXCL10, and CCL2 may suggest their involvement in the initiation and perpetuation of disease in severe G‐EAT thyroids, which progress to fibrosis.

Balance of proliferation and cell death between thyrocytes and myofibroblasts regulates thyroid fibrosis in granulomatous experimental autoimmune thyroiditis (G-EAT)

Severe granulomatous experimental autoimmune thyroiditis (G‐EAT), which progresses to fibrosis, is induced in DBA/1 mice by adoptive transfer of mouse thyroglobulin‐primed and ‐activated spleen cells. There is extensive destruction of thyrocytes and inflammatory cell infiltration including T cells, macrophages, neutrophils, and myofibroblasts (myofbs). Suppression of transforming growth factor‐β (TGF‐β) and deficiency of interferon‐γ (IFN‐γ) inhibit fibrosis, and inflammation eventually resolves. Thyrocyte destruction in wild‐type (WT) mice was a result of apoptosis, as many deoxynucleotide triphosphate nick‐end labeling + apoptotic thyrocytes were present in these thyroids. The balance of apoptosis and proliferation between thyrocytes and myofbs may be important factors determining the outcome of inflammation to fibrosis versus resolution. Apoptosis and proliferation in thyrocytes versus myofbs were evaluated by dual‐staining of cell‐proliferating marker (Ki‐67) or in situ cell death and cytokeratin or α‐smooth muscle actin and were analyzed by confocal microscopy. Apoptotic and antiapoptotic molecules in G‐EAT thyroids were detected by immunostaining. In WT thyroids, which develop fibrosis, only a few myofbs were apoptotic, and many myofbs were Ki‐67+, Fas‐associated death domain protein‐like interleukin‐1β‐converting enzyme‐like inhibitory protein (FLIP)+, and Bcl‐XL+. In contrast, proliferation was predominant on thyrocytes of IFN‐γ−/− mice or anti‐TGF‐β‐treated WT mice. These results indicate that apoptosis of inflammatory cells and regeneration of thyrocytes in IFN‐γ−/− mice and anti‐TGF‐β‐treated WT mice may limit development of fibrosis, whereas excessive proliferation of myofbs and loss of thyrocytes in WT mice may contribute to fibrosis.

FLIP and FasL expression by inflammatory cells vs thyrocytes can be predictive of chronic inflammation or resolution of autoimmune thyroiditis

Spontaneous autoimmune thyroiditis (SAT) in NOD.H-2h4 mice is a model of chronic inflammation of the thyroid, while granulomatous experimental autoimmune thyroiditis (G-EAT) is a model with spontaneous resolution of inflammation. In chronic inflammation (SAT), Fas, FasL, and FLIP were upregulated and predominant in inflammatory cells. There were few apoptotic cells, and low expression of active caspase-8 and -3. In resolving G-EAT in CBA/J and NOD.H-2h4 mice, FasL and FLIP were predominantly expressed by thyrocytes. There were many apoptotic inflammatory cells, and increased expression of active caspase-8 and -3. Depletion of CD8+ T cells inhibited G-EAT resolution and resulted in chronic inflammation. FLIP was expressed predominantly by inflammatory cells, and apoptosis of inflammatory cells and expression of active caspase-3 was reduced as in chronic SAT. Thus, differences in expression of pro- or antiapoptotic molecules in SAT or G-EAT were apparently related to the acute vs chronic nature of the inflammatory response rather than the method of disease induction. Upregulation of FLIP by inflammatory cells may block Fas-mediated apoptosis, contributing to chronic inflammation, whereas increased FLIP expression by thyrocytes in resolving G-EAT may protect thyrocytes from apoptosis, and FasL expression by thyrocytes may induce apoptosis of inflammatory cells, contributing to resolution.

Fas Ligand Is Required for Resolution of Granulomatous Experimental Autoimmune Thyroiditis

We previously suggested that CD8+ T cells promoted resolution of granulomatous experimental autoimmune thyroiditis (G-EAT) at least in part through regulation of Fas ligand (FasL) expression on thyroid epithelial cells. To directly evaluate the role of the Fas pathway in G-EAT resolution, Fas- and FasL-deficient mice on the NOD.H-2h4 background were used as recipients of activated G-EAT effector cells. When MTg-primed wild-type (WT) donor splenocytes were activated and transferred to WT recipients, thyroid lesions reached maximal severity on day 20 and resolved on day 50. Fas, FasL, and FLIP were up-regulated, and many apoptotic inflammatory cells were detected in recipient thyroids on day 20. Fas was predominantly expressed by inflammatory cells, and FasL and FLIP were mainly expressed by thyroid epithelial cells. After depletion of CD8+ T cells, G-EAT resolution was delayed, FLIP and FasL were predominantly expressed by inflammatory cells, and few inflammatory cells were apoptotic. When WT donor splenocytes were transferred to gld recipients, disease severity on day 20 was similar to that in WT recipients, but resolution was delayed. As in CD8-depleted WT recipients, there were few apoptotic inflammatory cells, and FLIP and FasL were expressed primarily by inflammatory cells. These results indicated that the expression of functional FasL in recipient mice was critical for G-EAT resolution. WT cells induced minimal disease in lpr recipients. This was presumably because donor cells were eliminated by the increased FasL on lpr recipient cells, because donor cells were not eliminated, and the mice developed G-EAT if lpr recipients were given anti-FasL mAb.