Expression of Class II Major Histocompatibility Complex Molecules on Thyrocytes Does Not Cause Spontaneous Thyroiditis but Mildly Increases
http://www.100md.com
内分泌学杂志 2005年第3期
Department of Pathology (H.K., M.K., S.-C.T., Y.-C.C., N.R.R., P.C.), The Johns Hopkins Medical Institutions, Baltimore, Maryland 21205; and Department of Microbiology (K.S.), Leprosy Research Center, National Institute of Infectious Diseases, Tokyo 189-0002, Japan
Address all correspondence and requests for reprints to: Patrizio Caturegli, M.D., M.P.H., Johns Hopkins Medical Institutions, Department of Pathology, Ross Building, Room 656, 720 Rutland Avenue, Baltimore, Maryland 21205. E-mail: pcat@jhmi.edu.
Abstract
Class II major histocompatibility complex (MHC) molecules are classically expressed on antigen-presenting cells of the hematopoietic lineage but have also been described on epithelial cells in association with autoimmunity. In this context, however, it remains debatable whether class II MHC molecules are the initiating event or rather the consequence of the autoimmune attack. In addition, the role of epithelial class II expression once the autoimmune attack has begun is unknown. We generated transgenic mice expressing in the thyroid follicular cells the class II transactivator, the master regulator of all the genes in the class II MHC pathway. The study used a cohort of 245 CBA/J mice (127 wild-type and 118 transgenic), both in basal conditions (n = 63) and at different time points after immunization with mouse thyroglobulin (n = 182). In basal conditions, transgenic mice were similar to wild-type controls and did not develop spontaneous autoimmune thyroiditis, despite the aberrant expression of class II MHC molecules on thyrocytes. After immunization, thyroiditis was 8% more severe in transgenics than controls (95% confidence interval from 1.8–13.4%; P = 0.033), especially during the florid stages of disease. These findings suggest that expression of class II MHC molecules on epithelial cells is not sufficient to initiate autoimmunity but mildly modulates an already established autoimmune attack against the target organ.
Introduction
AUTOIMMUNE DISEASES OF the thyroid gland, represented by Graves’ disease and Hashimoto’s thyroiditis and collectively referred to as autoimmune thyroiditis, are the most common autoimmune diseases (1) and an excellent model of organ-specific autoimmunity (2). They are chronic conditions initiated by loss of immunological tolerance to thyroid-restricted self-antigens (such as TSH receptor, thyroperoxidase, and thyroglobulin), a loss that leads to thyroid dysfunction via immune-mediated inflammation and ultimately to the clinical phenotype.
The loss of tolerance is considered to result from the combined effect of environmental factors, such as iodine, and gene variants. For example, individuals with an A-to-G polymorphism in the noncoding, 3' region of cytotoxic T lymphocyte antigen 4 gene have a 2.3-fold or 1.45-fold increased risk of developing Graves’ disease or Hashimoto’s thyroiditis, respectively (3). The first described and most potent genetic influence known to affect susceptibility to autoimmune thyroiditis, however, is the major histocompatibility complex (MHC). The association between MHC and autoimmunity was originally reported in 1971, first in patients with systemic lupus erythematosus (4) and then in mice with experimental autoimmune thyroiditis (5).
The MHC is located on the short arm of the human chromosome 6, spanning 3.6 megabases of complete and annotated DNA sequence (6). Its main functions are to guide the intracellular processing of antigens and to display antigenic peptides on the plasma membrane for recognition by T lymphocytes (7). It is the most gene-dense region of the genome, containing 224 loci (96 pseudogenes and 128 expressed genes) grouped in three classes. Class II MHC genes are constitutively expressed only on hematopoietic cells involved in antigen presentation (dendritic cells, macrophages, B lymphocytes, and cortical thymic epithelial cells) but can be induced by inflammatory stimuli on many other cell types (such as endothelial cells, hepatocytes, ?-cells of the pancreas, and thyrocytes). Little is known about the biological significance of this observation, although it has been implicated in allograft rejection (8) and subsequently in autoimmunity. In this field, Bottazzo and Feldman hypothesized that class II MHC molecules, when aberrantly expressed on epithelial cells, confer to them antigen-presenting functions favoring the initiation of autoimmune diseases (9, 10).
The MHC is also the region that has been associated with the greatest number of human diseases, ranging from narcolepsy (11) to primary immunodeficiencies (12) to most, if not all, autoimmune conditions (13, 14). The strength of this association, which is classically measured by relative risks (RR) or odds ratios, varies from low, such as a RR of 2.5 for myasthenia gravis in a person carrying the DR3 allele (15), to medium, such as a RR of 30 for DQ8 and type 1 diabetes (16), to high, such as a RR of 109 for A29 and birdshot retinochoroidopathy (17). For Graves’ disease, the MHC (DR3) association is mild, with RRs ranging from 1.5–3.5 (reviewed in Ref.18). For Hashimoto’s thyroiditis, the contribution of MHC haplotypes to disease is less definitive, although they may play a role in the familial clustering of thyroiditis and diabetes (19).
Given the biological plausibility, the replication of the findings, and the strength of many of these associations, MHC molecules are considered in the causal pathway that leads to autoimmune disease. There are, however, important factors that could confound the above-described association. First, MHC loci are found together at frequencies much higher than would be expected by random combination of their frequencies in the population. It is thus possible that the genes causing disease are not the MHC genes per se but rather genes that travel with the MHC. For example, polymorphisms of the LMP2 and LMP7 genes, which are embedded within the class II region, have been associated with ankylosing spondylitis (20). In addition, the same allele that is associated with a particular disease is also common in the normal, unaffected population, and actually the majority of patients with the susceptible allele do not develop autoimmune disease. Finally, it is unknown whether to initiate autoimmunity MHC molecules have to be expressed on professional antigen-presenting cells within secondary lymphoid organs or rather on nonhematopoietic cells of the target organ itself, such as the thyroid follicular cell in the case of autoimmune thyroiditis.
To investigate whether MHC class II molecules when expressed on thyrocytes play a causal role in autoimmune thyroiditis, we developed transgenic mice that express specifically in the thyroid gland the MHC class II transactivator (CIITA), the master regulator of the MHC class II pathway (21, 22). In this experimental model, the temporal relationship between exposure (i.e. expression of class II MHC molecules on thyrocytes) and disease (i.e. autoimmune thyroiditis) is clear and defined ab initio.
Materials and Methods
Construction and screening of the thyr-CIITA transgenic mice
The thyr-CIITA transgene was made by joining the rat thyroglobulin promoter, the rat CIITA type IV cDNA, the jellyfish green fluorescent protein (GFP), and part of the human GH gene (as a source of introns and polyadenylation signal) (Fig. 1A). GFP was linked bicistronically to CIITA, using an internal ribosomal entry site (IRES), to allow easier detection of transgene expression. The 6.5-kb transgene was excised by ClaI and SacII digestion, purified by electroelution, and resuspended at 10 ng/μl in 10 mM Tris/0.1 mM EDTA, pH 7.4. The transgene was injected into fertilized eggs from (CBA x C57BL6) F1 females and maintained as hemizygous by mating to wild-type CBA/J mice (from Jackson Laboratory, Bar Harbor, ME).
FIG. 1. Construction and screening of thyr-CIITA transgenic mice. A, Structure of the transgene. The rat thyroglobulin promoter, the rat type IV CIITA, the internal ribosomal entry site (IRES), the jellyfish GFP, and part of the human GH, as source of splice donor and acceptor sequences are indicated by rectangles of different shades. Primers used for screening are indicated by arrows. B, PCR screening using primers on the CIITA region (top) and GFP region (bottom): positive plasmid control (lane 1); 100-bp DNA marker (lane 2); water control (lane 3); wild-type littermate (lane 4); and line C and line D thyr-CIITA transgenic (lanes 5 and 6). C, Southern blot screening: -HindIII DNA marker (lane 1); positive plasmid control (lane 2); wild-type littermate (lane 3); line C CIITA transgenic (lane 4); and line D CIITA transgenic (lane 5).
Founders were identified by Southern hybridization. Eight micrograms of tail genomic DNA were digested with BamHI, separated by overnight electrophoresis and then transferred onto Hybond-N+ membranes (Amersham, Piscataway, NJ). A 305-bp DNA fragment from the GFP gene was used to generate the 32P-labeled probe, using a random priming labeling kit (InVitrogen, Carlsbad, CA). Hybridization was carried out overnight at 60 C in 0.25 M phosphate buffer (pH 7.2), supplemented with 7% SDS and 5% dextran sulfate, containing 1.5 x 106 trichloroacetic acid-precipitable counts per minute of the probe per milliliter of hybridization buffer. Blots were washed twice in 2x standard saline citrate (SSC) (1x SSC is 0.15 M NaCl plus 0.015 M sodium citrate) and then washed for 30 min at 42 C in 2x SSC/0.1% SDS and then for 30 min at room temperature in 0.2x SSC/0.1% SDS. Blots were wrapped in plastic and exposed overnight at –70 C to BioMax MS autoradiographic films (Eastman Kodak, Rochester, NY). After establishment of the transgenic lines, mice were screened by PCR on tail genomic DNA using one primer pair on the CIITA region and one on the GFP region. The primer sequences were as follows: CIITA forward primer, 5'-GTCAAGTGTTCTTGAACAGTAG-3'; CIITA reverse primer, 5'-AGTATGGCCTTGCAGGTAAG-3'; GFP forward primer, 5'-TGAACCGCATCGAGCTGAAG-3'; GFP reverse primer, 5'-TCCAGCAGGACCATGTGATC-3'. PCR products were fractionated by electrophoresis through a 1.5% agarose gel (Fig. 1B). Mice that contained both the CIITA amplicon (451 bp) and the GFP amplicon (305 bp) were labeled as transgenic.
Analysis of thyr-CIITA transgene expression
Thyroid-specific expression of CIITA was assessed by fluorescent microscopy to detect GFP and RT-PCR and immunohistochemistry to detect MCH class II. GFP expression was measured by freezing the thyroids, cutting 10-μm sections at the microtome, and observing directly the green fluorescence under a Zeiss UV microscope.
Messenger RNA was extracted from thyroid lobes using oligo(d)T magnetic beads (Dynal, Lake Success, NY), treated with RNase-free DNase I (Invitrogen), and reverse transcribed with Superscript II (Invitrogen) into cDNA. cDNA was then amplified by PCR with forward (5'-GACATTGAGGCCGACCACGTAG-3') and reverse (5'-ATTGGTAGCTGGGGTGGAATTTG-3') primers on the I-A locus of the mouse H-2 complex. Primers for glyceraldehyde-3-phosphate dehydrogenase (forward, 5'-GCATCTTGGGCTACACTGAG-3'; reverse, 5'-TCTCTTGCTCAGTGTCCTTG-3') were used to adjust for amount of loading.
Immunohistochemistry was performed on zinc-fixed, paraffin-embedded thyroid specimens. Five-micrometer sections were cut, mounted onto SuperFrost plus slides (Fisher Scientific, Pittsburgh, PA), deparaffinized, and rehydrated. After blocking antibody nonspecific binding (1 h in 2% normal goat serum), sections were incubated overnight in a humid chamber at 4 C with a biotin-conjugated monoclonal antibody recognizing I-Ak (clone 10-3.6 from PharMingen, San Diego, CA). Sections were washed in PBS and incubated in 3% hydrogen peroxide for 5 min to block endogenous peroxidase and then with peroxidase-conjugated streptavidin (Dako, Carpinteria, CA) for 30 min in the humid chamber. Reactions were visualized by the addition of a diaminobenzidine substrate (Sigma Chemical Co., St. Louis, MO). Finally, sections were rinsed in distilled water, counterstained with Mayer’s hematoxylin (Polyscientific, Bay Shore, NY), washed in running tap water, dehydrated through increasing concentrations of ethanol, and mounted with Cytoseal (Richard-Allan Scientific, Kalamazoo, MI).
Induction of experimental autoimmune thyroiditis and evaluation of thyroid pathology
The study population included a total of 245 mice on the CBA/J background: 127 wild-type (81 females and 46 males) and 118 thyr-CIITA transgenic (51 females and 67 males). Sixty-three mice (30 wild-type and 33 transgenic) were analyzed at different time points from 2 to 18 months of age in basal conditions, that is without immunization (d 0 in Table 1). The remaining 182 mice (97 wild-type and 85 transgenic) were immunized twice (on d 0 and 7) via sc injection of gel-purified murine thyroglobulin: 75 μg emulsified in complete Freund’s adjuvant, which contains 5 mg/ml of Mycobacterium tuberculosis strain H37 Ra (Difco Laboratories, Detroit, MI). Mice were then killed at different time points after the immunization, as indicated in Table 1. After euthanasia, tracheas with attached thyroids were removed and fixed for 2–3 d in zinc-based Beckstead’s solution (23). After processing and embedding in paraffin, six to eight nonsequential sections were cut from each specimen and stained with hematoxylin and eosin. The section with the greatest surface area was selected for digital imaging, performed as described (2). Briefly, after acquisition of the images with a SPOT RT video camera mounted on a Zeiss Axioplot microscope, images were analyzed with Carnoy software. The area corresponding to the entire thyroid gland was marked, measured, and used as the denominator (b). The areas of mononuclear cell infiltration were marked, summed, and used as the numerator (a). The thyroid histopathology score was then expressed as the percentage of the total thyroid area infiltrated by mononuclear cells: a/b x 100. All experimental protocols conformed to Johns Hopkins Animal Care and Use Committee guidelines.
TABLE 1. Study population. Distribution of mouse numbers in wild-type and thyr-CIITA transgenic mice, according to the day of killing after immunization with mouse thyroglobulin
Measurement of mouse thyroglobulin-specific IgG antibodies
Mice were bled from the retroorbital venous plexus at the beginning of the immunization and then 10, 14, 21, 28, 35, 70, or 100 d after the first immunization. Sera were diluted in PBS (1/100, 1/400) and incubated overnight in Immulon2 ELISA plates (Dynex Technologies, Chantilly, VA), precoated with mouse thyroglobulin (100 ng/well). After proper washing, mouse thyroglobulin-specific IgG subclasses were detected using secondary antibodies against IgG1, IgG2a, and IgG2b, conjugated to alkaline phosphatase. The color change of p-nitrophenol phosphate was measured at 405 nm using the Emax microplate reader (Molecular Devices, Sunnyvale, CA). Each plate included a homemade standard curve, derived from serial dilutions of a pool serum with known mouse thyroglobulin antibodies. The standard curve allowed us to express the results in arbitrary units (rather than absorbance), thus adjusting for intra- and interassay variability.
Flow cytometry analysis of draining cervical lymph nodes
Cervical lymph nodes were isolated and mechanically disrupted to prepare single-cell suspension. After 15 min of Fc block at 4 C, lymphoid cells were stained for 30 min at 4 C with the following monoclonal antibodies (all from PharMingen, San Diego, CA): fluorescein isothiocyanate-conjugated anti-B220, anti-CD69, and anti-CD8; phycoerythrin-conjugated anti-CD19 and anti-CD44; and cy-chrome-conjugated anti-CD4 and anti-CD3. Data were collected on a FACScalibur cytometer (Becton Dickinson, San Diego, CA), gated first on forward and side scatter and then on CD4, compensated, and displayed using CellQuest software (Becton Dickinson).
Assessment of thyroid function by total T4 and TSH levels
Total T4 level was measured by a commercial competitive RIA (GammaCoat [125I]T4, Diasorin, Stillwater, MN). Mouse TSH was measured with a highly sensitive double-antibody RIA, developed by A. F. Parlow (24). Briefly, the assay employs a highly purified rat TSH (AFP11542B) as the iodinated ligand, a selected guinea pig antimouse TSH (AFP98991 as the primary antibody, and a partially purified extract of mouse pituitary containing TSH (AFP5171.8MP) as the reference preparation. Cross-reactivity of either highly purified mouse FSH or mouse LH in this mouse TSH RIA was less than 1%.
Statistical analysis
Statistical analysis was performed in two phases. The first phase analyzed cross-sectionally whether the mean severity of thyroid lesions (the thyroid histopathology score) differed between wild-type and thyr-CIITA transgenic mice. The crude (unadjusted) effect of the covariate genotype on the outcome thyroid histopathology score was calculated by simple linear regression, ignoring the effect of all the other covariates. Then, the net (adjusted) effect of genotype on thyroid histopathology score, holding all other covariates fixed, was calculated by multiple linear regression. The covariates relating to thyroid histopathology score considered for inclusion in the model, in addition to genotype, were sex, transgenic line, day of killing after the first immunization, total T4 at d 0, TSH at d 0, and thyroglobulin-specific antibodies at d 0. Day of killing after immunization was treated as a dummy variable, with d 0 chosen as the reference category. Using stepwise backward selection, we removed sex, total T4 at d 0, TSH at d 0, and thyroglobulin-specific antibodies at d 0, based on P values greater than 0.05. The final multiple linear regression model, therefore, included genotype, transgenic line, and day of killing as covariates used to estimate the mean thyroid histopathology score. Analysis of the residuals after fitting this regression model indicated nonnormality and heteroscedasticity. We thus used bootstrapping with 2000 replications to estimate SE around each regression coefficient and perform valid hypothesis testing from regression analysis of the nonnormal thyroid histopathology scores. We also analyzed the cumulative incidence of thyroiditis in transgenic and controls, labeling as thyroiditis cases those mice with a thyroid histopathology score greater than 2.5%. RRs at each time point after immunization were then expressed as ratio of the incidence of thyroiditis in transgenic over the incidence in wild-type mice. Confidence intervals around the RRs were calculated in the logarithmic scale and results then exponentiated. Fisher’s exact test was used to test the null hypothesis that RRs were equal to 1.
The second phase used multiple linear regression with generalized estimating equations (GEE) (25) to analyze longitudinally how thyroid function (assessed by total T4 and TSH) and thyroglobulin antibodies evolved after the immunization in thyr-CIITA transgenic and controls. GEE allows for analysis with multiple values of T4 and TSH from the same mouse and corrects for the fact that the characteristics of a single animal over time are likely to be correlated with one another. In GEE analyses, repeated measures for each participant are clustered. Furthermore, GEE analysis takes into account the status or changing value of the covariates at each time point. The outcomes used in GEE analysis were total T4, TSH, IgG1, IgG2a, and IgG2b thyroglobulin-specific antibodies; the covariates, in addition to genotype, were sex, transgenic line, and day of killing.
Statistical analyses were performed using Stata 8 (from Stata Corp., College Station, TX).
Results
Transgene expression and baseline phenotype
The CIITA transgene was expressed under control of the rat thyroglobulin promoter (26), which has been shown to support transcription specifically in thyroid follicular cells. Four of 76 injected mice had the transgene integrated in their genome and passed it to their progeny. Two of these four founders, named thyr-CIITA transgenic lines C and D (Fig. 1C), were maintained as hemizygous and expanded by mating to normal CBA/J mice to form the cohort described in the present study.
Specific expression of the transgene in the thyroid gland was assessed by fluorescent microscopy, which showed the characteristic green positivity of GFP specifically within the thyroid follicular cells (Fig. 2A). RT-PCR revealed aberrant expression of class II MHC transcripts on thyrocytes from thyr-CIITA transgenic, but not on thyrocytes from wild-type littermates (Fig. 2, B and C). Immunohistochemistry confirmed at the protein level the expression of class II MHC molecules in transgenic thyrocytes (Fig. 2D) but not in wild-type thyrocytes (Fig. 2E).
FIG. 2. Thyroid expression and function of the thyr-CIITA transgene. A, Frozen section of a thyr-CIITA transgenic thyroid observed under the UV microscope for direct green fluorescence (x25 magnification, 1-min exposure). The inset represents the background autofluorescence of a wild-type thyroid (x25 magnification, 5-min exposure). B, Specific expression of MHC class II (I-Ak) mRNA in transgenic thyroids, as assessed by RT-PCR. C, Control gel for the expression of a common housekeeping gene (glyceraldehyde-3-phosphate dehydrogenase). In both B and C gels, the minus sign indicates that the RT reaction was performed in the absence of reverse transcriptase to control for possible genomic DNA contamination in the RNA samples. Lane 1, 100-bp DNA marker; lane 2, water control; lanes 3 and 4, wild-type littermate; lanes 5 and 6, line C thyr-CIITA transgenic; lanes 7 and 8, line D thyr-CIITA transgenic. D, Specific expression of MHC class II (I-Ak) protein in transgenic thyroids as assessed by immunohistochemistry. Note the diffuse cytosolic and plasma membrane brown staining in the epithelial cells lining the thyroid follicles of thyr-CIITA transgenics. Nuclei are negative and colored blue by the hematoxylin counterstain. E, Wild-type control shown for comparison.
Thyr-CIITA transgenic mice developed normally and had normal litter size, life span, T4 levels (4.15 μg/dl ± 1.08), and TSH levels (103.9 U ± 39.1). None of the 63 mice (30 wild-type and 33 transgenic) analyzed histologically at d 0 showed signs of thyroid pathology (Fig. 3A) or developed antibodies against mouse thyroglobulin (Fig. 4). Thus, under basal conditions, thyr-CIITA transgenic mice, despite the ectopic expression of MHC class II molecules on thyrocytes, were not different from wild-type littermates and did not develop spontaneous autoimmune thyroiditis.
FIG. 3. Thyroid histopathology at different time points after immunization with mouse thyroglobulin. A–D, Thyroids from CIITA transgenic mice on d 0 (A), 10 (B), 14 (C), and 21 (D); E and F, d-35 thyroids in wild-type (E) and CIITA transgenic (F); G and H, thyroids from CIITA transgenic mice on d 70 (G) and 100 (H).
FIG. 4. Time series of thyroglobulin-specific antibodies after immunization. A, IgG2a isotype; B, IgG2b; C, IgG1. There was no difference in antibody levels between wild-type and transgenics; therefore, these data represent the pooled mean (±SE) of both genotypes. The numbers below each graph represent the mice analyzed at each time point.
Thyroid histopathology after immunization with mouse thyroglobulin
We then immunized 182 mice (97 wild-type and 85 transgenic) with mouse thyroglobulin and assessed the severity of thyroid lesions (the thyroid histopathology score) at different time points thereafter. In both groups of mice, thyroid lesions were already present, although minimal (12% of the total thyroid area), at d 10 post immunization (Fig. 5). They then increased to 65% on d 14 and reached a peak of 90% on d 21 (Figs. 3, C and D, and 5). In wild-type mice, the thyroid histopathology score decreased to approximately 45% and remained around that value until d 100 (Fig. 5). In thyr-CIITA transgenic, instead, the score remained higher and became significantly higher than controls at d 35. Simple linear regression assessing the crude effect of genotype on score revealed that transgenic mice tended to have a more severe disease, with an average score that was 5% higher than that seen in wild-type, although this difference was not marked (P = 0.046). Multiple linear regression showed that, holding transgenic line and day of killing constant, thyr-CIITA transgenic mice had an average thyroid histopathology score that was 8% higher than that seen in wild-type (95% confidence interval, 1.8–13.4%; P = 0.033). Overall, these results indicate that thyr-CIITA transgenics developed thyroiditis that is moderately more severe than that of wild-type controls. The most pronounced separation between transgenic and wild-type was seen at d 35 (Figs. 3, E and F, and 5B); transgenics had an average thyroid histopathology score of 80% compared with the wild-type score of 55% (P = 0.032 by Wilcoxon rank sum test). In both groups of mice, the scores observed at d 14, 28, 35, 70, and 100 were significantly higher than those observed on d 0 and 10 (Figs. 3, A–H, and 5A) (P < 0.0001 comparing d 14, 28, 35, 70, or 100 with d 10 or 0).
FIG. 5. Thyroid histopathology score. A, Distribution of the mean scores in thyr-CIITA transgenic () and wild-type littermates () at d 0 and at several time points after immunization; B, individual scores in thyr-CIITA transgenic and wild-type littermates at d 35 after immunization.
In addition to severity, also the incidence of thyroiditis tended to be greater in thyr-CIITA transgenic mice, although at no time point did this difference reach statistical significance (Table 2).
TABLE 2. Incidence of experimental autoimmune thyroiditis in wild-type and thyr-CIITA transgenic mice after immunization with mouse thyroglobulin
Antibody response to thyroglobulin after immunization with mouse thyroglobulin
Longitudinal data regression analysis demonstrated that IgG1 and IgG2b isotypes specific for mouse thyroglobulin followed a similar pattern and reached similar levels in thyr-CIITA transgenic and controls. They began to increase at d 14 post immunization, became significantly higher at d 21, and reached a peak at d 28 (P = 0.0096 vs. d 21 and P = 0.001 vs. d 35) (Fig. 4, B and C). IgG1 and IgG2b then gradually decreased on d 35, 70, and 100, still remaining significantly higher than d 0.
The IgG2a isotype attained higher levels than the IgG1 and IgG2b isotypes and showed a different trend. It also reached a peak at d 21 but continued to stay high until d 70. No significant difference was observed between d 21, 28, 35, and 70, although the mean value looked higher at d 70 (Fig. 4A). IgG2a levels then decreased on d 100, still remaining significantly higher than d 0 (P = 0.0001). IgG2a levels did not differ between thyr-CIITA transgenic and controls.
Lymphocyte activation in the draining cervical lymph nodes after immunization with mouse thyroglobulin
It has been shown in experimental models of type 1 diabetes mellitus that the activation of islet-specific CD4 T cells that precedes the onset of diabetes occurs in the draining pancreatic lymph nodes or within the islets themselves (27). These observations have also been confirmed in primate experimental allergic encephalomyelitis and human multiple sclerosis (28). We therefore looked for CD4 T cell activation markers (increased expression of CD69 and CD44 and diminished display of CD62L) in cervical lymph nodes draining the thyroid gland. Before immunization with mouse thyroglobulin (d 0), there were few activated CD4 T cells (Fig. 6A) in both thyr-CIITA transgenic and wild-type littermates. At d 35 after immunization, there was a significant increase over d 0 (P < 0.001) in the percentage of activated CD4 T cells (Fig. 6B), with no difference, however, between transgenic and wild-type.
FIG. 6. Flow cytometry analysis of draining cervical lymph nodes for markers of T cell activation. Expression of CD69 and CD44 on gated CD4 T cells before (A) and 35 d after (B) immunization with mouse thyroglobulin. Four to six lymph nodes were collected from each mouse and pooled for the analysis, which was performed on 11 thyr-CIITA transgenic mice and seven wild-type controls.
Thyroid function after immunization with mouse thyroglobulin
Longitudinal regression analysis showed that total T4 significantly decreased from an average of 4.1 μg/dl at d 0 to an average of 2.2 μg/dl at d 14 (P = 0.0001). T4 then remained significantly low until d 35 (there was no difference among d 14, 21, 28, and 35) and then increased to normal levels on d 70 (P = 0.183 vs. d 0) and d 100 (P = 0.206 vs. d 0) (Fig. 7A). Holding all other covariates constant (transgenic line, sex, days of killing, and thyroglobulin-specific antibodies), there was no difference between thyr-CIITA transgenic and wild-type in the total T4 levels. Total T4 was a poor predictor of the thyroid histopathology score (P = 0.842 in the multiple linear regression analysis). The decrease in total T4 at d 14 was present before a rise in TSH, suggesting that this decrease represents the recovery phase from the euthyroid sick syndrome induced by the immunization procedure.
FIG. 7. Thyroid function. Total T4 levels (A) and TSH levels (B) in CIITA transgenic () and wild-type littermates () before and at several time points after immunization with mouse thyroglobulin. Dotted lines indicate the normal reference range in CBA/J mice, from 5th to 95th percentile. The range was 2.62–5.99 μg/dl for T4 and 63–182 U/ml for TSH.
TSH began to increase at 21 d after immunization and then became significantly higher than d 0 at d 35 (P = 0.045) and d 70 (P = 0.031) and returned to normal levels at d 100 (Fig. 7B). Holding all other covariates constant, thyr-CIITA transgenic tended to have higher TSH levels, although the difference did not reach statistical significance (P = 0.078). Similarly to what was observed for thyroid pathology, the greatest distinction between wild-type and thyr-CIITA transgenic was observed at d 35 post immunization (Fig. 5B). Among the humoral predictors measured in this study (total T4, TSH, IgG1, IgG2a, and IgG2b thyroglobulin antibodies), TSH was the best predictor of the thyroid histopathology score. Multiple linear regression showed that for every 100-U increase in serum TSH, the score increased by 6.1% (95% confidence interval, 4.2–8.0%; P < 0.00001).
Discussion
MHC molecules are undoubtedly associated with autoimmune diseases (13). In some cases, such as ankylosing spondylitis (29) and birdshot retinochoroidopathy (30), the strength of this association is so high (i.e. RRs around 100) that there should be little doubt about a causative role of MHC molecules in the pathway leading to autoimmunity. Yet, quoting from S. Beck and J. Trowsdale, "in no case can it be argued that a particular MHC allele is necessary or sufficient to cause autoimmune disease" (31). Also unknown is the importance of location; that is, whether MHC molecules play a causal role when expressed in secondary lymphoid organs or rather on parenchymal cells of the organ targeted by the autoimmune attack (32).
We focused on class II MHC molecules and used experimental autoimmune thyroiditis as a classical model of organ-specific autoimmunity. Class II molecules are primarily expressed on antigen-presenting cells of the hematopoietic lineage but can also be induced on endothelial and epithelial cells. This observation raised the question of whether nonhematopoietic parenchymal cells can function as antigen-presenting cells and whether they are required for damage to the target organ. Much of the work comes from the transplantation field and has generated conflicting results. On one side, Lakkis et al. (33) found that mice lacking all secondary lymphoid organs (spleen, lymph nodes, and mucosa-associated lymphoid tissues) are incapable of rejecting either skin allografts or primarily vascularized cardiac allografts, suggesting that secondary lymphoid organs are required for initiating the immune response. Similarly, Teshima et al. (34), using a model of graft-vs.-host disease in bone marrow chimeras expressing MHC molecules only on antigen-presenting cells but not on epithelial or endothelial cells, found that the graft was capable of inducing epithelial damage in the target organs (skin, gut, and liver) of the host that expressed MHC molecules only on the host antigen-presenting cells and not on the host target epithelium. These studies are consistent with recent direct visual evidence of the immune response, indicating that naive T cells constantly roam through secondary lymphoid organs where they encounter antigen-presenting cells that have taken up antigen from the target organs and brought it into the lymphoid organs (35, 36, 37, 38). These studies are also consistent with the notion that naive T cells possess receptors necessary for homing to secondary lymphoid organs rather than to nonlymphoid organs and with the notion that nonprofessional antigen-presenting cells usually lack the costimulatory molecules required for complete T cell activation. On the other side, however, Kreisel et al. (39) showed that naive CD8 T cells can indeed be activated in vitro and in vivo by nonprofessional antigen-presenting cells, most likely endothelial cells, and lead to prompt rejection of the cardiac allograft, suggesting that endothelial expression of MHC molecules can direct pathogenic immune responses independently of the expression of MHC molecules on professional antigen-presenting cells.
We expressed via transgenesis the transcriptional regulator CIITA specifically in a nonprofessional antigen-presenting cell, the thyrocyte. CIITA is a transcription factor that when defective causes a severe primary immunodeficiency called bare lymphocyte syndrome (22), entirely attributable to the absence of class II MHC expression. CIITA is a true coactivator because it does not bind directly to DNA but mediates its function on the class II MHC promoter through interaction with other proteins (40). It is considered a master regulator because it is necessary and sufficient for the expression of all the genes in the MHC class II pathway (21), although it has been recently shown to control other genes outside the MHC (41). CIITA has three different isoforms synthesized from different, cell-specific promoters (42); promoter 1 is used in dendritic cells, promoter 3 in B lymphocytes, and promoter 4 in nonhematopoietic cells. We have previously shown in vitro that promoter 4 mediates interferon--induced transcription of CIITA in thyrocytes (43), and thus we used the type 4 isoform of CIITA for the present study.
Under basal conditions, CIITA transgenic mice did not differ from wild-type littermates. They did not develop spontaneously autoimmune thyroiditis despite the aberrant expression of class II MHC molecules on thyrocytes. These findings are in agreement with a recent report by Li et al. (44), which shows that thyroidal expression of class II molecules driven by the TSH receptor promoter does not lead to apparent thyroid autoimmunity. The phenomenon does not seem to be limited to the thyroid gland. Stüve et al. (45) expressed CIITA, in fact, specifically in the central nervous system (using the promoter for the glial fibrillary protein), and Herkel et al. (46) expressed it in the liver (using the C-reactive protein promoter). In both cases, despite the aberrant expression of class II MHC molecules on astrocytes or hepatocytes, mice did not develop spontaneous autoimmunity. Taken together, these observations strongly suggest that the expression of class II MHC molecules on nonprofessional antigen-presenting cells is by itself not sufficient to initiate an autoimmune disease. We can postulate that this expression is more likely the results of inflammatory stimuli (such as interferon- and lipopolysaccharide) released once the immune cells infiltrate the target organ, rather than being the initiating event that recruits the immune cells to the target organ.
In addition to basal conditions, we were also interested in analyzing the role of class II molecules on thyrocytes during an autoimmune response. Upon immunization, thyr-CIITA transgenic mice developed thyroiditis that was moderately more severe and of greater incidence than that observed in wild-type littermates. The difference in severity, although higher in transgenics at all time points, became significantly higher only at d 35 after immunization. These results suggest that thyrocytes expressing class II MHC molecules play a secondary role in the initiation phase but can serve as better targets and facilitate greater immunopathology, once the process has already begun. Lack of significant difference in thyroiditis severity at earlier time points (d 10–28) can be explained by the fact that the immunization protocol we used was very effective in inducing thyroid lesions in the majority of the mice (incidence comprised between 86 and 95% in wild-type and 100% in transgenic).
The notion that class II MHC molecules on epithelial cells have a modulatory role once the autoimmune attack has begun is intriguing and may have possible relevance to Hashimoto’s thyroiditis and Graves’ disease (47). In these diseases, in fact, collectively referred to as autoimmune thyroiditis, the expression of class II molecules may indicate a particularly severe stage of the disease. This question is difficult to assess in humans; the use of a mouse model is essential because it allows us to analyze disease over a period of time, rather than at a single time point.
The greater severity of thyroid lesions displayed by thyr-CIITA transgenic was not a result of a greater number or function of activated T cells, considering that the CD4 T cells in the cervical lymph nodes expressed similar levels of activation markers in both wild-type and transgenics. The difference is more likely explained by the fact that thyrocytes expressing class II MHC molecules are better recognized by autoreactive CD4 T cells once these T cells have already expanded. In other words, thyrocytes may serve as endogenous immunogens that potentiate and expand the reactivity of T cells initially selected by the thyroglobulin immunization protocol. It is well established that T cell activation requires formation of a specialized cell-cell junction with the antigen-presenting cell, called the immunological synapse, which facilitates the outcome of immune recognition (48). The center of the synapse is occupied by T-cell receptor/CD3 complexes and CD4 coreceptors on the T cell side and by MHC/peptide complexes on the antigen-presenting side. Also in the center are costimulatory molecules, such as CD28 and CD40 ligand on T cells and B7 and CD40 on antigen-presenting cells. The periphery of the synapse is instead occupied by integrins, such as lymphocyte function-associated antigen-1 and CD2 on T cells and intercellular adhesion molecule-1 and CD58 on antigen-presenting cells. To function efficiently as antigen-presenting cell, the thyrocytes that aberrantly express MHC class II molecules should also express costimulatory molecules and integrins. Thyrocytes from patients with Graves’ disease and Hashimoto’s thyroiditis have been shown to express the integrins intercellular adhesion molecule-1 and CD58 (49, 50) and the costimulator CD40 (51, 52). The expression on thyrocytes of B7, probably the most potent costimulator, is less firmly established, although it was reported in Hashimoto’s thyroiditis (53) and differentiated thyroid tumors (54). B7 costimulation could also be supplied in trans from professional antigen-presenting cells to parenchymal cells (55). Thus, in theory, thyrocytes aberrantly expressing class II MHC molecules possess the signal transduction circuitry necessary for full T cell activation, although it is unknown whether individual components can associate in the supramolecular activation complex. Here we show that expression on thyrocytes of class II MHC is insufficient to initiate autoimmune thyroiditis. These results emphasize the importance of location in the initiation of the immune response, a location that likely requires the interaction between autoreactive T cell and antigen-presenting cells within the secondary lymphoid organs.
In conclusion, this study reports that expression of class II MHC molecules on epithelial thyroid cells is not required for the initiation of an autoimmune attack to the thyroid. The initiation, then, seems to be mainly mediated by the professional antigen-presenting cells in secondary lymphoid organs. Once the immune recognition has begun, however, the aberrant expression of class II on nonprofessional antigen-presenting cells can facilitate disease progression and yields a more severe disease. Additional studies are required to fully explain the intriguing association between MHC molecules and autoimmune diseases.
Acknowledgments
We are grateful to Drs. William D. M. Baldwin and James Tonascia for critical revisions of the manuscript.
References
Jacobson DL, Gange SJ, Rose N, Graham NMH 1997 Epidemiology and estimated population burden of selected autoimmune diseases in the United States. Clin Immunol Immunopathol 84:223–243
Caturegli P, Rose NR, Kimura M, Kimura H, Tzou SC 2003 Studies on murine thyroiditis: new insights from organ flow cytometry. Thyroid 13:419–426
Ueda H, Howson JM, Esposito L, Heward J, Snook H, Chamberlain G, Rainbow DB, Hunter KM, Smith AN, Di Genova G, Herr MH, Dahlman I, Payne F, Smyth D, Lowe C, Twells RC, Howlett S, Healy B, Nutland S, Rance HE, Everett V, Smink LJ, Lam AC, Cordell HJ, Walker NM, et al. 2003 Association of the T-cell regulatory gene CTLA4 with susceptibility to autoimmune disease. Nature 423:506–511
Grumet FC, Coukell A, Bodmer JG, Bodmer WF, McDevitt HO 1971 Histocompatibility (HL-A) antigens associated with systemic lupus erythematosus: a possible genetic predisposition to disease. N Engl J Med 285:193–196
Vladutiu AO, Rose NR 1971 Autoimmune murine thyroiditis: relation to histocompatibility (H-2) type. Science 174:1137–1139
Mungall AJ, Palmer SA, Sims SK, Edwards CA, Ashurst JL, Wilming L, Jones MC, Horton R, Hunt SE, Scott CE, Gilbert JG, Clamp ME, Bethel G, Milne S, Ainscough R, Almeida JP, Ambrose KD, Andrews TD, Ashwell RI, Babbage AK, Bagguley CL, Bailey J, Banerjee R, Barker DJ, Barlow KF, et al. 2003 The DNA sequence and analysis of human chromosome 6. Nature 425:805–811
Kumanovics A, Takada T, Lindahl KF 2003 Genomic organization of the mammalian MHC. Annu Rev Immunol 21:629–657
Dallman MJ, Mason DW 1983 Induction of Ia antigens on murine epidermal cells during the rejection of skin allografts. Transplantation 36:222–224
Bottazzo GF, Pujol-Borrell R, Hanafusa T, Feldmann M 1983 Role of aberrant HLA-DR expression and antigen presentation in induction of endocrine autoimmunity. Lancet 2:1115–1119
Mirakian R, Hammond LJ, Bottazzo GF 1998 Pathogenesis of thyroid autoimmunity: the Bottazzo-Feldmann hypothesis. Immunol Today 19:97–98
Lin L, Hungs M, Mignot E 2001 Narcolepsy and the HLA region. J Neuroimmunol 117:9–20
Kralovicova J, Hammarstrom L, Plebani A, Webster AD, Vorechovsky I 2003 Fine-scale mapping at IGAD1 and genome-wide genetic linkage analysis implicate HLA-DQ/DR as a major susceptibility locus in selective IgA deficiency and common variable immunodeficiency. J Immunol 170:2765–2775
Price P, Witt C, Allcock R, Sayer D, Garlepp M, Kok CC, French M, Mallal S, Christiansen F 1999 The genetic basis for the association of the 8.1 ancestral haplotype (A1, B8, DR3) with multiple immunopathological diseases. Immunol Rev 167:257–274
Klein J, Sato A 2000 The HLA system: second of two parts. N Engl J Med 343:782–786
Carlsson B, Wallin J, Pirskanen R, Matell G, Smith CI 1990 Different HLA DR-DQ associations in subgroups of idiopathic myasthenia gravis. Immunogenetics 31:285–290
Pociot F, McDermott MF 2002 Genetics of type 1 diabetes mellitus. Genes Immun 3:235–249
Levinson RD, Gonzales CR 2002 Birdshot retinochoroidopathy: immunopathogenesis, evaluation, and treatment. Ophthalmol Clin North Am 15:343–350, vii
Tomer Y, Davies TF 2003 Searching for the autoimmune thyroid disease susceptibility genes: from gene mapping to gene function. Endocr Rev 24:694–717
Levin L, Ban Y, Concepcion E, Davies TF, Greenberg DA, Tomer Y 2004 Analysis of HLA genes in families with autoimmune diabetes and thyroiditis. Hum Immunol 65:640–647
Maksymowych WP, Tao S, Vaile J, Suarez-Almazor M, Ramos-Remus C, Russell AS 2000 LMP2 polymorphism is associated with extraspinal disease in HLA-B27 negative Caucasian and Mexican Mestizo patients with ankylosing spondylitis. J Rheumatol 27:183–189
Mach B 1999 Perspectives: immunology. Regulating the regulator. Science 285:1367
Reith W, Mach B 2001 The bare lymphocyte syndrome and the regulation of MHC expression. Annu Rev Immunol 19:331–373
Beckstead JH 1995 A simple technique for preservation of fixation-sensitive antigens in paraffin-embedded tissues: addendum. J Histochem Cytochem 43:345
Schneider MJ, Fiering SN, Pallud SE, Parlow AF, St Germain DL, Galton VA 2001 Targeted disruption of the type 2 selenodeiodinase gene (DIO2) results in a phenotype of pituitary resistance to T4. Mol Endocrinol 15:2137–2148
Diggle PJ, Liang KY, Zeger SL 1995 Analysis of longitudinal data. Oxford, UK: Clarendon Press
Musti AM, Ursini VM, Avvedimento EV, Zimarino V, Di Lauro R 1987 A cell type specific factor recognizes the rat thyroglobulin promoter. Nucleic Acids Res 15:8149–8166
Hoglund P, Mintern J, Waltzinger C, Heath W, Benoist C, Mathis D 1999 Initiation of autoimmune diabetes by developmentally regulated presentation of islet cell antigens in the pancreatic lymph nodes. J Exp Med 189:331–339
de Vos AF, van Meurs M, Brok HP, Boven LA, Hintzen RQ, van der Valk P, Ravid R, Rensing S, Boon L, ’t Hart BA, Laman JD 2002 Transfer of central nervous system autoantigens and presentation in secondary lymphoid organs. J Immunol 169:5415–5423
Sims AM, Wordsworth BP, Brown MA 2004 Genetic susceptibility to ankylosing spondylitis. Curr Mol Med 4:13–20
LeHoang P, Ozdemir N, Benhamou A, Tabary T, Edelson C, Betuel H, Semiglia R, Cohen JH 1992 HLA-A29.2 subtype associated with birdshot retinochoroidopathy. Am J Ophthalmol 113:33–35
Beck S, Trowsdale J 2000 The human major histocompatability complex: lessons from the DNA sequence. Annu Rev Genomics Hum Genet 1:117–137
Lakkis FG 2003 Where is the alloimmune response initiated? Am J Transplant 3:241–242
Lakkis FG, Arakelov A, Konieczny BT, Inoue Y 2000 Immunologic ‘ignorance’ of vascularized organ transplants in the absence of secondary lymphoid tissue. Nat Med 6:686–688
Teshima T, Ordemann R, Reddy P, Gagin S, Liu C, Cooke KR, Ferrara JL 2002 Acute graft-versus-host disease does not require alloantigen expression on host epithelium. Nat Med [Erratum (2002) 8:1039] 8:575–581
Monks CR, Freiberg BA, Kupfer H, Sciaky N, Kupfer A 1998 Three-dimensional segregation of supramolecular activation clusters in T cells. Nature 395:82–86
Wulfing C, Davis MM 1998 A receptor/cytoskeletal movement triggered by costimulation during T cell activation. Science 282:2266–2269
Grakoui A, Bromley SK, Sumen C, Davis MM, Shaw AS, Allen PM, Dustin ML 1999 The immunological synapse: a molecular machine controlling T cell activation. Science 285:221–227
Itano AA, Jenkins MK 2003 Antigen presentation to naive CD4 T cells in the lymph node. Nat Immunol 4:733–739
Kreisel D, Krupnick AS, Gelman AE, Engels FH, Popma SH, Krasinskas AM, Balsara KR, Szeto WY, Turka LA, Rosengard BR 2002 Non-hematopoietic allograft cells directly activate CD8+ T cells and trigger acute rejection: an alternative mechanism of allorecognition. Nat Med 8:233–239
Ting JP, Trowsdale J 2002 Genetic control of MHC class II expression. Cell 109(Suppl):S21–S33
Reith W 2003 ComPLEXIN new targets for CIITA. Nat Immunol 4:819–820
Muhlethaler-Mottet A, Otten LA, Steimle V, Mach B 1997 Expression of MHC class II molecules in different cellular and functional compartments is controlled by differential usage of multiple promoters of the transactivator CIITA. EMBO J 16:2851–2860
Mori-Aoki A, Pietrarelli M, Nakazato M, Caturegli P, Kohn LD, Suzuki K 2000 Class II transactivator suppresses transcription of thyroid-specific genes. Biochem Biophys Res Commun 278:58–62
Li YS, Kanamoto N, Hataya Y, Moriyama K, Hiratani H, Nakao K, Akamizu T 2004 Transgenic mice producing major histocompatibility complex class II molecules on thyroid cells do not develop apparent autoimmune thyroid diseases. Endocrinology 145:2524–2530
Stuve O, Youssef S, Slavin AJ, King CL, Patarroyo JC, Hirschberg DL, Brickey WJ, Soos JM, Piskurich JF, Chapman HA, Zamvil SS 2002 The role of the MHC class II transactivator in class II expression and antigen presentation by astrocytes and in susceptibility to central nervous system autoimmune disease. J Immunol 169:6720–6732
Herkel J, Jagemann B, Wiegard C, Lazaro JF, Lueth S, Kanzler S, Blessing M, Schmitt E, Lohse AW 2003 MHC class II-expressing hepatocytes function as antigen-presenting cells and activate specific CD4 T lymphocytes. Hepatology 37:1079–1085
Weetman AP 2003 Autoimmune thyroid disease: propagation and progression. Eur J Endocrinol 148:1–9
Dustin ML 2002 The immunological synapse. Arthritis Res 4(Suppl 3):S119–S125
Zheng RQ, Abney ER, Grubeck-Loebenstein B, Dayan C, Maini RN, Feldmann M 1990 Expression of intercellular adhesion molecule-1 and lymphocyte function-associated antigen-3 on human thyroid epithelial cells in Graves’ and Hashimoto’s diseases. J Autoimmun 3:727–736
Bonita RE, Rose NR, Rasooly L, Caturegli P, Burek CL 2002 Adhesion molecules as susceptibility factors in spontaneous autoimmune thyroiditis in the NOD-H2h4 mouse. Exp Mol Pathol 73:155–163
Faure GC, Bensoussan-Lejzerowicz D, Bene MC, Aubert V, Leclere J 1997 Coexpression of CD40 and class II antigen HLA-DR in Graves’ disease thyroid epithelial cells. Clin Immunol Immunopathol 84:212–215
Metcalfe RA, McIntosh RS, Marelli-Berg F, Lombardi G, Lechler R, Weetman AP 1998 Detection of CD40 on human thyroid follicular cells: analysis of expression and function. J Clin Endocrinol Metab 83:1268–1274
Battifora M, Pesce G, Paolieri F, Fiorino N, Giordano C, Riccio AM, Torre G, Olive D, Bagnasco M 1998 B7.1 costimulatory molecule is expressed on thyroid follicular cells in Hashimoto’s thyroiditis, but not in Graves’ disease. J Clin Endocrinol Metab 83:4130–4139
Shah R, Banks K, Patel A, Dogra S, Terrell R, Powers PA, Fenton C, Dinauer CA, Tuttle RM, Francis GL 2002 Intense expression of the B7-2 antigen presentation coactivator is an unfavorable prognostic indicator for differentiated thyroid carcinoma of children and adolescents. J Clin Endocrinol Metab 87:4391–4397
Mandelbrot DA, Kishimoto K, Auchincloss Jr H, Sharpe AH, Sayegh MH 2001 Rejection of mouse cardiac allografts by costimulation in trans. J Immunol 167:1174–1178(Hiroaki Kimura, Miho Kimu)
Address all correspondence and requests for reprints to: Patrizio Caturegli, M.D., M.P.H., Johns Hopkins Medical Institutions, Department of Pathology, Ross Building, Room 656, 720 Rutland Avenue, Baltimore, Maryland 21205. E-mail: pcat@jhmi.edu.
Abstract
Class II major histocompatibility complex (MHC) molecules are classically expressed on antigen-presenting cells of the hematopoietic lineage but have also been described on epithelial cells in association with autoimmunity. In this context, however, it remains debatable whether class II MHC molecules are the initiating event or rather the consequence of the autoimmune attack. In addition, the role of epithelial class II expression once the autoimmune attack has begun is unknown. We generated transgenic mice expressing in the thyroid follicular cells the class II transactivator, the master regulator of all the genes in the class II MHC pathway. The study used a cohort of 245 CBA/J mice (127 wild-type and 118 transgenic), both in basal conditions (n = 63) and at different time points after immunization with mouse thyroglobulin (n = 182). In basal conditions, transgenic mice were similar to wild-type controls and did not develop spontaneous autoimmune thyroiditis, despite the aberrant expression of class II MHC molecules on thyrocytes. After immunization, thyroiditis was 8% more severe in transgenics than controls (95% confidence interval from 1.8–13.4%; P = 0.033), especially during the florid stages of disease. These findings suggest that expression of class II MHC molecules on epithelial cells is not sufficient to initiate autoimmunity but mildly modulates an already established autoimmune attack against the target organ.
Introduction
AUTOIMMUNE DISEASES OF the thyroid gland, represented by Graves’ disease and Hashimoto’s thyroiditis and collectively referred to as autoimmune thyroiditis, are the most common autoimmune diseases (1) and an excellent model of organ-specific autoimmunity (2). They are chronic conditions initiated by loss of immunological tolerance to thyroid-restricted self-antigens (such as TSH receptor, thyroperoxidase, and thyroglobulin), a loss that leads to thyroid dysfunction via immune-mediated inflammation and ultimately to the clinical phenotype.
The loss of tolerance is considered to result from the combined effect of environmental factors, such as iodine, and gene variants. For example, individuals with an A-to-G polymorphism in the noncoding, 3' region of cytotoxic T lymphocyte antigen 4 gene have a 2.3-fold or 1.45-fold increased risk of developing Graves’ disease or Hashimoto’s thyroiditis, respectively (3). The first described and most potent genetic influence known to affect susceptibility to autoimmune thyroiditis, however, is the major histocompatibility complex (MHC). The association between MHC and autoimmunity was originally reported in 1971, first in patients with systemic lupus erythematosus (4) and then in mice with experimental autoimmune thyroiditis (5).
The MHC is located on the short arm of the human chromosome 6, spanning 3.6 megabases of complete and annotated DNA sequence (6). Its main functions are to guide the intracellular processing of antigens and to display antigenic peptides on the plasma membrane for recognition by T lymphocytes (7). It is the most gene-dense region of the genome, containing 224 loci (96 pseudogenes and 128 expressed genes) grouped in three classes. Class II MHC genes are constitutively expressed only on hematopoietic cells involved in antigen presentation (dendritic cells, macrophages, B lymphocytes, and cortical thymic epithelial cells) but can be induced by inflammatory stimuli on many other cell types (such as endothelial cells, hepatocytes, ?-cells of the pancreas, and thyrocytes). Little is known about the biological significance of this observation, although it has been implicated in allograft rejection (8) and subsequently in autoimmunity. In this field, Bottazzo and Feldman hypothesized that class II MHC molecules, when aberrantly expressed on epithelial cells, confer to them antigen-presenting functions favoring the initiation of autoimmune diseases (9, 10).
The MHC is also the region that has been associated with the greatest number of human diseases, ranging from narcolepsy (11) to primary immunodeficiencies (12) to most, if not all, autoimmune conditions (13, 14). The strength of this association, which is classically measured by relative risks (RR) or odds ratios, varies from low, such as a RR of 2.5 for myasthenia gravis in a person carrying the DR3 allele (15), to medium, such as a RR of 30 for DQ8 and type 1 diabetes (16), to high, such as a RR of 109 for A29 and birdshot retinochoroidopathy (17). For Graves’ disease, the MHC (DR3) association is mild, with RRs ranging from 1.5–3.5 (reviewed in Ref.18). For Hashimoto’s thyroiditis, the contribution of MHC haplotypes to disease is less definitive, although they may play a role in the familial clustering of thyroiditis and diabetes (19).
Given the biological plausibility, the replication of the findings, and the strength of many of these associations, MHC molecules are considered in the causal pathway that leads to autoimmune disease. There are, however, important factors that could confound the above-described association. First, MHC loci are found together at frequencies much higher than would be expected by random combination of their frequencies in the population. It is thus possible that the genes causing disease are not the MHC genes per se but rather genes that travel with the MHC. For example, polymorphisms of the LMP2 and LMP7 genes, which are embedded within the class II region, have been associated with ankylosing spondylitis (20). In addition, the same allele that is associated with a particular disease is also common in the normal, unaffected population, and actually the majority of patients with the susceptible allele do not develop autoimmune disease. Finally, it is unknown whether to initiate autoimmunity MHC molecules have to be expressed on professional antigen-presenting cells within secondary lymphoid organs or rather on nonhematopoietic cells of the target organ itself, such as the thyroid follicular cell in the case of autoimmune thyroiditis.
To investigate whether MHC class II molecules when expressed on thyrocytes play a causal role in autoimmune thyroiditis, we developed transgenic mice that express specifically in the thyroid gland the MHC class II transactivator (CIITA), the master regulator of the MHC class II pathway (21, 22). In this experimental model, the temporal relationship between exposure (i.e. expression of class II MHC molecules on thyrocytes) and disease (i.e. autoimmune thyroiditis) is clear and defined ab initio.
Materials and Methods
Construction and screening of the thyr-CIITA transgenic mice
The thyr-CIITA transgene was made by joining the rat thyroglobulin promoter, the rat CIITA type IV cDNA, the jellyfish green fluorescent protein (GFP), and part of the human GH gene (as a source of introns and polyadenylation signal) (Fig. 1A). GFP was linked bicistronically to CIITA, using an internal ribosomal entry site (IRES), to allow easier detection of transgene expression. The 6.5-kb transgene was excised by ClaI and SacII digestion, purified by electroelution, and resuspended at 10 ng/μl in 10 mM Tris/0.1 mM EDTA, pH 7.4. The transgene was injected into fertilized eggs from (CBA x C57BL6) F1 females and maintained as hemizygous by mating to wild-type CBA/J mice (from Jackson Laboratory, Bar Harbor, ME).
FIG. 1. Construction and screening of thyr-CIITA transgenic mice. A, Structure of the transgene. The rat thyroglobulin promoter, the rat type IV CIITA, the internal ribosomal entry site (IRES), the jellyfish GFP, and part of the human GH, as source of splice donor and acceptor sequences are indicated by rectangles of different shades. Primers used for screening are indicated by arrows. B, PCR screening using primers on the CIITA region (top) and GFP region (bottom): positive plasmid control (lane 1); 100-bp DNA marker (lane 2); water control (lane 3); wild-type littermate (lane 4); and line C and line D thyr-CIITA transgenic (lanes 5 and 6). C, Southern blot screening: -HindIII DNA marker (lane 1); positive plasmid control (lane 2); wild-type littermate (lane 3); line C CIITA transgenic (lane 4); and line D CIITA transgenic (lane 5).
Founders were identified by Southern hybridization. Eight micrograms of tail genomic DNA were digested with BamHI, separated by overnight electrophoresis and then transferred onto Hybond-N+ membranes (Amersham, Piscataway, NJ). A 305-bp DNA fragment from the GFP gene was used to generate the 32P-labeled probe, using a random priming labeling kit (InVitrogen, Carlsbad, CA). Hybridization was carried out overnight at 60 C in 0.25 M phosphate buffer (pH 7.2), supplemented with 7% SDS and 5% dextran sulfate, containing 1.5 x 106 trichloroacetic acid-precipitable counts per minute of the probe per milliliter of hybridization buffer. Blots were washed twice in 2x standard saline citrate (SSC) (1x SSC is 0.15 M NaCl plus 0.015 M sodium citrate) and then washed for 30 min at 42 C in 2x SSC/0.1% SDS and then for 30 min at room temperature in 0.2x SSC/0.1% SDS. Blots were wrapped in plastic and exposed overnight at –70 C to BioMax MS autoradiographic films (Eastman Kodak, Rochester, NY). After establishment of the transgenic lines, mice were screened by PCR on tail genomic DNA using one primer pair on the CIITA region and one on the GFP region. The primer sequences were as follows: CIITA forward primer, 5'-GTCAAGTGTTCTTGAACAGTAG-3'; CIITA reverse primer, 5'-AGTATGGCCTTGCAGGTAAG-3'; GFP forward primer, 5'-TGAACCGCATCGAGCTGAAG-3'; GFP reverse primer, 5'-TCCAGCAGGACCATGTGATC-3'. PCR products were fractionated by electrophoresis through a 1.5% agarose gel (Fig. 1B). Mice that contained both the CIITA amplicon (451 bp) and the GFP amplicon (305 bp) were labeled as transgenic.
Analysis of thyr-CIITA transgene expression
Thyroid-specific expression of CIITA was assessed by fluorescent microscopy to detect GFP and RT-PCR and immunohistochemistry to detect MCH class II. GFP expression was measured by freezing the thyroids, cutting 10-μm sections at the microtome, and observing directly the green fluorescence under a Zeiss UV microscope.
Messenger RNA was extracted from thyroid lobes using oligo(d)T magnetic beads (Dynal, Lake Success, NY), treated with RNase-free DNase I (Invitrogen), and reverse transcribed with Superscript II (Invitrogen) into cDNA. cDNA was then amplified by PCR with forward (5'-GACATTGAGGCCGACCACGTAG-3') and reverse (5'-ATTGGTAGCTGGGGTGGAATTTG-3') primers on the I-A locus of the mouse H-2 complex. Primers for glyceraldehyde-3-phosphate dehydrogenase (forward, 5'-GCATCTTGGGCTACACTGAG-3'; reverse, 5'-TCTCTTGCTCAGTGTCCTTG-3') were used to adjust for amount of loading.
Immunohistochemistry was performed on zinc-fixed, paraffin-embedded thyroid specimens. Five-micrometer sections were cut, mounted onto SuperFrost plus slides (Fisher Scientific, Pittsburgh, PA), deparaffinized, and rehydrated. After blocking antibody nonspecific binding (1 h in 2% normal goat serum), sections were incubated overnight in a humid chamber at 4 C with a biotin-conjugated monoclonal antibody recognizing I-Ak (clone 10-3.6 from PharMingen, San Diego, CA). Sections were washed in PBS and incubated in 3% hydrogen peroxide for 5 min to block endogenous peroxidase and then with peroxidase-conjugated streptavidin (Dako, Carpinteria, CA) for 30 min in the humid chamber. Reactions were visualized by the addition of a diaminobenzidine substrate (Sigma Chemical Co., St. Louis, MO). Finally, sections were rinsed in distilled water, counterstained with Mayer’s hematoxylin (Polyscientific, Bay Shore, NY), washed in running tap water, dehydrated through increasing concentrations of ethanol, and mounted with Cytoseal (Richard-Allan Scientific, Kalamazoo, MI).
Induction of experimental autoimmune thyroiditis and evaluation of thyroid pathology
The study population included a total of 245 mice on the CBA/J background: 127 wild-type (81 females and 46 males) and 118 thyr-CIITA transgenic (51 females and 67 males). Sixty-three mice (30 wild-type and 33 transgenic) were analyzed at different time points from 2 to 18 months of age in basal conditions, that is without immunization (d 0 in Table 1). The remaining 182 mice (97 wild-type and 85 transgenic) were immunized twice (on d 0 and 7) via sc injection of gel-purified murine thyroglobulin: 75 μg emulsified in complete Freund’s adjuvant, which contains 5 mg/ml of Mycobacterium tuberculosis strain H37 Ra (Difco Laboratories, Detroit, MI). Mice were then killed at different time points after the immunization, as indicated in Table 1. After euthanasia, tracheas with attached thyroids were removed and fixed for 2–3 d in zinc-based Beckstead’s solution (23). After processing and embedding in paraffin, six to eight nonsequential sections were cut from each specimen and stained with hematoxylin and eosin. The section with the greatest surface area was selected for digital imaging, performed as described (2). Briefly, after acquisition of the images with a SPOT RT video camera mounted on a Zeiss Axioplot microscope, images were analyzed with Carnoy software. The area corresponding to the entire thyroid gland was marked, measured, and used as the denominator (b). The areas of mononuclear cell infiltration were marked, summed, and used as the numerator (a). The thyroid histopathology score was then expressed as the percentage of the total thyroid area infiltrated by mononuclear cells: a/b x 100. All experimental protocols conformed to Johns Hopkins Animal Care and Use Committee guidelines.
TABLE 1. Study population. Distribution of mouse numbers in wild-type and thyr-CIITA transgenic mice, according to the day of killing after immunization with mouse thyroglobulin
Measurement of mouse thyroglobulin-specific IgG antibodies
Mice were bled from the retroorbital venous plexus at the beginning of the immunization and then 10, 14, 21, 28, 35, 70, or 100 d after the first immunization. Sera were diluted in PBS (1/100, 1/400) and incubated overnight in Immulon2 ELISA plates (Dynex Technologies, Chantilly, VA), precoated with mouse thyroglobulin (100 ng/well). After proper washing, mouse thyroglobulin-specific IgG subclasses were detected using secondary antibodies against IgG1, IgG2a, and IgG2b, conjugated to alkaline phosphatase. The color change of p-nitrophenol phosphate was measured at 405 nm using the Emax microplate reader (Molecular Devices, Sunnyvale, CA). Each plate included a homemade standard curve, derived from serial dilutions of a pool serum with known mouse thyroglobulin antibodies. The standard curve allowed us to express the results in arbitrary units (rather than absorbance), thus adjusting for intra- and interassay variability.
Flow cytometry analysis of draining cervical lymph nodes
Cervical lymph nodes were isolated and mechanically disrupted to prepare single-cell suspension. After 15 min of Fc block at 4 C, lymphoid cells were stained for 30 min at 4 C with the following monoclonal antibodies (all from PharMingen, San Diego, CA): fluorescein isothiocyanate-conjugated anti-B220, anti-CD69, and anti-CD8; phycoerythrin-conjugated anti-CD19 and anti-CD44; and cy-chrome-conjugated anti-CD4 and anti-CD3. Data were collected on a FACScalibur cytometer (Becton Dickinson, San Diego, CA), gated first on forward and side scatter and then on CD4, compensated, and displayed using CellQuest software (Becton Dickinson).
Assessment of thyroid function by total T4 and TSH levels
Total T4 level was measured by a commercial competitive RIA (GammaCoat [125I]T4, Diasorin, Stillwater, MN). Mouse TSH was measured with a highly sensitive double-antibody RIA, developed by A. F. Parlow (24). Briefly, the assay employs a highly purified rat TSH (AFP11542B) as the iodinated ligand, a selected guinea pig antimouse TSH (AFP98991 as the primary antibody, and a partially purified extract of mouse pituitary containing TSH (AFP5171.8MP) as the reference preparation. Cross-reactivity of either highly purified mouse FSH or mouse LH in this mouse TSH RIA was less than 1%.
Statistical analysis
Statistical analysis was performed in two phases. The first phase analyzed cross-sectionally whether the mean severity of thyroid lesions (the thyroid histopathology score) differed between wild-type and thyr-CIITA transgenic mice. The crude (unadjusted) effect of the covariate genotype on the outcome thyroid histopathology score was calculated by simple linear regression, ignoring the effect of all the other covariates. Then, the net (adjusted) effect of genotype on thyroid histopathology score, holding all other covariates fixed, was calculated by multiple linear regression. The covariates relating to thyroid histopathology score considered for inclusion in the model, in addition to genotype, were sex, transgenic line, day of killing after the first immunization, total T4 at d 0, TSH at d 0, and thyroglobulin-specific antibodies at d 0. Day of killing after immunization was treated as a dummy variable, with d 0 chosen as the reference category. Using stepwise backward selection, we removed sex, total T4 at d 0, TSH at d 0, and thyroglobulin-specific antibodies at d 0, based on P values greater than 0.05. The final multiple linear regression model, therefore, included genotype, transgenic line, and day of killing as covariates used to estimate the mean thyroid histopathology score. Analysis of the residuals after fitting this regression model indicated nonnormality and heteroscedasticity. We thus used bootstrapping with 2000 replications to estimate SE around each regression coefficient and perform valid hypothesis testing from regression analysis of the nonnormal thyroid histopathology scores. We also analyzed the cumulative incidence of thyroiditis in transgenic and controls, labeling as thyroiditis cases those mice with a thyroid histopathology score greater than 2.5%. RRs at each time point after immunization were then expressed as ratio of the incidence of thyroiditis in transgenic over the incidence in wild-type mice. Confidence intervals around the RRs were calculated in the logarithmic scale and results then exponentiated. Fisher’s exact test was used to test the null hypothesis that RRs were equal to 1.
The second phase used multiple linear regression with generalized estimating equations (GEE) (25) to analyze longitudinally how thyroid function (assessed by total T4 and TSH) and thyroglobulin antibodies evolved after the immunization in thyr-CIITA transgenic and controls. GEE allows for analysis with multiple values of T4 and TSH from the same mouse and corrects for the fact that the characteristics of a single animal over time are likely to be correlated with one another. In GEE analyses, repeated measures for each participant are clustered. Furthermore, GEE analysis takes into account the status or changing value of the covariates at each time point. The outcomes used in GEE analysis were total T4, TSH, IgG1, IgG2a, and IgG2b thyroglobulin-specific antibodies; the covariates, in addition to genotype, were sex, transgenic line, and day of killing.
Statistical analyses were performed using Stata 8 (from Stata Corp., College Station, TX).
Results
Transgene expression and baseline phenotype
The CIITA transgene was expressed under control of the rat thyroglobulin promoter (26), which has been shown to support transcription specifically in thyroid follicular cells. Four of 76 injected mice had the transgene integrated in their genome and passed it to their progeny. Two of these four founders, named thyr-CIITA transgenic lines C and D (Fig. 1C), were maintained as hemizygous and expanded by mating to normal CBA/J mice to form the cohort described in the present study.
Specific expression of the transgene in the thyroid gland was assessed by fluorescent microscopy, which showed the characteristic green positivity of GFP specifically within the thyroid follicular cells (Fig. 2A). RT-PCR revealed aberrant expression of class II MHC transcripts on thyrocytes from thyr-CIITA transgenic, but not on thyrocytes from wild-type littermates (Fig. 2, B and C). Immunohistochemistry confirmed at the protein level the expression of class II MHC molecules in transgenic thyrocytes (Fig. 2D) but not in wild-type thyrocytes (Fig. 2E).
FIG. 2. Thyroid expression and function of the thyr-CIITA transgene. A, Frozen section of a thyr-CIITA transgenic thyroid observed under the UV microscope for direct green fluorescence (x25 magnification, 1-min exposure). The inset represents the background autofluorescence of a wild-type thyroid (x25 magnification, 5-min exposure). B, Specific expression of MHC class II (I-Ak) mRNA in transgenic thyroids, as assessed by RT-PCR. C, Control gel for the expression of a common housekeeping gene (glyceraldehyde-3-phosphate dehydrogenase). In both B and C gels, the minus sign indicates that the RT reaction was performed in the absence of reverse transcriptase to control for possible genomic DNA contamination in the RNA samples. Lane 1, 100-bp DNA marker; lane 2, water control; lanes 3 and 4, wild-type littermate; lanes 5 and 6, line C thyr-CIITA transgenic; lanes 7 and 8, line D thyr-CIITA transgenic. D, Specific expression of MHC class II (I-Ak) protein in transgenic thyroids as assessed by immunohistochemistry. Note the diffuse cytosolic and plasma membrane brown staining in the epithelial cells lining the thyroid follicles of thyr-CIITA transgenics. Nuclei are negative and colored blue by the hematoxylin counterstain. E, Wild-type control shown for comparison.
Thyr-CIITA transgenic mice developed normally and had normal litter size, life span, T4 levels (4.15 μg/dl ± 1.08), and TSH levels (103.9 U ± 39.1). None of the 63 mice (30 wild-type and 33 transgenic) analyzed histologically at d 0 showed signs of thyroid pathology (Fig. 3A) or developed antibodies against mouse thyroglobulin (Fig. 4). Thus, under basal conditions, thyr-CIITA transgenic mice, despite the ectopic expression of MHC class II molecules on thyrocytes, were not different from wild-type littermates and did not develop spontaneous autoimmune thyroiditis.
FIG. 3. Thyroid histopathology at different time points after immunization with mouse thyroglobulin. A–D, Thyroids from CIITA transgenic mice on d 0 (A), 10 (B), 14 (C), and 21 (D); E and F, d-35 thyroids in wild-type (E) and CIITA transgenic (F); G and H, thyroids from CIITA transgenic mice on d 70 (G) and 100 (H).
FIG. 4. Time series of thyroglobulin-specific antibodies after immunization. A, IgG2a isotype; B, IgG2b; C, IgG1. There was no difference in antibody levels between wild-type and transgenics; therefore, these data represent the pooled mean (±SE) of both genotypes. The numbers below each graph represent the mice analyzed at each time point.
Thyroid histopathology after immunization with mouse thyroglobulin
We then immunized 182 mice (97 wild-type and 85 transgenic) with mouse thyroglobulin and assessed the severity of thyroid lesions (the thyroid histopathology score) at different time points thereafter. In both groups of mice, thyroid lesions were already present, although minimal (12% of the total thyroid area), at d 10 post immunization (Fig. 5). They then increased to 65% on d 14 and reached a peak of 90% on d 21 (Figs. 3, C and D, and 5). In wild-type mice, the thyroid histopathology score decreased to approximately 45% and remained around that value until d 100 (Fig. 5). In thyr-CIITA transgenic, instead, the score remained higher and became significantly higher than controls at d 35. Simple linear regression assessing the crude effect of genotype on score revealed that transgenic mice tended to have a more severe disease, with an average score that was 5% higher than that seen in wild-type, although this difference was not marked (P = 0.046). Multiple linear regression showed that, holding transgenic line and day of killing constant, thyr-CIITA transgenic mice had an average thyroid histopathology score that was 8% higher than that seen in wild-type (95% confidence interval, 1.8–13.4%; P = 0.033). Overall, these results indicate that thyr-CIITA transgenics developed thyroiditis that is moderately more severe than that of wild-type controls. The most pronounced separation between transgenic and wild-type was seen at d 35 (Figs. 3, E and F, and 5B); transgenics had an average thyroid histopathology score of 80% compared with the wild-type score of 55% (P = 0.032 by Wilcoxon rank sum test). In both groups of mice, the scores observed at d 14, 28, 35, 70, and 100 were significantly higher than those observed on d 0 and 10 (Figs. 3, A–H, and 5A) (P < 0.0001 comparing d 14, 28, 35, 70, or 100 with d 10 or 0).
FIG. 5. Thyroid histopathology score. A, Distribution of the mean scores in thyr-CIITA transgenic () and wild-type littermates () at d 0 and at several time points after immunization; B, individual scores in thyr-CIITA transgenic and wild-type littermates at d 35 after immunization.
In addition to severity, also the incidence of thyroiditis tended to be greater in thyr-CIITA transgenic mice, although at no time point did this difference reach statistical significance (Table 2).
TABLE 2. Incidence of experimental autoimmune thyroiditis in wild-type and thyr-CIITA transgenic mice after immunization with mouse thyroglobulin
Antibody response to thyroglobulin after immunization with mouse thyroglobulin
Longitudinal data regression analysis demonstrated that IgG1 and IgG2b isotypes specific for mouse thyroglobulin followed a similar pattern and reached similar levels in thyr-CIITA transgenic and controls. They began to increase at d 14 post immunization, became significantly higher at d 21, and reached a peak at d 28 (P = 0.0096 vs. d 21 and P = 0.001 vs. d 35) (Fig. 4, B and C). IgG1 and IgG2b then gradually decreased on d 35, 70, and 100, still remaining significantly higher than d 0.
The IgG2a isotype attained higher levels than the IgG1 and IgG2b isotypes and showed a different trend. It also reached a peak at d 21 but continued to stay high until d 70. No significant difference was observed between d 21, 28, 35, and 70, although the mean value looked higher at d 70 (Fig. 4A). IgG2a levels then decreased on d 100, still remaining significantly higher than d 0 (P = 0.0001). IgG2a levels did not differ between thyr-CIITA transgenic and controls.
Lymphocyte activation in the draining cervical lymph nodes after immunization with mouse thyroglobulin
It has been shown in experimental models of type 1 diabetes mellitus that the activation of islet-specific CD4 T cells that precedes the onset of diabetes occurs in the draining pancreatic lymph nodes or within the islets themselves (27). These observations have also been confirmed in primate experimental allergic encephalomyelitis and human multiple sclerosis (28). We therefore looked for CD4 T cell activation markers (increased expression of CD69 and CD44 and diminished display of CD62L) in cervical lymph nodes draining the thyroid gland. Before immunization with mouse thyroglobulin (d 0), there were few activated CD4 T cells (Fig. 6A) in both thyr-CIITA transgenic and wild-type littermates. At d 35 after immunization, there was a significant increase over d 0 (P < 0.001) in the percentage of activated CD4 T cells (Fig. 6B), with no difference, however, between transgenic and wild-type.
FIG. 6. Flow cytometry analysis of draining cervical lymph nodes for markers of T cell activation. Expression of CD69 and CD44 on gated CD4 T cells before (A) and 35 d after (B) immunization with mouse thyroglobulin. Four to six lymph nodes were collected from each mouse and pooled for the analysis, which was performed on 11 thyr-CIITA transgenic mice and seven wild-type controls.
Thyroid function after immunization with mouse thyroglobulin
Longitudinal regression analysis showed that total T4 significantly decreased from an average of 4.1 μg/dl at d 0 to an average of 2.2 μg/dl at d 14 (P = 0.0001). T4 then remained significantly low until d 35 (there was no difference among d 14, 21, 28, and 35) and then increased to normal levels on d 70 (P = 0.183 vs. d 0) and d 100 (P = 0.206 vs. d 0) (Fig. 7A). Holding all other covariates constant (transgenic line, sex, days of killing, and thyroglobulin-specific antibodies), there was no difference between thyr-CIITA transgenic and wild-type in the total T4 levels. Total T4 was a poor predictor of the thyroid histopathology score (P = 0.842 in the multiple linear regression analysis). The decrease in total T4 at d 14 was present before a rise in TSH, suggesting that this decrease represents the recovery phase from the euthyroid sick syndrome induced by the immunization procedure.
FIG. 7. Thyroid function. Total T4 levels (A) and TSH levels (B) in CIITA transgenic () and wild-type littermates () before and at several time points after immunization with mouse thyroglobulin. Dotted lines indicate the normal reference range in CBA/J mice, from 5th to 95th percentile. The range was 2.62–5.99 μg/dl for T4 and 63–182 U/ml for TSH.
TSH began to increase at 21 d after immunization and then became significantly higher than d 0 at d 35 (P = 0.045) and d 70 (P = 0.031) and returned to normal levels at d 100 (Fig. 7B). Holding all other covariates constant, thyr-CIITA transgenic tended to have higher TSH levels, although the difference did not reach statistical significance (P = 0.078). Similarly to what was observed for thyroid pathology, the greatest distinction between wild-type and thyr-CIITA transgenic was observed at d 35 post immunization (Fig. 5B). Among the humoral predictors measured in this study (total T4, TSH, IgG1, IgG2a, and IgG2b thyroglobulin antibodies), TSH was the best predictor of the thyroid histopathology score. Multiple linear regression showed that for every 100-U increase in serum TSH, the score increased by 6.1% (95% confidence interval, 4.2–8.0%; P < 0.00001).
Discussion
MHC molecules are undoubtedly associated with autoimmune diseases (13). In some cases, such as ankylosing spondylitis (29) and birdshot retinochoroidopathy (30), the strength of this association is so high (i.e. RRs around 100) that there should be little doubt about a causative role of MHC molecules in the pathway leading to autoimmunity. Yet, quoting from S. Beck and J. Trowsdale, "in no case can it be argued that a particular MHC allele is necessary or sufficient to cause autoimmune disease" (31). Also unknown is the importance of location; that is, whether MHC molecules play a causal role when expressed in secondary lymphoid organs or rather on parenchymal cells of the organ targeted by the autoimmune attack (32).
We focused on class II MHC molecules and used experimental autoimmune thyroiditis as a classical model of organ-specific autoimmunity. Class II molecules are primarily expressed on antigen-presenting cells of the hematopoietic lineage but can also be induced on endothelial and epithelial cells. This observation raised the question of whether nonhematopoietic parenchymal cells can function as antigen-presenting cells and whether they are required for damage to the target organ. Much of the work comes from the transplantation field and has generated conflicting results. On one side, Lakkis et al. (33) found that mice lacking all secondary lymphoid organs (spleen, lymph nodes, and mucosa-associated lymphoid tissues) are incapable of rejecting either skin allografts or primarily vascularized cardiac allografts, suggesting that secondary lymphoid organs are required for initiating the immune response. Similarly, Teshima et al. (34), using a model of graft-vs.-host disease in bone marrow chimeras expressing MHC molecules only on antigen-presenting cells but not on epithelial or endothelial cells, found that the graft was capable of inducing epithelial damage in the target organs (skin, gut, and liver) of the host that expressed MHC molecules only on the host antigen-presenting cells and not on the host target epithelium. These studies are consistent with recent direct visual evidence of the immune response, indicating that naive T cells constantly roam through secondary lymphoid organs where they encounter antigen-presenting cells that have taken up antigen from the target organs and brought it into the lymphoid organs (35, 36, 37, 38). These studies are also consistent with the notion that naive T cells possess receptors necessary for homing to secondary lymphoid organs rather than to nonlymphoid organs and with the notion that nonprofessional antigen-presenting cells usually lack the costimulatory molecules required for complete T cell activation. On the other side, however, Kreisel et al. (39) showed that naive CD8 T cells can indeed be activated in vitro and in vivo by nonprofessional antigen-presenting cells, most likely endothelial cells, and lead to prompt rejection of the cardiac allograft, suggesting that endothelial expression of MHC molecules can direct pathogenic immune responses independently of the expression of MHC molecules on professional antigen-presenting cells.
We expressed via transgenesis the transcriptional regulator CIITA specifically in a nonprofessional antigen-presenting cell, the thyrocyte. CIITA is a transcription factor that when defective causes a severe primary immunodeficiency called bare lymphocyte syndrome (22), entirely attributable to the absence of class II MHC expression. CIITA is a true coactivator because it does not bind directly to DNA but mediates its function on the class II MHC promoter through interaction with other proteins (40). It is considered a master regulator because it is necessary and sufficient for the expression of all the genes in the MHC class II pathway (21), although it has been recently shown to control other genes outside the MHC (41). CIITA has three different isoforms synthesized from different, cell-specific promoters (42); promoter 1 is used in dendritic cells, promoter 3 in B lymphocytes, and promoter 4 in nonhematopoietic cells. We have previously shown in vitro that promoter 4 mediates interferon--induced transcription of CIITA in thyrocytes (43), and thus we used the type 4 isoform of CIITA for the present study.
Under basal conditions, CIITA transgenic mice did not differ from wild-type littermates. They did not develop spontaneously autoimmune thyroiditis despite the aberrant expression of class II MHC molecules on thyrocytes. These findings are in agreement with a recent report by Li et al. (44), which shows that thyroidal expression of class II molecules driven by the TSH receptor promoter does not lead to apparent thyroid autoimmunity. The phenomenon does not seem to be limited to the thyroid gland. Stüve et al. (45) expressed CIITA, in fact, specifically in the central nervous system (using the promoter for the glial fibrillary protein), and Herkel et al. (46) expressed it in the liver (using the C-reactive protein promoter). In both cases, despite the aberrant expression of class II MHC molecules on astrocytes or hepatocytes, mice did not develop spontaneous autoimmunity. Taken together, these observations strongly suggest that the expression of class II MHC molecules on nonprofessional antigen-presenting cells is by itself not sufficient to initiate an autoimmune disease. We can postulate that this expression is more likely the results of inflammatory stimuli (such as interferon- and lipopolysaccharide) released once the immune cells infiltrate the target organ, rather than being the initiating event that recruits the immune cells to the target organ.
In addition to basal conditions, we were also interested in analyzing the role of class II molecules on thyrocytes during an autoimmune response. Upon immunization, thyr-CIITA transgenic mice developed thyroiditis that was moderately more severe and of greater incidence than that observed in wild-type littermates. The difference in severity, although higher in transgenics at all time points, became significantly higher only at d 35 after immunization. These results suggest that thyrocytes expressing class II MHC molecules play a secondary role in the initiation phase but can serve as better targets and facilitate greater immunopathology, once the process has already begun. Lack of significant difference in thyroiditis severity at earlier time points (d 10–28) can be explained by the fact that the immunization protocol we used was very effective in inducing thyroid lesions in the majority of the mice (incidence comprised between 86 and 95% in wild-type and 100% in transgenic).
The notion that class II MHC molecules on epithelial cells have a modulatory role once the autoimmune attack has begun is intriguing and may have possible relevance to Hashimoto’s thyroiditis and Graves’ disease (47). In these diseases, in fact, collectively referred to as autoimmune thyroiditis, the expression of class II molecules may indicate a particularly severe stage of the disease. This question is difficult to assess in humans; the use of a mouse model is essential because it allows us to analyze disease over a period of time, rather than at a single time point.
The greater severity of thyroid lesions displayed by thyr-CIITA transgenic was not a result of a greater number or function of activated T cells, considering that the CD4 T cells in the cervical lymph nodes expressed similar levels of activation markers in both wild-type and transgenics. The difference is more likely explained by the fact that thyrocytes expressing class II MHC molecules are better recognized by autoreactive CD4 T cells once these T cells have already expanded. In other words, thyrocytes may serve as endogenous immunogens that potentiate and expand the reactivity of T cells initially selected by the thyroglobulin immunization protocol. It is well established that T cell activation requires formation of a specialized cell-cell junction with the antigen-presenting cell, called the immunological synapse, which facilitates the outcome of immune recognition (48). The center of the synapse is occupied by T-cell receptor/CD3 complexes and CD4 coreceptors on the T cell side and by MHC/peptide complexes on the antigen-presenting side. Also in the center are costimulatory molecules, such as CD28 and CD40 ligand on T cells and B7 and CD40 on antigen-presenting cells. The periphery of the synapse is instead occupied by integrins, such as lymphocyte function-associated antigen-1 and CD2 on T cells and intercellular adhesion molecule-1 and CD58 on antigen-presenting cells. To function efficiently as antigen-presenting cell, the thyrocytes that aberrantly express MHC class II molecules should also express costimulatory molecules and integrins. Thyrocytes from patients with Graves’ disease and Hashimoto’s thyroiditis have been shown to express the integrins intercellular adhesion molecule-1 and CD58 (49, 50) and the costimulator CD40 (51, 52). The expression on thyrocytes of B7, probably the most potent costimulator, is less firmly established, although it was reported in Hashimoto’s thyroiditis (53) and differentiated thyroid tumors (54). B7 costimulation could also be supplied in trans from professional antigen-presenting cells to parenchymal cells (55). Thus, in theory, thyrocytes aberrantly expressing class II MHC molecules possess the signal transduction circuitry necessary for full T cell activation, although it is unknown whether individual components can associate in the supramolecular activation complex. Here we show that expression on thyrocytes of class II MHC is insufficient to initiate autoimmune thyroiditis. These results emphasize the importance of location in the initiation of the immune response, a location that likely requires the interaction between autoreactive T cell and antigen-presenting cells within the secondary lymphoid organs.
In conclusion, this study reports that expression of class II MHC molecules on epithelial thyroid cells is not required for the initiation of an autoimmune attack to the thyroid. The initiation, then, seems to be mainly mediated by the professional antigen-presenting cells in secondary lymphoid organs. Once the immune recognition has begun, however, the aberrant expression of class II on nonprofessional antigen-presenting cells can facilitate disease progression and yields a more severe disease. Additional studies are required to fully explain the intriguing association between MHC molecules and autoimmune diseases.
Acknowledgments
We are grateful to Drs. William D. M. Baldwin and James Tonascia for critical revisions of the manuscript.
References
Jacobson DL, Gange SJ, Rose N, Graham NMH 1997 Epidemiology and estimated population burden of selected autoimmune diseases in the United States. Clin Immunol Immunopathol 84:223–243
Caturegli P, Rose NR, Kimura M, Kimura H, Tzou SC 2003 Studies on murine thyroiditis: new insights from organ flow cytometry. Thyroid 13:419–426
Ueda H, Howson JM, Esposito L, Heward J, Snook H, Chamberlain G, Rainbow DB, Hunter KM, Smith AN, Di Genova G, Herr MH, Dahlman I, Payne F, Smyth D, Lowe C, Twells RC, Howlett S, Healy B, Nutland S, Rance HE, Everett V, Smink LJ, Lam AC, Cordell HJ, Walker NM, et al. 2003 Association of the T-cell regulatory gene CTLA4 with susceptibility to autoimmune disease. Nature 423:506–511
Grumet FC, Coukell A, Bodmer JG, Bodmer WF, McDevitt HO 1971 Histocompatibility (HL-A) antigens associated with systemic lupus erythematosus: a possible genetic predisposition to disease. N Engl J Med 285:193–196
Vladutiu AO, Rose NR 1971 Autoimmune murine thyroiditis: relation to histocompatibility (H-2) type. Science 174:1137–1139
Mungall AJ, Palmer SA, Sims SK, Edwards CA, Ashurst JL, Wilming L, Jones MC, Horton R, Hunt SE, Scott CE, Gilbert JG, Clamp ME, Bethel G, Milne S, Ainscough R, Almeida JP, Ambrose KD, Andrews TD, Ashwell RI, Babbage AK, Bagguley CL, Bailey J, Banerjee R, Barker DJ, Barlow KF, et al. 2003 The DNA sequence and analysis of human chromosome 6. Nature 425:805–811
Kumanovics A, Takada T, Lindahl KF 2003 Genomic organization of the mammalian MHC. Annu Rev Immunol 21:629–657
Dallman MJ, Mason DW 1983 Induction of Ia antigens on murine epidermal cells during the rejection of skin allografts. Transplantation 36:222–224
Bottazzo GF, Pujol-Borrell R, Hanafusa T, Feldmann M 1983 Role of aberrant HLA-DR expression and antigen presentation in induction of endocrine autoimmunity. Lancet 2:1115–1119
Mirakian R, Hammond LJ, Bottazzo GF 1998 Pathogenesis of thyroid autoimmunity: the Bottazzo-Feldmann hypothesis. Immunol Today 19:97–98
Lin L, Hungs M, Mignot E 2001 Narcolepsy and the HLA region. J Neuroimmunol 117:9–20
Kralovicova J, Hammarstrom L, Plebani A, Webster AD, Vorechovsky I 2003 Fine-scale mapping at IGAD1 and genome-wide genetic linkage analysis implicate HLA-DQ/DR as a major susceptibility locus in selective IgA deficiency and common variable immunodeficiency. J Immunol 170:2765–2775
Price P, Witt C, Allcock R, Sayer D, Garlepp M, Kok CC, French M, Mallal S, Christiansen F 1999 The genetic basis for the association of the 8.1 ancestral haplotype (A1, B8, DR3) with multiple immunopathological diseases. Immunol Rev 167:257–274
Klein J, Sato A 2000 The HLA system: second of two parts. N Engl J Med 343:782–786
Carlsson B, Wallin J, Pirskanen R, Matell G, Smith CI 1990 Different HLA DR-DQ associations in subgroups of idiopathic myasthenia gravis. Immunogenetics 31:285–290
Pociot F, McDermott MF 2002 Genetics of type 1 diabetes mellitus. Genes Immun 3:235–249
Levinson RD, Gonzales CR 2002 Birdshot retinochoroidopathy: immunopathogenesis, evaluation, and treatment. Ophthalmol Clin North Am 15:343–350, vii
Tomer Y, Davies TF 2003 Searching for the autoimmune thyroid disease susceptibility genes: from gene mapping to gene function. Endocr Rev 24:694–717
Levin L, Ban Y, Concepcion E, Davies TF, Greenberg DA, Tomer Y 2004 Analysis of HLA genes in families with autoimmune diabetes and thyroiditis. Hum Immunol 65:640–647
Maksymowych WP, Tao S, Vaile J, Suarez-Almazor M, Ramos-Remus C, Russell AS 2000 LMP2 polymorphism is associated with extraspinal disease in HLA-B27 negative Caucasian and Mexican Mestizo patients with ankylosing spondylitis. J Rheumatol 27:183–189
Mach B 1999 Perspectives: immunology. Regulating the regulator. Science 285:1367
Reith W, Mach B 2001 The bare lymphocyte syndrome and the regulation of MHC expression. Annu Rev Immunol 19:331–373
Beckstead JH 1995 A simple technique for preservation of fixation-sensitive antigens in paraffin-embedded tissues: addendum. J Histochem Cytochem 43:345
Schneider MJ, Fiering SN, Pallud SE, Parlow AF, St Germain DL, Galton VA 2001 Targeted disruption of the type 2 selenodeiodinase gene (DIO2) results in a phenotype of pituitary resistance to T4. Mol Endocrinol 15:2137–2148
Diggle PJ, Liang KY, Zeger SL 1995 Analysis of longitudinal data. Oxford, UK: Clarendon Press
Musti AM, Ursini VM, Avvedimento EV, Zimarino V, Di Lauro R 1987 A cell type specific factor recognizes the rat thyroglobulin promoter. Nucleic Acids Res 15:8149–8166
Hoglund P, Mintern J, Waltzinger C, Heath W, Benoist C, Mathis D 1999 Initiation of autoimmune diabetes by developmentally regulated presentation of islet cell antigens in the pancreatic lymph nodes. J Exp Med 189:331–339
de Vos AF, van Meurs M, Brok HP, Boven LA, Hintzen RQ, van der Valk P, Ravid R, Rensing S, Boon L, ’t Hart BA, Laman JD 2002 Transfer of central nervous system autoantigens and presentation in secondary lymphoid organs. J Immunol 169:5415–5423
Sims AM, Wordsworth BP, Brown MA 2004 Genetic susceptibility to ankylosing spondylitis. Curr Mol Med 4:13–20
LeHoang P, Ozdemir N, Benhamou A, Tabary T, Edelson C, Betuel H, Semiglia R, Cohen JH 1992 HLA-A29.2 subtype associated with birdshot retinochoroidopathy. Am J Ophthalmol 113:33–35
Beck S, Trowsdale J 2000 The human major histocompatability complex: lessons from the DNA sequence. Annu Rev Genomics Hum Genet 1:117–137
Lakkis FG 2003 Where is the alloimmune response initiated? Am J Transplant 3:241–242
Lakkis FG, Arakelov A, Konieczny BT, Inoue Y 2000 Immunologic ‘ignorance’ of vascularized organ transplants in the absence of secondary lymphoid tissue. Nat Med 6:686–688
Teshima T, Ordemann R, Reddy P, Gagin S, Liu C, Cooke KR, Ferrara JL 2002 Acute graft-versus-host disease does not require alloantigen expression on host epithelium. Nat Med [Erratum (2002) 8:1039] 8:575–581
Monks CR, Freiberg BA, Kupfer H, Sciaky N, Kupfer A 1998 Three-dimensional segregation of supramolecular activation clusters in T cells. Nature 395:82–86
Wulfing C, Davis MM 1998 A receptor/cytoskeletal movement triggered by costimulation during T cell activation. Science 282:2266–2269
Grakoui A, Bromley SK, Sumen C, Davis MM, Shaw AS, Allen PM, Dustin ML 1999 The immunological synapse: a molecular machine controlling T cell activation. Science 285:221–227
Itano AA, Jenkins MK 2003 Antigen presentation to naive CD4 T cells in the lymph node. Nat Immunol 4:733–739
Kreisel D, Krupnick AS, Gelman AE, Engels FH, Popma SH, Krasinskas AM, Balsara KR, Szeto WY, Turka LA, Rosengard BR 2002 Non-hematopoietic allograft cells directly activate CD8+ T cells and trigger acute rejection: an alternative mechanism of allorecognition. Nat Med 8:233–239
Ting JP, Trowsdale J 2002 Genetic control of MHC class II expression. Cell 109(Suppl):S21–S33
Reith W 2003 ComPLEXIN new targets for CIITA. Nat Immunol 4:819–820
Muhlethaler-Mottet A, Otten LA, Steimle V, Mach B 1997 Expression of MHC class II molecules in different cellular and functional compartments is controlled by differential usage of multiple promoters of the transactivator CIITA. EMBO J 16:2851–2860
Mori-Aoki A, Pietrarelli M, Nakazato M, Caturegli P, Kohn LD, Suzuki K 2000 Class II transactivator suppresses transcription of thyroid-specific genes. Biochem Biophys Res Commun 278:58–62
Li YS, Kanamoto N, Hataya Y, Moriyama K, Hiratani H, Nakao K, Akamizu T 2004 Transgenic mice producing major histocompatibility complex class II molecules on thyroid cells do not develop apparent autoimmune thyroid diseases. Endocrinology 145:2524–2530
Stuve O, Youssef S, Slavin AJ, King CL, Patarroyo JC, Hirschberg DL, Brickey WJ, Soos JM, Piskurich JF, Chapman HA, Zamvil SS 2002 The role of the MHC class II transactivator in class II expression and antigen presentation by astrocytes and in susceptibility to central nervous system autoimmune disease. J Immunol 169:6720–6732
Herkel J, Jagemann B, Wiegard C, Lazaro JF, Lueth S, Kanzler S, Blessing M, Schmitt E, Lohse AW 2003 MHC class II-expressing hepatocytes function as antigen-presenting cells and activate specific CD4 T lymphocytes. Hepatology 37:1079–1085
Weetman AP 2003 Autoimmune thyroid disease: propagation and progression. Eur J Endocrinol 148:1–9
Dustin ML 2002 The immunological synapse. Arthritis Res 4(Suppl 3):S119–S125
Zheng RQ, Abney ER, Grubeck-Loebenstein B, Dayan C, Maini RN, Feldmann M 1990 Expression of intercellular adhesion molecule-1 and lymphocyte function-associated antigen-3 on human thyroid epithelial cells in Graves’ and Hashimoto’s diseases. J Autoimmun 3:727–736
Bonita RE, Rose NR, Rasooly L, Caturegli P, Burek CL 2002 Adhesion molecules as susceptibility factors in spontaneous autoimmune thyroiditis in the NOD-H2h4 mouse. Exp Mol Pathol 73:155–163
Faure GC, Bensoussan-Lejzerowicz D, Bene MC, Aubert V, Leclere J 1997 Coexpression of CD40 and class II antigen HLA-DR in Graves’ disease thyroid epithelial cells. Clin Immunol Immunopathol 84:212–215
Metcalfe RA, McIntosh RS, Marelli-Berg F, Lombardi G, Lechler R, Weetman AP 1998 Detection of CD40 on human thyroid follicular cells: analysis of expression and function. J Clin Endocrinol Metab 83:1268–1274
Battifora M, Pesce G, Paolieri F, Fiorino N, Giordano C, Riccio AM, Torre G, Olive D, Bagnasco M 1998 B7.1 costimulatory molecule is expressed on thyroid follicular cells in Hashimoto’s thyroiditis, but not in Graves’ disease. J Clin Endocrinol Metab 83:4130–4139
Shah R, Banks K, Patel A, Dogra S, Terrell R, Powers PA, Fenton C, Dinauer CA, Tuttle RM, Francis GL 2002 Intense expression of the B7-2 antigen presentation coactivator is an unfavorable prognostic indicator for differentiated thyroid carcinoma of children and adolescents. J Clin Endocrinol Metab 87:4391–4397
Mandelbrot DA, Kishimoto K, Auchincloss Jr H, Sharpe AH, Sayegh MH 2001 Rejection of mouse cardiac allografts by costimulation in trans. J Immunol 167:1174–1178(Hiroaki Kimura, Miho Kimu)