Reevaluating Thyrotropin Receptor-Induced Mouse Models of Graves’ Disease and Ophthalmopathy
http://www.100md.com
内分泌学杂志 2005年第2期
Centre for Endocrine and Diabetes Sciences (G.B., G.M., M.L.) and Departments of Ophthalmology (G.B.) and Medical Microscopy Sciences (C.v.R.), School of Medicine, Cardiff University, Cardiff CF14 4XN, United Kingdom
Address all correspondence and requests for reprints to: Marian Ludgate, Centre for Endocrine and Diabetes Sciences, School of Medicine, Cardiff University, Heath Park, Cardiff CF14 4XN, United Kingdom. E-mail ludgate@cf.ac.uk
Abstract
We aimed to establish and extend the characterization of murine models of thyroiditis and Graves’ ophthalmopathy, induced by transfer of TSH receptor (TSHR) primed T cells. Experiments were performed in a different animal unit but using female BALB/cbyJico mice from the same supplier as previously. We report our findings together with a reevaluation of the earlier studies. In the first experiment, genetic immunization or TSHR fusion protein induced TSHR antibodies in all nine mice. Some of the antibodies functioned as thyroid-stimulating antibodies and/or TSH binding inhibiting Igs with two of seven mice having elevated T4. Thyroiditis and orbital changes were absent. Splenocyte transfer induced no immune response in naive BALB/cbyJico recipients. Subsequently genetic immunization or fusion protein-treated mice were maintained in either local or Brussels conditions (water, chow, and bedding). TSHR antibodies were induced in nine of nine Brussels (with decreased T4 in one of nine) but five of nine local mice. No thyroiditis or orbital changes were induced, but misleading fixation artefacts in extraocular muscles were noted. Nonspecific in vitro stimulation induced more CD-4+/IL-4+ cells in Brussels maintained. TSHR stimulation produced a significant increase in IL-4 secretion in six of nine local but one of seven Brussels mice. Thyroids from many TSHR-treated and control mice contained ectopic thymus. Our results confirm that thyroiditis is required for disease transfer but indicate the heterogeneity in TSHR-induced immune response in an inbred strain. Ectopic thymus can masquerade as thyroiditis, and care is required to avoid muscle artefacts. Because neither animal unit is pathogen free, microbial environment may contribute to determining TSHR-induced responses.
Introduction
GRAVES’ DISEASE (GD) is a common autoimmune condition in which thyroid-stimulating antibodies (TSABs) cause hyperthyroidism (reviewed in Ref.1). It is often accompanied by Graves’ ophthalmopathy (GO) in which expansion of the orbital contents, by hypertrophy and hyperplasia, produces proptosis (reviewed in Ref.2). The TSH receptor (TSHR) is an important autoantigen in both disorders, being the target of pathogenic autoantibodies in GD (reviewed in Ref.3) and promoting lymphocyte homing to the orbit in GO (reviewed in Ref.2). TSHR expression is increased during adipogenesis (4, 5), a mechanism causing expansion in the orbital volume, as illustrated by the exacerbation of GO in a person with type 2 diabetes treated with the proadipogenic drug, pioglitazone (6). We have demonstrated that adipogenesis is not restricted to the orbit but also occurs in the neck of GD patients but is not modified by thyroid status (7).
A central question in GD and GO is how tolerance to the TSHR is overcome. To address this, in the absence of a spontaneous model, a number of TSHR-induced models have been developed (reviewed in Refs.8, 9, 10). One of the approaches employed transfer of TSHR-primed T cells to naive syngeneic recipients (11). Priming could be achieved by either using a TSHR fusion protein (FP) or genetic immunization (GI) with a plasmid encoding the human TSHR. The studies used BALB/cbyJico mice from a French supplier and found that disease was more severe in females, as predicted. The TSHR antibodies produced in the model were mainly TSH binding inhibiting Igs (TBIIs) and thyroid-blocking antibodies (TBABs); TSABs were only rarely induced. One of the most significant attributes of the model was the development of Th2 thyroiditis in approximately 60% of the recipients, providing thyroiditis had been induced in some of the animals supplying the TSHR-primed T cells. Furthermore, about one third of the animals with thyroiditis also displayed orbital changes, including immune infiltration and edema, similar to those seen in GO (12). The severity of the orbital changes correlated positively with the extent of the Th2 skew in the thyroid immune infiltrate of that animal.
The model is unusual in being one of only three examples in which thyroiditis and orbital changes have been reported (12, 13, 14). One of the others was also developed in Brussels but uses GI of NMRI outbred mice to induce TSABs, increased thyroxine levels, thyroiditis, and orbital changes in approximately 20% of the animals (14). The higher disease rate of the transfer model makes it more attractive as a tool to increase our understanding of GO, a condition that remains a challenge for endocrinologists and ophthalmologists alike.
Our initial aim was to extend the characterization of the transfer model, including whether the disease follows a Rundle’s curve (15) as in humans. Subsequently we had planned to use the model to investigate alternative therapies for GO, e.g. peroxisomal proliferator-activated receptor- antagonists. In our initial experiments we used GI and FP injection of the BALB/cbyJico mice available from our local animal unit. Despite both protocols inducing a significant humoral response to the TSHR in the majority of animals treated, none of the mice developed thyroiditis. We also failed to transfer disease to naive syngeneic recipients using splenocytes from the GI- or FP-treated mice as a source of primed T cells (data not shown).
Consequently we attempted to reproduce the conditions resulting in the induction of thyroiditis and orbital changes reported previously (11, 12). This included studying the BALB/cbyJico substrain from the same French supplier. Our results differed significantly and led us to reevaluate the earlier studies in the light of the current experiments. We now report our findings obtained exclusively with the French BALB/cbyJico substrain.
Materials and Methods
Immunization with extracellular domain (ECD)/maltose binding protein (MBP) or pcDNA3-TSHR
All procedures were performed in accordance with the Animals (Scientific Procedures) Act 1986 and appropriate project and personal license approval.
Immunization (under halothane anesthesia) of 6-wk-old female BALB/cbyJico (Iffa Credo, Les Oncins, France) was performed on d 0 (with preimmune blood sampling) and at 3, 6, and 9 wk. The antigens employed were the pcDNA3-TSHR plasmid, prepared using a standard maxiprep kit (Qiagen, Santa Clarita, CA) according to the manufacturer’s instructions, and an FP comprising the ECD of the TSHR fused to MBP as previously described (16). GI mice received 100 μg pcDNA3-TSHR in 25% sucrose shared equally between both anterior tibialis muscles at each time point. FP-treated animals had an initial 100 μg FP ip in a 200 μl volume of 0.9% saline with adjuvant, 0.1% aluminum oxide, 0.4% magnesium hydroxide (Maalox, Rhone-Poulenc Rorer, Paris, France), and 50 μl acellular Bordetella pertussis vaccine (Wyeth, Madison, NJ). Thereafter immunizations used 50 μg FP with adjuvant. Mice were killed after 14 wk; their serum stored; and spleens, thyroids, and orbital contents removed. Thyroid and orbital tissues were fixed in 10% phosphate-buffered formalin with 2% acetic acid for a minimum of 48 h for paraffin sections using standard techniques. The thyroids were stained with hemotoxylin and eosin and orbital tissue with periodic acid and Schiff’s reagent counterstained with hematoxylin.
T-cell transfer to naive recipients
Spleens removed from immunized and control animals were mechanically disrupted and frozen in 10% dimethylsulfoxide in fetal calf serum. Forty-eight hours before transfer, the frozen aliquots were rapidly thawed and viable splenocytes obtained from the opaque interface of a Ficoll-Hypaque gradient. After washing, in vivo primed and nonprimed cells underwent in vitro priming with FP, all as previously described (11).
Naive 9-wk mice were immunized via the tail veins, with 5 x 105 splenocytes in 100 μl PBS. Groups of seven each male and female BALB/cbyJico (Iffa Credo) mice, received FP or pcDNA3-TSHR primed or nonprimed cells. The latter came from two male and two female 9-wk BALB/cbyJico animals. Mice were killed at 10 wk, excluding six males and females that were reimmunized at 15 wk with 5 x 106 FP or pcDNA3-TSHR-primed cells and killed 12 wk later.
In all cases sera were stored and orbital tissues plus thyroids (one lobe each) were either fixed overnight in 0.2% glutaraldehyde or 4% formaldehyde for plastic sections or snap frozen for cryostat sectioning.
Environmental modification
A second group of 20 imported BALB/cbyJico female mice were initially all maintained on standard local water, feed (rat and mouse no. 1 maintenance, SDS Ltd., Witham, UK), and bedding. The mice were then immunized, 10 with pcDNA3-TSHR and 10 with FP [0.2 μg pertussis toxin (Sigma, St. Louis, MO) replaced acellular B. pertussis]. At immunization the two groups were subdivided; half continued on standard maintenance, whereas the other received imported water, bedding, and feed (RN-01–20K12, Carfil Nutrients) as used by Institut de Recherche en Biologie Humaine et Moleculaire (IRIBHM), Brussels. The mice were killed after 14 wk, their sera stored, splenocytes frozen, and thyroids plus orbital contents prepared for plastic sectioning.
Antibodies to the TSHR
ELISA using ECD-MBP.
Ninety-six-well plates were coated overnight at 4 C with 100 μl/well of FP (10 μg/ml) in carbonate/bicarbonate coating buffer. Sera were tested at 1:100 by incubation for 2 h at room temperature followed by 100 μl of an antimouse IgG horseradish peroxidase conjugate for 30 min at room temperature. Detection used 100 μl of substrate, and the OD was measured in a plate reader at 410 nm.
Flow cytometry using Chinese hamster ovary cells expressing human TSHR
We used Chinese hamster ovary cells expressing the TSHR-ECD anchored by a glycophosphatidylinositol link (17). The cells were maintained in Ham’s F12 medium supplemented with 10% fetal calf serum and antibiotics until 80–90% confluent and then detached with 5 mM EGTA/EDTA in PBS. They were resuspended in PBS/BSA 0.1% and 100-μl aliquots containing 2 x 105 cells incubated with 2 μl mouse serum for 1 h at room temperature. They were incubated at 4 C in the dark with fluorescein isothiocyanate (FITC)-conjugated F(ab)2 fragment of goat antimouse Igs (diluted 1:16, Dako, Carpinteria, CA) and counterstained with propidium iodide to gate out damaged cells. Fluorescence from 5000 cells was measured using a FACScan flow cytometer (Becton Dickinson, Lincoln Park, NJ) and compared with preimmune sera, also at 1:50 dilution. Histogram shifts were examined using the Kolmogorov-Smirnov test (18). This compares the cumulative distribution of the data with the expected cumulative Gaussian distribution and bases its value simply on the largest discrepancy-Dmax (Cellquest Pro software, Becton Dickinson). Cumulative FITC fluorescence curves were automatically plotted with test sera vs. preimmune to calculate a D value. The 97th centile of the D values from recipients of nonprimed T cells (n = 7) after transfer (19 wk old) was 0.3986; values above this were considered positive (0.4).
TBIIs
Glycophosphatidylinositol cells were plated in 96-well plates at 2 x 104 cells/well. The next day, 100 μl salt-free Hanks’ balanced salt solution was added to 100 μl I125-labeled bovine TSH tracer (kindly provided by BRAHMS Diagnostica, Berlin, Germany) and 3 μl serum (triplicates). The nonspecific binding was determined by adding 100 mU bovine TSH. After a 3-h incubation at room temperature, the cells were washed, lysed with 100 μl 1 N NaOH (30 min, 4 C), transferred to counting tubes, and radioactivity measured in a -counter. Results are expressed as a percentage of the counts per minute bound by the test sera relative to the preimmune value of 100% tracer binding after subtracting the nonspecific binding. The TBII indices were then plotted against 97th and third centiles constructed from individual nonprimed and preimmune sera (n = 9).
TSABs
TSABs were assayed using LULU* cells, which express both the human TSHR and a cAMP-responsive luciferase reporter construct, as previously described (19). Briefly, the cells were plated in Ham’s F12 medium supplemented with 10% charcoal-stripped serum and then incubated the next day for 4 h at 37 C in 50 μl serum-free Ham’s F12 per well, containing 5 μl of test serum (triplicates). After washes, plates were air dried and lysed by 20 min shaking in 50 μl lysis buffer (Promega, Madison, WI) per well. Luciferase assay reagent (Promega) was prepared according to the manufacturer’s instructions, added to the wells, and light output read using a luminometer (Tropix Microplate luminometer, Applied Biosystems, Foster City, CA). The output of relative light units from test sera was compared with preimmune sera to give a stimulation index (SI). Test sera were considered positive when greater than the 97th centile of the preimmune and nonprimed recipients (n = 9); this was calculated as SI greater than 1.8.
Measurement of serum T4
Total serum-free T4 was measured on duplicate 10-μl volumes of serum, using a commercially available competitive RIA (Amerlex, Amersham International, Amersham, UK) according to the manufacturer’s instructions. Results are reported as micrograms per deciliter.
Splenocyte stimulation using TSHR or phorbol-12-myristate-13-acetate (PMA)/ionomycin
Splenocytes from the environment modification group and two BALB/cbyJico mice in which the GI had been performed at IRIBHM, were thawed, resuspended in 5 ml RF10, and incubated overnight at 37 C. Trypan blue exclusion confirmed viability greater than 85%, the splenocytes were resuspended at 2 x 106 cells/ml in RF10 for polyclonal and antigen-specific stimulation. Splenocyte aliquots of 1 x 106 cells were prepared for intracytoplasmic cytokine staining (ICCS) and parallel measurements of cytokines released into the culture supernatant.
For ICCS, the control sample was incubated for 4 h at 37 C with monensin (7 μM) and the stimulated sample with PMA (25 ng/ml), ionomycin (1 μg/ml), and monensin (7 μM). Control and stimulated splenocytes were then washed and resuspended in 1 ml PBS/BSA. Surface antigen detection used 10 μl each of peridinin chlorophyll-a protein-conjugated rat antimouse CD4 (L3T4) and allophycocyanin-conjugated rat antimouse CD8a (Ly-2) monoclonal antibodies (PharMingen, San Diego, CA). After 30 min incubation at room temperature in the dark, cells were fixed and permeabilized using a kit (CPK-200, cell permeabilization kit, ImmunologicalsDirect.com, Kidlington, UK) according to the manufacturer’s instructions. Simultaneously 10 μl anticytokine monoclonal antibodies (PharMingen) were added, r-phycoerythrin (PE)-conjugated rat antimouse IL-4 and FITC-conjugated rat antimouse interferon (IFN)-, and incubated at room temperature in the dark for 30 min. Then 400 μl of 2% paraformaldehyde was finally added and after a minimum of 10 min either analyzed or kept overnight at 4 C. The splenocytes were analyzed using four-color flow cytometry to count 105 events (FACSCaliber, Becton Dickinson). Once gated according to surface markers, the percentage of cells expressing IL-4 and IFN from control and stimulated splenocytes was measured using arbitrary quadrants. The figure recorded was the percentage of cells expressing cytokines in the stimulated sample minus the unstimulated; differences were examined for significance between groups.
For supernatant cytokine measurements the cells were incubated as for ICCS but without monensin, centrifuged (400 x g, 5 min), supernatants aspirated, and stored at –80 C until assayed for cytokines.
In addition, we used two separate sources of TSHR, the ECD/MBP FP and also a pooled panel of synthetic TSHR peptides spanning the ECD (20). Ninety-six-well plates were loaded with 250 μl of splenocyte suspension (5 x 105 cells) in RF10 and incubated (in triplicate) at 37 C for 6 d with 10 μg/ml FP or 5 μg/ml pooled TSHR peptides (dissolved in dimethylsulfoxide). Supernatant replicates were pooled and stored at –80 C until assayed. Cytokines secreted from splenocytes stimulated by either polyclonal or specific stimuli were measured by commercial murine IFN and IL-4 ELISA (Duoset, R&D Systems, Minneapolis, MN) according to the manufacturer’s instructions.
Standard curves were plotted to calculate the concentration of cytokines in picograms per milliliter of sample supernatants. The value recorded is the amount of cytokine present in the stimulated sample minus that of the unstimulated sample.
Statistical analysis
Data were analyzed using software, SPSS for Windows (version 11.0.1, SPSS, Chicago, IL) or the DOS-based INSTAT (version 2.0, GraphPad Software, San Diego, CA). Categorical data were examined using 2 with Fisher’s exact test where required and continuous data using t tests, ANOVA, or Kruskal-Wallis where appropriate. Homogeneity of variance was assessed by Levene’s (SPSS) or Bartlett’s (INSTAT) statistics and post hoc testing mainly by Tukey honestly significantly different. Test values are reported with P values and test type.
Results
A strong humoral response to the TSHR is induced using plasmid or fusion protein
In the first group of female BALB/cbyJico mice, a strong humoral response was induced to the TSHR using either GI or injection of FP. All of the mice developed an IgG antibody response to the FP measured by ELISA. The FP animals developed a rapid response with antibodies detected as early as 2 wk and peaking around 6 wk. Their response is potentially to both components of the fusion protein (ECD and MBP), unlike that of the more slowly evolving antibody response seen in the pcDNA3-TSHR immunized group (Fig. 1).
FIG. 1. ELISA from the first immunized group showing immune response to ECD/MBP (GI, solid line; FP immunized, dotted line, and 97th centile of preimmune sera, fine dotted line).
The sera were then tested by flow cytometry, a method that detects antibodies predominantly to conformational epitopes of the TSHR. Representative histograms are shown in Fig. 2, and for comparison, a typical plot obtained using sera from a similarly aged female, pre- and post receipt of nonprimed T cells. All nine GI or FP-treated mice from group 1 demonstrated D values above the 97th centile of the nonprimed T-cell recipient control group (n = 7), as shown in Fig. 3.
FIG. 2. Representative anti-TSHR IgG flow histograms from immunized group 1. Outlined histograms represent preimmune sera and solid histograms sera at 14 wk. A and B, Two GI. C and D, Two FP-treated mice. E, Outlined histogram from a female mouse before and solid histogram 10 wk after transfer of nonprimed (NP) T cells.
FIG. 3. IgG class antibodies to the TSHR measured by flow cytometry and presented as D values for individual mice from GI and FP-immunized groups 1 and 2. The third and 97th centiles (fine dotted line) were calculated from recipients of nonprimed T cells 10 wk after transfer (n = 7).
When the biological activity of the TSHR antibodies was evaluated, using an assay for TBIIs (Fig. 4), three of five GI and one of four FPs were positive (falling below the third centile of the nonprimed T-cell recipient control group, n = 7).
FIG. 4. TBII activity in immunized groups 1 and 2. The third and 97th centiles (fine dotted line) were calculated from recipients of nonprimed T cells 10 wk after transfer (n = 7).
In a luminescent bioassay for TSABs, one of five of the GIs and one of four of the FPs were strongly positive with a further two of five GIs and one of four FPs having an SI greater than 1.8 and consequently positive for TSAB (Fig. 5).
FIG. 5. TSAB activity (SI) in immunized groups 1 and 2. The 97th centile (SI = 1.8, fine dotted line) was calculated from recipients of nonprimed T cells 10 wk after transfer (n = 7).
The response is associated with elevated T4 levels but not thyroiditis or orbital changes
We used preimmune sera from these mice and also sera from similarly aged female recipients of nonprimed T cells to obtain the normal range of T4 values (n = 9). This was very broad, 2.57 (third centile) and 8.91 (97th centile). Two of four of the GI mice tested had values higher than this range; none of the animals had T4 below the third centile. The results are shown in Fig. 6.
FIG. 6. Total T4 levels for individual mice from GI and FP-immunized groups 1 and 2. The third and 97th centiles (dotted lines) were calculated from recipients of nonprimed T cells 10 wk after transfer (n = 7).
None of the receptor-immunized mice displayed thyroiditis or orbital changes (described in more detail below).
Limited transfer of TSHR immune responsiveness to naive recipients
Splenocytes from group 1 were used as a source of primed T cells for transfer to naive French female BALB/cbyJico recipients. Using flow cytometry, none of the recipients developed antibodies to the TSHR. In contrast ,by ELISA, six of 14 mice demonstrated a weak response, three males and three females and in all but one, priming was with the FP so that antibodies to MBP would also be detected. When using bioassays for the TSHR antibodies, all mice were negative for TSAB and TBII activity. Two male recipients of FP-primed T cells displayed T4 levels above the 97th centile (none of these data shown). None of the animals developed thyroiditis or changes in their orbital tissues.
Does environmental modification change the TSHR-induced response?
The absence of induced thyroiditis or orbital changes (described below) combined with the minimal humoral reactivity obtained indicate no transfer of immune response to the TSHR. Because we were using mice identical with those used previously for TSHR-induced thyroiditis and transfer of disease; we investigated whether modifying the environment might alter the outcome. A second group of 20 French female BALB/cbyJico mice received either GI or treatment with the FP. Half were maintained on standard chow/bedding/water provided by the local animal unit. The other half were provided with chow, bedding, and water identical with that used at IRIBHM in Brussels. Neither the local nor Brussels animal units are pathogen free. All 10 of the GI and eight of 10 of the FP-treated mice survived and were available for analysis.
The plasmid/protein-induced humoral response is modified by environment
Using flow cytometry, all of the FP-treated mice had a strong immune response to the TSHR that developed by 9 wk. In the GI group, all of the mice in Brussels conditions (GIB) but only one of five in local conditions (GIL) had developed TSHR antibodies (Fig. 3).
None of the mice had TSAB or TBII activities (Figs. 4 and 5)
Significantly lower average T4 was found in group 2, compared with group 1, immunized mice (Fig. 6), 8.64 vs. 3.67 μg/dl (P < 0.0001, Mann-Whitney U). In group 2, 16 of 17 tested T4 levels fell within the broad normal range, with the exception of a FP-treated mouse in Brussels conditions (FPB) that had a value below the third centile.
The T-cell response is altered by immunization and modified by environment
Splenic T-cell profiles were examined using flow cytometry. As shown in Table 1, there is a higher percentage of CD4+ cells in the spleens of FP-treated animals, compared with genetic or nonimmunized mice, with a significant difference between FPB and GIL (P = 0.048). All immunization protocols induce a significant reduction in the percentage of CD8+ cells and consequently an increase in the CD4+ to CD8+ ratio. GI produced an increase in the ratio of non-T cell to T cell, and this was significant for mice in local conditions, compared with all other groups. Changes were due predominantly to the immunization method with modification of the environment having only a minimal effect on the proportions of CD4+, CD8+, or non-T splenocytes.
TABLE 1. Mean percentages (and SEM) of CD4+ and CD8+ cells and CD4+:CD8+ and non-T:T ratios in the splenocyte populations of the second environment-modified immunized group
Use of the nonspecific stimulus, PMA, and ionomycin produces a snapshot of the T-cell repertoire after treatment. The percentages of CD4+ and CD8+ cells expressing IL-4 and IFN were low in all mice studied. Percentages of CD4+/IL-4, CD4+/IFN, and CD8+/IL-4 cells were higher in FP-treated mice than in GI with Brussels maintained having the larger value in each case, as shown in Fig. 7. The difference between percentages of CD4+/IL4+ cells in FPB and mice treated with FP maintained in local conditions (FPL) were significantly different (P = 0.042) but did not reach significance for CD8+/IL4+ or CD4+/IFN. Percentages of CD8+/IFN-expressing cells were similar in all groups.
FIG. 7. Summary of the percentages of CD4+ and CD8+ splenocyte populations stimulated with PMA/ionomycin to express IFN and Il-4 analyzed by four-color flow cytometry after intracellular cytokine staining. The plots show the median (heavy horizontal line), interquartile range (boxed area), and extreme values (bars). BRUS, Mice treated by GI performed in Brussels; NORM, nontreated mice.
We then investigated antigen-specific responses by measuring cytokines in the supernatants of splenocytes incubated with a pooled panel of TSHR peptides or FP. Levels of IL-4 secretion were detectable but low. Compared with normal splenocytes, TSHR (FP and/or peptides)-specific IL-4 secretion was significantly increased in four of four FPL (range 4.7–25.6 pg/ml) and two of five GIL (5.9–6.8 pg/ml), compared with one of three FPB (12.2 pg/ml) and none of four GIB (<2.6 pg/ml) in Brussels conditions.
Finally, we examined the splenocyte profiles in two BALB/cbyJico mice in which the GI had been performed at IRIBHM. These mice displayed a similar proportion of CD4+ and CD8+ but significantly fewer non-T cells than GI mice-treated in Cardiff (Table 1) and the lowest percentage of CD4+ cells expressing IL4+ (Fig. 7). When assessing their antigen-specific responses, we obtained similar results to those of the GIB-treated mice, with IFN levels below the limit of detection and no significant increase in IL-4 secretion.
Immunized, transfer, and control mice demonstrated normal thyroid pathology and often contained ectopic thymus
Mice from immunized, transfer, and control groups had normal thyroid macroscopic appearances with no goiter formation or hypertrophy. Normal thyroid histology was seen in all three groups and comprised heterogeneity of follicular size and thyrocyte depth with occasional apical budding and colloid endocytosis (Fig. 8, A and B). Mast cells were resident in connective tissue adjacent to and within thyroids, with no greater frequency between immunized and control mice. Additionally, normal thyroids contained parathyroid glands and cysts that were seen frequently associated with highly organized clusters of encapsulated densely stained large nucleated cells (Fig. 8, C, F, I, and J). In further sections these clusters of cells had less nuclear staining (Fig. 8, F and J) and contained specialized cells, murine equivalents of Hassal bodies (Fig. 8L), identifying them as ectopic medullary thymus and the former dense staining cells as cortical thymus. The appearances were by no means uniform; the ectopic thymus could have a large extrathyroidal location (Fig. 8, G and H) but most commonly was located in the periphery of the thyroid gland. At times, ectopic thymus could appear as a lone cluster of cells (Fig. 8, D and E) but if thoroughly sectioned was always partnered by the parathyroid gland. TSHR immunization did not alter the apparent frequency of ectopic thymus and was present in 13 of 26 immunized mice, compared with six of nine control nonimmunized mice (P = 0.46, Fisher’s exact). However, routine sectioning sometimes missed the two parathyroid glands, and this might have led us to underreport the incidence of the congenital ectopic thymic anomaly.
FIG. 8. Normal follicular heterogeneity was present in mice from control, immunized, and transfer animals with occasional areas of active endocytosis (A and B; arrow in B shows endocytosis). Parathyroid glands (P) had a characteristic mosaic of dark, intermediate, and clear staining cells frequently containing ciliated cysts (C). Ectopic cortical thymus (T) could appear to be a lone cluster of inflammatory cells in the periphery of the thyroid (D and E). Medullary thymus contained cells with less staining than the cortex with a lymphoblastoid appearance (F). The range of appearances of ectopic thymus was wide, examples of large extrathyroidal locations were seen (G and H), and the usual clear separation of parathyroid from ectopic thymus was not always present (I). Small fragments of medullary thymus could be difficult to define (J) but had additional histological features of thymocytes clustering around venules (V in K) and murine equivalents of Hassall bodies (H in L). Semithin sections mounted in LR white, stained with toluidine blue. Scale bars A to K, 100 μm; L, 50 μm.
Orbital connective tissue and fat develop no inflammatory changes: different ocular muscle types and contraction artefact cause diagnostic difficulty
No significant macroscopic changes akin to thyroid eye disease were present in terms of proptosis, orbital fat expansion, or extraocular muscle hypertrophy.
Control and immunized ocular muscle in cross-sections had areas in which connective tissue was common between muscle fibers (Fig. 9A); this corresponded to the singly innervated orbital muscle layers without edema or accumulation of periodic acid and Schiff’s-positive material to suggest glycosaminoglycan deposition. Ocular muscle fibers with a closely packed appearance were the multiply innervated global muscle fibers (Fig. 9B). Both orbital and global layer muscle fibers had areas of contracted, buckled myofibrils, resulting from disinserting unfixed muscle from the globe. Contraction-induced buckling deformed the connective tissue, particularly around orbital fibers, and could give rise to a false impression of edema. Sections of myofibrils could appear normal adjacent to segments with contraction bands and knots (Fig. 9C). These contraction artefacts were not present in later control samples fixed with the muscles left inserted to the globe and attached to their origins around the optic nerve at the annulus of Zinn, which splinted the muscles and prevented their contracture (Fig. 9D).
FIG. 9. Ocular muscle from control and immunized mice comprised two fiber types histologically. Orbital fibers (MF) were clustered with a few fibers surrounded by connective tissue and were interspersed with large myelinated nerve fibers (N in A). Global muscle fibers were clustered in large numbers with less connective tissue and nerve fibers (B). Postmortem muscle contraction artifacts of knots (K) and contraction bands (CB) occurred adjacent to normal segments of muscle (N), resultant disruption of connective tissue could be misinterpreted as edema (C). Fixation before disinsertion from the muscle attachments removed such fixation artifacts (D). Occasional multiloculated adipocytes appeared within the connective tissues of both control and immunized mice (E). Orbital fat had a normal lobular organization with mast cells (MC) in the adjacent connective tissues without hypertrophy in the immunized groups (F). None of the orbital tissues showed inflammatory changes. Semithin sections mounted in LR white stained with toluidine blue. Scale bars, 50 μm.
Orbital fat appearances were similar in control and immunized animals. Organized lobules of fat were bordered by connective tissue, within which scatterings of mast cells were resident (Fig. 9F). Sporadic multivacuolated adipocytes were found within areas of connective tissue in control and immunized mice representing the normal background of adipogenesis (Fig. 9E). No lymphocytic infiltration was present in the muscles, fat, and connective tissues of the orbits of any immunized or control mice.
Discussion
Our initial aim was to establish a previously reported TSHR-induced animal model, using the BALB/cbyJico strain, and apply it to answer questions pertinent to the pathogenesis of GD and GO. The model induced in Cardiff is very different from that of Brussels, despite both units using conventional (as opposed to SPF-specific pathogen-free) housing and even with mice from the same supplier. We confirmed the induction of antibodies recognizing the native receptor by GI in the majority of BALB/c mice, in contrast to other investigators (21, 22, 23). However, we have not induced thyroiditis in any of the TSHR-treated mice.
Furthermore, our extensive series of experiments illustrates the qualitative heterogeneity of the TSHR response induced. For example, TSABs were induced in five of nine mice in the first experiment (including mice immunized with FP) and in none of 18 in the second, despite a robust humoral response to the receptor induced in 23 of 27 of the mice. This was accompanied by significantly higher T4 levels in group 1, compared with group 2. Both experiments were performed in the late autumn/early winter so that seasonal variation is not the explanation, although mice in group 1 were approximately 4 wk younger than group 2 at the start of immunization. Different batches of plasmid were used for the immunizations, but all were prepared using the same standard maxiprep kit. The different response induced using FP may be explained by a change in the adjuvant used (to single antigen pertussis toxin) because of unavailability of the acellular B. pertussis vaccine (which contains several antigenic components).
Our results indicate that care should be taken in interpreting changes in a TSHR-induced immune reaction attributed to, for example, an added cytokine, and that might simply be the consequence of the innate heterogeneity of the response.
We confirmed the previous study reporting that thyroiditis is a prerequisite in donor mice for successful transfer of disease to naive recipients (11).
We immunized 15 BALB/cbyJico mice using TSHR GI and the same using a TSHR FP of which 15 and 12, respectively, survived for analysis. From previous reported incidences of induced thyroiditis, we would have expected at least nine mice to have thyroiditis and three of these to display orbital changes akin to GO. None of the animals developed thyroiditis or GO. Part of the reason for this is the possible overestimation of thyroiditis in the previous studies, caused by the high level of ectopic thymus we found in the substrain from this supplier. In mammals the thymus and parathyroid are third pharyngeal pouch partners that would normally separate around d 14–15 gestation; failed separation can result in ectopic remains of thymus within the parathyroid (24). Previously published thyroid histology (Figs. 2 and 3 of Ref.11 ; Figs. 5 and 6 of Ref.16 ; and Fig. 7 of Ref.25) demonstrates "inflammatory clusters at the periphery of the thyroid," that bear all the hallmarks of ectopic thymus tissue being encapsulated with and adjacent to the parathyroid gland. It should be stressed that the references also state that cell-infiltrates were also found between the follicles, and these are clearly visible in the relevant figures. However, recent flow cytometry studies of normal whole-thyroid digests demonstrate significant amounts of resident immune cells (26). Such a quantifiable technique would differentiate more convincingly between populations of resident and infiltrating immune cells.
Our report illustrates, at high resolution, the wide range of possible appearances of ectopic thyroidal thymic tissue. Any experimental animal with peripheral focal thyroiditis should have its relationship to the parathyroid described to prevent misinterpretation of ectopic thymus for a pathological lymphocytic infiltration, particularly given the high incidence of thymic ectopy reported in BALB/c (and CBA/J) mice (26, 27, 28).
Similarly, ocular muscle anatomy is complex and easily disturbed if muscle contraction occurs before fixation (29). We obtained contraction artifacts, which could be misinterpreted as edema, in both immunized and control mice. Changes were not restricted to ocular muscles; careful examination of the strap muscles adjacent to thyroidectomy specimens show identical changes and have been illustrated previously but not commented on (Fig. 7C of Ref.14). Similar contraction bands in the myofibrils of ocular muscles are present in the published figures from both the BALB/c (Fig. 3, C and D, of Ref.12) and NMRI (Fig. 6, B and C, of Ref.14) ophthalmopathy models, and the outlined method indicates muscle dissection is performed before fixation. Inadequate fixation in these specimens may also have caused the loss of the characteristic myofibril light and dark bands and are unlikely to represent a pathological process, given these bands are conserved in human ophthalmopathy until the latest fibrotic stage. Such fixation artifacts could be avoided by perfusion fixation or excising the globe en bloc, leaving the muscles attached to the annulus of Zinn, splinted by the optic nerve, and then fixed before excising and embedding. We present our findings to raise awareness of an apparently common pitfall in modeled ophthalmopathy. Whereas there are convincing TSHR-induced models of thyroid hypertrophy (23, 30, 31), we still lack a convincing ophthalmopathy model in which macroscopic changes of proptosis and muscle hypertrophy occur. The only other model to show orbital fat hypertrophy are adiponectin transgenic FVB mice that are hypothyroid relative to other mouse strains (32, 33).
Other possible reasons for the absence of thyroiditis relate to differences in the environment of the mice in the present and previous studies. Our attempts to recreate Brussels could extend only as far as the chow, water, and bedding. Even with this level of modification, we observed some variation in the induced response. Antibodies to the TSHR were induced in all animals maintained in Brussels conditions, and this was accompanied by a significantly higher percentage of CD4+ cells expressing IL-4. However, this indicator of a generalized Th2 milieu did not extend to the antigen-specific T-cell response because significant IL-4 secretion was observed in six of nine mice in local conditions, compared with one of seven in Brussels. One potential difference between the two conditions is in the iodine intake. The local feed (rat and mouse (RM) no. 1 maintenance, SDS) has 1 mg/kg iodine, whereas the feed used in Brussels (RN-01, Carfil, Oud-Turnhout, Belgium) has 4.5 mg/kg. An enriched iodine diet (water containing 6.5 mg/liter) can transiently increase the amount of thyroid-associated ectopic thymic tissue in Wistar rats (34). However, we are not aware of the iodide concentrations in Cardiff and Brussels water, and to confirm the higher iodine intake of the Brussels maintenance regimen would require measurement of urinary iodine.
Although levels of IL-4 were low, they were measurable, unlike secretion of IFN, which remained largely below the limit of detection in all mice and in all conditions. This contrasts with the findings of Pichurin et al. (21) in which genetic immunization of BALB/c mice did not induce antibodies to the TSHR and splenocytes stimulated with a receptor preparation produced significant IFN but no IL-4. These experiments were performed in a SPF facility, unlike those of the present study and the previous reports from IRIBHM. When we compared the splenocyte profile of two BALB/cbyJico mice in which the genetic immunization was performed at IRIBHM with the equivalent from Cardiff, we observed no difference in the antigen-specific immune response. There were some minor differences, such as a decrease in the number of non-T cells in the former, but this is unlikely to account for the absence of thyroiditis in the latter.
In conclusion, the TSHR-induced animal model developed in Cardiff differs from that previously reported. We suggest that the microbial environment and iodine may have a significant impact in determining what we find to be a heterogenous, TSHR-induced immune response. Thymic ectopy needs to be excluded when focal thyroiditis is identified; more quantitative techniques should be employed in assessing thyroid infiltrates, and methods to prevent fixation artifacts in ocular muscles need to be followed.
Acknowledgments
We thank the Wales Office for Research and Development and The Leonardo di Capua Association for financial support. We also thank Dr. Sabine Costagliola (Institut de Recherche en Biologie Humaine et Moleculaire, Brussels, Belgium) for provision of activated T cells and constructive comments in the preparation of the manuscript. We also thank Dr. Phil Watson and Professor Tony Weetman (Northern General Hospital, Sheffield, United Kingdom) for the Gpi cell line and the panel of TSH receptor peptides. The continued support of BRAHMS in providing radiolabeled TSH is warmly appreciated.
References
Paschke R, Ludgate M 1997 The thyrotropin receptor in thyroid disease. N Engl J Med 337:1675–1681
Ludgate M, Baker GR 2002 Unlocking the immunological mechanisms of orbital inflammation in thyroid eye disease. Clin Exp Immunol 127:193–198
Prabhakar BS, Bahn RS, Smith TJ 2003 Current perspective on the pathogenesis of Graves’ disease and ophthalmopathy. Endocr Rev 24:802–835
Valyasevi R, Erickson D, Harteneck D, Dutton C, Heufelder A, Jyonouchi S, Bahn R 1999 Differentiation of human orbital preadipocyte fibroblasts induces expression of functional thyrotropin receptor. J Clin Endocrinol Metab 84:2557–2562
Crisp MS, Starkey KJ, Lane C, Ham J, Ludgate M 2000 Adipogenesis in thyroid eye disease. Invest Ophthalmol Vis Sci 41:3249–3255
Starkey K, Heufelder A, Baker G, Joba W, Evans M, Davies S, Ludgate M 2003 PPAR in thyroid eye disease: contra-indication for thiazolidenedione use? J Clin Endocrinol Metab 88:55–59
Starkey KJ, Janezic A, Jones G, Jordan N, Baker G, Ludgate M 2003 Adipose thyrotropin receptor expression is elevated in Graves’ and thyroid eye diseases ex vivo and indicates adipogenesis in progress in vivo. J Mol Endocrinol 30:369–380
Seetharamaiah GS 2003 Animal models of Graves’ hyperthyroidism. Autoimmunity 36:381–387
Ludgate M 2000 Animal models of Graves’ disease. Eur J Endocrinol 142:1–8
Ruwhof C, Drexhage HA 2001 Iodine and thyroid autoimmune disease in animal models. Thyroid 11:427–436
Costagliola S, Many MC, Stalmansfalys M, Vassart G, Ludgate M 1996 Transfer of thyroiditis, with syngeneic spleen-cells sensitized with the human thyrotropin receptor, to naive BALB/C and Nod mice. Endocrinology 137:4637–4643
Many MC, Costagliola S, Detrait M, Denef JF, Vassart G, Ludgate M 1999 Development of an animal model of autoimmune thyroid eye disease. J Immunol 162:4966–4974
Yamada M, Li AW, West KA, Chang CH, Wall JR 2002 Experimental model for ophthalmopathy in BALB/c and outbred (CD-1) mice genetically immunized with G2s and the thyrotropin receptor. Autoimmunity 35:403–413
Costagliola S, Many MC, Denef JF, Pohlenz J, Refetoff S, Vassart G 2000 Genetic immunization of outbred mice with thyrotropin receptor cDNA provides a model of Graves’ disease. J Clin Invest 105:803–811
Rundle FF, Wilson 1944 Development and course of exophthalmos and ophthalmoplegia in Graves’ disease with special reference to the effect of thyroidectomy. Clin Sci 5:177–194
Costagliola S, Many MC, Stalmansfalys M, Tonacchera M, Vassart G, Ludgate M 1994 Recombinant thyrotropin receptor and the induction of autoimmune thyroid-disease in Balb/C mice—a new animal-model. Endocrinology 135:2150–2159
Metcalfe R, Jordan N, Watson P, Gullu S, Wiltshire M, Crisp M, Evans C, Weetman A, Ludgate M 2002 Demonstration of IgG, IgA and IgE autoantibodies to the human thyrotropin receptor using flow cytometry. J Clin Endocrinol Metab 87:1754–1761
Young IT 1977 Proof without prejudice: use of the Kolmogarov-Smirnov test for the analysis of histograms from flow systems and other sources. J Histochem Cytochem 25:935–941
Jordan N, Rinderle C, Ashfield J, Morgenthaler N, Lazarus J, Ludgate M, Evans C 2001 A luminescent bioassay for thyroid blocking antibodies. Clin Endocrinol (Oxf) 54:355–364
Tandon N, Freeman MA, Weetman AP 1992 T-cell responses to synthetic TSH receptor peptides in Graves disease. Clin Exp Immunol 89:468–473
Pichurin P, Yan XM, Farilla L, Guo J, Chazenbalk GD, Rapoport B, McLachlan SM 2001 Naked TSH receptor DNA vaccination: a TH1 T cell response in which interferon- production, rather than antibody, dominates the immune response in mice. Endocrinology 142:3530–3536
Rao PV, Watson PF, Weetman AP, Carayanniotis G, Banga JP 2003 Contrasting activities of thyrotropin receptor antibodies in experimental models of Graves’ disease induced by injection of transfected fibroblasts or deoxyribonucleic acid vaccination. Endocrinology 144:260–266
Nagayama Y, Kita-Furuyama M, Ando T, Nakao K, Mizuguchi H, Hayakawa T, Eguchi K, Niwa M 2002 A novel murine model of Graves’ hyperthyroidism with intramuscular injection of adenovirus expressing the thyrotropin receptor. J Immunol 168:2789–2794
Cordier A, Haumont S 1980 Development of thymus, parathyroids, and ultimo-branchial bodies in NMRI and nude mice. Am J Anat 157:227–263
Costagliola S, Rodien P, Many MC, Ludgate M, Vassart G 1998 Genetic immunization against the human thyrotropin receptor causes thyroiditis and allows production of monoclonal antibodies recognising the native receptor. J Immunol 160:1458–1465
Caturegli P, Rose N, Kimura M, Kimura H, Tzou S 2003 Studies on murine thyroiditis: new insights from organ flow cytometry. Thyroid 13:419–426
Dunn TB, Moloney JB, Green AW, Arnold B 1961 Pathogenesis of a virus induced leukaemia in mice. J Natl Cancer Inst 26:189–221
Vladutiu A, Rose N 1972 Aberrant thymus tissue in rat and mouse thyroid. Experientia 28:79–81
Muhlendyck H, Syed A 1978 Fixation artefacts in the external eye muscles in biopsy examinations. (A light microscopy and electron microscopy study). Buch Augenarzt 73:181–191
Shimojo N, Kohno Y, Yamaguchi K-i, Kikuoka S-i, Hoshioka A, Niimi H, Hirai A, Tamura Y, Saito Y, Kohn LD, Tahara K 1996 Induction of Graves-like disease in mice by immunization with fibroblasts transfected with the thyrotropin receptor and a class II molecule. Proc Natl Acad Sci USA 93:11074–11079
Kaithamana S, Fan J, Osuga Y, Liang SG, Prabhakar BS 1999 Induction of experimental autoimmune Graves’ disease in BALB/c mice. J Immunol 163:5157–5164
Combs TP, Pajvani UB, Berg AH, Lin Y, Jelicks LA, Laplante M, Nawrocki AR, Rajala MW, Parlow AF, Cheeseboro L, Ding YY, Russell RG, Lindemann D, Hartley A, Baker GRC, Obici S, Deshaies Y, Ludgate M, Rossetti L, Scherer PE 2004 A transgenic mouse with a deletion in the collagenous domain of adiponectin displays elevated circulating adiponectin and improved insulin sensitivity. Endocrinology 145:367–383
Pohlenz J, Maqueem A, Cua K, Weiss R, Van Sande J, Refetoff S 1999 Improved radioimmunoassay for measurement of mouse thyrotropin in serum: strain differences in thyrotropin concentration and thyrotroph sensitivity to thyroid hormone. Thyroid 9:1256–1271
Mooij P, de Wit H, Drexhage H 1994 A high iodine intake in Wistar rats results in the development of a thyroid-associated ectopic thymic tissue and is accompanied by a low thyroid autoimmune reactivity. Immunology 81:309–316(Glynn Baker, Gherardo Maz)
Address all correspondence and requests for reprints to: Marian Ludgate, Centre for Endocrine and Diabetes Sciences, School of Medicine, Cardiff University, Heath Park, Cardiff CF14 4XN, United Kingdom. E-mail ludgate@cf.ac.uk
Abstract
We aimed to establish and extend the characterization of murine models of thyroiditis and Graves’ ophthalmopathy, induced by transfer of TSH receptor (TSHR) primed T cells. Experiments were performed in a different animal unit but using female BALB/cbyJico mice from the same supplier as previously. We report our findings together with a reevaluation of the earlier studies. In the first experiment, genetic immunization or TSHR fusion protein induced TSHR antibodies in all nine mice. Some of the antibodies functioned as thyroid-stimulating antibodies and/or TSH binding inhibiting Igs with two of seven mice having elevated T4. Thyroiditis and orbital changes were absent. Splenocyte transfer induced no immune response in naive BALB/cbyJico recipients. Subsequently genetic immunization or fusion protein-treated mice were maintained in either local or Brussels conditions (water, chow, and bedding). TSHR antibodies were induced in nine of nine Brussels (with decreased T4 in one of nine) but five of nine local mice. No thyroiditis or orbital changes were induced, but misleading fixation artefacts in extraocular muscles were noted. Nonspecific in vitro stimulation induced more CD-4+/IL-4+ cells in Brussels maintained. TSHR stimulation produced a significant increase in IL-4 secretion in six of nine local but one of seven Brussels mice. Thyroids from many TSHR-treated and control mice contained ectopic thymus. Our results confirm that thyroiditis is required for disease transfer but indicate the heterogeneity in TSHR-induced immune response in an inbred strain. Ectopic thymus can masquerade as thyroiditis, and care is required to avoid muscle artefacts. Because neither animal unit is pathogen free, microbial environment may contribute to determining TSHR-induced responses.
Introduction
GRAVES’ DISEASE (GD) is a common autoimmune condition in which thyroid-stimulating antibodies (TSABs) cause hyperthyroidism (reviewed in Ref.1). It is often accompanied by Graves’ ophthalmopathy (GO) in which expansion of the orbital contents, by hypertrophy and hyperplasia, produces proptosis (reviewed in Ref.2). The TSH receptor (TSHR) is an important autoantigen in both disorders, being the target of pathogenic autoantibodies in GD (reviewed in Ref.3) and promoting lymphocyte homing to the orbit in GO (reviewed in Ref.2). TSHR expression is increased during adipogenesis (4, 5), a mechanism causing expansion in the orbital volume, as illustrated by the exacerbation of GO in a person with type 2 diabetes treated with the proadipogenic drug, pioglitazone (6). We have demonstrated that adipogenesis is not restricted to the orbit but also occurs in the neck of GD patients but is not modified by thyroid status (7).
A central question in GD and GO is how tolerance to the TSHR is overcome. To address this, in the absence of a spontaneous model, a number of TSHR-induced models have been developed (reviewed in Refs.8, 9, 10). One of the approaches employed transfer of TSHR-primed T cells to naive syngeneic recipients (11). Priming could be achieved by either using a TSHR fusion protein (FP) or genetic immunization (GI) with a plasmid encoding the human TSHR. The studies used BALB/cbyJico mice from a French supplier and found that disease was more severe in females, as predicted. The TSHR antibodies produced in the model were mainly TSH binding inhibiting Igs (TBIIs) and thyroid-blocking antibodies (TBABs); TSABs were only rarely induced. One of the most significant attributes of the model was the development of Th2 thyroiditis in approximately 60% of the recipients, providing thyroiditis had been induced in some of the animals supplying the TSHR-primed T cells. Furthermore, about one third of the animals with thyroiditis also displayed orbital changes, including immune infiltration and edema, similar to those seen in GO (12). The severity of the orbital changes correlated positively with the extent of the Th2 skew in the thyroid immune infiltrate of that animal.
The model is unusual in being one of only three examples in which thyroiditis and orbital changes have been reported (12, 13, 14). One of the others was also developed in Brussels but uses GI of NMRI outbred mice to induce TSABs, increased thyroxine levels, thyroiditis, and orbital changes in approximately 20% of the animals (14). The higher disease rate of the transfer model makes it more attractive as a tool to increase our understanding of GO, a condition that remains a challenge for endocrinologists and ophthalmologists alike.
Our initial aim was to extend the characterization of the transfer model, including whether the disease follows a Rundle’s curve (15) as in humans. Subsequently we had planned to use the model to investigate alternative therapies for GO, e.g. peroxisomal proliferator-activated receptor- antagonists. In our initial experiments we used GI and FP injection of the BALB/cbyJico mice available from our local animal unit. Despite both protocols inducing a significant humoral response to the TSHR in the majority of animals treated, none of the mice developed thyroiditis. We also failed to transfer disease to naive syngeneic recipients using splenocytes from the GI- or FP-treated mice as a source of primed T cells (data not shown).
Consequently we attempted to reproduce the conditions resulting in the induction of thyroiditis and orbital changes reported previously (11, 12). This included studying the BALB/cbyJico substrain from the same French supplier. Our results differed significantly and led us to reevaluate the earlier studies in the light of the current experiments. We now report our findings obtained exclusively with the French BALB/cbyJico substrain.
Materials and Methods
Immunization with extracellular domain (ECD)/maltose binding protein (MBP) or pcDNA3-TSHR
All procedures were performed in accordance with the Animals (Scientific Procedures) Act 1986 and appropriate project and personal license approval.
Immunization (under halothane anesthesia) of 6-wk-old female BALB/cbyJico (Iffa Credo, Les Oncins, France) was performed on d 0 (with preimmune blood sampling) and at 3, 6, and 9 wk. The antigens employed were the pcDNA3-TSHR plasmid, prepared using a standard maxiprep kit (Qiagen, Santa Clarita, CA) according to the manufacturer’s instructions, and an FP comprising the ECD of the TSHR fused to MBP as previously described (16). GI mice received 100 μg pcDNA3-TSHR in 25% sucrose shared equally between both anterior tibialis muscles at each time point. FP-treated animals had an initial 100 μg FP ip in a 200 μl volume of 0.9% saline with adjuvant, 0.1% aluminum oxide, 0.4% magnesium hydroxide (Maalox, Rhone-Poulenc Rorer, Paris, France), and 50 μl acellular Bordetella pertussis vaccine (Wyeth, Madison, NJ). Thereafter immunizations used 50 μg FP with adjuvant. Mice were killed after 14 wk; their serum stored; and spleens, thyroids, and orbital contents removed. Thyroid and orbital tissues were fixed in 10% phosphate-buffered formalin with 2% acetic acid for a minimum of 48 h for paraffin sections using standard techniques. The thyroids were stained with hemotoxylin and eosin and orbital tissue with periodic acid and Schiff’s reagent counterstained with hematoxylin.
T-cell transfer to naive recipients
Spleens removed from immunized and control animals were mechanically disrupted and frozen in 10% dimethylsulfoxide in fetal calf serum. Forty-eight hours before transfer, the frozen aliquots were rapidly thawed and viable splenocytes obtained from the opaque interface of a Ficoll-Hypaque gradient. After washing, in vivo primed and nonprimed cells underwent in vitro priming with FP, all as previously described (11).
Naive 9-wk mice were immunized via the tail veins, with 5 x 105 splenocytes in 100 μl PBS. Groups of seven each male and female BALB/cbyJico (Iffa Credo) mice, received FP or pcDNA3-TSHR primed or nonprimed cells. The latter came from two male and two female 9-wk BALB/cbyJico animals. Mice were killed at 10 wk, excluding six males and females that were reimmunized at 15 wk with 5 x 106 FP or pcDNA3-TSHR-primed cells and killed 12 wk later.
In all cases sera were stored and orbital tissues plus thyroids (one lobe each) were either fixed overnight in 0.2% glutaraldehyde or 4% formaldehyde for plastic sections or snap frozen for cryostat sectioning.
Environmental modification
A second group of 20 imported BALB/cbyJico female mice were initially all maintained on standard local water, feed (rat and mouse no. 1 maintenance, SDS Ltd., Witham, UK), and bedding. The mice were then immunized, 10 with pcDNA3-TSHR and 10 with FP [0.2 μg pertussis toxin (Sigma, St. Louis, MO) replaced acellular B. pertussis]. At immunization the two groups were subdivided; half continued on standard maintenance, whereas the other received imported water, bedding, and feed (RN-01–20K12, Carfil Nutrients) as used by Institut de Recherche en Biologie Humaine et Moleculaire (IRIBHM), Brussels. The mice were killed after 14 wk, their sera stored, splenocytes frozen, and thyroids plus orbital contents prepared for plastic sectioning.
Antibodies to the TSHR
ELISA using ECD-MBP.
Ninety-six-well plates were coated overnight at 4 C with 100 μl/well of FP (10 μg/ml) in carbonate/bicarbonate coating buffer. Sera were tested at 1:100 by incubation for 2 h at room temperature followed by 100 μl of an antimouse IgG horseradish peroxidase conjugate for 30 min at room temperature. Detection used 100 μl of substrate, and the OD was measured in a plate reader at 410 nm.
Flow cytometry using Chinese hamster ovary cells expressing human TSHR
We used Chinese hamster ovary cells expressing the TSHR-ECD anchored by a glycophosphatidylinositol link (17). The cells were maintained in Ham’s F12 medium supplemented with 10% fetal calf serum and antibiotics until 80–90% confluent and then detached with 5 mM EGTA/EDTA in PBS. They were resuspended in PBS/BSA 0.1% and 100-μl aliquots containing 2 x 105 cells incubated with 2 μl mouse serum for 1 h at room temperature. They were incubated at 4 C in the dark with fluorescein isothiocyanate (FITC)-conjugated F(ab)2 fragment of goat antimouse Igs (diluted 1:16, Dako, Carpinteria, CA) and counterstained with propidium iodide to gate out damaged cells. Fluorescence from 5000 cells was measured using a FACScan flow cytometer (Becton Dickinson, Lincoln Park, NJ) and compared with preimmune sera, also at 1:50 dilution. Histogram shifts were examined using the Kolmogorov-Smirnov test (18). This compares the cumulative distribution of the data with the expected cumulative Gaussian distribution and bases its value simply on the largest discrepancy-Dmax (Cellquest Pro software, Becton Dickinson). Cumulative FITC fluorescence curves were automatically plotted with test sera vs. preimmune to calculate a D value. The 97th centile of the D values from recipients of nonprimed T cells (n = 7) after transfer (19 wk old) was 0.3986; values above this were considered positive (0.4).
TBIIs
Glycophosphatidylinositol cells were plated in 96-well plates at 2 x 104 cells/well. The next day, 100 μl salt-free Hanks’ balanced salt solution was added to 100 μl I125-labeled bovine TSH tracer (kindly provided by BRAHMS Diagnostica, Berlin, Germany) and 3 μl serum (triplicates). The nonspecific binding was determined by adding 100 mU bovine TSH. After a 3-h incubation at room temperature, the cells were washed, lysed with 100 μl 1 N NaOH (30 min, 4 C), transferred to counting tubes, and radioactivity measured in a -counter. Results are expressed as a percentage of the counts per minute bound by the test sera relative to the preimmune value of 100% tracer binding after subtracting the nonspecific binding. The TBII indices were then plotted against 97th and third centiles constructed from individual nonprimed and preimmune sera (n = 9).
TSABs
TSABs were assayed using LULU* cells, which express both the human TSHR and a cAMP-responsive luciferase reporter construct, as previously described (19). Briefly, the cells were plated in Ham’s F12 medium supplemented with 10% charcoal-stripped serum and then incubated the next day for 4 h at 37 C in 50 μl serum-free Ham’s F12 per well, containing 5 μl of test serum (triplicates). After washes, plates were air dried and lysed by 20 min shaking in 50 μl lysis buffer (Promega, Madison, WI) per well. Luciferase assay reagent (Promega) was prepared according to the manufacturer’s instructions, added to the wells, and light output read using a luminometer (Tropix Microplate luminometer, Applied Biosystems, Foster City, CA). The output of relative light units from test sera was compared with preimmune sera to give a stimulation index (SI). Test sera were considered positive when greater than the 97th centile of the preimmune and nonprimed recipients (n = 9); this was calculated as SI greater than 1.8.
Measurement of serum T4
Total serum-free T4 was measured on duplicate 10-μl volumes of serum, using a commercially available competitive RIA (Amerlex, Amersham International, Amersham, UK) according to the manufacturer’s instructions. Results are reported as micrograms per deciliter.
Splenocyte stimulation using TSHR or phorbol-12-myristate-13-acetate (PMA)/ionomycin
Splenocytes from the environment modification group and two BALB/cbyJico mice in which the GI had been performed at IRIBHM, were thawed, resuspended in 5 ml RF10, and incubated overnight at 37 C. Trypan blue exclusion confirmed viability greater than 85%, the splenocytes were resuspended at 2 x 106 cells/ml in RF10 for polyclonal and antigen-specific stimulation. Splenocyte aliquots of 1 x 106 cells were prepared for intracytoplasmic cytokine staining (ICCS) and parallel measurements of cytokines released into the culture supernatant.
For ICCS, the control sample was incubated for 4 h at 37 C with monensin (7 μM) and the stimulated sample with PMA (25 ng/ml), ionomycin (1 μg/ml), and monensin (7 μM). Control and stimulated splenocytes were then washed and resuspended in 1 ml PBS/BSA. Surface antigen detection used 10 μl each of peridinin chlorophyll-a protein-conjugated rat antimouse CD4 (L3T4) and allophycocyanin-conjugated rat antimouse CD8a (Ly-2) monoclonal antibodies (PharMingen, San Diego, CA). After 30 min incubation at room temperature in the dark, cells were fixed and permeabilized using a kit (CPK-200, cell permeabilization kit, ImmunologicalsDirect.com, Kidlington, UK) according to the manufacturer’s instructions. Simultaneously 10 μl anticytokine monoclonal antibodies (PharMingen) were added, r-phycoerythrin (PE)-conjugated rat antimouse IL-4 and FITC-conjugated rat antimouse interferon (IFN)-, and incubated at room temperature in the dark for 30 min. Then 400 μl of 2% paraformaldehyde was finally added and after a minimum of 10 min either analyzed or kept overnight at 4 C. The splenocytes were analyzed using four-color flow cytometry to count 105 events (FACSCaliber, Becton Dickinson). Once gated according to surface markers, the percentage of cells expressing IL-4 and IFN from control and stimulated splenocytes was measured using arbitrary quadrants. The figure recorded was the percentage of cells expressing cytokines in the stimulated sample minus the unstimulated; differences were examined for significance between groups.
For supernatant cytokine measurements the cells were incubated as for ICCS but without monensin, centrifuged (400 x g, 5 min), supernatants aspirated, and stored at –80 C until assayed for cytokines.
In addition, we used two separate sources of TSHR, the ECD/MBP FP and also a pooled panel of synthetic TSHR peptides spanning the ECD (20). Ninety-six-well plates were loaded with 250 μl of splenocyte suspension (5 x 105 cells) in RF10 and incubated (in triplicate) at 37 C for 6 d with 10 μg/ml FP or 5 μg/ml pooled TSHR peptides (dissolved in dimethylsulfoxide). Supernatant replicates were pooled and stored at –80 C until assayed. Cytokines secreted from splenocytes stimulated by either polyclonal or specific stimuli were measured by commercial murine IFN and IL-4 ELISA (Duoset, R&D Systems, Minneapolis, MN) according to the manufacturer’s instructions.
Standard curves were plotted to calculate the concentration of cytokines in picograms per milliliter of sample supernatants. The value recorded is the amount of cytokine present in the stimulated sample minus that of the unstimulated sample.
Statistical analysis
Data were analyzed using software, SPSS for Windows (version 11.0.1, SPSS, Chicago, IL) or the DOS-based INSTAT (version 2.0, GraphPad Software, San Diego, CA). Categorical data were examined using 2 with Fisher’s exact test where required and continuous data using t tests, ANOVA, or Kruskal-Wallis where appropriate. Homogeneity of variance was assessed by Levene’s (SPSS) or Bartlett’s (INSTAT) statistics and post hoc testing mainly by Tukey honestly significantly different. Test values are reported with P values and test type.
Results
A strong humoral response to the TSHR is induced using plasmid or fusion protein
In the first group of female BALB/cbyJico mice, a strong humoral response was induced to the TSHR using either GI or injection of FP. All of the mice developed an IgG antibody response to the FP measured by ELISA. The FP animals developed a rapid response with antibodies detected as early as 2 wk and peaking around 6 wk. Their response is potentially to both components of the fusion protein (ECD and MBP), unlike that of the more slowly evolving antibody response seen in the pcDNA3-TSHR immunized group (Fig. 1).
FIG. 1. ELISA from the first immunized group showing immune response to ECD/MBP (GI, solid line; FP immunized, dotted line, and 97th centile of preimmune sera, fine dotted line).
The sera were then tested by flow cytometry, a method that detects antibodies predominantly to conformational epitopes of the TSHR. Representative histograms are shown in Fig. 2, and for comparison, a typical plot obtained using sera from a similarly aged female, pre- and post receipt of nonprimed T cells. All nine GI or FP-treated mice from group 1 demonstrated D values above the 97th centile of the nonprimed T-cell recipient control group (n = 7), as shown in Fig. 3.
FIG. 2. Representative anti-TSHR IgG flow histograms from immunized group 1. Outlined histograms represent preimmune sera and solid histograms sera at 14 wk. A and B, Two GI. C and D, Two FP-treated mice. E, Outlined histogram from a female mouse before and solid histogram 10 wk after transfer of nonprimed (NP) T cells.
FIG. 3. IgG class antibodies to the TSHR measured by flow cytometry and presented as D values for individual mice from GI and FP-immunized groups 1 and 2. The third and 97th centiles (fine dotted line) were calculated from recipients of nonprimed T cells 10 wk after transfer (n = 7).
When the biological activity of the TSHR antibodies was evaluated, using an assay for TBIIs (Fig. 4), three of five GI and one of four FPs were positive (falling below the third centile of the nonprimed T-cell recipient control group, n = 7).
FIG. 4. TBII activity in immunized groups 1 and 2. The third and 97th centiles (fine dotted line) were calculated from recipients of nonprimed T cells 10 wk after transfer (n = 7).
In a luminescent bioassay for TSABs, one of five of the GIs and one of four of the FPs were strongly positive with a further two of five GIs and one of four FPs having an SI greater than 1.8 and consequently positive for TSAB (Fig. 5).
FIG. 5. TSAB activity (SI) in immunized groups 1 and 2. The 97th centile (SI = 1.8, fine dotted line) was calculated from recipients of nonprimed T cells 10 wk after transfer (n = 7).
The response is associated with elevated T4 levels but not thyroiditis or orbital changes
We used preimmune sera from these mice and also sera from similarly aged female recipients of nonprimed T cells to obtain the normal range of T4 values (n = 9). This was very broad, 2.57 (third centile) and 8.91 (97th centile). Two of four of the GI mice tested had values higher than this range; none of the animals had T4 below the third centile. The results are shown in Fig. 6.
FIG. 6. Total T4 levels for individual mice from GI and FP-immunized groups 1 and 2. The third and 97th centiles (dotted lines) were calculated from recipients of nonprimed T cells 10 wk after transfer (n = 7).
None of the receptor-immunized mice displayed thyroiditis or orbital changes (described in more detail below).
Limited transfer of TSHR immune responsiveness to naive recipients
Splenocytes from group 1 were used as a source of primed T cells for transfer to naive French female BALB/cbyJico recipients. Using flow cytometry, none of the recipients developed antibodies to the TSHR. In contrast ,by ELISA, six of 14 mice demonstrated a weak response, three males and three females and in all but one, priming was with the FP so that antibodies to MBP would also be detected. When using bioassays for the TSHR antibodies, all mice were negative for TSAB and TBII activity. Two male recipients of FP-primed T cells displayed T4 levels above the 97th centile (none of these data shown). None of the animals developed thyroiditis or changes in their orbital tissues.
Does environmental modification change the TSHR-induced response?
The absence of induced thyroiditis or orbital changes (described below) combined with the minimal humoral reactivity obtained indicate no transfer of immune response to the TSHR. Because we were using mice identical with those used previously for TSHR-induced thyroiditis and transfer of disease; we investigated whether modifying the environment might alter the outcome. A second group of 20 French female BALB/cbyJico mice received either GI or treatment with the FP. Half were maintained on standard chow/bedding/water provided by the local animal unit. The other half were provided with chow, bedding, and water identical with that used at IRIBHM in Brussels. Neither the local nor Brussels animal units are pathogen free. All 10 of the GI and eight of 10 of the FP-treated mice survived and were available for analysis.
The plasmid/protein-induced humoral response is modified by environment
Using flow cytometry, all of the FP-treated mice had a strong immune response to the TSHR that developed by 9 wk. In the GI group, all of the mice in Brussels conditions (GIB) but only one of five in local conditions (GIL) had developed TSHR antibodies (Fig. 3).
None of the mice had TSAB or TBII activities (Figs. 4 and 5)
Significantly lower average T4 was found in group 2, compared with group 1, immunized mice (Fig. 6), 8.64 vs. 3.67 μg/dl (P < 0.0001, Mann-Whitney U). In group 2, 16 of 17 tested T4 levels fell within the broad normal range, with the exception of a FP-treated mouse in Brussels conditions (FPB) that had a value below the third centile.
The T-cell response is altered by immunization and modified by environment
Splenic T-cell profiles were examined using flow cytometry. As shown in Table 1, there is a higher percentage of CD4+ cells in the spleens of FP-treated animals, compared with genetic or nonimmunized mice, with a significant difference between FPB and GIL (P = 0.048). All immunization protocols induce a significant reduction in the percentage of CD8+ cells and consequently an increase in the CD4+ to CD8+ ratio. GI produced an increase in the ratio of non-T cell to T cell, and this was significant for mice in local conditions, compared with all other groups. Changes were due predominantly to the immunization method with modification of the environment having only a minimal effect on the proportions of CD4+, CD8+, or non-T splenocytes.
TABLE 1. Mean percentages (and SEM) of CD4+ and CD8+ cells and CD4+:CD8+ and non-T:T ratios in the splenocyte populations of the second environment-modified immunized group
Use of the nonspecific stimulus, PMA, and ionomycin produces a snapshot of the T-cell repertoire after treatment. The percentages of CD4+ and CD8+ cells expressing IL-4 and IFN were low in all mice studied. Percentages of CD4+/IL-4, CD4+/IFN, and CD8+/IL-4 cells were higher in FP-treated mice than in GI with Brussels maintained having the larger value in each case, as shown in Fig. 7. The difference between percentages of CD4+/IL4+ cells in FPB and mice treated with FP maintained in local conditions (FPL) were significantly different (P = 0.042) but did not reach significance for CD8+/IL4+ or CD4+/IFN. Percentages of CD8+/IFN-expressing cells were similar in all groups.
FIG. 7. Summary of the percentages of CD4+ and CD8+ splenocyte populations stimulated with PMA/ionomycin to express IFN and Il-4 analyzed by four-color flow cytometry after intracellular cytokine staining. The plots show the median (heavy horizontal line), interquartile range (boxed area), and extreme values (bars). BRUS, Mice treated by GI performed in Brussels; NORM, nontreated mice.
We then investigated antigen-specific responses by measuring cytokines in the supernatants of splenocytes incubated with a pooled panel of TSHR peptides or FP. Levels of IL-4 secretion were detectable but low. Compared with normal splenocytes, TSHR (FP and/or peptides)-specific IL-4 secretion was significantly increased in four of four FPL (range 4.7–25.6 pg/ml) and two of five GIL (5.9–6.8 pg/ml), compared with one of three FPB (12.2 pg/ml) and none of four GIB (<2.6 pg/ml) in Brussels conditions.
Finally, we examined the splenocyte profiles in two BALB/cbyJico mice in which the GI had been performed at IRIBHM. These mice displayed a similar proportion of CD4+ and CD8+ but significantly fewer non-T cells than GI mice-treated in Cardiff (Table 1) and the lowest percentage of CD4+ cells expressing IL4+ (Fig. 7). When assessing their antigen-specific responses, we obtained similar results to those of the GIB-treated mice, with IFN levels below the limit of detection and no significant increase in IL-4 secretion.
Immunized, transfer, and control mice demonstrated normal thyroid pathology and often contained ectopic thymus
Mice from immunized, transfer, and control groups had normal thyroid macroscopic appearances with no goiter formation or hypertrophy. Normal thyroid histology was seen in all three groups and comprised heterogeneity of follicular size and thyrocyte depth with occasional apical budding and colloid endocytosis (Fig. 8, A and B). Mast cells were resident in connective tissue adjacent to and within thyroids, with no greater frequency between immunized and control mice. Additionally, normal thyroids contained parathyroid glands and cysts that were seen frequently associated with highly organized clusters of encapsulated densely stained large nucleated cells (Fig. 8, C, F, I, and J). In further sections these clusters of cells had less nuclear staining (Fig. 8, F and J) and contained specialized cells, murine equivalents of Hassal bodies (Fig. 8L), identifying them as ectopic medullary thymus and the former dense staining cells as cortical thymus. The appearances were by no means uniform; the ectopic thymus could have a large extrathyroidal location (Fig. 8, G and H) but most commonly was located in the periphery of the thyroid gland. At times, ectopic thymus could appear as a lone cluster of cells (Fig. 8, D and E) but if thoroughly sectioned was always partnered by the parathyroid gland. TSHR immunization did not alter the apparent frequency of ectopic thymus and was present in 13 of 26 immunized mice, compared with six of nine control nonimmunized mice (P = 0.46, Fisher’s exact). However, routine sectioning sometimes missed the two parathyroid glands, and this might have led us to underreport the incidence of the congenital ectopic thymic anomaly.
FIG. 8. Normal follicular heterogeneity was present in mice from control, immunized, and transfer animals with occasional areas of active endocytosis (A and B; arrow in B shows endocytosis). Parathyroid glands (P) had a characteristic mosaic of dark, intermediate, and clear staining cells frequently containing ciliated cysts (C). Ectopic cortical thymus (T) could appear to be a lone cluster of inflammatory cells in the periphery of the thyroid (D and E). Medullary thymus contained cells with less staining than the cortex with a lymphoblastoid appearance (F). The range of appearances of ectopic thymus was wide, examples of large extrathyroidal locations were seen (G and H), and the usual clear separation of parathyroid from ectopic thymus was not always present (I). Small fragments of medullary thymus could be difficult to define (J) but had additional histological features of thymocytes clustering around venules (V in K) and murine equivalents of Hassall bodies (H in L). Semithin sections mounted in LR white, stained with toluidine blue. Scale bars A to K, 100 μm; L, 50 μm.
Orbital connective tissue and fat develop no inflammatory changes: different ocular muscle types and contraction artefact cause diagnostic difficulty
No significant macroscopic changes akin to thyroid eye disease were present in terms of proptosis, orbital fat expansion, or extraocular muscle hypertrophy.
Control and immunized ocular muscle in cross-sections had areas in which connective tissue was common between muscle fibers (Fig. 9A); this corresponded to the singly innervated orbital muscle layers without edema or accumulation of periodic acid and Schiff’s-positive material to suggest glycosaminoglycan deposition. Ocular muscle fibers with a closely packed appearance were the multiply innervated global muscle fibers (Fig. 9B). Both orbital and global layer muscle fibers had areas of contracted, buckled myofibrils, resulting from disinserting unfixed muscle from the globe. Contraction-induced buckling deformed the connective tissue, particularly around orbital fibers, and could give rise to a false impression of edema. Sections of myofibrils could appear normal adjacent to segments with contraction bands and knots (Fig. 9C). These contraction artefacts were not present in later control samples fixed with the muscles left inserted to the globe and attached to their origins around the optic nerve at the annulus of Zinn, which splinted the muscles and prevented their contracture (Fig. 9D).
FIG. 9. Ocular muscle from control and immunized mice comprised two fiber types histologically. Orbital fibers (MF) were clustered with a few fibers surrounded by connective tissue and were interspersed with large myelinated nerve fibers (N in A). Global muscle fibers were clustered in large numbers with less connective tissue and nerve fibers (B). Postmortem muscle contraction artifacts of knots (K) and contraction bands (CB) occurred adjacent to normal segments of muscle (N), resultant disruption of connective tissue could be misinterpreted as edema (C). Fixation before disinsertion from the muscle attachments removed such fixation artifacts (D). Occasional multiloculated adipocytes appeared within the connective tissues of both control and immunized mice (E). Orbital fat had a normal lobular organization with mast cells (MC) in the adjacent connective tissues without hypertrophy in the immunized groups (F). None of the orbital tissues showed inflammatory changes. Semithin sections mounted in LR white stained with toluidine blue. Scale bars, 50 μm.
Orbital fat appearances were similar in control and immunized animals. Organized lobules of fat were bordered by connective tissue, within which scatterings of mast cells were resident (Fig. 9F). Sporadic multivacuolated adipocytes were found within areas of connective tissue in control and immunized mice representing the normal background of adipogenesis (Fig. 9E). No lymphocytic infiltration was present in the muscles, fat, and connective tissues of the orbits of any immunized or control mice.
Discussion
Our initial aim was to establish a previously reported TSHR-induced animal model, using the BALB/cbyJico strain, and apply it to answer questions pertinent to the pathogenesis of GD and GO. The model induced in Cardiff is very different from that of Brussels, despite both units using conventional (as opposed to SPF-specific pathogen-free) housing and even with mice from the same supplier. We confirmed the induction of antibodies recognizing the native receptor by GI in the majority of BALB/c mice, in contrast to other investigators (21, 22, 23). However, we have not induced thyroiditis in any of the TSHR-treated mice.
Furthermore, our extensive series of experiments illustrates the qualitative heterogeneity of the TSHR response induced. For example, TSABs were induced in five of nine mice in the first experiment (including mice immunized with FP) and in none of 18 in the second, despite a robust humoral response to the receptor induced in 23 of 27 of the mice. This was accompanied by significantly higher T4 levels in group 1, compared with group 2. Both experiments were performed in the late autumn/early winter so that seasonal variation is not the explanation, although mice in group 1 were approximately 4 wk younger than group 2 at the start of immunization. Different batches of plasmid were used for the immunizations, but all were prepared using the same standard maxiprep kit. The different response induced using FP may be explained by a change in the adjuvant used (to single antigen pertussis toxin) because of unavailability of the acellular B. pertussis vaccine (which contains several antigenic components).
Our results indicate that care should be taken in interpreting changes in a TSHR-induced immune reaction attributed to, for example, an added cytokine, and that might simply be the consequence of the innate heterogeneity of the response.
We confirmed the previous study reporting that thyroiditis is a prerequisite in donor mice for successful transfer of disease to naive recipients (11).
We immunized 15 BALB/cbyJico mice using TSHR GI and the same using a TSHR FP of which 15 and 12, respectively, survived for analysis. From previous reported incidences of induced thyroiditis, we would have expected at least nine mice to have thyroiditis and three of these to display orbital changes akin to GO. None of the animals developed thyroiditis or GO. Part of the reason for this is the possible overestimation of thyroiditis in the previous studies, caused by the high level of ectopic thymus we found in the substrain from this supplier. In mammals the thymus and parathyroid are third pharyngeal pouch partners that would normally separate around d 14–15 gestation; failed separation can result in ectopic remains of thymus within the parathyroid (24). Previously published thyroid histology (Figs. 2 and 3 of Ref.11 ; Figs. 5 and 6 of Ref.16 ; and Fig. 7 of Ref.25) demonstrates "inflammatory clusters at the periphery of the thyroid," that bear all the hallmarks of ectopic thymus tissue being encapsulated with and adjacent to the parathyroid gland. It should be stressed that the references also state that cell-infiltrates were also found between the follicles, and these are clearly visible in the relevant figures. However, recent flow cytometry studies of normal whole-thyroid digests demonstrate significant amounts of resident immune cells (26). Such a quantifiable technique would differentiate more convincingly between populations of resident and infiltrating immune cells.
Our report illustrates, at high resolution, the wide range of possible appearances of ectopic thyroidal thymic tissue. Any experimental animal with peripheral focal thyroiditis should have its relationship to the parathyroid described to prevent misinterpretation of ectopic thymus for a pathological lymphocytic infiltration, particularly given the high incidence of thymic ectopy reported in BALB/c (and CBA/J) mice (26, 27, 28).
Similarly, ocular muscle anatomy is complex and easily disturbed if muscle contraction occurs before fixation (29). We obtained contraction artifacts, which could be misinterpreted as edema, in both immunized and control mice. Changes were not restricted to ocular muscles; careful examination of the strap muscles adjacent to thyroidectomy specimens show identical changes and have been illustrated previously but not commented on (Fig. 7C of Ref.14). Similar contraction bands in the myofibrils of ocular muscles are present in the published figures from both the BALB/c (Fig. 3, C and D, of Ref.12) and NMRI (Fig. 6, B and C, of Ref.14) ophthalmopathy models, and the outlined method indicates muscle dissection is performed before fixation. Inadequate fixation in these specimens may also have caused the loss of the characteristic myofibril light and dark bands and are unlikely to represent a pathological process, given these bands are conserved in human ophthalmopathy until the latest fibrotic stage. Such fixation artifacts could be avoided by perfusion fixation or excising the globe en bloc, leaving the muscles attached to the annulus of Zinn, splinted by the optic nerve, and then fixed before excising and embedding. We present our findings to raise awareness of an apparently common pitfall in modeled ophthalmopathy. Whereas there are convincing TSHR-induced models of thyroid hypertrophy (23, 30, 31), we still lack a convincing ophthalmopathy model in which macroscopic changes of proptosis and muscle hypertrophy occur. The only other model to show orbital fat hypertrophy are adiponectin transgenic FVB mice that are hypothyroid relative to other mouse strains (32, 33).
Other possible reasons for the absence of thyroiditis relate to differences in the environment of the mice in the present and previous studies. Our attempts to recreate Brussels could extend only as far as the chow, water, and bedding. Even with this level of modification, we observed some variation in the induced response. Antibodies to the TSHR were induced in all animals maintained in Brussels conditions, and this was accompanied by a significantly higher percentage of CD4+ cells expressing IL-4. However, this indicator of a generalized Th2 milieu did not extend to the antigen-specific T-cell response because significant IL-4 secretion was observed in six of nine mice in local conditions, compared with one of seven in Brussels. One potential difference between the two conditions is in the iodine intake. The local feed (rat and mouse (RM) no. 1 maintenance, SDS) has 1 mg/kg iodine, whereas the feed used in Brussels (RN-01, Carfil, Oud-Turnhout, Belgium) has 4.5 mg/kg. An enriched iodine diet (water containing 6.5 mg/liter) can transiently increase the amount of thyroid-associated ectopic thymic tissue in Wistar rats (34). However, we are not aware of the iodide concentrations in Cardiff and Brussels water, and to confirm the higher iodine intake of the Brussels maintenance regimen would require measurement of urinary iodine.
Although levels of IL-4 were low, they were measurable, unlike secretion of IFN, which remained largely below the limit of detection in all mice and in all conditions. This contrasts with the findings of Pichurin et al. (21) in which genetic immunization of BALB/c mice did not induce antibodies to the TSHR and splenocytes stimulated with a receptor preparation produced significant IFN but no IL-4. These experiments were performed in a SPF facility, unlike those of the present study and the previous reports from IRIBHM. When we compared the splenocyte profile of two BALB/cbyJico mice in which the genetic immunization was performed at IRIBHM with the equivalent from Cardiff, we observed no difference in the antigen-specific immune response. There were some minor differences, such as a decrease in the number of non-T cells in the former, but this is unlikely to account for the absence of thyroiditis in the latter.
In conclusion, the TSHR-induced animal model developed in Cardiff differs from that previously reported. We suggest that the microbial environment and iodine may have a significant impact in determining what we find to be a heterogenous, TSHR-induced immune response. Thymic ectopy needs to be excluded when focal thyroiditis is identified; more quantitative techniques should be employed in assessing thyroid infiltrates, and methods to prevent fixation artifacts in ocular muscles need to be followed.
Acknowledgments
We thank the Wales Office for Research and Development and The Leonardo di Capua Association for financial support. We also thank Dr. Sabine Costagliola (Institut de Recherche en Biologie Humaine et Moleculaire, Brussels, Belgium) for provision of activated T cells and constructive comments in the preparation of the manuscript. We also thank Dr. Phil Watson and Professor Tony Weetman (Northern General Hospital, Sheffield, United Kingdom) for the Gpi cell line and the panel of TSH receptor peptides. The continued support of BRAHMS in providing radiolabeled TSH is warmly appreciated.
References
Paschke R, Ludgate M 1997 The thyrotropin receptor in thyroid disease. N Engl J Med 337:1675–1681
Ludgate M, Baker GR 2002 Unlocking the immunological mechanisms of orbital inflammation in thyroid eye disease. Clin Exp Immunol 127:193–198
Prabhakar BS, Bahn RS, Smith TJ 2003 Current perspective on the pathogenesis of Graves’ disease and ophthalmopathy. Endocr Rev 24:802–835
Valyasevi R, Erickson D, Harteneck D, Dutton C, Heufelder A, Jyonouchi S, Bahn R 1999 Differentiation of human orbital preadipocyte fibroblasts induces expression of functional thyrotropin receptor. J Clin Endocrinol Metab 84:2557–2562
Crisp MS, Starkey KJ, Lane C, Ham J, Ludgate M 2000 Adipogenesis in thyroid eye disease. Invest Ophthalmol Vis Sci 41:3249–3255
Starkey K, Heufelder A, Baker G, Joba W, Evans M, Davies S, Ludgate M 2003 PPAR in thyroid eye disease: contra-indication for thiazolidenedione use? J Clin Endocrinol Metab 88:55–59
Starkey KJ, Janezic A, Jones G, Jordan N, Baker G, Ludgate M 2003 Adipose thyrotropin receptor expression is elevated in Graves’ and thyroid eye diseases ex vivo and indicates adipogenesis in progress in vivo. J Mol Endocrinol 30:369–380
Seetharamaiah GS 2003 Animal models of Graves’ hyperthyroidism. Autoimmunity 36:381–387
Ludgate M 2000 Animal models of Graves’ disease. Eur J Endocrinol 142:1–8
Ruwhof C, Drexhage HA 2001 Iodine and thyroid autoimmune disease in animal models. Thyroid 11:427–436
Costagliola S, Many MC, Stalmansfalys M, Vassart G, Ludgate M 1996 Transfer of thyroiditis, with syngeneic spleen-cells sensitized with the human thyrotropin receptor, to naive BALB/C and Nod mice. Endocrinology 137:4637–4643
Many MC, Costagliola S, Detrait M, Denef JF, Vassart G, Ludgate M 1999 Development of an animal model of autoimmune thyroid eye disease. J Immunol 162:4966–4974
Yamada M, Li AW, West KA, Chang CH, Wall JR 2002 Experimental model for ophthalmopathy in BALB/c and outbred (CD-1) mice genetically immunized with G2s and the thyrotropin receptor. Autoimmunity 35:403–413
Costagliola S, Many MC, Denef JF, Pohlenz J, Refetoff S, Vassart G 2000 Genetic immunization of outbred mice with thyrotropin receptor cDNA provides a model of Graves’ disease. J Clin Invest 105:803–811
Rundle FF, Wilson 1944 Development and course of exophthalmos and ophthalmoplegia in Graves’ disease with special reference to the effect of thyroidectomy. Clin Sci 5:177–194
Costagliola S, Many MC, Stalmansfalys M, Tonacchera M, Vassart G, Ludgate M 1994 Recombinant thyrotropin receptor and the induction of autoimmune thyroid-disease in Balb/C mice—a new animal-model. Endocrinology 135:2150–2159
Metcalfe R, Jordan N, Watson P, Gullu S, Wiltshire M, Crisp M, Evans C, Weetman A, Ludgate M 2002 Demonstration of IgG, IgA and IgE autoantibodies to the human thyrotropin receptor using flow cytometry. J Clin Endocrinol Metab 87:1754–1761
Young IT 1977 Proof without prejudice: use of the Kolmogarov-Smirnov test for the analysis of histograms from flow systems and other sources. J Histochem Cytochem 25:935–941
Jordan N, Rinderle C, Ashfield J, Morgenthaler N, Lazarus J, Ludgate M, Evans C 2001 A luminescent bioassay for thyroid blocking antibodies. Clin Endocrinol (Oxf) 54:355–364
Tandon N, Freeman MA, Weetman AP 1992 T-cell responses to synthetic TSH receptor peptides in Graves disease. Clin Exp Immunol 89:468–473
Pichurin P, Yan XM, Farilla L, Guo J, Chazenbalk GD, Rapoport B, McLachlan SM 2001 Naked TSH receptor DNA vaccination: a TH1 T cell response in which interferon- production, rather than antibody, dominates the immune response in mice. Endocrinology 142:3530–3536
Rao PV, Watson PF, Weetman AP, Carayanniotis G, Banga JP 2003 Contrasting activities of thyrotropin receptor antibodies in experimental models of Graves’ disease induced by injection of transfected fibroblasts or deoxyribonucleic acid vaccination. Endocrinology 144:260–266
Nagayama Y, Kita-Furuyama M, Ando T, Nakao K, Mizuguchi H, Hayakawa T, Eguchi K, Niwa M 2002 A novel murine model of Graves’ hyperthyroidism with intramuscular injection of adenovirus expressing the thyrotropin receptor. J Immunol 168:2789–2794
Cordier A, Haumont S 1980 Development of thymus, parathyroids, and ultimo-branchial bodies in NMRI and nude mice. Am J Anat 157:227–263
Costagliola S, Rodien P, Many MC, Ludgate M, Vassart G 1998 Genetic immunization against the human thyrotropin receptor causes thyroiditis and allows production of monoclonal antibodies recognising the native receptor. J Immunol 160:1458–1465
Caturegli P, Rose N, Kimura M, Kimura H, Tzou S 2003 Studies on murine thyroiditis: new insights from organ flow cytometry. Thyroid 13:419–426
Dunn TB, Moloney JB, Green AW, Arnold B 1961 Pathogenesis of a virus induced leukaemia in mice. J Natl Cancer Inst 26:189–221
Vladutiu A, Rose N 1972 Aberrant thymus tissue in rat and mouse thyroid. Experientia 28:79–81
Muhlendyck H, Syed A 1978 Fixation artefacts in the external eye muscles in biopsy examinations. (A light microscopy and electron microscopy study). Buch Augenarzt 73:181–191
Shimojo N, Kohno Y, Yamaguchi K-i, Kikuoka S-i, Hoshioka A, Niimi H, Hirai A, Tamura Y, Saito Y, Kohn LD, Tahara K 1996 Induction of Graves-like disease in mice by immunization with fibroblasts transfected with the thyrotropin receptor and a class II molecule. Proc Natl Acad Sci USA 93:11074–11079
Kaithamana S, Fan J, Osuga Y, Liang SG, Prabhakar BS 1999 Induction of experimental autoimmune Graves’ disease in BALB/c mice. J Immunol 163:5157–5164
Combs TP, Pajvani UB, Berg AH, Lin Y, Jelicks LA, Laplante M, Nawrocki AR, Rajala MW, Parlow AF, Cheeseboro L, Ding YY, Russell RG, Lindemann D, Hartley A, Baker GRC, Obici S, Deshaies Y, Ludgate M, Rossetti L, Scherer PE 2004 A transgenic mouse with a deletion in the collagenous domain of adiponectin displays elevated circulating adiponectin and improved insulin sensitivity. Endocrinology 145:367–383
Pohlenz J, Maqueem A, Cua K, Weiss R, Van Sande J, Refetoff S 1999 Improved radioimmunoassay for measurement of mouse thyrotropin in serum: strain differences in thyrotropin concentration and thyrotroph sensitivity to thyroid hormone. Thyroid 9:1256–1271
Mooij P, de Wit H, Drexhage H 1994 A high iodine intake in Wistar rats results in the development of a thyroid-associated ectopic thymic tissue and is accompanied by a low thyroid autoimmune reactivity. Immunology 81:309–316(Glynn Baker, Gherardo Maz)