Conceived and designed the experiments: EMJ YT PK JVT. Performed the experiments: EMJ EC KH JVT YT. Analyzed the data: EMJ EC KH YT PK JVT. Contributed reagents/materials/analysis tools: EMJ YT PK JVT. Wrote the paper: EMJ YT.
The authors have declared that no competing interests exist.
Thyroglobulin (Tg) represents one of the largest known self-antigens involved in autoimmunity. Numerous studies have implicated it in triggering and perpetuating the autoimmune response in autoimmune thyroid diseases (AITD). Indeed, traditional models of autoimmune thyroid disease, experimental autoimmune thyroiditis (EAT), are generated by immunizing mice with thyroglobulin protein in conjunction with an adjuvant, or by high repeated doses of Tg alone, without adjuvant. These extant models are limited in their experimental flexibility, i.e. the ability to make modifications to the Tg used in immunizations. In this study, we have immunized mice with a plasmid cDNA encoding the full-length human Tg (hTG) protein, in order to generate a model of Hashimoto's thyroiditis which is closer to the human disease and does not require adjuvants to breakdown tolerance. Human thyroglobulin cDNA was injected and subsequently electroporated into skeletal muscle using a square wave generator. Following hTg cDNA immunizations, the mice developed both B and T cell responses to Tg, albeit with no evidence of lymphocytic infiltration of the thyroid. This novel model will afford investigators the means to test various hypotheses which were unavailable with the previous EAT models, specifically the effects of hTg sequence variations on the induction of thyroiditis.
Autoimmune conditions targeting the thyroid are a fairly common occurrence, with a population prevalence of 1–2%
Despite the knowledge that has been gained by studying EAT and its special place in scientific history as the first model of experimentally induced autoimmunity, the method suffers from some limitations. EAT is not a model that can discriminate between subtle differences in immune responses especially to different thyroglobulin molecules. In fact we have previously shown that specific Tg amino acid variants confer susceptibility to AITD as well as EAT
Several overlapping fragments of the human
DNA was freshly prepared using the BioRad maxiprep kit. DNA was ethanol precipitated then resuspended and adjusted to a concentration of 1.5 µg/µl.
Mouse and human thyroglobulin were a gift from the laboratory of Dr. Terry F. Davies, James J. Peters VA Medical Center.
C3H/Hen mice were obtained from Taconic laboratories. Mice which were used were exclusively female and were 8 weeks of age. All protocols were approved by the Mount Sinai School of Medicine Institutional Animal Care and Use Committee (animal use protocol #09-00149).
Female C3H/Hen mice were injected with bupivacaine into the right
All electroporation experiments were conducted under anesthesia as follows: Mice were anesthetized with Avertin (2,2,2, tribromoethanol, Sigma-Aldrich T4, 840.2). Avertin was administered i.p. at a dosage of 0.4–0.6 mg/gram of body weight. After mice were fully anesthetized, muscle was dissected and exposed. Following dissection, mice were injected into the specified muscle group with either 50 µl of 1.5 µg/µl hTg cDNA (a total of 75 µg) or 50 µl of 1.5 µg/µl of empty vector (total of 75 µg)−pcDNA3.1+ (Invitrogen). DNA was delivered in PBS/1.67% mannitol). Immediately after injection of Tg cDNA or empty vector, a current was administered to the site of injection using a square wave generator. During the exploratory phase of this work (comparison of electroporation
C3H/HeN mice received 50 µl of human Tg in 100 µl of PBS. This Tg was administered through a tail vein. 2–3 hours after the Tg, 20 µl LPS was administered. Three immunizations, spaced by approximately two week intervals, were administered. Animals were sacrificed two weeks after the third and final round of immunization.
Spleens were collected from mice and kept on ice in Hanks Balanced Salt Solution (HBSS). To harvest lymphocytes, the spleens were placed in complete RPMI medium [RPMI 1640 medium supplemented with fetal bovine serum (FBS; 10%), L-glutamine (2 mM) and penicillin-streptomycin (100 U/ml, 100 µg/ml)] (Hyclone, Fisher Scientific), and cut in several places. These were then pressed in a circular motion with a plunger from a 6 ml syringe until only fibrous tissue remained. To further disperse clumps in the suspension, it was drawn up and expelled several times through a 6 ml syringe using a 19 gauge needle, and then the suspension was filtered twice through a nylon screen (70 µm cell strainer) into a 50 ml falcon tube. The cell suspension was then spun for 10 minutes at 200× g and the supernatant was discarded. Cells were again suspended in complete RPMI medium and centrifuged a second time, and the media were discarded. To remove remaining non-lymphocytic cells from spleen, 5 ml of ACK lysis buffer was added to the cells for 5 minutes, after which complete RPMI medium was added to fill the tube, and centrifuged for 10 minutes. The remaining cell pellet was washed and resuspended in 10 ml of complete RPMI medium for counting and plating.
Cells isolated from spleens of mice were cultured in RPMI 1640 medium supplemented with fetal bovine serum (FBS; 10%), L-glutamine (2 mM) and penicillin-streptomycin (100 U/ml, 100 µg/ml) (Hyclone, Fisher Scientific). Cells were plated 2×105 cells per well in medium in a total volume of 200 µl. Cells were treated with either PBS (negative control), thyroglobulin (40 µg/ml) or concavalin A (2 µg/ml; positive control). After 48 hrs, cells were pulsed with 1 µCi/well of [3H]thymidine (MP biomedical, Costa Mesa, CA). Cells were harvested 18 h later, and [3H]thymidine incorporation was measured in a scintillation counter (TopCount·NXT™; PE life sciences, Boston, MA). All assays were performed in quadruplicates. Data are expressed as stimulation index. We calculated the stimulation index by using the following formula: Stimulation index = (CPM of the antigen treated lymphocytes−background)/(CPM of PBS treated lymphocytes−background).
10 µg/ml of mouse thyroglobulin in carbonate buffer (pH 9.6) (total volume 100 µl) was used to coat wells of an ELISA plate, and incubated overnight at 4°C. The plate was washed 4 times with PBST (PBS+0.05% tween). 200 µl of blocking buffer (PBST+2.5% bovine serum albumin) was added to wells and incubated for 1 hour at 37°C. The plate was then washed 4–6 times with PBST. Sera were diluted in 1∶100 in PBST/1%BSA and added in a total volume of 100 µl, in triplicates. Binding was allowed to proceed for 2 hours at room temperature. Unbound sera were removed by washing 4–6 times with PBST. Then 100 µl of goat anti-mouse IgG-HRP (Sigma-Aldrich), diluted 1∶500 in PBST/1% BSA was added. The binding was allowed to proceed for 30 minutes at 37°C. Unbound antibody was removed by washing with PBST 4 times. 100 µl of freshly prepared PNPP substrate was added to each well. Readings were performed by an ELISA reader at 405 nm.
In order to test for the isotype of the detected anti-Tg antibodies, a modified version of the methodology described by Chronopoulou
In order to calculate the amount of a specific IgG subclasses in the sera, a serum with known amounts of IgG subclasses (MP Biomedicals catalog #64091) was employed as follows: on the same plate which was used to read the immunized mouse sera, a series of wells were coated with anti-mouse IgG (Sigma M6898), diluted to 50 µg/ml, in the carbonate pH 9.6 buffer and was added at 100 µl per well. Washing was performed as described above. The standard reference serum was diluted 1∶10 in PBST to establish a working solution. Next, 7 serial dilutions of reference serum solution were performed (1∶5, 1∶25, 1∶125, 1∶625, 1∶3125, 1∶15625, and 1∶78,125). 100 µl of diluted reference serum was added to the wells coated with anti-mouse IgG. These reference dilutions were performed in triplicate to construct the calibration curve. Standard curves typically gave an R2≥0.95.
Interferon gamma and interleukin 4 levels in the supernatant of stimulated lymphocytes were determined by a commercial ELISA kits (BD Biosciences, OptEIA™).
The differences between T cell proliferation indexes and antibody levels at different conditions were analyzed using Student's t-test. P-values of <0.05 were considered significant.
Despite the immense potential of skeletal muscle to uptake and express a multitude of plasmid cDNAs
Shown is an ELISA measuring total IgG (all subclasses), directed against murine thyroglobulin (Tg). The results demonstrate a significant anti-Tg response in mice immunized with hTg cDNA. For this analysis, we compared 2 non-immunized mice, the 2 best responders, out of a group of 6 that were immunized with hTg without electroporation, and 6 hTg cDNA immunized/electroporated mice. Relative to directly immunized mice, the use of an electric current (6 pulses total, 60 V/cm, and 50 ms pulse length, with a 200 ms respite between pulses) caused a significant increase in the level of anti-Tg antibodies (p = 0.037). All readings were performed in triplicate. Data are shown as mean + standard error of the mean (SEM). Animals immunized with empty vector (pCDNA3.1+) exhibited no response (data not shown) to mouse Tg.
Splenocytes isolated from
Splenocytes isolated from immunized (2–3 months after the final boost) and non-immunized C3H/HeNmice were incubated with increasing concentrations of hTg protein for 72 hrs. Splenocytes from hTg immunized/electroporated mice showed significantly increased proliferation, compared to similarly challenged splenocytes from non-immunized mice. The x axis represents the concentration (µg/ml) of hTg which was used to stimulate the splenocytes. All T-cell proliferation experiments were performed in triplicates. Data, which were initially recorded as cpm, were subsequently expressed relative to splenocytes which did not receive thyroglobulin, are shown as mean + SEM.
To determine if cDNA immunization with hTG resulted in a Th1 or Th2 polarization of the anti-Tg response we further characterized the T cell responses by examining cytokine production. IFN-gamma production was significantly increased in response to hTg cDNA (
Splenocytes were isolated from immunized/electroporated (2–3 months after the final boost) and non-immunized mice. 2×105 cells were challenged with 0, 5, or 20 µg/ml of human thyroglobulin. Supernatants were collected and were analyzed for the presence of cytokines. hTg-treated splenocytes from immunized/electroporated animals secreted significantly higher levels of interferon-gamma than controls. Mice immunized with empty vector showed no differences compared to the non-immunized controls (data not shown). All readings were performed in triplicate. Data are shown as mean + SEM.
All the mice that were immunized with hTg cDNA and electroporated developed a robust antibody response to mouse thyroglobulin (mTg) demonstrating the development of a
Three, 8-week old C3H/HeN mice received hTg immunization/electroporation while 4 mice received an empty pcDNA plasmid electroporation (denoted by ‘**’). Anti-mouse Tg antibodies of the IgG2a subclass were the most prevalent isotype, followed by IgG1. There were very low levels of anti-mTg antibodies of the IgG2b class. All ELISA experiments were performed in triplicates. Data are presented as mean + SEM.
In order to compare our new model to the classical EAT model, we also induced classical EAT in the mice by immunizing them i.v. with hTg protein followed by i.v. LPS 2–3 hrs later. As expected, we saw a robust anti-Tg antibody response. Intriguingly, the classical EAT model showed a different IgG subtype response than the cDNA immunized/electroporated mice. While IgG2a showed high levels in the classical EAT similar to the cDNA immunization model, IgG2b showed very high levels in the classical EAT, while its levels remained very low in the cDNA immunized mice (
Three, 8-week old C3H/HeN mice were intravenously immunized with 50 µg of hTg in 100 µl of PBS into the tail vein. 2–3 hours later, 20 µl of LPS was injected into the tail vein. Three immunizations, spaced two weeks apart, were administered. Animals were sacrificed two weeks after the third and final round of immunization. At sacrifice, sera were collected and were analyzed for the presence of anti-mTg and the specific isotype of the antibodies. All ELISA experiments were performed in triplicates. Data are shown as mean + SEM.
Hematoxylin-eosin staining of thin sections from thyroid glands of mice from the different experiments did not show any evidence of lymphocytic infiltration into the thyroid gland (data not shown).
Several cDNA vaccine-based animal models have been recently developed, the most notable of which is experimental autoimmune Graves' disease (EAGD), induced by immunization of BALB/c mice with an Adenovirus vector carrying the TSH receptor
Up until 1980, it was believed that Tg reactive T-cells were completely deleted by central tolerance mechanisms. Therefore, the dogma was that EAT can only be induced with adjuvant which stimulated non-Tg reactive T-cells to bypass the tolerance to Tg and provide help for Tg-reactive B-cells. However, in 1980–1981, Rose and colleagues published seminal studies clearly demonstrating that repeated immunizations with high doses of Tg without adjuvant could induce a strong anti-Tg antibody response and lymphocytic infiltration of the thyroid
Traditionally, cDNA vaccines are best known as agents that can generate protective immunity against infectious disease and experimental cancer, or can supply a therapeutic dose of a gene product (reviewed in
Thyroglobulin is a 660 kDA homodimeric, iodinated glycoprotein that serves as a precursor for thyroid hormones
The current new model we have developed demonstrates that it is possible to breakdown tolerance to Tg, using cDNA immunization coupled with electroporation. However, despite the breakdown in tolerance, the mice did not develop lymphocytic infiltration of the thyroid and hypothyroidism. One potential explanation for the lack of lymphocytic infiltration of the thyroid is that we have immunized with human Tg. Even though there is 70% amino acid sequence identity between human and mouse Tg, enough differences may exist that it may not be sufficient, using
Despite the lack of lymphocytic infiltration of the thyroid, our findings furnish significant implications and afford new options to future studies focusing on the etiology of AITD. The main advantage of our model is that it enables immunization with human Tg and it facilitates manipulating the sequence of the hTg cDNA (through site directed mutagenesis), in order to investigate the importance of different parts of the molecule in inducing autoimmunity. Indeed, our genetic studies have demonstrated a strong association between amino acid variants in the thyroglobulin gene and AITD, suggesting that subtle sequence changes in hTg may have a significant effect on disease susceptibility
In summary, for the first time we have produced a mouse model of thyroid autoimmunity using cDNA immunization, coupled with electroporation, without the need for a strong adjuvant such as CFA or LPS. Our novel mouse model is unique because it involves immunization by intramuscular injection of Tg cDNA resulting in production of non-iodinated Tg in muscle cells. Because we utilized cDNA immunization our model is amenable to genetic modification of Tg used in the immunization. This model should prove useful for future studies on the role of different Tg sequence variants or Tg domains in the induction of thyroid autoimmunity. Furthermore, the milieu afforded by the mouse myocyte will allow testing, through the co-immunization and electroporation of their respective cDNA's, of the role of pro-inflammatory and pro-apoptotic factors in the pathoetiology of Hashimoto thyroiditis. Overall, this new model of thyroid autoimmunity is a model with much potential and flexibility.
We are thankful to Drs. Mari Arufe and David Sassoon for technical assistance. We thank Drs. Rauf Latif and Takao Ando for their expert advice and sharp insights.