Establishment and Comparison of Two Different Animal Models of Graves’ Orbitopathy

Abstract

Purpose: The purpose of this study was to construct an animal model of Graves’ ophthalmopathy (GO) by comparing recombinant adenovirus expressing human thyrotropin receptor A subunit (Ad-TSHR A) gene immunization and dendritic cell (DC) immunization. We evaluated the animal models that are closer to the pathology of human GO, and laid the foundation for the study of GO.

Materials and Methods: Ad-TSHR A was injected intramuscularly into female BALB/c mice to induce the GO animal model. A GO animal model was constructed using TSHR combined with IFN-γ-modified primary DC immunized female BALB/c. The animal models constructed by the above two methods were evaluated in terms of ocular appearance, serology, pathology, and imaging to assess the modeling rate of the animal models, respectively.

Results: Both modeled mice exhibited increased serological indexes of free thyroxine (FT4) and TSH receptor antibodies (TRAbs) levels and decreased TSH (P < 0.01). Thyroid pathology analysis revealed the number of thyroid follicles increases, the size varies, and the follicular epithelial cells proliferate to varying degrees in a cuboidal or tall columnar pattern, with a small amount of lymphocytic infiltration visible. Adipose tissue behind the eyeball was accumulated, the muscle outside the eyeball was broken and fibrotic, and hyaluronic acid (HA) behind the eyeball was increased. The animal model of GO constructed by immunization of TSHR with IFN-γ-modified DC had a modeling rate of 60%, whereas that of Ad-TSHR A gene immunization was 72%.

Conclusions: Both gene immunization and cellular immunization can be used to construct GO models, and the modeling rate of gene immunization is higher than that of cellular immunization.

Translational Relevance: In this study, two innovative methods, cellular immunity and gene immunity, were used to establish GO animal models, which improved the success rate to a certain extent. To our knowledge, this study presents the first cellular immunity modeling idea of TSHR combined with IFN-γ for the GO animal model, which provides an animal model basis for understanding the pathogenesis of GO and developing new treatment methods.

Introduction

Graves’ ophthalmopathy (GO) is an autoimmune disease that affects the eyes, and is closely associated with autoimmune thyroid disease. It is also known as thyroid eye disease or thyroid-associated ophthalmopathy (TAO).^1–3 GO is primarily seen in patients with Graves’ disease, but can also develop in patients with normal thyroid function, Hashimoto's thyroiditis, or thyroid cancer.^3–5 GO is a chronic condition that can cause severe eye problems, including blindness, and significantly impact patients’ daily lives, resulting in significant economic and spiritual burdens.^6,7 In recent years, the incidence of GO has been increasing, and it is now the most common orbital disease among adults.^8–10 

GO is an organ-specific autoimmune disease that is triggered by genetic and environmental factors and activated by cellular and humoral immunization.^11,12 The disease primarily affects orbital fibroblasts, which interact with pro-inflammatory factors, chemokines, and other immune cell-secreted factors, leading to pathological changes characterized by retrobulbar and periorbital tissue infiltration, including soft tissue inflammation, fibrous adipose tissue expansion, remodeling, and fibrosis.^10,13,14 Although the pathogenesis of GO is not fully understood, current research suggests that the thyrotropin receptor (TSHR) and TSH receptor antibodies (TRAbs) play a crucial role in the disease process.^15,16 TSI, in particular, is the only specific biomarker of Graves’ disease and GO,¹⁵ and its level is closely related to the clinical features of the disease.^16,17 

Despite ongoing research, effective animal models for GO remain elusive. Existing methods, such as direct immunity, genetic immunity, cellular immunity, drug induction, and protein immunity, have limited effectiveness due to low modeling rates, long modeling cycles, complex operations, and poor reproducibility.¹⁸ In this study, we modified the experimental protocol, improved the existing genetic immunization method, and developed a new cellular immunization method. We used these methods to construct two GO models by immunization with TSHRA recombinant adenovirus gene and TSHR combined with IFN-γ-modified dendritic cell (DC). We evaluated the models from various perspectives, including ocular appearance, serological testing, imaging assessment, and pathological testing, in order to develop better animal models and promote further study of GO. 

Materials and Methods

Animals Study

We confirm the study adherence to the ARVO Statement for the Use of Animals in Ophthalmic and Vision. The experimental procedures performed on mice were conducted in accordance with the approved guidelines in the ethical permit approved by the Longhua Hospital Shanghai University of Traditional Chinese Medicine Animal Welfare and Ethics committee (No. LHERAW-19067). The National Institutes of Health Guide for the Care and Use of Laboratory Animals served as the reference for this study. Six-week-old female BALB/C mice (specific pathogen-free grade, animal license number: SCXK [Shanghai] 2018-0003) were obtained from Shanghai Ling Chang Biotechnology Laboratory (Shanghai, China). BALB/C mice were kept in a constant temperature and 12/12-hours light-dark cycle environment and allowed free access to sufficient water and food without restriction. 

Establishment of GO Model by Immunization With TSHR Combined With IFN-γ-Modified Primary DC

Group Allocation

A total of 80 mice were randomly divided into four groups based on body weight. Twenty mice were immunized with TSHR and IFN-γ gene-modified DC (TSHR + IFN-γ), 20 mice were immunized with TSHR gene-modified DC (TSHR), 20 mice were immunized with IFN-γ gene-modified DC (IFN-γ), and 20 mice were injected with saline at the same site as a control group. 

DC Culture

Bone marrow cells were collected from the femur and tibia of 6-week-old BALB/c mice and spread in RPMI-1640 medium containing 10% fetal bovine serum (FBS; Cat #16140071; Gibco, Grand Island, NY, USA), penicillin-streptomycin solution (100×, Cat #C0222; Beyotime Biotechnology, Shanghai, China), granulocyte-macrophage colony stimulating factor (GM-CSF; 20 ng/mL, Cat #415-ML-010; R&D Systems, Shanghai, China), and interleukin-4 (IL-4; 15 ng/mL, Cat #404-ML-010; R&D Systems, Shanghai, China).¹⁹ The medium was changed every other day, and the cells were collected in suspension on the seventh day as primary cultured DC, and the purity of DC was detected by flow cytometry. 

Infection of DC

The recombinant adenovirus used in the experiment was synthesized and packaged by Shanghai Genechem Technology Laboratory (Shanghai, China). The multiplicity of infection (MOI) was calculated as (virus titer × virus volume)/cell number. Before the experiment, cells were ensured to be in optimal growth conditions and were counted after resuspension. Different MOIs (25, 50, 100, 200, 300, and 400) were tested, and the cell suspension (1 × 10⁴ cells per well of a 96-well plate), virus dilution, and Hitrans GP virus infection reagent (Cat #REVG005; Jikai, Shanghai, China) were mixed and added to the 96-well plate. The fluorescence expression was observed by inverted fluorescence microscopy after 48 hours. The infection conditions and MOI values corresponding to an infection efficiency of about 70% and good cell status were selected as the basis for subsequent infection experiments. 

Cellular Immunization Protocol

Recombinant adenovirus encoding TSHR and IFN-γ was utilized to infect primary DCs. A total of seven immunizations were performed in mice every 3 weeks, with each injection consisting of 50 µL PBS containing 1 × 10⁶ DCs via tail vein injection. Control mice were injected with phosphate-buffered saline (PBS) only. The entire injection period lasted for 18 weeks and mice were euthanized at 4 weeks after the last immunization. 

Establishment of GO Model by Immunization With TSHR A Subunit Recombinant Adenovirus Gene

Group Allocation

Seventy-five mice were randomly allocated into three groups based on their body weight. Twenty-five mice received an intramuscular injection of Ad-TSHR A, whereas another 25 mice were injected with the same dose of adenovirus expressing green fluorescent protein (Ad-EGFP). The remaining 25 mice were injected with saline at the same site and served as the control group. 

Genetic Immunization Protocol

The immunization protocol consisted of four initial injections administered at 3-week intervals, followed by 7 booster injections administered at 4-week intervals. A volume of 25 µL of adenovirus was injected into the left and right femoral muscles each time, with a concentration of 10¹⁰ PFU. The total injection period was 37 weeks and mice were euthanized at 4 weeks after the last immunization. 

Detection of Serum FT4, TSH, and TRAb

To detect serum levels of free thyroxine (FT4), TSH, and TRAb, enzyme-linked immunosorbent assay (ELISA) kits were used (Cat #s YKE30223, YKE30226, and YKE30224, respectively, from Yankobio, Hubei, China). The ELISA kit was removed in advance and equilibrated to room temperature. Sample wells and standard wells were set up, and 50 µL of standards at different concentrations (48, 24, 12, 6, 3, and 1.5 U/L) were added to the standard wells. Then, 50 µL of the sample to be tested was added to the sample wells, and the blank wells were left empty. Horseradish peroxidase-labeled detection antibody (100 µL) was added to each well except the blank wells, which were sealed with a sealing membrane and incubated at 37°C for 1 hour. The liquid was then discarded, and 350 µL of washing solution was added to each well, left for 1 minute, and then discarded, and the plate was washed 5 times in total. Substrate A (50 µL) and Substrate B (50 µL) were added to each well, and the plate was incubated for 15 minutes at 37°C, protected from light. Termination solution (50 µL) was then added to each well, and the optical density (OD) value was detected at 450 nm using an enzyme marker. 

Animal Sampling and Sample Processing

In this study, mice were subjected to intraperitoneal injection of 3% pentobarbital sodium to induce anesthesia, sterilized with 75% ethanol, and placed on a material plate. The chest cavity was then opened, and the heart was fully exposed. Blood was drawn from the heart using a 1 mL syringe and collected in a 1.5 mL centrifuge tube. The collected blood was kept at room temperature for 1 hour and then transferred to a 4°C refrigerator for 5 hours. The supernatant was separated by centrifugation at 3000 rpm for 10 minutes at 4°C and stored in an −80°C constant temperature freezer for subsequent ELISA detection. 

After blood collection, the hair around the eyes was removed, and the head and neck skin were cut off in an inverted T-shape. The orbital bone was then removed using ophthalmic scissors, and the optic nerve was protected. The trachea was fully exposed by cutting the neck skin, and the thyroid tissue, including the trachea, was removed with ophthalmic scissors. The complete thyroid, eyeball, and retrobulbar tissues were placed in a 4% paraformaldehyde (PFA) solution. The tissues were fixed in the PFA solution for 48 hours at 4°C. The eyeball was removed with ophthalmic scissors and a surgical blade after fixation, preserving the retrobulbar tissue and optic nerve. 

Some tissues were dehydrated in 10%, 20%, and 30% sucrose solutions, and the dehydrated fundus tissue was covered with optimal cutting temperature (OCT) frozen section embedding agent, and cut into 20 µm slices for subsequent Oil Red O staining. Thyroid tissue and another part of fundus tissue were placed in an automatic tissue dehydrator for gradient dehydration. The dehydrated fundus tissue was routinely embedded in paraffin and cut into 5 µm slices. The paraffin sections were baked for 3 hours at 64°C and stored at room temperature for later use in hematoxylin and eosin (H&E) staining, Masson and Standard alcian blue staining. 

Histological Examination of Thyroid and Orbit Tissues

The H&E staining kit (Beyotime Biotechnology, Jiangsu, China) was utilized to evaluate the histomorphological changes of the thyroid gland and a portion of the orbital tissue. The Masson staining kit (Solarbio, Beijing, China) was used on orbital sections to detect fibrosis of the extraocular muscles. Frozen sections were stained with the Oil Red O staining kit (Solarbio, Beijing, China) to detect adipose tissue. The Standard alcian blue staining kit (Solarbio, Beijing, China) was used to detect hyaluronic acid (HA). 

For the H&E staining of the thyroid gland and orbital tissue, the slices were subjected to gradient dewaxing, followed by rinsing with distilled water for 5 minutes, for 3 times in total. Hematoxylin staining was performed for 7 minutes, and then the slices were rinsed with running water for 10 minutes. Hydrochloric acid ethanol fractionation solution was then soaked for 1 minute, and the slices were rinsed with tap water for 10 minutes. Eosin staining was performed for 2 minutes, followed by rinsing with running water for 10 minutes. Finally, neutral gum was used to seal the slices after gradient dehydration. 

For the Oil Red O staining of the orbital tissue frozen sections, the slices were washed in distilled water for 20 seconds, followed by washing in 60% isopropyl alcohol for 20 seconds. Oil Red O staining was performed for 15 minutes, and then the slices were washed in 60% isopropyl alcohol for 2 seconds to remove the stain. Mayer hematoxylin stain was applied for 2 minutes, and the slices were slightly rinsed in distilled water before being sealed with glycerin gelatin. 

For the Masson staining of the orbital tissue, the slices were routinely dewaxed to water, and 4% PFA was used to fix the tissue for 15 minutes. Mordant staining solution was then added for dip-staining, and the tissue was subjected to mordant staining for 1 hour at 60°C in a warm oven, followed by rinsing in running water for 10 minutes. Azurite blue staining solution was drip-stained for 2.5 minutes, and the slices were washed twice with water for 15 seconds each time. Hematoxylin staining solution was then applied dropwise for 2 minutes, followed by washing twice with water for 15 seconds each time. Acidic ethanol differentiation solution was used for differentiation for 1 minute until the tissue turned completely red, and the differentiation was terminated by 10 minutes of water washing. Lichon red magenta staining solution was then applied dropwise for 15 minutes, followed by washing with weak acid solution. Phosphomolybdic acid solution was used for 10 minutes, and then the top solution was poured off, and aniline blue staining solution was added dropwise for 30 seconds. After washing off the Aniline Blue solution with weak acid, the sections were treated with drops of weak acid for 2 minutes. Finally, after gradient dehydration, the sections were sealed with neutral gum. 

For the Standard alisin blue staining of orbital tissue, the slices were routinely removed from the wax from xylene to water. Prepared Alcian acid solution: Alcian acid solution: distilled water = 1:2. Soaked in Alcian acid solution for 5 minutes. Stained in Alcian staining solution for 30 minutes. Rinsed with running water for 5 minutes. Counterstained with Nuclear Fast Red stain for 5 minutes. Rinsed with running water for 5 minutes. Dehydrated with absolute ethanol I and II each for 1 minute. Cleared with xylene I and II each for 1 minute. Mounted with neutral gum. 

MRI Scanning

One week after the last immunization, ocular imaging of mice was performed using a Biospec mouse micro-magnetic resonance imaging (MRI) instrument (Bruker, 7.0 t/20 cm). The T2-weighted images were taken (fast spin-echo sequence, echo delay time = 36 ms; repetition time = 2000 ms; 256 [h] × 256 [v] matrix; flip angle α = 180 degrees; slice thickness = 1.5 mm; and field of view = 30 mm × 60 mm). Mice were first anesthetized (isoflurane + pure oxygen, 3% induction anesthesia, 1 to approximately 1.5% maintenance dose) and their respiration and heart rate were closely monitored during the examination. 

Statistical Analyses

All data are presented as means and standard deviations (SDs). Differences between groups were analyzed using ANOVA or the independent samples t-test. All statistical analyses were performed using SPSS (SPSS 25.0), and P < 0.05 was considered significant. 

Results

TSHR Combined With IFN-γ-Modified Primary DC Immune Construct GO Model

The TSHR combined with IFN-γ-modified primary DC (pDC) immune construct GO model was utilized in the DC immunization protocol and sacrificed time (Fig. 1A). The surface molecule CD11c on DCs cultured up to the eighth day was detected using flow cytometry. The flow cytometry plots revealed that the purity of CD11c + pDCs had reached 93.3% (Fig. 1B). Fluorescence staining results indicated that TSHR and IFN-γ adenovirus were capable of infecting pDCs with an infection efficiency of over 70%, thereby confirming the success of the co-infection for subsequent experiments (Fig. 1C). During the experiment, the daily food intake normalized to body weight was higher in the TSHR and IFN-γ + TSHR group than in the control group and the IFN-γ group (Fig. 1D). 

Figure 1.

 

(A) DC immunization protocol. (B) Flow cytometry plots showing the purity of DCs. (C) Fluorescence microscopy of primary dendritic cells and their transfection efficiency. (D) Daily food intake normalized to body weight during the entire experiment. (E) Gross images of eyes in control, IFN-γ, TSHR, IFN-γ + TSHR immunized group mice. (F) MRI scans of control and IFN-γ + TSHR immunized group mice orbits. (G) Quantitative analysis of eye protrusion in the control and IFN-γ + TSHR groups in MRI. **: P < 0.01. (H) MRI of the extraocular muscles in the control and IFN-γ + TSHR groups, the red arrows indicate the extraocular muscles. For all the statistical analyses in this figure, significance was determined by the independent samples t-test, and results were shown as mean ± SD.

In order to verify whether the mice displayed specific manifestations of GO, their ocular appearance changes and orbit MRI scans were observed. The mice in the control group, IFN-γ group, and TSHR group exhibited normal ocular appearance, without any abnormal changes, such as protruding eyeballs and widening of the eye fissure. However, out of the 20 mice in the IFN-γ + TSHR group, 11 mice displayed orbital conjunctival congestion and protruding eyeballs (Fig. 1E). In the axial and coronal MRI images of mouse orbital, we observed that the eyes of IFN-γ + TSHR group were more protruding than those of the control group (Fig. 1F). In addition, we quantified the distance between the midpoint of the mouse eye and the anterior midline using three images of the same plane from the coronal scans of the mouse orbital MRI with the same sequence (Fig. 1G). Furthermore, we noticed that the extraocular muscle was hypertrophic in the IFN-γ + TSHR group compared with the control group (Fig. 1H). 

The TRAb levels were significantly higher in the IFN-γ + TSHR group when compared to the other groups (Fig. 2A). Furthermore, immunization with TSHR and IFN-γ + TSHR not only suppressed the expression of TSH but also upregulated FT4 expression in the serum, when compared to the control and IFN-γ group (Figs. 2B, 2C). H&E staining revealed serious destruction of normal thyroid follicles in the thyroid gland following immunization with TSHR and IFN-γ + TSHR, when compared to the control group and the IFN-γ group. Specifically, 12 out of 20 mice in the TSHR group and 15 out of 20 mice in the IFN-γ + TSHR group displayed a large number of disrupted thyroid follicles, proliferated follicular epithelial cells, and occasional small lymphocyte infiltration (Fig. 2D). 

Figure 2.

 

(A, B, C) ELISA and quantitative analysis of FT4, TSH, and TRAb in the serum. ****: P < 0.0001. (D) H&E staining of the thyroid gland tissue of control, IFN-γ, TSHR, and IFN-γ + TSHR immunized group mice (× 200). Scale bar = 100 µm. (E) H&E, Oil Red O, Masson, and Alcian blue staining of the orbital tissue of the control, IFN-γ, TSHR, and IFN-γ + TSHR immunized group mice (× 40). Scale bar = 100 µm. In H&E and Oil Red O staining of the orbital tissue, the black arrows indicate the adipose tissue. In Masson staining of the orbital tissue, the black arrows indicate the muscle fibers. (F) Quantitative analysis of adipose tissue in H&E staining of the orbital tissue, the y-axis represents the value of the stained area/total area. **: P < 0.01. (G) Quantitative analysis of orbital fibrosis in Masson staining of the orbital tissue, the y-axis represents the value of the stained area/total area. **: P < 0.01. For all the statistical analyses in this figure, significance was determined by 1-way ANOVA followed by Tukey's multiple comparisons test, and results were shown as mean ± SD.

H&E staining and Oil Red O staining demonstrated that immunization with IFN-γ + TSHR could increase the adipose tissue which distributed around the optic nerve in orbital tissue (Figs. 2E, 2F). Specifically, 12 out of 20 mice in the IFN-γ + TSHR group had significant orbital adipose tissue hyperplasia, accompanied by a few distributed lymphocytes. Alcian blue staining showed that there was no blue staining in the retrobulbar adipose interstice of the control group, the TSHR group, and IFN-γ group mice, whereas there were 9 of the 20 mice in the TSHR + IFN-γ group whose retrobulbar adipose interstitial space was stained blue, indicating that the production of HA increased (Fig. 2E). Additionally, Masson staining indicated the fibrosis of extraocular muscles of the orbit, and disorder and fracture of muscle fibers, which were visible in 8 out of 20 mice in the IFN-γ + TSHR group (Figs. 2E, 2G). 

Ad-TSHR A Subunit Recombinant Adenovirus Gene Immunization to Construct GO Model

Ad-TSHR A subunit recombinant adenovirus gene immunization was employed to construct the GO model, following the Ad-TSHR A immunization protocol and euthanization time (Fig. 3A). During the experiment, the daily food intake normalized to body weight was higher in the Ad-TSHR A group than in the control and the Ad-EGFP group (Fig. 3B). Out of 25 mice in the Ad-TSHR A group, 17 exhibited specific GO changes, such as eye protrusion and widening of the eye fissure, which were not observed in the control and Ad-EGFP groups (Fig. 3C). In the coronal view scans, we measured the distance between the midpoint of the eye and the anterior midline in three identical scan planes, the results showed that the eyes of the Ad-TSHR A group were more protruded than those in the control (Figs. 3D, 3E). Additionally, in the axial view scans, we also observed that the eyes of the Ad-TSHR A group were more protruded (Fig. 3D). Furthermore, we observed that the extraocular muscles in the Ad-TSHR A group were hypertrophic compared to those in the control group (Fig. 3F). 

Figure 3.

 

(A) Ad-TSHR A immunization protocol. (B) Daily food intake normalized to body weight during the entire experiment. (C) Gross images of eyes in the control, Ad-EGFP, and Ad-TSHR A groups of immunized mice. (D) MRI scans of control and Ad-TSHR A group mice orbits. (E) Quantitative analysis of eye protrusion in the control and Ad-TSHR A groups in MRI. **: P < 0.01. (F) MRI of the extraocular muscles in the control and Ad-TSHR A groups, the red arrows indicate the extraocular muscles. For all the statistical analyses in this figure, significance was determined by the independent samples t-test, and results were shown as mean ± SD.

Serum ELISA results indicated that Ad-TSHR A injection suppressed the expression of TSH and upregulated TRAb and FT4 expression equally (Figs. 4A–C). Furthermore, H&E staining of thyroid gland revealed that thyroid follicles in the control and Ad-EGFP group were structurally intact, with a small number of follicles, low cuboidal follicular epithelial cells, and no lymphocytic infiltration in the follicular interstitium. In contrast, 19 out of 25 mice in the Ad-TSHR A group exhibited the number of thyroid follicles is increased, with varying degrees of follicular epithelial cell hyperplasia in a cuboidal or hypercolumnar pattern, and a small amount of lymphocytic infiltration is seen (Fig. 4D). H&E staining and Oil Red O staining showed that Ad-TSHR A injection caused orbital adipose tissue hyperplasia and lymphocytic infiltration in 18 out of 25 mice (Figs. 4E, 4F). Alcian blue staining indicated that there was no blue staining in the retrobulbar adipose interstice of the control and Ad-EGFP group mice, and there were 10 of the 25 mice in the Ad-TSHR A group whose retrobulbar adipose interstitial space were stained blue, indicating that the production of HA increased (see Fig. 4E). Masson staining indicated that the orbital extraocular muscles of 12 out of 25 mice in the Ad-TSHR A group were stained blue, suggesting extraocular muscle fibrosis, with disorganized and broken extraocular muscle fibers being observed (see Figs. 4E, 4G). 

Figure 4.

 

(A, B, C) ELISA and quantitative analysis of FT4, TSH, and TRAb in the serum. ****: P < 0.0001. (D) H&E staining of the thyroid gland tissue of the control, Ad-EGFP, Ad-TSHR A groups of immunized mice (× 200). Scale bar = 100 µm. (E) H&E, Oil Red O, Masson, and Alcian blue staining of the orbital tissue of the control, Ad-EGFP, and Ad-TSHR A groups of immunized mice. The black arrow indicates retrobulbar adipose tissue (× 40). Scale bar = 100 µm. In H&E and Oil Red O staining of the orbital tissue, the black arrows indicate the adipose tissue. In Masson staining of the orbital tissue, the black arrows indicate the muscle fibers. (F) Quantitative analysis of adipose tissue in H&E staining of the orbital tissue, the y-axis represents the value of the stained area/total area. **: P < 0.01. (G) Quantitative analysis of orbital fibrosis in Masson staining of the orbital tissue, the y-axis represents the value of the stained area/total area. **: P < 0.01. For all the statistical analyses in this figure, significance was determined by 1-way ANOVA followed by Tukey's multiple comparisons test, and results were shown as mean ± SD.

Discussion

The etiology and pathogenesis of GO remain incompletely understood, largely due to the absence of widely recognized animal models. Presently, the primary approaches to GO modeling include adenovirus, plasmid, cellular, and other immunizations. Plasmids are double-stranded DNA molecules that express genetic information and serve as vehicles for gene immunity, which can be transcribed and expressed effectively by host cells. BALB/c mice were immunized using the extracellular structural domain (ECD) of human TSHR to generate TBII in the serum and induce pathological changes in the thyroid gland, but no changes in ocular signs were detected.²⁰ Female BALB/c mice were also immunized by electroporation with a plasmid-conjugated TSHR A subunit, resulting in a 20% to 30% increase in orbital volume after 22 weeks, but the pathology was not well established.²¹ Male and female mice immunized with human TSHR A subunit plasmids exhibited hyperthyroidism, characterized by TSHR-stimulating antibodies, elevated T4, pathological changes in the thyroid gland, and cardiac hypertrophy. Ocular protrusion and inflammation were also observed, along with increased fat volume and glycosaminoglycan deposition, but significant ocular protrusion was only observed in a portion of the mice.²² Immunization of female BALB/c mice using TSHR A subunit plasmids combined with electroporation at two experimental centers with identical environments resulted in differing serological and thyroid pathology outcomes, indicating a lack of reproducibility of the GO model constructed using this method.²³ Female BALB/c mice immunized with insulin-like growth factor 1 receptor a (IGF-1Ra) subunit plasmids did not exhibit hyperthyroidism or pathological changes in the thyroid gland or orbit, but anti-IGF-1Ra antibodies were detected.²⁴ 

Adenovirus is the primary vector used in gene therapy due to its higher infection efficiency, wider host range, ability to infect dividing or non-dividing cells and primary cells, and greater safety profile. TSHR has a distinctive molecular structure as a G protein-coupled receptor, consisting of an extracellular A subunit, a transmembrane structural domain, and an intracellular B subunit linked by disulfide bonds to form a heterodimer.^25–27 Overexpression of the TSHR A subunit (aa.1-289) in vivo increases the incidence of Graves’ disease compared to full-length TSHR immunization.²⁸ Genetic immunization of TSHR A subunits using adenoviral vectors can increase the incidence of hyperthyroidism to approximately 65% to 80%, suggesting that overexpression of TSHR A subunits is a highly effective method of inducing hyperthyroidism. Adenovirus immunization of various strains of mice resulted in elevated serum thyroxine T4 levels, and follicular cell hypertrophy and hyperplasia in the thyroid gland, but no significant changes in extraocular muscles.^29,30 Hyperthyroidism was induced by using recombinant adenovirus that encoded either the full-length or a subunit of human TSHR. The success rate of establishing animal models of GO can be increased by prolonging the duration of routine immunization. Therefore, we attempted to extend the injection time and increase the number of injections of TSHR A subunit recombinant adenovirus to induce an animal model of GO. 

Female BALB/c mice were subcutaneously injected with a recombinant adenovirus expressing TSHR after transfection of DCs. One injection was given at 3 weeks. After 2 and 3 injections, 8% and 35% of mice, respectively, developed hyperthyroidism characterized by the production of TSHR-stimulating antibodies, elevated serum thyroxine levels, and diffuse goiter with thyroid epithelial cell hyperplasia. Compared to the present study, it had a low success rate of hyperthyroidism and no damage to the intact extraocular muscles.³¹ BALB/c mice were intraperitoneally injected with the adenovirus of the TSHR A subunit after DC transfection. One injection was given at 3-week intervals. After 3 injections, 70% of BALB/c mice developed hyperthyroidism. Most hyperthyroid mice exhibited detectable TSAb activity, and thyroid H&E staining showed a diffusely enlarged thyroid with hypertrophic hyperplasia of thyroid epithelial cells. However, there was no lymphocytic infiltration and no ocular appearance was described.³² Few studies have been conducted on the modeling approach of cellular immunity, but it provides a new perspective on the modeling approach. Hence, this study attempts to use the cellular immunity approach for a contemporaneous comparison with genetic immunity. 

Immune mechanisms play a critical role in the pathogenesis of GO, which is characterized by infiltration of various immune cells in the orbit. In the early stages of GO, Th1 cells and their associated factors, such as TNF-α, IL-1β, IFN-γ, IL-2, and IL-12, dominate and promote disease progression into the active phase, leading to activation and proliferation of orbital fibroblasts (OFs) and increased production of glycosaminoglycans (GAGs). In the late stages of GO, Th2 cells mediate humoral immunity and autoantibody production, and their main cytokines include IL-4, IL-5, and IL-10, which are associated with disease maintenance, remission, and fibrosis.³³ The main function of IFN-γ is to participate in the immune regulation of the body. It is secreted by Th1 cell activation and mediates cellular immunity by promoting antigen expression in various cells. On one hand, it enhances the ability of specialized antigen-presenting cells to process and present antigens, and on the other hand, it leads to a Th1/Th2 deviation toward Th1. The mRNA levels of pro-inflammatory cytokines IL-1β and IFN-γ are significantly elevated in the orbital tissues of patients with GO. When orbital fibroblasts are activated by IL-1β and IFN-γ, they can express high levels of IL-6, PGE2, and MHCII.^34,35 which are closely related to the activation of T and B cells. IFN-γ can activate Th1 cells, thereby promoting the secretion of cytokines, such as IFN-γ and TNF-α, which can, in turn, inhibit Th2 cells. The level of IFN-γ expression in patients with GO is significantly higher than in controls, and IFN-γ expression is positively correlated with GO activity, suggesting that the higher the level of IFN-γ, the more severe the clinical manifestations.³³ Given the crucial role of IFN-γ in autoimmune diseases and the T1/T2 imbalance in GO pathogenesis, we hypothesized that IFN-γ cotransfection could induce a stronger immune response, disrupting T1/T2 homeostasis and inducing the onset of GO. 

DCs are a unique type of immune cell that belong to the class of antigen-presenting cells. They possess the ability to recognize, extract, process, and present antigens, and serve as the starting point of the adaptive immune response. DCs play a critical bridging role between the intrinsic and adaptive immune responses, and participate in various immune regulatory processes in the body, including infections, tumors, and autoimmune diseases.³⁶ Notably, DCs not only initiate antigen-specific immune responses but also induce immune tolerance and regulate immune homeostasis, making them essential in initiating T- and B-cell-mediated immune responses. Studies have utilized flow cytometry to detect DC subpopulations in the peripheral blood of patients with GO and healthy controls, revealing that blood myeloid DCs (mDCs) are negatively correlated with the incidence of GO. This finding suggests that alterations in mDC ratios may be associated with the development of GO. Galectin(gal), a protein that plays a crucial role in regulating DC maturation and immunity,³⁷ has been shown to have reduced expression in the DCs of patients with autoimmune thyroid disease, particularly in those with GO. Moreover, the severity of ocular symptoms in patients with GO is negatively correlated with Gal-9 expression in DCs.³⁸ As antigen-presenting cells, DCs can initiate T- and B-cell-mediated immune responses, which contribute to the development of autoimmune diseases. Therefore, we induced the GO model in BALB/c female mice by using TSHR combined with IFN-γ modified primary DCs. 

In this experiment, both cellular and genetic immunity approaches were used to construct GO animal models. The models generated from both methods performed similarly and were consistent with human disease characteristics. The modeling rate was 60% for the TSHR combined with IFN-γ-modified DC approach and 72% for the recombinant adenovirus immunization approach, with the success rate of genetic immunization being higher compared to cellular immunization. 

In conclusion, the TSHR combined with IFN-γ-modified DC immunological approach is a novel method that presents a new modeling idea for GO animal models. The evaluation criteria for these two modeling techniques are more comprehensive than those of existing models and include the presence or absence of active TSHR antibodies, orbital MRI, pathological changes in the thyroid gland and eye, and persistent and stable eye appearance. Additionally, the modeling rates of both methods are higher. However, the main limitations of this study include the prolonged modeling time of cell immunization and genetic immunization methods, as well as the lack of construction of the model in multiple centers simultaneously due to time and budget constraints, to validate the reproducibility of the modeling methods. Furthermore, simultaneous comparison of GO phenotypes at different time points is lacking, and the determination of the properties of retrobulbar adipose tissue after modeling still needs further exploration in the future. The pathogenesis of GO is complex and deserves further exploration in future modeling studies. 

Acknowledgments

The authors thank the staff of the Central Laboratory of Longhua Hospital affiliated to Shanghai University of Traditional Chinese Medicine for their technical guidance in this study. 

Authors’ Contributions: The authors’ responsibilities were as follows: W.W. and H.L. were responsible for designing and planning of the study. W.W. and J.W.-Z. performed the experiments. W.W., Y.J.-Q., and H.Y.-L. were responsible for data analysis and interpretation. H.L. provided administrative support. W.W. and Y.X.-D. drafted the first and revised manuscript with contributions from the other authors. All authors read and approved the final manuscript. 

Availability of Data and Materials: The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation. 

Supported by the Shanghai Committee of Science and Technology Research Projects (grant nos. 19140904600 and 18401900800) and Chinese Medicine Research Project of Shanghai Health Care Commission (grant nos. 2020 JQ001). 

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. 

Disclosure: W. Wang, None; J.-W. Zhang, None; Y.-J. Qin, None; H.-Y. Li, None; Y.-X. Dai, None; H. Li, None