Epigenetic Modulation of Thyroid-Stimulating Hormone Receptor Expression in Orbital Fibroblasts From Graves' Ophthalmopathy: Target for Treatment

Abstract

Purpose

This study aims to investigate the expression and function of DNA methyltransferases (DNMTs) and histone methyltransferase EZH2 in regulating the expression of thyroid-stimulating hormone receptor (TSHR) and fibrotic markers by orbital fibroblasts from Graves’ ophthalmopathy (GO) patients.

Methods

DNMTs mRNA levels were measured in orbital fibroblasts from healthy controls (control orbital fibroblasts), active GO fibroblasts, and inactive GO fibroblasts (iGOFs). TSHR and other fibrotic markers were measured in iGOFs treated with platelet-derived growth factor (PDGF)-BB and decitabine (a DNMT inhibitor) or DZNep (an EZH2 inhibitor) or small interfering RNA by RT-qPCR or Western blotting.

Results

DNMT1 and DNMT3A mRNA expressions were significantly higher in active GOFs compared with iGOFs and control orbital fibroblasts. Among all DNMTs, DNMT1 expression was at the highest levels. Upon PDGF-BB stimulation, DNMT1 and DNMT3A mRNA levels in iGOFs were significantly induced. Decitabine significantly inhibited PDGF-BB–induced cell viability and TSHR mRNA levels in iGOFs, whereas DNMT1 silencing partly reduced TSHR protein. Because a positive correlation between DNMT1 and EZH2 expression in iGOF was observed, the regulatory role of EZH2 on TSHR expression was also investigated. DZNep slightly decreased TSHR mRNA expression in iGOFs and iGO orbital tissue, whereas EZH2 silencing partly affects TSHR expression.

Conclusions

This study demonstrates that DNMT1 is the major DNMT during the active stage of GO. Moreover, DNMT1 is inducible upon PDGF-BB stimulation in iGOFs, suggesting an important role of PDGF-BB and DNMT1 during an active stage of GO. Our findings also highlighted that the inhibition of DNMT or EZH2 potentially suppressed TSHR mRNA expression, suggesting potential therapies targeting DNMT/EZH2 in controlling GO progression.

Graves’ ophthalmopathy (GO) is an autoimmune disease known as thyroid eye disease.^1,2 Several studies showed that GO orbital fibroblasts (GOFs) play an important role in GO pathogenesis as they express autoantigens, thyroid-stimulating hormone receptor (TSHR) and insulin-like growth factor 1 receptor, which can be stimulated by the circulating autoantibodies.^2–4 Moreover, platelet-derived growth factor (PDGF)-B was reported to be upregulated in GO orbital tissues in both inactive and active stages, as well as an increase in PDGF-receptor alpha expression by GOFs.^5,6 As a result, PDGF-BB–treated GOFs increased proliferation, TSHR expression, cytokines, and hyaluronan production and enhanced adipogenesis,^5,7,8 leading to orbital tissues inflammation and fibrosis.⁴ Common clinical features of GO are, therefore, presented with conjunctival chemosis, bilateral erythema, and proptosis, which could lead to severe symptoms such as corneal ulcers, optic neuritis, and loss of vision.²

DNA methyltransferases (DNMTs) are enzymes that induce the methylation of CpG-rich regions on the promoters to repress transcription. In contrast, although not fully understood, gene-body DNA methylation, especially on exons, correlates with transcriptional activation and alternative splicing.⁹ Several studies demonstrated that DNA methylation plays important roles in fibroblast activation in pulmonary,¹⁰ hepatic,¹¹ and renal fibrosis.¹² Moreover, our group and other groups recently revealed the role of DNA methylation in controlling several GOF activities^13,14 by demonstrating that global DNA methylation levels in GOF are significantly higher than in control orbital fibroblasts (COFs).¹³ Pathway analysis showed that TSHR and PDGF-BB signaling were reported with the highest score linking to hypermethylated genes in active GOFs (aGOFs).¹³ Therefore, the aim of this study was to investigate the role of DNMTs in regulating GOF activities through PDGF-BB activation.

Methods

Orbital Fibroblast Isolation

Orbital tissues were collected from Thai participants including 6 healthy controls who underwent plastic surgery, 3 patients with active GO, and 16 patients with inactive GO undergoing orbital decompression surgery at King Chulalongkorn Memorial Hospital. Informed consents were obtained from all participants. Ethical approval was obtained in accordance with the Declaration of Helsinki (2013 revision) from the Institutional Review Board of the Faculty of Medicine, Chulalongkorn University (Bangkok, Thailand), under the protocol number 401/61 (COA 1090/2023).

Orbital fibroblasts were isolated from healthy controls (COFs), patients with active GO (aGOFs), or patients with inactive GO (iGOFs), as described previously.¹⁵ Briefly, orbital tissues were cut into small pieces and placed onto six-well plates. The media containing Dulbecco's modified Eagle's medium (DMEM; GE Lifesciences, Westborough, MA, USA) with 20% fetal bovine serum (FBS; Life Technologies, Inc., Darmstadt, Germany) (20% FBS/DMEM), and 25 mg/mL plasmocin (InvivoGen, San Diego, CA, USA) were added onto the orbital tissues. The culture was then incubated at 37°C and 5% CO₂, and the media were refreshed twice a week. Isolated GOFs were maintained in 10% FBS/DMEM with 50 mg/mL gentamicin (Gibco, Shanghai, China). Cells at passages three to five were used for the experiments reported herein.

PDGF-BB–Stimulated Orbital Fibroblasts

COFs, aGOFs, or iGOFs were seeded at a density of 2.5 × 10⁵ cells per well in six-well plates with 1% FBS/DMEM containing 50 mg/mL gentamicin and incubated overnight to allow attachment. On the next day, the medium was refreshed in unstimulated condition, whereas PDGF-BB (50 ng/mL) (BioLegend, Inc., San Diego, CA, USA) was added for an additional 24 hours for gene expression, or 24 and 48 hours for protein studies.

MTT Assay

iGOFs were seeded at a density of 6 × 10³ cells per well in 96-well plates containing 1% FBS/DMEM and gentamicin and incubated overnight. Next, the cells were pretreated with decitabine at the concentrations of 2.5, 5.0, 10.0, or 15.0 µM for 24 hours. After that, cells were stimulated with 50 ng/mL PDGF-BB in the presence of decitabine for another 24 hours.

For MTT experiments, cells were incubated with 5 mg/mL thiazolyl blue tetrazolium bromide (Thermo Fisher Scientific Inc., Waltham, MA, USA) for 4 hours at 37°C and 5% CO₂. Subsequently, DMSO was added to dissolve formazan crystal from viable cells and the absorbance was measured at 540 nm. The percentage of cell viability and metabolism was calculated as indicated using the medium without cell as the blank and cells in unstimulated condition as negative controls (100% of cell viability and metabolism). A value of greater than 100% indicates an increase of cell viability and metabolism, whereas values of less than 100% indicate a reduction of cell viability and metabolism:

$$\begin{array}{ccc}& & \%\text{Cell}\text{viability}\text{and}\text{metabolism}\hfill \\ & & =\frac{(\text{OD}\text{test}\text{sample}-\text{OD}\text{blank})\times 100}{\text{OD}\text{negative}\text{control}-\text{OD}\text{blank}}\hfill \end{array}$$

Lactate Dehydrogenase (LDH) Assay

The culture supernatants of decitabine-treated iGO orbital tissues were collected to determine cytotoxicity by LDH assay according to the manufacturer's protocol (Biovision, Inc., Milpitas, CA, USA). Percentages of cytotoxicity were calculated and compared with negative control (culture supernatant from untreated condition) and positive control (culture supernatant from orbital tissue treated with lysis buffer). A 100% cytotoxicity indicates the highest cytotoxicity from positive control.

$$\begin{array}{ccc}& & \%\text{Cytotoxicity}\hfill \\ & & =\frac{\left(\text{OD}\text{of}\text{test}\text{sample}-\text{negative}\text{control}\right)\times 100}{\text{OD}\text{of}\text{positive}\text{control}-\text{negative}\text{control}}\hfill \end{array}$$

Decitabine- or DZNep-Treated iGOFs

iGOFs were seeded at a density of 2.5 × 10⁵ cells per well into six-well plates containing 1% FBS/DMEM and gentamicin and incubated overnight. Then, cells were pretreated with 2.5 µM decitabine or 6 µM DZNep for 24 hours. Subsequently, cells were stimulated with 50 ng/mL PDGF-BB in the presence of decitabine or DZNep for another 24 hours. Then, culture supernatants were collected to measure IL-6 levels using ELISA kits (BioLegend Inc.) according to the manufacturer's protocol. Total RNA was extracted from cells and proceeded to RT-qPCR.

Decitabine- or DZNep-Treated iGO Orbital Tissues

iGO orbital tissues were cut into small pieces and cultured in 1% FBS/DMEM and gentamicin with either 15 µM decitabine or 6 µM DZNep for 24 hours. Total RNA was extracted from the orbital tissues using Aurum Total RNA Fatty and Fibrous Tissue Kit (Bio-Rad Laboratories, Inc., Hercules, CA, USA), according to the manufacturer's protocol. cDNA was synthesized from total RNA and analyzed by RT-qPCR. TSHR/ABL mRNA expression of treated orbital tissues was compared with untreated tissues.

DNMT1- and EZH2-Silenced iGOFs

iGOFs were seeded at a density of 3 × 10⁵ cells per well in six-well plates for mRNA and protein expression studies. For MTT assay, cells were seeded at 6 × 10³ cells per well in 96-well plates. All cultures used 1% FBS/DMEM without antibiotics. We prepared 25 nM of human ON-TARGETplus small interfering RNA targeting DNMT1, EZH2, or a nontargeting control pool (Horizon Discover, UK) in transfection medium with a transfection reagent (Dharmacon, Inc., Lafayette, CO, USA) according to manufacturer's protocol and added to iGOF for 48 hours. Next, the cells were stimulated with PDGF-BB (50 ng/mL) in the presence of small interfering RNA for 24 hours (for MTT assay) or 48 hours (for protein extraction).

Gene Expression

Total RNA was extracted using the RNeasy Mini Kit (QIAGEN, Hilden, Germany) and cDNA was synthesized using the iScript cDNA Synthesis Kit (Bio-Rad Laboratories, Inc.) following the manufacturer's protocol. RT-qPCR was performed using TaqMan Gene Expression Assays with SsoAdvanced Universal Probes Supermix (Bio-Rad Laboratories, Inc.) by the CFX Connect real-time PCR detection system (Bio-Rad Laboratories, Inc.). The mRNA expression levels of DNMT1 (Hs00154749_m1), DNMT3A (Hs01027166_m1), DNMT3B (Hs00171876_m1), TSHR (Hs01053846_m1) (Applied Biosystems Inc., Waltham, MA, USA), and ACTA2 (forward primer, 5′-CACTGCCTTGGTGTGTGACAAT-3′; reverse primer, 5′-CGTAGCTGTCTTTTTGTCCCATTC-3′; and TaqMan Probe, FAM-TGTTTTCCCATCCATTGTGGGACGTC-TAMRA) were determined and normalized to ABL (forward primer, 5′ TGGAGATAACATCTAAGCATAACTAAAGGT-3′; reverse primer, 5′-GATGTAGTTGCTTGGGACCA-3′; and TaqMan Probe, FAM-CCATTTTTGGTTTGGGCTTCACACCATT-TAMRA).

Protein Expression

Total proteins were extracted from cells using 50 µL of 1× high-salt RIPA lysis buffer (50 mM Tris-HCl, pH 7.4, 500 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.5% sodium dodecyl sulfate) containing 1X protease/phosphatase inhibitor (Cell Signaling Technology, Inc., Danvers, MA, USA). Total protein concentration was determined by Pierce BCA Protein Assay Kit (Thermo Fisher Scientific Inc.). We used 30 µg of protein for sodium dodecyl sulfate polyacrylamide gel electrophoresis. After that, the proteins were transferred to PVDF membrane (Bio-Rad Laboratories, Inc.) using semidry transfer Trans-Blot SD (Bio-Rad Laboratories, Inc.). The following primary antibodies and secondary antibody from Cell Signaling Technology were used, including 1:1000 DNMT1 (Catalog no. 5032), 1:1000 EZH2 (Catalog no. 5246), 1:4000 β-actin (Catalog no. 4970), 1:1000 H3K27me3 (Catalog no. 9733), 1:4000 H3 (Catalog no. 4499), and 1:4000 mouse anti-rabbit IgG horseradish peroxidase-conjugate secondary antibody (Catalog no. 5127). 1:1000 TSHR (Catalog no. 14450-1-AP) was purchased from Proteintech (San Diego, CA, USA). Protein expression levels were detected using Alliance Q9 Chemiluminescence Imaging System (Uvitec Inc., Cambridge, UK) or Hyperfilm. The protein band intensities were quantified with ImageJ software.

Statistical Analysis

Statistical analysis in this study was performed using GraphPad Prism version 10.5.0. (GraphPad Software, San Diego, CA, USA). The differences among three or more groups were evaluated using one-way ANOVA followed by Bonferroni's multiple comparisons test and the differences between two groups were assessed by two-tailed paired Student t-tests. Correlation was tested using two-tailed Pearson's correlation test. Statistical significance was indicated with a P value of less than 0.05.

Results

DNMTs Were Differentially Expressed and Upregulated Upon PDGF-BB Stimulation

To investigate the levels of DNMT expression in orbital fibroblasts, DNMTs mRNA expression in aGOFs, iGOFs, and COFs were determined. The results showed that DNMT1 was expressed at the highest level compared with all DNMTs (Figs. 1A–C). Moreover, DNMT1, DNMT3A, and DNMT3B mRNA expression were higher in aGOFs compared with other groups (Figs. 1A–C).

Figure 1.

DNMTs expression in orbital fibroblasts. (A) DNMT1, (B) DNMT3A, and (C) DNMT3B mRNA expression levels in COFs (●), iGOFs (■), and aGOFs (▲). (D) DNMT1, (E) DNMT3A, and (F) DNMT3B mRNA expression levels in PDGF-BB–induced COFs and iGOFs. Each dot represents an individual patient or healthy control. Horizontal lines indicate mean and SEM. Statistical analysis was performed using one-way ANOVA followed by Bonferroni's multiple comparisons test. **P < 0.01, when compared between groups. #P < 0.05 and ##P < 0.01, respectively, when compared with the unstimulated condition.

We further investigated DNMTs expression in PDGF-BB–stimulated COFs and iGOFs at 24 hours, which represented a surrogate model for aGOFs. DNMT1 and DNMT3B mRNA levels were significantly upregulated in the PDGF-BB–stimulated condition compared with the unstimulated condition in COFs, whereas only DNMT1 showed significant upregulation in iGOFs (Figs. 1D, 1F). However, there were no significant differences between DNMT1 or DNMT3B mRNA expression levels between COFs and iGOFs (Figs. 1D, 1F). In contrast, PDGF-BB did not induce DNMT3A mRNA expression in either group (Fig. 1E). Therefore, PDGF-BB–stimulated iGOFs were used to represent aGOFs in this study.

Decitabine Inhibited TSHR mRNA Expression in PDGF-BB–Stimulated iGOFs and iGO Orbital Tissues

To evaluate the role of DNMTs in GOFs, PDGF-BB–activated iGOFs were treated with decitabine, a pan-DNMT inhibitor. A significant reduction in cell viability and metabolism was observed in dose-dependent manner (Fig. 2A). Decitabine at 2.5 µM, the lowest concentration that blocked PDGF-BB–induced cell viability and metabolism, was selected for further validation of DNMT function. However, decitabine did not affect ACTA2 mRNA levels or IL-6 production in PDGF-BB–stimulated iGOFs (Figs. 2B, 2C), but significantly decreased TSHR mRNA levels in PDGF-BB–induced iGOFs (Fig. 2D). Subsequently, the effect of the DNMT inhibitor on TSHR mRNA expression was examined in ex vivo experiments with iGO orbital tissues with decitabine at a concentration of 15 µM, the highest noncytotoxic concentration of decitabine on orbital tissues (data not shown). Upon decitabine treatment, TSHR mRNA levels showed a decreasing trend in iGO orbital tissues (Fig. 2E).

Figure 2.

The effect of decitabine on iGOFs and iGO orbital tissues. (A) Percentages of cell viability and metabolism in iGOFs after treatment with different concentrations of decitabine. The mRNA expression levels of (B) ACTA2 and (D) TSHR, as well as (C) IL-6 production levels by decitabine-treated iGOFs. (E) TSHR mRNA expression levels from decitabine-treated iGO orbital tissues. Each dot represents an individual patient, with horizontal lines indicating mean and SEM. Bar graphs show mean values. Statistical analysis was performed using one-way ANOVA followed by Bonferroni's multiple comparisons test (A) and two-tailed paired Student t-tests (B–E). *P < 0.05; ****P < 0.0001.

DZNep Decreased TSHR mRNA Expression in PDGF-BB–Stimulated iGOFs and iGO Orbital Tissues

Previous studies demonstrated aberrant histone modifications, particularly in the H3K27me3 position, induced by histone methyltransferase EZH2 in iGOFs.^15–17 Moreover, an interplay between DNA and histone methylation was reported.^18,19 This finding led to our hypothesis that histone methyltransferase EZH2 might also facilitate DNMT in regulating TSHR expression. Therefore, the correlation between the protein expression level of DNMT1 and EZH2 was first examined in PDGF-BB–stimulated iGOFs. Our results showed that DNMT1 positively correlated with EZH2 expression levels (Figs. 3A, 3B). Subsequently, DZNep was used to inhibit EZH2 in PDGF-BB–stimulated iGOFs and iGO orbital tissues. DZNep showed a trend toward a reduction in TSHR mRNA expression for both PDGF-BB–stimulated iGOFs (P = 0.13) (Fig. 3C) and iGO orbital tissues (P = 0.11) (Fig. 3D).

Figure 3.

The effect of DZNep on TSHR expression by iGOFs and iGO orbital tissues. (A) Positive correlation between DNMT1 and EZH2 protein expression in PDGF-BB–stimulated iGOFs. (B) Representative protein bands of DNMT1 and EZH2 in iGOFs. (C) TSHR mRNA expression in DZNep-treated, PDGF-BB–stimulated iGOFs. (D) TSHR mRNA expression in DZNep-treated iGO orbital tissues. Each dot represents an individual patient, with bar graphs showing mean values. Statistical analysis was performed using two-tailed Pearson's correlation test (A) and two-tailed paired Student t-tests (C and D).

DNMT1 and EZH2 Silencing Did Not Significantly Affect TSHR Protein Expression in iGOFs

Because decitabine significantly reduced TSHR mRNA expression and DNMT1 mRNA expressed at the highest level in PDGF-BB–stimulated iGOFs, DNMT1 silencing was performed to investigate its effects on PDGF-BB–induced TSHR expression. DNMT1 silencing condition, with a 90% reduction in DNMT1 protein expression (Supplementary Figs. S1A, S1C), had no significant effect on TSHR protein levels (Figs. 4A, 4C).

Figure 4.

The effect of DNMT1 and EZH2 knockdown on PDGF-BB–stimulated iGOFs. TSHR protein expression levels upon DNMT1 small interfering RNA (siRNA) (A) and EZH2 siRNA (B) treatment in the presence of PDGF-BB, with (C) representative protein bands. Each dot represents an individual patient; bar graphs indicate mean values. Statistical analysis was performed using two-tailed paired Student t-tests.

Because EZH2 protein expression positively correlated with DNMT1 levels and DZNep treatment showed a trend toward reduced TSHR mRNA levels, we hypothesized that EZH2 might also regulate TSHR in iGOFs. However, an EZH2 silencing condition in the presence of PDGF-BB stimulation, with a 69% reduction in EZH2 protein levels (Supplementary Figs. S1B, S1D), did not significantly alter TSHR protein levels (Figs. 4B, 4C). To further investigate the interplay between DNA and histone methylation, the H3K27me3 level was also measured. However, DNMT1 knockdown did not alter global H3K27me3 protein levels (Supplementary Figs. S2A, S2B).

Discussion

Our group previously demonstrated that global DNA methylation in GOFs was significantly higher than in COFs.¹³ This report suggested that DNA methylation play an important role in GO pathogenesis, especially in GOFs, although its mechanistic role in GOFs has never been investigated. Therefore, this study aimed to investigate the expression and functional roles of DNMTs in GO. Our data demonstrated that DNMTs expression were differentially expressed among GO stages, indicating the highest expression levels in aGOFs. These data align with our previous findings¹³ and also several reports on other fibrotic diseases that DNA methylation serves as an early biomarker. Moreover, upregulated DNMT expression and their activities correlate with active stage of the disease.^10,20 Nevertheless, decompression surgery in active GO patients is only performed when the condition becomes sight threatening, particularly with compressive optic neuropathy.²¹ Consequently, orbital tissues from active GO patients were limited in this study, and PDGF-BB–stimulated iGOFs were instead used as a representative model for aGOFs, even though this approach may not replicate the active microenvironment of the orbital tissues.

To identify functions of DNMTs in GOFs, decitabine, a pan-DNMT inhibitor, was used to treat PDGF-BB–activated iGOFs. Our data showed that decitabine reduced cell viability and metabolism and TSHR mRNA expression in PDGF-BB–stimulated iGOFs. This finding is in concordance with the role of DNMTs in regulating TSHR expression in thyroid cancer.²² Moreover, DNMT1 silencing confirmed its regulation of cell viability and metabolism, which is consistent with previous studies in several cell types (Supplementary Fig. S1E).^23–25 Unexpectedly, DNMT1 silencing did not significantly affect TSHR expression in this study. Conflicting data between decitabine and DNMT1 silencing suggests several possibilities including off-target effects from decitabine, post-transcriptional regulation by miRNA, differential translational regulation of the enzyme, optimization of silencing experiments and also the roles of other DNMTs within the family such as DNMT3A and 3B in regulating TSHR expression by GOF which require further investigation.

As interplay between DNA and histone methylation has been widely reported,^18,19 and EZH2 was reported by our group as an important histone methyltransferase regulating cell proliferation, hyaluronan, and cytokine production in iGOFs,¹⁶ EZH2 might also regulate TSHR expression. Our results showed that DZNep treatment reduced TSHR mRNA expression, and EZH2 silencing had no significant effect on TSHR protein levels. Several factors have been proposed to explain the discrepancy on the data between the inhibitor and gene silencing. In addition, other polycomb repressive complex 2 components, such as EZH1, SUZ12, and EED, might also contribute to H3K27me3 in regulating TSHR and other gene expression levels.²⁶⁻²⁸

To study the interplay between DNA and histone methylation in iGOFs in repressing gene expression, DNMT1 was silenced and H3K27me3 protein levels were assessed. However, the level of H3K27me3 remained unchanged. These data are consistent with a previous report that showed no changes in the level of global H3K27me3.²⁹ However, to identify precise TSHR regulatory regions affected by DNA and histone methylation that control TSHR expression levels in iGOFs, several OMICS analyses such as bisulfite and chromatin immunoprecipitation sequencing are needed in further study. In addition, we observed that the responses upon DNMT and EZH2 silencing were patient specific, suggesting complexity in epigenetic regulation in controlling TSHR expression in each individual.

Conclusions

Our findings provide supportive evidence and highlight the possible involvement of DNMTs and EZH2 in regulating TSHR autoantigen in GO. Thus, we proposed that DNMTs and EZH2 represent as potential therapeutic targets for GO treatment.