Changes in the DNA Methylation and Hydroxymethylation Status of the Intercellular Adhesion Molecule 1 Gene Promoter in Thyrocytes from Autoimmune Thyroiditis Patients

Abstract

Background:

The intercellular adhesion molecule 1 (ICAM1) gene is not expressed in normal thyroid tissue but was shown to be expressed in the thyroid tissue of autoimmune thyroiditis (AIT) patients.

Methods:

This study aimed to explore whether the DNA methylation and hydroxymethylation status of the ICAM1 promoter are aberrantly altered in the thyroid cells of AIT patients and whether this change is associated with dysfunctional expression of ICAM1. A total of 35 AIT patients and 35 sex- and age-matched controls were studied. After the isolation of thyrocytes via density-gradient centrifugation, ICAM1 mRNA expression was measured using real-time PCR. The DNA methylation and hydroxymethylation status were assessed using quantitative PCR following T4 β-glucosyltransferase treatment and MspI/HpaII cleavage at −937 bp, −701 bp, −226 bp, and −65 bp upstream of the transcription start site (TSS). The DNA methylation level was verified via pyrosequencing.

Results:

The AIT group showed increased DNA hydroxymethylation at −937 bp and −226 bp and decreased methylation at −937 bp, −701 bp, and −226 bp upstream of the TSS. Pyrosequencing also revealed DNA hypomethylation at −708 bp, −692 bp, −690 bp, and −688 bp upstream of the TSS. The DNA methylation status at −708 bp, −692 bp, and −226 bp upstream of the TSS was negatively associated with ICAM1 mRNA expression.

Conclusion:

In summary, we identified aberrant DNA methylation and hydroxymethylation of the ICAM1 gene promoter in the thyrocytes of AIT patients. This aberrant epigenetic modification is associated with increased expression of the ICAM1 gene.

Introduction

Autoimmune thyroid disease (AIT) is a common disease, characterized by positivity for thyroid autoantibodies, infiltration of lymphocytes in the thyroid, and aberrant expression of immune-related genes in thyrocytes. Recent epidemiological data indicated that the prevalence of positivity for thyroid autoantibodies (anti-thyroid peroxidase [TPOAb] and anti-thyroglobulin [TgAb]) is 9–10% in China. The thyroid cells of AIT patients express immune-related molecules such as HLA-DR (1), ICAM1 (2), CD40, and FAS/FASL (3). Genetic analyses, including linkage studies and association studies, have identified several candidate AIT-related genes, including the HLA family, CTLA4, CD40, PTPN22, FOXP3, CD25, and thyroid-specific genes (4). Recently, several new genes were identified in young-age-of-onset Graves' disease (GD) patients in addition to the known GD susceptibility loci, including BTNL2, NOTCH4, TNFAIP3, and CXCR4 (5). However, these candidate genes are not associated with a significantly increased risk for the disease, which indicates that factors other than genetic factors, such as epigenetic modifications, affect the disease process (6). Previous studies have found increased expression of the intercellular adhesion molecule 1 (ICAM1) gene in the thyroid tissue of AIT patients. To the best of our knowledge, there was insufficient evidence demonstrating an association between AIT and SNPs of the ICAM1 gene (7). Clinical observation and animal experimentation showed that the expression of ICAM1 can be enhanced by interferon-gamma (IFN-γ), tumor necrosis factor alpha, excessive iodine, and other environmental factors (8,9). These findings indicate that the ICAM1 gene could be influenced by environmental factors and aberrantly changed by epigenetic modification.

DNA methylation regulates tissue-specific gene expression patterns by silencing the expression of function-irrelevant genes. DNA hydroxymethylation is a recently discovered type of DNA epigenetic modification. In addition to passive demethylation during DNA duplication, hydroxymethylation is a new mechanism of active demethylation that converts 5-methylcytosine (5-mc) to 5-hydroxymethylcytosine (5-hmc), followed by conversion to unmethylated cytosine by the ten-eleven-translocase (TET) family (10). Hydroxymethylation was initially found to be concentrated in embryonic cells and neurons and was subsequently shown to be distributed among all cell types and to be involved in various biological processes (11,12). In contrast to DNA methylation, promoter hydroxymethylation is related to transcriptional activation (13). Notably, DNA methylation and hydroxymethylation cannot be discriminated through traditional bisulfite sequencing, except via oxidation or glycosylation (14,15). Aberrant changes in methylation and hydroxymethylation have been observed in multiple diseases, including various cancers and autoimmune disorders (16,17).

This study aimed to explore the DNA methylation and hydroxymethylation status of the ICAM1 promoter. Polymorphisms of DNA methyltransferase genes and the whole-genome DNA methylation pattern of the CD4+ and CD8+ T cells of autoimmune thyroid disease patients have been studied previously (18,19). The DNA methylation status of the ICAM1 gene in CD4+ T cells was also shown to be associated with GD in genomic studies. However, the expression and DNA methylation patterns are tissue specific, and epigenetic modification of the target organ in thyroid disease has not yet been elucidated. This study analyzed the DNA modification of a specific gene in thyrocytes and focused on AIT patients for the first time.

Methods and Materials

Subjects

The study was approved by the Ethics Committee of the First Affiliated Hospital of China Medical University. The subjects underwent surgery for thyroid adenomas or nodules, or for a partial thyroidectomy to relieve tracheal compression caused by goiter in Hashimoto thyroiditis. Written, informed consent was obtained before surgery. Serum samples were also obtained before surgery to test thyroid function and detect autoantibodies, which were assessed using an electrochemiluminescence method (Architect i2000SR, Abbott Laboratories). Thyroid tissue was obtained through surgical excision and immediately placed on ice. All the patients included in the case group were females who showed positivity for at least one thyroid autoantibody (TPOAb or TgAb) and received a pathologic diagnosis of Hashimoto disease. Age- and sex-matched controls negative for the presence of autoantibodies were included. Thyroid tissue for the control group was obtained from para-adenoma tissue. All of the subjects were euthyroid and not exposed to radioactive iodine, antithyroid drugs, thyroid hormone replacement, and they did not have systemic autoimmune disease or any severe illnesses. Subjects with malignant nodules were excluded. Subjects who had positive thyroid autoantibodies but were not diagnosed with Hashimoto thyroiditis in surgical pathology were also excluded. The baseline clinical characteristics are summarized in Table 1.

Table 1.

Baseline Clinical Characteristics of the Study Subjects

	AIT group (n = 35)	Control group (n = 35)	Reference interval ^a	p-Value^**
Age (years)	52.20 ± 3.00	55.61 ± 1.77		0.17
Thyroglobulin antibody (IU/mL)	79.19 (21.86, 296.88)	1.19 (1.04, 3.03)	0–4.11	<0.001^**
Thyroid peroxidase antibody (IU/mL)	99.65 (4.83, 888.24)	0.31 (0.21, 0.94)	0–5.61	<0.001^**
Thyrotropin (mIU/L)	1.69 (0.60, 2.79)	1.19 (0.62, 1.50)	0.35–4.94	0.128
Thyroxine (pmol/L)	14.12 ± 1.17	14.33 ± 0.45	9.01–19.05	0.62
Triiodothyronine (pmol/L)	3.81 ± 0.78	4.60 ± 0.80	2.63–5.70	0.045^**

Method: electrochemiluminescence method (Architect i2000SR, Abbott Laboratories, Chicago, IL).

p-Value <0.05 was considered to represent a statistically significant difference.

Isolation of thyrocytes

Thyrocytes were isolated via Percoll density-gradient centrifugation as previously described (20) with minor modifications. Fresh thyroid tissue obtained during surgery was minced into 1 × 1 × 1 mm pieces and incubated with 1.25 mg/mL collagenase IV (Sigma, C1889) in Hank's solution at 37°C. The digested materials were filtered through a nylon mesh with a 75 μm pore size every 30 minutes, and undigested fragments were collected to repeat the digestion procedure. The filtrate was centrifuged at 100 g for 5 minutes at 4°C and resuspended in 0.25% Trypsin-EDTA. A single-cell suspension was obtained via incubation at 37°C for 10 min. Enzyme activity was stopped by the addition of RPMI-1640 medium containing fetal bovine serum. The single-cell suspension was then gently transferred to a three-layer discontinuous Percoll gradient. The densities were 1.030 g/mL, 1.061 g/mL, and 1.077 g/mL from top to bottom, which corresponded to 20%, 45%, and 60% Percoll, respectively. After centrifugation at 400 g for 25 minutes at 4°C, thyrocytes sedimented between the first and second layers. The purity of the thyrocytes was tested via flow cytometry by staining for thyroglobulin. Thyrocyte preparations with a purity greater than 90% were used for experiments. The thyroglobulin antibody was purchased from Abcam (cat. ab76765). Subjects with insufficient tissue for thyrocyte isolation or a thyroid cell purity of less than 90% were excluded.

RNA and DNA preparation

DNA was extracted via the phenol–chloroform method. RNA was extracted using the TRIzol reagent (Thermo, 15596-026). The concentration and 260/280 OD value were determined using a NanoDrop 2000. All of the 260/280 OD values for the DNA and RNA samples were between 1.7 and 1.9 or 1.8 and 2.0, respectively. Reverse transcription of RNA was performed using PrimeScript RT Master Mix (TAKARA, RR036). Genomic DNA and cDNA samples were stored at −20°C until analysis.

DNA treatment

The DNA methylation and hydroxymethylation status of the ICAM1 promoter was analyzed as described in the manufacturer's instructions for the EpiMark 5-hmC and 5-mC Analysis Kit (E3317, NEB). Briefly, the genomic DNA was divided into two equal aliquots, one of which was treated with T4 β-glucosyltransferase in 80 mM uridine diphosphate–glucose at 37°C for 18 hours, while the control was treated with ddH₂O. Each of these aliquots was then divided into three equal aliquots and treated with MspI/HpaII or no enzyme at 37°C for 16 hours. The MspI enzyme can recognize and cleave all CCGG sequences except for glycosylated 5-hmc, while HpaII can recognize only unmodified cytosines.

Quantitative PCR

The primers for the detection of ICAM1 mRNA levels and 5-mc or 5-hmc levels in the ICAM1 promoter region were designed using Primer3 (http://primer3.sourceforge.net/) and are listed in Table 2. The primers for 5-mc and 5-hmc detection were designed to flank the CCGG sequences at −937 bp, −701 bp, −226 bp, and −65 bp relative to the transcription start site (TSS). Real-time PCR was carried out using SYBR Premix Ex Taq (TAKARA, RR820) and a LightCycler 480 (Roche) and was repeated at least three times. If there was an obvious outlier in the three repeats, the outlier was deleted. If not, the average cycle number was used for subsequent quantification. The mRNA level was measured via relative quantification using GAPDH as a reference. The levels of 5-mC and 5-hmC at CCGG sites were calculated according to the formula provided by the manufacturer.

Table 2.

Primer Sequence

Target sequence	Forward	Reverse
ICAM1 mRNA	AGAGACCCCGTTGCCTAAAA	CAGTACACGGTGAGGAAGGT
Site 1 in ICAM1 promoter	ATATGCGTCTCTGGATGGCC	TGACCCCAAGACTCGACATC
Site 2 in ICAM1 promoter	GATGTCGAGTCTTGGGGTCA	TTGGATCTTGACGAGGGAGG
Site 3 in ICAM1 promoter	GAAATGCCCGTGTCAGCTAG	AGTCCTCCCTCTCCAGCTTA
Site 4 in ICAM1 promoter	TAAGCTGGAGAGGGAGGACT	GGCTAAGCTTGAATCACGGT

Pyrosequencing primer	Forward	Reverse	Sequence
Site 2 in ICAM1 promoter	TTTGGTGTTAGGAGATTTAAGAGTAT	AAAACAACAACCCCCAAAC	ATTAGTTTATATTTAGTGTATGAG
Site 3 in ICAM1 promoter	AGGGGGAGGGGTATTTTTTA	AAATCCTCCCTCTCCAACTTAC	TTTAGTGGAGGGAGT

Pyrosequencing

To validate the DNA methylation status measured using the EpiMark Kit, bisulfite pyrosequencing was performed for 20 subjects randomly selected from each group. The primers were designed using PyroMark Assay Design 2.0 and are listed in Table 2. The genomic DNA was reacted with bisulfite and analyzed on a PyroMark Q96 ID (Qiagen). The results were evaluated with Pyro Q-CpG software.

Statistics

The distribution normality was tested via the Kolmogorv-Smirnov method. Age and serum thyrotropin, free thyroxine, free triiodothyronine, and ICAM1 mRNA expression levels were normally distributed. These data were analyzed with Student's t test and the results are expressed as mean ± standard error. The data on TPOAb and TgAb and the percentages of 5-mc and 5-hmc within the ICAM1 promoter failed the normality test. These variables were analyzed with the Mann–Whitney U test and the results are expressed as medians with interquartile ranges. Correlation analysis was carried out to examine the association between the ICAM1 mRNA level and the 5-mc or 5-hmc level. The analysis was conducted with the AIT group and controls. For the cases with undetectable levels of 5-hmc or 5-mc, the outliers were excluded. Statistical analysis was carried out using SPSS 21 software. p-Values less than 0.05 were regarded as statistically significant.

Results

ICAM1 mRNA expression levels

The ICAM1 mRNA level was calculated relative to that of GAPDH mRNA. As shown in Figure 1, the ICAM1 mRNA expression level in the AIT group was approximately 80% higher than that in the control group [0.053 ± 0.010 vs. 0.089 ± 0.015 (mean ± standard error), p = 0.046].

FIG. 1.

ICAM1 mRNA expression. (a) Relative expression levels of ICAM1 mRNA in the autoimmune thyroid disease (AIT) (n = 35) and control groups (n = 35) (mean ± standard error [SE]; *p < 0.05) shown in scatter plot. ICAM1 mRNA levels were calculated relative to glyceraldehyde 3-phosphate dehydrogenase mRNA. (b) Relative expression levels of ICAM1 shown in bar chart.

Methylation and hydroxymethylation status of the ICAM1 promoter

The DNA methylation status and hydroxymethylation status at positions 1 (−937 bp), 2 (−701 bp), 3 (−226 bp), and 4 (−65 bp) upstream of the TSS were tested with an EpiMark Kit and expressed as the percentages of 5-mc and 5-hmc. The percentage of 5-hmc at positions 1 and 3 increased significantly in the AIT group (Fig. 2a, c). The median percentages of 5-hmc (with the interquartile range) at positions 1 and 3 were 6.41% (0, 33.61%) and 1.49% (0, 30.98%) in the AIT group and 0% (0, 0.48%) and 1.46% (0, 5.31%) in the control group, respectively (p < 0.05). The percentages of 5-hmc at the other positions did not show a statistically significant difference. The methylation levels at positions 1, 2, and 3 were decreased significantly in the AIT group (Fig. 2e–g). The median percentages of 5-mc (with the interquartile range) were 0% (0, 2.98%), 0% (0, 20.21%), and 14.89% (0, 34.51%) in the AIT group and 3.21% (0, 18.66%), 29.13% (0, 61.44%), and 56.74% (33.96%, 76.97%) in the control group, respectively (p < 0.05). The methylation levels at the other positions did not show a statistically significant difference. In the subjects selected for pyrosequencing verification, all of the tested positions (−708 bp, −692 bp, −690 bp, −688 bp, and −226 bp upstream of the TSS) showed an increased methylation level in the AIT group (Fig. 2i). This change was statistically significant. The relative positions of the methylated or hydroxymethylated CpG sites and transcription factor binding sites are shown in Figure 3, and the results are summarized in Table 3.

FIG. 2.

ICAM1 promoter methylation and hydroxymethylation status. (a, b, c, d) Percentages of 5-hydroxymethylcytosine (5-hmc) at CCGG sequences 1–4 in the AIT and control groups (median with interquartile, *p < 0.05). (e, f, g, h) Percentages of 5-mc at CCGG sequences 1–4 in the AIT and control groups (median with interquartile, *p < 0.05). (i) Pyrosequencing results validating the percentages of 5-methylcytosine (5-mc) at CG sites 2, 2a, 2b, 2c, and 2d in the AIT (n = 20) and control (n = 20) groups (mean ± SE, *p < 0.05).

FIG. 3.

The tested sites in the ICAM1 promoter. (a) Schematic representation of the promoter region of the ICAM1 gene. Arrows indicate the transcription start site (TSS) and coding sequence (CDS) start site. The transcription factor binding sites, the relative distance to the TSS, and the position codes of the tested sites are also indicated. Positions 1, 2, 3 and 4 are located at −937 bp, −701 bp, −226 bp and −65 bp upstream of the TSS, respectively. The methylation and hydroxymethylation status was tested using an EpiMark Kit. Positions 2a, 2b, 2c, and 2d are located −708 bp, −692 bp, −690 bp and −688 bp upstream of the TSS, respectively. The number 2 positions and their methylation status were verified through pyrosequencing. (b) Sequence of the ICAM1 promoter. The sequences of the sites tested for their methylation or hydroxymethylation status are indicated in bold. The transcription factor binding sites are indicated in gray. The arrows indicate the TSS and CDS start site.

Table 3.

ICAM1 mRNA Expression Level (Mean ± SE) and the 5-mc and 5-hmc Level in ICAM1 Promoter (Median with Interquartile)

DNA modification ^a	AIT group	Control group	p-Value^**
ICAM1 mRNA	0.089 ± 0.015	0.053 ± 0.010	<0.05^**
5-hmc at 1 site 1	6.41% (0, 33.61%)	0% (0, 0.48%)	<0.05^**
5-hmc at 2 site 2	0.22%(0, 16.87%)	2.01%(0, 43.23%)	0.85
5-hmc at 3 site 3	1.49% (0, 30.98%)	1.46% (0, 5.31%)	<0.05^**
5-hmc at 4 site 4	0%(0, 1.27%)	0%(0, 0.66%)	0.74
5-mc at 1 site 1	0% (0, 2.98%)	3.21% (0, 18.66%)	<0.05^**
5-mc at 2 site 2	0% (0, 20.21%)	29.13% (0, 61.44%)	<0.05^**
5-mc at 3 site 3	14.89% (0, 34.51%)	56.74%(33.96%,76.97%)	<0.05^**
5-mc at 4 site 4	0.85%(0, 4.86%)	1.90%(0, 4.01%)	0.86
5-mc at 2a site 2a	7.69%(6.12%,11.92%)	11.61%(10.92%,13.87%)	<0.05^**
5-mc at 2b site 2b	20.16%(12.68%,23.18%)	22.71%(20.41%,29.06%)	<0.05^**
5-mc at 2c site 2c	7.01% (5.47%, 9.01%)	9.82% (8.89%, 10.78%)	<0.05^**
5-mc at 2d site 2d	15.65%(13.79%,22.54%)	24.75%(22.70%,27.57%)	<0.05^**

#1 site is at −937bp upstream of TSS, #2 site is at −701bp upstream of TSS, #3 site is at −226bp upstream of TSS. #2a, #2b, #2c, #2d site are at −708 bp, −692 bp, −690 bp and −688 bp upstream of the TSS respectively.

p < 0.05 was considered to represent a statistically significant difference.

Correlation analysis was conducted between ICAM1 mRNA levels and 5-hmc or 5-mc percentages. The correlation analysis was conducted with the AIT and the control group. Since the 5-hmc and 5-mc percentages failed the normality distribution test, the Spearman correlation analysis for not normally distributed parameters was used. Cases with undetectable 5-hmc and 5-mc, were regarded as outliers and not included in the analysis. The methylation levels at positions 2a, 2b, and 3 showed moderate negative association with ICAM1 mRNA expression. The corresponding Spearman's correlation coefficients were r = −0.445 (p = 0.015, n = 40), r = −0.381 (p = 0.042, n = 40) and r = −0.401 (p = 0.031, n = 40), respectively.

Discussion

In this study, we found an abnormal decrease in DNA methylation and an increase in DNA hydroxymethylation at the ICAM1 promoter; these changes were associated with ICAM1 mRNA expression. The CpG sites that we tested in the ICAM1 promoter, from −931 bp to −226 bp upstream of the TSS, showed hyperhydroxymethylation and hypomethylation. The loss of DNA methylation at these sites in the thyroid of AIT patients was demonstrated via a glycosylation and enzyme digestion–based method and pyrosequencing. Enzymatic digestion after glycosylation is capable of discriminating 5-hmc from 5-mc. Our findings provide the first evidence regarding the DNA hydroxymethylation status of AIT patients, specifically showing that the level of 5-hmc modification at −921 bp and −226 bp upstream of the TSS is elevated. The DNA methylation observed at −708 bp, −692 bp and −226 bp upstream of the TSS was decreased and showed a correlation with ICAM1 mRNA levels. However, the hydroxymethylation status did not show any association with mRNA expression. The reason for this lack of association may lie in the fact that hydroxymethylation is a relatively temporary modification of cytosine. Additionally, the hyperhydroxymethylation did not interfere with the pyrosequencing results showing hypomethylation at −931 bp and −226 bp upstream of the TSS. The lack of abundance of 5-hmc may explain why the percentage of 5-hmc at the ICAM1 promoter is generally less than 10% and is significantly lower than that of 5-mc in a given patient.

ICAM1 is expressed on the surface of immunological cells and mediates homing and trafficking of these cells to target organs. It also plays an important role in cell interactions during antigen presentation or cell signaling transduction. In physiological conditions, ICAM1 is expressed at a low concentration on thyrocyte membranes and regulates the localization of immunological cells in thyroid tissue. ICAM1 expression is upregulated in thyrocytes of AIT patients (2). It acts as the binding site for inflammatory cells and enhances their recruitment to the thyroid at an early stage of the disease. ICAM1 gene expression can be increased by IFN-γ, excessive iodine and other environmental factors (8,21,22). Recently, several studies have found that increased soluble ICAM1 (sICAM1) in circulation is a hallmark of Hashimoto thyroiditis and GD (23 –25). However, the mechanism of ICAM1 overexpression in thyrocytes has not been elucidated.

Epigenetic regulation links environmental factors and gene expression, which could be one of the mechanisms involved in enhancing thyroiditis. The loss of promoter hypermethylation could activate previously silenced immune-related genes in thyrocytes. The pathways for DNA demethylation include passive demethylation during replication and active demethylation via oxidation by the TET family. A recent study demonstrated that TET2 can induce hypomethylation in the ICAM1 promoter and enhance ICAM1 gene expression in hepatocytes (26,27). We identified DNA hyperhydroxymethylation and hypomethylation in the ICAM1 promoter in AIT thyrocytes. Since DNA demethylation always leads to transcriptional activation (28), we hypothesized that the hypermethylated cytosine site is then hydroxymethylated by TET and converted to demethylated cytosine, leading to ICAM1 gene expression.

To the best of our knowledge, this is the first report on the DNA modification of a specific gene in thyrocytes from AIT patients. The involvement of epigenetic mechanisms in autoimmune thyroid disease has been reported by previous studies. In addition to the studies concentrating on DNA methylation, Stefan et al. (29,30) discovered that the transcription factor binding motifs in the promoter regions of the thyroglobulin and thyroid stimulating hormone receptor genes were enriched with mono- and trimethylated Lys-4 residues of histone H3 (H3K4me1), which is a type of active chromatin modification that enables transcription factor binding and gene expression. In addition, genomic H3K4me1 enrichment can be enhanced after IFN-γ treatment. Similar studies have identified an association of miRNA expression with autoimmune thyroid disease (31 –34). The data reported here demonstrate DNA modification through hydroxymethylation of the ICAM1 gene for the first time in AIT research. However, there are some limitations in our study. First, this is a cross-sectional observation. Correlation analysis indicate an association between the DNA methylation and mRNA expression of ICAM1, but this is not a strong evidence for a causal relationship. Second, our study focused on the DNA hydroxymethylation and methylation status of one specific gene. However, the biological behavior of a specific type of cell is regulated by the epigenetic modification and gene expression at the whole-genome level. Further studies are needed to elucidate the DNA modification pattern of the thyroid in AIT patients at the genomic scale.

In summary, our findings show that hydroxymethylation and methylation of the ICAM1 promoter was aberrantly changed and the methylation status is associated with expression of ICAM1 mRNA. Our results highlight the importance of studying epigenetics to elucidate the mechanism of environmental influence in disease onset. Further studies are needed to determine how risk factors affect the disease process of AIT at the genomic scale.

Footnotes

Acknowledgments

This work was supported by the National Natural and Science Foundation of China (grant number 81430018). We thank all the volunteers for their participation in this study and the donation of samples. We also thank Yushu Li, Haixia Guan, and Jing Li, physicians from the Department of Endocrinology and Metabolism of the First Affiliated Hospital of China Medical University, for their comments and advice.

Author Disclosure Statement

No competing financial interests exist.

References

Jacobson

, Huber

, Tomer

. 2008. The HLA gene complex in thyroid autoimmunity: from epidemiology to etiology. J Autoimmun, 30:58–62.

Weetman

, Cohen

, Makgoba

, Borysiewicz

. 1989. Expression of an intercellular adhesion molecule, ICAM-1, by human thyroid cells. J Endocrinol, 122:185–191.

Jacobson

, Tomer

. 2007. The CD40, CTLA-4, thyroglobulin, TSH receptor, and PTPN22 gene quintet and its contribution to thyroid autoimmunity: back to the future. J Autoimmun, 28:85–98.

Lee

, Li

, Hammerstad

, Stefan

, Tomer

. 2015. Immunogenetics of autoimmune thyroid diseases: a comprehensive review. J Autoimmun, 64:82–90.

Brown

, Lombardi

, Hasham

, Greenberg

, Gordon

, Concepcion

, Hammerstad

, Lotay

, Zhang

, Tomer

. 2014. Genetic analysis in young-age-of-onset Graves' disease reveals new susceptibility loci. J Clin Endocrinol Metab, 99:E1387–E1391.

Tomer

. 2014. Mechanisms of autoimmune thyroid diseases: from genetics to epigenetics. Annu Rev Pathol, 9:147–156.

Akman

, Kanmaz-Ozer

, Vural

, Ozderya

, Karadag

, Dogru-Abbasoglu

, Uysal

. 2012. G241R and K469E polymorphisms of intercellular adhesion molecule 1 (ICAM-1) could predispose to Hashimoto thyroiditis. Mol Biol Rep, 39:10723–10729.

Yamazaki

, Tanigawa

, Suzuki

, Yamada

, Takano

, Obara

, Sato

. 2010. Iodide-induced chemokines and genes related to immunological function in cultured human thyroid follicles in the presence of thyrotropin. Thyroid, 20:67–76.

Sharma

, Traore

, Trush

, Rose

, Burek

. 2008. Intracellular adhesion molecule-1 up-regulation on thyrocytes by iodine of non-obese diabetic H2(h4) mice is reactive oxygen species-dependent. Clin Exp Immunol, 152:13–20.

10.

Ito

, D'Alessio

, Taranova

, Hong

, Sowers

, Zhang

. 2010. Role of Tet proteins in 5mC to 5hmC conversion, ES-cell self-renewal and inner cell mass specification. Nature, 466:1129–1133.

11.

Koh

, Yabuuchi

, Rao

, Huang

, Cunniff

, Nardone

, Laiho

, Tahiliani

, Sommer

, Mostoslavsky

, Lahesmaa

, Orkin

, Rodig

, Daley

, Rao

. 2011. Tet1 and Tet2 regulate 5-hydroxymethylcytosine production and cell lineage specification in mouse embryonic stem cells. Cell Stem Cell, 8:200–213.

12.

, Liu

. 2011. Distribution of 5-hydroxymethylcytosine in different human tissues. J Nucleic Acids, 2011:870726.

13.

Pastor

, Pape

, Huang

, Henderson

, Lister

, Ko

, McLoughlin

, Brudno

, Mahapatra

, Kapranov

, Tahiliani

, Daley

, Liu

, Ecker

, Milos

, Agarwal

, Rao

. 2011. Genome-wide mapping of 5-hydroxymethylcytosine in embryonic stem cells. Nature, 473:394–397.

14.

Huang

, Pastor

, Shen

, Tahiliani

, Liu

, Rao

. 2010. The behaviour of 5-hydroxymethylcytosine in bisulfite sequencing. PLoS One, 5:e8888.

15.

Booth

, Branco

, Ficz

, Oxley

, Krueger

, Reik

, Balasubramanian

. 2012. Quantitative sequencing of 5-methylcytosine and 5-hydroxymethylcytosine at single-base resolution. Science, 336:934–937.

16.

Taylor

, Smeriglio

, Dhulipala

, Rath

, Bhutani

. 2014. A global increase in 5-hydroxymethylcytosine levels marks osteoarthritic chondrocytes. Arthritis Rheumatol, 66:90–100.

17.

Sui

, Tan

, Yang

, Yan

, Lin

, Ou

, Xue

, Chen

, Zou

, Jing

, Guo

, Cao

, Sun

, Cui

, Dai

. 2015. Genome-wide analysis of 5-hmC in the peripheral blood of systemic lupus erythematosus patients using an hMeDIP-chip. Int J Mol Med, 35:1467–1479.

18.

Limbach

, Saare

, Tserel

, Kisand

, Eglit

, Sauer

, Axelsson

, Syvanen

, Metspalu

, Milani

, Peterson

. 2016. Epigenetic profiling in CD4+ and CD8+ T cells from Graves' disease patients reveals changes in genes associated with T cell receptor signaling. J Autoimmun, 67:46–56.

19.

Cai

, Muhali

, Song

, Qin

, Wang

, Shi

, Jiang

, Xiao

, Li

, Zhang

. 2015. Genome-wide DNA methylation analysis in Graves' disease. Genomics, 105:204–210.

20.

Baur

, Wall

, Schleusener

. 1983. Isolation and characterization of mononuclear cells from various thyroid tissue specimens. Life Sci, 32:55–65.

21.

Kostic

, Toffoletto

, Fontanini

, Moretti

, Cesselli

, Beltrami

, Ambesi Impiombato

, Curcio

. 2009. Influence of iodide excess and interferon-gamma on human primary thyroid cell proliferation, thyroglobulin secretion, and intracellular adhesion molecule-1 and human leukocyte antigen-DR expression. Thyroid, 19:283–291.

22.

Burek

, Rose

. 2008. Autoimmune thyroiditis and ROS. Autoimmun Rev, 7:530–537.

23.

Beumer

, Effraimidis

, Drexhage

, Wiersinga

, Drexhage

. 2013. Changes in serum adhesion molecules, chemokines, cytokines, and tissue remodeling factors in euthyroid women without thyroid antibodies who are at risk for autoimmune thyroid disease: a hypothesis on the early phases of the endocrine autoimmune reaction. J Clin Endocrinol Metab, 98:2460–2468.

24.

Jublanc

, Beaudeux

, Aubart

, Raphael

, Chadarevian

, Chapman

, Bonnefont-Rousselot

, Bruckert

. 2011. Serum levels of adhesion molecules ICAM-1 and VCAM-1 and tissue inhibitor of metalloproteinases, TIMP-1, are elevated in patients with autoimmune thyroid disorders: relevance to vascular inflammation. Nutr Metab Cardiovasc Dis, 21:817–822.

25.

Ozderya

, Aydin

, Temizkan

, Abbasoglu

, Vural

, Altuntas

. 2016. High circulating levels of sICAM-1 and sVCAM-1 in the patients with Hashimoto's thyroiditis. Endocr Res, 2016:1–7.

26.

Fan

, Zhang

, Pascuzzi

, Andrisani

. 2016. Hepatitis B virus X protein induces EpCAM expression via active DNA demethylation directed by RelA in complex with EZH2 and TET2. Oncogene, 35:715–726.

27.

Chen

, Kazemier

, de Groote

, Ruiters

, Xu

, Rots

. 2014. Induced DNA demethylation by targeting Ten-Eleven Translocation 2 to the human ICAM-1 promoter. Nucleic Acids Res, 42:1563–1574.

28.

Han

, Cortez

, Yang

, Nichols

, Jones

, Liang

. 2011. DNA methylation directly silences genes with non-CpG island promoters and establishes a nucleosome occupied promoter. Hum Mol Genet, 20:4299–4310.

29.

Stefan

, Wei

, Lombardi

, Li

, Concepcion

, Inabnet

3rd , Owen

, Zhang

, Tomer

. 2014. Genetic-epigenetic dysregulation of thymic TSH receptor gene expression triggers thyroid autoimmunity. Proc Natl Acad Sci USA, 111:12562–12567.

30.

Stefan

, Jacobson

, Huber

, Greenberg

, Li

, Skrabanek

, Conception

, Fadlalla

, Ho

, Tomer

. 2011. Novel variant of thyroglobulin promoter triggers thyroid autoimmunity through an epigenetic interferon alpha-modulated mechanism. J Biol Chem, 286:31168–31179.

31.

Wang

, Fan

, Zhang

, Lin

, Lu

, Bai

, Li

, Zhao

, Zhang

. 2014. Integrative analysis of mRNA and miRNA array data reveals the suppression of retinoic acid pathway in regulatory T cells of Graves' disease. J Clin Endocrinol Metab, 99:E2620–E2627.

32.

Liu

, Ma

, Xu

, Wang

, Jiang

, Zhu

, Cui

, Ning

, Lin

, Wang

. 2012. Differential microRNA expression in peripheral blood mononuclear cells from Graves' disease patients. J Clin Endocrinol Metab, 97:E968-E972.

33.

Yamada

, Itoh

, Hiratsuka

, Hashimoto

. 2014. Circulating microRNAs in autoimmune thyroid diseases. Clin Endocrinol, 81:276–281.

34.

Kagawa

, Watanabe

, Inoue

, Otsu

, Saeki

, Katsumata

, Takuse

, Iwatani

. 2016. Increases of microRNA let-7e in peripheral blood mononuclear cells in Hashimoto's disease. Endocr J, 63:375–380.