Subjects

This study was approved by the Ethics Committee of the First Affiliated Hospital of China Medical University. Thyroid specimens were obtained from the non-affected lobes of patients who underwent thyroidectomy. Serum samples were also obtained to detect thyroid antibodies and measure thyroid function using an electrochemiluminescence method. All subjects were female and were euthyroid at the moment of surgery, with no thyroid hormone replacement therapy or anti-thyroid drugs and with no history of systemic autoimmune disease or any severe illnesses. Patients in the HT group were positive for at least one thyroid antibody (TG-Ab, TPO-Ab) and had pathologically confirmed HT during surgery. Subjects in the control group were age-matched and were both negative for thyroid antibody and pathological diagnosis of HT. Signed informed consent was obtained from all donors before surgery. The clinical characteristics of patients and controls are listed in Table I.

Primary cultures of human thyroid cells

Thyroid tissue was cut into small pieces, washed with Hanks’ balanced salt solution, and then incubated with collagenase type IV (1.25 mg/ml) (Sigma Aldrich , St. Louis, MO, USA) in RPMI 1640 medium for 2 h at 37°C. The cells were then filtered and centrifuged at 1,000 rpm for 5 min. The cells were finally resuspended in a complete medium consisting of RPMI 1640 medium, supplemented with 10% fetal calf serum (FCS) (Gibco, Grand Island, NY, USA), bovine insulin (1 μg/ml), transferrin (5 μg/ml), hydrocortisone (10^–8 mM), thyroid stimulating hormone (30 μU/ml, non-essential amino acids (1 mM) and 1% penicillin-streptomycin (Sigma).

Chromatin immunoprecipitation sequencing (ChIP-seq) using deepTools

Chromatin immunoprecipitation was performed using a ChIP-IT High Sensitivity Kit (Active Motif, Carlsbad, CA, USA). After fixation with formaldehyde, the cells were collected by scraping into a 15 ml centrifuge tube, incubated with micrococci nuclease (Active Motif) at 37°C for 20 min. The nuclei were resuspended and sonicated to obtain the appropriate chromatin. Each chromatin sample was incubated overnight at 4°C with H3K4me3 antibody (CST, Beverly, MA, USA), and then incubated with protein G beads in a shaker for 2 h at 4°C and separated using a magnetic stand. After de-crosslinking and purification, the qualified DNA was used for second-generation sequencing.

Bioinformatics analyses

We sequenced 12 samples (including 6 inputs) for hg19 species using Chip-seq technology, on average generating 32272581 raw reads and 32109156 clean reads, which were mapped to the human hg19 reference genome using SOAPaligner/soap2 [8]. Only alignments within two mismatches were considered in peak calling, based on a defined analysis using Model-based Analysis of ChIP-Seq (MACS), and MAnorm was used to identify differential peaks between the HT and control groups [9]. For one of the ChIP-ed control samples, the mapping rate was much lower than for the others and this sample did not qualify for further analysis.

Western blotting

The total protein extraction was performed using Protein Extraction Kit (KeyGEN, Jiangsu, China). Immunoblotting was performed according to the standard protocol and the proteins were transferred to PVDF membranes (Millipore, Burlington, MA, USA), blocked with TBST + 5% skimmed milk, incubated overnight at 4°C with the antibodies anti-H3K4me3 (Abcam, Cambridge, MA, USA), anti-β-actin (Abcam) and anti-MLL1 (CST), then washed three times with TBST and incubated with HRP-conjugated IgG antibody. Immunoblot bands were quantified relative to β-actin using Image Pro Plus software (Media Cybernetics, Inc.) [10].

Real-time RT-PCR

Total RNA was extracted using TRIzol (Gibco). Reverse transcription was carried out using a PrimeScript RT Reagent Kit with gDNA Eraser (Takara, Bio Inc., Shiga, Japan). cDNA amplification was then performed using SYBR Premix Ex Taq (Takara) with a Light Cycler 480 instrument (Roche, Basel Switzerland). mRNA expression for each sample was analyzed using glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as an internal control. Primer sequences are listed in Table II.

Immunohistochemistry

After dewaxing and hydration, paraffin sections were incubated with 3% H₂O₂ for 15 min, placed in a microwave oven for 10 min for antigen retrieval, blocked with goat serum for 20 min, and then incubated overnight with MLL1 antibody (CST). The sections were then dehydrated in gradient alcohol, sealed and observed using a microscope (Olympus BX51, Tokyo, Japan).

Statistical analysis

Statistical analyses were performed using SPSS Software 20.0 (SPSS, Inc., Chicago, IL, USA). Results are shown as means ± standard deviation and an independent sample t-test was used to determine whether differences between the patient group and the control group were statistically significant. P-values < 0.05 are viewed as statistically significant.

Differences in H3K4me3 patterns between HT patients and control subjects

Based on their location (UCSC annotation data), peaks were classified and assigned to the following genomic regions: intergenic, introns, downstream, upstream and exons. In agreement with the previously described canonical H3K4me3 pattern [11], all of our samples showed clear enrichment of H3K4me3 at the TSSs, suggesting that the genomic profiles were highly reproducible and that the average depth of H3K4me3 showed a bimodal distribution, from 1,000 bp upstream to 1,000 bp downstream, of the TSSs. The profiles of H3K4me3 identified in this study varied between the HT and control groups. Most of the sites with differential enrichment were located in intergenic and intron regions (Figure 1).

Figure 1

Genomic pattern of H3K4me3 peaks detected in six samples. A – H3K4me is mainly enriched around TSSs; B – Differential peak distribution on gene elements between Hashimoto’s thyroiditis group and age-matched control group. For one of the ChIP-ed control samples, the mapping rate was much lower than for the others and this sample did not qualify for further analysis

HT – Hashimoto’s thyroiditis, N – age- and sex-matched controls, TSS – transcriptional start site, H3K4me3 – trimethylated histone 3 lysine 4.

Gene ontology (GO) annotation and KEGG pathway analysis were used to further explore the function of the genes. The GO analysis showed that the differential peak-related genes were mainly involved in biological processes such as responses to stimuli and regulation of cell adhesion. KEGG pathway analysis suggested that these genes were remarkably enriched in cell adhesion molecules, AITD, pathways in cancer and the Ras signaling pathway (Table III). The most important differential genes with higher signals for H3K4me3 peaks in the HT group were those enriched in the AITD pathway, including TG, TSHR, MHCII, CTLA, CD80/86 and Fas/Fasl. It is noteworthy that the TG promoter of TFCs from HT patients had, on average, a much greater enrichment signal for H3K4me3 compared with the control group. The Integrative Genomics Viewer (IGV) was used to visualize H3K4me peaks in the TG promoter (Figure 2). Chemokines, including CCL2 (C-C motif chemokine ligand 2), CXCL8 (C-X-C motif chemokine ligand 8), CXCL10 (C-X-C motif chemokine ligand 10) and IL-18 (interleukin-18), adhesion molecules, including ICAM1 (intercellular adhesion molecule 1) and ITGA4 (integrin subunit α4), and apoptosis factors, including FASLG (Fas ligand) and TRAIL (TNF superfamily member 10), also showed increased H3K4me3 enrichment in the promoter region around the TSSs (Figure 3).

Figure 2

H3K4me3 profiles of thyroglobulin gene in HT and control group indicated by integrative genomics viewer. Regulation of H3K4me3 around thyroglobulin gene transcriptional start sites are shown in the tracks with significant differential peaks

Figure 3

Distribution of H3K4me3 enrichment around gene transcriptional start sites of eight genes

HT – Hashimoto’s thyroiditis, N – age- and sex-matched controls.

Gene expression

It is accepted that H3K4me3 is broadly associated with RNA polymerase II occupancy at the TSSs of active gene expression [12] and H3K4me3 was thought to regulate gene transcription [13]. Here, RT-PCR was used to determine the expression of eight HT-associated genes from our gene list (CCL2, ICAM1, CXCL8, CXCL10, IL-18, TRAIL, ITGA4 and FASLG), all of which showed H3K4me3 enrichment in the promoter. The mRNA expression of three genes (ICAM1, CCL2 and CXCL8) was elevated in primary thyrocytes from HT patients (n = 16, p < 0.05). No significant difference of mRNA expression was seen for the remaining five genes (p > 0.05) (Figure 4).

Figure 4

RT-PCR validation for transcriptional expression of eight genes with increased H3K4me3 enrichment. mRNA level normalized to control group (n = 16). Statistically significant differences are shown as *p < 0.05, **p < 0.01. Fold-changes for CCL2, CXCL8 and ICAM1 were 1.81, 1.67 and 1.71, respectively. No differences in expression were found for the other five genes (p > 0.05)

HT – Hashimoto’s thyroiditis, CON – age- and sex-matched controls.

Global H3K4me3 expression in thyrocytes and thyroid tissues

We next investigated global H3K4me3 levels of TFCs (n = 8) and thyroid tissues (n = 12). Global H3K4me3 protein levels were slightly, but not significantly (p > 0.05), higher in primary cultured thyrocytes from HT patients (Figure 5). The difference in H3K4 trimethylation of TFCs was focused mainly on certain genes, whereas overall methylation was unchanged. Compared with the control group, however, global H3K4me3 protein levels were significantly elevated in thyroid tissues from HT patients (p < 0.05).

Figure 5

Detection of global H3K4me3 protein expression by immunoblotting. A – Global H3K4me3 protein expression was markedly enhanced in HT thyroid tissues (n = 12). B – Bar graphs depicting global H3K4me3 protein expression from tissues. C – Global H3K4me3 protein expression of primary cultured thyrocytes in HT was slightly, but not significantly, higher (n = 8). D – Bar graphs depicting global H3K4me3 protein expression in thyrocytes (p > 0.05)

Statistically significant differences are shown as *p < 0.05, **p < 0.01. HT – Hashimoto’s thyroiditis, N – age- and sex-matched controls.

Detection of H3K4 methyltransferases

Because the MLL family of methyltransferases is known to be mainly responsible for methylation of H3K4 to influence genome transcription, we next investigated whether family members, including MLL1, LL2, MLL3, MLL4 and other methylation state-modifying enzymes, including SET7/9 (SET domain containing 7) and ASH1L (ASH1-like histone lysine methyltransferase), are associated with H3K4 trimethylation in thyroid tissues. We found that mRNA expression of MLL1 was markedly increased (n = 60, p < 0.05) (Figure 6). Immunoblotting and immunohistochemistry were then used to examine protein levels of MLL1 from a small number of samples (n = 16). Consistent with mRNA expression levels, protein expression of MLL1 was significantly increased in thyroid tissues from HT patients (p < 0.05) (Figure 7).

Figure 6

mRNA levels of H3K4 methyltransferases in thyroid tissues from HT patients and controls. Transcriptional expression of MLL1 was elevated in thyroids from HT patients (n = 60). GAPDH was used as an endogenous reference

Statistically significant differences are shown as *p < 0.05. HT – Hashimoto’s thyroiditis, CON – age- and sex-matched controls.

Figure 7

Protein expression of MLL detected in thyroid tissues from HT patient and controls. A – Protein expression of MLL1 in thyroid tissue from HT patients and controls (n = 12); B – bar graphs depicting MLL1 protein expression by immunoblotting; C – protein expression of MLL1 in thyroid tissue from HT patients and controls (n = 16); D – bar graphs depicting protein expression of MLL1 by immunohistochemistry

Statistically significant differences are shown as *p < 0.05, **p < 0.01, ***p < 0.01. HT – Hashimoto’s thyroiditis, N – age- and sex-matched controls.