GCM2 Silencing in Parathyroid Adenoma Is Associated With Promoter Hypermethylation and Gain of Methylation on Histone 3

Abstract

Context

Glial cells missing 2 (GCM2), a zinc finger-transcription factor, is essentially required for the development of the parathyroid glands.

Objective

We sought to identify whether the epigenetic alterations in GCM2 transcription are involved in the pathogenesis of sporadic parathyroid adenoma. In addition, we examined the association between promoter methylation and histone modifications with disease indices.

Methods

Messenger RNA (mRNA) and protein expression of GCM2 were analyzed by reverse transcriptase–quantitative polymerase chain reaction (RT-qPCR) and immunohistochemistry in 33 adenomatous and 10 control parathyroid tissues. DNA methylation and histone methylation/acetylation of the GCM2 promoter were measured by bisulfite sequencing and chromatin immunoprecipitation–qPCR. Additionally, we investigated the role of epigenetic modifications on GCM2 and DNA methyltransferase 1 (DNMT1) expression in parathyroid (PTH)-C1 cells by treating with 5-aza-2′-deoxycytidine (DAC) and BRD4770 and assessed for GCM2 mRNA and DNMT1 protein levels.

Results

mRNA and protein expression of GCM2 were lower in sporadic adenomatous than in control parathyroid tissues. This reduction correlated with hypermethylation (P < .001) and higher H3K9me3 levels in the GCM2 promoter (P < .04) in adenomas. In PTH-C1 cells, DAC treatment resulted in increased GCM2 transcription and decreased DNMT1 protein expression, while cells treated with the BRD4770 showed reduced H3K9me3 levels but a nonsignificant change in GCM2 transcription.

Conclusion

These findings suggest the concurrent association of promoter hypermethylation and higher H3K9me3 with the repression of GCM2 expression in parathyroid adenomas. Treatment with DAC restored GCM2 expression in PTH-C1 cells. Our results showed a possible epigenetic landscape in the tumorigenesis of parathyroid adenoma and also that DAC may be a promising avenue of research for parathyroid adenoma therapeutics.

Primary hyperparathyroidism (PHPT) occurs because of the enlargement of one or more parathyroid glands, which causes overproduction of parathyroid hormone (PTH) (1). However, there is limited literature on the molecular mechanism involved in parathyroid tumorigenesis. Glial cells missing 2 (GCM2) is a master regulator required for the development of parathyroid glands in mammals (2-4). In GCM2-knockout mice, missing parathyroid glands develop hypocalcemia and hyperphosphatemia, as observed in hypoparathyroidism (5). Since parathyroid glands are not formed in GCM2-knockout mice, parathyroid precursor cells undergo apoptosis and die soon after birth (3, 6).

GCM2 is a zinc-type nuclear transcription factor containing a DNA binding domain at the N-terminus that binds to the consensus motif 5′-ATGCGGGT-3′ (2, 7-9). GCM2 is known to be expressed in parathyroid glands; however ,the downstream targets of GCM2 and its deregulation in parathyroid adenomas have not been explored (10). Inactivating mutations in the GCM2 gene can lead to decreased secretion of PTH, resulting in congenital isolated hypoparathyroidism (5, 11, 12).

In silico studies have shown that transcription factors, that is, GCM2-functional binding elements, are present in the calcium-sensing receptor (CASR) promoter 1 (–451 to –441 bp relative to the transcription start site [TSS]) and promoter 2 (–166 to –156 bp relative to the TSS), and are able to transactivate CASR expression (13-15). Previous studies reported decreased GCM2 expression both in vivo in mice and in human parathyroid cultured cells treated with GCM2 small interfering RNA that is associated with reduction of the CaSR gene, a marker of differentiation for parathyroid cells (16).

It has also been demonstrated that GCM2 interacts with other transcription factors such as GATA binding protein 3 (GATA3) and musculoaponeurotic fibrosarcoma oncogene homolog B (MAFB), and determines PTH expression in cultured cells (8). Nucleotide sequence analysis revealed binding sites for the transcription factors GATA3 (–55 to –18 relative to the TSS) and paired box 1 (PAX1; +8 to +17 relative to the TSS), on the promoter region of GCM2, and GATA3 and PAX1 both are essential for parathyroid gland development during embryogenesis (17, 18). However, little is known about the molecular mechanisms involved in the transcriptional regulation of GCM2 and in the downstream activities of GCM2 protein in normal parathyroid tissue or why there is decreased GCM2 expression in cultured parathyroid cells isolated from chronic kidney disease patients (14). Hence, we asked whether epigenetic modifications such as DNA methylation and histone posttranslational changes regulate GCM2 expression. DNA methylation is mediated by DNA methyltransferases (DNMTs) that modify cytosine-guanine dinucleotide (CpG) sites, which may result in gene silencing (19). Histone modifications are posttranslational changes to DNA packaging histone proteins that promote chromatin remodeling and changes in gene expression. Methylation and acetylation are the 2 well-characterized histone modifications associated with transcriptional repression and activation of many other target genes. Because of the reversibility of epigenetic modifications, pharmacological agents like 5-aza-2′-deoxycytidine (DAC) and BRD4770 are used to reverse this process, making them perfect targets for prevention and therapy. The potent and specific DNA methylation inhibitor is DAC, a deoxyribose analogue of nucleoside that is converted to deoxynucleotide triphosphate and is then incorporated in place of cytosine into replicating DNA (20). DAC is used to inhibit DNA methylation in tumor cells and helps to increase the transcription of CASR in colon cancer (21, 22). The formal name of BRD4770 is 2-(benzoylamino)-1-(3-phenylpropyl)-1H-benzimidazole-5-carboxylic acid, and it is used as a histone methyltransferase. Using specific concentrations of BRD4770 will decrease demethylation and trimethylation of lysines on histone 3. A previous study on pancreatic cancer showed the expected effects of BRD770, as it stops the proliferation of pancreatic cancer cells without resulting in apoptosis (23).

Accordingly, in the present study we explored whether DNA methylation and histone modifications affect the regulation of GCM2 and promote parathyroid tumorigenesis. In addition, we explored the mechanism of epigenetic control of GCM2 by using DNA and histone methyltransferase–inhibiting drugs in a rat cell line model of parathyroid adenoma, and assessed the effects on GCM2 mRNA expression.

Materials and Methods

Baseline Participant Characteristics

Participants were enrolled, sampled, and analyzed after giving written informed consent. This study was approved by the institutional ethics committee at the Post Graduate Institute of Medical Education and Research, Chandigarh, India. Participants were diagnosed with PHPT if they presented with elevated plasma PTH and hypercalcemia. A family history suggestive of genetic forms of PHPT such as multiple endocrine neoplasia 1 (MEN1), multiple endocrine neoplasia type 2A (MEN2A), multiple endocrine neoplasia type 4 (MEN4), familial hypocalciuric hypercalcemia syndrome, neonatal severe hyperparathyroidism, HPT-jaw tumor, and familial isolated PHPT were excluded from the study. The differential diagnosis can also be assisted by clinical features, and those related to genetic PHPT were excluded by appropriate clinical examination such as acromegaly, galactorrhea, neurofibromatosis, thyroid cancer surgery, jaw tumor, and appropriate hormonal measurements (serum prolactin and insulin-like growth factor 1). However, gene testing of MEN1 is the desired diagnostic criterion because it identifies germline mutations even without apparent family history of PHPT.

In total, we collected 40 tissue samples from PHPT patients with histologically confirmed parathyroid adenoma (PA) and, from 10 patients, normal parathyroid (control) during required thyroid surgeries. The tissue samples were routinely fixed in 10% buffered formalin and processed for hematoxylin and eosin stain in the department of histopathology of the institution. The slides were examined to confirm the diagnosis either as adenoma or as normal parathyroid tissue by an experienced endocrine pathologist (U.N.S.).

For each sample, relevant clinical details were recorded and biochemical measurements such as serum calcium (reference range [RR], 8.2-10.2 mg/dL), serum phosphorus (RR, 2.7-4.5 mg/dL), serum creatinine (RR, 0.50-1.20 mg/dL), serum albumin (RR, 3.5-5.5 g/dL), and alkaline phosphatase (RR, 40-129 U/L) were measured by an Olympus autoanalyzer, and the serum calcium level was adjusted with the serum albumin level. Plasma intact PTH (iPTH; RR, 15-65pg/mL) and 25-hydroxyvitamin D (RR, 11.1-42.9 ng/mL) were measured using immunochemiluminescence (ELECSYS-2010; Roche Diagnostics).

Multiple endocrine neoplasia 1 (MEN1) Screening of Study Participants and In Silico Analysis of the Pathogenic Effect of MEN1 variants

MEN1 mutation screening was performed for all study participants from the genomic DNA samples. DNA was isolated from the patient’s blood using a Qiagen DNA isolation kit and subjected to polymerase chain reaction (PCR) amplification of all coding exons of the MEN1 gene. The MEN1 mutations were identified by Sanger sequencing. Post sequencing, the chromatograms were visualized and analyzed using the CodonCode Aligner. Patients with a MEN1 variation were further studied; each variation was analyzed in silico for any possible benign or pathogenic effect using the MutPred bioinformatics tool (http://mutpred.mutdb.org/). Only patients with no MEN1 mutation (n = 33) were included in this study to rule out a genetic contribution to disease pathology.

Analysis of Glial Cells Missing 2 and DNA Methyltransferase 1 Gene Expression

Total RNA was isolated from each parathyroid adenoma and normal parathyroid tissue and reverse transcribed as described previously (24). GCM2 and DNMT1 mRNA levels were quantified by reverse transcriptase–quantitative PCR (RT-qPCR) on an ABI StepOnePlus Real-Time PCR System (Applied Biosystems) using 18s ribosomal RNA as a housekeeping gene for normalization and specific primers: GCM2 forward primer 5′AAGGCATGCCCTAACTGTCA-3′ and GCM2 reverse primer 5′ATGAACTCCCTTGGCCTGAA-3′; DNMT1 forward primer 5′TACCTGGACGACCCTGACCTC-3′ and DNMT1 reverse primer 5′-CGTTGGCATCAAAGATGGACA-3′. RT-qPCR reaction conditions were similar to those described in our previous article (25). Experiments were performed in duplicate with 2 nontemplate controls as a negative control. The 2^–ΔΔCT method was used to quantify the relative expression of GCM2 and DNMT1 as a fold change.

Immunohistochemistry

The paraffin-embedded parathyroid adenoma and normal parathyroid tissue blocks were retrieved from the Department of Histopathology and 4- to 5-µ sections were cut for immunohistochemistry (IHC). The IHC protocol was performed as described previously (26). Slides were graded semiquantitatively based on the percentage of positive cells and staining intensity by an endocrine pathologist (U.N.S.) for GCM2 (Abcam, catalog No. ab201170, RRID:AB_2891136 [27]) protein expression. Nuclear staining for GCM2 was regarded as positive. Approximately 200 cells were counted in 6 different regions for each section under 40× magnification. The staining intensity for the parathyroid adenoma and normal parathyroid samples was assessed and graded semiquantitatively: weakly positive (+; < 40%), moderately positive (++; 40%-60%), and strongly positive (+++; > 60%) (24).

Bisulfite-Specific Polymerase Chain Reaction and Sequencing

Genomic DNA from parathyroid adenoma and normal parathyroid tissues were isolated and bisulfite-converted by using the EZ DNA Methylation-Gold Kit (Zymo Research Corp) as previously described (25).

For bisulfite-specific PCR, bisulfite-converted DNA was amplified by using specific primer sets for GCM2: forward primer 5′-TTTGGGTAATAGAGTGAGATAAAAA-3′ and reverse primer 5′-AAAAACTAAAAACAATAATTATAAACCC-3′. Primers were used to amplify the 365-bp sequence containing 17 CpG sites located at nucleotides (nts) –300 to +240 relative to the GCM2 TSS. PCR reaction was carried out using EpiTaq polymerase enzyme (Takara). Amplified PCR products were separated by electrophoresis on 2% agarose gels and analyzed by bisulfite sequencing (ABI 3730 XL DNA Analyzer; Thermo Fisher Scientific). The sequencing data were evaluated with the online software Bisulfite sequencing and DNA Methylation analysis (BISMA) (25). A methylation density of greater than 10% was considered methylated for the specific gene examined (28).

Chromatin Immunoprecipitation of H3K9me3, H3K27me3, and H3K9ac

Levels of H3K9m3 (Abcam, catalog No. ab8898, RRID:AB_306848 [29]), H3K27me3 (Abcam, catalog No. ab6002, RRID:AB_305237 [30]), and H3K9ac (Cell Signaling Technology, catalog No. 9649, RRID:AB_823528 [31]) on the GCM2 promoter region were quantified by chromatin immunoprecipitation (ChIP)-qPCR assay. Twenty-three parathyroid adenoma samples were processed for each ChIP-qPCR assay, and the experiment was performed as described previously (22, 32). In brief, 25 to 30 mg of homogenized frozen tissue was fixed with 1% paraformaldehyde at room temperature for 10 minutes, stopped with 0.125-M glycine, and lysed for 10 minutes on ice. The lysate was digested in 1-mL lysis buffer using a tissue homogenizer (NaCl 150 mM, Tris HCl [pH 7.5] 25 mM, EDTA 5 mM, Triton X 100 1%, sodium dodecyl sulfate 0.1%, sodium deoxycholate 0.5%, phenylmethylsulfonyl fluoride 1 mM, and sodium butyrate 10 mM) 1X PIC followed by resuspension in lysis buffer for sonication 15 cycles (30 seconds on and 30 seconds off) using a Diagenode sonicator. Cross-linked DNA was immunoprecipitated with nonspecific immunoglobulin G (IgG) antibodies (45% of total supernatant [Abcam, catalog No. ab171870, RRID:AB_2687657] [33]) for isotype control and immunoprecipitation with anti-H3K9m3, anti-H3K27me3, and anti-H3K9ac antibodies (45% of total supernatant) was performed overnight at 4 °C. The next day, samples were pulled down with protein A/G beads, washed, reverse cross-linked, and ChIP-DNA was extracted with phenol-chloroform isoamyl alcohol as described previously. For ChIP-qPCR assay, primers were designed by using primer blast provided by National Center for Biotechnology Information and sequences used were GCM2 forward primer 5′-TGTGCTCCTACGGGATGCAG-3′ and GCM2 reverse primer 5′-CCCTCCTTCGGATGGACAGC-3′. qPCR with ChIP primers was carried out and performed in duplicate, and levels of H3K9me3, H3K27me3, and H3K9ac were calculated by the fold enrichment method relative to the nonspecific antibody and normalized to the input DNA.

Parathyroid-C1 cell line

Because a human parathyroid cell line is not available, the parathyroid-C1 (PTH-C1) cell line, which was established from rat parathyroid tissue, was used for in vitro treatment assays (34). The PTH-C1 cells have an epithelial morphology and exhibit parathyroid functional features (34). The cells were cultured in 25 cm² or 75 cm² tissue culture flasks supplemented with nutrient mixture F-12 Ham Coon’s modification (Sigma-Aldrich) containing 10% fetal bovine serum (Gibco, Thermo Fisher Scientific), and 100-U/mL penicillin/streptomycin (Gibco, Thermo Fisher Scientific) and then incubated at 37 °C in a 5% CO₂ incubator.

In vitro treatments and 3-(4, 5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide assay

A stock solution of DAC or BRD4770 (Sigma-Aldrich) was made in vehicle and diluted to appropriate concentrations for treatment. Cells treated with vehicle control were taken as controls. Approximately 1 × 10⁴ cells were seeded per well in a 96-well plate and then subsequently treated with different concentrations of DAC for 72 hours and BRD4700 for 48 hours alone, respectively. At the end of the treatments, 100 µL of 3-(4, 5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) reagent was added to each well from a 5-mg/mL stock in phosphate-buffered saline (PBS). The 96-well plates were incubated for 3 hours at 37 °C in the dark. The supernatant was taken out and formazan crystals were dissolved in 100-µL dimethyl sulfoxide at 37 °C for 10 minutes with gentle agitation. Absorbance from the plates was read at 570 nm and reference was read at 620 nm with an enzyme-linked immunosorbent assay reader.

Impact of 5-aza-2′-deoxycytidine treatment on glial cells missing 2 expression in parathyroid-C1

To study the impact of DAC on the expression of GCM2 in PTH-C1 cells by RT-qPCR, cells were treated with 10-µM DAC for 72 hours (the fresh drug was changed every 24 hours). Total RNA was isolated from treated and untreated PTH-C1 cells, reverse transcribed, and GCM2 mRNA expression was assessed by RT-qPCR as described earlier for tissue samples (25). In treatment experiments, rat-specific GCM2 primers were used: forward primer 5′-TGCCTCAGGAACCAGCTCACT-3′ and reverse primer 5′-AGGTGGCGCTGAGCTTTCTTC-3′. Experiments were performed in triplicate with 2 nontemplate samples as a negative control. Data were normalized using β-actin (forward primer 5’CCCATCTATGAGGGTTACGC-3′ and reverse primer 5′-TTTAATGTCACGCACGATTTC -3′) as a housekeeping gene and relative expression by fold change was calculated using 2^–ΔΔCT value.

Impact of BRD4770 treatment on glial cells missing 2 expression in parathyroid-C1

After treatment with 5-µM BRD4770 in PTH-C1 cells for 48 hours, H3K9me3 levels assessed by ChIP-qPCR and mRNA expression of GCM2 were quantified by RT-qPCR. ChIP-qPCR assay was performed on ChIP-DNA isolated from treated and untreated PTH-C1 cells. Rat-specific GCM2 primers (forward primer 5′-GGAGCCTGAGGTAATGGGCT-3′ and reverse primer 5′-ATTAGAGCCCCGAGCCAAGG-3′) were used in the qPCR. Finally, ChIP-qPCR was performed, and analysis was carried out as described earlier for human parathyroid tissue samples.

Immunofluorescence staining of DNA methyltransferase 1 in parathyroid-C1 cells

After treatment with DAC, PTH-C1 cells were grown on 8-well glass-chambered slides and fixed with 4% paraformaldehyde in 1X PBS for 20 minutes, then washed extensively with 1X PBS. Cells were permeabilized with 0.2% Triton-X in 1X PBS for 20 minutes and blocked with 2% bovine serum albumin in PBS for 1 hour. Primary antibodies for proteins of interest (DNMT1; Abcam, catalog No. ab188453, RRID:AB_2877711 [35]) were diluted in 2% bovine serum albumin and applied to the cells. Slides were incubated for 4 hours at room temperature. After washing with 1X PBS, a secondary antibody-labeled fluorochrome (1:500) was applied to the cells for 1 hour at room temperature in the dark. The cells were washed once for 5 minutes with 1X PBS. Nuclear staining and mounting were performed with Anti Fade Gold reagent (Thermo Fisher Scientific). Whole-slide images were acquired using an EVOS microscope with appropriate fluorescent labels.

Statistical analysis

Graph Pad Prism version 6.0 was used to analyze the data. qPCR data are expressed as mean ± SD and ChIP-qPCR data expressed as mean ± SEM. In vitro data represent 3 independent trials and are shown as mean with SEM. A P value of less than .05 was considered statistically significant. Mann-Whitney, chi-square, and t tests were used as appropriate to compare the differences in study data between the control and adenoma samples. The Pearson correlation test was used to examine the relationship among various continuous variables.

Results

Characteristics of Primary Hyperparathyroidism Patients

The mean age of the patients with confirmed parathyroid adenoma was 43 ± 13.5 years (range, 18-65 years) with a female to male ratio of 3:1 (31 women and 9 men). The mean age of the patients with normal parathyroid function who had undergone required surgery for benign thyroid disease was 43 ± 13.3 years (range, 25-61 years; 7 women and 3 men). The presence of bone pains, weakness, and fatigue, kidney stones, gallstone disease, fractures, pancreatitis, and vitamin D deficiency (25-hydroxyvitamin D level < 20 ng/mL) at the time of presentation was categorized as symptomatic PHPT. All the patients had hypercalcemia ,with mean serum calcium of 12.0 ± 1.0 (RR, 10.5-14.5 mg/dL), elevated median iPTH level of 558 pg/mL (first interquartile range [IQR]-third IQR), and median parathyroid adenoma weight was 2 g (first IQR-third IQR) (22).

Transcriptional Pattern of Glial Cells Missing 2 in Sporadic Parathyroid Adenomas

RT-qPCR analysis showed GCM2 mRNA levels were significantly lower in 29 (88%) of the 33 sporadic parathyroid adenomas relative to the control parathyroid samples, with a mean fold-change reduction of 0.25 ± 0.25 (P < .0001) (Fig. 1A). IHC analysis revealed nuclear staining for GCM2 protein both in the adenomatous and control parathyroid gland samples. In the control parathyroid samples, we observed 3+ intensity. In contrast, in the parathyroid adenomatous samples, we observed 1+ intensity in 16 of 33 (48%), 2+ intensity in 14 of 33 (42%), and 3+ intensity in 3 of 33 (9%) (Fig. 1B). The mean proportion of GCM2-positive cells was 44 ± 22% in sporadic parathyroid adenomas and 77.8 ± 3.4% in the control parathyroid samples. GCM2 protein expression was significantly reduced in the parathyroid adenomas relative to the control parathyroid samples (P < .0001).

Figure 1.

Glial cells missing (GCM2) messenger RNA (mRNA) and protein expression analysis in sporadic parathyroid adenoma (N = 33). a, Bar graph representing relative mRNA expression of GCM2 in parathyroid adenoma compared to control parathyroid samples. b, Representative photomicrographs showing GCM2 protein levels in A, parathyroid tissue sections control, and B, C, and D, show intensity in parathyroid adenoma as 1+, 2+, and 3+, respectively. Images captured at 40× magnification, scale bar = 400 µm.

Glial Cells Missing 2 Promoter Hypermethylation in Sporadic Parathyroid Adenomas

Bisulfite-converted DNA sequence analysis showed that 17 CpG sites (nt –300 to +240 relative to the TSS) located upstream of the GCM2 promoter (Fig. 2A) had high levels of methylation in 57.5% (n = 19) of the parathyroid adenomas with a mean methylation density of 20.7 ± 6% (range, 11.1%-33.3%). While 14 parathyroid adenomas had a mean methylation density of less than 10%, 3 parathyroid adenomas did not show methylation at any CpG sites in the region of the GCM2 promoter. Normal control parathyroid samples showed complete methylation at none of the 17 CpG sites, except for a single sample that showed an average of 6.7% methylation. Representative unmethylated and methylated CpG sites in the GCM2 promoter region are shown in Fig. 2B. The promoter region of GCM2 in parathyroid adenoma was highly methylated, showing complete methylation at 13 of 17 CpG sites. Also notable in the parathyroid adenomas, the eighth CpG site, which is a putative GATA3 binding site (nt –55 to –18 relative to +1 TSS), showed an average methylation of 56.4%, and the twelfth CpG site, a PAX1 binding site (nt +8 to +17 relative to +1 TSS), showed only 3.4% average methylation (Fig. 2C). Correlation analysis revealed a significant negative association between GCM2 mRNA expression and the level of methylation in sporadic parathyroid adenomas (r = –0.52; P < .001) (Fig. 2D). Thus, the reduced expression of GCM2 in parathyroid adenoma tissues was the effect of the methylation.

Figure 2.

Glial cells missing (GCM2) promoter methylation in sporadic parathyroid adenoma (N = 33). A, GCM2 promoter region selected for bisulfite sequencing polymerase chain reaction showing putative GATA3 and PAX1 binding sites and the transcription start site (TSS). B, Representative chromatograms showing methylated sites (blue arrow) in parathyroid adenoma compared to control parathyroid. C, Tabular representation of 17 cytosine-guanine dinucleotide (CpG) sites at GCM2 promoter showing average percentage methylation at each CpG site. D, Scatter plot showing the negative association between gene expression of GCM2 and DNA promoter methylation in sporadic parathyroid adenoma cases (N = 33).

DNA Methyltransferase 1 Expression in Parathyroid Adenomas

Because we observed such a high level of methylation in the GCM2 promoter in parathyroid adenomas, we next examined the gene expression of the maintenance methyltransferase DNMT1. We found that DNMT1 mRNA levels were significantly upregulated in sporadic parathyroid adenomas, with a mean fold change of 5.1 ± 3.7 (P < .0006) compared with the control parathyroid samples (Fig. 3). We found a significantly positive association between DNMT1 mRNA levels and GCM2 promoter hypermethylation (r = 0.39; P < .03) in sporadic parathyroid adenomas. Further, we observed that DNMT1 mRNA expression was positively associated with tumor weight (r = 0.40; P < .03), although it was not associated with any other biochemical parameters of parathyroid adenomas.

Figure 3.

Bar graph representing relative messenger RNA (mRNA) expression of DNA methyltransferase 1 (DNMT1) in parathyroid adenomas compared to control parathyroid samples.

H3K9me3, H3K27me3, and H3K9ac Modifications in Sporadic Parathyroid Adenomas

Next, we investigated levels of posttranslational methylation and acetylation on histone 3 within the GCM2 promoter in adenomatous and control parathyroid tissues using 23 adenomas and 10 control parathyroid samples. ChIP-qPCR analysis revealed higher H3K9me3 levels relative to IgG in the GCM2 promoter in parathyroid adenomas compared to control parathyroid, with a mean fold increase of 7.67 ± 3.08 in parathyroid adenoma and 0.52 ± 0.16 in control parathyroid samples relative to IgG. Levels of H3K27me3 relative to IgG were high in 19 of the 23 (83%) parathyroid adenomas with an overall mean fold increase of 31.2 ± 22.4, whereas the mean increase was only 1.15 ± 0.23 in control parathyroid samples. Thus, sporadic parathyroid adenomas exhibited 15-fold (P < .02) and 26-fold (P < .004) higher levels of H3K9me3 and H3K27me3, respectively, compared to control parathyroid samples (Fig. 4A and 4B). In contrast, H3K9ac levels were lower in parathyroid adenomas, with mean fold increases of 3.03 ± 0.58 in 20 of the 23 (87%) and 14.6 ± 9.40 in the control parathyroid samples relative to IgG (Fig. 4C). Thus, the active histone marker H3K9ac was decreased by 5-fold (P < .02) in sporadic parathyroid adenomas compared to control parathyroid samples.

Figure 4.

Chromatin immunoprecipitation–quantitative polymerase chain reaction for Glial cells missing (GCM2) promoter region in sporadic parathyroid adenomas. Quantification of fold enrichment in A, H3K9me3; B, H3K27me3; and C, H3K9ac (levels relative to nonspecific immunoglobulin G as control and normalized with input DNA). Bars represent mean fold enrichment and SEM (n = 23).

The Pearson correlation analysis between GCM2 mRNA expression and H3K9me3 methylation, a repressive modification associated with gene repression, showed a significantly negative association in parathyroid adenomas (r = –0.49, P < .04). The other repressive modification, H3K27me3, was significantly associated with serum calcium, whereas the activating modification, H3K9ac, showed a significant inverse relationship with plasma iPTH in parathyroid adenomas.

5-Aza 2′-deoxycytidine and BRD4770 cytotoxicity detection using the 3-(4, 5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide assay

To detect the cytotoxicity of DAC and BRD4770, we performed the MTT assay. As shown in Fig. 5, there was no significant difference between the experimental group and negative control. We conclude that 10-μM DAC and 5-μM BRD4770 produced no cytotoxicity in PTH-C1 cells, respectively.

Figure 5.

Cell viability was analyzed with 3-(4, 5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay in parathyroid (PTH)-C1 cells after using different concentrations of A, 5-aza 2′ deoxycytidine (DAC) for 72 hours. Data are shown as means of 3 experiments in triplicate (P < .78). B and C, Representative bright-field images of control and DAC-treated PTH-C1 cells for 72 hours; and D, 2-(benzoylamino)-1-(3-phenylpropyl)-1H-benzimidazole-5-carboxylic acid (BRD4770), for 48 hours. Data are shown as means of 3 experiments in triplicate (P < .40). E and F, Representative bright-field images of control and BRD4770-treated PTH-C1 cells for 48 hours at 20×, scale bar = 200 μM.

Inhibition of DNA Methyltransferase 1 With 5-Aza 2′-Deoxycytidine Results in Higher Glial Cells Missing 2 Transcription in Parathyroid-C1 cells

To test whether inhibition of DNMT1 would result in increased GCM2 transcription by putatively relieving epigenetic inhibition, we treated the transformed rat parathyroid cell line, PTH-C1, with DAC, and measured the mRNA levels of GCM2 by RT-qPCR. In PTH-C1 cells, we observed that the expression of GCM2 was significantly increased (P < .001) (Fig. 6A) after treatment with 10-μM DAC for 72 hours relative to untreated cells. Immunofluorescence microscopy revealed reduced protein levels of DNMT1 after treatment with 10-μM DAC for 72 hours in PTH-C1 cells compared to untreated cells (Fig. 6B).

Figure 6.

The effect of 5-aza-2′-deoxycytidine (DAC) treatment in a parathyroid cancer cell line (PTH-C1). A, Expression of glial cells missing (GCM2) messenger RNA (mRNA) relative to beta actin in PTH-C1 cells treated with 10 µM DAC vs untreated cells at 72 hours as determined by reverse transcriptase–quantitative polymerase chain reaction. The results are shown as mean and SEM values and P values were evaluated using t test from 3 independent experiments. B, Representative immunofluorescence staining of DNA methyltransferase 1 (DNMT1; red) in PTH-C1 cells (untreated with DAC, upper panel; treated with DAC, bottom panel). Nuclei were stained with 4′, 6-diamidino-2-phenylindole (DAPI; blue). Images captured at 20× magnification. Scale bar = 200 µm.

Histone Methylation (H3K9me3) and Glial Cells Missing 2 Expression in Parathyroid-C1 cells in response to BRD4770

Next, to show that treatment with the histone methyltransferase inhibitor BRD4770 affects the H3K9me3 modification within the GCM2 promoter, a ChIP-qPCR assay was performed to measure H3K9me3 levels. In ChIP-qPCR analysis, we observed a decrease of H3K9me3 levels in the GCM2 promoter region in BRD4770-treated PTH-C1 cells compared to untreated cells (Fig. 7A). However, we found a slight but not significant increase of GCM2 mRNA in BRD4770-treated PTH-C1 cells relative to untreated cells (Fig. 7B).

Figure 7.

The effect of 2-(benzoylamino)-1-(3-phenylpropyl)-1H-benzimidazole-5-carboxylic acid (BRD4770) treatment in a parathyroid cancer cell line (PTH-C1). PTH-C1 cells were treated with 5 µM of BRD4770 for 48 hours. A, Presence of H3K9me3 modification in the glial cells missing (GCM2) promoter quantified by chromatin immunoprecipitation–quantitative polymerase chain reaction. After treatment with BRD4770, decreased in trimethylation of H3K9 level was observed in the GCM2 promoter region. B, Expression of GCM2 messenger RNA (mRNA) relative to actin in the PTH-C1 cell line before and after BRD4770 treatment. Results are shown as mean and SEM values, and P values were evaluated by t test from 3 independent experiments.

Discussion

In the present study, we demonstrate that GCM2, a master regulator for parathyroid gland development (36), is hypermethylated in parathyroid adenomas and results in downregulated expression of GCM2 mRNA. We also found that H3K9me3 and H3K27me3 levels were higher, and H3K9ac was lower, in adenomatous compared to control parathyroid samples. Furthermore, treatment with DAC, a DNA methyltransferase inhibitor, can restore GCM2 expression in PTH-C1 cells. To our knowledge, this is the first study to fully describe the epigenetic landscape of the GCM2 promoter in human parathyroid adenomas, and the first to report that promoter DNA hypermethylation contributes to the regulation of GCM2 in sporadic parathyroid adenomas.

We observed significantly reduced mRNA and protein expression of GCM2 in sporadic parathyroid adenomas relative to control parathyroid samples. These findings are in agreement with previous reports that showed similar underexpression of GCM2 in parathyroid tumors (14, 37).

One limitation of the present study is that the genetic status of GCM2 was not analyzed in sporadic parathyroid adenomas. Previous studies by Kebebew and colleagues in 2004 and Mannstadt et al in 2011 showed underexpression or overexpression of the GCM2 gene in parathyroid adenomas, respectively (10, 38). Activating mutations might account for the overexpression of the GCM2 gene. Another study by Guan et al in 2016 that studied the genetic status of the GCM2 identified the non–C-terminal conserved inhibitory domain (CCID) variant in the Italian population; however, it did not appear in Asian/Indian population (39). In a subset of families with familial isolated PHPT, germline-activating mutations affecting the GCM2 CCID variants were observed (39) and have never been found in MEN1 or HPT-jaw tumor syndrome. Of note, Guan et al in 2017 concluded that GCM2-activating variants were seen in the sporadic PHPT patients who had multigland involvement or persistent disease (40). However, Riccardi and colleagues in a 2019 study reported that the known GCM2-activating variants may not intend to develop a sporadic adenoma, though increased disease risk would be expected in carriers of such variants compared with the general population (41). But the exact mechanism of deregulated GCM2 expression in parathyroid tumorigenesis is still undetermined.

Interestingly, Canaff et al in 2007 identified functional GCM2-binding sites on CaSR promoters (promoter 1 and promoter 2); thus, there appears to be an association between GCM2 and CaSR (14). Our recent study also showed hypermethylation of a CpG site located within the GCM2-binding region on the CaSR promoter 2 affected CaSR expression in patients with sporadic parathyroid adenomas (22, 24, 37). Thus, the methylation analysis of a specific CpG site might play a key determinant role in the reduced gene expression. In the present study, the GCM2 promoter was hypermethylated with a significant inverse relationship between GCM2 mRNA expression and promoter methylation. Previously, it has been reported that loss of GCM2 expression in renal cell carcinoma is due to genetic deletions and that tumor-specific methylation correlates with survival (42). The GCM2 promoter contains putative GATA3 and PAX1-binding sites, and both are developmental transcription factors that regulate GCM2 expression (6, 8). A noteworthy observation was that the eighth (GATA3-binding site) and twelfth CpG sites (PAX1-binding site) showed 57.5% methylation and 3.4% methylation, respectively, in sporadic parathyroid adenoma. Similarly, others have reported that GATA3 acts via GCM2 regulation in parathyroid adenomas, as the GCM2 promoter is enriched with GATA3-binding motifs (6, 43). Another study reported that GATA3 interacts with GCM2 and MAFB transcription factors resulting in stimulation of PTH gene expression (8). Also, there is a strong link between PAX1 and GCM2, as PAX1 also helps to determine cell fate in the development of the parathyroid glands (17). Our results showed that GCM2 promoter hypermethylation, as well as specific methylated CpG sites, might lead to steric hindrance for transcription factor (GATA3 and PAX1)-binding, causing decreased GCM2 expression. This observation is of key interest for future investigations exploring predictive and prognostic biomarkers for sporadic parathyroid adenomas.

Our study showed significantly increased mRNA expression of DNMT1 in sporadic parathyroid adenomas compared to control parathyroid samples. A previous study on MEN showed DNA hypermethylation to be mediated by enhanced DNMT1 activity, thereby regulating the transcriptional silencing of menin 1 (MEN1). In this study, we found that DNMT1 expression was strongly associated with GCM2 promoter methylation patterns in sporadic parathyroid adenomas, suggesting that reduced expression of GCM2 may be due to DNMT1-mediated methylation. Correlation analysis suggested a significant association between DNMT1 mRNA expression and tumor weight in sporadic parathyroid adenomas. Earlier reports have suggested giant or large parathyroid tumor size correlates well with the severity of the disease (44, 45). This suggests that large tumor weight results in a more pronounced clinical picture in sporadic PHPT patients, with severe hypercalcemia and elevated iPTH levels.

Additionally, to understand the reason for reduced GCM2 expression, histone modifications have been studied. However, this is the first study to analyze histone changes in the promoter region of GCM2 in parathyroid adenoma, and we observed significantly higher H3K9me3 and H3K27me3 levels and lower H3K9ac levels. In a similar manner, we observed a significant inverse relationship of promoter hypermethylation and H3K9me3 levels with reduced GCM2 expression in sporadic parathyroid adenomas, thus linking both with transcriptional silencing. This is in line with a previous study that reported an interplay of histone modification and DNA methylation in decreased RASSF1A expression in breast cancer (46). Thus, we propose a possible role for H3K9me3 as a chief regulator in parathyroid tumorigenesis that should be investigated further.

We further examined the role of GCM2 promoter DNA and histone methylation by using the epigenetic inhibition drugs DAC and BRD4770 in the PTH-C1 cell line to observe whether they might cause reactivation of silenced GCM2 in PTH-C1 cell line. Several reports in colorectal cancer cell lines such as HCT116, SW48, and SW480 have shown that DAC treatment inhibits DNMT1 activity, causing reexpression of genes such as clusterin, interleukin-23 receptor, transketolase-like 1, and tubulin α-3C (47). We saw a reduction of DNMT1 protein in DAC-treated PTH-C1 cells and a concomitant significant increase in GCM2 gene expression after DAC treatment. The restoration of GCM2 expression suggests that inhibition of DNMT1 is sufficient and necessary to activate the GCM2 promoter. Previous studies have reported that DAC has been recognized as a therapeutic epigenetic-modifying drug and has been approved for the treatment of cancer (21, 47, 48).

In addition, we also further explored the role of H3K9me3 in PTH-C1 cells, since increased H3K9me3 levels were observed in the GCM2 promoter in human sporadic parathyroid adenomas. Thus, PTH-C1 cells treatment with BRD4770 resulted in lower H3K9me3 levels in the GCM2 promoter as observed by ChIP-qPCR, but changes in GCM2 transcription were not significantly different between treated and untreated cells, despite a trend suggestive of increased expression. BRD4770 has shown success in reducing H3K9me3 levels in pancreatic and liver cancer cells without inducing apoptosis (23).

The data on PTH-C1 cells, while not significant, may indicate a promising topic for future, more detailed research on the role of H3K9me3 histone modification in parathyroid disease.

Taken together, DNA hypermethylation and higher histone H3K9me3 levels in the GCM2 promoter in sporadic parathyroid adenomas both suggest that epigenetic modifications play a key role in the transcriptional silencing of GCM2, perhaps contributing to parathyroid tumorigenesis.

The limitations of this study, first, we did not determine the genetic status of the GCM2 gene in sporadic parathyroid adenomas. Second, in vitro experiments were performed in the rat parathyroid cell line PTH-C1 (34). As of yet, there is no human parathyroid cell line available for ideal in vitro experiments; however, transformed rat parathyroid cell lines may serve as an excellent model to study potential therapeutics. Third, protein expression of DNMT1 was not performed in parathyroid adenomas.

Conclusion

Our results demonstrate for the first time the importance of epigenetics in downregulating the expression of GCM2 through its promoter hypermethylation and histone methylation in human sporadic parathyroid adenomas. Strategies to upregulate GCM2 expression by targeting DNA and histone methylation may be useful to prevent and/or treat sporadic parathyroid adenoma.

Glossary

Abbreviations

BRD4770

2-(benzoylamino)-1-(3-phenylpropyl)-1H-benzimidazole-5-carboxylic acid

CASR

calcium-sensing receptor

ChIP

chromatin immunoprecipitation

CpG

cytosine-guanine dinucleotide

DAC

5-aza 2′-deoxycytidine

DNMTs

DNA methyltransferases

DNMT1

DNA methyltransferase 1

GATA3

GATA binding protein 3

GCM2

glial cells missing 2

IgG

immunoglobulin G

IHC

immunohistochemistry

iPTH

plasma intact parathyroid hormone

IQR

interquartile range

MAFB

musculoaponeurotic fibrosarcoma oncogene homolog B

MEN

multiple endocrine neoplasia

mRNA

messenger RNA

MTT

3-(4, 5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide

nt

nucleotide

PAX1

paired box 1

PBS

phosphate-buffered saline

PHPT

primary hyperparathyroidism

PTH

parathyroid hormone

PTH-C1

parathyroid-C1 cell line

RR

reference range

RT-qPCR

reverse transcriptase–quantitative polymerase chain reaction

TSS

transcription start site

Additional Information

Disclosures: The authors have no potential conflicts of interest to disclose.

Data Availability

All relevant data are within in the manuscript. The data sets generated during and/or analyzed during the present study are available from the corresponding author on reasonable request.