This article is mentioned in:

Introduction

Parathyroid hormone is the principal physiologic regulator of calcium homeostasis. Hyperparathyroidism is a result of increased activity of the parathyroid glands, either from an intrinsic change altering excretion of parathyroid hormone (primary hyperparathyroidism, pHPT) or from an extrinsic change affecting calcium homeostasis stimulating production of parathyroid hormone (secondary hyperparathyroidism, sHPT) (1). Clinically, hyperparathyroidism leads to skeletal and renal complications in addition to an impairment in quality of life (2–4). Prolonged oversecretion of parathyroid hormone is accompanied by histologically abnormal parathyroid glands that are typically enlarged and hypercellular with decreased stromal fat (5,6).

Parathyroid glands are derived from the third and fourth pharyngeal pouches and are endodermal in origin (7). Generally, pHPT is caused by a single adenoma (80–85%) or four-gland hyperplasia (10–15%). Parathyroid carcinomas are rare and account for <1% of pHPT cases. By contrast, four-gland hyperplasia is the rule in sHPT, ranging from diffuse chief cell hyperplasia to nodular formations. Significantly increased proliferation and apoptosis were both observed in pHPT and sHPT in comparison with normal parathyroids (8). Nonetheless, different genetic alterations are implicated in the development of different types of hyperparathyroidism (9,10).

Although pHPT mainly occurs as a sporadic disease, it may be part of a hereditary syndrome (e.g., multiple endocrine neoplasia types 1 and 2A). On the other hand, chronic renal failure is the primary cause of sHPT. The phenomenon that different pathophysiology and genetic alterations of hyperparathyroidism lead to partially overlapping histological phenotype is notable. However, to the best of our knowledge, there have been no reports comparing transcriptional alterations in different types of hyperparathyroidism. In the present study, the gene expression differences between pHPT and sHPT were analyzed and molecular pathways that are dysregulated in different contexts of hyperparathyroidism were identified.

Materials and methods

Patients and tissue samples

The present study was approved by the Institutional Review Board of MacKay Memorial Hospital (Taipei, Taiwan; approval no. 11MMHIS194), and all patients gave written informed consent. Parathyroid samples were obtained from patients undergoing surgical treatment of hyperparathyroidism at MacKay Memorial Hospital, Taipei, Taiwan (11). All samples were snap frozen in liquid nitrogen within 10 min of resection and stored at −80°C. Diagnosis was histologically confirmed by a senior endocrine pathologist using hematoxylin-eosin staining.

RNA extraction

Total RNA was extracted from homogenized frozen tissue samples using TRIzol reagent (Life Technologies, ThermoFisher Scientific, Inc., Waltham, MA, USA) and purified using the RNeasy Plus Mini Kit (Qiagen, Valencia, CA, USA) according to the manufacturer's recommendations (12). Sample purity was confirmed by measuring ratios of sample absorbance at 260 and 280 nm (ranging from 1.8 to 2.2). The quality of RNA was determined before labeling using the 2,100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA).

Microarray hybridizations

A total of 200 µg of total RNA was amplified by a Low Input Quick Amp Labeling Kit (Agilent Technologies) and labeled with Cy3 during the in vitro transcription process. Cy3-labled cRNA (600 µg) was fragmented to an average size of ~50–100 nucleotides by incubation with fragmentation buffer at 60°C for 30 min. Fragmented labeled cRNA was then pooled and hybridized to Agilent SurePrint G3 Human Gene Expression v2 8×60K Microarray at 65°C for 17 h. After washing and drying, microarrays were scanned with an Agilent microarray scanner at 535 nm for Cy3. Scanned images were analyzed by Feature Extraction software version 10.5.1.1 (Agilent Technologies) to quantify signal and background intensity.

Data analysis and comparison with public microarray data

The microarray data were subjected to linear normalization to allow comparison between arrays. Hierarchical cluster analysis was performed with Cluster 3.0 (bonsai.hgc.jp/~mdehoon/software/cluster/software.htm), and heat maps were constructed with Java Treeview software (www.princeton.edu/~abarysh/treeview/). A public microarray dataset (GSE10317) was retrieved from the National Center for Biotechnology Information Gene Expression Omnibus (www.ncbi.nlm.nih.gov/geo/). GSE10317 comprises gene expression data of a case of pHPT (13). Gene expression levels of the parathyroid tumor and normal parathyroid tissue were analyzed using Affymetrix Human Genome U133 Plus 2.0 Arrays. (Affymetrix, Inc., Santa Clara, CA, USA)

t statistics were used to estimate the significance of expression difference between pHPT and sHPT. R software version 3.0.2 (www.r-project.org) was used for Bayes-regularized t tests. Associated P-values were adjusted for multiple testing by controlling for a false discovery rate <5% using the Benjamini-Hochberg procedure (14), and adjusted P<0.05 was considered to indicate a statistically significant difference. For the GSE10317 data, probes with a differential expression of at least 2-fold were considered to be significant. A meta-signature that characterized the intersection of differentially expressed genes from both datasets were constructed. Genes that demonstrated significantly altered expression changes in the same direction for both dataset were considered to be pHPT-associated. The intersection of differentially expressed genes of the two dataset in the opposite direction was considered to be sHPT-associated.

Gene Ontology (GO) terms and Kyoto Encyclopedia of Genes and Genomes (KEGG; www.genome.jp/kegg/) pathway analyses were performed to annotate the biological functions and pathways in which the aberrantly expressed genes of pHPT and sHPT were involved.

Results

Microarray gene expression analyses were performed in parathyroid tissues from 2 pHPT and 3 sHPT patients. The 2 pHPT patients were female and had single parathyroid adenoma. All sHPT patients including 2 women and 1 man had four-gland nodular hyperplasia. Unsupervised hierarchical clustering analysis for the expression of all genes revealed two natural subgroups containing pHPT and sHPT, respectively.

A meta-signature was constructed to represent an intersection of two sets of differential expression profile. Based on predefined criteria, 339 genes were upregulated and 261 genes were downregulated in pHPT. The ten most common leading-edge genes are summarized in Tables I and II. A total of 218 genes were upregulated and 367 genes were downregulated in sHPT. The top upregulated and downregulated genes are shown in Tables III and IV, respectively. A heat map generated from the most differently expressed genes is presented in Fig. 1.

Figure 1. Hierarchical clustering of microarray data in patients with hyperparathyroidism.

Table I. Upregulated genes in primary hyperparathyroidism.

Table I.

Upregulated genes in primary hyperparathyroidism.

Probe	Gene	Accession #	Description
A_23_P130333	TTR	NM_000371	Homo sapiens transthyretin
A_33_P3814721	INSC	NM_001031853	Homo sapiens inscuteable homolog (Drosophila), transcript variant 1
A_32_P224525	COL6A6	NM_001102608	Homo sapiens collagen, type VI, alpha 6
A_33_P3400273	SELL	NM_000655	Homo sapiens selectin L, transcript variant 1
A_24_P252364	NRCAM	NM_001037132	Homo sapiens neuronal cell adhesion molecule, transcript variant 1
A_23_P157333	EPHA1	NM_005232	Homo sapiens EPH receptor A1
A_23_P350396	CDSN	NM_001264	Homo sapiens corneodesmosin
A_33_P3244728	LRP2	NM_004525	Homo sapiens low density lipoprotein receptor-related protein 2
A_23_P374689	GAD1	NM_000817	Homo sapiens glutamate decarboxylase 1 (brain, 67kDa), transcript variant GAD67
A_23_P71946	BSPRY	NM_017688	Homo sapiens B-box and SPRY domain containing

Table II. Downregulated genes in primary hyperparathyroidism.

Table II.

Downregulated genes in primary hyperparathyroidism.

Probe	Gene	Accession #	Description
A_23_P36658	MGST1	NM_145791	Homo sapiens microsomal glutathione S-transferase 1, transcript variant 3
A_33_P3300253	PTPN20B	NM_001042357	Homo sapiens protein tyrosine phosphatase, non-receptor type 20B, transcript variant 1
A_23_P74609	G0S2	NM_015714	Homo sapiens G0/G1 switch 2
A_33_P3251522	AQPEP	NM_173800	Homo sapiens laeverin
A_33_P3400763	PLIN4	NM_001080400	Homo sapiens perilipin 4
A_23_P23783	MYOC	NM_000261	Homo sapiens myocilin, trabecular meshwork inducible glucocorticoid response
A_21_P0000096	CPXM1	NM_019609	Homo sapiens carboxypeptidase X (M14 family), member 1, transcript variant 1
A_23_P258310	PXDNL	NM_144651	Homo sapiens peroxidasin homolog (Drosophila)-like
A_23_P55270	CCL18	NM_002988	Homo sapiens chemokine (C-C motif) ligand 18 (pulmonary and activation-regulated)
A_23_P86599	DMBT1	NM_007329	Homo sapiens deleted in malignant brain tumors 1, transcript variant 2

Table III. Upregulated genes in secondary hyperparathyroidism.

Table III.

Upregulated genes in secondary hyperparathyroidism.

Probe	Gene	Accession #	Description
A_24_P268685	SLC6A1	NM_003042	Homo sapiens solute carrier family 6 (neurotransmitter transporter), member 1
A_23_P503064	KL	NM_004795	Homo sapiens klotho
A_24_P73577	ALDH1A2	NM_170697	Homo sapiens aldehyde dehydrogenase 1 family, member A2, transcript variant 3
A_23_P1682	TMEM45B	NM_138788	Homo sapiens transmembrane protein 45B
A_23_P95930	HMGA2	NM_003483	Homo sapiens high mobility group AT-hook 2, transcript variant 1
A_24_P240187	LRRN1	NM_020873	Homo sapiens leucine rich repeat neuronal 1
A_23_P143348	OVOL2	NM_021220	Homo sapiens ovo-like zinc finger 2
A_32_P199429	NCAM2	NM_004540	Homo sapiens neural cell adhesion molecule 2
A_23_P370830	KLHL14	NM_020805	Homo sapiens kelch-like family member 14
A_23_P99253	LIN7A	NM_004664	Homo sapiens lin-7 homolog A (C. elegans)

Table IV. Downregulated genes in secondary hyperparathyroidism.

Table IV.

Downregulated genes in secondary hyperparathyroidism.

Probe	Gene	Accession #	Description
A_23_P144778	CKMT2	NM_001825	Homo sapiens creatine kinase, mitochondrial 2 (sarcomeric), transcript variant 1
A_33_P3319248	ZNF705G	NM_001164457	Homo sapiens zinc finger protein 705G
A_23_P131801	SGK2	NM_170693	Homo sapiens serum/glucocorticoid regulated kinase 2, transcript variant 1
A_23_P30554	KCNIP1	NM_001034837	Homo sapiens Kv channel interacting protein 1, transcript variant 1
A_23_P129085	SPESP1	NM_145658	Homo sapiens sperm equatorial segment protein 1
A_23_P363954	THRSP	NM_003251	Homo sapiens thyroid hormone responsive
A_32_P134007	XKR4	NM_052898	Homo sapiens XK, Kell blood group complex subunit-related family, member 4
A_33_P3423230	PCDH19	NM_001184880	Homo sapiens protocadherin 19, transcript variant 3
A_33_P3227793	CGREF1	NM_006569	Homo sapiens cell growth regulator with EF-hand domain 1, transcript variant 1
A_23_P58676	NPR3	NM_001204375	Homo sapiens natriuretic peptide receptor C/guanylate cyclase C (atrionatriuretic peptide receptor C), transcript variant 1

The gene function annotations were evaluated according to the GO and KEGG pathway databases. For genes with differential expression in pHPT, involved molecular functions in order were: Fatty-acid ligase activity, calcium ion binding, alkali metal ion binding, ligase activity forming carbon-sulfur bonds, and symporter activity. Involved biological processes were: Ion transport, cell adhesion, biological adhesion, cation transport, and metal ion transport. Dysregulated pathways in pHPT are presented in Table V.

Table V. Pathway analysis of genes with differential expression in primary hyperparathyroidism.

Table V.

Pathway analysis of genes with differential expression in primary hyperparathyroidism.

Pathway	Count	P-value	Genes
Cell adhesion molecules	9	0.022	HLA-DQB1, NRCAM, ALCAM, SDC1, NRXN3, CD40LG, SELL, ITGA8, ITGA4
PPAR signaling pathway	6	0.034	HMGCS2, APOA5, APOC3, SLC27A6, ACSL3, SLC27A2
Neuroactive ligand-receptor interaction	13	0.035	CSH1, S1PR3, GABRG2, PTGER3, PRLR, RXFP1, HTR7, GRIN2A, TAAR1, ADRA1A, GABBR2, PTGFR, GCGR

Molecular interaction and networks contributing to sHPT were identified. Molecular functions involved in sHPT were: Retinal dehydrogenase activity, steroid dehydrogenase activity acting on the CH-OH group of donors NAD or NADP as acceptor, carbohydrate binding, sulfotransferase activity, and acetylgalactosaminyltransferase activity. Involved biological processes were: Cell adhesion, biological adhesion, secondary metabolic process, skeletal system development, and regulation of nucleotide metabolic process. Dysregulated pathways in sHPT are listed in Table VI.

Table VI. Pathway analysis of genes with differential expression in secondary hyperparathyroidism.

Table VI.

Pathway analysis of genes with differential expression in secondary hyperparathyroidism.

Pathway	Count	P-value	Genes
Tryptophan metabolism	6	0.003	TDO2, CYP1B1, MAOA, AOX1, ALDH2, INMT
Tight junction	10	0.007	PRKCQ, INADL, MYH11, ACTN1, CLDN10, MYH7, CLDN11, CTNNA3, CTNNA2, MYL9
Renin-angiotensin system	4	0.009	ACE2, MAS1, ANPEP, CTSG
Steroid hormone biosynthesis	5	0.030	AKR1C3, CYP3A5, CYP1B1, AKR1C4, CYP19A1
O-glycan biosynthesis	4	0.041	GALNT3, GCNT1, GALNT13, GALNT14

Discussion

The exact mechanism underlying the development of pHPT remains poorly understood (15). In patients with multiglandular pHPT, independent genetic events may be present in separate glands within the same individual (16). Previous studies have indicated that parathyroid adenomas typically harbor few somatic variants (17,18). Mutations in the MEN1 tumor suppressor gene and alterations in the CCND1 (cyclin D1/PRAD1) oncogene represent the major driver in sporadic parathyroid tumorigenesis. Cell cycle regulators (including CDC73 aka HRPT2), growth factors, apoptosis-inducing ligands, death receptors, and other transmitter substances have also been implicated in the pathogenesis (19,20). Although there is an abundance of data examining the gene expression profiles of parathyroid adenoma, few studies have been performed in comparing gene expressions of pHPT with sHPT tissues.

In a previous study examining clonality in pHPT and sHPT, 7/8 pHPT glands (6 adenomas and 2 hyperplasias) exhibited monoclonal proliferation (21). One parathyroid adenoma demonstrated a polyclonal pattern. In an sHPT patient, one of the 3 hyperplastic lesions was monoclonal and the other 2 lesions were polyclonal. This finding underscores the complexity of pHPT pathogenesis. The results from later studies also attest that pHPT can arise by clonal and polyclonal mechanisms (22). This is different from sHPT in which polyclonal growth transforms to monoclonal proliferation during disease progression (23). Nonetheless, there remains the possibility that similar mechanisms regulate parathyroid cell growth in both entities. The current study represents the first effort to compare the transcriptional profiles between pHPT and sHPT using modern microarray technology.

One of the limitations of the study design is the lack of transcriptional data from normal parathyroid controls. To overcome this ethical constraint, additional data were extracted from a publicly available database and combined to form a meta-signature. Regardless, the aberrantly regulated pathways that the present study identified in different types of hyperparathyroidism provide critical insights into the differences in pathophysiology. For instance, cell adhesion molecules were upregulated in pHPT but downregulated in sHPT. It is well known that nodular hyperplasia in patients with chronic kidney disease is associated with progressive downregulation of calcium-sensing receptor and vitamin D receptor (6). Activation of calcium-sensing receptor may potentiate cell adhesion by promoting integrin binding to a fibronectin-rich matrix (24). It is therefore reasonable that downregulated tight junction and cell adhesion were among the dysregulated pathways in sHPT. Conversely, expression of some adhesion molecules including selectin L and neuronal cell adhesion molecule were significantly increased in pHPT (Table I).

It was also noted that renin-angiotensin system was among the dysregulated pathways in sHPT. Increasing evidence links the renin-angiotensin-aldosterone system to calcium regulatory systems (25). A high calcium diet was demonstrated to downregulate angiotensin-converting enzyme of the kidney in experimental renal failure (26). In the present study, it was demonstrated that the expression of angiotensin I converting enzyme 2 (ACE2) was downregulated in sHPT. Recently, it has been shown that angiotensin II infusion acutely stimulated the secretion of parathyroid hormone in a dose-dependent manner (27). It is possible that downregulation of angiotensin-converting enzymes in sHPT results from a negative feedback mechanism to reduce further stimulation from angiotensin II. Renin-angiotensin system may be a potential target for therapeutic intervention in hyperparathyroidism.

The result of the present study point to alterations in peroxisome proliferator-activated receptor pathway in addition to fatty acid and amino acid metabolism in pHPT and sHPT. Metabolic aberrations have already been established as serving essential roles in imaging studies of hyperparathyroidism. Uptake and accumulation of technetium-99m sestamibi in mitochondria-rich oxyphil cells are employed as the basis of scintigraphic detection of hyperfunctional parathyroids (28). In addition, various tracers for positron emission tomographic scan have been exploited in localization of abnormal parathyroid glands (29,30). Metabolic reprogramming in neoplasms has recently been indicated as another general hallmark of cancer. Nonetheless, metabolic rearrangements in hyperparathyroidism remain a virtually untapped area of investigation. Elucidating the complex interplay between calcium homeostasis and parathyroid metabolic activity is an exciting challenge for future research.

In conclusion, the present study demonstrates that different pathophysiology led to differential gene profiling in hyperparathyroidism. Systemic analysis and annotated pathway resources were used to identify several pathways that are dysregulated in hyperparathyroidism which may be targets of interest.

Acknowledgements

The present study was supported by a grant (grant no. MOST-103-2314-B-195-015-MY3) from the Ministry of Science and Technology of Taiwan. The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Parts of this paper were presented at the 17th European Congress of Endocrinology (Dublin, Ireland, May 2015).

References