New role for ZFX in oncogenesis

Parathyroid adenomas are common endocrine neoplasms. They occur in both sexes and are especially frequent in postmenopausal women. The estimated prevalence is 1–4 cases per 1000 persons overall, and up to 20 cases per 1000 women between ages 55–75.¹ True malignancies of the parathyroid glands do occur but they are rare, while typical parathyroid adenomas are benign, causing morbidity through their metabolic activity. The excessive secretion of parathyroid hormone (PTH) from these tumors can result in hypercalcemia, osteoporosis, fractures, and kidney stones. In addition, parathyroid adenomas also constitute an important model system whose study has already yielded fundamental insights. For example, one of modes in which the cyclin D1 gene (PRAD1, CCND1) was originally discovered, and the way cyclins were first directly linked to human neoplasia, was by its identification as a parathyroid adenoma oncogene activated by DNA rearrangement with the upstream regulatory region of the PTH gene.² Of great interest are the still-unknown reasons why cyclin D1 activation in the parathyroid typically yields benign neoplasia, while the same oncogene drives malignant neoplasia in other cell types such as breast, B-lymphoid, and squamous cells. Interest in parathyroid adenomas as potentially holding keys to understanding, reversing, or preventing the development of malignant properties in tumors is heightened by evidence that typical parathyroid adenomas almost never evolve into malignant parathyroid carcinomas,² contrasting with well-known models for such progression e.g., colorectal cancer.

Beyond cyclin D1, which is overexpressed in 20–40% of parathyroid adenomas, their molecular pathogenesis is only partially understood. Tumor suppressor gene MEN1 is also a fully established driver, subject to biallelic inactivation in 12–20%.² Germline and somatic mutations of tumor suppressor HRPT2/CDC73 are important in parathyroid carcinoma, but not in typical adenomas. Infrequent mutations in other candidate tumor genes, most notably EZH2 in 1% of adenomas,³ have been detected through whole-exome sequencing, but have not yet been shown to drive hyperparathyroidism in vivo. Of great interest from a cell cycle perspective is the special role of cyclin-dependent kinase inhibitor (CDKI) genes in parathyroid and other endocrine tumors. Germline mutation in the gene encoding p27kip1 causes multiple endocrine tumors in rodents and humans; variants in the p15, p18, and p21 genes are associated with similar human phenotypes, and are also found in sporadic parathyroid adenomas, suggesting that germline CDKI mutations may confer a low-penetrance predisposition to hyperparathyroidism.² Such findings, including the identification of p27 mutations in small intestine neuroendocrine tumors⁴ and the dearth of such mutations in most human cancers, suggest that the endocrine cell context provides an enhanced susceptibility to tumorigenesis and selective advantage when a CDKI mutation occurs. Future studies of emerging cell cycle-targeting therapeutic agents such as CDK4/6 inhibitors, may find selected tumors of endocrine-related tissue types to be especially responsive.

Given the need for a more comprehensive and integrated view of the molecular landscape responsible for parathyroid tumors, the recent report in Oncoscience implicating transcription factor ZFX in parathyroid tumorigenesis is significant.⁵ Starting with a cohort of parathyroid adenomas subjected to whole exome sequencing, ZFX mutations were identified in nearly 5% of tumors (6/130), one of the more fruitful outcomes of unbiased sequencing approaches in this disease. ZFX, best known for its role in regulating stem cell renewal and differentiation, belongs to the Krueppel C2H2-type zinc finger protein family and contains 13 zinc finger subdomains. Striking in their specificity, each mutation affected one of just 2 consecutive, highly conserved, arginine residues located in the 13^th zinc finger subdomain, converting positively charged R767 or R768 to glutamine, threonine, or leucine. Mutations were somatic in all cases when this could be determined, and their clonality, predicted damage to protein function, and absence from lists of normal variants, all suggest that they contributed a tumorigenic selective advantage and are not passenger alterations. Furthermore, the narrow spectrum of affected codons, and the presence in females of heterozygous mutation of this X-linked but non-X-inactivated gene, suggests that mutant ZFX is functioning as a direct-acting dominant oncogene, conferring a gain-of-function which might involve accentuation of normal transcriptional activities or a qualitative change in target genes.⁵

Regarding these questions of how ZFX mutation may drive neoplasia, other intriguing links between ZFX and tumorigenesis have been reported. In addition to its required role in self-renewal of haematopoietic and embryonic stem cells, ZFX contributes to MYC or NOTCH-induced leukemia induction in mice, and ZFX was also identified as a crucial regulator of specific Hedgehog-induced malignancies in vivo.⁶ Whether these pathways are involved in parathyroid neoplasia merits further study, as does the question of how often similar mutations will be encountered in other types of human tumors – interestingly, an identical ZFX mutation in an endometrial carcinoma, of previously uncertain significance in COSMIC, can now be viewed as probably pathogenetic. Finally, ZFX is a transcriptional target of cyclin D1,⁷ directly linking ZFX to an established parathyroid oncogenic pathway and to cell cycle regulation. Thus, the Oncoscience report will certainly trigger further studies of the biochemical and oncologic consequences of ZFX mutations in relevant experimental systems, and their outcome may well impact the future of personalized/precision medicine.