DNA methylation and gene expression patterns in adrenal medullary tumors

Abstract

Pheochromocytomas and abdominal paragangliomas are neuroendocrine tumors of the autonomous nervous system, arising in adrenal medulla and sympathetic ganglia throughout the abdomen. These tumors originate from catecholamine-producing chromaffin cells. Neuroblastomas are common childhood tumors in the adrenal medulla and abdominal ganglia. They derive from an unidentified subset of immature sympathetic nerve cells. Like most other cancers, pheochromocytomas, paragangliomas and neuroblastomas arise as a result of a vast array of genetic and epigenetic events that allow them to circumvent the normal mechanisms governing cell growth and death.

It is the purpose of this thesis and the works contained herein to further elucidate the mechanisms that propel the development of these tumors.

It was the purpose of Paper I to assess the involvement of the CDKN2A gene locus and its two resident tumor suppressor genes (TSGs), p16INK4A and p14ARF, previously implicated in the pathogenesis of pheochromocytomas and paragangliomas. p16INK4A promoter hypermethylation was prominent in a subset of malignant paragangliomas, and sequence alterations were observed in the gene. Suppressed p16INK4A RNA and protein expression was also evident in these tumors. SDHB gene mutation status was assessed, and overlapped fully with p16INK4A hypermethylation. In contrast, we found little evidence for p14ARF involvement in pheochromocytomas and paragangliomas.

In Paper II we assessed global and gene specific methylation levels in pheochromocytomas and paragangliomas in relation to clinical phenotype. CpG islands in the promoter regions of 11 tumor suppressor genes were assessed using a quantitative method. A CpG island methylator phenotype (CIMP; defined as hypermethylation in three or more of the assessed TSGs) was found in the same subset of paragangliomas that displayed p16INK4A hypermethylation in Paper I.

In Paper III we strived to verify concerted TSG promoter hypermethylation in a subset of malignant paragangliomas with SDHB mutation in an independent sample series, and to shed light on the temporal occurrence of genetic and epigenetic events in these tumors. Epigenetic changes in the new tumor set mirrored the results from Papers I and II. As in the previous papers, TSG hypermethylation was strongly associated with SDHB mutation, and we show that mutation precedes aberrant TSG methylation and tumor formation.

The purpose of Paper IV was to evaluate the involvement of two RAS effector proteins with tumor suppressor functions, RASSF1A and NORE1A, in pheochromocytomas and paragangliomas. Suppressed RASSF1A and NORE1A mRNA were observed in tumors in comparison to normal reference. In agreement with the previous papers we found RASSF1A hypermethylation in malignant paragangliomas. Reconstitution of Nore1a in rat pheochromocytoma cell line suppressed the ability to grow in soft agar and induced apoptosis, supporting a tumor suppressive function for Nore1a in pheochromocytoma.

In Paper V we assessed global and gene specific methylation of 14 TSGs in neuroblastomas in relation to clinical phenotype. Frequent TSG hypermethylation was observed with neuroblastomas, but no clear-cut correlations could be made with specific tumor features.

In Paper VI we assessed expression and gene methylation of the TSG RIZ in neuroblastomas. We observed suppressed RIZ1 expression in a subset of neuroblastomas with adverse features.

In conclusion we find substantial evidence for epigenetic involvement in TSG inactivation in paragangliomas and neuroblastomas. In light of these findings we propose the evaluation of demethylating agents in the treatment of these tumors.