ARMC5 mutations in macronodular adrenal hyperplasia with Cushing's syndrome

Abstract

Background: Corticotropin-independent macronodular adrenal hyperplasia may be an incidental finding or it may be identified during evaluation for Cushing's syndrome. Reports of familial cases and the involvement of both adrenal glands suggest a genetic origin of this condition.

Methods: We genotyped blood and tumor DNA obtained from 33 patients with corticotropin-independent macronodular adrenal hyperplasia (12 men and 21 women who were 30 to 73 years of age), using single-nucleotide polymorphism arrays, microsatellite markers, and whole-genome and Sanger sequencing. The effects of armadillo repeat containing 5 (ARMC5) inactivation and overexpression were tested in cell-culture models.

Results: The most frequent somatic chromosome alteration was loss of heterozygosity at 16p (in 8 of 33 patients for whom data were available [24%]). The most frequent mutation identified by means of whole-genome sequencing was in ARMC5, located at 16p11.2. ARMC5 mutations were detected in tumors obtained from 18 of 33 patients (55%). In all cases, both alleles of ARMC5 carried mutations: one germline and the other somatic. In 4 patients with a germline ARMC5 mutation, different nodules from the affected adrenals harbored different secondary ARMC5 alterations. Transcriptome-based classification of corticotropin-independent macronodular adrenal hyperplasia indicated that ARMC5 mutations influenced gene expression, since all cases with mutations clustered together. ARMC5 inactivation decreased steroidogenesis in vitro, and its overexpression altered cell survival.

Conclusions: Some cases of corticotropin-independent macronodular adrenal hyperplasia appear to be genetic, most often with inactivating mutations of ARMC5, a putative tumor-suppressor gene. Genetic testing for this condition, which often has a long and insidious prediagnostic course, might result in earlier identification and better management. (Funded by Agence Nationale de la Recherche and others.).

Figure 1. Chromosomal Alterations in Nodules Identified by Means of Single-Nucleotide Polymorphism (SNP) Arrays

Panel A shows the number of patients with corticotropin-independent macronodular adrenal hyperplasia who had a chromosomal alteration (gain, loss, or copy-neutral loss of heterozygosity) in at least one nodule. The most common event, the copy-neutral loss of heterozygosity in 16p, was detected in 7 of 26 patients. Panel B shows chromosome 16 from a nodule obtained from Patient 5 with a copy-neutral loss of heterozygosity in 16p. The upper part of the panel shows genotypes of the SNPs expressed as the B allele frequency. The lower part of the panel shows the DNA copy number expressed on a base-2 log scale (log ratio), with the red line corresponding to two copies of DNA.

Figure 2. ARMC5 Alterations in Leukocyte and Tumor DNA

Panel A shows the two ARMC5 alterations detected in the tumor DNA obtained from 18 patients. The analysis of leukocyte DNA (left column) shows the germline alteration (blue box). Each tumor under “Somatic Variants in Different Tumors” is shown as a box with a blue rectangle (showing the alteration in leukocyte DNA) and a yellow rectangle (showing the alteration that is detected only in the tumor DNA). Germline DNA from 4 patients was not available (bottom right). Panel B shows Western blot analysis of ARMC5 protein in adrenocortical human cell line H295R transfected with control small interfering RNA (siRNA) or ARMC5 siRNA (left), 3 normal adrenal glands (middle), and 17 samples obtained from patients with corticotropin-independent macronodular adrenal hyperplasia (right): 7 without an ARMC5 alteration and 10 with various types of ARMC5 alterations. LOH denotes loss of heterozygosity, LT left tumor, and RT right tumor.

Figure 3. Analysis of Multiple Nodules Obtained from the Same Patient

Computed tomographic scans and the various nodules present on both adrenal glands in Patient 5 are shown. Each nodule showed the germline defect (blue). A second alteration (yellow) differed between the two adrenals.

Figure 4. Analysis of the Role of ARMC5

Panels A and B show the effects of ARMC5 inactivation by siRNA on the expression of ARMC5, NR5A1, NR0B1, MC2R, CYP17A1, and CYP21A2 and on basal and forskolin-stimulated cortisol synthesis in adrenocortical H295R cells. NS denotes not significant. Panel C shows immunofluorescence staining of transfected H295R cells with different ARMC5-FLAG constructs (a nonmutated construct and two missense mutants: p.R898W and p.L548P). After 6 hours, the cells expressing the ARMC5 constructs had normal morphologic features. After 14 hours, the cells expressing the nonmutant ARMC5 construct (white arrows) were apoptotic (coexpression of the cleaved caspase 3 [green stain] and altered cell morphology with condensed nuclei), and the cells expressing the missense ARMC5 constructs (p.R898W and p.L548P) were not apoptotic. Red staining shows FLAG antibodies (ARMC5 constructs), and blue staining shows 4′,6-diamidino-2-phenylindole (DAPI) (cell nuclei).