Genetic Profiling of Synchronous Pituitary Corticotroph Adenomas

Abstract

Purpose: Double or multiple pituitary adenomas account for only 1.6-3.3% of all corticotroph tumors. We sought to better understand the underlying molecular pathogenesis of 2 distinct corticotroph adenomas that were encountered in a 43-year-old female.

Methods: Two distinct histopathologically confirmed corticotroph adenomas were submitted for whole exome sequencing (WES) together with blood sample. The functional effects of identified pathogenic variants on murine corticotroph tumor pro-opio-melanocortin (POMC) transcription and proliferation were characterized.

Results: WES demonstrated a loss-of-function variant in the G-protein coupled receptor 162 [GPR162 (R218*)] in the right corticotroph tumor, and a novel missense variant in ubiquitin specific peptidase 8 [USP8 (P681Q)] in the left corticotroph tumor. Compared to wild-type GPR162 which potently suppressed POMC transcription, the premature stop-gain GPR162 variant (R218*) found in our patient exhibited a reduced POMC transcription inhibitory effect. The novel USP8 variant (P681Q) found in the contra-lateral tumor led to increased POMC transcription similar to the well characterized USP8 hotspot variant S718P. Interestingly, the patient also had a germline variant in the 21-alpha-hydroxylase gene (CYP21A2 p.A392T) although she did not exhibit a phenotype consistent with congenital adrenal hyperplasia. The CYP21A2 transcript and protein were absent in both corticotroph tumors from the index case whereas the protein expression was demonstrated in a series of 9 corticotroph adenomas.

Conclusion: We hypothesize that the germline CYP21A2 variant by increasing corticotroph cell stimulation may have acted in a permissive way to facilitate the additional somatic mutations which led to development of the 2 distinct corticotroph tumors.

Keywords: CYP21A2; Cushing disease; GPR162; USP8; single cell RNAseq; whole exosome sequencing.

Figure 1. MRI images of index case and introduction of sequencing and analysis workflow.

(A) Contrast-enhanced T1-weighted coronal (right) and sagittal (left) MRI images in the index case demonstrating left-sided hypoenhancing lesion (arrowed). (B) Schematic workflow summarizing the steps in pituitary tumor retrieval and whole-exome sequencing and data processing using automated pipelines in the index case corticotroph tumors. (C) Illustration of the analytic Alters used to identify somatic and germline variants. (D) Summary of identified somatic and germline variants.

Figure 2. Functional studies of the somatic variant GP162 R218* identified in the right tumor.

(A) Murine corticotroph tumor AtT20 cells were transiently transfected with siRNA GPR162 and siRNA control. GPR612 knockdown efficiency (left) and POMC mRNA expression (right) were analyzed by real time PCR. (B) AtT20 cells were transiently transfected with constructs expressing wild type (WT) GPR162 and vector controls. Twenty-four hours later, the second round of transfection was performed to introduce siRNA GPR162 for rescue experiments. POMC mRNA expression was analyzed by real time PCR. (C & D) Constructs expressing wild type (WT) and GPR 162 (R218*) variant were transiently transfected into AtT20 cells. The effects of the ectopic expression of WT and mutant GPR162 on basal POMC mRNA expression were assessed by real time PCR (C), and cell proliferation rate was assessed by CellTiterGlo assay (D). (E) Dex-induced POMC inhibition were evaluated by real time PCR following transient transfection of WT-GPR162 and R218*. The Dex treatment was at 10nM and 100nM for 24h, and DMSO was used as vehicle control.

Figure 3. Role of USP8 variant Q681P identified in the left tumor in POMC regulation and EGFR ubiquitination.

(A & B) Wild type and mutant USP8 (P681Q and S718P) plasmids were transiently transfected into AtT20 cells, and the changes in POMC mRNA expression (A) and proliferation rate (B) were assessed by real time PCR (A) and CellTiterGlo assay (B). (C) AtT20 cells were transiently transfected with GFP-EGFR in combination with either WT-USP8, P681Q, S718P variants or empty vector. Twenty-four hours later, corticotroph tumor cell transfectants were treated with (+) or without (−) EGF (at 100ng/mL for 5 min). EGFR ubiquitination was detected by immunoprecipitating with anti-GFP antibody linked agarose bead, and immunoblotting with the indicated antibodies.

Figure 4. Illustration of features of germline variant of CYP21A2.

(A) Location of amino acid changes caused by the CYP21A2 c.1174G>A (p. A392T) germline variant identified in the index case. (B) Depiction of the 3-dimensional structure of CYP21A2 crystal structure and highlighted site of identified CYP21A2 variant. (C) The list of this variant in COSMIC dataset.

Figure 5. Validation of CYP21A2 expression at RNA and protein levels.

(A) scRNAseq demonstrated CYP21A2 mRNA expression in 4 additional sporadic corticotroph tumors (CD1, 2, 4 & 5), but absent CYP21A2 transcript in the index corticotroph tumor (CD3). (B) Using normal adrenal gland as positive control, immunocytochemical CYP21A2 expression was found to be absent in both corticotroph tumors from the index case, but present in 9 additional sporadic corticotroph tumors. magnification x 100. (C) CYP21A2 immunostaining intensity was scored using the histoscore (range, 0 to 400), derived from the product of staining intensity (absent, 0; weak, 100 to 200; intermediate, 200 to 300; strong, 300 to 400) and the percentage of tumor cell staining (range, 0 to 100).