Transcriptome analysis of MENX-associated rat pituitary adenomas identifies novel molecular mechanisms involved in the pathogenesis of human pituitary gonadotroph adenomas

Abstract

Gonadotroph adenomas comprise 15-40% of all pituitary tumors, are usually non-functioning and are often large and invasive at presentation. Surgery is the first-choice treatment, but complete resection is not always achieved, leading to high recurrence rates. As gonadotroph adenomas poorly respond to conventional pharmacological therapies, novel treatment strategies are needed. Their identification has been hampered by our incomplete understanding of the molecular pathogenesis of these tumors. Recently, we demonstrated that MENX-affected rats develop gonadotroph adenomas closely resembling their human counterparts. To discover new genes/pathways involved in gonadotroph cells tumorigenesis, we performed transcriptome profiling of rat tumors versus normal pituitary. Adenomas showed overrepresentation of genes involved in cell cycle, development, cell differentiation/proliferation, and lipid metabolism. Bioinformatic analysis identified downstream targets of the transcription factor SF-1 as being up-regulated in rat (and human) adenomas. Meta-analyses demonstrated remarkable similarities between gonadotroph adenomas in rats and humans, and highlighted common dysregulated genes, several of which were not previously implicated in pituitary tumorigenesis. Two such genes, CYP11A1 and NUSAP1, were analyzed in 39 human gonadotroph adenomas by qRT-PCR and found to be up-regulated in 77 and 95% of cases, respectively. Immunohistochemistry detected high P450scc (encoded by CYP11A1) and NuSAP expression in 18 human gonadotroph tumors. In vitro studies demonstrated for the first time that Cyp11a1 is a target of SF-1 in gonadotroph cells and promotes proliferation/survival of rat pituitary adenoma primary cells and cell lines. Our studies reveal clues about the molecular mechanisms driving rat and human gonadotroph adenomas development, and may help identify previously unexplored biomarkers for clinical use.

Fig. 1

Validation of selected genes by qRT-PCR and by IHC on rat PA tissues. a, b RNA was extracted from 11 macrodissected pituitary tumors of adult mutant rats and from five normal pituitaries of wild-type (wt/wt) rats. qRT-PCR was performed using TaqMan primer and probe sets specific to a rat genes up-regulated by array analysis: Ass1, Bmp7, p57, NeuroD1, Lyn, Id2, Cpy11a1, Pttg, Nusap, Nr5a1 and b rat genes down-regulated by array analysis: Lhβ, Fshβ, Pomc, Cga (αGSU). The relative mRNA expression level of the target genes was normalized for input RNA using the housekeeping β2-microglobulin gene and a calibrator RNA always run in parallel and was calculated with the 2^−ΔΔCt formula. The obtained relative value was normalized against the average level in normal pituitary, arbitrarily set to 1. The boundary of the box closest to zero indicates the 25th percentile, the line within the box marks the median, and the boundary of the box farthest from zero indicates the 75th percentile. Error bars above and below the box indicate the 99th and 1th percentiles. *P < 0.05, **P < 0.001. c Formalin-fixed, paraffin-embedded (FFPE) pituitaries from adult wild-type (WT) (left) and mutant (MUT) (right) rats were used. IHC was performed using antibodies against ASS1 and NuSAP and counterstained with hematoxylin. Immunofluorescence was performed with antibodies against P450scc or Cyclin E and nuclei were counterstained with DAPI. T tumor area, N normal adjacent area. Original magnification: ×200; insets: ×400

Fig. 2

Overlap of differentially expressed genes. Venn diagram displaying the overlap among lists of differentially expressed genes from gene expression studies in human gonadotroph/non-functioning PAs [28, 31, 32] and rat PAs (Dataset 1). For details, see “Materials and methods”

Fig. 3

Overexpression of CYP11A1 and NUSAP1 and their encoded proteins in human gonadotroph tumors. a RNA was extracted from frozen pituitary tumors obtained after transsphenoidal surgery. qRT-PCR was performed using TaqMan primer and probe sets specific to human NR5A1, CYP11A1, NUSAP1 and NEUROD1. The relative mRNA expression level of the target genes was normalized for input RNA using human TBP gene expression (housekeeping gene) and a calibrator human brain RNA always run in parallel and was calculated with the 2^−ΔΔCt formula. The obtained relative value was normalized against the average expression of normal pituitary arbitrarily set to 1. The boundary of the box closest to zero indicates the 25th percentile, the line within the box marks the median, and the boundary of the box farthest from zero indicates the 75th percentile. Error bars above and below the box indicate the 99th and 1th percentiles. b Human normal pituitary (top) and gonadotroph adenoma samples were used. IHC was performed using antibodies against P450scc (left) or against NuSAP (right) and counterstained with hematoxylin. Original magnification: ×200; insets: ×400. c Scoring of the expression of NuSAP, Ki-67 and P450scc in a series of human gonadotroph adenomas

Fig. 4

Expression of Ki-67 and NuSAP in rat and human pituitary adenoma cells. a Example of immunofluorescent staining with antibodies against Ki-67 (red) and NuSAP (green) performed on FFPE pituitary tumor tissues from a MENX rat (MENX) and from a patient with gonadotroph adenoma (Human). Nuclei were counterstained with DAPI. Only the tumor area is shown. Original magnification: ×100. b Correlation between NuSAP and Ki-67 labeling index (LI) in 12 human gonadotroph adenoma samples

Fig. 5

Effect of Cyp11a1 on tumor cell proliferation. a Infection of GH3 cells with lentiviral vectors expressing shCyp11a1-GFP or GFP only (MOCK). b GH3 cells were infected with the lentiviral vectors as in “a” and the level of Cyp11a1 was analyzed by qRT-PCR 72 h later. c In samples parallel to “b”, proteins were extracted, and Western blotting was performed to monitor Cyp11a1 expression. The ratios of the band intensities for P450scc versus α-tubulin, normalized against the ratio in the MOCK control (ratio = 1), is indicated. d In samples parallel to “b”, cell proliferation was assessed 72 h after infection using the WST-1 assay. Data were analyzed independently with six replicates each and were expressed as the mean ± SEM. e Growth curve of GH3 cells infected with lentiviral vectors expressing shCyp11a1-GFP or GFP only (MOCK). Cells were trypsinized, stained with Trypan Blue and viable cells were counted. Values are the mean of 3 independent experiments ± SD. f In samples parallel to “d”, the expression of p21 was assessed by qRT-PCR as indicated in the legend of Fig. 1, and is reported relative to the expression level in mock-transfected cells arbitrarily set to 1. g In samples parallel to “d”, caspase 3/7 activity was measured to monitor apoptosis 72 h post-infection. Data were analyzed independently with six replicates each and were expressed as the mean ± SEM. **P < 0.01, ***P < 0.001, versus MOCK

Fig. 6

Effect of Cyp11a1 on primary pituitary tumor cell proliferation. a Infection of rat primary pituitary tumor cells with lentiviral vectors expressing shCyp11a1-GFP or GFP only (MOCK). b Primary pituitary tumor cells from three mutant rats were infected with the lentiviral vectors as in “a” and the level of Cyp11a1 was analyzed by qRT-PCR as indicated in the legend of Fig. 1, and is reported relative to the expression level in mock-transfected cells arbitrarily set to 100. c In samples parallel to “b”, cell proliferation was assessed 72 h after infection using the WST-1 assay. Data were analyzed independently with six replicates each and were expressed as the mean ± SEM. d In samples parallel to “c”, the expression of p21 was assessed by qRT-PCR as indicated in the legend of Fig. 1, and is reported relative to the expression level in mock-transfected cells arbitrarily set to 1. e Caspase 3/7 activity was measured in samples in parallel to “c” to monitor apoptosis in GH3 cells 72 h post-infection. *P < 0.05, **P < 0.01 versus MOCK

Fig. 7

Co-localization of P450scc and SF-1 in rat PAs and reduction of Cyp11a1 expression following down-regulation or inhibition of SF-1 in gonadotroph cells. a Immunofluorescent staining with antibodies against P450scc (red) and SF-1 (green) was performed on FFPE pituitary tumor tissues from MENX mutant rats. Nuclei were counterstained with DAPI. The inset in MERGE shows the co-localization of SF-1 and P450scc in rat tumor cells. Original magnification: ×200; inset: ×400. b LβT2 cells were transfected with scrambled (MOCK) or siRNA oligos against the mouse Nr5a1 gene (siNr5a1). SF-1 and α-tubulin expression levels were assessed by Western blotting 24 and 48 h after transfection. c In samples parallel to “b”, Cyp11a1 expression level was assessed by qRT-PCR as indicated in the legend of Fig. 1, and is reported relative to the expression level in mock-transfected cells arbitrarily set to 100. d LβT2 cells were treated with different concentrations of the SF-1 inhibitor IsoQ (0.1 M and 10 nM) or left untreated (CON). After 24 h, we determined Cyp11a1 expression levels as in “c”. e Primary pituitary tumor cells from mutant rats (n = 3) were incubated with IsoQ as in “d”. After 24 h, we assessed Cyp11a1 expression levels by qRT-PCR as in “c”. Data from primary cultures were analyzed independently with six replicates each and were expressed as the mean ± SEM. *P < 0.05, ***P < 0.001 versus MOCK or CON