Transcriptomic Profiling of Gene Expression Associated with Granulosa Cell Tumor Development in a Mouse Model

Abstract

Ovarian granulosa cell tumors (GCTs) are rare sex cord-stromal tumors, accounting for ~5% ovarian tumors. The etiology of GCTs remains poorly defined. Genetically engineered mouse models are potentially valuable for understanding the pathogenesis of GCTs. Mice harboring constitutively active TGFβ signaling (TGFBR1-CA) develop ovarian GCTs that phenocopy several hormonal and molecular characteristics of human GCTs. To determine molecular alterations in the ovary upon TGFβ signaling activation, we performed transcriptomic profiling of gene expression associated with GCT development using ovaries from 1-month-old TGFBR1-CA mice and age-matched controls. RNA-sequencing and bioinformatics analysis coupled with the validation of select target genes revealed dysregulations of multiple cellular events and signaling molecules/pathways. The differentially expressed genes are enriched not only for known GCT-related pathways and tumorigenic events but also for signaling events potentially mediated by neuroactive ligand-receptor interaction, relaxin signaling, insulin signaling, and complements in TGFBR1-CA ovaries. Additionally, a comparative analysis of our data in mice with genes dysregulated in human GCTs or granulosa cells overexpressing a mutant FOXL2, the genetic hallmark of adult GCTs, identified some common genes altered in both conditions. In summary, this study has revealed the molecular signature of ovarian GCTs in a mouse model that harbors the constitutive activation of TGFBR1. The findings may be further exploited to understand the pathogenesis of a class of poorly defined ovarian tumors.

Keywords: TGFBR1; gene ontology; granulosa cell tumors; signaling pathways.

Figure 1

RNA-seq analysis of ovaries from TGFBR1-CA and control mice at the age of 1 month. (A) Volcano plot depicting genes identified by RNA-seq. Genes with Log₂ fold change greater than 1 (red) or less than −1 (blue) and adjusted p-values less than 0.05 were defined as DE genes. A total of 715 DE genes were identified, with 499 upregulated genes and 216 downregulated genes in the ovaries of TGFBR1-CA mice compared with age-matched controls. n = 5 per group. (B) Heatmap of the top 20 DE genes. Heatmap was generated using row Z scores. A full list of DE genes is shown in Supplementary Table S1.

Figure 2

GO analysis of differentially expressed genes in ovaries between TGFBR1-CA and control mice. (A–C) Top 10 GO biological processes, molecular functions, and cellular components revealed by Enrichr analysis using DE genes from RNA-seq. Full lists are shown in Supplementary Tables S2–S4.

Figure 3

KEGG pathways enriched among DE genes in ovaries between TGFBR1-CA and control mice at the age of 1 month. (A) The top 10 differentially regulated signaling pathways. (B) Clustergram depicting DE genes associated with the top 20 differentially regulated signaling pathways. Note that only 40 out of 715 DE genes are shown. (C) Statistical parameters of the top 20 significantly regulated pathways. A full list of pathways is shown in Supplementary Table S5.

Figure 4

Upregulation of Wisp1 and WNT signaling activation in TGFBR1-CA ovaries. (A) Log₂ fold changes of a subset of WNT-related genes from RNA-seq analysis. (B) qRT-PCR analysis of Wisp1, Bgn, Dcn, Creb5, Cxxc4, and Dach1 in ovaries from TGFBR1-CA and control mice at the age of 1 month. n = 4. * p-value < 0.05, ** p-value < 0.01, and *** p-value < 0.001. (C–H) Immunohistochemical staining of WISP1 and non-phospho CTNNB1 in ovaries from TGFBR1-CA and control mice at the age of 1 month. Isotype-matched rabbit IgG was included as a negative control. n = 3. Scale bar = 50 µm (C–H).

Figure 5

Identification of common regulators of ovarian GCT development between mice and humans. (A) Comparison of DE genes in TGFBR1-CA ovaries with those between the human progressive GCT and normal ovaries. Gene expression of human GCTs was previously reported [28]. (B) Comparison of DE genes in TGFBR1-CA ovaries with those between advanced stage and early stage GCTs. Identification of tumor stage-associated genes was previously documented [27]. (C,D) Comparison of differentially expressed genes in TGFBR1-CA ovaries with those between FOXL2^C134W and vehicle or FOXL2^WT-transfected HGrC1. An immortalized human granulosa cell line was used to reveal genes associated with FOXL2 mutation [11]. Overlapping genes between panels (C,D) are bolded.