Fig 2: FHL2-GLI2 expression results in the expression of sex cord markers in vitro.a Quantitative assessment of CALB2 transcripts in immortalized mesenchymal stem cells (MSCs) and HEK-293 cells stably expressing empty vector (control), wild-type FHL2 (FHL2), wild-type GLI2 (GLI2), truncated GLI2 (tGLI2), or the FHL2-GLI2 fusion. Expression levels were normalized to GAPDH expression, and comparisons of mRNA expression levels were performed relative to control. b Representative western blot analysis of calretinin protein levels in MSCs and HEK-293 cells stably expressing control, FHL2, GLI2, tGLI2, or FHL2-GLI2. Tubulin was used as protein loading control. Quantification (bottom) of protein levels as compared to control. c Representative confocal micrographs of immunofluorescence analysis of calretinin (green) and 4–6-diamidino-2-phenylindole (DAPI, blue) in MSCs and HEK-293 cells stably expressing control, FHL2, GLI2, tGLI2, or FHL2-GLI2 (scale bars, 50 μm). Quantification (bottom) of calretinin intensity/cell relative to control. d Quantitative assessment of FOXL2 transcripts in MSCs and HEK-293 cells stably expressing control, FHL2, GLI2, tGLI2, or FHL2-GLI2. Expression levels were normalized to GAPDH expression, and comparisons of mRNA expression levels were performed relative to control. In a–d, data are representative of at least three independent experiments. Error bars, s.d. of mean; n.s., not significant; *P < 0.05, **P < 0.01, ***P < 0.001; two-tailed unpaired t-test.

Fig 3: FHL2-GLI2 activates the Sonic Hedgehog pathway through DNA binding in GLI target promoters.a Representative confocal micrographs of immunofluorescence analysis of FLAG (red), 4–6-diamidino-2-phenylindole (DAPI, blue), and GFP (green) in HEK-293 cells stably expressing empty vector (control), wild-type FHL2 (FHL2), wild-type GLI2 (GLI2), truncated GLI2 (tGLI2), or FHL2-GLI2. Scale bars, 10 µm. b GLI response element (GLI-RE) luciferase reporter assay of HEK-293 cells stably expressing control, FHL2, GLI2, tGLI2, or FHL2-GLI2 (top), and GLI-RE promoter activity in HEK-293 transiently transfected with control, FHL2-GLI2 and FHL2-GLI2 with R338A-K339A mutations in the Zinc Finger 5 of GLI2 required for DNA binding (FHL2-GLI2 Mutated; bottom). SV40-Renilla was used to normalize transfection efficiency. c Immunoprecipitation assay with SUFU antibody of HEK-293 cells stably expressing control, FHL2, GLI2, tGLI2, or FHL2-GLI2. Western blot analysis using anti-FLAG and anti-SUFU antibodies, and tubulin as loading control (left). GLI-RE promoter activity in HEK-293 stably expressing control, FHL2, GLI2, tGLI2, or FHL2-GLI2 transfected with SUFU or control (right). d Cell titer blue proliferation assay of HEK-293, DAOY and BCC cells stably expressing control, FHL2, GLI2, tGLI2, or FHL2-GLI2 treated with 20 µM GANT61 or vehicle control (DMSO). GANT, GANT61. e FLAG Chromatin Immunoprecipitation (ChIP) assay of GLI1 and PTCH1 promoters (promoter 1 and 2) in MSC and HEK-293 cells stably expressing either control or FHL2-GLI2. GLI1 and PTCH1 gene body and MYOD1, gene promoters not under GLI regulation, were used as negative controls. In b–d, data are representative of at least three independent experiments. Error bars, s.d. of mean; n.s., not significant; *P < 0.05, **P < 0.01, ***P < 0.001; two-tailed unpaired t-test.

Fig 4: Recurrent FHL2-GLI2 fusion gene in sclerosing stromal tumors of the ovary.a Photograph of the cut section of an ovarian sclerosing stromal tumor (SST; left) displaying classic SST appearance with yellow tissue at periphery and white, central fibrotic depression, and micrographs of hematoxylin & eosin stained representative section at low (top right) and high (bottom right) magnification. Scale bars, 1 cm (left), 200 µm (top right), 50 µm (bottom right). b Schematic representation of the FHL2-GLI2 fusion transcript including the exons and domains involved. The breakpoint of the 5' and 3' partner genes are represented as black vertical lines. Spanning reads are depicted and aligned to the predicted junction sequence. c Schematic representation showing the Reads Per Kilobase per Million (RPKM) mapped read counts of each GLI2 exon. The GLI2 fusion breakpoint is represented as a red dashed line. d Fluorescence in situ hybridization (FISH) of two representative SSTs using a three-color FHL2-GLI2 probe, with 5' GLI2 (orange), 3' GLI2 (red), and 5' FHL2 (green), showing the presence of the FHL2-GLI2 fusion (white arrows). e Representative Sanger sequencing electropherograms of the genomic FHL2-GLI2 breakpoint. f RNA in situ hybridization (RNA-ISH) using custom FHL2-GLI2 probes (red) showing the chimeric FHL2-GLI2 mRNA expression in two representative SSTs harboring the FHL2-GLI2 fusion. g Frequency of the FHL2-GLI2 fusion gene and GLI2 rearrangements in 26 SSTs from this study. h Frequency of the FHL2-GLI2 fusion gene and GLI2 rearrangements in 26 SSTs and frequency of the FHL2-GLI2 fusion gene in 48 other ovarian sex cord-stromal tumors from this study. aGCT, adult-type granulosa cell tumor. i Frequency of FHL2-GLI2 fusion gene in 26 SSTs from this study and in 9950 tumors from 33 cancer types from The Cancer Genome Atlas (TCGA). AML acute myeloid leukemia, PCPG pheochromocytoma and paraganglioma.

Fig 5: Expression of the FHL2-GLI2 fusion results in the activation of the Sonic Hedgehog pathway.a Differential gene expression analysis of human sclerosing stromal tumors (SSTs) subjected to RNA-sequencing (n = 8, this study) and high-grade serous ovarian carcinomas (n = 16; The Cancer Genome Atlas) and other sex cord-stromal tumors (SCSTs, n = 11; this study). Gene expression fold-change is color-coded according to the legend. Only genes significantly differentially expressed (P < 0.05; two-tailed unpaired t-test) are shown. CPM, count per million. b Expression levels of Sonic Hedgehog (SHH) pathway genes in human SSTs (n = 11) and other sex-cord stromal tumors (n = 9) as defined using NanoString. Expression levels and SHH enrichment scores are color-coded according to the legends. ***P < 0.001, Wilcoxon rank test. Hierarchical clustering was performed using complete linkage and Euclidian distance. c Quantitative assessment of the Sonic Hedgehog pathway PTCH1 and GLI1 transcripts in immortalized mesenchymal stem cells (MSCs), HEK-293 and medulloblastoma (DAOY) cells and of PTCH1 and CCND1 transcripts in human basal cell carcinoma (BCC) cells stably expressing empty vector (control), wild-type FHL2 (FHL2), wild-type GLI2 (GLI2), truncated GLI2 (tGLI2), or FHL2-GLI2. Expression levels were normalized to GAPDH expression, and comparisons of mRNA expression levels were performed relative to control. d Representative western blot analysis of PTCH1 and GLI1 protein expression in MSC, HEK-293 and DAOY cells and of PTCH1 and CCND1 in BCC cells stably expressing control, FHL2, GLI2, tGLI2, or FHL2-GLI2. Tubulin was used as protein loading control. Quantification (below) of protein levels as compared to control. In c–d, data are representative of at least three independent experiments. Error bars, s.d. of mean; n.s., not significant; *P < 0.05, **P < 0.01, ***P < 0.001; two-tailed unpaired t-test.