RUNX2 inhibition disrupts a PAX3::FOXO1-RUNX2 feed-forward loop and dismantles oncogenic gene programs in fusion-positive rhabdomyosarcoma

Abstract

Fusion-positive rhabdomyosarcoma is an aggressive pediatric cancer of skeletal muscle lineage, with a 5-year overall survival of <30% for high-risk disease, and <8% when metastatic. The PAX3::FOXO1 fusion gene, resulting from t(2:13), is a signature driver of fusion-positive rhabdomyosarcoma, but similar to other transcription-factor based fusion genes in other cancers, not currently pharmacologically tractable. To identify novel druggable proteins in fusion-positive rhabdomyosarcoma tumor tissue and cell lines, we performed mRNA-seq of RMS patient tumors and utilizing the human FP-RMS cell lines Rh30 and Rh4, found that the RUNX2 transcription factor was the top druggable dependency. In vitro loss of function studies using genetic (RNAi) or pharmacologic (small molecule CADD522) inhibition showed that RUNX2 suppression inhibited FP-RMS cell growth, induced myogenic differentiation and apoptosis, and phenocopied PAX3::FOXO1 suppression. In vivo loss of function studies using conditional (dox-inducible) or pharmacologic (small molecule CADD522) blockade of tumor growth in a xenograft model system showed that RUNX2 suppression inhibited tumor growth. Mechanistically, we identify a PAX3::FOXO1 feed-forward loop whereby PAX3::FOXO1 binds a RUNX2 enhancer to upregulate gene expression alongside MYOD1, while RUNX2 expression supports the expression of PAX3::FOXO1 at the mRNA and protein level.

Keywords: CADD522; PAX3::FOXO1; RMS; RUNX2; Rhabdomyosarcoma.

Figure 1.

mRNA-seq performed on archival fresh frozen RMS tissues classified according to their histological interpretation. A, Quality check of mapped short read sequencing reads to the human genome in n=5 ARMS (sample IDs: R01, R02, R05, R08, R13) n=4 ERMS (sample IDs: R07, R09, R10, R12), n=3 SCRMS (sample IDs: R06, R14, R14), n=1 PRMS (sample ID: R11) and n=14 SkM (sample IDs: R17–28 and R30) tissues. Most of the reads mapped to protein coding mRNA loci. B, Pearson correlation coefficient (PCC) plot shows a linear correlation between the sample sets. C, Data reduction to two-dimensions via biplot principal component analysis (PCA) shows consistent groups along the PC1 axis that correspond to ARMS (light red), ERMS (olive), SCRMS (blue), PRMS (green) and SkM (pink) tissues. D,E Heat map based hierarchical cluster analysis of selected differentially expressed genes (x-axis) across ARMS and combined ERMS, SCRMS, and PRMS tissues (y-axis) when compared to the SkM with the log₂ fold change in expression also shown. Z-score refers to high (red) and low (blue) gene expression using normalized values when compared to the mean of total sequencing reads. F, Candidate selection of differentially expressed genes across the current study (stage 1), the cancer DepMap (stage 2), Rh30 and Rh4 P3F targets (stage 3). Candidate genes of interest included those with the greatest gene set intersection (stage 4) and with pharmacological inhibitors (stage 5) available. Top candidate is bolded. G, UpSet plot to visualize intersected genes (and therefore generate gene/s-of-interest for downstream studies) across the current study, the cancer DepMap, and Rh30 and Rh4 cell line P3F targets.

Figure 2.

RUNX2 knockdown in human FP-RMS cells (Rh30) impairs classical oncogenic phenotypes in vitro. A, qRT-PCR for RUNX2 quantification in the negative control (NT) Rh30 cells and the three independent knockdown cells (sh1, sh2, sh4). Each black dot represents an independent biological replicate. B, Normalized RUNX2 protein expression in the negative control (NT) Rh30 cells and the three independent knockdown cells (sh1, sh2, sh4) measured at 0 h, 24 h and 48 h. C, Cell growth in negative control and RUNX2 stable knockdown Rh30 cells. Black dots are independent biological replicates. D, Colony formation in negative control and RUNX2 stable knockdown Rh30 cells. E, Colony numbers in negative control and RUNX2 stable knockdown Rh30 cells. Black dots are independent technical replicates. F, RUNX2 knockdown through three independent stably expressed shRUNX2 lentiviral preparations (sh1, sh2, sh4) plus negative control (shNT) in Rh30 cells. Black arrows point to visible areas of differentiation. Cells were imaged at 20X magnification and captured at 48 h and 72 h. Scale bars are 25 μm. G,H, qRT-PCR for known drivers of mesenchymal stem cell differentiation; MYOD1 and OC, following RUNX2 stable knockdown. Black dots are independent biological replicates. I, Cell viability of negative control and RUNX2 stable knockdown Rh30 cells. Black dots are independent technical replicates. J, Quantification of caspase 3/7 mediated apoptosis in negative control and RUNX2 stable knockdown Rh30 cells. Black dots are independent technical replicates. K, Cell cycle analysis in negative control and RUNX2 stable knockdown Rh30 cells. Black dots are independent biological replicates consisting of three technical replicates each.

Figure 3.

Transcriptomic alterations following RUNX2 stable knockdown in vitro. A,B, Data reduction to two-dimensions via biplot PCA shows consistent groups along the PC1 axis that correspond to Rh30 and Rh4 cells RUNX2 knockdown at 0 h post selection (light red), 24 h post selection (olive), 48 h post selection (green), and negative control (NT) cells at identical time points (turquoise- 0 h, blue- 24 h, and pink- 48 h). C,D, Differentially expressed genes between shNT and sh4 RUNX2 knockdown Rh30 and Rh4 cells at 0 h post selection. X-axis= −Log₂FC. Y-axis= −Log10 AdjP. E,F, Heat map based hierarchical cluster analysis following gene set enrichment analysis (GSEA) of Rh30 and Rh4 cells at selected timepoints (x-axis), and cancer hallmark (n=49) and NCI (n=546) gene-sets (y-axis). Key represents normalized enrichment score (NES). G,H Enrichment plots for Rh30 and Rh4 cells following RUNX2 knockdown at 0 h post selection. Selected gene-sets characterize the effect of RUNX2 knockdown on differentiation, cell cycle, apoptosis, and FP-RMS regulatory circuitry. Enrichment score (ES), P-values, and false discovery rate (FDR) are shown.

Figure 4.

Doxycycline inducible RUNX2 knockdown in Rh30 FP-RMS cells. A, qRT-PCR for RUNX2 72 h after 1 μg/mL doxycycline treatment. Black dots represent technical replicates. B, RUNX2 protein expression in RUNX2 knockdown cells (sh2, sh4) following treatment with 1 μg/mL doxycycline for 140 h. C, Doxycycline inducible knockdown of RUNX2 in Rh30 cells using two different shRNA constructs (sh2, sh4). Black arrows point to visible differentiated phenotypes. Cells were imaged at 10X magnification. Scale bars are 100 μm. D, qRT-PCR for PAX3::FOXO1 expression 72 h after 1 μg/mL doxycycline treatment in the RUNX2 knockdown cells (sh2, sh4). Black dots are technical replicates. E, Tumor volume growth curves in mice engrafted with sh2 doxycycline inducible RUNX2 knockdown Rh30 cells and with or without access to doxycycline chow. Stars above each day represent a significant difference between means of control versus treated tumors on that day. F, Tumors excised from the mice and weighed (g). Black dots represent independent biological replicates. G, Representative H&E images of excised tumors taken at 40X magnification plus immunohistochemistry for Ki-67 and CC3. Scale bars are 50 μm. H,I, Three representative tumors from each group (one shown) stained for Ki-67 and CC3 and quantified using ImageJ. J, qRT-PCR for RUNX2, MYOD1, OC, and PAX3::FOXO1 expression following RUNX2 knockdown in vivo. Blak dots represent independent biological replicates.

Figure 5.

PAX3::FOXO1 and RUNX2 reciprocally regulate one another in FP-RMS. A,B, Stable knockdown of PAX3::FOXO1 in Rh4 cells affect on RUNX2 and RUNX1 expression. C,D, qRT-PCR for PAX3::FOXO1 and RUNX2 expression in PAX3::FOXO1 transient knockdown Rh30 cells. Black dots are independent biological replicates. E, Genome browser view of RUNX2. Peaks for the top 9 rows reflect degree of DNA-binding of H3K27ac, H3K9ac, H3K4me3, H3K27me3, P300, MYOD1, and PAX3::FOXO1 using ChIP-seq. Row 10 is open chromatin peaks from ATAC-seq, and rows 11 and 12 are mRNA-seq reads. All rows are in Rh4 cells, counts are in reads per million. F, PAX3::FOXO1, FOXO1, and RUNX2 protein expression in negative control (NT) and puromycin selected RUNX2 knockdown Rh4 cells after 48 h selection. G, PAX3::FOXO1, FOXO1 and RUNX2 protein expression in negative control (NT) and puromycin selected RUNX2 knockdown Rh30 cells after 72 h selection. H, PAX3::FOXO1, FOXO1 and RUNX2 protein expression in negative control (NT) and puromycin selected RUNX2 knockdown Rh30 cells 144 h after selection.

Figure 6.

Rh30 cells treated with the RUNX2 small molecule inhibitor CADD522 in vitro and in vivo. A, qRT-PCR for RUNX2 performed following 24 h of treatment with 0.5 μM and 1.0 μM of CADD522 in vitro. Black dots are independent biological replicates. B, CADD522 treatment effects on colony formation in vitro. Cells were treated with a 12 μM dose every other day for 14 d. C, Colony number after CADD522 treatment. Black dots represent technical replicates. D, Rh30 cells stained with the myogenic marker sarcomere myosin (MF20) when treated with a negative control (DMSO) and different CADD522 concentrations over 72 h. Images were captured at 20X magnification. Scale bars are 50 μm. E, Quantification of MF20 positive cells (dark brown) versus MF20 negative cells (light brown) following CADD522 treatment. F, G, H qRT-PCR for MYOD1, OC, and PAX3::FOXO1 performed following 24 h of treatment with 0.5 μM and 1.0 μM of CADD522 in vitro. Black dots are independent biological replicates. I, Tumor volume growth curves with (n=9) or without 10 mg/kg CADD522 (n=8). J, Kaplan-Meier survival curve for mice treated with 10 mg/kg CADD522. Stars above each day represent a significant difference between means of control versus treated tumors on that day. K, H&E stains on excised tumors and immunohistochemistry (IHC) for Ki-67 and CC3 with and without CADD522 treatment. Images were captured at 40X magnification. Scale bars are 50 μm. L,M, Ki-67 and CC3 quantification. N, qRT-PCR for RUNX2, MYOD1, OC, and PAX3::FOXO1 expression following CADD522 treatment in vivo. Black dots represent independent biological replicates

Figure 7.

Schematic to show PAX3::FOXO1 transcriptionally upregulates RUNX2 expression at the chromatin to prevent FP-RMS terminal differentiation and apoptosis via the defined PAX3::FOXO1 core regulatory circuitry.