Refractory testicular germ cell tumors are highly sensitive to the targeting of polycomb pathway demethylases KDM6A and KDM6B

Abstract

Testicular germ cell tumors (TGCTs) can be treated with cisplatin-based therapy. However, a clinically significant number of cisplatin-resistant patients die from progressive disease as no effective alternatives exist. Curative cisplatin therapy results in acute and life-long toxicities in the young TGCT patient population providing a rationale to decrease cisplatin exposure. In contrast to genetic alterations, recent evidence suggests that epigenetics is a major driving factor for TGCT formation, progression, and response to chemotherapy. Hence, targeting epigenetic pathways with "epidrugs" is one potential relatively unexplored strategy to advance TGCT treatment beyond cisplatin. In this report, we demonstrate for the first time that targeting polycomb demethylases KDM6A and KDM6B with epidrug GSK-J4 can treat both cisplatin-sensitive and -resistant TGCTs. While GSK-J4 had minimal effects alone on TGCT tumor growth in vivo, it dramatically sensitized cisplatin-sensitive and -resistant TGCTs to cisplatin. We validated KDM6A/KDM6B as the target of GSK-J4 since KDM6A/KDM6B genetic depletion had a similar effect to GSK-J4 on cisplatin-mediated anti-tumor activity and transcriptome alterations. Pharmacologic and genetic targeting of KDM6A/KDM6B potentiated or primed the p53-dominant transcriptional response to cisplatin, with also evidence for basal activation of p53. Further, several chromatin modifier genes, including BRD4, lysine demethylases, chromodomain helicase DNA binding proteins, and lysine methyltransferases, were repressed with cisplatin only in KDM6A/KDM6B-targeted cells, implying that KDM6A/KDM6B inhibition sets the stage for extensive chromatin remodeling of TGCT cells upon cisplatin treatment. Our findings demonstrate that targeting polycomb demethylases is a new potent pharmacologic strategy for treating cisplatin resistant TGCTs that warrants clinical development.

Keywords: GSK-126; GSK-J4; Testicular cancer; chemotherapy resistance; cisplatin; epigenetics; polycomb; preclinical; transcriptomics.

Figure 1. Pharmacologic inhibition or induction of the polycomb pathway alters cisplatin sensitivity in TGCT cells but not in glioblastoma, colon, and breast cancer cells.

Parental TGCT cell line 2102EP was treated with EZH2 inhibitor GSK-126 (0.5 μM) for 3 days and cisplatin-resistant TGCT cell line 2102EP-C1 was treated with KDM6A/KDM6B inhibitor GSK-J4 (1.0 μM) for 3 days before 6-hour cisplatin treatments. Cells were assayed for viability 3 days later. U87-MG glioblastoma, HCT116 colon cancer, and MDA-MB-231 and MCF7 breast cancer cells were treated similarly, except cisplatin dosages were higher due to less inherent sensitivity to cisplatin compared to TGCT cells.

Figure 2. Knockdown of polycomb components EZH2 and BMI1 confer cisplatin resistance in TGCT cells, while knockdown of polycomb demethylases KDM6A/KDM6B sensitizes TGCT cells to cisplatin.

(A) RT-PCR demonstrating efficient knockdown of EZH2 and BMI1 in TGCT cells NT2/D1, 2102EP, and 833K. (B) Western blot demonstrating efficient knockdown of EZH2 and BMI1 in 2012EP and 833K cells and repression of H3K27me3 levels with EZH2 knockdown. (C) Cell proliferation and viability assays of control pLKO.1 and EZH2- and BMI1-knockdown cells treated with cisplatin. (D,E) RT-PCR demonstrating single and dual KDM6A- and KDM6B-knockdown in 2102EP, 2012EP-C1, NT2/D1, and NT2/D1-A4 cells and repression of DNMT3B expression upon KDM6A/KDM6B-knockdown. (F) Western blot demonstrating induction of H3K27me3 and repression of DNMT3B expression with single and dual knockdown of KDM6A and KDM6B in 2102EP-C1 cells. (G,H) Cell proliferation and viability assays of control PLK pLKO.1 and KDM6A/KDM6B-knockdown cells treated with cisplatin.

Figure 3. KDM6A/KDM6B inhibitor GSK-J4 and dual KDM6A/KDM6B-knockdown sensitize human TGCT tumor xenografts to cisplatin.

(A) Schematic of GSK-J4, cisplatin, or GSK-J4 + cisplatin treatment schedule for cisplatin-resistant 2102EP-C1 and cisplatin-sensitive NT2/D1 cells xenografts (left) and schematic of cisplatin treatment schedule for KDM6A/KDM6B dual knockdown or control 2102EP-C1 cells (right). (B-J) Depicted are tumor volume, percent change in tumor volume from the day prior to treatment initiation (day 0), and survival of mouse xenografts for 2102EP-C1 and NT2/D1 tumors treated with GSK-J4, cisplatin, or GSK-J4 + cisplatin (left) or xenograft tumors of KDM6A/KDM6B dual knockdown or control 2102EP-C1 cells (right). TV, tumor volume.

Figure 4. Transcriptome analysis of GSK-J4 treated and KDM6A/KDM6B knockdown cells reveals the importance of basal activation of gene expression and p53 signaling in cisplatin sensitization.

(A) Multidimensional scaling (MDS) plots of the three RNA-seq experiments NT2/D1-A4 and 2102EP-C1 cells treated with GSK-J4, cisplatin, or GSK-J4 + cisplatin and KDM6A/KDM6B dual knockdown or control 2102EP-C1 cells treated with cisplatin. (B) Enhanced volcano plots of the three RNA-seq experiments. For the first two experiments, each treatment (cisplatin, GSK-J4, or the combination) is compared to untreated vehicle control. For the third experiment, PLK control + cisplatin, untreated sh6A+6B, and sh6A+6B + cisplatin cells are compared to untreated PLK cells. Significant cutoff is fold-change >1.3 and FDR < 0.05. The number of up- and downregulated genes are provided. (C) Gene set enrichment analysis (GSEA) results corresponding to the volcano plot comparisons for upregulated genes. Top 5 gene sets as determined by normalized enrichment score (NES) from the MSigDB C2 collection are provided. P53 target gene collections are highlighted. The top 20 gene sets enriched for upregulated and downregulated genes for each experimental arm for all three RNA-seq experiments are provided in Supplemental Table S1. Red text highlights p53 target gene sets, including KERLEY_RESPONSE_TO_CISPLATIN, which we previously identified to be a p53 dominate gene set in cisplatin treated NT2/D1 cells (37).

Figure 5. Polycomb demethylase targeting potentiates p53 target gene activation and represses DNMT3B expression in TGCT cells.

(A,B,C) Expression of select p53 target genes across the 4 experimental arms of the three RNA-seq experiments of Figure 4. (D,E,F) DNMT3B expression across the 4 experimental arms of the three RNA-seq experiments of Figure 4.

Figure 6. Transcriptome analysis of KDM6A/KDM6B-knockdown cells reveals cisplatin sensitization is associated with alterations in chromatin remodeling genes upon cisplatin treatment.

(A) Venn diagram comprised of comparing genes upregulated in PLKO.1 cisplatin-treated vs PLKO.1 untreated, sh6A/KDM6B untreated vs PLKO.1 untreated, and sh6A/KDM6B cisplatin vs sh6A/6B untreated. Also depicted are gene set enrichment analysis (GSEA) results corresponding to indicated comparison groups. Top 10 gene sets from GeneOverlap analysis as determined by p value from the MSigDB C2 collection are provided. P53 target gene collections are highlighted. malignant peripheral nerve sheath tumors (MPNSTs), B) Venn diagram comprised of comparing genes downregulated in PLKO.1 cisplatin-treated vs PLKO.1 untreated, sh6A/6B untreated vs PLKO.1 untreated, and sh6A/6B cisplatin-treated vs sh6A/6B untreated. Also depicted are GeneOveralp results corresponding to indicated comparison groups. Top 10 gene sets as determined by p value from the MSigDB C2 collection are provided. The REACTOME_CHROMATIN_MODIFYING_ENZYMES gene set is highlighted. (C) Expression of select chromatin-modifying enzymes and proteins across the 4 experimental arms of the RNA-seq experiment. P53 target gene sets are in red text, H3K27me3/polycomb gene sets are in blue text, REACTOME_CHROMATIN_MODIFYING_ENZYMES gene set is in green text.