HDAC3 genetic and pharmacologic inhibition radiosensitizes fusion positive rhabdomyosarcoma by promoting DNA double-strand breaks

Abstract

Radiotherapy (RT) plays a critical role in the management of rhabdomyosarcoma (RMS), the prevalent soft tissue sarcoma in childhood. The high risk PAX3-FOXO1 fusion-positive subtype (FP-RMS) is often resistant to RT. We have recently demonstrated that inhibition of class-I histone deacetylases (HDACs) radiosensitizes FP-RMS both in vitro and in vivo. However, HDAC inhibitors exhibited limited success on solid tumors in human clinical trials, at least in part due to the presence of off-target effects. Hence, identifying specific HDAC isoforms that can be targeted to radiosensitize FP-RMS is imperative. We, here, found that only HDAC3 silencing, among all class-I HDACs screened by siRNA, radiosensitizes FP-RMS cells by inhibiting colony formation. Thus, we dissected the effects of HDAC3 depletion using CRISPR/Cas9-dependent HDAC3 knock-out (KO) in FP-RMS cells, which resulted in Endoplasmatic Reticulum Stress activation, ERK inactivation, PARP1- and caspase-dependent apoptosis and reduced stemness when combined with irradiation compared to single treatments. HDAC3 loss-of-function increased DNA damage in irradiated cells augmenting H2AX phosphorylation and DNA double-strand breaks (DSBs) and counteracting irradiation-dependent activation of ATM and DNA-Pkcs as well as Rad51 protein induction. Moreover, HDAC3 depletion hampers FP-RMS tumor growth in vivo and maximally inhibits the growth of irradiated tumors compared to single approaches. We, then, developed a new HDAC3 inhibitor, MC4448, which showed specific cell anti-tumor effects and mirrors the radiosensitizing effects of HDAC3 depletion in vitro synergizing with ERKs inhibition. Overall, our findings dissect the pro-survival role of HDAC3 in FP-RMS and suggest HDAC3 genetic or pharmacologic inhibition as a new promising strategy to overcome radioresistance in this tumor.

Fig. 1. High HDAC3 expression sustain radioresistance in FP-RMS cells.

A Box plot depicting Class-I HDACs among normal skeletal muscle tissue (n = 5), FP- (n = 32) and FN-RMS (n = 66) patients using GSE108022 dataset. B Kaplan–Meyer plot depicting correlation between Class-I HDACs expression level [Williamson dataset and FN- (n = 57) and FP-and FN-RMS (n = 45) patients’ survival (E-TABM-1202)]. C Representative western blot (n = 3) of all Class-I HDACs in RD and RH30 cells transfected with either Control (siCTR) or Class-I HDACs siRNAs (siHDAC1, siHDAC2, siHDAC3, siHDAC8) at 48 h post-transfection. Vinculin was used as loading control. D Representative images of RD and RH30 colonies stained with crystal violet 12 days post seeding. RMS cells were transfected as in B and 48 h later irradiated with 4 Gy. The cells were processed 6 h post IR. E Histograms depicting the plating efficiency of RD and RH30 cells treated as in (D). Graph represents the mean of three independent experiments ±SD, one-way ANOVA. Exact p values are reported in the histograms.

Fig. 2. HDAC3 depletion radiosensitizes FP-RMS cells inducing Caspase3/7 and PARP dependent apoptosis.

A Representative western blot (n = 3) depicting HDAC3 protein level of RH30 and RH4 cells stably expressing Cas9 infected with either sgControl (sgCTR) or sgRNA against HDAC3 (sgHDAC3) at 72 h post-infection and then treated or not with 6 Gy of IR. Vinculin was used as loading control. B Representative images of RH30 and RH4 colonies stained with crystal violet 12 days post seeding. FP-RMS cells were infected as in (A) and 72 h later irradiated or not with 6 Gy. The cells were processed 6 h post IR. C Histograms depicting the plating efficiency of RH30 and RH4 treated as in (B). Graph represents the mean of three independent experiments ±SD, one-way ANOVA. Exact p values are reported in the figure. D Histogram depicting the percentage of Annexin-V positive/7-AAD single- and double-positive RH30 and RH4 cells treated as in (B). Graphs represent the mean of three independent experiments ±SD, one-way ANOVA. Exact p values are reported in the figure. E Histogram depicts the quantification of Caspase 3/7 activity RH30 and RH4 cells treated as in (B). Graphs represent the mean of three independent experiments ±SD, one-way ANOVA. Exact p values are reported in the figure. F Representative western blot (n = 3) depicting levels of Cleaved- PARP1, Bcl2; BAX, p-JNK, and JNK of RH30 and RH4 cells infected with either sgCTR or sgHDAC3 at 72 h post-infection and the treated or not with 6 Gy of IR. Vinculin was used as loading control. G Representative western blot (n = 3) depicting levels of BIP, IRE1α, PERK, ATF6, PDI, Calnexin, CHOP and ERO1α of RH30 and RH4 cells infected with either sgCTR or sgHDAC3 at 72 h post-infection and the treated or not with 6 Gy of IR. GAPDH was used as loading control.

Fig. 3. HDAC3 depletion in combination with IR reduces stemness and tumorigenicity in vitro and suppresses growth in vivo.

A Representative light microscopy pictures of cancer stem cell formation assay on RH30 and RH4 cells infected with either sgControl (sgCTR) or sgRNA against HDAC3 (sgHDAC3) at 72 h post-infection and then treated or not with 6 Gy of IR. Scale bar = 50 μm. B Histogram depicts the quantification of rhabdospheres number per field. n = 3 independent experiments, data presented as mean values ± SD, one-way ANOVA. C Representative western blot (n = 3) of RH30 and RH4 cells treated as in (A) reporting MYCN, pERK and ERK protein levels. Vinculin was used as loading control. D Tumor volume of sgCTR (n = 9), sgHDAC3 (n = 6) of RH30 cells xenografts assessed by caliper measurement represented in mm³ followed for 48 days post-inoculation. 13 days after inoculation mice were irradiated each 3 days for 5 times with a single dose of 2 Gy of IR. Data presented as mean values ± SD, two-way ANOVA. E Tumor weight of transplanted RH30 cell xenografts treated as in (D). Data presented as mean values ± SD, two-way ANOVA.

Fig. 4. HDAC3 depletion in FP-RMS cell lines combined to IR induces DNA damage and impairs DNA repair.

A Representative immunofluorescence images of γH2AX foci (green, left) (n = 3 independent biological replicates) in HDAC3 depleted -RH30 and -RH4 cells 6 h post IR with 6 Gy. DAPI (blue) was used as nuclear counterstain. Images were taken using confocal microscopy with a ×60 oil immersion objective lens. Scale Bar = 10 μm. Scatter plot of γH2AX average intensity per cell number (right) in HDAC3 depleted -RH30 and -RH4 cells after 6 h post IR with 6 Gy. Graph represents the mean of three independent experiments ±SEM, ANOVA one-way test. Exact p-values are reported in the figure. B Analysis of DSBs accumulation by the neutral Comet assay. Representative images are reported above. RH30 and RH4 HDAC3 KO cells were exposed or not to 6 Gy of IR and after 6 h subjected to the neutral Comet assay. In the scatter plot below, data are presented as mean tail moment ±SEM, ANOVA one-way test. Exact p values are reported in the figure. C Representative western blot (n = 3 independent biological replicates) depicting the effect of IR (6 Gy), HDAC3 depletion and their combination on DNA damage. The cells were processed 6 h post IR. pATM (Ser 1981), total ATM, pDNA-PKcs (Thr2609), total pDNA-PKcs, Rad51, and Ku70 protein levels were detected. All protein levels were normalized on Vinculin.

Fig. 5. MC4448 is a new potent and highly selective HDAC3 inhibitor.

A Biochemical data of MC4448 against HDAC1, −3, −4, −6 and −8 at three fixed doses (0.01, 0.3, and 7 µM). B Dose-response curves of RH30 and RH4 cells treated with increasing concentration of MC4448. Graph represents the mean of three independent experiment ±SEM. C Representative western blot (n = 3) of RH30 and RH4 treated with MC4448 IC50 for 24 h showing levels of acetylated H3. Vinculin was used as loading control. D Growth curve analysis of RH30 and RH4 cells treated as in (C). n = 3 independent experiments, data presented as mean values ± SD, Student’s two-tailed t test. E Heatmap depicting dose response effect of MC4448 treatment for 72 h on FP-RMS cells (RH30 and RH4), and Normal Human Skeletal Muscle Myoblast (HSMM) and Normal Fetal Human Lung Fibroblast (MRC5).

Fig. 6. MC4448 treatment mirrors HDAC3 depletion effects radiosensitizing FP-RMS cells.

A Representative images of RH30 and RH4 colonies stained with crystal violet 12 days post seeding. FP-RMS cells were treated with either MC4448 IC50 (40 nM for RH30 and 58 nM for RH4) or DMSO for 24 h and then irradiated or not with 6 Gy. The cells were processed 6 h post IR. B Histograms depicting the plating efficiency of RH30 and RH4 treated as in (A). Graph represents the mean of three independent experiments ±SD, one-way ANOVA. Exact p values are reported in the figure. C Histogram depicting the percentage of Annexin-V positive/7-AAD single- and double-positive RH30 and RH4 cells treated as in (A). Graphs represent the mean of three independent experiments ±SD, one-way ANOVA. Exact p values are reported in the figure. D Histogram depicts the quantification of Caspase 3/7 activity in RH30 and RH4 cells treated as in (A). Graphs represent the mean of three independent experiments ±SD, one-way ANOVA. Exact p-values are reported in the figure. E. Representative western blot (n = 3) depicting levels of Bcl2 and BAX of RH30 and RH4 cells treated with either DMSO or MC4448 IC50 for 24 h and then treated or not with 6 Gy of IR. Vinculin was used as loading control. F Representative western blot (n = 3) depicting levels of BIP, IRE1α, PERK, ATF6, PDI, Calnexin, CHOP, and ERO1α of RH30 and RH4 cells infected with either DMSO or MC4448 at for 24 h and then treated or not with 6 Gy of IR. GAPDH was used as loading control.

Fig. 7. MC4448 treatment in combination with IR reduces stemness and counteracts IR-dependent induction of MYCN and p-ERK.

A Representative light microscopy pictures of cancer stem cell formation assay on RH30 and RH4 cells treated with either DMSO or MC4448 IC50 for 24 h and then treated or not with 6 Gy of IR and grown for 2 weeks. Scale bar = 50 μm. B Histogram depicts the quantification of rhabdosphere numbers per field. n = 3 independent experiments, data presented as mean values ± SD, one-way ANOVA. C Representative western blot (n = 3) of RH30 and RH4 cells treated as in A reporting MYCN, pERK and ERK protein levels. Vinculin was used as loading control.

Fig. 8. The combination of IR and MC4448 inhibitor induces DNA damage and impairs DNA repair.

A Representative immunofluorescence images of γH2AX (green, left) (n = 3 independent biological replicates) in MC4448 treated RH30 and RH4 cells 6 h post IR with 6 Gy. DAPI (blue) was used as nuclear counterstain. Scale Bar = 10 μm. Scatter plot of γH2AX average intensity per cell number (right) in MC4448 treated RH30 and RH4 cells after 6 h post IR with 6 Gy. Graph represents the mean of three independent experiments ±SEM, ANOVA one-way test. Exact p values are reported in the figure. B Analysis of DSBs accumulation by the neutral Comet assay. Representative images are reported above. RH30 and RH4 cells were treated with MC4448 and then exposed to 6 Gy of IR. After 6 h FP-RMS cells were subjected to the neutral Comet assay. In the scatter plot below, data are presented as mean tail moment ± SEM, ANOVA one-way test. Exact p values are reported in the figure. C Representative western blot (n = 2 independent biological replicates) of FP-RMS cells treated with MC4448 and IR (6 Gy). The cells were processed 6 h post IR to evaluate the effect of combination of treatments on DNA damage response. pATM (Ser 1981), total ATM, pDNA-PKcs (Thr2609), total pDNA-PKcs, Rad51, and Ku70 protein levels were detected. All protein levels were normalized on Vinculin.