Targeting SOD1 induces synthetic lethal killing in BLM- and CHEK2-deficient colorectal cancer cells

Abstract

Cancer is a major cause of death throughout the world, and there is a large need for better and more personalized approaches to combat the disease. Over the past decade, synthetic lethal approaches have been developed that are designed to exploit the aberrant molecular origins (i.e. defective genes) that underlie tumorigenesis. BLM and CHEK2 are two evolutionarily conserved genes that are somatically altered in a number of tumor types. Both proteins normally function in preserving genome stability through facilitating the accurate repair of DNA double strand breaks. Thus, uncovering synthetic lethal interactors of BLM and CHEK2 will identify novel candidate drug targets and lead chemical compounds. Here we identify an evolutionarily conserved synthetic lethal interaction between SOD1 and both BLM and CHEK2 in two distinct cell models. Using quantitative imaging microscopy, real-time cellular analyses, colony formation and tumor spheroid models we show that SOD1 silencing and inhibition (ATTM and LCS-1 treatments), or the induction of reactive oxygen species (2ME2 treatment) induces selective killing within BLM- and CHEK2-deficient cells relative to controls. We further show that increases in reactive oxygen species follow SOD1 silencing and inhibition that are associated with the persistence of DNA double strand breaks, and increases in apoptosis. Collectively, these data identify SOD1 as a novel candidate drug target in BLM and CHEK2 cancer contexts, and further suggest that 2ME2, ATTM and LCS-1 are lead therapeutic compounds warranting further pre-clinical study.

Keywords: BLM; CHEK2; SOD1; drug targeting; synthetic lethality.

Figure 1. BLM and CHEK2 are synthetic lethal with SOD1

A. Western blot depicting SOD1 silencing in HCT116 cells with either individual (siSOD1–2 and siSOD1–3) or pooled (siSOD1-P) siRNA duplexes relative to controls (untransfected and siGAPDH); α-Tubulin serves as a loading control. B. Representative low-resolution images depicting the decrease in Hoechst stained nuclei (bottom right quadrant) following SOD1 silencing in BLM-deficient cells. Arrowheads identify nuclei exhibiting apoptotic hallmarks. Scale bars represent 100 μm. C. Graph depicting the statistically significant decrease in BLM- (left) or CHEK2-deficient cells (right) following SOD1 silencing relative to controls. The statistical significance is indicated (ns, not significant; ****, p-value < 0.0001). GAPDH serves as the negative control, while PLK1 is an essential gene used as a positive control for death and a transfection indicator. D. Graphs depicting the SL interaction observed following simultaneous silencing of BLM (left) or CHEK2 (right) with SOD1 in HCT116 cells. Presented are the mean normalized percentages (± SD) for the individual silencing of either BLM (solid squares) or CHEK2 (open squares) and SOD1 (open triangles), and the expected value (grey circles) determined for the dual combined siRNAs as calculated using a multiplicative model. Solid circles identify the actual observed values for the simultaneous dual silencing (i.e. BLM and SOD1, or CHEK2 and SOD1) and are lower than the expected values indicating a SL phenotype.

Figure 2. BLM- and CHEK2-deficient cells are hypersensitive to 2ME2, ATTM and LCS-1

Standard dose response curves for cells (indicated at top) treated with varying concentrations of 2ME2 A. ATTM B. and LCS-1 C. Data are normalized to the respective DMSO-treated controls.

Figure 3. NAC administration rescues the hypersensitivity of BLM-deficient and CHEK2-deficient cells to 2ME2, ATTM and LCS-1

A. RTCA growth curves for control (left), BLM- (middle) and CHEK2-deficient (right) cells treated with DMSO, 2ME2, ATTM or LCS-1. Arrowheads identify the timepoints at which the chemicals were administered. B. RTCA depicting NAC rescue of BLM- (top panels) or CHEK2-deficient cells (bottom panels) treated with 2ME2 (left), ATTM (middle) or LCS-1 (right). Arrowheads identify the timepoints at which the chemicals were administered. Note that NAC addition restores growth back to approximately that of the corresponding DMSO-treated controls for all three compounds and in both cell lines.

Figure 4. DNA DSBs persist in BLM-deficient cells treated with 2ME2, ATTM and LCS-1

A. Representative low-resolution (10×) images presenting the qualitative changes in γ-H2AX and 53BP1 signal intensities within control (left) and BLM-deficient cells (right) treated with DMSO, bleomycin (positive control), 2ME2, ATTM, or LCS-1. Cells were imaged after 2 h (t = 2 h; bleomycin) or 6 h (t = 6 h; DMSO, 2ME2, ATTM and LCS-1) treatments, or following treatment, washout and a 36 h recovery phase (t = 42 h). Nuclei were counterstained with Hoechst, and images were acquired using identical exposure times at each wavelength so that qualitative and quantitative analyses could be performed. Hoechst, γ-H2AX and 53BP1 are pseudo-colored blue, green, and red, respectively, within the merged images. Scale bars represent 100 μm. Note the persistence of γ-H2AX and 53BP1 signal intensities within the BLM-deficient cells following washout and recovery relative to controls. B. Graphs presenting the mean normalized γ-H2AX (left) and 53BP1 (right) signal intensities (± SD) within control and BLM-deficient cells treated with DMSO, bleomycin, 2ME2, ATTM, or LCS-1 or following washout and a 36 h recovery phase (t = 42 h). All data are presented relative to the DMSO-treated controls. Raw, unprocessed images were used to determine γ-H2AX and 53BP1 signal intensities. Note the persistence and statistically significant differences observed for γ-H2AX and 53BP1 following washout and recovery within the BLM-deficient cells relative to controls (ns, not significant; ****, p-value < 0.0001).

Figure 5. 2ME2, ATTM and LCS-1 induce apoptosis in BLM- and CHEK2-deficient cells

A. Representative low-resolution images (10×) presenting the qualitative differences in cleaved Caspase 3 signal intensities within control, BLM- and CHEK2-deficient cells treated with DMSO, staurosporine (positive control), 2ME2, ATTM and LCS-1. Cells were labeled for cleaved Caspase 3, while nuclei were counterstained with Hoechst. All images were collected using identical exposure times at each wavelength so that qualitative and quantitative analyses could be performed. Hoechst and cleaved Caspase 3 are pseudo-colored red and green, respectively within the merged images. Scale bars represent 30 μm. Note the visually striking increases in cleaved Caspase 3 signal intensities within the BLM- and CHEK2-deficient cells treated with 2ME2, ATTM and LCS-1 relative to controls. B. Bar graphs depicting the mean normalized cleaved Caspase 3 signal intensities (± SD) within control, BLM- (left) and CHEK2-deficient (right) cells treated with DMSO, staurosporine, 2ME2, ATTM and LCS-1. All data are presented relative to DMSO treated controls. Cleaved Caspase 3 signal intensities were determined from raw, unprocessed images. Note the statistically significant increase in cleaved Caspase 3 signal intensities within the BLM- and CHEK2-deficient cells relative to controls (ns, not significant; ****, p-value < 0.0001).

Figure 6. 2ME2, ATTM and LCS-1 inhibit growth of 2D and 3D cultures of BLM- and CHEK2-deficient cells

A. Bar graphs depicting statistically significant decreases in the mean number of BLM- and CHEK2-deficient colonies following 2ME2, ATTM and LCS-1 treatments relative to controls (ns, not significant; ****, p-value < 0.0001). Cells were treated for 28-days and the data are presented relative to DMSO treated controls (± SD). B. Bar graphs presenting statistically significant decreases in the size of BLM- and CHEK2-deficient 3D tumor spheres treated with 2ME2, ATTM and LCS-1 relative to controls (ns, not significant; ****, p-value < 0.0001). All spheres were treated for 14-days and the data are presented relative to DMSO treated controls (± SD).