ARID1A is frequently mutated in CRC cells. Thus, it is of fundamental clinical importance to understand its molecular functions and determine whether ARID1A deficiency can be used in CRC therapy. Usually, the loss of ARID1A compromises DNA damage repair, and induced DNA damage burdens may increase the reliance on PARP or ATR-dependent DNA repair pathways of cancer cells to maintain genome integrity and render these cells susceptible to ATR inhibitors (ATRi) and PARP inhibitors (PARPi). The present study evaluated the effect of IR in combination with ATRi, PARPi and Aurora Kinases A/B inhibitor in colon carcinoma cell lines with ARID1A^- and ARID1A⁺. Besides, recent studies from our laboratory showed that its homolog ARID1B is synthetically lethal with ARID1A^-. Thus, the effect of IR in combination with these inhibitors on the sensitivity after ARID1B knockdown was assessed in ARID1A^-and ARID1A⁺ CRC cell lines.

In the present work CRC cell lines with ARID1A⁺ (HCT15, HCT116, Colon320) and ARID1A^- (RKO, SW48, LS180) were used and treated with ATRi, PARPi and Aurora Kinases A/B inhibitor and the consequence on radiation sensitivity was measured by clonogenic survival. In addition, the effect of ARID1B knock-down with siRNA in combination with ATRi on radiation sensitivity was further assessed. The effect on DSB repair signaling as a molecular endpoint for non-homologous and homologous repair was measured by immunofluorescence staining of g-H2AX and RAD51-foci, respectively. In addition, a plasmid based direct-repeat green fluorescent protein (DR-GFP) repair system was used to evaluate its effect specifically on the homologous recombination repair (HRR) of DNA double-strand breaks (DSBs). The concept of selective vulnerability of ARID1A deficient CRC cells to ATRi was also evaluated in an ex-vivo system in cells from untreated CRC patients with ARID1A^- and ARID1A⁺ genotypes using the ATP-tumor chemosensitivity assay (ATP-TCA).

The results showed and confirmed that CRC cell lines with mt ARID1A were selectively sensitized to IR after the silencing of ARID1B by siRNA (P< 0.01). Moreover, our results reconfirm that ATR, PARP and AURKA inhibitors had synthetic lethality in ARID1A^- colon cancer. Among these inhibitors, ATR inhibitors were more effective in ARID1A^- CRC cell lines, which showed reduced viability compared to the viability of ARID1A⁺ (P<0.01) also after the combined treatment with ATRi and IR (P<0.001). Therefore, ATR inhibitors were used for further experiments. Downregulation of ARID1B in addition to ATRi in ARID1A^- CRC cell lines further increased the radiosensitization effect compared with ARID1A⁺ CRC cell lines (P <0.01). Mechanistically, the treatment with ATRi in ARID1A deficient cells significantly decreased the maintenance of radiation induced G2 checkpoint (P <0.01), decreased the number of RAD51 foci induced by radiation in ARID1A^- CRC cell lines compared with ARID1A⁺ CRC cell lines (P<0.01) and thus negatively affect HRR. DR-GFP homologous recombination reporter system confirmed that knock-down of ARID1A selectively decreased HRR in ARID1A^- cell lines. Ex-vivo experiments using the ATP-TCA assay showed that ATRi exerted the highest sensitizing effect in ARID1A negative cells of CRC patients.

In conclusion, the present results provide clues for a pre-clinical and clinical mechanism of action of ARID1A defects as a biomarker for ATR inhibitor response as a single agent or as a synthetic lethal approach in combination with ionizing radiation for CRC with ARID1A^-.