Skip to main content

Exploring ’synthetic lethality’ and ’synthetic viability’ to elucidate responses of breast and prostate cancer cells to DNA damage and treatment resistance

Final Report Summary - SYNVIA (Exploring ’synthetic lethality’ and ’synthetic viability’ to elucidate responses of breast and prostate cancer cells to DNA damage and treatment resistance)

The key overall objective of this Marie Curie project was to contribute to better understanding of the DNA damage response (DDR) mechanisms in human cells to derive new therapies, and help predict individual responses to anti-cancer therapies that exploit DDR deficiencies, which occur commonly in most types of cancer (Ashworth et al., 2011; Bartkova et al., 2005; Hoeijmakers, 2001, Jackson and Bartek, 2009). The recent use of specific inhibitors to down regulate particular DDR pathways in cancer therapy is exemplified by the highly successful clinical trials with PARP1/2 inhibitors on BRCA1/2-deficient breast/ovarian tumours based on synthetic lethality concept (Fong et al., 2009). Recent efforts focus also on the small molecule inhibitors of the major DDR signaling kinases ATM, ATR, DNA-PK, Chk1 and Chk2 (Jackson and Bartek, 2009; Toledo et al., 2011).
On the other hand, also targeted therapies such as PARP1/2 inhibition face the daunting problem of potential acquired resistance, which mechanisms are still not fully clear.

We performed set of through high-content functional siRNA based screens, to identify novel examples of synthetic lethality and synthetic viability combinations, using promising inhibitors of the DDR factors, namely PARP1/2 (olaparib AZD2281/KU0059436) and ATMi ( AZ12618466-012 / KU-60019). ATM inhibition based screens identified a group of potential hits (e.g. BARD1, FBXL18, NEDD4L, RNF168, RNF145, SENP3, SHFM1, SIK2, SUMF1, TOPBP1, TRIM41, UCHL5, USP12, ZBTB25). Selected hits (TOPBP, SHFM1, BARD1, NEDD4L, RNF168) were validated and functionally characterized, using a spectrum of approaches and human cell culture models. Synthetic lethality screens based on PARP1/2 inhibition by olaparib revealed a group of potential hits (e.g. CBLC, CUL1, DHX9, FBXO5, PHF21A, PIAS4,RNF17, TOPBP1, VCP) from which TOPBP1 was selected, validated and subjected to detailed study. We found out that PARP1/2 inhibition in combination with TOPBP1 depletion induces micronucleation and leads to DNA damage measured by γH2AX foci formation. TOPBP1 silencing sensitizes U2OS and Cal51 cancer cells to olaparib, leading to cell death. Our results clearly show that TOBP1depletion renders tumour cells sensitive to PARP1/2 inhibitors. This interaction is a new example of synthetic lethality (Kaelin, 2005), where the lack of one pathway of genome maintenance in the tumour in combination with the subsequent inhibition of another (SSB repair involving PARP1) promotes selective tumour cell death, because of functional redundancy (Aly and Ganesan, 2011; Ashworth et al., 2011; Jackson and Bartek, 2009). Therapeutic potential of discovered interaction will be assessed after completion of ongoing study regarding TOPBP1 level in breast cancer patients samples.
Detailed study revealed that TOPBP1 knockdown delays DNA repair and TOPBP1 is a regulator of RAD51 loading to chromatin. We showed that TOPBP1 interacts with RAD51 and its C-terminal part is required for RAD51 foci formation upon DNA damage. On the other hand, TOPBP1 is not required for DNA resection, RPA loading on chromatin, BRCA1 and BRCA2 recruitment upon DNA damaging agents and irradiation. All obtained data regarding sensitivity to PARP1/2 inhibitors and TOPBP1 role as a regulator of RAD51 loading to chromatin will be published within coming months (manuscript in preparation). Additionally to that, a few more hits from siRNA based screens mentioned above were validated, however synthetic lethal interaction between PARP1 either ATM or selected genes was not as robust as screen results suggested.

A lot effort was put during realization of that project to characterize molecular basis of the example of synthetic viability for the BRCA-deficient tumours with loss of 53BP1. Study was performed on breast cancer cell lines deficient in BRCA1 protein (MDA MB 436 and SUM PT149 and SUM PT135) as well as BRCA1 proficient cell lines (Cal51, MCF7) modified by lentiviral transfection (BRAC1 shRNA and 53BP1 shRNA). In used models effect of HR restoration upon 53BP1 depletion (measured by DNA ends resection, Rad51 foci formation) was not observed, contrary to different model systems (mouse) (Bunting et al., 2010; Bouwman et al., 2010). Better survival of 53BP1 downregulated cells seems not to results from improvement of DNA damage repair capacity because no changes of DNA damage repair dynamics were observed. Decreased mortality of BRCA1 and 53BP1 deficient human breast cell lines is under continuous investigation and obtained results suggest efficient cell cycle arrest as a mechanism protecting those cells from cell death upon olaparib treatment. It should be mentioned here that mechanisms describing molecular basis of competition between BRCA1 and 53BP1 pathways and role of Rif1 protein in this interplay have been already published during duration of that project (Zimmermann et al., 2013; Di Virgilio et al., 2013; Chapman et al., 2013).

Based on data obtained during realization of objectives 2 and 3 of the project, analysis of clinical specimens is ongoing to assess a potential of selected candidate biomarkers (53BP1 and spontaneous PARylation) to predict PARPi based therapy outcome.
From among few commercially available antibodies the most specific were chosen based on extended study on panel of cell lines (Cal51, HCT116, MDA 231, MDA MB 436, SUM PT149, U2OS, BJ fibroblasts) using immunofluorescence and protein level analysis on WB. Optimalization of staining conditions allowed us to applied immunohistochemistry on a large cohort of breast tumour specimens to assess the feasibility of monitoring 53BP1 abundance and localization, in parallel with the BRCA1 and BRCA2 proteins to gain insights into the value of such markers as potential biomarkers for future studies, here correlated with basic clinical parameters and responses to chemotherapy, analogous to previous studies (Bartkova et al., 2007; Bartkova et al., 2008; Jiang et al., 2009; Fagerholm et al., 2008). This task is a part of collaboration between the host laboratory and the Danish Centre for Translational Breast Cancer Research (DCTB), under supervision of the senior researcher J. Bartkova. Preliminary results are promising, however study has not been finished yet, due to relatively limited number of samples to analyze coming from patients before treatment with PARPi.