Computational analysis of data from a genome-wide screening identifies new PARP1 functional interactors as potential therapeutic targets

Abstract

Knowledge of interaction network between different proteins can be a useful tool in cancer therapy. To develop new therapeutic treatments, understanding how these proteins contribute to dysregulated cellular pathways is an important task. PARP1 inhibitors are drugs used in cancer therapy, in particular where DNA repair is defective. It is crucial to find new candidate interactors of PARP1 as new therapeutic targets in order to increase efficacy of PARP1 inhibitors and expand their clinical utility. By a yeast-based genome wide screening, we previously discovered 90 candidate deletion genes that suppress growth-inhibition phenotype conferred by PARP1 in yeast. Here, we performed an integrated and computational analysis to deeply study these genes. First, we identified which pathways these genes are involved in and putative relations with PARP1 through g:Profiler. Then, we studied mutation pattern and their relation to cancer by interrogating COSMIC and DisGeNET database; finally, we evaluated expression and alteration in several cancers with cBioPortal, and the interaction network with GeneMANIA. We identified 12 genes belonging to PARP1-related pathways. We decided to further validate RIT1, INCENP and PSTA1 in MCF7 breast cancer cells. We found that RIT1 and INCENP affected PARylation and PARP1 protein level more significantly in PARP1 inhibited cells. Furthermore, downregulation of RIT1, INCENP and PSAT1 affected olaparib sensitivity of MCF7 cells. Our study identified candidate genes that could have an effect on PARP inhibition therapy. Moreover, we also confirm that yeast-based screenings could be very helpful to identify novel potential therapy factors.

Keywords: PARP1; cancer therapy targets; functional interactors; genome wide screening.

Figure 1. The genes identified by the yeast-based screening belong to several cellular pathways and are mutated in different kinds of cancer

(A) “Cellular metabolism” and “amino-acid and sugar transportation” are the pathways where most genes belong. “mRNA processing and transport” and “Chromatin assembly” are pathways where PARP1 is deeply involved. Also, pathways such as “Chromosome segregation” and “MAPK” are related to PARP1. (B) Total mutations are retrieved from COSMIC. Data are grouped for type of mutation. The number of mutations is reported at the top of any histogram. (C) Number of mutations found in different cancers (from COSMIC) where PARP1 inhibition is currently used in therapy (breast, ovarian and prostate cancer) or where clinical trials are ongoing (intestine, lung, pancreas, melanoma and gastric cancer). Missense and nonsense mutations are shown for each kind of cancer. The number of nonsense and missense mutations of all the genes is reported for any kind of cancer. (D) all the genes are reported to carry somatic mutations in Cancer (COSMIC). Total number of somatic mutations for each gene is indicated. Genes are divided between probable “oncogene” and probable “tumor suppressor” applying 20/20 rule [44].

Figure 2. Functional interaction and correlation between PARP1 and the 12 selected genes

(A) Heat-map of amplification correlation between PARP1 and the selected 12 genes in cancer from cBioPortal. Lowest p-value means highest correlation and most frequent de-regulation of their pathway in cancer. (B) Heat-map of expression correlation between PARP1 and the 12 genes in cancer samples from cBioPortal. Higher Pearson coefficient (p) means high correlation indicating that PARP1 and the selected gene are highly expressed in that kind of cancer/tissue. (C) GeneMANIA network between the selected proteins and PARP1 (left). Physical and genetic interactions, co-expression, shared protein domains co-localization and common pathway are shown by different colors. On the right, network between RIT1 and PARP1 is shown.

Figure 3. Effect of RIT1, INCENP and PSAT1 on PARylation and PARP1 protein level, and olaparip sensitivity

(A) Expression of INCENP, RIT1 and PSAT1 in triple negative and hormone receptor positive (+) breast cancer cell lines: (B) Expression of INCENP, RIT1 and PSAT1 in high grade serous and ovarian clear cell cancer lines. Data were compared and statically analyzed by Student's t-test as indicated: ns, not significant, ^*p<0.05, ^***p<0.001. (C) Expression level measured by qRT-PCR of total RNA extracted from siRNA-transfected cells. Results are the mean of 3-4 experiments ± SD. (D) Western blot analysis of total protein extracts from MCF7 cells transfected with siRNA. Each lane was loaded with 30μg of protein as reported in Materials and Methods. Primary and secondary antibodies are described in the Materials and Methods. Tubulin was evaluated as loading control. (E) PARylation and PARP1 level in cell transfect with siRNA and treated with the PARP inhibitor “olaparib”. Western blot analysis of total protein extracts from siRNA-transfected MCF7 cells to measure PARylation and PARP1 level in olaparib not-treated (left) and treated (right) cells. Each lane was loaded with 30μg of protein as reported in Materials and Methods. Antibodies are described in the Materials and Methods. Tubulin was evaluated as loading control. (F) Densitometry was carried out by direct normalization on tubulin level. In panels 1 and 2, data on PARylation level are shown; in panels 3 and 4, data on PARP1 protein level. Panels 1 and 3 are referred to not-treated cells, panels 2 and 4 to olaparib-treated cells. Results are the mean of 3-4 experiments ± SD. Statistical analysis was performed using the Student's t-test; ns, not significant, ^*p< 0.05, ^**p< 0.01, ^***p< 0.001. (G) Effect on RIT1, INCENP and PSAT1 on olaparib sensitivity in MCF7 cells. Cells are transfected with specific siRNA and after 6 hours re-plated in presence of different doses of olaparib (0, 5, 1, 2, 5, 10 μM) or DMSO. After 48 hours cell viability is calculated as described in Materials and Methods. Results are reported as mean of 3-4 experiments ± SD. Statistical analysis was performed using the Student's t-test by comparing data from siRNA-not-transfected (NT) to siRNA-transfected cells; ^*p< 0.05, ^**p< 0.01, ^***p< 0.001.