Poly(ADP)-Ribosylation Inhibition: A Promising Approach for Clear Cell Renal Cell Carcinoma Therapy

Abstract

Poly(ADP-ribose) polymerase 1 (PARP-1) and glycohydrolase (PARG) enzymes regulate chromatin structure, transcription activation, and DNA repair by modulating poly(ADP-ribose) (pADPr) level. Interest in PARP-1 inhibitors has soared recently with the recognition of their antitumor efficacy. We have shown that the development of clear cell renal cell carcinoma (ccRCC) is associated with extreme accumulation of pADPr caused by the enhanced expression of PARP-1 and decreased PARG levels. The most severe misregulation of pADPr turnover is found in ccRCC specimens from metastatic lesions. Both, classical NAD-like and non-NAD-like PARP-1 inhibitors reduced viability and clonogenic potential of ccRCC cell lines and suppressed growth of ccRCC xenograft tumors. However, classical NAD-like PARP-1 inhibitors affected viability of normal kidney epithelial cells at high concentrations, while novel non-NAD-like PARP-1 inhibitors exhibited activity against malignant cells only. We have also utilized different approaches to reduce the pADPr level in ccRCC cells by stably overexpressing PARG and demonstrated the prominent antitumor effect of this "back-to-normal" intervention. We also generated ccRCC cell lines with stable overexpression of PARG under doxycycline induction. This genetic approach demonstrated significantly affected malignancy of ccRCC cells. Transcriptome analysis linked observed phenotype with changes in gene expression levels for lipid metabolism, interferon signaling, and angiogenesis pathways along with the changes in expression of key cancer-related genes.

Keywords: PARG; PARP-1 inhibitors; PARylation; RCC; cancer cell; poly(ADP-ribose).

Figure 1

ccRCC cell lines and tumors affected by pADPr turnover. Western blot analysis for pADPr and PARP-1 in ccRCC cell lines (A) and patient-derived tumors (B). “N,” “T,” and “M” indicate “normal,” “tumor,” and “metastatic” samples, respectively. They were obtained from the same patients and are presented in corresponding order. All studied ccRCC samples demonstrate a high level of pADPr. Actin and Tubulin levels are shown as a loading control. (C) Immunohistochemical staining of cryosectioned normal kidney, ccRCC tumor, and metastatic samples from patients reveals significant accumulation of pADPr (brown) in nucleus of tumor and metastatic cells. Representative images are shown. White scale 20 µm. The uncropped Western Blot images can be found in Figure S1.

Figure 2

PARP-1 inhibitors affect ccRCC viability, proliferation and xenograft growth. (A) NAD-like and non-NAD-like PARP-1 inhibitors olaparib and 5F02, respectively, affect patient-derived ccRCC PNX0010 cell line viability. 5F02 eliminated PNX0010 cells (IC50—1.8 μM) with almost no cytotoxicity to normal cells, while olaparib suppressed viability of both normal (IC50—14.7 μM) and cancer cells (IC50—7.2 μM). (B) PARP-1 inhibitors suppress malignancy potential of cancer-derived cells PNX0010. Calculation of cell survival rate was based on clonogenic assays. Cells were plated into 24-well plates. Cells were allowed to adhere overnight and were treated with a non-NAD-like inhibitor (5F02) (magenta), olaparib (blue), and both (red) for 14 days. Colonies were counted and plotted on the graph. Inhibition of pADPr by olaparib and 5F02 is shown in Figure S3. (C) Tumor growth of PNX0010 xenografts is diminished by PARP-1 inhibitor treatments with olaparib (blue) and 5F02 (red), as well as the classical anti-ccRCC drug sunitinib (purple). Growth of tumor without any treatment is shown in green. The most profound effect is demonstrated for 5F02 compared to control group (p < 0.01) and olaparib (p < 0.05) group. (D) Western blot analysis shows the reduction of pADPr level in tumors upon olaparib and 5F02 PARPs inhibition. Actin level is shown as loading control. (E) Treatment of ccRCC cell lines 786-O, 769-P, SK-RC-45, SK-RC-26b and normal NK677 with PARP-1 inhibitor rucaparib demonstrates the potential reduction of proliferation rate. Rucaparib eliminated normal cells (IC50—15.2), SK-RC-45 (IC50—1.38 μM), SK-RC-26b (IC50—1.31 μM), 786-O (IC50—1.29 μM), 769-P (IC50—0.98 μM), PNX0010 cells (IC50—1.4 μM). The uncropped Western Blot images can be found in Figure S3.

Figure 3

PARG overexpression reduces tumorigenicity of ccRCC cell lines. (A) Schematic representation of doxycycline inducible system for PARG overexpression experiment. HIS-tagged PARG cDNA was cloned into lentivirus plasmid pLVX-TetOne™-Puro (Takara) after doxycycline TET-On inducible promoter, followed by P2A cleavable mClover3 green fluorescence protein for expression visualization. Puromycin selection was utilized to expand positively transduced cells. (B) pADPr inhibition in PNX0010 cell lines overexpressing PARG under 500 ng/uL doxycycline stimulation for 72 h. Tubulin level is shown as a loading control. (C) cRCC cells were seeded at 500 cells per well at 6-well plate in triplicates and grown without/with 500 ng/mL doxycycline. Number of colonies were counted after 5–7 days of culture when the colonies reached the size around 50 cells/colony. ccRCC under PARG overexpression form less colonies when grown on plastic, p-value < 0.05. Sample wells with colonies for studied cell lines are shown in Figure S12. (D) ccRCC PARG Tet-On cell lines were grown for 72 h without/with 500 ng/mL doxycycline in triplicates and cell cycle analysis was performed with DNA propidium iodide staining and flow cytometry. Percentages of cells on each G0/G1, S, and G2/M stages were calculated with FlowJo software. Representative histograms are present in Figure S11. For all studied ccRCC cell lines PARG overexpression caused accumulation of cells at G0/G1 stage. (F) cRCC cells were seeded at 2000 cells per well at 6-well plate concentration in 0.3% agar in media without/with 500 ng/mL doxycycline. Fresh 0.3% agar was added each third day. Number of colonies were counted after 2 weeks of culture. 769-P cell line was unable to form colonies. Sample PNX0010 colonies at lower and higher magnification are shown in (E) and sample wells for studied cell lines are presented in Figure S12. White scale bar 150 um. ccRCC under PARG overexpression form less colonies when grown in an unanchored state in agar, (p-value < 0.05). The uncropped Western Blot images can be found in Figure S6.

Figure 4

Lipid biosynthesis, interferon signaling, angiogenesis pathways, and several cancer-driving genes are affected under PARG overexpression in PNX0010 cell line. (A) Differently expressed genes with FDR-corrected p-value < 0.05 and maximum group expression value >1 were subjected to STRING analysis. Three main clusters were identified that are related to lipid metabolism, interferons signaling, and angiogenesis. (B) Differently expressed genes with FDR-corrected p-value < 0.05 and maximum expression value > 1 are plotted and genes with fold change >1.5 are named. (C) GO terms for selected overrepresented pathways (for the whole list see (for the whole list see Figure S17). (D) IPA analysis shows activation of cholesterol biosynthesis and interferon signaling pathways.

Figure 5

pADPr reduction by PARP-1 inhibition and PARG overexpression affects gene expression in similar way. (A) Changes in expression of selected genes measured by qPCR and RNAseq methods in PNX0010 cell line under PARG overexpression. Correspondence of two analyses is demonstrated. * p-value < 0.05 (B) qPCR analysis of genes from (A) in ccRCC cell lines under PARP-1 inhibition with 10 uM rucaparib for 24 h. Downregulation of cancer-related genes ID1-ID3 is shown for both PARG overexpression and PARP-1 inhibition. Opposite effect was shown for metabolism-related genes and colony-stimulation factor 2 coding gene. Serpine1 gene shows concordant results under PARP-1 inhibition and PARG overexpression for 786-O and SK-RC-45 cell lines. * p-value < 0.05.