Identification of ribonucleotide reductase M2 as a potential target for pro-senescence therapy in epithelial ovarian cancer

Abstract

Epithelial ovarian cancer (EOC) is the leading cause of gynecological-related cancer deaths in the United States. There is, therefore, an urgent need to develop novel therapeutic strategies for this devastating disease. Cellular senescence is a state of stable cell growth arrest that acts as an important tumor suppression mechanism. Ribonucleotide reductase M2 (RRM2) plays a key role in regulating the senescence-associated cell growth arrest by controlling biogenesis of 2'-deoxyribonucleoside 5'-triphosphates (dNTPs). The role of RRM2 in EOC remains poorly understood. Here we show that RRM2 is expressed at higher levels in EOCs compared with either normal ovarian surface epithelium (P<0.001) or fallopian tube epithelium (P<0.001). RRM2 expression significantly correlates with the expression of Ki67, a marker of cell proliferation (P<0.001). Moreover, RRM2 expression positively correlates with tumor grade and stage, and high RRM2 expression independently predicts a shorter overall survival in EOC patients (P<0.001). To delineate the functional role of RRM2 in EOC, we knocked down RRM2 expression in a panel of EOC cell lines. Knockdown of RRM2 expression inhibits the growth of human EOC cells. Mechanistically, RRM2 knockdown triggers cellular senescence in these cells. Notably, this correlates with the induction of the DNA damage response, a known mediator of cellular senescence. These data suggest that targeting RRM2 in EOCs by suppressing its activity is a novel pro-senescence therapeutic strategy that has the potential to improve survival of EOC patients.

Keywords: DNA damage response; cell proliferation; cellular senescence; epithelial ovarian cancer; ribonucleotide reductase M2 (RRM2).

Figure 1. RRM2 is upregulated in EOC cells and tumor specimens. (A) Expression of RRM2 protein in 3 individual isolations of normal human ovarian surface epithelial (HOSE) cells and the indicated human EOC cell lines by immunoblotting. β-actin was used as a loading control. (B) Immunohistochemical staining of RRM2 protein expression in normal ovary, normal fallopian tube, and human EOC tissue specimens. Shown is an example of a high-grade serous EOC tissue specimen. Arrows point to the RRM2 positively stained EOC cells.

Figure 2. RRM2 expression correlates with tumor grade and stage and is a poor prognostic marker in EOC patients. (A) RRM2 is upregulated in EOC specimens (n = 99) compared with normal ovarian surface epithelium samples (n = 4). * P < 1 × 10⁻⁷, false discovery rate (FDR) < 1 × 10⁻⁷. (B). RRM2 is upregulated in invasive serous EOC specimens (n = 246) compared with non-invasive, borderline serous tumors of low malignant potential (n = 18). *P < 1 × 10⁻⁷, FDR < 1 × 10⁻⁷. (C) RRM2 is upregulated in advanced stage (stage 3 and 4) EOC specimens (n = 239) compared with early stage (stage 1 and 2) EOC specimens (n = 42). *P = 0.0199, FDR = 0.0199. (D) RRM2 is upregulated in grade 2 EOCs (n = 97) compared with grade 1 EOCs (n = 19) and in grade 3 EOCs (n = 164) compared with grade 1 or grade 2 EOCs. * P < 0.001. Grade 1 vs. grade 2: P = 0.0003, FDR = 0.0006. Grade 1 vs. grade 3: P < 1 × 10⁻⁷, FDR < 1 × 10⁻⁷. Grade 2 vs. grade 3: P = 3.04 × 10⁻⁵ FDR = 3.04 × 10⁻⁵. (E) RRM2 is upregulated in serous (41 samples), endometrioid (n = 37), mucinous (n = 13), and clear cell (n = 8) histotypes of EOCs compared with normal human ovarian surface epithelium (n = 4). *P < 0.0002. Endometrioid vs. normal P = 2.3 × 10⁻⁶, FDR = 4.6 × 10⁻⁶. Serous vs. normal P = 4.4 × 10⁻⁶ FDR = 8.8 × 10⁻⁶. Mucinous vs. normal P = 0.00013, FDR = 0.0003. Clear cell vs. normal P = 8.4 × 10⁻⁵, FDR = 0.00017. (F) RRM2 is higher in C1 (n = 83), C2 (n = 50), C4 (n = 46), and C5 (n = 36) EOC molecular subtypes compared with C3 (n = 28), and C6 (n = 8). * P < 1 × 10⁻⁷, FDR < 1 × 10⁻⁷. C3 vs. C1, C3 vs. C2, C3 vs. C4, and C3 vs. C5: all are statistically significant at the level of P < 0.01 and FDR < 0.01. C6 vs. C1, C6 vs. C2, C6 vs. C4, and C6 vs. C5: all are statistically significant at the level of P < 0.01 and FDR < 0.01. (G) High RRM2 expression predicates a shorter overall survival in EOC patients. Kaplan–Meier curves of overall survival for EOC patients with high RRM2 expression (C1, C2, C4, and C5 EOC molecular subtypes) or low RRM2 expression (C3 and C6 molecular subtypes). Median overall survival 44 mo for high RRM2 vs. 69 mo for low RRM2 in EOC patients. Log rank P < 0.001, Cox P = 0.004 and hazard ratio (HR) for high RRM2 EOC patients = 1.77. Adjusted Cox P = 0.02 and HR = 1.58.

Figure 3. Knockdown of RRM2 inhibits cell growth and proliferation of EOC cells. (A) SKOV3 EOC cells were infected with control or 2 individual shRRM2-encoding lentivirus and selected with 3 μg/ml puromycin. The expression of RRM2 was determined in drug-selected cells by immunoblotting. β-actin was used as a loading control. (B) Same as (A) but labeled with 10 μM BrdU for 30 min to identify the cells that are actively undergoing DNA replication, and the incorporated BrdU was visualized by immunofluorescence staining. DAPI counterstaining was used to visualize cell nuclei. (C) Quantification of (B) 200 cells from each of the indicated groups were examined for BrdU incorporation. Mean of 3 independent experiments with SEM *P < 0.01 compared with controls. (D) Same as (A) but an equal number of cells (3000 cells/well) were seeded in 6-well plates, and the number of cells was counted at the indicated time points. Mean of 3 independent experiments with SD *P < 0.05 compared with controls. (E) Same as (D), but after 2 weeks of culture the plates were stained with 0.05% crystal violet in PBS to visualize focus formation. Shown are representative images of 3 independent experiments. (F) Same as (D), but cells were seeded into soft agar, and 2 weeks later, the colonies were visualized with bright field fluorescence. Shown are representative images of 3 independent experiments. (G) Quantification of (F). Colonies were stained with 1% crystal violet and counted. Mean of 3 independent experiments with SEM *P < 0.05 compared with controls.

Figure 4. Knockdown of RRM2 induces senescence of EOC cells, which correlates with an increase in DNA damage. (A) SKOV3 cells were infected with control or 2 individual shRRM2-encoding lentivirus and selected with 3 μg/ml puromycin. After 7 d in culture, cells were stained for SA-β-gal activity. (B) Quantification of (A). Two hundred cells from each of the indicated groups were examined for SA-β-gal activity. Mean of 3 independent experiments with SEM *P < 0.05 compared with controls. (C) Same as (A), but RRM2 and γH2AX expression were determined by immunoblotting. β-actin was used as a loading control. (D) Same as (A), but γH2AX and 53BP1 foci formation were determined by immunofluorescence. (E) Quantification of (D). Two hundred cells from each of the indicated groups were examined for γH2AX and 53BP1 foci. Mean of 3 independent experiments with SEM *P < 0.05 compared with controls. (F) SKOV3 cells were infected with control or an shRRM2-encoding lentivirus with or without addition of 250 nM of exogenous nucleosides and selected with in with 3 μg/ml puromycin. After 7 d in culture, cells were stained for SA-β-gal activity. (G) Quantification of (F). Two hundred cells from each of the indicated groups were examined for SA-β-gal activity. Mean of 3 independent experiments with SEM. *P < 0.05 compared with controls. #P < 0.05 compared with shRRM2 alone cells. (H) Same as (F) but γH2AX and RRM2 protein expression was determined by immunoblotting. β-actin was used as a loading control. (I) Same as (F) but γH2AX and 53BP1 foci formation were determined by immunofluorescence. (J) Quantification of (I). Two hundred cells from each of the indicated groups were examined for γH2AX and 53BP1 foci. Mean of 3 independent experiments with SEM *P < 0.05 compared with controls. ^#P < 0.05 compared with shRRM2 alone cells.