HGF/c-Met axis drives cancer aggressiveness in the neo-adjuvant setting of ovarian cancer

Abstract

Ovarian cancer is the most lethal gynecologic malignancy. Recently, NACT (Neo Adjuvant Chemotherapy) has been tested as alternative approach for the management of ovarian cancer patients. A biological predictor helpful in selecting patients for NACT would be desirable. This study was aimed at identifying actionable mechanisms of resistance to NACT. Expression of a panel of microRNAs was screened in a discovery set of 85 patients. Analysis of the potential targets was conducted in the same RNAs by calculating significant correlations between microRNAs and genes. Quantitative fluorescent immunohistochemistry was employed in a validation set of 109 patients. MiR-193a-5p was significantly overexpressed in the NACT setting. Analysis of its potential targets demonstrated that this microRNA is also significantly correlated with HGF and MET genes. Analysis of protein expression in samples taken before and after NACT demonstrated that both HGF and c-Met are increased after NACT. Patients who relapse shortly after NACT exhibited the highest relative basal expression of both HGF and c-Met, while the opposite phenomenon was observed in the best responders. Mir-193a-5p, HGF and c-Met expression may help select eligible patients for this modality of treatment. Moreover, inhibitors of this pathway may improve the efficacy of NACT.

Fig. 1

A: Dot plot showing the results of microRNA expression in the discovery set. In x- and y-axis the fold increase (NACT/PDS-CT) is reported along with the statistical significance calculated with Wilcoxon test. In black, red and green are reported the microRNAs not significantly modulated, significantly decreased in NACT and significantly increased in NACT, respectively. B: Dot plot showing the results of gene expression analysis in the discovery set. In x- and y-axis the fold increase (NACT/PDS-CT) is reported along with the statistical significance calculated with Wilcoxon test. Only the genes (red and blue are upregulated and downregulated in NACT, respectively) significantly modulated are shown. C: ANOM analysis of the target significantly modulated in NACT. Y-axis indicates the perecentages of significant target for each microRNA. The blue area indicates the upper/lower decision level for each micro-RNA. Only miR-193a-5p exhibited a significant proportion of targets modulated in the NACT setting. D: Diamond chart showing the expression of miR-193a-5p, HGF and MET in the validation set of 109 NACT patients. The top and bottom of each diamond represent the confidence interval for each group mean. The mean line across the middle of each diamond represents the group mean. A statistically significant increase was noticed for both miR193a-5p and HGF. A significant decrease was also reported for MET (Wilcoxon assay).

Fig. 2

A: Kaplan-Meier curves of the survival of the validation setting for patients grouped according to PFI. Green, blue, and red line is the survival curve for patients belonging to the refractory (PFI 0-3 months), resistant (PFI 3-12 months) and sensitive setting (PFI >12 months). Difference among the three groups is highly significant (p<0.001, Wilcoxon test). B-D: Diamond chart showing the expression of miR-193a-5p (B), HGF (C) and MET (D) according to response to NACT. The top and bottom of each diamond represent the confidence interval for each group mean. The mean line across the middle of each diamond represents the group mean. E-G: Diamond chart reporting the expression of miR-193a-5p (E), HGF (probe 210997_at, F) and MET (probe 211599_x_at, G) in the TCGA dataset. Analysis was restricted to stage IV patients. The top and bottom of each diamond represent the confidence interval for each group mean. The mean line across the middle of each diamond represents the group mean. In B-G double and single asterisks indicate a significant difference at a p value <0.001 or <0.05, respectively (Wilcoxon test).

Fig. 3

A: Representative qPCR analysis of the expression of miR-193a-5p (left column) and let-7g (right column) in OV2774 (top) and SKOV3 (bottom). The red lines marks the cells treated with the transfecting medium, while green yellow and blue lines are for 1, 5 and 10 nM of transfected miR-193a-5p after 48 hours of culture, respectively. B: Bar chart reporting the results of two independent experiments performed in triplicate samples. Expression was normalized for the negative control (only transfection medium=1). Bar and error bars represents mean and SD, respectively. C: Representative qPCR analysis of the expression of MET (left column), HGF (middle column) and TUBB (right column) in OV2774 (top) and SKOV3 (bottom). The red lines marks the cells treated with the transfecting medium, while green yellow and blue lines are for 1, 5 and 10 nM of transfected miR-193a-5p after 48 hours of culture, respectively. D: Bar chart reporting the results of two independent experiments performed in triplicate samples. Expression was normalized for the negative control (only transfection medium=1). Bar and error bars represents mean and SD, respectively.

Fig. 4

A Representative fluorescent immunohistochemistry for HGF (left column) and c-Met. From top to bottom; DAPI channel, CD68 (HGF) and vimentin (c-Met) channel, pan-cytokeratin, HGF and c-Met, merged image. As expected, cytokeratin and vimentin/CD68 have a non-overlapping pattern of staining, while HGF and c-Met are expressed in both cancer and macrophage or stromal compartments. B: Chart summarizing the number of cancer cells in specimens used for proteomic analysis in the validation set. Data point and bar represent the average and the 5^th-95^th percentile range. Patients were grouped according to PFI and the number of cancer cells in each specimen quantified using the cytokeratin mask. There was no difference in the number of cancer cells in the refractory and resistant group. There was a significant decrease of cancer cells (p<0.05, Wilcoxon test) in the patients with sensitive disease.

Fig. 5

A-B: Diamond chart reporting the expression of HGF (A) and c-Met (B) in specimens collected before and after NACT (validation set). The top and bottom of each diamond represent the confidence interval for each group mean. The mean line across the middle of each diamond represents the group mean. Double asterisks indicate a strong statistical significance of the difference (p<0.001, Wilcoxon test). C-E Diamond chart reporting the expression of HGF in specimens (validation set) collected before (C-D) and after (E-F) NACT in cancer cells (C&E) and macrophage CD68+ (D&F). The top and bottom of each diamond represent the confidence interval for each group mean. The mean line across the middle of each diamond represents the group mean. Double asterisks indicate a strong statistical significance of the difference (p<0.001, Wilcoxon test). Single asterisk indicates significant difference at a p value <0.05.

Fig. 6

A-D Diamond chart depicting the expression of c-Met in specimens (validation set) collected before (A-C) and after (B-D) NACT in cancer cells (A-C) and stromal cells which are vimentin+ (B-D). The top and bottom of each diamond represent the confidence interval for each group mean. The mean line across the middle of each diamond represents the group mean. Double asterisks indicate a strong statistical significance of the difference (p<0.001, Wilcoxon test). E: diamond chart depicting the PFI (y-axis) across four groups (double positive, double negative for HGF/c-Met, single positive for HGF or c-Met). Double asterisks indicate a strong statistical significance of the difference (p<0.001, Wilcoxon test, double positive vs. double negative).