Fig. 5

FRS2 duplications predict poor prognosis and promoter tumor angiogenesis in UBC. a qPCR analysis in an additional 196 UBCs confirms the high-level duplication of FRS2 in UBCs. b Immunohistochemical staining of 50 UBC tumors (25 tumors with FRS2 high copy and others with FRS2 low copy) displaying consequences of copy-number gain in FRS2. FRS2 gene duplication leads to FRS2 protein overexpression in tumor tissues. c Immunohistochemical analysis of microvessel density in two different clusters of UBC tumors: one cluster with FRS2 gene copy > 3 (n = 14) and the other cluster with FRS2 gene copy ≤ 3 (n = 174). The microvessel densities were determined as described in methods. The three horizontal lines represent mean ± SD for the subjects. ***P < 0.001 vs. low-copy group, two-sided unpaired t-test with Welch’s correction (b) or Mann–Whitney test (c). d Kaplan–Meier survival curves for two clusters of patients (as mentioned in part c). The former cluster patients have poorer prognosis as compared with patients of the latter cluster in both nonmuscle-invasive bladder cancer (NMIBC) and muscle-invasive bladder cancer (MIBC) subcohorts. Statistical significance was determined by log-rank test. e, f Knockdown of FRS2 in 5637 and SW780 cells attenuated their abilities to recruit human umbilical vein endothelial cells (e) and induce tube formation (f). Scale bars, 100 µm. Error bars represent the SEM. The data shown represent averages from three independent experiments and were statistically analyzed by two-sided t test. *P < 0.05, **P < 0.01, ***P < 0.001 vs. siCon