Fig 2: NECTIN4 regulation of A431 cell proliferation via ERK signaling. (A) The phosphorylation status of ERK and Akt was assessed in control and NECTIN4 siRNA-transfected cells. Representative blots are shown, in addition to the mean ± SD of pAkt/Akt and pERK/ERK ratios (n = 3). ** p < 0.01. (B) Inhibition of ERK phosphorylation by U0126 was confirmed by western blotting. Representative blots are shown, as well as the mean ± SD of pERK/ERK ratio in the DMSO (vehicle control)- and U0126-treated groups (n = 3). *** p < 0.001. (C) Viability of DMSO (0.1%)- and U0126 (10 µM)-treated cells on days 1–5 after treatment, as assessed using the CCK-8 cell viability assay. Experiments were independently repeated three times and performed in five wells per condition. *** p < 0.001. (D,E) Mean (±SD) relative cyclin D1 expression at the mRNA (D) and protein (E) levels in DMSO- and U0126-treated cells (n = 3). *** p < 0.001. (F) Representative immunocytochemical images of Ki67 staining (left) and the mean ± SD percentage of Ki67-positive cells (right) in DMSO- and U0126-treated cells. DAPI was used to detect cell nuclei. Three different areas per well were randomly selected and three independent wells were observed. *** p < 0.001. Scale bars = 200 µm.

Fig 3: A zoomable version of this network is provided as pdf in the Supplement as Figure S39. Network of significant Nectin 4 correlated gene-expression co-association clusters (cf. Table 5) correlated to other omics (proteomics), FACS (immune cell populations) and Luminex (cyto/chemokines) data (cf. Table 4). The gene-expression co-association clusters are represented by their first principal components (of all gene expression values), PC1s. The distribution of the percentages of explained variations (PEVs) of these PC1s is shown at the right border. Below a network of the correlations of the PC1s of the gene-expression co-association clusters is shown as estimated by graphical Gaussian modelling (GGM). The main network shows all significant (FDR < 10%) correlations of the PC1s of the clusters with all other data. Edge colors represent the direction of the correlation (red, positive; blue, negative correlation), the color of the nodes indicates the log2 fold-changes (FCs) of the analytes/cell-populations with the percentages of Nectin 4 positive cells (Nectin 4 score), and the size of the nodes indicates if these FCs are significant (large, yes; small, no). The type and source of the analytes/cell-populations is coded in the node-label and node border color, respectively. Analytes and cell-populations are given as node-labels (for complete lists of analyzed factors cf. Table 4 and Table S5). A comprehensive legend is shown on the left border of the figure.

Fig 4: Validation of the impact of Nectin 4 protein expression on overall survival and correlation to Nectin 4 cluster (c1_143) expression. (A) Non-linear correlation of percentage Nectin 4 positive cells (x-axis) with relative hazard for death (y-axis) corrected for age, FIGO stage, and residual tumor mass, as estimated by fractional polynomials Cox regression. (B) Correlations of gene expressions of the c1_143 cluster with NECTIN4 as hub-gene with Nectin 4 protein abundances (Nectin 4 score) in corresponding tumor tissues (colors represent log2 fold-changes, FCs). (C) Barcode enrichment plot showing the ranked statistics of log2 FCs from subfigure B. (D) Kaplan-Meier estimate for overall survival dichotomizing Nectin 4 percentages following the cutoff determined in subfigure A (>50% versus =50%). (E) Survival curves of the multiple COX regression model (cf. Table 3), dichotomized as in subfigure D. As these survival curves represent a multiple Cox model, no censored patients are indicated.

Fig 5: Effects of NECTIN4 inhibition on the migration of A431 cells. (A,B) Cells were transfected with control or NECTIN4 siRNA and the expression of ZEB1, ZEB2, and SNAIL was assessed at the mRNA level after 48 h (A) and at the protein level after 72 h (B). Data are the mean ± SD. (n = 3). ** p < 0.01, *** p < 0.001. (C) Monolayers of control or NECTIN4 siRNA-transfected A431 cells were scratched, and cell migration was monitored up to 10 h at 2-h intervals. Representative images of the wound area at 0 and 10 h are shown for control and NECTIN4 siRNA-transfected A431 cells. The graph shows the mean ± SD of relative wound area. The migration assay was independently repeated three times and performed in 18 wells per condition. * p < 0.05, ** p < 0.01, *** p < 0.001 compared with control siRNA-transfected cells.