KLK4 expression in the progression of prostate cancer and in foci of atypical suspicious prostate glands. (A) IHC staining of KLK4 in prostate tissues. (a) PIN lesion and adjacent normal glands (Normal, arrows). (b) Low‐grade PIN (LGPIN) with positive staining of basal cells (closed arrows) and no staining in stromal cells (open arrow head). (c) High‐grade PIN (HGPIN) lesions with positive staining in both the nucleus (open arrows) and cytoplasm (arrows) of the secretory cells. (d–f) Strong intensity of KLK4 staining in prostate cancer lesions (Ca or arrows) compared to weak staining of an adjacent normal glands and stroma (Normal and open arrow heads). (g) Anti‐KLK4 antibody absorbance showing no staining, as a negative control. Scale bars are as indicated. (B) Comparison of KLK4 staining intensity in prostate tissues with different Gleason grades. Average staining intensity (●) and SD are shown. Number of region analysed per lesion type is indicated after each lesion name. One‐way ANOVA test, *P < 0.05 compared to normal and ^# P < 0.05 compared to benign prostatic hyperplasia (BPH). Details can be found in Tables and . ¹Note that the ‘normal’ prostate sample comprises scores for the single normal prostate tissue specimen and 15 tumour‐adjacent normal prostate tissue regions. (C) (a–d) Low and high magnification of cells in adjacent normal gland showing strong KLK4 staining (box), with the high KLK4 foci appearing to have no basal cells (arrow). (e–g) Low and high magnifications of double staining of high molecular weight cytokeratin 34βE12 and KLK4 in normal prostate gland, showing the KLK4 expressing foci (brown) with disappearing basal cells (arrow) compared to the basal cell layer expressing 34βE12 (pink) in the surrounding area. Scale bars are as indicated. PIN, prostatic intraneoplasia; Ca, cancer.

KLK4 can activate PAR1 in prostate‐derived stromal cells. (A) Activation of PARs in WPMY1 cells was analysed by calcium flux assay. Cells were treated with AP1, AP2 (100 μm), KLK4 and mKLK4 (300 nm). Results are expressed as ratio of fluorescence emission between excitation at 340 and 380 nm (ratio 340/380, y‐axis) over time (x‐axis). (B) Calcium flux assay as in (A). Fluorescence emission was monitored for 600 s and two successive stimulations were made: first stimulation at 30 s with AP1 or AP2 to desensitize PAR1 or PAR2 and second stimulation at 480 s with AP1, AP2 or KLK4. (C) Calcium flux as in A using WPMY1 cells control‐ or PAR1‐siRNA (left and right panel, respectively).

KLK4 regulates gene expression through PAR1 in prostate‐derived stromal cells. (A) Gene expression was studied by RTqPCR in WPMY1 cells treated for 6, 12 or 18 h with mKLK4, KLK4 (20 nm), AP1 or AP2 (100 μm). PBS treatment was used as reference for each time point. Results are presented as mean ± SD of three biological replicates. (B) Gene expression was investigated by RTqPCR in WPMY1 cells transfected with PAR1‐siRNA or control‐siRNA treated with KLK4 or mKLK4 (20 nm) for 18 h. Expression in WPMY1 cells control‐siRNA treated with mKLK4 was used as reference. Results are presented as mean of relative mRNA expression ± SD of three biological replicates. (C) Gene expression was determined by RTqPCR in WPMY1 cells treated for 6 h with mKLK4, KLK4 (20 nm), AP1 or AP2 (100 μm) in the presence of 0.3 and 0.7 μm of PAR1 inhibitor (SHC79797) or vehicle control (DMSO). Gene expression after mKLK4 treatment was used as reference for each concentration of inhibitor. Results are presented as mean ± SD of three biological replicates. *P < 0.05, **P < 0.01, ***P < 0.001 compared to reference.

KLK4 regulates protein expression through PAR1 in prostate‐derived stromal cells. (A–B) FGF1 protein in 30 μg of total cellular proteins was determined using FGF1‐ELISA in (A) WPMY1 cells treated for 6, 12 or 18 h with mKLK4, KLK4 (20 nm) or AP1 (100 μm), or (B) WPMY1 cells transfected with PAR1‐siRNA or control‐siRNA and treated with KLK4 or mKLK4 (20 nm) for 24 h. Results are expressed as mean ± SD from three biological replicates. (C) TAGLN expression was determined by immunofluorescent detection in WPMY1 cells treated for 48 h with mKLK4, KLK4 (20 nm) or AP1 (100 μm). Nuclei were stained using DAPI. Representative images are shown, scale bar: 20 μm. (D) TAGLN expression was determined by western blot in WPMY1 cells transfected with PAR1‐siRNA or control‐siRNA and treated for 24 and 48 h with mKLK4 or KLK4 (20 nm). Densitometry analysis was performed using imagej software on three independent experiments. (E–G) WPMY1 cells transfected with FGF1‐siRNA or control‐siRNA were treated for 24 h with mKLK4 or KLK4 (20 nm). (E). Gene expression was obtained by RTqPCR with expression observed for WPMY1 cells control‐siRNA treated with mKLK4 as reference. Results are presented as mean ± SD of three biological replicates. (F) The amount of FGF1 protein in 30 μg of total cellular proteins was determined using FGF1‐ELISA. Results are expressed as mean ± SD calculated on three biological replicates. (G) TAGLN protein expression was determined by western blot as in D. Densitometry analysis was performed using imagej software on three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001 compared to reference.

KLK4 regulates gene expression in prostate‐derived NPF/CAF. (A) NPFs isolated from two patients were treated for 24 h with KLK4 and mKLK4 (20 nm) and gene expression analysed (RTqPCR). Gene expression observed for NPF cells treated with mKLK4 was used as reference for each patient. Results are presented as mean ± SD of two biological replicates. (B) The amount of FGF1 protein in 30 μg of total cellular protein in NPF/CAFs from 2 different patients was determined using FGF1‐ELISA. Results are expressed as mean ± SD calculated on three biological replicates. (C) TAGLN protein expression was determined by western blot (as in Fig. D). Mann–Whitney test with *P < 0.05, **P < 0.01, ***P < 0.001. (D) Gene expression was analysed in matched NPF/CAF isolated from five patients. Gene expression observed for NPF cells was used as reference for each patient. Results are presented as mean ± SD of three technical replicates. Statistical analysis was performed using one‐way ANOVA test and Bonferroni's multiple comparison, *P < 0.05, **P < 0.01, ***P < 0.001.

KLK4 induces CAF‐related features in prostate stromal cells through PAR1. (A) Gene expression of αSMA, ESR1 and SFRP1 was determined by RTqPCR in WPMY1 cells treated with mKLK4, KLK4 or AP1 for 24 h. Results are expressed as mean ± SD calculated on three biological replicates, *P < 0.05, **P < 0.01, ***P < 0.001. (B–C) WPMY1 cells were treated every 48 h with mKLK4, KLK4 (20 nm) or AP1 (100 μm) for 6 days. (B) αSMA and TAGLN expression was determined by western blot (left panel). Densitometry analysis was performed using imageJ on three independent experiments. αSMA expression was also determined by immunofluorescent staining (right panel) and the fluorescence quantified (Fig. B). (C). Proliferation was measured by direct cell counting using InCell analyzer and cellprofiler software based on nuclei staining (DAPI). Results are presented as mean ± SD of three biological replicates, *P < 0.05, **P < 0.01, ***P < 0.001. (D) Conditioned media (CM) of WPMY1 cells treated for 48 h with KLK4 and mKLK4 (20 nm) were analysed using a protein array (Human XL Cytokine array). For each factor analysed, results are expressed as mean ± SD of relative intensity of duplicate spots compared to mean intensity of six positive control spots present on each array (%). Full analysis of the cytokine array can be found in Table . (E) IL8 and VEGF concentrations were determined by ELISA in conditioned media (CM) from wild‐type WPMY1 cells or WPMY1 cells transfected with PAR1‐siRNA or control‐siRNA treated for 24 h with mKLK4 or KLK4 (20 nm). Results are expressed as mean ± SD calculated on three biological replicates. Statistical analysis was performed using t‐test, Mann–Whitney with *P < 0.05, **P < 0.01, ***P < 0.001 compared to reference. (F) Effect of CM prepared in E on proliferation of HUVEC cells was analysed. Left panel: HUVEC growth in the presence of WPMY1‐derived CM was followed by live cell imaging (Incucyte) for 48 h. Relative confluence was calculated using confluence at 24 h of HUVEC cells treated with mKLK4‐treated WPMY1's CM as reference. Right panel: HUVEC growth in the presence of WPMY1‐derived CM or endothelial cells growth medium (EGM) was analysed by DNA assay after 48 h of treatment in the presence of an IgG isotype control or a VEGF‐neutralizing IgG. Relative fluorescence intensity was calculated using mKLK4‐treated WPMY1's CM containing IgG isotype control as reference. Results are presented as mean ± SD of three biological replicates, *P < 0.05, **P < 0.01, ***P < 0.001 compared to reference.

Schematic for the possible involvement of KLK4 in early stages of prostate cancer. KLK4 is produced by premalignant cells in benign prostatic hyperplasia and PIN lesions and acts on prostate fibroblast stromal cells through activation of PAR1 as well as other undetermined pathways conducting to gene regulation of IL8, FGF1, LOX, αSMA, TAGLN and VEGFA (PAR1 dependent) and FGF5, ESR1 and SFRP1 (PAR1 independent). In response to KLK4 stimulation, prostate fibroblast stromal cells present a higher proliferation rate and a modification of their secretome (increase in Dkk‐1, GDF15, HGF/SF and decrease in IGFBP3, MCP1 and PDGF‐AA) in favour of a proangiogenic response (increase in IL8 and VEGF), which could ultimately lead to the development of a proangiogenic microenvironment necessary for prostate cancer progression. We could also hypothesize that modification of prostate fibroblast secretome after KLK4 stimulation could directly influence proliferation/migration/survival of prostate cancer cells.