WLS-Wnt signaling promotes neuroendocrine prostate cancer

Abstract

Neuroendocrine prostate cancer (NEPC) is a lethal prostate cancer subtype arising as a consequence of more potent androgen receptor (AR) targeting in castration-resistant prostate cancer (CRPC). Its molecular pathogenesis remains elusive. Here, we report that the Wnt secretion mediator Wntless (WLS) is a major driver of NEPC and aggressive tumor growth in vitro and in vivo. Mechanistic studies showed that WLS is a transcriptional target suppressed by AR that activates the ROR2/PKCδ/ERK signaling pathway to support the neuroendocrine (NE) traits and proliferative capacity of NEPC cells. Analysis of clinical samples and datasets revealed that WLS was highly expressed in CRPC and NEPC tumors. Finally, treatment with the Wnt secretion inhibitor LGK974 restricted NE prostate tumor xenograft growth in mice. These findings collectively characterize the contribution of WLS to NEPC pathogenesis and suggest that WLS is a potential therapeutic target in NEPC.

Keywords: Biological Sciences; Cancer; Cell Biology; Transcriptomics.

Graphical abstract

Figure 1

Wnt signaling and WLS expression are upregulated in ENZ^R NEPC cells (A) GSEA plots of enrichment of gene signatures related to androgen response and neuronal system in ENZ^R C4-2B (C4-2B^ENZR) cells compared to ENZ-sensitive control C4-2B cells. (B) Western blot analysis of PSA, CHGA, and NSE protein expression in control and ENZ^R C4-2B cells. (C) Representative images of control and ENZ^R C4-2B cell morphology. Scale bars: 50 μm. (D) Quantification of per-cell number of neurites and average neurite length in control and ENZ^R C4-2B cells (n = 50 cells per cell line), which is shown as a representative of 3 independent experiments. Data represent the mean ± SEM. ∗∗p < 0.01, unpaired t test. (E) KEGG pathway analysis of the top 57 pathways upregulated in ENZ^R C4-2B cells compared to controls. KEGG pathways were clustered based on functional relation, indicated by different colors (red, signal transduction; green, human diseases; light blue, cellular processes; dark blue, metabolism; purple, organismal systems). The Wnt signaling pathway is pointed out by an arrow. (F) RT-qPCR analysis of Wnt gene expression categorized based on the ability to activate the canonical or noncanonical Wnt signaling pathways in control and ENZ^R C4-2B cells (n = 3). Data represent the mean ± SEM. ∗p < 0.05, ∗∗p < 0.01, unpaired t test. (G) GSEA analysis of select positively enriched Wnt-related gene sets in C4-2B^ENZR cells compared to controls. (H) RT-qPCR analysis of WLS mRNA expression in control and ENZ^R C4-2B cells (n = 3). Data represent the mean ± SEM. ∗∗p < 0.01, unpaired t test. (I) Western blot analysis of WLS protein expression in cell lysates and secreted Wnt5A and Wnt11 protein levels in conditioned media of control and ENZ^R C4-2B cells. See also Figures S1 and S2.

Figure 2

WLS is highly expressed in human CRPC and NEPC (A) IHC images of WLS protein expression in representative CRPC versus hormone-sensitive counterpart (HSPC) patient samples from the cohorts described in (B). Scale bars: 20 μm. (B) Quantification of WLS protein expression between HSPC (n = 39) and CRPC (n = 21) cohorts by average cell-based IHC staining intensity counts analyzed by inForm software. Data represent the mean ± SEM. ∗∗p < 0.01, unpaired t test. (C) IHC images of WLS and CHGA protein staining in two representative CRPC patient samples. Scale bars: 20 μm. (D) Protein co-expression correlation between WLS and CHGA in the CRPC cohort (n = 21). Average cell-based staining intensity counts for each individual protein were analyzed by inForm software to access the co-expression correlation by Pearson correlation. (E) Quantification of WLS mRNA expression between CRPC (n = 39) and NEPC (n = 12) cohorts in the Beltran dataset. Data represent the mean ± SEM. ∗p < 0.05, unpaired t test. (F) mRNA co-expression correlation of WLS with NE marker genes, AR and REST represented by positive (blue) or inverse (red) correlations in the Beltran dataset. ∗p < 0.05, ∗∗p < 0.01, Pearson correlation. (G) Relative WLS and CHGA mRNA levels normalized to ACTB in human PC cell lines toward an increasing degree of NE differentiation from CellExpress. See also Figure S3.

Figure 3

WLS is directly transcriptionally repressed by AR (A) Western blot analysis of WLS and PSA protein expression in LNCaP and C4-2B cells grown in normal growth media (FBS, 5 days), CSS media (5 days), or CSS media (2 days) followed by R1881 treatment (1 nM, 72 hr). (B) Western blot analysis of WLS protein expression in LNCaP cells grown in normal media treated with 20 μM ENZ for different times. (C) Western blot analysis of WLS protein expression in LNCaP and C4-2B cells transfected with a scrambled siRNA or siRNA against AR (siAR) for 48 hr prior to treatment with either CSS media or CSS media supplemented with 1 nM R1881 for additional 72 hr. (D) RT-qPCR analysis of WLS mRNA levels in LNCaP cells grown in CSS media for 48 hr followed by treatment with or without 1 nM R1881, with ethanol as a vehicle (Veh), for additional 72 hr (n = 3). Data represent the mean ± SEM. ∗p < 0.05, unpaired t test. (E) Relative WLS mRNA expression in LNCaP cells exposed to long-term culture in CSS media measured by RNA-seq from dataset GSE8702. Data represent the mean ± SEM. ∗∗p < 0.01, one-way ANOVA. (F) Genomic browser representation of AR binding in the distal enhancer region of WLS gene from VCaP cells and PC patient samples extracted from ChIP-seq datasets GSE55062 and GSE70079, respectively. Each track depicts ChIP-seq AR binding intensity for a given sample. The sequences of androgen response element (ARE) at each locus are shown with nucleotides homologous to the canonical ARE underlined. (G) ChIP-qPCR assays of AR occupancy of two ARE-centric enhancers of WLS gene (n = 3). Data represent the mean ± SEM. ∗p < 0.05, unpaired t test. See also Figure S4.

Figure 4

WLS is required for NEPC cell growth and NE marker expression (A) Western blot analysis of WLS, NSE and CD56 protein expression and Wnt5A secretion in lysates or conditioned media of control (shCon) and WLS-KD (shWLS) C4-2B^ENZR cells. (B) Western blot analysis of NSE and CD56 protein expression and Wnt5A secretion in C4-2B^ENZR cells treated with LGK974 at different concentrations for 7 days. (C) Western blot analysis of CD56 protein expression in control, WLS-KD and WLS-KD/recombinant Wnt5A protein-added (100 ng/mL, 3 days) C4-2B^ENZR cells. (D) Proliferation curves of control and WLS-KD C4-2B^ENZR and 22Rv1 cells by crystal violet staining (n = 4). Data represent the mean ± SEM. ∗∗p < 0.01, one-way ANOVA. (E) Colony formation assays of control and WLS-KD C4-2B^ENZR and 22Rv1 cells (n = 3) with the number of colonies in respective control groups set as 100%. Representative images from each group are shown. Data represent the mean ± SEM. ∗∗p < 0.01, one-way ANOVA. (F) Proliferation curves of C4-2B^ENZR and 22Rv1 cells treated with LGK974 at different concentrations by crystal violet staining (n = 4). Data represent the mean ± SEM. ∗∗p < 0.01, one-way ANOVA. See also Figure S5.

Figure 5

WLS mediates NEPC cell growth through the PKCδ/ERK pathway (A) Phospho antibody array analysis of phosphoproteins implicated in Wnt signaling in WLS-KD C4-2B^ENZR cells. All phosphoproteins with significantly (p < 0.05) decreased levels in WLS-KD cells compared to controls are presented. All phospho signals were normalized to their total forms from a single array with 6 replicate spots (n = 6). Data represent the mean ± SEM. ∗p < 0.05, ∗∗p < 0.01, unpaired t test. (B) Western blot analysis of p-PKCδ, PKCδ, p-ERK and ERK protein expression in control and WLS-KD C4-2B^ENZR cells, with quantification of normalized p-PKCδ and p-ERK levels presented separately (n = 3). Data represent the mean ± SEM. ∗∗p < 0.01, unpaired t test. (C) Western blot analysis of p-ERK and ERK protein expression in control and WLS-KD C4-2B^ENZR cells treated with BJE-106 (BJE, 0.1 μM, 3 hr). (D and E) Proliferation curves of C4-2B^ENZR and 22Rv1 cells treated with BJE-106 (D) or UO126 (UO, E) at different concentrations by crystal violet staining (n = 4). Data represent the mean ± SEM. ∗∗p < 0.01, one-way ANOVA. (F) Colony formation assays of C4-2B^ENZR and 22Rv1 cells treated with BJE-106 or UO126 at different concentrations (n = 3). Representative images from each group are shown. Data represent the mean ± SEM. ∗∗p < 0.01, one-way ANOVA. (G) Proliferation curves of C4-2B^ENZR cells treated with PKCδ or ERK1/2 siRNAs by MTS assays (n = 4). Data represent the mean ± SEM. ∗p < 0.05, unpaired t test. (H) Western blot analysis of CD56 protein expression in C4-2B^ENZR cells treated by BJE-106 (0.1 μM, 7 days) or UO126 (6.25 μM, 7 days). (I) Western blot analysis of CD56 protein expression in C4-2B^ENZR cells treated by PKCδ or ERK1/2 siRNAs. See also Figures S6 and S7.

Figure 6

ROR2 mediates WLS induction of the PKCδ/ERK pathway (A) Western blot analysis of ROR2 protein expression in control and WLS-KD C4-2B^ENZR cells. (B) Co-expression correlation between WLS and ROR2 mRNA in the Beltran dataset by Pearson correlation. (C) Proliferation curves of C4-2B^ENZR cells treated with a scrambled siRNA (siCon) or ROR2 siRNAs (siROR2) by crystal violet staining (n = 4). Data represent the mean ± SEM. ∗∗p < 0.01, one-way ANOVA. (D) Colony formation assays of C4-2B^ENZR cells treated with a scrambled siRNA or ROR2 siRNAs (n = 3). Data represent the mean ± SEM. ∗∗p < 0.01, one-way ANOVA. (E) Western blot analysis of ROR2, CD56, p-PKCδ, PKC, p-ERK, and ERK protein expression in control and siRNA-mediated ROR2-KD C4-2B^ENZR cells. (F) Proliferation curves of control, WLS-KD and WLS-KD/ROR2-overexpressing (OE) C4-2B^ENZR cells by crystal violet staining (n = 4). Data represent the mean ± SEM. ∗∗p < 0.01, one-way ANOVA. (G) Western blot analysis of CD56 protein expression in control, WLS-KD and WLS-KD/ROR2-OE C4-2B^ENZR cells. See also Figure S8.

Figure 7

WLS silencing inhibits NEPC tumor growth in mice (A) Growth curves of 22Rv1 tumors (shCon, n = 14; shWLS#1, n = 14; shWLS#2, n = 12) implanted subcutaneously in castrated SCID mice. Data represent the mean ± SEM. ∗p < 0.05, one-way ANOVA. (B) Tumor weights measured for each group in (A) at the experimental endpoint. Data represent the mean ± SEM. ∗p < 0.05, one-way ANOVA. (C) IHC representative images of WLS, Ki-67, cleaved caspase 3 (cCas3), CD34, CD56, p-PKCδ and p-ERK in tumor samples from each group in (A). Scale bars: 20 μm. (D) Quantitation of WLS, CD34, CD56, p-PKCδ and p-ERK IHC per-cell staining intensity and % of Ki-67⁺ or fold change of cCas3⁺ cells in tumor samples from each group in (A) (n = 15). Data represent the mean ± SEM. ∗∗p < 0.01, one-way ANOVA. (E) Growth curves of 22Rv1 subcutaneous tumors receiving corn oil (Veh, n = 7) or LGK974 (5 mg/kg, orally, n = 8) daily in castrated nude mice. Data represent the mean ± SEM. ∗∗p < 0.01, unpaired t test. (F) Tumor weights measured for each group in (E) at respective experimental endpoints (Veh, day 14; LGK974, day 21). Data represent the mean ± SEM. ∗∗p < 0.01, unpaired t test. (G) Anatomic images of tumors from each group in (E) at respective experimental end points. (H) IHC representative images of WLS, Ki-67, cCas3, CD34, CD56, p-PKCδ and p-ERK staining in tumor samples from each group in (E). Scale bars: 20 μm. (I) Quantitation of % of Ki-67⁺ or fold change of cCas3⁺ cells and CD34, CD56, p-PKCδ and p-ERK IHC per-cell staining intensity from each group in (E) (n = 15). Data represent the mean ± SEM. ∗p < 0.05, ∗∗p < 0.01, unpaired t test.