Roles of KHDRBS2 and KHDRBS3 in prostate cancer progression and AR-V7 regulation Authors: XuDong Yang, Xia Liu, Ran Xiao, YiLin Cao Published: 2025 ABSTRACT AbstractObjectiveTo investigate the biological functions of KHDRBS2 and KHDRBS3 in prostate cancer (PCa) progression and their potential roles in regulating androgen receptor splice variant 7 (AR-V7) expression, a key factor in castration-resistant prostate cancer (CRPC).MethodsWe performed a comprehensive analysis of publicly available datasets to examine the expression patterns of KHDRBS family members in PCa, including CRPC and neuroendocrine subtypes. In vitro experiments were conducted using AR-positive 22RV1 and AR-negative PC3 cell lines to assess the expression and regulatory interactions of KHDRBS2 and KHDRBS3. Functional assays evaluated their effects on cell proliferation and tumor growth in vivo. Additionally, KHDRBS2 protein levels were manipulated in 22RV1 cells to assess their impact on AR-V7 and full-length AR expression.ResultsOur dataset analysis revealed distinct expression patterns of KHDRBS2 and KHDRBS3, with higher alteration frequencies in CRPC and neuroendocrine PCa. In vitro, KHDRBS2 and KHDRBS3 exhibited mutually exclusive expression, with KHDRBS2 predominantly found in AR-positive 22RV1 cells and KHDRBS3 in AR-negative PC3 cells. Reciprocal regulation between the two proteins was observed in both cell lines. Functional studies showed that both KHDRBS2 and KHDRBS3 promoted cell proliferation and tumor growth. Notably, silencing KHDRBS2 in 22RV1 cells led to a selective reduction in AR-V7 expression, without affecting full-length AR levels.ConclusionsThese findings uncover novel roles for KHDRBS2 and KHDRBS3 in PCa progression, with KHDRBS2 identified as a potential key regulator of AR-V7 expression. Our results provide new insights into AR splice variant regulation and highlight potential therapeutic targets for PCa treatment.Supplementary InformationThe online version contains supplementary material available at 10.1007/s12672-025-04294-1.Keywords: Androgen receptor (AR), Alternative splicing, AR-V7, Castration-resistant prostate cancer, KHDRBS family, Prostate cancer Introduction Prostate cancer, represents one of the most complex solid malignancies, with its progression intricately regulated by androgen and its nuclear receptor, androgen receptor [1, 2], alongside other signaling pathways [3, 4]. Although endocrine therapy targeting androgen production and AR function initially prolongs patient survival, the majority of cases inevitably progress to CRPC within 14–20 months [5]. This progression is primarily attributed to the reactivation of AR signaling through various mechanisms, including AR mutations, gene amplification, protein over-expression, intratumoral androgen synthesis, aberrant co-regulator expression, and alternative activation by cytokines and growth factors [6, 7]. Recent studies have proposed the critical role of C-terminal truncated AR variants (AR-Vs) in CRPC progression [8–10]. These variants, produced by alternative splicing (AS) and regulated by RNA-binding proteins (RBPs), lack the ligand-binding domain (LBD) and function as constitutively active transcription factors [8]. Among the twenty identified AR-Vs, AR-V7 and ARV-567es are particularly implicated in tumor initiation [11], progression [12, 13], and therapeutic resistance [14, 15]. Clinical studies have demonstrated significantly shorter progression-free and overall survival in AR-V7-positive patients treated with next-generation AR-targeted therapies [16–18], highlighting AR-Vs as both prognostic biomarkers and potential therapeutic targets. The regulation of AR splicing by RBPs, particularly those of the signal transduction and activation of RNA (STAR) protein family, has emerged as a key area of investigation. KHDRBS1, a well-studied member of this family, promotes AR-V7 expression [19, 20] and drives PCa cell proliferation [21]. However, the roles of two other STAR family members, KHDRBS2 and KHDRBS3, in PCa remain largely unexplored. Similar to KHDRBS1, both KHDRBS2 and KHDRBS3 possess a highly conserved KH domain that is essential for their alternative splicing function, and their interactions have been confirmed through co-immunoprecipitation studies [22, 23]. The functional significance of these structural similarities is evidenced by their cooperative regulation of neurexin AS in the mouse central nervous system, demonstrating their synergistic roles in neuronal development and synaptic plasticity [24–26], suggesting potential collaborative functions in other contexts. Notably, genetic alterations of KHDRBS2 and KHDRBS3 have been observed in various malignancies. High-throughput sequencing and microarray analyses have identified mutations, amplifications, deletions, and gene fusions of these genes in oral leukoplakia [27], breast cancer [28], prostate cancer [29, 30], renal cell carcinoma [31], and lung adenocarcinoma [32]. These genomic alterations suggest potential involvement of KHDRBS2 and KHDRBS3 in disease processes. Given the structural similarities and interactions among these STAR family proteins, KHDRBS2 and KHDRBS3 probably play significant roles in regulating AR splicing and PCa progression, similar to KHDRBS1. To address this question, we investigated the expression patterns and functions of KHDRBS2 and KHDRBS3 in PCa cell lines. Our study reveals a novel, predominantly mutually exclusive expression pattern and reciprocal modulation between KHDRBS2 and KHDRBS3 in 22RV1 and PC3 cells. Furthermore, we demonstrate their positive influence on cell proliferation both in vitro and in vivo. Notably, KHDRBS2 knockdown in 22RV1 cells results in AR-V7 down-regulation, suggesting a potential regulatory role in AR splicing. These findings not only expand our knowledge of the complex regulatory network governing AR splicing in PCa but also identify KHDRBS2 and KHDRBS3 as potential new players in disease progression, which provide new ideas for therapeutic intervention in PCa. Materials and methods Cell culture All experiments were performed with mycoplasma-free cells that were authenticated by STR profiling. Cell line sources are detailed in the Supplementary Materials-1. All media were supplemented with 10% heat-inactivated fetal bovine serum (FBS, Corning, USA). Cells were maintained at 37 °C in a humidified atmosphere with 5% CO2 and sub-cultured at 80% confluence. RNA extraction and quantitative real-time PCR (RT-qPCR) Total RNA was extracted using TRIzol reagent (ThermoFisher, USA) following the manufacturer’s protocol. RNA concentration was determined spectrophotometrically. cDNA was synthesized using Oligo d(T)23 VN and ProtoScript II (NEB, USA). qRT-PCR was performed on a LightCycler480 system (Roche, Switzerland) using the KAPA SYBR FAST qPCR Kit (KAPA Biosystems, USA). The amplification program consisted of: 95 °C for 2 min (pre-denaturation), followed by 40 cycles of 95 °C for 15 s (denaturation) and 60 °C for 30 s (annealing and extension). Melting curve analysis was performed at 95 °C for 1 min and 65 °C for 1 min. GAPDH served as the internal control, and relative mRNA expression was calculated using the 2 − ΔΔCT method. All assays were performed in technical triplicates with three independent cDNA syntheses. Primer information is listed in supplementary material-1. Protein extraction and Western blotting (WB) Cells were lysed in RIPA buffer containing 1 mM PMSF. Protein concentrations were determined using the BCA Protein Assay Kit (Beyotime, China). Equal amounts of protein were separated by 10% SDS-PAGE and transferred to PVDF membranes (Millipore, USA). Membranes were blocked with 5% nonfat milk for 1 h at room temperature, then incubated overnight at 4 °C with primary antibodies. After washing with TBS-T, membranes were incubated with HRP-conjugated secondary antibodies for 1 h at room temperature. Immunoreactive proteins were visualized using ECL substrate (ThermoFisher, USA) and quantified using UVP Biospectrum, with β-actin as the loading control.Primary antibodies are listed in supplementary material-1. Plasmid construction and lentiviral transduction Over-expression plasmids and shRNAs for KHDRBS2 and KHDRBS3 were obtained from (GeneCopoeia, Inc. Guangzhou, China). Lentiviral particles were produced in HEK293T cells using psPAX2 and pMD2.G packaging plasmids. Transfections were performed using JetPRIME reagent (Polyplus Transfection, France) according to the manufacturer’s instructions. For transient transfections, 2 µg plasmid DNA was used per well of a 6-well plate. For lentiviral production, 10 µg total plasmid DNA (5 µg shRNA, 3.75 µg psPAX2, 1.25 µg pMD2.G) was used per 100 mm dish. Lentiviral transductions were performed using 1 mL virus-containing medium with 8 µg/mL polybrene. The plasmids and shRNA used are listed in supplementary material-1. Cell proliferation assay Cell proliferation was assessed using the CCK-8 assay (Dojindo Laboratories, Japan). Cells were seeded at 1,500 cells per well in 96-well plates and allowed to adhere for 12 h. Absorbance at 450 nm was measured daily for 7 consecutive days following the manufacturer’s protocol. Xenograft tumor model Six-week-old male BALB/c nude mice were injected subcutaneously with 5 × 106 cells in each flank. Tumor growth was monitored for 30 days, after which mice were sacrificed, and tumors were excised, weighed, and processed for further analysis. cBioPortal data analysis Genomic data from 11 PCa datasets were analyzed using the cBioPortal online platform (http://cbioportal.org) [33]. Expression and alteration frequencies of target genes were extracted and analyzed across the selected datasets, which are detailed in supplementary material-1. Androgen stimulation assay 22RV1 cells were cultured in phenol red-free RPMI 1640 (Gibco) supplemented with 5% charcoal-stripped FBS (Biological Industries, Israel) for 24 h. Cells were then treated with or without 1 nM R1881 for an additional 24 h before analysis. Cells cultured in standard medium served as controls. Further experimental details are shown in Figure S4 which is included in supplementary material-1. Statistical analysis Data were analyzed using GraphPad Prism software (version 9.0). Comparisons between groups were performed using two-tailed unpaired t-tests for two-group comparisons and one-way ANOVA followed by Tukey’s post-hoc test or two-way repeated measures ANOVA based on the general linear model (GLM) followed by Šídák’s for multiple group comparisons. Survival analyses were performed using the Kaplan-Meier method, and differences between groups were evaluated using the log-rank test. p values < 0.05 were considered statistically significant. Sample sizes (n) refer to biological replicates unless stated otherwise. Results Expression patterns and genetic alterations of KHDRBS family members in prostate cancer We conducted an analysis of the KHDRBS family genes using eleven prostate cancer sequencing datasets, encompassing 4,353 samples from 15 distinct studies through the cBioPortal platform. Analysis of mRNA expression levels in the TCGA prostate cancer cohort revealed distinct patterns for each family member. KHDRBS1 showed relatively consistent expression with a median log2 value of 11.56 (interquartile range [IQR]: 11.43–11.70; Fig. 1A). KHDRBS2 exhibited generally lower expression levels (median: − 1.52, IQR: − 2.5 to − 0.45; Fig. 1B), while KHDRBS3 displayed intermediate expression levels with greater variability (median: 7.71, IQR: 6.91–8.95; Fig. 1C). Expression patterns and alteration frequencies of KHDRBS1-3 across different PCa studies. (A-C) RNA sequencing expression data (log2) distribution for KHDRBS1 (A), KHDRBS2 (B), and KHDRBS3 (C) genes across cancer samples. Blue dots represent samples without mutations, and red dots represent samples with missense mutations. The grey boxes indicate the interquartile range, with the median shown as a horizontal line. (D-F) Alteration frequencies of KHDRBS1 (D), KHDRBS2 (E), and KHDRBS3 (F) across multiple PCa studies. The y-axis shows alteration frequency as a percentage, and the x-axis lists different PCa studies with their sample sizes (n) indicated. Different types of alterations are color-coded: mutations (green), deletions (blue), amplifications (red), and multiple alterations (grey). The presence (+) or absence (-) of mutation and CNA (Copy Number Alteration) data for each study is indicated below the bars. NEPC shows notably higher alteration frequencies, particularly for KHDRBS3, with amplification being the predominant type of alteration. Cancer study abbreviations include NEPC (Neuroendocrine Prostate Cancer), PRAD (Prostate Adenocarcinoma), TCGA (The Cancer Genome Atlas), and other institutional cohorts. Sample sizes are shown for each study (e.g., “n = 114”) Genetic alteration frequencies exhibited significant heterogeneity across datasets and family members. KHDRBS1 alterations were detected in 4 datasets (Fig. 1D), while KHDRBS2 and KHDRBS3 showed more widespread alterations, being present in 8 and 10 datasets, respectively (Fig. 1E, F). Notably, KHDRBS3 consistently demonstrated higher alteration frequencies compared to KHDRBS1 in datasets where both were altered. The neuroendocrine prostate cancer (NEPC) cohort exhibited the highest alteration frequencies for all three family members, with particularly striking rates for KHDRBS3 (43%), followed by KHDRBS2 (13%) and KHDRBS1 (10%). Detailed analysis of genetic alterations across four key datasets revealed distinct patterns among KHDRBS family members (Figure S1, Supplementary material-2). In the neuroendocrine prostate cancer (NEPC) cohort (Trento/Cornell/Broad 2016), we observed remarkably high alteration frequencies: KHDRBS1 (50%), KHDRBS2 (15%), and KHDRBS3 (64%) (Figure S1A). The Fred Hutchinson Cancer Research Center (FHCRC) cohort showed lower but still substantial alteration rates: KHDRBS1 (9%), KHDRBS2 (8%), and KHDRBS3 (34%) (Figure S1B). Both TCGA cohorts (Provisional and Cell 2015) demonstrated similar patterns, with alteration frequencies of approximately 5%, 7%, and 12–13% for KHDRBS1, KHDRBS2, and KHDRBS3, respectively (Figure S1C-D). Notably, the NEPC cohort consistently exhibited the highest alteration frequencies across all three family members (Figure S1A), suggesting a potential association between KHDRBS alterations and neuroendocrine differentiation. To assess the clinical implications of these alterations, we performed survival analyses using the TCGA dataset (Figure S2, Supplementary material-3). The Kaplan-Meier curves revealed that alterations in KHDRBS1 (Figure S2A-B) and KHDRBS2 (Figure S2C-D) showed no significant association with patient outcomes (log-rank test, p > 0.05 for both overall survival [OS] and disease-free survival [DFS]). In contrast, KHDRBS3 alterations emerged as a significant predictor of poor prognosis (Figure S2E-F). Specifically, patients with KHDRBS3 alterations demonstrated significantly shortened OS (median: 86.5 versus 151.2 months; hazard ratio [HR] = 1.76, 95% confidence interval [CI]: 1.12–2.78; log-rank test, p = 0.0007; Figure S2E). However, KHDRBS3 alterations did not significantly impact DFS (log-rank test, p = 0.134; Figure S2F). These findings suggest that KHDRBS3 alterations may serve as a potential prognostic marker for overall survival in prostate cancer patients. Expression patterns and reciprocal regulation of KHDRBS2 and KHDRBS3 in prostate cancer cell lines Following our analysis of KHDRBS family expression in TCGA datasets, we investigated both the mRNA levels and protein expression of KHDRBS family members across four prostate cancer cell lines (22RV1, LNCaP, DU145, and PC3). Quantitative PCR analysis revealed distinct expression patterns among family members at the transcriptional level (Fig. 2A-C). KHDRBS1 showed the highest overall mRNA expression across all cell lines, while KHDRBS2 exhibited comparatively lower expression levels. Notably, KHDRBS2 expression was significantly elevated in 22RV1 cells compared to other cell lines (p < 0.01, Fig. 2B). In contrast, KHDRBS3 demonstrated markedly higher expression in PC3 cells compared to 22RV1, LNCaP, and DU145 (p < 0.0001, Fig. 2C). Western blot analysis confirmed that protein expression patterns of KHDRBS family members were consistent with their mRNA levels across all cell lines (Fig. 2D), further validating the differential expression profiles observed at the transcriptional level. Expression of KHDRBS family genes in prostate cancer cell lines. (A–C) mRNA levels of KHDRBS1, KHDRBS2, and KHDRBS3 in 22RV1, LNCaP, DU145, and PC3 cell lines, as determined by RT-qPCR. (D) Western blot analysis of KHDRBS1,KHDRBS2 and KHDRBS3 protein levels in 22RV1, LnCaP, DU145, and PC3 cell lines. (E–F) Effects of KHDRBS3 overexpression on KHDRBS2 mRNA levels in 22RV1 cells, and effects of KHDRBS2 overexpression on KHDRBS3 mRNA levels. Significant changes were observed. (G) Western blot analysis of KHDRBS2 and KHDRBS3 protein levels after overexpression of KHDRBS2 or KHDRBS3 in 22RV1 cells. (H–I) Effects of KHDRBS3 knockdown on KHDRBS2 expression, and effects of KHDRBS2 knockdown on KHDRBS3 expression in 22RV1 cells. Statistical significance is indicated. (J–K) Western blot analysis of KHDRBS2 and KHDRBS3 protein levels in 22RV1 cells after knockdown of KHDRBS2 or KHDRBS3. Error bars represent mean ± SD (n = 3). Statistical analysis was performed using one-way ANOVA followed by Tukey’s post-hoc test for multiple group comparisons and two-tailed unpaired t-tests for two-group comparisons (*p < 0.05, **p < 0.01, ****p < 0.0001). The original images of western-blot were cropped for favouring figure presentation. Raw images are provided in the Supplementary matarial-5 for reference To investigate potential regulatory relationships between KHDRBS2 and KHDRBS3, 22RV1 cells were transfected with plasmids overexpressing KHDRBS2 or KHDRBS3, and mRNA level and protein expression of the other protein were respectively detected by RT-qPCR and WB. As Fig. 2E and G shown, KHDRBS3 over-expression significantly decreased mRNA and protein levels of endogenous KHDRBS2, and vice versa (Fig. 2F and G). Besides, we also applied KHDRBS2 or KHDRBS3 specific shRNA to reduced endogenous protein expression and observe the changes of mRNA and protein levels of KHDRBS2 and KHDRBS3. To avoid off-target effect, we respectively used two pairs of shRNA targeting KHDRBS2 or KHDRBS3 to reduce protein expression. Consistently, either KHDRBS2 or KHDRBS3 knockdown enhanced the other protein expression at both mRNA (Fig. 2I and J) and protein levels (Fig. 2H and K). Same procedures were performed in PC3 cells and similar regulatory patterns were observed (Figure S3, Supplementary material-4). KHDRBS3 over-expression caused the decrease of KHDRBS2 mRNA level (p < 0.05, Figure S3A), and KHDRBS2 over-expression also reduced KHDRBS3 mRNA (p < 0.05, Figure S3B), with corresponding protein changes confirmed by WB (Figure S3C). Knockdown experiments in PC3 cells showed that reducing KHDRBS3 increased KHDRBS2 mRNA (p < 0.05, Figure S3D), and KHDRBS2 knockdown enhanced KHDRBS3 mRNA (p < 0.05, Figure S3E), with protein levels verified by WB analysis (Figure S3F and S3G). These findings reveal a consistent pattern of reciprocal regulation between KHDRBS2 and KHDRBS3 in both cell lines, suggesting a compensatory mechanism that may contribute to their distinct expression profiles in prostate cancer cells. KHDRBS2 and KHDRBS3 regulate prostate cancer cell growth in vitro and in vivo To investigate the functional roles of KHDRBS2 and KHDRBS3 in prostate cancer progression, we performed both in vitro and in vivo studies using AR-positive 22RV1 and AR-negative PC3 cell lines. CCK8 proliferation assays revealed that KHDRBS2 knockdown significantly suppressed 22RV1 cell growth in a time-dependent manner (Fig. 3A). The growth inhibition became apparent at 96 h post-transfection and continued throughout the observation period (192 h), with the most significant difference observed at the endpoint (p < 0.0001). Effects of KHDRBS2 and KHDRBS3 knockdown on prostate cancer cell proliferation in vitro and in vivo. (A, D) Proliferation of 22RV1 (A) and PC3 (C) cells after knockdown of KHDRBS2 or KHDRBS3, respectively, as measured by the CCK-8 assay at multiple time points. Knockdown of KHDRBS2 or KHDRBS3 significantly suppressed cell proliferation compared to the scramble control. (B, E) Representative images of tumor xenografts from nude mice injected with 22RV1 (B) or PC3 (E) cells transfected with scramble or shRNA targeting KHDRBS2 or KHDRBS3, respectively. Tumors on the right side correspond to shRNA-treated groups, while the left side corresponds to scramble controls. (C, F) Tumor weights of xenografts derived from 22RV1 (C) or PC3 (F) cells. Knockdown of KHDRBS2 or KHDRBS3 significantly reduced tumor weight compared to the scramble control. Error bars represent mean ± SD (n = 3). Statistical analysis was performed using two-way repeated measures ANOVA based on the general linear model (GLM) followed by Šídák’s correction method or two-tailed unpaired t-tests for two-group comparisons (*p < 0.05, **p < 0.01, ****p < 0.0001) To validate these findings in vivo, we established xenograft models using nude mice (n = 3 per group). Consistent with our in vitro observations, KHDRBS2-depleted 22RV1 xenografts exhibited markedly reduced tumor growth (Fig. 3B). Quantitative analysis revealed significantly decreased tumor weights in the shKHDRBS2 group compared to the scramble control (0.18 ± 0.02 g vs. 0.44 ± 0.15 g, p < 0.05; Fig. 3C). Parallel experiments conducted with KHDRBS3 knockdown in PC3 cells yielded similar results. In vitro proliferation was significantly impaired in KHDRBS3-depleted PC3 cells (Fig. 3D), with the growth inhibition becoming statistically significant at 48 h (p < 0.01) and maintaining this trend throughout the experiment (p < 0.0001 at 192 h). The in vivo xenograft model further confirmed these findings, with KHDRBS3 knockdown resulting in substantially reduced tumor burden (Fig. 3E). Final tumor weights in the shKHDRBS3 group were significantly lower than those in the control group (0.16 ± 0.08 g vs. 0.70 ± 0.30 g, p < 0.01; Fig. 3F). No significant changes were observed in cell cycle distribution, migration, or invasion assays for either KHDRBS2 knockdown in 22RV1 cells or KHDRBS3 knockdown in PC3 cells (figures are not showed here because of negative results). These findings demonstrate that KHDRBS2 and KHDRBS3 specifically regulate cell proliferation in AR-positive 22RV1 and AR-negative PC3 prostate cancer cells, respectively, as evidenced by both in vitro and in vivo studies. Expression patterns of AR variants and their regulation by KHDRBS2/3 To investigate the relationship between AR variants and KHDRBS family members, we first profiled AR variant expression across different prostate cancer cell lines. Quantitative PCR analysis revealed distinct expression patterns between AR-positive and AR-negative cell lines (Fig. 4A-C). Among the tested cell lines, 22RV1 exhibited significantly higher expression of AR-FL, AR-V7, and AR-V567es compared to LNCaP, DU145, and PC3 cells (p < 0.05). While both 22RV1 and LNCaP cells expressed detectable levels of AR variants, DU145 and PC3 cells showed minimal expression, consistent with their AR-negative status. Effects of KHDRBS2 and KHDRBS3 on AR variants expression in 22RV1 cells. (A–C) Basal mRNA levels of AR-FL, AR-V7, and AR-V567es in PCa cell lines 22RV1, LNCaP, DU145, and PC3, as determined by RT-qPCR. AR-V7 and AR-V567es were highly expressed in 22RV1 cells compared to other cell lines. (D–F) Effects of over-expression of KHDRBS2 and KHDRBS3 on mRNA levels of AR-FL, AR-V7, and AR-V567es in 22RV1 cells. Over-expression did not significantly alter expression levels. (G) Western blot analysis showing protein levels of AR, AR-V7, and AR-V567es after over-expression of KHDRBS2 or KHDRBS3 in 22RV1 cells. (H–J) Effects of knockdown of KHDRBS2 and KHDRBS3 on mRNA levels of AR-FL, AR-V7, and AR-V567es in 22RV1 cells. Knockdown of KHDRBS2 significantly increased AR-V7 expression, while no significant changes were observed for AR-FL or AR-V567es. (K) Western blot analysis showing protein levels of AR, AR-V7, and AR-V567es after knockdown of KHDRBS2 or KHDRBS3 in 22RV1 cells. Error bars represent mean ± SD (n = 3). Statistical analysis was performed using one-way ANOVA followed by Tukey’s post-hoc test for multiple group comparisons (*p < 0.05, **p < 0.01, ns: not significant). The original images of western-blot were cropped for favouring figure presentation. Raw images are provided in the Supplementary matarial-5 for reference To elucidate the regulatory roles of KHDRBS2 and KHDRBS3 in AR variant expression, we performed both gain- and loss-of-function studies in 22RV1 cells. Notably, transient over-expression of either KHDRBS2 or KHDRBS3 did not significantly alter the mRNA levels of AR-FL, AR-V7, or AR-V567es (p > 0.05, Fig. 4D-F). These findings were further validated at the protein level through WB analysis (Fig. 4G). However, gene silencing experiments revealed specific regulatory relationships. KHDRBS2 knockdown resulted in a significant decrease in AR-V7 mRNA (p < 0.01, Fig. 4I), while AR-FL and AR-V567es mRNA levels remained unchanged (Fig. 4H and J). Conversely, KHDRBS3 knockdown led to increased AR-V7 mRNA (p < 0.05, Fig. 4I), suggesting an antagonistic relationship between KHDRBS2 and KHDRBS3 in regulating AR-V7 mRNA. WB analysis confirmed these transcriptional changes at the protein level (Fig. 4K). These results indicate that KHDRBS2 and KHDRBS3 specifically modulate AR-V7 expression through distinct mechanisms, while having minimal impact on other AR variants. Androgen-dependent regulation of AR variants and KHDRBS family members in 22RV1 cells We investigated the expression patterns of AR variants and KHDRBS family members under different androgenic conditions in 22RV1 cells. Under androgen-deprived conditions (CSS treatment for 48 h), AR-FL, AR-V7, and AR-V567es showed significant up-regulation compared to FBS controls. Subsequent treatment with R1881 (1nM) for 24 h partially attenuated this up-regulation, with significant decreases observed for AR-V567es (p < 0.05) and modest reductions in AR-FL and AR-V7 levels (Fig. 5A-C). Analysis of AR variants and KHDRBS family gene expression under different androgen conditions in 22RV1 cells. (A-C) RT-qPCR analysis of AR-FL, AR-V7, and AR-V567es mRNA levels in 22RV1 cells treated with FBS (+/48hr), CSS (+/48hr), or CSS + R1881 (1nM, +/24hr). (D-F) Expression analysis of KHDRBS1, KHDRBS2, and KHDRBS3 mRNA levels under the same treatment conditions as above. (G-J) Regulatory relationships between KHDRBS family members under different androgen conditions. (G) KHDRBS2 expression following KHDRBS3 overexpression. (H) KHDRBS3 expression following KHDRBS2 overexpression. (I) KHDRBS2 expression following KHDRBS3 knockdown. (J) KHDRBS3 mRNA levels following KHDRBS2 knockdown. (K-L) Impact of KHDRBS2 downregulation on AR variant expression under different androgen conditions. (K) AR-FL mRNA levels. (L) AR-V7 mRNA levels following KHDRBS2 knockdown. Error bars represent mean ± SD (n = 3). Statistical analysis was performed using one-way ANOVA followed by Tukey’s post-hoc test for multiple group comparisons (*p < 0.05, **p < 0.01, ns: not significant). FBS: fetal bovine serum; CSS: charcoal-stripped serum; R1881: synthetic androgen. Further experimental details are shown in Figure S4 Expression analysis of KHDRBS family members revealed parallel responses to androgenic manipulation. KHDRBS1 expression increased significantly under androgen-deprived conditions (p < 0.001) and was partially reversed by R1881 treatment (p < 0.05, Fig. 5D). Similarly, KHDRBS2 showed significant up-regulation in androgen-deprived conditions (p < 0.01) with subsequent reduction upon R1881 treatment (p < 0.05, Figure E). While KHDRBS3 displayed a similar trend, the changes did not reach statistical significance (Fig. 5F). To explore the regulatory relationship between KHDRBS2/3 and AR variants under different androgenic conditions, reciprocal regulation was observed through over-expression and knockdown experiments. Over-expression of either factor significantly suppressed the other’s expression, while knockdown of one led to increased expression of the other, in both CSS and R1881 conditions (p < 0.05 or p < 0.01, Fig. 5G-J). Notably, while KHDRBS2 knockdown did not significantly affect AR-FL mRNA (Fig. 5K), it markedly reduced AR-V7 levels under both CSS and R1881 conditions (p < 0.01, Fig. 5L). This selective effect on AR-V7 suggests a specific regulatory role for KHDRBS2 in AR splice variant expression. These findings demonstrate that androgen deprivation coordinately regulates AR variants and KHDRBS family members in 22RV1 cells, with KHDRBS2 playing a particularly important role in maintaining AR-V7 expression regardless of androgenic conditions. Discussion The regulation of AR alternative splicing represents a critical mechanism underlying CRPC progression. Our study reveals important functional roles for KHDRBS2 and KHDRBS3, two understudied members of the STAR protein family, in AR splice variant regulation and prostate cancer pathogenesis. Notably, the expression pattern of KHDRBS family members in PCa cell lines resembles that reported in mouse brain cells [24–26], suggesting evolutionarily conserved regulatory networks. We focused on 22RV1 and PC3 cell lines based on their contrasting AR status [34, 35] (AR-positive and AR-negative, respectively) and differential KHDRBS2/KHDRBS3 expression, enabling investigation of context-dependent functions in distinct PCa backgrounds. Gene silencing experiments revealed reciprocal regulation between KHDRBS2 and KHDRBS3, characterized by mutually exclusive expression patterns. This phenomenon parallels the well-characterized PTBP1/PTBP2 regulatory circuit in neuronal development, where competitive binding and auto-regulatory feedback maintain developmental transitions [36, 37]. Importantly, reciprocal regulation between KHDRBS2 (SLM1) and KHDRBS3 (SLM2) has been extensively characterized in neuronal cells. Iijima et al. demonstrated that these proteins exhibit mutually exclusive expression patterns in different neuronal cell types through post-transcriptional regulatory mechanisms involving RNA-binding and alternative splicing control [25]. As members of the STAR/KH domain protein family, KHDRBS proteins function primarily as RNA-binding proteins that regulate alternative splicing and mRNA metabolism [24]. In our prostate cancer cell experiments, we observed coordinate directional changes in both mRNA and protein levels upon overexpression or knockdown, consistent with RNA-level regulatory mechanisms. However, we note that direct quantitative comparison of fold changes between mRNA (qPCR-based) and protein (Western blot-based) measurements can be influenced by technical factors including primer/antibody specificity, detection range differences, and the presence of alternatively spliced isoforms. Given the differential AR status of 22RV1 (AR-positive) and PC3 (AR-negative) cells and the AR variant expression profile in 22RV1 cells [38, 39], this reciprocal regulation may represent an adaptive mechanism allowing cells to modulate AR signaling in response to selective pressure during disease progression. The molecular basis for this cross-regulation likely involves competition for shared binding sites or regulatory factors, consistent with the known functions of STAR family proteins as splicing regulators. KHDRBS proteins may directly or indirectly influence each other’s expression through splicing-dependent or post-transcriptional mechanisms. Detailed mapping of binding sites through CLIP-seq and identification of shared regulatory networks will be essential to fully elucidate these regulatory circuits. Androgen deprivation and R1881 stimulation experiments in 22RV1 cells revealed coordinate changes in AR, AR-V7, AR-V567es, KHDRBS1, and KHDRBS2 expression, indicating shared androgen-responsive regulatory mechanisms. Importantly, R1881 treatment failed to fully reverse androgen deprivation-induced elevation of AR variants, which may reflect the incomplete resensitization observed clinically during intermittent androgen deprivation therapy, ultimately contributing to treatment failure in CRPC patients [40, 41]. Critically, KHDRBS2 knockdown in 22RV1 cells selectively reduced AR-V7 expression without affecting full-length AR, indicating specific involvement in alternative splicing rather than general AR transcription. This finding expands the repertoire of AR-V7 splicing regulators beyond previously identified factors including U2AF65 [40], DDX39B [42], AURKA [43], CLK1 [44] and KHDRBS1 [19, 20]. The STAR protein family, characterized by conserved KH domains essential for RNA recognition, has emerged as an important regulator of AR variant generation. While KHDRBS1 functions as a positive regulator of AR-V7 through direct AR pre-mRNA interaction [19, 20], our findings suggest KHDRBS2 operates through similar but potentially non-redundant pathways. The selective impact on AR-V7 without affecting full-length AR indicates that KHDRBS2 likely promotes inclusion of cryptic exons or utilization of alternative splice sites specific to variant generation. Based on established mechanisms of STAR protein function, we propose that KHDRBS2 facilitates AR-V7 generation through recognition of splicing regulatory elements within AR introns 2–3, where cryptic splice sites are activated during variant formation. STAR proteins typically bind purine-rich enhancer sequences and recruit core splicing machinery to promote exon definition [23]. In the context of AR-V7, KHDRBS2 may stabilize binding of U2AF65 or other splicing factors at the cryptic 3’ splice site in intron 3, promoting premature polyadenylation and generation of constitutively active truncated receptors [8, 40]. This model (illustrated as figure S5) is consistent with the functional outcomes we observed but requires validation through direct RNA-binding assays and splicing reporter experiments to map precise interaction sites and define the sequence determinants of KHDRBS2-mediated regulation. The predominant expression of KHDRBS2 in AR-positive 22RV1 cells and its specific role in AR-V7 regulation positions it as a potential therapeutic target for CRPC patients who develop resistance to next-generation antiandrogens through AR variant expression. Furthermore, the elevated alteration frequencies of KHDRBS family members in neuroendocrine prostate cancer (NEPC) samples suggest that these proteins may play crucial roles in the transdifferentiation process that leads to treatment-emergent NEPC. This finding opens new avenues for understanding and potentially preventing this lethal form of treatment resistance. While our study establishes important functional roles for KHDRBS2 and KHDRBS3, several mechanistic questions remain. Additionally, the precise binding sites within AR pre-mRNA and the identity of interacting splicing factors remain undefined.Future studies employing CLIP-seq or related RNA-binding mapping technologies will be essential to identify direct KHDRBS2/3 binding sites on AR transcripts and define sequence-specific recognition motifs. Proteomic approaches such as proximity labeling or co-immunoprecipitation coupled with mass spectrometry can identify protein interaction networks that modulate KHDRBS splicing activity. Furthermore, development of small-molecule inhibitors or antisense oligonucleotides targeting KHDRBS2 function represents a promising therapeutic strategy that should be explored in preclinical models. In conclusion, our study establishes KHDRBS2 and KHDRBS3 as functionally important regulators of prostate cancer progression through their roles in AR variant regulation and cellular proliferation control. By identifying novel components of the AR splicing regulatory network, this work provides a foundation for developing therapeutic strategies targeting RNA processing machinery in CRPC, potentially offering new treatment options for patients with advanced disease resistant to conventional androgen-targeted therapies. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1. Cell line sources and ATCC cell culture recommendations. Primer information used in the study. Primary antibodies used in the study. The plasmids and shRNA used in the study. Datasets used for cBioPortal Data Analysis. Figure S4. Experimental design of androgen stimulation assay in 22RV1 cells. Figure S5. KHDRBS2/3-Mediated AR Splicing Regulation Mechanis. Supplementary Material 2. Figure S1. Genetic Alterations in KHDRBS1-3 Genes in Prostate Cancer. Supplementary Material 3. Figure S2. Genetic Alterations in KHDRBS1-3 Genes and Their Clinical Implications in Prostate Cancer. Supplementary Material 4. Figure S3. Expression of KHDRBS family genes in prostate cancer cell lines. Supplementary Material 5. 1. Raw images of western-blot of Fig. 2-D. 2.Raw images of western-blot of Fig. 2-G. 3. Raw images of western-blot of Fig. 2-J. 4. Raw images of western-blot of Fig. 2-K. 5.Raw images of western-blot of Fig. 4-G. 6. Raw images of western-blot of Fig. 4-K. 7. Raw images of western-blot of Figure S3-C. 8.Raw images of western-blot of Figure S3-F. 9. Raw images of western-blot of Figure S3-G. Acknowledgements We would like to express our sincere gratitude to Prof. Jinming Zhou (Institute of Medicinal Biotechnology, Chinese Academy of Medical Sciences & Peking Union Medical College) for kindly providing the LNCaP and 22RV1 human prostate cancer cell lines. Our thanks also go to Prof. Dong Wang (School of Medicine, Tsinghua University) for generously gifting the PC3 cells. Author contributions XuDong Yang: Study conception and design, data collection and statistical analysis, manuscript writing and revision.Xia Liu: Experimental guidance, standardization, and quality management.Ran Xiao: Statistical analysis guidance, manuscript review and revision.YiLin Cao: Study conception and design, manuscript review and critical revisions, project supervision. Funding No funds, grants, or other support was received. Data availability All relevant data supporting the findings of this study are available from the corresponding author upon reasonable request. Declarations Ethics approval and consent to participate All animal procedures were approved by the Institutional Animal Care and Use Committee of Plastic Surgery Hospital (Institute) of Union Medical University, Beijing (Approval No. 2016-093). The maximum permitted tumor burden was 2000 mm³ or a tumor diameter of 1.5 cm in any direction, or earlier if animals showed signs of distress, ulceration, or impaired mobility. Tumor growth was closely monitored throughout the study. In a small subset of animals, tumor volume marginally exceeded the permitted limit due to unexpectedly rapid tumor progression during the final observation window. These cases were promptly identified, and animals were humanely euthanized immediately upon detection in accordance with IACUC-approved humane endpoint criteria. All methods were conducted in accordance with relevant institutional, national, and international guidelines and regulations. Not applicable (no human participants involved). Consent for publication Not applicable (no human participants involved). Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information XuDong Yang, Email: lastsir.yee@163.com. YiLin Cao, Email: yilincao163@163.com. References Associated Data Supplementary Materials Supplementary Material 1. Cell line sources and ATCC cell culture recommendations. Primer information used in the study. Primary antibodies used in the study. The plasmids and shRNA used in the study. Datasets used for cBioPortal Data Analysis. Figure S4. Experimental design of androgen stimulation assay in 22RV1 cells. Figure S5. KHDRBS2/3-Mediated AR Splicing Regulation Mechanis. Supplementary Material 2. Figure S1. Genetic Alterations in KHDRBS1-3 Genes in Prostate Cancer. Supplementary Material 3. Figure S2. Genetic Alterations in KHDRBS1-3 Genes and Their Clinical Implications in Prostate Cancer. Supplementary Material 4. Figure S3. Expression of KHDRBS family genes in prostate cancer cell lines. Supplementary Material 5. 1. Raw images of western-blot of Fig. 2-D. 2.Raw images of western-blot of Fig. 2-G. 3. Raw images of western-blot of Fig. 2-J. 4. Raw images of western-blot of Fig. 2-K. 5.Raw images of western-blot of Fig. 4-G. 6. Raw images of western-blot of Fig. 4-K. 7. Raw images of western-blot of Figure S3-C. 8.Raw images of western-blot of Figure S3-F. 9. Raw images of western-blot of Figure S3-G. Data Availability Statement All relevant data supporting the findings of this study are available from the corresponding author upon reasonable request.