Stromal Steroid 5 Alpha-Reductase 2 Promotes Prostate Growth through WNT5A–Lymphoid Enhancer–Binding Factor 1–Insulin-Like Growth Factor 1 Signaling in Benign Prostatic Hyperplasia

Abstract

Steroid 5 α-reductase 2 (SRD5A2) is a key enzyme in androgen metabolism and a pharmacologic target in benign prostatic hyperplasia. Although SRD5A2 is known to mediate stromal-epithelial interactions that influence prostate growth, the relationship between baseline SRD5A2 expression and prostate volume remains unclear. In this study, SRD5A2 expression was analyzed in human prostate tissues from the Medical Therapy of Prostatic Symptoms trial and institutional biorepository cohorts. Quantitative assessments were performed and correlations were evaluated between expression level of SRD5A2, WNT5A, prostate volume, and tissue signaling profiles. SRD5A2 expression was significantly associated with total prostate and transition zone volume. Stromal-specific WNT5A expression showed a strong positive correlation with SRD5A2, whereas neither serum nor tissue dihydrotestosterone levels correlated with SRD5A2 expression. In Srd5a2-null mice, Wnt5a expression in the prostate stroma was dependent on Srd5a2 and showed region-specific regulation. Mechanistically, SRD5A2 overexpression in human prostate stromal cells up-regulated WNT5A and lymphoid enhancer–binding factor 1, activated insulin-like growth factor 1 (IGF1) signaling, increased proliferation, and reduced apoptosis. Conditioned media from these cells enhanced epithelial proliferation through paracrine IGF1 activity. This study provides the first evidence that SRD5A2 promotes prostate growth through a stromal WNT5A–lymphoid enhancer–binding factor 1–IGF1 paracrine signaling axis independent of androgen levels, suggesting a novel therapeutic mechanism relevant for patients with benign prostatic hyperplasia with resistance to conventional 5 α-reductase inhibitor therapy.

Graphical abstract

Benign prostatic hyperplasia (BPH) is a highly prevalent condition among aging men and a leading cause of lower urinary tract symptoms. One of the primary drivers of prostate growth in adulthood is androgen signaling, mediated by the binding of androgens to the androgen receptor (AR), which activates downstream transcriptional programs that promote cellular proliferation and differentiation.¹ The steroid 5 α-reductase 2 (SRD5A2), which is highly expressed in prostate tissue, catalyzes the conversion of testosterone into dihydrotestosterone (DHT), a more potent androgen that stimulates both epithelial and stromal proliferation.² Elevated SRD5A2 activity has been associated with increased intraprostatic DHT levels and larger prostate volumes.^3,4

Clinically, 5 α-reductase inhibitors (5ARIs), such as finasteride and dutasteride, are commonly prescribed to reduce prostate volume by blocking DHT synthesis and dampening androgen-driven prostate growth.⁵ However, finasteride reduces the progression of lower urinary tract symptoms by only approximately 34%, suggesting that androgen-independent mechanisms may also contribute to prostate enlargement and therapeutic resistance.⁶

Among the key pathways implicated in tissue growth regulation is the WNT signaling pathway, which controls epithelial proliferation, stem cell maintenance, and differentiation in multiple organs.^7,8 In the prostate, canonical WNT signaling promotes basal stem cell proliferation via activation of the β-catenin–T-cell factor (TCF)/lymphoid enhancer–binding factor (LEF) transcription complex, whereas noncanonical WNT signaling, particularly through WNT5A, facilitates stromal-epithelial interactions via calcium-dependent signaling cascades.^9,10 WNT5A, secreted predominantly by prostate stromal cells, has been shown to promote epithelial proliferation through paracrine mechanisms, independent of classic AR activation.^10,11 Importantly, WNT5A signaling becomes up-regulated in conditions of androgen deprivation, further supporting its role in maintaining prostate growth under low-androgen states, such as during 5ARI therapy.¹² In response to reduced androgen levels, stromal cells increase WNT5A secretion, supporting continued epithelial proliferation.¹² Given that a subset of patients with BPH do not respond to androgen-targeting therapies, WNT5A and its downstream signaling components may provide an alternative pathway that promotes prostate growth independently of DHT levels.

Despite advances in the understanding of SRD5A2 and WNT signaling individually, the molecular relationship between these pathways and their combined contribution to prostate growth remain poorly defined. Specifically, it is unclear whether WNT5A acts downstream of SRD5A2 or constitutes an independent regulatory mechanism. Although SRD5A2's role in DHT synthesis is well established, emerging data suggest that prostate enlargement may also be driven by alternative stromal-derived signals, such as WNT5A, that function independently of androgen levels.^11,13

In this study, we investigated the interplay between SRD5A2 and WNT signaling, with a particular focus on WNT5A as a potential mediator of SRD5A2-driven prostate growth through the integration of clinical data sets, preclinical models, and molecular assays. We hypothesize that WNT5A serves as a key DHT-independent effector linking SRD5A2 expression to prostate volume expansion. The aim of this study is to elucidate a novel stromal signaling axis that contributes to BPH pathogenesis and may offer a therapeutic target for patients unresponsive to conventional androgen-deprivation therapies.

Materials and Methods

Clinical Samples

De-identified human prostate biopsy samples (n = 61) collected during the Medical Therapy of Prostatic Symptoms (MTOPS) trial were obtained from the NIH/National Institute of Diabetes and Digestive and Kidney Diseases Central Repository under an approved X01 application. Relevant clinical data from the MTOPS trial, including total prostate volume, transition zone volume, and serum DHT levels, were obtained and used for correlative analysis.

Surgical prostate tissue specimens and blood samples were obtained from the Beth Israel Deaconess Medical Center (BIDMC) Prostate Tissue Biorepository under an institutional review board–approved protocol (2019P000728). All patients provided written informed consent for tissue collection and research use. Fresh prostate tissue was collected from patients undergoing transurethral resection of the prostate surgery, and processed within 2 hours after the surgery. Samples were stored at –80°C until use. A total of 49 BPH samples from patients aged 49 to 89 years were included. All specimens were de-identified before analysis. Tissues were selected on the basis of availability of matched clinical data and sufficient tissue quality. Relevant clinical data, including total prostate volume and transition zone volume, derived from prostate magnetic resonance imaging, computed tomography scans of the abdomen and pelvis, and genitourinary ultrasounds before the surgical date, were collected and used for correlative analysis. Tissues were used for immunohistochemistry, enzyme-linked immunosorbent assay (ELISA), RNA extraction, and Western blot analysis. Serum was used for DHT test.

Animal Studies

Generation of Srd5a2-null (Srd5a2^–/–) and heterozygous (Srd5a2^+/–) mice model was previously described.¹⁴ Mice were housed in standard polycarbonate cages with free access to normal food and water. Biological males were only studied because the prostate is the organ specific to biological male animals. All animal work described in this study was in adherence to the NIH Guide for the Care and Use of Laboratory Animals¹⁵ and approved by the Beth Israel Deaconess Medical Center Institutional Animal Care and Use Committee (protocol number 049-2022). No randomization was used.

Two-Dimensional Cell Culture

Benign human prostate epithelial cell line 1 (BHPrE1), benign human prostate stromal cell line 1 (BHPrS1), and BPH-1 cell line were gifted by Dr. Simon Hayward (North Shore University Health System Research Institute, Chicago, IL) since 2019. BHPrS1 and BPH-1 cells were maintained in RPMI 1640 medium (Gibco, Thermo Fisher Scientific, Waltham, MA; 11875119) with 10% v/v fetal bovine serum (FBS; Gibco; A5256801) and 1% penicillin/streptomycin (30-002-CI; Corning, Corning, NY). Vector control and SRD5A2-overexpressing BHPrS1 stable cell lines were established by transfecting cells with 1% SRD5A2-sequence coded lentiviral particles (GeneCopoeia; Rockville, MD; A3930) and 0.1% polybrene (Millipore, Burlington, MA; TR1003). Positive cells were selected and maintained using 0.5% puromycin (Gibco; A1113802). Transfection efficiency was tested using quantitative PCR. BHPrE1 cell was cultured in Dulbecco’s modified Eagle’s medium/F-12 (Gibco; 11320033) with 5% FBS, 1% v/v penicillin/streptomycin, 0.4% Bovine Pituitary Extract (Gibco; 13028014), 1% Insulin-Transferrin-Selenium (Gibco; 41400045), and 10 ng/mL Human EGF (Gibco; PHG0314). All cell lines were cultured in corresponding medium in a humidified incubator with 5% CO₂ at 37°C. All cell lines used in this study were authenticated and intermittently evaluated in-house for Mycoplasma contamination (ant-mpt; InvivoGen, San Diego, CA).

For cell seeding, cells were washed once with sterile 1× phosphate-buffered saline (PBS; SH30256.FS; Cytiva, Marlborough, MA) and detached with trypsin (Gibco; 2520114). Detached cells were counted using hemocytometer, and then seeded onto a 6-well plate.

For two-dimensional stromal-epithelial transwell co-culture, polycarbonate membrane cell culture inserts (Thermo Fisher Scientific; 140660) in 6-well plates were used following manufacturer's protocol with minor modifications. Briefly, 1.5 × 10⁵ cells/well of BPH-1 cells were seeded into the bottom of the 6-well plates, and 1 × 10⁵ cells/well of BPH-1 or BHPrS1 cells were seeded into inserts above. The insert and bottom of the well were each maintained with 1 mL of RPMI 1640 medium with 10% FBS. The cells in the inserts were treated with drugs or siRNAs, and the BPH-1 epithelial cells in the bottom well were collected for further analysis.

RNA Transfection and Drug Treatment

For gene silencing with siRNAs, BHPrS1 cells were seeded in 6-well plates and cultured for 24 hours before transfection with control (sc-37007; Santa Cruz Biotechnology, Dallas, TX) or LEF1 (sc-35804; Santa Cruz Biotechnology) siRNAs using Lipofectamine RNAiMAX Reagent (13778150; Invitrogen, Waltham, MA) and Opti-MEM (Gibco; 51985034), according to the manufacturer's instructions. The transfected cells were cultured for 48 hours at 37°C with 5% CO₂ before proceeding with immunocytochemistry, RNA extraction, MTT assay, or sample preparation steps for Western blot analysis.

For drug treatment and AR activation, BHPrS1 cells seeded in 6-well plates were starved for 24 hours before adding the corresponding drugs. Starved cells were then treated for 48 to 72 hours with 100 μmol/L finasteride (F1293; SigmaAldrich, St. Louis, MO), 100 μmol/L WNT5A-derived N-butyloxycarbonyl hexapeptide (BOX5; SigmaAldrich; 681673), or 50 nmol/L picropodophyllin (PPP; Thermo Fisher Scientific; J65916MA), with fresh drugs and medium changed every 24 hours. Similarly, androgen stimulation was performed by preparing the corresponding concentration of the testosterone and DHT in RPMI 1640 medium with 10% charcoal-stripped FBS instead of regular FBS, and adding to the cells for 72 hours of treatment after starvation with fresh androgen and medium changed every 24 hours.

Organoid and Spheroid Culture

The BHPrE1 organoid and BHPrS1 spheroid culture were performed using Matrigel (Corning; 356230) based on the manufacturer's instruction and previous publications16, 17, 18, 19 with minor modifications. Briefly, 700 cells/well of BHPrS1, 1400 cells/well of BHPrE1, or 700 BHPrS1 cells plus 1400 BHPrS1cells per well for co-culture were suspended in 10 μL of Dulbecco’s modified Eagle’s medium/F-12 medium and then mixed with 290 μL of 5 μg/mL of Matrigel. The 300 μL of cell-Matrigel mixture was then added to a 24-well plate with 100 μL precasted Matrigel matrix base in each well. After matrix gel formed, organoid cultures were maintained in 500 μL of Dulbecco’s modified Eagle’s medium/F-12 media that was changed every 3 to 4 days. The organoids and spheroids were imaged under Nikon (Minato City, Tokyo, Japan) Eclipse TS2 microscope with CaptaVision+ software version 2.0.3.0 (ACCU-SCOPE, Commack, NY). The surface area of organoids and spheroids occupied in ×4 microscopic images was measured by analyze particle function of ImageJ version 1.54p (NIH, Bethesda, MD; https://imagej.net/ij). The BHPrS1 spheroid colonies were treated with 100 μmol/L finasteride or dimethyl sulfoxide for 72 hours after 14 days and collected after 17 days, following a previously published protocol.^18,20 The fixed spheroid colonies were analyzed with immunofluorescence for Ki-67 and LEF1.

Softy Agar Colony Formation Assay

BPH-1 growth in different conditioned media was determined by assaying colony formation in soft agar.²¹ Briefly, 5000 BPH-1 cells suspended in 2× RPMI 1640 medium (Gibco; 31800022) with 20% FBS were mixed with 0.6% Noble Agarose (A20460; RPI, Mount Prospect, IL) and then seeded on top of 1% noble agarose layer in triplicate in 6-well plates. The cells were topped with 500 μL condition medium and cultured for 21 days in a humidified incubator containing 5% CO₂ at 37°C with medium replacement every 3 to 4 days. Colonies formed were then stained with 1 mg/mL crystal violet in 1× PBS with 10% methanol overnight. Images of colonies were taken under a white view using Nikon Eclipse TS2 microscope and CaptaVision+ software version 2.0.3.0.

Cell Viability Assay

The proliferation of BHPrS1 stromal cells was measured with MTT assay (11465007001; Roche, Basel, Switzerland), as previously described,²² Briefly, 1000 cells/well of BHPrS1 cells in 200 μL/well of culture medium were seeded in 96-well plates and then incubated for 72 hours with corresponding treatments. The proliferation of cells was determined by the relative absorbance reading of 590 nm measured using a Molecular Device (San Jose, CA) SpectraMax id5 plate reader.

Immunocytochemistry

Cells seeded on coated coverslips (7229011; Electron Microscopy Sciences, Hatfield, PA) in 12-well plates were fixed in 4% paraformaldehyde (Electron Microscopy Sciences; 15710), and non-specific binding of antibodies was blocked using 3% bovine serum albumin (P-753; Boston Bioproducts, Milford, MA) in 1× PBS. Fixed cells were permeabilized with 0.1% Triton X-100 (MP Biomedical, Santa Ana, CA; 04807426) in 1× PBS and incubated with primary antibody diluted in 3% bovine serum albumin in 1× PBS for 90 minutes, followed up with corresponding secondary antibodies for 60 minutes at room temperature. After excess antibodies washed away, samples were mounted with antifade reagent with DAPI (Invitrogen; S36939). For quantitative analysis, the fluorescence of the corresponding antibodies was imaged with Keyence (Osaka, Japan) BZ-X 2000 microscope, and the intensity was measured with ImageJ version 1.54p. Sources for all antibodies used in this study are listed in Table 1.

Table 1.

Antibodies

Target	Company	Catalog no.	Dilution	Application
SRD5A2	Abcam (Cambridge, UK)	124877	1:1000	Western blot analysis
WNT5A	Abcam	235966	1:1000	Western blot analysis
β-Catenin	Cell Signaling Technology (Danvers, MA)	8480	1:1000	Western blot analysis
PARP1	Cell Signaling Technology	9542	1:1000	Western blot analysis
AR	Abcam	133273	1:1000	Western blot analysis
PKCα	Cell Signaling Technology	5975	1:1000	Western blot analysis
Phosphorylated PKC	Cell Signaling Technology	9371	1:1000	Western blot analysis
GAPDH	Cell Signaling Technology	2118	1:5000	Western blot analysis
β-Actin	Cell Signaling Technology	4967	1:5000	Western blot analysis
Anti-rabbit HRP	Cell Signaling Technology	7074	1:2500–1:5000	Western blot analysis
Ki-67	Abcam	Ab16667	1:1000	ICC
Cleaved caspase-3	Cell Signaling Technology	9664	1:400	ICC
LEF1	Cell Signaling Technology	2230	1:400	ICC
Rhodamine phalloidin (actin stain 535)	Cytoskeleton (Denver, CO)	PHDR1	1:500	ICC
Donkey anti-rabbit Alexa Fluor 488	Invitrogen	A21206	1:800	ICC
SRD5A2	Invitrogen	PA5-42550	1:200	IHC
WNT5A	NovusBio (Centennial, CO)	NBP3-24567	1:500	IHC

AR, androgen receptor; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; HRP, horseradish peroxidase; ICC, immunocytochemistry; IHC, immunohistochemistry; PARP1, poly (ADP-ribose) polymerase 1; PKC, protein kinase C.

Immunohistochemistry

Immunohistochemistry staining on MTOPS biopsy samples and BIDMC biorepository surgical samples was performed as previously described.¹⁴ Briefly, fixed prostate tissue sections undergo deparaffinization, antigen retrieval, blocking, antibody application (Table 1), and counterstaining for visualizing SRD5A2 and WNT5A. Probed sections were imaged, and average immunoreactive score was calculated following the same protocol established in the previous publication.¹⁴

Immunofluorescence

Immunofluorescence staining on BHPrS1 spheroid-like colonies was performed similarly to immunocytochemistry described in Immunocytochemistry, with minor modifications. Briefly, cell colonies were collected and fixed on coverslips in 12-well plates with 4% paraformaldehyde. Rinsed plates were frozen at –20°C degrees for storage. On thawing, the colonies on coverslips were blocked, permeabilized, probed with corresponding antibodies (Table 1), and imaged following the immunocytochemistry protocol.

Cell Lysate Preparation and Western Blot Analysis

Cells cultured in 6-well plates were lysed in 20 volume size of 1× M-PER buffer (Thermo Fisher Scientific; 78501) with 0.1% SDS (Boston BioProducts, Milford, MA; BM-230) and 1% protein and protease inhibitor cocktail (Thermo Fisher Scientific; 78445), and chilled on ice for 30 minutes with constant shaking. Insoluble cell debris was removed after centrifugation at 13,200 × g for 15 minutes at 4°C, and the protein concentration of the clear supernatant was measured with the bicinchoninic acid assay (Thermo Fisher Scientific; 23227).

Western blot analysis was performed as previously described²³ with minor modifications. Cell lysate samples were diluted with lysis buffer to reach the same protein concentration and then denatured in 6× Laemmli Buffer (J61337; AlfaAesar, Ward Hill, MA) with 10% 2-mercaptoethanol (SigmaAldrich; M3148) for 5 minutes at 95°C. Denatured protein samples were run on 8% to 16% SDS-PAGE gels (4568104; Bio-Rad, Hercules, CA) at 90 V for 90 minutes and subsequently transferred to polyvinylidene difluoride membrane (Millipore; IPVH00010) for 3 hours on 250 mA at 4°C. The membrane was blocked with 5% (w/v) nonfat dry milk (M0841; Labscientific, Danvers, MA) in 1× Tris-buffered saline buffer (diluted from 20× Tris-buffered saline buffer; Boston Bioproducts; BM-301X), followed by probing with corresponding antibodies diluted in 5% bovine serum albumin in 1× Tris-buffered saline buffer with 0.05% Tween-20 (BP337-500; FisherScientific, Pittsburgh, PA). Probed membranes were developed using enhanced chemiluminescence reagents (Cytiva; RPN2232) and imaged with LI-COR (Lincoln, NE) Odyssey M imager. Antibodies used in this study were obtained from various sources and are listed in Table 1.

qPCR and PCR Microarray

Real-time quantitative PCRs (qPCRs) were performed as previously described²³ and run on QuantStudio 6 Flex (Applied Biosystems, Waltham, MA). GraphPad Prism 10 was used to calculate the ratio of mRNA levels of assayed genes to values of glyceraldehyde-3-phosphate dehydrogenase control using the ΔC_T method (2^−ΔΔC_T). Primers were ordered from IDT (Newark, NJ) and are listed in Table 2.

Table 2.

PCR Primers

Species	Gene symbols	Forward sequence	Reverse sequence
Human	SRD5A2	5′-GCAGTGTCTTAGTTGATGAG-3′	5′-TGTGGTTATTAAAACCTGGC-3′
Human	WNT5A	5′-GGCAAGTTGGTACAGGTCAA-3′	5′-ATTGCGCACGCAGTAGT-3′
Human	CTNNB1	5′-CACAAGCAGAGTGCTGAAGGTG-3′	5′-GATTCCTGAGAGTCCAAAGACAG-3′
Human	LEF1	5′-CTACCCATCCTCACTGTCAGTC-3′	5′-GGATGTTCCTGTTTGACCTGAGG-3′
Human	TCF7L1	5′-TCAAGGACACGAGGTCACCATC-3′	5′-GGAGAAGTGGTCATTGCTGTAGG-3′
Human	TCF7L2	5′-CCCTTTGCTCTTATATCTGACTTCT-3′	5′-TTCTGCACGCGTGAGTTT-3′
Human	CCND1	5′-TCTACACCGACAACTCCATCCG-3′	5′-TCTGGCATTTTGGAGAGGAAGTG-3′
Human	MYC	5′-CCTGGTGCTCCATGAGGAGAC-3′	5′-CAGACTCTGACCTTTTGCCAGG-3′
Human	CYCS	5′-AAGGGAGGCAAGCACAAGACTG-3′	5′-CTCCATCAGTGTATCCTCTCCC-3′
Human	CASP3	5′-GGAAGCGAATCAATGGACTCTGG-3′	5′-GCATCGACATCTGTACCAGACC-3′
Human	CASP9	5′-GTTTGAGGACCTTCGACCAGCT-3′	5′-CAACGTACCAGGAGCCACTCTT-3′
Human	PARP1	5′-TGATAGCAGCAAGGATCCCAT-3′	5′-CCGTGCCACAGCAATCTTCG-3′
Human	CACNA2D3	5′-GAACATCCCGATGTGTCCTTGG-3′	5′-ACTGGAGCAGAGGTTCTTTGCC-3′
Human	CAMKIIA	5′-GAGCCATTCTCACCACGATGCT-3′	5′-TGGTGTTGGTGCTCTCTGAGGA-3′
Human	PTPRK	5′-GGACTGAATCCTGGCACTTT-3′	5′-GCCAATCTCTACCCGTGAATC-3′
Human	RYK	5′-CAGCAAGACCTGGTACACATGG-3′	5′-CAAGTCTCTGGAGAGGGCATTG-3′
Human	ROR1	5′-GAGGCAACCAAAACACGTCAGAG-3′	5′-GGCACACTCACCCAATTCTTCC-3′
Human	ROR2	5′-GTACGCATGGAACTGTGTGACG-3′	5′-AAAGGCAAGCGATGACCAGTGG-3′
Human	LRP6	5′-CAGTTGGAGTGGTGCTGAAAGG-3′	5′-CCATCCAAAGCAGCCCGTTCAA-3′
Human	TGFBR1	5′-TCGAGATAGGCCGTTTGTATG-3′	5′-GCAATGGTCCTGATTGCAG-3′
Human	IGF1	5′-CTCTTCAGTTCGTGTGTGGAGAC-3′	5′-CAGCCTCCTTAGATCACAGCTC-3′
Human	IGF2	5′-TGGCATCGTTGAGGAGTGCTGT-3′	5′-ACGGGGTATCTGGGGAAGTTGT-3′
Human	VEGFA	5′-CGAGGTTCACCCTCAGTTTC-3′	5′-GCACGACTCCTTCTCCAAATA-3′
Human	VEGFB	5′-AAGGACAGTGCTGTGAAGCCAG-3′	5′-TGGAGTGGGATGGGTGATGTCA-3′
Human	TGFB1	5′-TACCTGAACCCGTGTTGCTCTC-3′	5′-GTTGCTGAGGTATCGCCAGGAA-3′
Human	TGFB2	5′-GATGACGACGACAACGATGA-3′	5′-GTCTGAACTAGTACCGCCTTT-3′
Human	TGFB3	5′-CTAAGCGGAATGAGCAGAGGATC-3′	5′-TCTCAACAGCCACTCACGCACA-3′
Human	GAPDH	5′-ACAGTTGCCATGTAGACC-3′	5′-TTGAGCACAGGGTACTTTA-3′
Mouse	Srd5a2	5′-CGGTGTCTTCTTCTTTATTCTG-3′	5′-ATCCATTCAATAATCTCGCC-3′
Mouse	Wnt5a	5′-ATCTTGGGCACAGGGATTG-3′	5′-CAAATGTGGGCGTGATTGTG-3′
Mouse	Lef1	5′-ACTGTCAGGCGACACTTCCATG-3′	5′-GTGCTCCTGTTTGACCTGAGGT-3′
Mouse	Tcf7l1	5′-GACCTCGGAGCGAACGA-3′	5′-TTGACCAGGGACGACTTGA-3′
Mouse	Xrcc4	5′-GACATGGCTATGGAGAAAGGAA-3′	5′-GAGAAATGCCGAGACTCCTTAG-3′
Mouse	Zfp143	5′-GCGCCCATCACCCTATTT-3′	5′-ACGAGTGAATTATGCCGATTCT-3′
Mouse	Zmiz1	5′-GCAGAGCAAGCGAGCAG-3′	5′-GCGAGAGCGCGGAAATC-3′
Mouse	Ptprk	5′-CTGCCTACCATGAAGGAGAATG-3′	5′-GGATATTCAAAGTGCCAGGGT-3′
Mouse	Ryk	5′-ACGACCCAGTATCTGAGAGCTG-3′	5′-GGCTTTGGCTTCCAGAAGAGTG-3′
Mouse	Ror2	5′-AACGTACGCGTGGAACTGTGTG-3′	5′-CAGACGAGGAAGAACAGGCAAG-3′
Mouse	Lrp6	5′-CTGAATGCTGACAACAGGACCTG-3′	5′-GACGTTCCGAAGGCTGTGGATA-3′
Mouse	Tgfb1	5′-CCTGTGAGACAGATGGTCTTTG-3′	5′-TGGCCTGTCTCGAGGAATTA-3′
Mouse	Tgfb2	5′-CAGATCCTGAGCAAGCTGAA-3′	5′-GTCCCTGGTACTGTTGTAGATG-3′
Mouse	Vegfa	5′-CGGGATTGCACGGAAACT-3′	5′-CAGCTCCGATCGGTTTGTC-3′
Mouse	Gapdh	5′-GACAACTTTGGCATTGTGGAAG-3′	5′-TGCTTCACCACCTTCTTGATGT-3′

The WNT signaling pathway RT² PCR Microarray was performed using Human WNT signaling pathway RT² profiler PCR array (330231 PAHS-243ZA; Qiagen, Hilden, Germany) and corresponding reverse transcription kit (Qiagen; 330401) and RT2 SYBR green qPCR master mix (Qiagen; 330500) following the manufacturer's protocol.

ELISA

WNT5A and insulin-like growth factor 1 (IGF1) levels in cell and culturing medium were quantified using human WNT5A ELISA kit (ANTIA77507-96; VWR, Radnor, PA) or IGF1 ELISA kit (SigmaAldrich; RAB0228) following the manufacturer's protocol. The absolute concentration was calculated by plotting the mean absorbance values of the unknowns onto the standard curve in GraphPad Prism version 10.0.0 (Dotmatics, Boston, MA), and the cellular concentration was calculated by dividing the value of absolute concentration over the protein concentration of corresponding cell lysate samples.

The DHT levels in the BIDMC biorepository serum samples and frozen surgical prostate tissues were measured with human DHT ELISA kit (LS-F10533-1; LS Bio, Newark, CA) following the manufacturer's standard protocol, as previously described.¹⁴

Single-Cell RNA Sequencing

Single-cell RNA sequencing (scRNA-seq) and data analysis were performed as previously described.¹⁴ Briefly, 23,000 single cells from fresh prostate tissues were loaded onto separate lanes of a single-cell A chip and generated the libraries submitted for sequencing on Illumina (San Diego, CA) NextSeq 500 and NovaSeq 6000 sequencers.

RNA Fluorescence in Situ Hybridization

RNA fluorescence in situ hybridization analysis of prostate tissue was performed as previously described.¹⁴ Briefly, three of each age-matched Srd5a2-null, heterozygous, and wild-type mice were sacrificed, and whole prostate tissues were fixed and sliced for staining conducted at the Neurobiology Imaging Facility, Harvard Medical School (Boston, MA). Image acquisition and quantitative analysis of Wnt5a signals follow the same protocol established in a previous publication.¹⁴

Statistical Analysis

All data included in this study were presented as means ± SD with at least three independent experiments. P ≤ 0.05 was considered significant. The statistical methods for each corresponding experiment are described in the figure legends.

Data and Resource Availability

The scRNA-seq data are available in the National Center for Biotechnology Information's Gene Expression Omnibus (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE235514; accession number GSE235514) under previous publication.¹⁴ The bulk RNA-seq data are available in National Center for Biotechnology Information's Gene Expression Omnibus (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE304244; accession number GSE304244).

Results

The Expression of SRD5A2 Is Correlated with Prostate Volume

Previous work demonstrated an inverse correlation between SRD5A2 expression and the clinical efficacy of 5ARIs in treating BPH.¹⁴ To address if baseline SRD5A2 expression in prostatic tissue is associated with prostate volume, 61 prostate biopsy samples from the MTOPS study cohort were first examined (Table 3),⁶ and a statistically significant positive correlation was observed between SRD5A2 expression and total prostate volume (Figure 1, A and B). Notably, SRD5A2 expression was also significantly correlated with the transition zone volume, but not with the peripheral zone volume (Figure 1, C and D), which aligns with the earlier report.²⁴ Interestingly, despite SRD5A2's role in DHT synthesis, there was no significant correlation between clinically recorded DHT levels and SRD5A2 expression in the prostate (Figure 1E).

Table 3.

Clinical Demographic Information of the MTOPS Cohort (n = 61)

Variable	Mean value (range)
Age, years	62.95 (50 to 75)
BMI, kg/m2	27.80 (18.8 to 35.8)
Total prostate volume, cc	60.62 (9.3 to 150.7)
Transition zone volume, cc	28.81 (3.8 to 114)
Peripheral zone volume, cc	31.82 (4.5 to 112)
Blood DHT level, ng/dL	43.09 (95 to 108)

BMI, body mass index; cc, cubic centimeter; DHT, dihydrotestosterone; MTOPS, medical therapy of prostatic symptoms.

Figure 1.

Prostate size is correlated with SRD5A2 expression in clinical samples. A: Representative image of high versus low SRD5A2 expression in the Medical Therapy of Prostatic Symptoms (MTOPS) cohort. B–E: Correlation plot of SRD5A2 protein expression with total prostate volume (B), transition zone volume (C), peripheral zone volume (D), and serum dihydrotestosterone (DHT) level (E) in the MTOPS cohort. F: Representative image of high versus low SRD5A2 expression in the Beth Israel Deaconess Medical Center biorepository cohort. G–K: Correlation plot of SRD5A2 protein expression with total prostate volume (G), transition zone volume (H), peripheral zone volume (I), serum DHT level (J), and tissue DHT level (K) in the BIDMC biorepository cohort. SRD5A2 expression was measured by immuoreactive score quantified from immunohistochemistry. Serum and tissue DHT levels of BIDMC biorepository samples were measured by enzyme-linked immunosorbent assay kit. Correlation analysis was performed using Pearson correlation. n = 61 (B–E); n = 35 (G); n = 20 (H and I); n = 23 (J); n = 43 (K). Scale bars: 50 μm (A, left panels); 20 μm (A, right panels, and F). Original magnifications: ×10 (A, left panels); ×20 (A, right panels, and F).

To validate these findings, an independent cohort of 49 BPH prostate specimens from the BIDMC institutional biorepository was analyzed (Table 4). Consistent with the MTOPS data, SRD5A2 expression in these samples was strongly correlated with total prostate volume and transition zone volume (Figure 1, F–H). Peripheral zone volume was also correlated with SRD5A2 expression in this cohort (Figure 1I). However, neither serum nor tissue DHT levels showed significant association with SRD5A2 expression (Figure 1, J and K). Together, these results demonstrated that SRD5A2 protein expression is positively associated with prostate volume, particularly in the transition zone.

Table 4.

Clinical Demographic Information of the Beth Israel Deaconess Medical Center Biorepository Cohort

Variable	Mean value (range)
Age, years (n = 49)	70.7 (52 to 85)
BMI, kg/m2 (n = 37)	27.6 (20.7 to 36.3)
Total prostate volume, cc (n = 35)	141.6 (16 to 330)
Transition zone volume, cc (n = 20)	89.6 (19.5 to 247.9)
Peripheral zone volume, cc (n = 20)	44.7 (25.4 to 82.1)
DHT in serum, pg/mg (n = 23)	17.1 (3.3 to 43.5)
DHT in prostate tissue, ng/mg (n = 43)	403.8 (210.9 to 1047.6)

BMI, body mass index; DHT, dihydrotestosterone.

SRD5A2 Knockout Alters Stromal-Epithelial Crosstalk and Activates Muscle-Related Signaling Pathways in the Mouse Prostate

To understand the molecular mechanism of SRD5A2 expression in prostate growth, Srd5a2-null (Srd5a2^–/–) mice were generated,¹⁴ and bulk RNA sequencing was performed on wild-type (Srd5a2^+/+) controls, heterozygous (Srd5a2^+/–) controls, and Srd5a2^–/– mouse prostate tissues to identify putative pathways regulated by SRD5A2. Hierarchical clustering effectively distinguished the three groups based on their differential gene expression profiles (Figure 2A and Supplemental Figure S1). Gene Ontology analysis identified 133 significantly up-regulated and 2 down-regulated genes in Srd5a2^–/– mice (Supplemental Figure S2). Notably, up-regulated genes were highly enriched in biological processes related to the muscle system. Key enriched terms included those related to muscle system processes, contractile fibers, and striated muscle development, highlighting a shift toward muscle-related cellular architecture and function (Figure 2A).

Figure 2.

Gene and cell type profiling of Srd5a2^–/– mouse prostate. A: Dot plot showing the Gene Ontology (GO) enrichment analysis conducted on prostate tissue from 8-week–old mice compared with three control and three Srd5a2^–/– samples. The x axis represents the gene ratio (the proportion of genes associated with each GO term relative to the total gene set analyzed), and the y axis lists the GO terms grouped by their biological processes, molecular functions, or cellular components. The size of the dots indicates the number of genes (count) involved in each term, and the color gradient represents the adjusted P value (Padj). Color represents the category of enrichment: blue, molecular function; green, cellular component; and red, biological process. B: The bar graphs showing the number of outgoing (secreted/paracrine/autocrine) signals and the number of incoming (ligand binding/receptor activation) signals among the cell clusters. C: The CellPhone-DB analysis of the ligand-receptor interactions in different mouse prostate cell populations. Each dot represents the interaction under a specific condition, with the size indicating the statistical significance [−log₁₀ (P value)] and the color representing the log₂ mean expression level of the molecules. P values are derived from one-sided permutation tests (significant P < 0.05). Larger and darker-colored dots signify stronger and more significant interactions. Boxed dots highlight the differential expression in Srd5a2^–/– luminal epithelial cell 1 (LE1) and LE2 compared with other luminal subtypes. D: The t-distributed stochastic neighbor embedding (t-SNE) dimensionality reduction plot showing the expression localization of Wnt5a in different cell types of mouse prostate tissue. The expression levels are shown in color gradient within cells. n = 9 (C). BE, basic endothelial; Endo, endothelial; Fib, fibroblast; IGF, insulin-like growth factor; Mac, macrophage; myoFib, myofibroblast; SM, smooth muscle; SR, sarcoplasmic reticulum; SV, seminal vesicle.

To further delineate how Srd5a2 deletion affects cell-cell communication, scRNA-seq was performed on mouse prostate tissues. It is well established that paracrine signaling between stroma and epithelium governs prostate development and regeneration.¹⁰ Therefore, the interactions between luminal epithelial cells (LEs) and stromal cell populations were focused on. With CellPhoneDB2 mapping the intercellular communication networks across genotypes,²⁵ stromal cells emerged as major signal senders, and particularly, myofibroblasts showed the highest outbound signaling in both Srd5a2^–/– mice and Srd5a2^+/– controls. Notably, the total weighted ligand activity of fibroblasts increased substantially in Srd5a2^–/– animals (from 366 to 493), suggesting enhanced fibroblast-mediated signaling on Srd5a2 deletion (Figure 2B).

On the receiving end, Srd5a2 knockout led to a striking reduction in receptor expression in LE1 epithelial cells (from 205 to 96), whereas LE2 cells maintained a stable number of receptors (Figure 2B). When examining stromal-derived ligands that may influence LE1 and LE2 behaviors, nine significant pairs with potential regulatory roles were identified (Figure 2C). Among these interactions, WNT5A^–/^– protein tyrosine phosphatase receptor type K (PTPRK) stood out as the candidate with differential activity across luminal subtypes (Figure 2C). Pathway analysis further confirmed activation of WNT signaling in LE2 cells on Srd5a2 loss.¹⁴ As a particular interest, WNT5A is a noncanonical WNT ligand that is known to modulate prostate growth and epithelial behavior.^10,25, 26, 27, 28, 29 Wnt5a was predominantly expressed in stromal compartments, particularly myofibroblast and fibroblast populations (Figure 2D).¹⁰ In addition, several transcription factors, including Tcf7l1, a downstream transcription factor of the WNT canonical pathway, were also observed to be enriched in LE2 cells and decreased in knockout mice compared with the wild type confirmed by qPCR (Supplemental Figure S3 and Supplemental Figure S4A). These spatially restricted expression patterns suggested a targeted influence on nearby luminal cells.

Spatial Expression and Regulation of WNT5A in the Mouse and Human Prostate Tissue

Next, RNA fluorescence in situ hybridization was used to assess Wnt5a spatial expression. Representative images revealed that Wnt5a transcripts in red puncta were predominantly localized in the stromal compartment, with minimal signal observed in epithelial cells (Figure 3A). There was a significant reduction of Wnt5a expression in the periurethral region of Srd5a2^−/− prostates compared with both Srd5a2^+/+ and Srd5a2^+/– controls (Figure 3B). Although the dorsal and lateral lobes exhibited no statistically significant differences between genotypes (Figure 3C), anterior lobes (AP) showed significant increase of Wnt5a signal in Srd5a2^−/− prostates (Figure 3D). Dissecting mouse prostate morphology and comparison with human prostate have confirmed that the AP of mouse prostate is more similar to the transition zone of human prostate, where BPH mostly occurs and enriches epithelial cells, whereas the periurethral region and dorsal and lateral lobes are higher in stromal cell populations compared with AP.30, 31, 32 Thus, the spatial expression of Wnt5a between AP and non-AP was differentiated, corresponding to epithelium and stroma regions of mouse prostate. Srd5a2 knockout down-regulated the mRNA level of Wnt5a in whole prostate (Figure 3, G and H), and the difference in spatial expression of Wnt5a was further confirmed by qPCR in AP and non-AP lobes (Figure 3, E and F). Other stromal-derived ligands, such as transforming growth factor beta 2 (Tgfb2) and vascular endothelial growth factor A (Vegfa), were also observed to be decreased on Srd5a2 knockout (Supplemental Figure S4B). More interestingly, both canonical and noncanonical Wnt receptors, such as receptor like tyrosine kinase (Ryk), receptor tyrosine kinase like orphan receptor 2 (Ror2), and LDL receptor related protein 6 (Lrp6), were decreased in Srd5a2^−/− prostates, yet the mRNA level of Ptprk, the pair of Wnt5a ligand found in CellPhoneDB2, remained unchanged (Supplemental Figure S4C). These results suggested that Wnt5a expression was dependent on Srd5a2 function, and exhibited regionally distinct regulation across different prostate compartments, particularly in the periurethral (proximal) region.^10,31

Figure 3.

SRD5A2 level positively correlates with the expression of WNT5A in mouse and human prostate. A: The representative RNA fluorescence in situ hybridization (FISH) images showing Wnt5a (white), luminal epithelial cell 2–specific marker estrogen receptor 1 (Esr1; red), and plasmacytoma variant translocation 1 (Pvt1; green) expression in the periurethral, dorsal lateral, and anterior lobes of control and Srd5a2^–/– mice. B–D: Quantification of Wnt5a RNA-FISH signal per cell in each lobe shown in A. E and F: Real-time quantitative PCR (qPCR) analysis of Wnt5a in nonanterior and anterior prostate tissue from Srd5a2^+/+ and Srd5a2^–/– mice. G and H: qPCR analysis of Srd5a2 and Wnt5a in total prostate tissue from Srd5a2^+/+, Srd5a2^+/–, and Srd5a2^–/– mice. I: Western blot analysis of SRD5A2, WNT5A, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) loading control in eight institutional biorepository patient prostate samples. The number labeling of samples is retrieved from the biorepository patient record. J–M: The correlation plot of SRD5A2 mRNA expression with WNT5A (J), protein tyrosine phosphatase receptor type K (PTPRK; K), transforming growth factor beta 2 (TGFB2; L), and transforming growth factor beta receptor 1 (TGFBR1; M) mRNA expression level in 18 patient samples. All experiments were replicated at least three times. Significance in all quantifications was calculated using two-tailed t-tests. Data on dot plots are presented as means ± SD (B–H and J–M). N ≥ 3 (B–H and J–M). ∗P ≤ 0.05, ∗∗P ≤ 0.01, ∗∗∗P ≤ 0.001, and ∗∗∗∗P ≤ 0.0001 versus Srd5a2^+/+; ^##P ≤ 0.01, ^####P ≤ 0.0001 versus Srd5a2^+/–. Scale bar = 5 μm (A). Original magnification, ×100 (A). NS, nonsignificant.

Similar with that in mouse prostate tissue, the expression of SRD5A2 corresponded with WNT5A in human prostate both in protein and mRNA levels (Figure 3, I and J). Additionally, SRD5A2 expression was also significantly correlated with the PTPRK (Figure 3K). In contrast, no significant correlation was observed between SRD5A2 and TGFBR1 or TGFB2 (Figure 3, L and M); both were among the nine significant pairs with potential regulatory roles and associated with stromal-epithelial interactions (Figure 2C). Collectively, these data demonstrated that there was a regulatory relationship between SRD5A2 and WNT5A in both the human and the mouse prostate.

SRD5A2 Enhances Prostate Stromal Cell Growth and Survival through Activation of WNT5A Pathways

To investigate the regulatory effect of SRD5A2 on WNT5A expression, a SRD5A2-overexpressing immortalized normal prostate cell line (BHPrS1) was used to assess WNT5A expression at both the mRNA and the protein levels. qPCR and Western blot analyses confirmed robust overexpression of SRD5A2 compared with vector control (Figure 4, A and B). Notably, SRD5A2 overexpression resulted in a substantial increase in both WNT5A transcript and protein levels within stromal cells (Figure 4, A and B, and Supplemental Figure S5A). Furthermore, ELISA quantification demonstrated significantly elevated WNT5A concentrations both intracellularly and in the culture supernatant (Figure 4, C and D), indicating enhanced secretion.

Figure 4.

SRD5A2 regulates prostate stromal cell growth via WNT5A. A: Real-time quantitative PCR (qPCR) analysis of SRD5A2, WNT5A in vector control (Vec), and SRD5A2-overexpressing (OE) BHPrS1 cells. B: Western blot analysis of SRD5A2, WNT5A, and actin loading control in BHPrS1 cells. C and D: The level of WNT5A in BHPrS1 cell lysate and culturing medium was measured by enzyme-linked immunosorbent assay (ELISA). E: Representative immunocytochemistry images of BHPrS1 cells showing the expression level of Ki-67 and cleaved caspase 3 (CC-3). Ki-67 and CC-3 protein (green) are detectable in the whole cell, phalloidin-labeled actin filaments (red) delineate cell structure, and DAPI (blue) marks nuclear localization. F and G: Quantification of Ki-67 and CC-3 signals in E. H: qPCR analysis of proliferation and apoptosis-related markers in BHPrS1 cells. I: Western blot analysis of SRD5A2, WNT5A, β-catenin, poly (ADP-ribose) polymerase 1 (PARP1), cleaved PARP1 (arrow), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) loading control in BHPrS1 cells treated with dimethyl sulfoxide (DMSO) or 100 μmol/L finasteride (Fin) for 72 hours. J and K: The level of WNT5A in DMSO or finasteride-treated BHPrS1 cell lysate and culturing medium was measured by ELISA. L: Cell viability of DMSO or Fin-treated BHPrS1 cells measured by MTT assay. M: Quantification of the BHPrS1 spheroid-like colony size from day 4 to day 18. N and O: Quantification of BHPrS1 spheroid-like colony size change from day 12 to day 15 and from day 12 to day 18 as BHPrS1 OE cells are treated with 100 μmol/L finasteride and/or 100 μmol/L WNT5A antagonist BOX5 for 72 or 144 hours. All experiments were replicated at least three times. The t-test and one-way analysis of variance with Tukey test were performed. Data on dot plots are presented as means ± SD (A, C, D, F–H, and J–O). N ≥ 3 (A, C, D, F–H, and J–O). ∗P ≤ 0.05, ∗∗P ≤ 0.01, ∗∗∗P ≤ 0.001, and ∗∗∗∗P ≤ 0.0001 versus untreated Vec; ^#P ≤ 0.05, ^##P ≤ 0.01, and ^####P ≤ 0.0001 versus untreated OE; ^††††P ≤ 0.0001 versus finasteride-treated Vec. Scale bar = 25 μm (E). Original magnification, ×40 (E). CASP, caspase; CCND1, cycline D1; CYCS, cytochrome C; MYC, myelocytomatosis oncogene; NS, nonsignificant; PARP1, poly(ADP-ribose) polymerase 1.

To analyze the effect of SRD5A2 on prostate cell growth, immunofluorescence staining was performed for the proliferation marker Ki-67 and the apoptosis marker cleaved caspase-3. The data revealed a significant increase in Ki-67 and decrease in cleaved caspase-3 staining in SRD5A2 overexpression cells compared with the vector control (Figure 4, E–G). Meanwhile, SRD5A2 overexpression greatly increased the expression of cleaved poly (ADP-ribose) polymerase 1 (PARP1) (Figure 4H), an essential regulator of cell apoptosis and downstream of WNT canonical signaling, particularly TCF/LEF transcription factor family.^33,34

Meanwhile, pharmacologic inhibition of SRD5A2 was simultaneously investigated to see if it impacts on WNT5A and its downstream protein expression. Compared with the dimethyl sulfoxide vehicle, SRD5A2 inhibitor finasteride suppressed the up-regulation of WNT5A, β-catenin, and PARP1 in SRD5A2-overexpressed stromal cells (Figure 4I). ELISA quantification also revealed significantly elevated WNT5A protein levels in both stromal cells and culture medium following SRD5A2 overexpression, and this was markedly reduced with finasteride treatment (Figure 4, J and K). From the perspective of noncanonical WNT signaling, overexpression of SRD5A2 increased calcium level, which was also abrogated with finasteride treatment (Supplemental Figure S5C). In addition, finasteride significantly reduced the viability of prostate stromal cells with normal levels of SRD5A2, whereas SRD5A2-overexpressing cells maintained significantly higher viability and showed partial resistance to finasteride-induced cytotoxicity (Figure 4L), suggesting that SRD5A2 overexpression confers a prosurvival advantage in the presence of 5ARI treatment.

To evaluate the long-term impact of SRD5A2 overexpression on prostate stromal cell proliferation in three-dimensional culture, spheroid-like colony formation assay was performed on BHPrS1 cells. At day 4, no significant difference was observed between vector control and overexpression cells. However, from day 8 onward, SRD5A2-overexpressing cells formed significantly larger spheroid colonies than control cells, with exponential growth observed through day 18 (Figure 4M). Furthermore, treatment with BOX5, finasteride, or a combination of the two between day 12 and day 15 showed significantly reduced colony expansion (Figure 4N), and this inhibitory effect was further pronounced comparing the change between day 18 and day 12 (Figure 4O), indicating sensitivity to both androgen metabolism and WNT5A signaling disruption. These findings highlighted that SRD5A2 promoted stromal spheroid growth through a WNT5A-dependent mechanism, and that targeting either 5 α-reductase or WNT5A signaling can significantly impair the expansion of SRD5A2-driven stromal colonies.

LEF1 Mediates the SRD5A2-Induced WNT Canonical Signaling Pathway

To explore the molecular components of WNT signaling pathway that are regulated by SRD5A2, a targeted PCR microarray analysis following SRD5A2 overexpression was performed. Among the 84 differentially expressed genes, LEF1, WNT5A, IGF2, and others were significantly up-regulated (Figure 5A). qPCR further confirmed that IGF1, IGF2, along with WNT5A receptors RYK, ROR1, ROR2, and LRP6 were greatly up-regulated (Supplemental Figure S5, B and C).

Figure 5.

SRD5A2 up-regulates the expression of LEF1 and IGF1. A: Volcano plot showing differentially expressed WNT-related genes in BHPrS1-overexpressing (OE) versus vector (Vec) cells with log₁₀ (fold change) indicating mean gene expression levels. Each dot represents a single gene. B: T-distributed stochastic neighbor embedding (t-SNE) dimensionality reduction plot showing the expression localization of Lef1 in different cell types of mouse prostate tissue and its expression levels within cells. C: Dot plot showing the expression of Lef1 and Wnt5a in different cell types from single-cell RNA sequencing of mouse prostate tissue. The size of the dots represents the proportion of positively expressing cells, and the color indicates the average expression levels. D: Real-time quantitative PCR (qPCR) analysis of T-cell factor (TCF)/LEF family transcription factors in BHPrS1 cells. E: Representative immunocytochemistry (ICC) images of BHPrS1 cells showing the expression level of LEF1. LEF1 protein (green) is detectable in the nuclei, phalloidin-labeled actin filaments (red) delineate cell structure, and DAPI (blue) marks nuclear localization. F: Quantification of LEF1 signals in E. G: qPCR analysis of LEF1, poly (ADP-ribose) polymerase 1 (PARP1), IGF1, and IGF2 in BHPrS1 Vec and OE cells transfected with scramble (Scr) or LEF1 (siLEF1) siRNAs. H: Representative ICC images showing the expression level of LEF1 in siRNA-transfected BHPrS1 cells. I: Quantification of LEF1 signals in H. J: Western blot analysis of SRD5A2, WNT5A, β-catenin, PARP1, cleaved PARP1 (arrow), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) loading control in siRNA-transfected BHPrS1 cells. K and L: The cellular level of IGF1 in BHPrS1 cell transfected with siRNAs or treated with 100 μmol/L finasteride and/or 100 μmol/L WNT5A antagonist BOX5 for 72 hours, measured by human IGF1 enzyme-linked immunosorbent assay kit. M and N: Cell viability of siRNA-transfected or 72-hour picropodophyllin (PPP)–treated BHPrS1 cells measured by MTT assay. All experiments were replicated at least three times. The t-test and one-way analysis of variance with Tukey test were performed. Data on dot plots are presented as means ± SD (D, F, G, I, and K–N). N ≥ 3.(D, F, G, I, and K–N) ∗P ≤ 0.05, ∗∗P ≤ 0.01, ∗∗∗P ≤ 0.001, and ∗∗∗∗P ≤ 0.0001 versus untreated Vec or Vec Scr; ^#P ≤ 0.05, ^###P ≤ 0.001, and ^####P ≤ 0.0001 versus untreated OE or OE Scr; ^†P ≤ 0.05, ^††††P ≤ 0.0001 versus treated Vec or Vec siLEF1. Scale bars: 25 μm (E); 50 μm (H). Original magnifications: ×40 (E); ×60 (H). BE, basic endothelial; Endo, endothelial; Fib, fibroblast; LE, luminal epithelial cell; Mac, macrophage; myoFib, myofibroblast; NS, nonsignificant; SM, smooth muscle; SV, seminal vesicle.

The Lef1 expression within the mouse prostate was further validated by re-analyzing the scRNA-seq data set. Lef1 signal was shown as primary enrichment in stromal populations, particularly in myofibroblasts and fibroblasts (Figure 5B). Dot plot analysis further confirmed these cell type–specific expression patterns for both Lef1 and Wnt5a (Figure 5C), suggesting potential functional interaction or coregulation between Lef1 and Wnt5a in these cell types. Notably, as a key transcriptional mediator of canonical WNT/β-catenin signaling,³⁵ LEF1 demonstrated one of the highest fold changes and statistical significance levels at both mRNA and protein level, suggesting strong activation of WNT transcriptional programs downstream of SRD5A2 (Figure 5, A and D–F). The activation of canonical WNT signaling was consistent with observations on SRD5A2 down-regulated β-catenin degradation. The protein level of β-catenin was increased on SRD5A2 overexpression and reduced by finasteride (Figure 4I). Interestingly, other downstream transcriptional regulators of canonical WNT signaling besides LEF1, specifically TCF7, TCF7L1, and TCF7L2, showed no significant change at the mRNA level (Figure 5D). These findings suggested activation of the canonical WNT signaling pathway by SRD5A2 was specific to the transcription factor LEF1 rather than the β-catenin–bound TCF/LEF family in general.

To determine whether LEF1 mediates the downstream effects of SRD5A2 overexpression, qPCR, immunofluorescence, and Western blot analysis were subsequently performed on transfected BHPrS1 stromal cells with either scrambled or LEF1-targeting siRNA. LEF1 was markedly reduced after LEF1-targeting siRNA transfection in SRD5A2 overexpression cells, confirming efficient knockdown at the mRNA and protein level (Figure 5, G–I). Western blot analysis results showed that SRD5A2 and WNT5A protein levels remained high in SRD5A2 overexpression group and were unaffected by LEF1-targeting siRNA (Figure 5J), and silencing LEF1 reduced SRD5A2-induced up-regulation of PARP1 expression at both mRNA and protein levels in both vector and SRD5A2-overexpression groups compared with corresponding cells transfected with scramble siRNA control (Figure 5, G and J). LEF1 knockdown significantly reduced cell viability in both cells, although to a lesser extent in SRD5A2 overexpression cells, indicating that LEF1 partially mediates SRD5A2-induced resistance to cell death (Figure 5M).

LEF1 Regulates IGF1 Expression in Prostate Stromal Cells through a WNT5A-Dependent Pathway

Given that the IGF1 promoter contains a WNT response element and that IGF1 has been reported as a direct WNT target gene,³⁶ it was next investigated whether IGF1 expression is regulated downstream of SRD5A2 via the WNT5A-LEF1 axis. Both mRNA and protein levels of IGF1 were significantly elevated in SRD5A2-overexpressing stromal cells, and markedly reduced following LEF1 knockdown (Figure 5, G and K). Furthermore, pharmacologic inhibition using either finasteride or BOX5, a WNT5A antagonist, significantly suppressed IGF1 expression (Figure 5L), supporting the notion that SRD5A2 enhances IGF1 transcription in a WNT5A/LEF1-dependent manner.

Functionally, treatment with picropodophyllin, a selective IGF1 receptor inhibitor, led to a significant reduction in cell viability in both vector control and SRD5A2-overexpressing cells (Figure 5N). Together, these findings demonstrated that LEF1 and IGF1 signaling were essential for SRD5A2-mediated survival in prostate stromal cells, and that targeting these pathways may overcome stromal cell–mediated resistance mechanisms.

SRD5A2-Overexpressing Stromal Cells Promote Epithelial Proliferation in an IGF1 Paracrine Manner

To evaluate the SRD5A2-induced ligand-receptor interaction change between stromal cells and adjacent epithelial cells, a co-culture system of BPH1 prostate epithelial cells with BHPrS1 stromal cells was used and analyzed epithelial expression profiles by qPCR and Western blot analysis. WNT receptors were not changed in BPH1 cells co-cultured with SRD5A2 overexpression stromal cells, and WNT5A and LEF1 were slightly reduced compared with BPH1 cells co-cultured with vector-expressing stromal cell (Figure 6A and Supplemental Figure S5, D and E). Notably, SRD5A2-overexpressing stromal cells induced a strikingly increase in IGF1 mRNA level in BPH1 cells (Figure 6A). The change of IGF1 expression at protein level was further confirmed by ELISA (Figure 6B). The results suggested that SRD5A2-overexpressing stromal cells might modulate epithelial growth through enhancing IGF1 receptor activation.

Figure 6.

SRD5A2-overexpression stromal cells increase epithelial cell IGF1 expression and cell proliferation. A: Real-time quantitative PCR analysis of WNT5A, β-catenin, LEF1, cyclin D1, poly (ADP-ribose) polymerase 1 (PARP1), and WNT-related growth factors in benign prostatic hyperplasia 1 (BPH1) cells co-cultured with BPH1 cells, BHPrS1 control cells, or BHPrS1 SRD5A2-overexpressing (OE) cells. B: The cellular level of IGF1 in BPH1 cells co-cultured with BPH1 cells, BHPrS1 control cells, or BHPrS1 SRD5A2-overexpressing cells measured by human IGF1 enzyme-linked immunosorbent assay (ELISA) kit. C: Cell viability of BPH1 cells cultured in control or drug-treated BHPrS1-conditioned medium (C.M.) for 72 hours measured by MTT assay. D: Representative images taken on day 21 of BPH1 soft agar colonies in control or drug-treated BHPrS1-conditioned medium starting on day 1. Red dashed boxed areas (top panels) indicate the regions shown at higher magnification (bottom panels). E and F: Quantification of the colony numbers (E) and colony size (F) of BPH1 cells shown in D. G: The cellular level of IGF1 in BHPrE1 cells co-cultured with BHPrE1 cells, BHPrS1 control cells, or SRD5A2-overexpressing BHPrS1 cells for 72 hours measured by human IGF1 ELISA kit. H: Representative images taken on day 8 and day 14 of BHPrE1 organoid co-cultured with BHPrS1 cells. Red dashed boxed areas (top panels) indicate the regions shown at higher magnification (bottom panels). I and J: Quantification of the size of BHPrE1 organoids on day 8 (I) and day 14 (J) shown in H. All experiments were replicated at least three times. The t-test and one-way analysis of variance with Tukey test were performed to calculate the significance. Data on dot plots are presented as means ± SD (A–C, E–G, I, and J). N ≥ 3 (A–C, E–G, I, and J). ∗P ≤ 0.05, ∗∗P ≤ 0.01, ∗∗∗P ≤ 0.001, and ∗∗∗∗P ≤ 0.0001 versus monoculture (Mono); ^#P ≤ 0.05, ^##P ≤ 0.01, ^###P ≤ 0.001, and ^####P ≤ 0.0001 versus untreated vector (Vec); ^††P ≤ 0.01, ^††††P ≤ 0.0001 versus untreated OE. Scale bars: 500 μm (D and H, top panels); 100 μm (D and H, bottom panels). Original magnifications: ×4 (D and H, top panels); ×10 (D and H, bottom panels). CCND1, cycline D1; CTNNB1, beta-catenin; DMSO, dimethyl sulfoxide; Fin, finasteride; NC, negative control; PPP, picropodophyllin.

To further assess the effect of SRD5A2-overexpressing stromal cells on epithelial proliferation, BPH cell line BPH1 was treated with conditioned media (CM) from both vector control and SRD5A2-overexpressing BHPrS1 cells and analyzed for cell viability and spheroid growth. BPH1 cell viability increased with CM from both groups, but CM from SRD5A2-overexpressing stromal cells induced a significantly greater increase (Figure 6C). Treatment with either finasteride or picropodophyllin significantly reduced the CM's viability-promoting effect on BPH1 cells, with the strongest inhibition observed on the combined treatment (Figure 6C). CM from LEF1 siRNA transfected stromal cells had much lower to almost no epithelial viability boosting effect compared with CM from the scramble control cells, further validating that the paracrine effect of stromal cells on epithelial viability was LEF1 dependent (Supplemental Figure S5F). Additionally, soft agar assays showed that CM significantly increased both the number and size of BPH1 colonies compared with regular medium (Figure 6, D–F). CM from SRD5A2-overexpressing stromal cells induced the greatest colony growth, which was significantly reduced by finasteride (Figure 6, E and F). These findings suggested that stromal SRD5A2 promoted epithelial expansion via soluble factors, such as IGF1, in a finasteride-sensitive manner.

To validate the functional impact of stromal SRD5A2 on epithelial morphogenesis, organoid-forming BHPrE1 cells were co-cultured with BHPrS1 stromal cells in a three-dimensional matrix. Consistent with BPH1 results, co-culture with SRD5A2-overexpressing stromal cells increased IGF1 expression (Figure 6G), and led to the formation of larger, more complex organoids by day 8 (Figure 6, H and I), with even more pronounced size difference by day 14 (Figure 6, H and J). These results supported a role for stromal SRD5A2 in promoting epithelial proliferation and prostate tissue remodeling through paracrine signaling.

WNT5A Expression Correlates with SRD5A2 and Total Prostate Volume in Human Prostate Tissues

The clinical relevance of WNT5A in the human prostate was evaluated by analyzing biopsy samples from the MTOPS clinical trial cohort. Immunohistochemistry staining showed stromal-enriched expression of WNT5A with distinct intensity differences across samples (Figure 7A). Quantitative analysis revealed a strong positive correlation between WNT5A and SRD5A2 expression (P < 0.001, R = 0.62) (Figure 7B), suggesting potential coregulation of these genes within the prostate stroma. To assess whether WNT5A expression was associated with total prostate volume, a marginal significant correlation was observed (P = 0.02, R = 0.39) (Figure 7C), indicating that WNT5A levels may not directly reflect prostate size.

Figure 7.

Prostate size and SRD5A2 expression level are correlated with WNT5A expression in clinical samples. A: Representative image of high versus low WNT5A expression in the Beth Israel Deaconess Medical Center biorepository cohort. B and C: Correlation plot of WNT5A expression with SRD5A2 expression (B) and total prostate volume (C) in the BIDMC biorepository cohort. D: Representative image of high versus low WNT5A expression in the Medical Therapy of Prostatic Symptoms (MTOPS) cohort. E and F: Correlation plot of WNT5A expression with SRD5A2 expression (E) and total prostate volume (F) in the MTOPS cohort. WNT5A and SRD5A2 expressions were measured by immunoreactive score quantified from immunohistochemistry. Correlation analysis was performed using Pearson correlation. n = 41 (B); n = 37 (C); n = 25 (E and F). Scale bars = 20 μm (A and D). Original magnification, ×20 (A and D). cc, cubic centimeter.

To validate these findings, an independent biorepository cohort was also examined. Similar to the MTOPS cohort, WNT5A staining remained stromal specific (Figure 7D) and was strongly correlated with SRD5A2 expression (P < 0.001, R = 0.66) (Figure 7E). Again, there was no significant correlation between WNT5A expression and total prostate volume (P = 0.58, R = 0.10) (Figure 7F), reinforcing the conclusion that WNT5A is not a determinant of glandular size. Together, these results demonstrated that WNT5A and SRD5A2 were tightly co-expressed in the prostate stroma, but WNT5A levels did not correlate with prostate volume, suggesting its role was more likely related to tissue signaling than to structural enlargement.

Discussion

In this study, important new insights regarding the role of SRD5A2 in BPH are presented. A significant positive correlation between SRD5A2 protein expression and prostate volume was observed, notably within the transition zone (Figure 1), where most hyperplastic changes occur. This finding aligns well with known clinical features of BPH, where enlargement of the prostate transition zone primarily drives lower urinary tract symptoms characteristic of BPH.^37,38 Although SRD5A2 mediates the conversion of testosterone into DHT, a potent androgen responsible for prostate growth, a direct relationship between enzyme expression and prostate volume appears biologically plausible. These findings in two independent cohorts suggest that SRD5A2 expression might serve as an important biomarker for prostate growth and potentially predict responsiveness to 5ARI therapy.

Interestingly and somewhat unexpectedly, the analyses revealed no significant correlation between SRD5A2 and intraprostatic or serum DHT concentrations (Figure 1, E, J, and K). This is contrary to the traditional paradigm that directly links SRD5A2 expression to elevated local DHT levels within prostate tissue.^3,39 This result could reflect complex compensatory mechanisms within the prostate microenvironment, including alternate androgen metabolic pathways, such as the estrogen signaling pathway.^14,40 It is also possible that factors beyond enzymatic activity, such as enzyme kinetics, cellular localization, or variations in androgen receptor sensitivity and signaling pathways, might affect intraprostatic androgen levels.^41,42 Therefore, this study highlights SRD5A2 as a potential determinant of prostate size isolated from traditional androgen-dependent mechanisms. Further studies exploring these mechanisms are warranted to better elucidate how SRD5A2 contributes to prostate tissue growth independent of local androgen concentration.

The present work identifies a novel regulatory pathway involving SRD5A2, WNT5A, LEF1, and IGF1 in the stromal compartment of BPH (Supplemental Figure S6). SRD5A2 overexpression significantly increases the production and secretion of WNT5A, activating downstream signaling via LEF1, PARP1, and IGF1 to promote stromal cell proliferation and resistance to apoptosis (Figures 5 and 6). LEF1 is known to drive transcriptional programs critical for cell proliferation and survival.⁴³ scRNA-seq analysis localized LEF1 predominantly to prostate stromal cells, specifically fibroblasts and myofibroblasts. This stromal specificity corroborates previous studies suggesting stromal cells as essential mediators of prostate growth through paracrine signaling pathways.⁴⁴

The functional relevance of LEF1 was validated through siRNA-mediated knockdown, which significantly reduced SRD5A2-induced PARP1 and IGF1 expression and attenuated stromal cell viability (Figure 5). PARP1 is involved in repairing DNA damage and is a well-known therapeutic target for inducing apoptosis of cancer cells to reduce prostate cancer growth and invasion.³⁴ IGF1 is a well-characterized growth-promoting and anti-apoptotic factor involved in prostate biology and BPH pathology.^23,45,46 Notably, pharmacologic inhibition of IGF1 signaling via picropodophyllin, an IGF1 receptor inhibitor, significantly decreased viability of SRD5A2-overexpressing stromal cells, further underlining the critical role of IGF1 in stromal cell proliferation and survival mediated by SRD5A2-WNT5A-LEF1 signaling.

Importantly, the observed partial resistance to finasteride in SRD5A2-overexpressing stromal cells suggests that targeting the downstream signaling components, such as WNT5A or IGF1, might result in a synergic effect and offer additional therapeutic benefits. In conclusion, this study presents compelling evidence for an SRD5A2-WNT5A-LEF1-IGF1 signaling axis within prostate stromal cells that enhances cell proliferation and survival independently of classic androgen receptor signaling.

Although the present study suggests that WNT5A acts primarily as a signaling mediator rather than a direct driver of stromal or glandular growth, further validation of the AR-independent role of WNT5A and IGF1 and their importance in hyperplasia development is needed, especially from an in vivo perspective. As one of essential stromal-derived factors, WNT5A and IGF1 are known to profoundly influence epithelial biology through paracrine mechanisms in the prostate.^11,47, 48, 49 A recent study further reported correlations between androgen, IGF1, and WNT/β-catenin signaling, as well as a dynamic regulatory loop of IGF1-WNT initiated by the AR through prostatic oncogenesis development in human prostate cancer samples,⁵⁰ whereas other studies found IGF1 and WNT5a enhanced both androgen-dependent and androgen-independent prostatic development in both cancer and noncancerous cells.^11,51,52 Thus, the effects of paracrine WNT5A and IGF1 on organ size and underlying mechanism are complex and context dependent. Therefore, future in vivo studies are necessary to directly assess the impact of WNT5A and the WNT5A-LEF1-IGF1 signaling axis on prostate growth and development.

Currently, 5ARIs represent mainstay medical treatments for BPH, yet patient responses remain highly variable.^53,54 5ARIs also showed synergic effects with α-blockers or phosphodiesterase type 5 (PDE5) inhibitors that address the smooth muscle contraction, indicating a good potential for combination therapy with other drugs.55, 56, 57 Earlier work demonstrated an inverse correlation between SRD5A2 expression and clinical response to 5ARIs.¹⁴ The current findings extend this understanding, suggesting that SRD5A2 expression, independent of local DHT concentrations, may inform treatment decisions or the development of complementary therapeutic strategies aimed at other stromal-epithelial signaling pathways. In conclusion, the present findings highlight the critical stromal co-expression of WNT5A and SRD5A2 and uncover an essential paracrine role for stromal SRD5A2 in regulating prostate epithelial proliferation and morphogenesis via the IGF1 signaling pathway. These findings highlight a critical stromal-epithelial axis in prostate tissue remodeling and provide new insights for more precise therapeutic interventions in BPH management.

Declaration of Generative AI and AI–Assisted Technologies in the Writing Process

During the preparation of this work, the authors used GPT-4o mini (OpenAI, San Francisco, CA) to improve language and readability. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

Disclosure Statement

None declared.

Supplemental Data

Supplemental Figure S1

Hierarchical clustering and differential gene expression analysis in Srd5a2 genotypes. To evaluate the molecular changes caused by SRD5A2 knockout (KO), bulk RNA sequencing was conducted on Srd5a2^+/+, Srd5a2^+/–, Srd5a2^–/– prostate tissues to identify putative pathways involved in these processes. Hierarchical clustering successfully separated these groups based on differential gene expression. Significantly up-regulated (133) and down-regulated (2) genes for KO mice were determined using Gene Ontology analysis with Fisher exact test.

Supplemental Figure S2

Gene Ontology (GO) enrichment analysis of differentially expressed genes (DEG) in Srd5a2 genotypes. Bulk RNA sequencing was performed on prostate tissues from Srd5a2^+/+, Srd5a2^+/−, and Srd5a2^−/− mice. A: GO enrichment analysis was performed on significantly differentially expressed genes between Srd5a2^+/+ and Srd5a2^−/− prostate tissues. The top enriched GO terms are categorized into biological processes (red), cellular components (green), and molecular functions (blue). The x axis shows the GO terms, whereas the y axis represents the –log10 [adjusted P value (padj)] of enrichment significance. The numbers above the bars indicate the count of genes associated with each GO term. B: DEG analysis between Srd5a2^+/+, Srd5a2^+/−, and Srd5a2^−/− mice prostate tissues. Left: A volcano plot illustrates significantly up-regulated and down-regulated genes in the Srd5a2^−/− mice compared with Srd5a2^+/+ controls, using a threshold of |log₂foldchange| > 1 and padj < 0.05. Right: A volcano plot illustrates significantly up-regulated and down-regulated genes in the Srd5a2^−/− mice compared with Srd5a2^+/− controls. Horizontal dashed line: padj < 0.05. Vertical dashed line: |log₂ fold change| >1. n = 127 up-regulated genes (B, left); n = 24 down-regulated genes (B, left); n = 133 up-regulated genes (B, right); n = 2 down-regulated genes (B, right).

Supplemental Figure S3

Transcription factor regulatory networks and gene expression signatures across luminal epithelial subtypes in mouse prostate. A: Heat map displaying the activity of transcription factors across five luminal epithelial (LE1 to LE5) clusters identified from single-cell RNA sequencing. Row z-scores indicate the relative activity of key transcription factors in each cluster. B: Motif logos and regulatory networks of selected transcription factors linked to their target genes. Colored edges represent different regulon groups associated with luminal epithelial identity. Dot plot showing expression patterns of representative lineage, erythroblast transformation-specific-related gene (ERG) signaling, mitogen-activated protein kinase (MAPK) signaling, WNT signaling, stemness, and duct secretion genes across LE1 to LE5 clusters. The color represents expression level (row z-score), and the size of the dot reflects the proportion of expressing cells within each cluster. C: Spatial distribution of key regulon activities across mouse prostate cells. Uniform manifold approximation and projection plots showing the spatial distribution of regulon activity for selected transcription factors. Cells with higher predicted regulon activity are shown in purple, whereas cells with low or absent activity are shown in gray. Each plot highlights the distinct cellular localization patterns of transcription factor regulatory networks within the prostate epithelial cell population.

Supplemental Figure S4

SRD5A2 knockout impaired prostate mouse prostate gene profiles. Real-time quantitative PCR analysis of transcription factors (A), secreted ligands (B), and WNT5A receptors (C) identified from single-cell RNA-sequencing analysis in Srd5a2^+/+, Srd5a2^+/−, and Srd5a2^−/− mice whole prostate. The t-test and one-way analysis of variance were performed. Data are given as means ± SD (A–C). N ≥ 3 in each condition (A–C). ∗P ≤ 0.05, ∗∗P ≤ 0.01, ∗∗∗P ≤ 0.001, and ∗∗∗∗P ≤ 0.0001 versus +/+; ^##P ≤ 0.01, ^###P ≤ 0.001, and ^####P ≤ 0.0001 versus +/–. Lef1, lymphoid enhancer factor 1; Lrp6, LDL receptor-related protein 6; Ptprk, protein tyrosine phosphatase receptor type K; Rel., relative; Ror2, receptor tyrosine kinase-like orphan receptor 2; Ryk, receptor-like protein tyrosine kinase; Tcf7l1, transcription factor 7 like 1; Tgfb, transforming growth factor-beta; Vegfa, vascular endothelial growth factor A; Xrcc4, x-ray repair cross complementing 4; Zfp143, zinger finger protein 143; Zmiz1, zinc finger MIZ domain-containing protein 1.

Supplemental Figure S5

SRD5A2-WNT5A-LEF1 regulatory axis in prostate cells. A: Western blot analysis of SRD5A2, androgen receptor (AR), WNT5A, and β-actin in control and SRD5A2-overexpressing (OE) BHPrS1 cells. B and C: Real-time quantitative PCR analysis of WNT-related growth factors (B) and WNT5A receptors (C) in control and SRD5A2-overexpressing BHPrS1 cells. D: Representative images of BHPrS1 spheroid-like colonies from day 4 to day 18. On day 12, 12 wells of BHPrS1 spheroid-like colonies were treated with vehicle control (Vec; 3), WNT5A antagonist (BOX5;3), finasteride (Fin; 3), or the combination of two (3). E: Western blot analysis of SRD5A2, AR, WNT5A, β-catenin, cyclin D1, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) loading control in benign prostatic hyperplasia 1 (BPH1) cells co-cultured with BPH1 cells, BHPrS1 control cells, or BHPrS1 SRD5A2-overexpressing cells. F: Real-time quantitative PCR analysis of WNT5A receptors in BPH1 cells co-cultured with BPH1 cells, BHPrS1 control cells, or BHPrS1 SRD5A2-overexpressing cells. G: Cell viability of BPH1 cells cultured in siRNA-transfected BHPrS1-conditioned medium (C.M.) for 72 hours measured by MTT assay. The t-test and one-way analysis of variance with Tukey test were performed. Data are given as means ± SD (B, C, F, and G). N ≥ 3 (B, C, F, and G). ∗P ≤ 0.05, ∗∗P ≤ 0.01, ∗∗∗P ≤ 0.001, and ∗∗∗∗P ≤ 0.0001 versus BHPrS1 Vec (B and C), and versus BPH1 monoculture (mono) (G); ^####P ≤ 0.0001 versus Vec scramble (scr); ^††††P ≤ 0.0001 versus OE scr. Scale bar = 500 μm (D). Original magnification, ×4 (D). Lrp6, LDL receptor-related protein 6; N/A, not applicable; NC, negative control; PTPRK, protein tyrosine phosphatase receptor type K; Rel., relative; Ror2, receptor tyrosine kinase-like orphan receptor 2; Ryk, receptor-like protein tyrosine kinase; T, testosterone.

Supplemental Figure S6

Proposed schematic model of SRD5A2-regulated WNT5A expression and SRD5A2 regulating prostate growth via poly (ADP-ribose) polymerase 1 (PARP1) and IGF-1. High expression of SRD5A2 in the prostate stromal compartment stimulates the expression of WNT5A and activates the β-catenin/LEF1 transcriptional activator complex. This active complex increases the expression of IGF1 and PARP1, thereby promoting prostate stromal growth by enhancing proliferation and reducing apoptosis. Paracrine IGF1 further stimulates prostatic epithelial cells to proliferate and produce additional IGF1. Treatment of benign prostatic hyperplasia (BPH) with the SRD5A2 inhibitor finasteride prevents WNT5A overexpression and downstream LEF1 activation. Moreover, combining finasteride with a potential LEF1 or IGF1 inhibitor could further suppress prostatic cell proliferation. Together, these effects help reduce prostate growth and restore prostatic homeostasis of patients with BPH.