Targeting SWI/SNF ATPases in enhancer-addicted prostate cancer

Abstract

The switch/sucrose non-fermentable (SWI/SNF) complex has a crucial role in chromatin remodelling¹ and is altered in over 20% of cancers^2,3. Here we developed a proteolysis-targeting chimera (PROTAC) degrader of the SWI/SNF ATPase subunits, SMARCA2 and SMARCA4, called AU-15330. Androgen receptor (AR)⁺ forkhead box A1 (FOXA1)⁺ prostate cancer cells are exquisitely sensitive to dual SMARCA2 and SMARCA4 degradation relative to normal and other cancer cell lines. SWI/SNF ATPase degradation rapidly compacts cis-regulatory elements bound by transcription factors that drive prostate cancer cell proliferation, namely AR, FOXA1, ERG and MYC, which dislodges them from chromatin, disables their core enhancer circuitry, and abolishes the downstream oncogenic gene programs. SWI/SNF ATPase degradation also disrupts super-enhancer and promoter looping interactions that wire supra-physiologic expression of the AR, FOXA1 and MYC oncogenes themselves. AU-15330 induces potent inhibition of tumour growth in xenograft models of prostate cancer and synergizes with the AR antagonist enzalutamide, even inducing disease remission in castration-resistant prostate cancer (CRPC) models without toxicity. Thus, impeding SWI/SNF-mediated enhancer accessibility represents a promising therapeutic approach for enhancer-addicted cancers.

Fig. 1. AU-15330, a specific degrader of SWI/SNF ATPases, exhibits preferential cytotoxicity in enhancer-binding transcription factor-driven cancers.

a, Structure of AU-15330 and schematic of SMARCA2, SMARCA4 and PBRM1 domains. AU-15330-targeted bromodomains (BD) are shown. QLQ, conserved Gln, Leu, Gln motif containing domain; HSA, helicase/SANT-associated domain; BRK, Brahma and Kismet domain; SnAC, Snf2 ATP coupling domain; BAH1, bromo-adjacent homology domain 1; BAH2, bromo-adjacent homology domain 2. b, Immunoblots of SMARCA2, SMARCA4 and PBRM1 on treatment of HEK 293 and HeLa cells with AU-15330 at increasing concentrations or time durations. Vinculin is used as a loading control, and is probed on a representative immunoblot. This experiment was repeated independently twice. c, IC₅₀ of AU-15330 in a panel of human-derived cancer or normal cell lines after 5 days of treatment. Known SMARCA4 loss-of-function (LOF) alterations and multiple myeloma (MM) cell lines with MYC rearrangements (MYC-R'ed) are identified below the graph. AR and FOXA1 scores quantify their transcriptional activities using cognate multi-gene signatures. Source data

Fig. 2. SWI/SNF ATPase degradation disrupts physical chromatin accessibility at the core-enhancer circuitry to disable oncogenic transcriptional programs.

a, ATAC-seq read-density heat maps from VCaP cells treated with DMSO or AU-15330 for indicated durations (n = 2 biological replicates). b, Genome-wide changes in chromatin accessibility upon AU-15330 treatment for 4 h in VCaP cells along with genomic annotation of sites that lose physical accessibility (lost) or remain unaltered (retained). c, d, ChIP–seq read-density heat maps for AR and FOXA1 (c) and H3K27Ac (d) at the AU-15330 (AU)-compacted genomic sites in VCaP cells after treatment with DMSO or AU-15330 (1 μM) for indicated times and stimulation with R1881 (1 nM, 3 h). e, RNA-seq heat maps for classical AR target genes in LNCaP, VCaP and LAPC4 prostate cancer cells with or without 24 h of AU-15330 treatment. Source data

Fig. 3. SWI/SNF ATPase degradation disrupts enhancer–promoter loops to temper supra-physiologic expression of driver oncogenes.

a, Immunoblots of indicated proteins in VCaP cells treated with DMSO for 24 h or AU-15330 (1 μM) for increasing time durations. GAPDH is used as a loading control, and is probed on a representative immunoblot. This experiment was repeated independently twice. b, H3K27Ac ChIP–seq signal rank-ordered list of super-enhancers in VCaP cells with select cis-coded driver oncogenes denoted (HOMER). c, Normalized read density of ATAC-seq at super-enhancers (n = 32,545 sites) in VCaP cells treated with DMSO or AU-15330 (1 μM) for 1 or 4 h (two-sided t-test). In box plots, the centre line shows median, box edges mark quartiles 1–3, and whiskers span quartiles 1–3 ± 1.5 × interquartile range. d, H3K4me3 HiChIP–seq heat maps within the AR gene locus in VCaP cells with or without AU-15330 (1 μM) treatment for 4 h (bin size = 25 kb). ATAC-seq read-density tracks from the same treatment conditions are overlaid. Grey highlights mark enhancers; blue highlights the AR promoter. Loops indicate read-supported cis interactions within the locus. IR, interaction reads. e, APA plots for H3K4me3 and H3K27Ac HiChIP–seq data for all possible interactions between putative enhancers and gene promoters in VCaP cells with or without with AU-15330 treatment (1 μM, 4 h). Source data

Fig. 4. AU-15330 inhibits tumour growth in preclinical models of CRPC and synergizes with enzalutamide.

a, Tumour volume (measured twice per week using callipers) in the VCaP-CRPC model with AU-15330 alone or in combination with enzalutamide (two-sided t-test). Data are mean ±s.e.m. (vehicle: n = 18; AU-15330: n = 20; enzalutamide: n = 18; AU-15330 + enzalutamide: n = 16). b, Waterfall plot depicting change in tumour volume after 33 days of treatment. Response evaluation criteria in solid tumours (RECIST) was used to stratify tumours: progressive disease (PD), at least a 20% increase in tumour size; stable disease (SD), increase of <20% to a decrease of <30%; partial response (PR), at least a 30% decrease. The vehicle and enzalutamide groups have 100% PD; the AU-15330 group has 61% PD, 33% SD and 6% PR; and the AU-15330 + enzalutamide group has 0% PD, 12% SD and 88% PR. c, Representative haematoxylin and eosin (H&E) staining and immunohistochemistry from the VCaP-CRPC xenograft study (n = 2 tumours per condition). Insets in the H&E images show expanded views of apoptotic cells. d, Tumour volume measurements showing efficacy of AU-15330, enzalutamide or combined treatment in C4-2B-derived CRPC xenografts (n = 20 per condition; two-sided t-test). Data are mean ± s.e.m. e, Tumour volume measurements showing the effect of enzalutamide alone or in combination with AU-15330 in the castration-resistant MDA-PCa-146-12 PDX study (two-sided t-test). Data are mean ± s.e.m. f, Mechanism of action of AU-15330-triggered cytotoxicity in AR–FOXA1-signalling-driven prostate cancer. SWI/SNF ATPase degradation induces a rapid, targeted loss in chromatin accessibility at the core-enhancer circuitry of AR, FOXA1, ERG and MYC, thereby attenuating their cancer-promoting transcriptional programs and tempering the enhancer-wired supra-physiologic expression of driver oncogenes. Source data

Extended Data Fig. 1. Conformational model of AU-15330 target interaction and activity profile in diverse cell lines.

(a) Docking model of AU-15330 (cyan sticks) with the SMARCA2 and VHL complex. AU-15330 is suggested to fit into the pocket of SMARCA2 and VHL and capture several key interactions. Key hydrogen bond interactions with protein residues (pink sticks in SMARCA2, white sticks in VHL) are shown by yellow dashes. (b) Effects of AU-15330 (1 μM, 4h) on the proteome of VCaP cells. Data plotted Log2 of the fold change (FC) versus DMSO (dimethyl sulfoxide) control against –Log10 of the p-value per protein (FDR, false discovery rate) from n = 3 independent experiments. All t-tests performed were two-tailed t-tests assuming equal variances. TMT, tandem mass tag. (c) Heatmap showing TMT-based MS abundance of detectable SWI/SNF components after 4h of treatment with AU-15330 at 1 μM. Data from three independent replicates are shown. (d) Heatmap of relative abundance of several bromodomain-containing proteins detected via Tandem Mass Tag (TMT)-based quantitative MS upon 4h AU-15330 treatment. DMSO, dimethyl sulfoxide (vehicle). (e) Heatmap of mammalian SWI/SNF (BAF) complex subunits split into three constituent modules detected in SMARCC1 (also known as BAF155) nuclear co-immunoprecipitation followed by MS. Direct AU-15330 targets are in bold. (f) Dose-response curves of cells treated with AU-15330 and AU-16235 (inactive epimer of AU-15330). Data are presented as mean +/− SD (n = 6) from one-of-three independent experiments. (g) Crystal violet staining showing the effect of AU-15330 on colony formation. This experiment was repeated independently twice. (h) Dose-response curves and IC₅₀ of cells treated with AU-15330, ACBI1, and BRM014. Data are presented as mean +/− SD (n = 6) from one-of-three independent experiments. (i) Immunoblots of noted proteins in VCaP cells treated with AU-15330, ACBI1, or BRM014 at increasing concentrations for 24h. Vinculin is the loading control probed on all immunoblots. This experiment was repeated independently twice. (j) Representative immunohistochemistry images showing expression of indicated proteins in patient-derived breast cancer cell lines. (k) Immunoblots of noted proteins in WA-72-P or WA-72-As breast cancer cells treated with DMSO or AU-15330 at noted concentrations for 24h, Vinculin is the loading control probed on a representative immunoblot. This experiment was repeated independently twice. Source data

Extended Data Fig. 2. Verification of PROTAC design of AU-15330 and confirmation of on-target growth effects.

(a) Immunoblots for indicated proteins in normal (RWPE) or PCa cells (LNCaP, VCaP, 22RV1, and LAPC4) treated with AU-15330 at varied concentrations. Vinculin is the loading control probed on a representative immunoblot. This experiment was repeated independently twice. (b) Western blot analysis showing the time-dependent effect of AU-15330 on SMARCA2, SMARCA4, and PBRM1 in RWPE, LNCaP, and VCaP cells. Vinculin is the loading control probed on a representative immunoblot. This experiment was repeated independently twice. (c) Immunoblots in LNCaP and VCaP cells examining time-dependent cleavage of PARP upon AU-15330 treatment. Vinculin is the loading control probed on a representative immunoblot. This experiment was repeated independently twice. (d) Dose-response curves of VCaP, LNCaP, PNT2, PNT2, BPH1, Bin67, and HEK293 cells treated with AU-15330, AU-15139, or AU-16235. Data are presented as mean +/− SD (n = 6) from one-of-three independent experiments. (e) Growth curves of non-neoplastic or PCa cells upon treatment with increasing concentrations of AU-15330. Bottom, rightmost panel shows real-time assessment of apoptotic signals in LNCaP cells after treatment with DMSO or increasing AU-15330 concentrations. Data are presented as mean +/− SD (n = 5) from one-of-three independent experiments. (f) (top) Chemical structure of AU-15330, AU-16235 (an epimer control of AU-15330), and AU-15139 (parent bromodomain-binding ligand of AU-15330). (bottom) Immunoblots for SMARCA4 and PBRM1 in LNCaP and VCaP cells treated with AU-15330, AU-15139, or AU-16235 at indicated concentrations. Vinculin is the loading control probed on all immunoblots. This experiment was repeated independently twice. (g) Immunoblots of SMARCA4 and PBRM1 in VCaP and LNCaP cells pre-treated with VL285, MLN4924, bortezomib, or thalidomide for 1h, then treated with AU-15330 at noted concentrations for 4h. Vinculin is the loading control probed on all immunoblots. This experiment was repeated independently twice. (h) Real-time measure showing the rescue effect of VHL ligand on AU-15330-mediated growth inhibition in VCaP and LNCaP cells. Data are presented as mean +/− SD (n = 4) from one-of-three independent experiments. Source data

Extended Data Fig. 3. SWI/SNF ATPases, SMARCA2 and SMARCA4, mediate chromatin accessibility at numerous sites across the genome in PCa cells.

(a, b) ATAC-seq read-density heatmaps from VCaP cells treated with DMSO (solvent control), AU-15330, or ZBC-260 (a BRD4 degrader) for indicated durations at genomic sites that are compacted (a) or remain unaltered (b) upon AU-15330 treatment. Immunoblots show loss of target proteins upon treatment of cancer cells with AU-15330 (1 μM) for increasing durations or ZBC-260 (10 nM) for 4h. Vinculin is the loading control probed on all immunoblots. This experiment was repeated independently twice. Barplot shows the changes in mRNA expression (RNA-seq) of AU-15330 (1 μM) target genes in VCaP cells treated for noted durations. (c) Schematic outlining the CRISPR/Cas9 and shRNA-based generation of LNCaP cells with either independent or simultaneous inactivation of SWI/SNF ATPases, SMARCA2 and SMARCA4. Immunoblots showing the decrease in target expression in the genetic models shown above. Vinculin is the loading control probed on a representative immunoblot. This experiment was repeated independently twice. (d) ATAC-seq read-density heatmaps from genetically engineered LNCaP cells with SMARCA2 and/or SMARCA4 functional inactivation at AU-15330-compacted genomic sites. (e) Binding analysis for the regulation of transcription (BART) prediction of specific transcription factors mediating the observed transcriptional changes upon AU-15330 treatment in LNCaP or VCaP cells. The top 10 significant and strong (z-score) mediators of transcriptional responses are labeled (BART, Wilcoxon rank-sum test). (f) Top ten de novo motifs (ranked by p-value) enriched within AU-15330-compacted genomic sites (HOMER, hypergeometric test) in VCaP cells. (g) De novo motif analysis with top 10 motifs (ranked by p-value) enriched within genomic sites that retain chromatin accessibility upon AU-15330 treatment in VCaP cells (HOMER hypergeometric test). Source data

Extended Data Fig. 4. SWI/SNF inhibition condenses chromatin at enhancer sites bound by oncogenic transcription factors AR and FOXA1 in PCa cells.

(a) ATAC-seq read-density heatmaps from LNCaP cells treated with DMSO or AU-15330 for indicated durations at all genomic sites that lose physical accessibility upon AU-15330 treatment. (b) Genome-wide changes in chromatin accessibility upon AU-15330 treatment for 12h in LNCaP cells, along with genomic annotation of sites that are lost or retained in the AU-15330-treated cells. (c) De novo motif analysis with top 10 motifs (ranked by p-value) enriched within AU-15330-compacted or unaltered genomic sites in LNCaP cells (HOMER, hypergeometric test). (d) ChIP-seq read-density heatmaps for ERG at the AU-15330-compacted genomic sites in VCaP cells after treatment with DMSO or AU-15330 (1 μM) for indicated times and stimulation with R1881 (1 nM, 3h). (e) Genome-wide changes in AR and FOXA1 ChIP-seq peaks upon AU-15330 treatment (1 μM, 6h) in VCaP cells stimulated with R1881, a synthetic androgen (1 nM, 3h). (f) Immunoblots showing the changes in indicated chemical histone marks upon treatment with AU-15330. Vinculin is the loading control probed on a representative immunoblot. This experiment was repeated independently twice. (g) ChIP-seq read-density heatmaps for AR, FOXA1, and H3K27Ac at the compacted genomic sites in LNCaP cells after indicated durations of treatment with AU-15330 (1 μM). (h) Genome-wide changes in AR and FOXA1 ChIP-seq peaks upon AU-15330 treatment (1 μM, 6h) in LNCaP cells stimulated with R1881 (1 nM, 3h). (i) ChIP-seq tracks for AR, FOXA1, and H3K27Ac within the KLK2/3 gene locus in R1881-stimulated VCaP and LNCaP cells with or without AU-15330 (AU).

Extended Data Fig. 5. The SWI/SNF complex is a common chromatin cofactor of the central transcriptional machinery in PCa cells.

(a) The overlap between AR, FOXA1, ERG, and SMARCC1 ChIP-seq peaks in VCaP cells. (b) Genomic annotation of oncogenic transcription factor and SWI/SNF (SMARCC1) chromatin binding sites. (c) The overlap between transcription factor and SWI/SNF complex shared genomic sites (from a) and H3K27Ac ChIP-seq peaks along with the genomic annotations of the shared binding sites. (d) Left: volcano plot showing the AR interacting proteins identified from AR immunoprecipitation followed by MS. Significantly enriched SWI/SNF subunits are highlighted in red (two-sided t-test). Right: Overlap between AR, FOXA1, and ERG interacting proteins identified from in-house or publicly available datasets. (e) Immunoblots for indicated proteins followed by nuclear co-immunoprecipitation (IP) of AR, FOXA1, ERG, or SMARCC1 (a core SWI/SNF subunit) in VCaP and LNCaP cells after DHT (dihydrotestosterone) stimulation (10 nM, 3h). This experiment was repeated independently twice. Source data

Extended Data Fig. 6. The canonical SWI/SNF complex is the primary cofactor of enhancer-binding transcription factors and is essential for enabling their oncogenic gene programs.

(a, c) Genome-wide ChIP-seq read-density heatmaps and Venn diagrams for CTCF in VCaP (a) or LNCaP (c) cells treated with either DMSO or AU-15330 (1 μM) for 6h. Vinculin is the loading control probed on a representative immunoblot. (b, d) Immunoblots of indicated proteins in VCaP (b) or LNCaP (d) cells treated with AU-15330 (1 μM) for increasing time durations. Total histone H3 is the loading control probed on all immunoblots. This experiment was repeated independently twice. (e) GSEA plots for ERG, FOXA1, and MYC-regulated genes using the fold change rank-ordered gene signature from AU-15330-treated (1 nM, 24h) VCaP cells. NES, net enrichment score; adj P, adjusted p-value; DEGs, differentially expressed genes. (f, g) GSEA of FOXA1, MYC, or ARID1A-regulated genes (see Methods for gene sets) in the fold change rank-ordered AU-15330 gene signature in indicated PCa cells. DEGs, differentially expressed genes. (n = 2 biological replicates, GSEA enrichment test) (h, i) Expression of indicated genes (qPCR) in VCaP (h) or LNCaP (i) cells upon treatment with DMSO, AU-15330, dBRD7 (BRD7 degrader), or dBRD7/9 (dual BRD7 and BRD9 degrader) at 1 μM for 24h. Data are presented as mean +/− SD (n = 3, technical replicates) from one-of-two independent experiments. (j) Immunoblots for indicated proteins in LNCaP and VCaP cells treated with AU-15330, dBRD9 (BRD9 degrader), or VZ185 (BRD7/9 degrader) at indicated concentrations. Vinculin is the loading control probed on all immunoblots. This experiment was repeated independently twice. Source data

Extended Data Fig. 7. SWI/SNF inhibition down-regulates the expression of oncogenic drivers through disruption of promoter and super-enhancer interactions.

(a, b) RNA expression (RNA-seq) heatmaps from VCaP or LNCaP cells treated with DMSO, AU-15330 (1 μM), or ZBC-260 (BRD4 degrader) for the noted durations. n = 2, biological replicates. (c) RNA expression (qPCR) of indicated genes in stable CRISPR-engineered LNCaP-sgNC (control) or LNCaP-sgSMARCA2 (SMARCA2 inactivation) cells that were treated with a non-target control shRNA or two distinct shRNAs targeting the SMARCA4 gene. Data are presented as mean +/− SD (n = 3, technical replicates) from one-of-two independent experiments. Right, immunoblots showing expression of the indicated protein in CRISPR/shRNA-engineered LNCaP cells. Vinculin is the loading control probed on a representative immunoblot. This experiment was repeated independently twice. (d) Normalized read density of AR, FOXA1 and H3K27Ac ChIP-seq signal at the super-enhancer sites (n = 1,551 sites) in VCaP cells treated with DMSO or AU-15330 (1 μM) for 4h or H3K27Ac with 24h AU-15330 (two-sided t-test). For all box plots, the center shows median, box marks quartiles 1–3, and whiskers span quartiles 1–3 ± 1.5 × IQR. Source data

Extended Data Fig. 8. Enhancer-promoter interactions at loci of oncogenic transcription factors with AU-15330.

(a) ATAC-seq and ChIP-seq tracks for AR, FOXA1, and H3K27Ac within the AR gene locus in VCaP cells with or without AU-15330 treatment (1 μM for 6h for AR and FOXA1; 1 μM for 24h for H3K27Ac). (b) H3K27Ac HiChIP-seq heatmaps within the FOXA1 gene locus in VCaP cells plus/minus treatment with AU-15330 (1 μM) for 4h (bin size = 25Kb). ATAC-seq read-density tracks from the same treatment conditions are overlaid. Grey highlights mark enhancers, while the blue highlight marks the FOXA1 promoter. (c) Aggregate peak analysis (APA) plots for H3K4me3 (active promoter mark) HiChIP-seq data for all possible interactions between putative enhancers and gene promoters in VCaP cells plus/minus treatment with AU-15330 (1 μM) for noted durations. (d) APA plots for CTCF HiChIP-seq data for all possible interactions between CTCF-bound insulator elements in VCaP cells plus/minus treatment with AU-15330 (1 μM, 4h). TAD, topologically associating domain. (e) CTCF HiChIP-seq heatmaps in a gene locus at Chr14, including the FOXA1 topologically associating domain (TAD), in VCaP cells plus/minus treatment with AU-15330 (1 μM) for 4h (bin size = 100Kb). CTCF ChIP-seq read-density tracks from VCaP cells plus/minus AU-15330 treatment (1 μM) for 6h are overlaid. Source data

Extended Data Fig. 9. AU-15330 is well tolerated in mice and induces on-target degradation of SMARCA2, SMARCA4, and PBRM1.

(a) Immunoblots of indicated proteins in B16F10 and MC38 cells treated with DMSO or AU-15330 (100 nM or 1 μM). Vinculin is the loading control probed on a representative immunoblot. This experiment was repeated independently twice. (b) Schematic outlining the AU-15330 in vivo study in non-tumor bearing CD-1 mice. Male mice were treated with vehicle (control) or AU-15330 at the indicated concentration throughout the experiment. (c) Pharmacokinetics profile of AU-15330 following intravenous (IV) injection in CD-1 mice. Mice received a single injection at indicated concentration of AU-15330, and plasma levels were determined by HPLC. Data are presented as mean +/− SD (n = 6, biological replicates). (d) Immunohistochemistry staining of SMARCA4/BRG1 was carried out using lung, small intestine, and prostate sections after necropsy to show on-target efficacy of AU-15330 in vivo (n = 2, biological replicates). (e) Body weight measurements showing AU-15330 does not affect weight of non-tumor bearing CD-1 mice. Data are presented as mean +/− SD (n = 6, biological replicates). (f) Major organ weight measurements (taken after necropsy) showing AU-15330 does not affect their weight in non-tumor bearing CD-1 mice. Data are presented as mean +/− SD (n = 6, biological replicates). (g) Complete blood count showing AU-15330 does not affect the hematologic system. Non-tumor bearing CD-1 mice were treated with vehicle or AU-15330 at the indicated concentration throughout the treatment period, and whole blood was then collected and processed. WBC, white blood cells; RBC, red blood cells; PLT, platelets. Data are presented as mean +/− SD (n = 6, biological replicates). Source data

Extended Data Fig. 10. Combined in vivo treatment with AU-15330 and enzalutamide causes tumor regression in PCa xenografts without toxic effects on other organs.

(a) Schematic outlining the AU-15330 in vivo efficacy study using the VCaP-CRPC xenograft model. VCaP cells were subcutaneously grafted in immunocompromised mice that were castrated after 2 weeks of tumor growth to induce disease regression. This was eventually followed by tumor re-growth in the androgen-depleted conditions, generating the aggressive, castration-resistant tumors. (b) Individual tumors and weights from vehicle, enzalutamide, AU-15330, and AU-15330+enzalutamide groups from VCaP-CRPC study (two-sided t-test). Data are presented as mean+/−SEM (vehicle: n = 18, enzalutamide: n = 20, AU-15330: n = 18, AU-15330+enzalutamide: n = 16). For all box plots, the center shows median, box marks quartiles 1–3, and whiskers span the range. (c) Immunoblots of direct AU-15330 targets (upper) and oncogenic transcription factors (bottom) from VCaP-CRPC xenografts (n = 4 tumors/arm) after 5 days of in vivo treatment. Vinculin is the loading control probed on a representative immunoblot. (d) Representative immunohistochemistry images from the VCaP-CRPC xenograft study (n = 2 tumors/arm) for SMARCA2 and SMARCA4. (e) Box plot of the percent of cells with positive Ki-67 staining. Two-sided t-test shows significant differences between vehicle vs. enzalutamide, AU-15330, or AU-15330+enzalutamide groups. Data are presented as mean +/− SEM (n = 4, biological replicates). For all box plots, the center shows median, box marks quartiles 1–3, and whiskers span the range. (f) Percent body weight measurement showing the effect of vehicle, enzalutamide, AU-15330, and combination of AU-15330 and enzalutamide throughout the treatment period (two-sided t-test). Data are presented as mean +/− SEM (vehicle: n = 9, enzalutamide: n = 10, AU-15330: n = 9, AU-15330 + enzalutamide: n = 8). (g) H&E staining was carried out to examine the effect of AU-15330 in vivo using colon, spleen, liver, and kidney sections after necropsy. Representative images of H&E staining are shown. (h) Immunohistochemistry staining of SMARCA4/BRG1 was carried out using liver and kidney sections after necropsy to show on-target efficacy of AU-15330 in vivo. Source data

Extended Data Fig. 11. AU-15330 inhibits CRPC growth and synergizes with the AR antagonist enzalutamide.

(a) Schematic outlining the AU-15330 in vivo efficacy study using the C4-2B (CRPC) xenograft model. C4-2B-xenograft bearing male mice were castrated and, upon tumor regrowth, randomized into various treatment arms. (b) Body weight measurements showing the effect of the indicated treatments on animal weight. Tumor-bearing SCID mice were treated with the indicated drug throughout the treatment period, and the body weight was measured at endpoint. Data are presented as mean +/− SEM (n = 10, biological replicates). (c) Individual tumor volumes from different treatment groups with p-values are shown (two-sided t-test). Data are presented as mean +/− SEM (n = 20, biological replicates). (d) Immunoblots of direct AU-15330 targets (SMARCA2, SMARCA4, and PBRM1) in the whole cell lysate from C4-2B xenografts from all treatment arms after 5 days of in vivo treatment (n = 4, biological replicates). Vinculin is the loading control probed on a representative immunoblot. (e-g) VCaP, C4-2B, and LNCaP cells were treated with AU-15330 and/or enzalutamide at varied concentrations to determine the effect on cell growth and drug synergism, with assessments using the Bliss Independence method. Red peaks in the 3D-plots denote synergy with the average synergy scores noted above. The mean of three biological replicates is shown on top. Data are presented as mean (n = 4) from one-of-three independent experiments. (h) Crystal violet staining showing the synergistic effect of AU-15330 and enzalutamide on colony formation in VCaP and LNCaP. (i, j) Dose–response curves of VCaP cells treated with enzalutamide in combination with DMSO or AU-15330 at indicated concentrations. Data are presented as mean +/− SD (n = 4) from one-of-three independent experiments. (k) Dose-response curves of VCaP_Parental and VCaP_EnzR cells treated with enzalutamide or AU-15330. Data are presented as mean +/− SD (n = 6) from one-of-three independent experiments. (l) IC₅₀ for AU-15330 in enzalutamide-resistant (EnzR) LNCaP and VCaP cells after 5 days of treatment. Source data

Extended Data Fig. 12. AU-15330 inhibits tumor growth of an enzalutamide-resistant patient-derived xenograft (PDX) model without evident toxicities.

(a) Schematics outlining the AU-15330 in vivo efficacy studies using the MDA-PCa-146-12 (top) or the MDA-PCa-146-12-CRPC (bottom) xenograft model. MDA-PCa-146-12-CRPC xenograft-bearing male mice were castrated and, upon tumor regrowth, randomized into various treatment arms that were administered vehicle, enzalutamide, or the combination of AU-15330+enzalutamide at indicated concentrations. (b) Tumor volume measurements (caliper twice per week) showing efficacy of enzalutamide alone or in combination with AU-15330 in the enzalutamide-resistant MDA-PCa-146-12 PDX model (n = 20/arm; two-sided t-test). Data are presented as mean +/− SEM (vehicle: n = 18, enzalutamide: n = 18, AU-15330+enzalutamide: n = 16). (c) Individual tumor weights from different treatment groups from the MDA-PCa-146-12 PDX study with p-values indicated (two-sided t-test). Data are presented as mean +/− SEM (vehicle: n = 18, enzalutamide: n = 18, AU-15330+enzalutamide: n = 8). (d) Waterfall plot showing percent change from baseline of individual tumors from the MDA-PCa-146-12-CRPC model with indicated treatment group after 43 days of treatment. (e, f) Animal body weight measurements showing the effect of vehicle, enzalutamide, and combination of AU-15330 and enzalutamide on animal weight in the (e) MDA-PCa-146-12 or the (f) MDA-PCa-146-12-CRPC PDX models. Tumor-bearing SCID mice were treated with vehicle, enzalutamide, or a combination of AU-15330 and enzalutamide at the indicated concentration throughout the treatment period. Data are presented as mean +/− SEM (for e, vehicle: n = 9, AU-15330: n = 9, AU-15330+enzalutamide: n = 8; for f, vehicle: n = 7, AU-15330: n = 8, AU-15330+enzalutamide: n = 8). (g) Representative Alcian blue staining images from the large intestinal tract harvested at the VCaP-CRPC efficacy study endpoint (n = 2/treatment group). Right, quantification of goblet:epithelial cell densities in the colon (two-sided t-test). Data are presented as mean +/− SEM (n = 6, biological replicates). (h) Top, Representative H&E of the testis gland harvested from DMSO or AU-15330-treated intact male mice after 21 days of in vivo treatment. Right, quantification of germ cell density and maturation carried out using the Johnsen scoring system (two-sided t-test). Bottom, gross images of the testis glands. Data are presented as mean +/− SEM (n = 6, biological replicates). For all box plots, the center shows median, box marks quartiles 1–3, and whiskers span the range. (i) Individual testes weight and images from different treatment groups of the C4-2B xenograft efficacy study at endpoint (i.e., after 24 days of treatment) with p-values indicated (two-sided t-test). Data are presented as mean +/− SEM (vehicle: n = 9, enzalutamide: n = 10, AU-15330: n = 10, AU-15330+enzalutamide: n = 10). Source data