Abstract

Background

   Resistance to poly (ADP-ribose) polymerase inhibitors (PARPi) poses a
   major challenge to therapeutic efficacy in castration-resistant
   prostate cancer (CRPC). Although circular RNAs (circRNAs) have emerged
   as critical regulators in cancer biology, their involvement in PARPi
   resistance remains largely uncharacterized.

Objective

   This study aims to elucidate the molecular mechanism by which
   hsa_circ_0038737 modulates PARPi resistance in CRPC through
   post-transcriptional regulatory pathways.

Methods

   We employed a comprehensive set of in vitro and in vivo approaches,
   including qRT-PCR, RNA sequencing, RNA-protein pull-down, RNA
   immunoprecipitation, functional assays, and xenograft/organoid models,
   to investigate the biological function and mechanistic role of
   hsa_circ_0038737 in CRPC progression and therapeutic response.

Results

   We identified hsa_circ_0038737 as a nuclear-enriched circRNA
   significantly upregulated in CRPC, with expression levels correlating
   with poor prognosis and aggressive clinical features. Mechanistically,
   hsa_circ_0038737 interacts with RNA-binding protein (RBP) IGF2BP3,
   enhancing the stability of DNPH1 mRNA, a nucleotide sanitizer critical
   for DNA repair. The circRNA-RBP-mRNA regulatory axis promotes PARPi
   resistance by facilitating DNA damage repair capacity. Moreover, we
   revealed that reverse-complementary Alu elements mediate circRNA
   biogenesis, with HNRNPDL facilitating this process. Pharmacologic
   inhibition of DNPH1 effectively restored PARPi sensitivity both in
   vitro and in vivo.

Conclusion

   Our findings reveal a novel hsa_circ_0038737/IGF2BP3/DNPH1 axis driving
   PARPi resistance in CRPC, offering promising potential biomarkers and
   therapeutic targets to overcome resistance and improve treatment
   outcomes in advanced prostate cancer.

Supplementary Information

   The online version contains supplementary material available at
   10.1186/s12943-025-02447-y.

   Keywords: Castration-resistant prostate cancer (CRPC), Circular RNA
   (circRNA), PARP inhibitor, IGF2BP3, DNPH1 stabilization

Introduction

   Prostate cancer (PCa) is the second most commonly diagnosed cancer
   among men, with an estimated 1,466,680 new cases and approximately
   396,792 deaths worldwide in 2022 [[44]1, [45]2]. Castration-resistant
   prostate cancer (CRPC) denotes an advanced stage of PCa characterized
   by disease progression despite castrate levels of testosterone.
   Virtually all prostate cancer-related mortalities are attributable to
   CRPC, frequently with metastatic progression (mCRPC) [[46]3].
   Epidemiologically, CRPC constitutes 10%−20% of all PCa cases [[47]4,
   [48]5]. Upon the onset of metastasis, the median survival duration is
   less than three years [[49]6]. Approximately 28% of mCRPC patients
   harbor mutations in homologous recombination repair (HRR) genes, such
   as BRCA1/2, making them candidates for poly-ADP-ribose polymerase
   inhibitors (PARPi) [[50]7, [51]8]. PARPi, such as olaparib and
   rucaparib, exploit synthetic lethality by targeting HRR-deficient
   tumors, leading to DNA damage accumulation and cell death [[52]9,
   [53]10]. However, the clinical efficacy of PARPi is often limited by
   the development of resistance, which remains a significant challenge in
   CRPC treatment [[54]11]. Mechanisms of PARPi resistance include
   reversion mutations in BRCA genes, upregulation of drug efflux pumps,
   and alterations in DNA repair pathways [[55]12–[56]14]. Understanding
   these resistance mechanisms is crucial for developing strategies to
   overcome therapeutic limitations and improve patient outcomes.

   Circular RNAs (circRNAs) have emerged as key regulators of
   tumorigenesis and cancer progression due to their unique circular
   structure and functional versatility [[57]15–[58]17]. Unlike linear
   RNAs, circRNAs are highly stable and resistant to degradation, allowing
   them to modulate gene expression through interactions with RNA-binding
   proteins (RBPs) and microRNAs [[59]18]. In PCa, circRNAs have been
   implicated in regulating critical pathways such as cell proliferation,
   apoptosis, and drug resistance [[60]16]. For instance, METTL3 enhances
   the stability of CircGLIS3 through m6A modification, and CircGLIS3
   sequesters miR-661 to upregulate MDM2 mRNA levels, thereby regulating
   the p53 signaling pathway to promote the proliferation, migration, and
   invasion of PCa [[61]19]. Additionally, circ_0004087 interacts with the
   transcriptional coactivator SND1, stimulating the transactivation of
   MYB and enhancing the expression of its downstream target BUB1, which
   in turn promotes mitosis and increases PCa resistance to docetaxel
   [[62]20]. These findings underscore the potential of circRNAs as
   biomarkers and therapeutic targets in PCa.

   In this study, we investigate the role of hsa_circ_0038737 in mediating
   PARPi resistance in CRPC. We demonstrate that hsa_circ_0038737 enhances
   the stability of DNPH1 mRNA by binding to IGF2BP3, a well-characterized
   RBP involved in mRNA stabilization. This interaction leads to increased
   DNPH1 expression, which in turn contributes to PARPi resistance in CRPC
   cells. Our findings provide novel insights into the molecular
   mechanisms underlying PARPi resistance and highlight the potential of
   targeting circRNA-RBP interactions as a therapeutic strategy to
   overcome resistance in CRPC.

Materials and methods

Human tissue specimens and cell lines

   Between 2015 and 2022, a total of 78 PCa tissue samples were collected
   from patients who underwent radical prostatectomy at the Fudan
   University Shanghai Cancer Center (FUSCC). The collection and use of
   these tissue samples were approved by the hospital’s ethics committee,
   and written informed consent was obtained from all participating
   patients. Clinicopathological features of the 78 PCa patients were
   subsequently analyzed. The study was conducted in strict accordance
   with the guidelines of the Declaration of Helsinki. Additionally, five
   PCa cell lines (LNCaP, DU145, C4-2, PC-3, and 22RV1) and one human
   normal prostate epithelial cell line (RWPE-1) were obtained from the
   Chinese Academy of Sciences Type Culture Collection (Shanghai, China)
   for further experimental investigations.

Establishment of patient-derived tumor organoids

   PCa specimens were obtained from Fudan University Shanghai Cancer
   Center. Immediately after surgical resection, fresh tumor samples were
   placed in RPMI 1640 medium containing antibiotics, supplemented with 10
   µM Y-27,632 (MCE, USA) and 10% FBS, and transported to the laboratory
   on ice at 4 °C for immediate processing. Upon arrival, tumors were
   gently washed at least three times with pre-cooled 1X DPBS containing
   3% penicillin-streptomycin solution to remove surface contaminants. The
   tumor tissue was then minced into small pieces using surgical scissors
   and subjected to enzymatic digestion using a tissue dissociation system
   (D1Med, China) to generate a single-cell suspension. The digested
   suspension was passed through a 70 μm cell strainer to remove any
   undigested debris, followed by centrifugation at 500 × g for 5 min. The
   resulting cell pellet was resuspended in Matrigel (D1Med, China) and
   plated in 24-well culture dishes. After allowing the Matrigel to
   solidify for 15 min at room temperature, conditioned medium (provided
   by D1 Medical Technology, China) was added to support cell growth and
   maintenance.

Cell culture and transfection

   The RWPE-1 cell line was cultured in Keratinocyte Serum-free Medium
   (K-SFM; Gibco, USA). The prostate cancer cell lines LNCaP, DU145, C4-2,
   PC-3, and 22RV1 were cultured in Roswell Park Memorial Institute medium
   (RPMI-1640; Gibco, USA). All cell lines were Maintained in their
   respective media supplemented with 10% fetal bovine serum (FBS; Gibco,
   USA), 100 U/mL penicillin, and 100 µg/mL streptomycin (Invitrogen,
   USA). Cells were incubated at 37 °C in a humidified atmosphere
   containing 95% air and 5% CO₂.

Plasmid construction, transfection and lentivirus infection assay

   The hsa_circ_0038737 knockdown and overexpression lentiviral vectors
   were obtained from HANBIO (Shanghai, China). When 22RV1 or PC-3 cells
   reached approximately 50% confluence, they were transfected with either
   hsa_circ_0038737 knockdown or overexpression lentiviruses, as well as
   their respective negative control lentiviruses. Following transfection,
   cells were selected using puromycin at concentrations of 4–8 µg/mL for
   two rounds to ensure stable integration and expression. The
   overexpression plasmids and short hairpin RNAs (shRNAs) targeting
   IGF2BP3, HNRNPDL, and DNPH1 were acquired from GenePharma (Shanghai,
   China). Transfections of these constructs into 22RV1 or PC-3 cells were
   performed using the Lipofectamine 3000 reagent kit (Invitrogen, USA)
   according to the manufacturer’s instructions. The sequences of the
   shRNAs used in this study are listed in TableS1.

RNA extraction and quantitative real time-PCR (qRT-PCR)

   Total RNA was isolated from PCa tissues or cell lines using TRIzol
   reagent (Invitrogen, USA) following the manufacturer’s instructions.
   The extracted RNA was subsequently reverse-transcribed into cDNA using
   the HiScript II Reverse Transcriptase Kit (Vazyme, China). qRT-PCR was
   performed to quantify the expression levels of circRNA or mRNA using
   the SYBR Green Master Mix (Vazyme, China). The qRT-PCR experiments were
   conducted on either the StepOne Plus Real-Time PCR System (Applied
   Biosystems, USA) or the LightCycler 480 (Roche, USA). The primers used
   in this study were synthesized by TsingKe (China), and their detailed
   sequences are also provided in TableS1.

Protein isolation and Western blot assays

   Proteins were extracted from tissue or cell samples using RIPA buffer
   (Sigma, USA) supplemented with protease and phosphatase inhibitors to
   ensure protein integrity. Protein concentrations were quantified using
   the BCA Protein Assay Kit (Beyotime, China) according to the
   manufacturer’s instructions. Equal amounts of protein were separated by
   SDS-PAGE based on molecular weight and then transferred onto PVDF
   membranes (Millipore, USA). The membranes were blocked with 5% skim
   milk in Tris-buffered saline with Tween-20 at room temperature for 1 h
   to prevent nonspecific binding. Subsequently, the membranes were
   incubated with primary antibodies overnight at 4 °C. After extensive
   washing with TBST, the membranes were incubated with horseradish
   peroxidase-conjugated secondary antibodies for 1 h at room temperature.
   Chemiluminescent signals were detected using an enhanced
   electrochemiluminescence system (Thermo Fisher Scientific, USA) and
   visualized using a ChemiDoc MP Imaging System (Bio-Rad, USA).

RNase R treatment assay

   Total RNA was treated with RNase R (0.2 µl/µg, Epicenter) at 37 °C for
   30 min. Subsequently, the expression levels of hsa_circ_0038737 and
   GTF3C1 mRNA were detected by qRT-PCR.

Actinomycin D treatment and RNA stability assays

   The 22RV1 and PC-3 cell lines were treated with actinomycin D (1 µg/ml;
   Abcam, UK) at designated time points (0, 2, 4, 6, 8, and 10 h).
   Following treatment, total RNA was extracted from the cells, and
   qRT-PCR was performed to assess the expression levels of
   hsa_circ_0038737, GTF3C1, and DNPH1.

RNA-protein pulldown, silver staining and mass spectrometry assays

   RNA-protein pull-down assays were conducted using the Pierce Magnetic
   RNA-Protein Pull-Down Kit (Thermo Fisher Scientific, USA) according to
   the manufacturer’s instructions. Biotinylated probes specific to the
   back-splice junction of hsa_circ_0038737 were synthesized by GenePharma
   (GenePharma, China). Cells were lysed in IP lysis buffer and incubated
   with the biotinylated probes at room temperature for 4 h. The
   RNA-protein complexes were captured using streptavidin-coated magnetic
   beads. Subsequently, Western blotting was performed to identify the
   proteins that were pulled down by the probes. Silver staining was
   carried out using the Quick Silver Staining Kit (Beyotime, China) to
   visualize the protein bands. Further identification and quantification
   of the proteins were performed by mass spectrometry analysis at
   Keygentec (Keygentec, China), utilizing the Proteome Discoverer
   Software 2.4 (Thermo Fisher Scientific, USA) for data processing.

Immunofluorescence and fluorescence in situ hybridization (IF-FISH) assays

   Cell smears were fixed with 4% paraformaldehyde for 15 min at room
   temperature and then permeabilized with 0.1% Triton X-100 for 10 min.
   Cy3-labeled hsa_circ_0038737 probes (Genepharma, China) were hybridized
   with the cells in hybridization buffer at 37 °C for 16 h. After
   hybridization, cells were blocked with 5% bovine serum albumin for 1 h
   at room temperature to reduce nonspecific binding. Subsequently, cells
   were incubated sequentially with primary antibodies against IGF2BP3 and
   corresponding secondary antibodies (Beyotime, China) for 1 h each at
   room temperature. Nuclei were counterstained with DAPI (MCE, USA).
   Confocal microscopy images were captured using a TCS SPE-II confocal
   microscopy system (Leica, Germany) to visualize the localization of
   hsa_circ_0038737 and IGF2BP3 within the cells.

RNA immunoprecipitation (RIP) assay

   For RIP assays, cells were collected and lysed in IP lysis buffer. The
   Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore, USA)
   was used according to the manufacturer’s protocol. Specific antibodies
   against IGF2BP3 (Proteintech, China), FLAG (Proteintech, China), HNRNPL
   (Proteintech, China), and control IgG (Proteintech, China) were
   employed to precipitate RNA-protein complexes. Total RNA was extracted
   from both the input and IP samples using TRIzol reagent (Invitrogen,
   USA). The levels of specific RNAs were subsequently quantified by
   quantitative real-time PCR (qRT-PCR).

Cell proliferation and cloning formation assays

   To evaluate the proliferation capacity of PCa cells, 22RV1 or PC-3
   cells were plated at a density of 1,000 cells per well in a 96-well
   plate. Cell proliferation was assessed using the Cell Counting Kit-8
   (CCK-8) (MCE, USA), and absorbance values were measured at 450 nm with
   a microplate reader (Tecan, Switzerland).

   For colony formation assays, approximately 1,000 22RV1 or PC-3 cells
   were Seeded into each well of a 6-well plate. After 14 days of
   incubation, the cells were fixed with 4% paraformaldehyde and stained
   with 0.1% crystal violet solution. The colonies were visualized and
   quantified using ImageJ software (NIH, USA).

Cell cycle analysis

   Cells were harvested and fixed in a 70% ethanol solution at 4 °C for
   2 h. After fixation, the cells were washed with cold phosphate-buffered
   saline and stained with propidium iodide buffer (Beyotime, China) at
   room temperature for 30 min. PCa cells were then analyzed for cell
   cycle distribution using a FACS Scan flow cytometer (BD Biosciences,
   USA). The data were processed and analyzed using ModFit LT 2.0 software
   (Verity Software House, USA).

Half-maximal inhibitory concentration (IC50) assay

   Cells were Seeded in 96-well plates and treated with varying
   concentrations of olaparib or niraparib for 7 days. Cell viability was
   assessed using the CCK8 assay and IC50 values were calculated by
   plotting dose-response curves and determining the concentration that
   inhibits 50% cell growth compared to untreated controls.

Comet assay

   Cells were treated, harvested, and mixed with low-melting agarose using
   the Keygenbio Comet Assay Kit (Keygenbio, China). Lysed cells were
   subjected to electrophoresis, stained with DNA dye, and visualized
   under a fluorescence microscope (Olympus FV3000, Olympus Corp.; Japan)
   to evaluate DNA damage.

Proximity ligation assay (PLA)

   PC-3 and 22RV1 prostate cancer cells were plated in chambered
   coverslips and cultured for 24 h. Following fixation with 4%
   paraformaldehyde (15 min) and permeabilization in 0.3% Triton X-100
   (10 min), samples were blocked with Duolink blocking solution (Sigma,
   USA; 60 min, 37 °C). Primary antibodies against PARP1 and γ-H2AX were
   applied overnight at 4 °C. Species-specific PLA probes (Duolink In Situ
   PLA Probe Anti-Rabbit PLUS, Sigma; Anti-Mouse MINUS, Sigma) were
   incubated for 120 min at 37 °C. Oligonucleotide ligation (30 min,
   37 °C) and rolling-circle amplification (100 min, 37 °C) were performed
   using Duolink In Situ Detection Reagents Red (Sigma, USA). Nuclei were
   counterstained with DAPI (MCE, USA) in ProLong Gold mounting medium
   (MCE, USA). Imaging was conducted on an Olympus FV3000 confocal
   microscope with a 63× oil objective.

Chromatin fractionation

   To isolate chromatin-bound proteins, a subcellular fractionation
   procedure was conducted employing a commercial kit (Thermo Fisher
   Scientific, USA). Treated cells were harvested, washed with ice-cold
   PBS, and subjected to sequential extraction. Cytosolic components were
   first released by resuspending the cell pellet in Cytoplasmic
   Extraction Buffer (CEB) for a 10-minute incubation on ice, followed by
   centrifugation (500 × g, 5 min, 4 °C). Subsequent treatment of the
   resulting pellet with Membrane Extraction Buffer (MEB) for 10 min on
   ice and centrifugation (3,000 × g, 5 min, 4 °C) yielded the membrane
   protein fraction. The remaining insoluble material was then exposed to
   Nuclear Extraction Buffer (NEB) for 30 min on ice. Centrifugation
   (5,000 × g, 5 min, 4 °C) separated the soluble nuclear proteins in the
   supernatant. Finally, the chromatin-enriched pellet was digested with
   NEB containing nuclease for 15 min at ambient temperature. High-speed
   centrifugation (16,000 × g, 5 min) recovered the chromatin-associated
   proteins in the supernatant. All steps strictly adhered to the
   manufacturer’s protocol.

Immunohistochemistry (IHC) assay

   Paraffin-embedded tissue samples were cut into 4 μm sections and
   mounted onto glass slides. The sections were deparaffinized and
   rehydrated through a graded ethanol series. Tissue sections underwent
   heat-mediated antigen retrieval in 10 mM citrate buffer (pH 6.0) for
   20 min at 95 °C. The buffer was prepared by dissolving citric acid
   (MCE, USA) in ultrapure water, adjusted to pH 6.0 with NaOH. Following
   this, the sections were treated with 3% hydrogen peroxide for 10 min to
   quench endogenous peroxidase activity. Subsequently, the sections were
   incubated overnight at 4 °C with primary antibodies targeting Ki-67,
   IGF2BP3, PSMA, and AR (all from Proteintech, China). After washing, the
   sections were incubated with secondary antibodies for 30 min at room
   temperature. Immunostaining was visualized using diaminobenzidine (MCE,
   USA) as the chromogen, followed by counterstaining with hematoxylin to
   highlight cell nuclei. The stained sections were examined and imaged
   under a light microscope.

RNA sequencing

   Total RNA was extracted from three pairs of sh-hsa_circ_0038737 and
   sh-NC PC3 cells. Following stringent quality control measures, RNA
   sequencing was performed on the Illumina platform at Oebiotech
   (Shanghai, China). Differentially expressed genes were identified using
   the OECloud tool available at Oebiotech Cloud Platform
   ([63]https://cloud.oebiotech.com), with the criteria of q-value < 0.05
   and |log2 fold change (FC)| >0.5. Functional enrichment analysis was
   subsequently conducted using the ‘clusterProfiler’ R package to
   elucidate potential pathways associated with the observed gene
   expression changes.

Nucleic and cytoplasmic RNA assay

   Cellular and nuclear RNA were isolated separately using the Cytoplasmic
   & Nuclear RNA Purification Kit (Norgen Biotek, Thorold, ON, USA). 22RV1
   and PC-3 cells were lysed on ice in Lysis Buffer J, followed by
   centrifugation at 14,000 × g for 10 min at 4 °C. The supernatant,
   containing cytoplasmic RNA, was carefully transferred to an RNase-free
   tube. The pellet, containing nuclear RNA, was retained for further
   analysis. qRT-PCR was performed to quantify the RNA levels in both
   cytoplasmic and nuclear fractions. U6 and GAPDH mRNA were used as
   reference genes for the nuclear and cytoplasmic fractions,
   respectively, to evaluate the relative distribution of the target RNA
   between these compartments.

Xenograft experiments

   Male BALB/c nude mice, aged 6 to 7 weeks, were obtained from the Animal
   Medical Center Affiliated with Fudan University. To establish
   subcutaneous xenograft models, 22Rv1 cells (vehicle, hsa_circ_0038737
   shRNA, or hsa_circ_0038737 overexpression) were mixed 1:1 with Matrigel
   (Corning, USA) and injected subcutaneously into the dorsal flank of the
   mice. Experimental treatments included vehicle, olaparib (50 mg/kg), or
   a combination of olaparib and DNPH1 inhibitor (25 mg/kg). Both drugs
   were administered orally once Daily. Olaparib was given for 5
   consecutive days followed by a 2-day rest period, while the DNPH1
   inhibitor was administered continuously for 4 days each week, starting
   the day after olaparib treatment initiation. Tumor size and volume were
   measured every 5 days. After 3 weeks of treatment, animals were
   euthanized, and tumors were excised for further analysis. The animal
   experimental protocol was approved by the Institutional Animal Care and
   Use Committee of the Shanghai Veterinary Research Institute (Approval
   No. FUSCC-IACUC-2024293).

Analysis of GO and KEGG enrichment and network-based functional annotations

   To explore the functional characteristics of the 270 proteins detected
   by mass spectrometry, we conducted Gene Ontology (GO) and Kyoto
   Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses.
   Using clusterProfiler and org.Hs.eg.db, we performed GO enrichment
   analysis on the gene set, covering three main categories: Biological
   Process (BP), Cellular Component (CC), and Molecular Function (MF),
   with the p-value adjustment method being the Benjamini-Hochberg
   correction. Additionally, we employed the enrichKEGG function to carry
   out KEGG pathway enrichment analysis on the same gene set. All
   significantly enriched pathways were ranked based on -log10(p-value)
   and Rich Factor (defined as the ratio of the number of genes
   participating in a pathway to the total number of background genes in
   that pathway). The results were visualized using ggplot2 in a dual-axis
   bar chart, displaying the significance of p-values (-log10(pvalue)) and
   the degree of pathway enrichment (Rich Factor), thereby revealing key
   functional pathway characteristics. Further enrichment analysis based
   on WikiPathways database and visualization using the ClueGO (version
   2.5.9) and CluePedia (version 1.5.9) plugins within the Cytoscape
   software (version 3.9.1). The enrichment test employed was the
   two-sided hypergeometric test, with multiple testing correction applied
   using the Benjamini-Hochberg method (FDR < 0.05). The parameters for
   network construction were set as follows: a minimum of 3 pathway genes
   and a minimum gene proportion of 4%, with a functional connectivity
   (Kappa score) set at 0.4.

Cohort validation for the prognostic implications of DNPH1 expression

   Two independent PCa cohorts were collected from cbioportal database
   ([64]https://www.cbioportal.org/) in this study. Only samples with
   complete prognostic information and available RNA-seq transcriptomic
   data were retained for downstream analysis. The MSKCC cohort (n = 140)
   comprised patients with primary PCa, while the SU2C 2019 cohort
   (n = 81) consisted of CRPC cases. Progression-free survival (PFS) and
   overall survival were used as the index to evaluate prognosis of PCa
   patients, respectively. All of the expression data was annotated by
   corresponding platform and expressed as Transcripts Per Million (TPM).
   To evaluate the prognostic significance of the target gene, we utilized
   the Survminer package to determine the optimal cutoff value for gene
   expression based on the maximum selected rank statistic. Subsequently,
   patients were divided into high-expression and low-expression groups
   according to this cutoff value. The Kaplan-Meier survival curves were
   plotted and the differences between groups were compared using the
   log-rank test.

Statistical analysis

   Continuous variables are presented as mean ± standard deviation
   (mean ± SD), while categorical variables are described using counts and
   proportions. Statistical analysis and visualization were conducted
   using GraphPad Prism 10.0 software. Data were derived from public
   databases and sequencing results. Statistical comparisons were
   performed using t-tests, Mann-Whitney U tests, and χ² tests. A p-value
   of less than 0.05 was considered statistically significant, with
   significance levels indicated as follows: * for p < 0.05, ** for
   p < 0.01, *** for p < 0.001, and **** for p < 0.0001.

Results

Hsa_circ_0038737 is up-regulated in PCa tissues and correlated with
aggressive progression of PCa

   To elucidate the potential clinical relevance of circular RNAs in PCa,
   we first investigated the expression profile of hsa_circ_0038737, a
   previously uncharacterized circRNA. Analysis of five independent,
   publicly available transcriptomic datasets revealed that
   hsa_circ_0038737 is consistently and significantly upregulated in PCa
   tissues compared to matched adjacent normal tissues (Fig. [65]1A). This
   finding was further validated in our cohort using qRT-PCR, and the
   back-splicing junction characteristic of circular RNAs was confirmed by
   Sanger sequencing (Fig. [66]1B). Genomic alignment indicated that
   hsa_circ_0038737 is generated from exons 35 and 36 of the GTF3C1 gene,
   located on chromosome 16, through a canonical back-splicing mechanism.

Fig. 1.

   [67]Fig. 1
   [68]Open in a new tab

   Upregulation and Oncogenic Association of hsa_circ_0038737 in Prostate
   Cancer. (A) Detection of differentially expressed circRNAs in prostate
   cancer through five public sequencing datasets. (B) Identification of
   the reverse-splicing junction of hsa_circ_0038737 by Sanger sequencing.
   (C) Detection of hsa_circ_0038737 in cDNA and gDNA of 22RV1 and PC-3
   cells using divergent and convergent primers, respectively. (D) The
   half-life of hsa_circ_0038737 in 22RV1 cells determined by actinomycin
   D treatment. (E) The relative expression levels of hsa_circ_0038737 in
   22RV1 and PC-3 cells with or without RNase R treatment. (F-G) The
   subcellular localization of hsa_circ_0038737 in 22RV1 and PC-3 cells
   determined by nuclear and cytoplasmic fractionation and Fish. (H) The
   expression levels of hsa_circ_0038737 in different prostate cell lines.
   (I) The expression levels of hsa_circ_0038737 in normal and tumor
   prostate tissues. (J) The expression levels of hsa_circ_0038737 in
   different T stage prostate cancer tissues. (K) The expression levels of
   hsa_circ_0038737 in different Gleason scores prostate cancer tissues.
   Notes: Data are presented as mean ± SD. *P < 0.05, **P < 0.01,
   ***P < 0.001, ****P < 0.0001, ns, no significance

   To verify its circular topology, we designed convergent and divergent
   primers targeting the splice junction. The circRNA was successfully
   amplified from complementary DNA (cDNA), but not from genomic DNA
   (gDNA), confirming the absence of linear genomic contamination and the
   authenticity of its circular structure (Fig. [69]1C). Functional assays
   revealed that hsa_circ_0038737 exhibits enhanced stability, with a
   prolonged half-life under actinomycin D treatment compared to its
   linear counterpart GTF3C1 mRNA (Fig. [70]1D), and showed marked
   resistance to RNase R digestion, a hallmark of circRNAs (Fig. [71]1E).

   Subcellular fractionation and FISH assays demonstrated that
   hsa_circ_0038737 is predominantly localized in the nucleus
   (Figs. [72]1F, G), suggesting a potential role in transcriptional or
   epigenetic regulation. Expression analysis across a panel of PCa cell
   lines revealed significantly higher levels of hsa_circ_0038737 in
   malignant cells compared to the normal prostate epithelial cell line
   RWPE-1 (Fig. [73]1H). Consistently, in a cohort of 78 paired clinical
   PCa samples, hsa_circ_0038737 was markedly upregulated in tumor tissues
   (Fig. [74]1I), with expression levels positively correlating with
   advanced T stage and higher Gleason scores in PCa patients
   (Fig. [75]1J, K). Collectively, these findings position
   hsa_circ_0038737 as a stable, nucleus-enriched circRNA with oncogenic
   potential in prostate cancer. Its strong association with aggressive
   clinical features and enhanced expression in tumor tissues suggests a
   functional role in disease progression and highlights its potential
   utility as a biomarker for prognosis stratification and a candidate
   target for therapeutic intervention.

Hsa_circ_0038737 enhances proliferation of PCa cells via promotion of G2/M
cell cycle progression

   To investigate the functional role of hsa_circ_0038737 in PCa
   proliferation independent of its parental linear transcript GTF3C1, we
   established both loss- and gain-of-function models in two
   representative PCa cell Lines, 22RV1 and PC-3. Lentiviral vectors
   encoding either shRNAs targeting hsa_circ_0038737 or circRNA-specific
   overexpression constructs were transduced into the cells. Efficient
   knockdown and overexpression of hsa_circ_0038737 were confirmed by
   qRT-PCR (Fig. [76]2A, B; Figure S1 A-D), without significant alteration
   of GTF3C1 mRNA expression, thereby ensuring circRNA specificity.

Fig. 2.

   [77]Fig. 2
   [78]Open in a new tab

   Hsa_circ_0038737 Enhances Proliferation of Prostate Cancer Cells via
   Promotion of G2/M Cell Cycle Progression. (A-B) The relative expression
   levels of hsa_circ_0038737 were detected in 22RV1 and PC-3 cells after
   transfection with shRNAs (Sh1, Sh2, Sh3) or overexpression constructs.
   (C-E) CCK-8 and colony formation assays were performed to evaluate the
   impact of hsa_circ_0038737 knockdown and overexpression on the
   proliferation of 22RV1 and PC-3 cells. (F-G) Cell cycle analysis was
   conducted to investigate the effect of hsa_circ_0038737 knockdown and
   overexpression on cell cycle distribution in PCa cells. Data are
   presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001,
   ****P < 0.0001, ns, no significance

   To assess the impact of hsa_circ_0038737 on tumor cell proliferation,
   we conducted CCK-8 cell viability assays and colony formation assays.
   The results consistently demonstrated that silencing hsa_circ_0038737
   markedly suppressed cell proliferation, while its overexpression
   significantly enhanced clonogenic capacity in both cell lines
   (Fig. [79]2C-E). To further elucidate the underlying mechanism, we
   performed cell cycle analysis via flow cytometry. Knockdown of
   hsa_circ_0038737 led to a prominent accumulation of cells in the G2
   phase, indicative of a G2/M checkpoint arrest (Fig. [80]2F, G), whereas
   circRNA overexpression decreased the G2-phase population, suggesting
   facilitated progression into mitosis. These results imply that
   hsa_circ_0038737 may regulate tumor growth by accelerating the G2-to-M
   phase transition, a critical control point in cell division.

   Collectively, these data demonstrate that hsa_circ_0038737 functions as
   a positive regulator of PCa cell proliferation through modulation of
   cell cycle progression, independent of its linear host gene GTF3C1.
   This proliferative advantage may contribute to the aggressive clinical
   phenotype observed in tumors with elevated hsa_circ_0038737 expression.
   Given the essential role of cell cycle control in tumorigenesis, these
   findings underscore the potential of hsa_circ_0038737 as a mechanistic
   driver and therapeutic vulnerability in prostate cancer.

Hsa_circ_0038737 directly interacts with IGF2BP3 via KH domains to mediate
functional activity in prostate cancer

   Recent evidence suggests that nuclear-localized circRNAs may exert
   regulatory functions by acting as scaffolds for RNA-binding proteins,
   thereby modulating post-transcriptional gene regulation. To identify
   the potential RBP partners of hsa_circ_0038737, we employed a
   biotinylated RNA pull-down assay using probes specifically designed to
   target the back-splice junction of the circRNA. Silver staining of the
   pull-down complexes followed by mass spectrometry revealed a panel of
   candidate proteins that interact with hsa_circ_0038737, among which
   IGF2BP3 emerged as a prominent interactor (Fig. [81]3A and Table S2).

Fig. 4.

   [82]Fig. 4
   [83]Open in a new tab

   The hsa_circ_0038737-IGF2BP3 Complex Stabilizes DNPH1 mRNA to Promote
   DNA Damage Response Signaling. (A) RNA-seq analysis was performed on
   three pairs of hsa_circ_0038737 knockdown and control 22RV1 cells to
   identify differentially expressed genes. (B) GO enrichment analysis was
   conducted to determine the biological processes in which these
   differentially expressed genes were primarily enriched. (C-D)
   WikiPathways, Reactome, and KEGG pathway enrichment analyses were
   carried out to identify significant enrichment of differentially
   expressed genes in pathways related to DNA repair and DNA damage
   response. (E) Integrative analysis was performed using downregulated
   genes from RNA-seq data, publicly available IGF2BP3 CLIP-seq data, and
   Starbase predictions to identify DNPH1 as a potential target. (F) RIP
   assays were conducted to confirm the specific binding of IGF2BP3 to
   DNPH1 mRNA in 22RV1 and PC-3 cells. (G) The effects of hsa_circ_0038737
   knockdown or overexpression on DNPH1 mRNA levels were assessed. (H)
   Changes in DNPH1 protein levels upon IGF2BP3 knockdown or
   overexpression were observed. (I) The stability of DNPH1 mRNA following
   knockdown of hsa_circ_0038737 or IGF2BP3 in both cell lines was
   evaluated. Data are presented as mean ± SD. *P < 0.05, **P < 0.01,
   ***P < 0.001, ****P < 0.0001

Fig. 3.

   [84]Fig. 3
   [85]Open in a new tab

   Hsa_circ_0038737 Directly Interacts with IGF2BP3 via KH Domains to
   Mediate Functional Activity in Prostate Cancer. (A) RNA-protein
   pull-down experiments were performed using biotin-labeled probes
   targeting the back-splice site of hsa_circ_0038737 in 22RV1 cells.
   Silver staining was conducted to identify proteins that specifically
   interact with hsa_circ_0038737. (B) Gene Ontology, KEGG and Wikipathway
   analysis were carried out to examine the enrichment of these proteins
   in pathways associated with mRNA binding and regulation of mRNA
   stability. (C) Data from RBPsuite, CircNet, CSCD2 databases, and mass
   spectrometry results were integrated to predict IGF2BP3 as a primary
   protein partner of hsa_circ_0038737. (D-E) RNA-FISH and IF
   co-localization assays, and RIP experiments using anti-IGF2BP3
   antibodies were performed to verify the direct binding between
   hsa_circ_0038737 and IGF2BP3 in 22RV1 and PC-3 cells. Data are
   presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001,
   ****P < 0.0001, ns, no significance

   To gain insight into the biological relevance of these interactions, we
   performed Gene Ontology (GO), KEGG, and WikiPathway enrichment
   analyses, which indicated that the identified proteins were
   significantly enriched in pathways associated with mRNA binding, mRNA
   stabilization, and RNA metabolism (Fig. [86]3B). To further prioritize
   functional interactors, we integrated our mass spectrometry data with
   computational predictions from three independent circRNA-RBP
   interaction databases: RBPsuite
   ([87]http://www.csbio.sjtu.edu.cn/bioinf/RBPsuite), a deep-learning
   platform for RBP binding site prediction; CircNet
   ([88]https://awi.cuhk.edu.cn/~CircNet/php/index.php), which maps
   circRNA-mediated regulatory networks; and CSCD2
   ([89]https://geneyun.net/CSCD2/) databases, a cancer-specific circRNA
   database that annotates RBP interaction motifs. All three resources
   convergently predicted IGF2BP3 as a high-confidence binding partner of
   hsa_circ_0038737 (Fig. [90]3C), strongly supporting its functional
   significance.

   To experimentally validate this interaction, we performed RNA-FISH
   combined with immunofluorescence staining. The results revealed
   co-localization of hsa_circ_0038737 and IGF2BP3 in the nuclear
   compartment of both 22RV1 and PC-3 cells (Fig. [91]3D). Additionally,
   RNA pull-down followed by western blot and RIP assays using
   anti-IGF2BP3 antibodies confirmed the physical binding between the
   circRNA and the protein (Fig. [92]3E; Figure S1E). To delineate the
   interaction interface, we generated a series of FLAG-tagged IGF2BP3
   truncation constructs corresponding to distinct RNA-binding domains
   (Figure S2A, B). RIP assays with anti-FLAG antibodies revealed that
   hsa_circ_0038737 predominantly binds to the KH1 and KH2 domains of
   IGF2BP3 (Figure S2C), which are known to mediate sequence-specific RNA
   recognition.

   Together, these results establish hsa_circ_0038737 as a functional
   interactor of IGF2BP3, acting through its KH domains to potentially
   modulate downstream mRNA stability. The circRNA-RBP interaction forms
   the molecular foundation for subsequent regulatory events, laying the
   groundwork for the circRNA’s role in therapeutic resistance and cell
   proliferation in CRPC.

IGF2BP3 functions as a proliferative driver and indicator of poor prognosis
in prostate cancer

   To functionally characterize the role of IGF2BP3, a key RNA-binding
   partner of hsa_circ_0038737, in PCa, we established both gain- and
   loss-of-function models in 22RV1 and PC-3 cell lines.
   Lentiviral-mediated overexpression and knockdown of IGF2BP3 were
   confirmed at the mRNA and protein levels (Figure S3A, B). The CCK-8
   assay revealed that IGF2BP3 knockdown significantly suppressed the
   proliferation rate of both 22RV1 and PC-3 cells, whereas its
   overexpression markedly enhanced cell proliferation compared to
   controls (Figure S3C, D). Consistent with these findings, colony
   formation assays demonstrated that IGF2BP3 depletion substantially
   decreased the number of viable colonies, while its overexpression
   resulted in a dramatic increase in clonogenic capacity in both cell
   lines (Figure S3E, F). These results suggest a negative regulatory role
   of IGF2BP3 in PCa cell proliferation, which is mechanistically
   intriguing given its interaction with hsa_circ_0038737.

   To explore the clinical relevance of IGF2BP3 expression, we performed
   survival analyses using patient data from two large-scale prostate
   cancer cohorts: Stand Up to Cancer (SU2C) and Memorial Sloan Kettering
   Cancer Center (MSKCC). Kaplan–Meier survival curves demonstrated that
   patients with high IGF2BP3 expression exhibited significantly shorter
   overall survival (OS) and disease-free survival (DFS) compared to those
   with low expression levels (Figure S3G, H), underscoring its potential
   prognostic value. Taken together, these data identify IGF2BP3 as a
   context-dependent regulator of prostate cancer cell proliferation and a
   marker of poor clinical outcomes. Although classically described as an
   RNA stabilizer that enhances oncogenic mRNA expression, our findings
   suggest that in the presence of specific molecular regulators, such as
   hsa_circ_0038737-IGF2BP3 may play a dual role, potentially acting as a
   regulatory node within a broader RNA-protein interaction network.

The hsa_circ_0038737-IGF2BP3 complex stabilizes DNPH1 mRNA to promote DNA
damage response signaling

   To elucidate the downstream functional consequences of hsa_circ_0038737
   in prostate cancer cells, we performed RNA-seq on 22RV1 cells following
   knockdown of hsa_circ_0038737, with Matched controls. Differential gene
   expression analysis identified 101 upregulated and 314 downregulated
   genes in the knockdown group (Fig. [93]4A and Table S3). Gene Ontology
   (GO) enrichment revealed that these transcripts were significantly
   associated with mRNA stability regulation and DNA damage response
   pathways (Fig. [94]4B). Moreover, pathway enrichment analyses using
   WikiPathways, Reactome, and KEGG databases consistently highlighted DNA
   repair, DNA damage checkpoint signaling, and genomic stability
   maintenance as the most affected biological processes (Fig. [95]4C, D),
   implicating hsa_circ_0038737 in the modulation of cellular responses to
   genotoxic stress.

Fig. 7.

   [96]Fig. 7
   [97]Open in a new tab

   Hsa_circ_0038737 Promotes Tumor Growth and Mediates PARP Inhibitor
   Resistance via DNPH1 in Vivo. (A-D) Subcutaneous xenograft models were
   established in nude mice to evaluate the effects of hsa_circ_0038737
   overexpression or knockdown on tumor growth and Olaparib resistance,
   with or without DNPH1i co-treatment. Mouse body weight was monitored
   throughout the experiment. (E-G) IHC analysis was performed on
   xenograft tissues to detect DNPH1 and Ki-67 levels in the
   hsa_circ_0038737 overexpression and knockdown groups. (H-I) IHC
   analysis was conducted on patient-derived tissues to examine the
   correlation between IGF2BP3 and DNPH1 expression. (J-K) Patient-derived
   organoids were used to assess Olaparib sensitivity in those with high
   hsa_circ_0038737 expression, with or without DNPH1i treatment. Data are
   presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001,
   ****P < 0.0001, ns, no significance

   Although manipulation of hsa_circ_0038737 expression did not alter
   IGF2BP3 protein levels in either 22RV1 or PC-3 cells (Figure S3I), we
   hypothesized that hsa_circ_0038737 may function as a molecular
   scaffold, facilitating IGF2BP3-mediated stabilization of specific
   target mRNAs. To identify these targets, we employed an integrative
   computational and experimental strategy. Specifically, we intersected
   (1) the downregulated genes from our RNA-seq dataset, (2) publicly
   available IGF2BP3 CLIP-seq data [[98]21], and (3) StarBase-predicted
   mRNA targets of IGF2BP3 (Fig. [99]4E). This analysis pinpointed DNPH1
   as a high-confidence target, an mRNA encoding a nucleotide sanitizer
   enzyme involved in DNA repair and PARPi response modulation [[100]22,
   [101]23].

   To validate this interaction, RIP assays using IGF2BP3 antibodies
   confirmed direct binding between IGF2BP3 and DNPH1 mRNA in both 22RV1
   and PC-3 cells (Fig. [102]4F). Functionally, knockdown of
   hsa_circ_0038737 significantly reduced DNPH1 mRNA levels, whereas
   overexpression of hsa_circ_0038737 led to a marked increase in DNPH1
   abundance (Fig. [103]4G). Parallel experiments showed that IGF2BP3
   knockdown decreased DNPH1 protein expression, while its overexpression
   elevated DNPH1 levels (Fig. [104]4H), consistent with a
   post-transcriptional regulatory mechanism. Finally, to determine
   whether hsa_circ_0038737 and IGF2BP3 cooperatively regulate mRNA
   stability, we conducted mRNA decay assays following transcriptional
   inhibition with actinomycin D. Notably, knockdown of either
   hsa_circ_0038737 or IGF2BP3 substantially decreased the half-life of
   DNPH1 mRNA, confirming their role in stabilizing DNPH1 transcripts
   (Fig. [105]4I).

   Collectively, these findings reveal a novel regulatory axis in which
   hsa_circ_0038737 serves as a functional adaptor that facilitates
   IGF2BP3 binding to DNPH1 mRNA, thereby enhancing its stability. Given
   the established role of DNPH1 in DNA repair and PARPi sensitivity, this
   mechanism provides important insight into how circRNA-RBP complexes
   contribute to therapeutic resistance and highlights a potential
   targetable pathway in castration-resistant prostate cancer.

DNPH1 enhances proliferation and confers resistance to PARP inhibitors in
CRPC cells

   Although DNPH1 has been implicated in nucleotide metabolism and DNA
   repair, its functional role and clinical relevance in CRPC remain
   largely uncharacterized. To investigate its biological impact, we
   established DNPH1-overexpressing and DNPH1-silenced 22RV1 and PC-3 cell
   models. Functional assays demonstrated that DNPH1 overexpression
   significantly enhanced cell proliferation and colony-forming ability,
   while DNPH1 knockdown markedly suppressed tumor cell growth
   (Fig. [106]5A-D), suggesting a positive role in tumor progression. To
   explore the clinical significance of DNPH1 in PCa, we analyzed
   transcriptomic and survival data from two independent patient cohorts.
   In the SU2C cohort, high DNPH1 expression was significantly associated
   with poorer OS (log-rank P = 0.00012, Figure S4A). Similarly, in the
   MSKCC cohort, high DNPH1 expression was associated with markedly
   reduced DFS (log-rank P = 0.037, Figure S4B), highlighting its
   potential utility as a prognostic biomarker.

Fig. 8.

   Fig. 8
   [107]Open in a new tab

   Proposed model of the hsa_circ_0038737/IGF2BP3/DNPH1 regulatory axis in
   mediating PARP inhibitor resistance in CRPC. Hsa_circ_0038737 is
   generated through Alu element-mediated back-splicing, a process
   facilitated by the RNA-binding protein HNRNPDL. Once formed,
   nuclear-localized hsa_circ_0038737 interacts with IGF2BP3 via its KH
   domains, stabilizing DNPH1 mRNA. Elevated DNPH1 expression enhances DNA
   damage repair capacity, thereby reducing the efficacy of PARP
   inhibitors in CRPC

Fig. 5.

   [108]Fig. 5
   [109]Open in a new tab

   DNPH1 Enhances Proliferation and Confers Resistance to PARP Inhibitors
   in Prostate Cancer Cells. (A-D) CCK-8 and colony formation assays were
   performed to evaluate the effects of DNPH1 overexpression or knockdown
   on the proliferation of 22RV1 and PC-3 cells. (E-F) IC50 assays were
   conducted to determine the sensitivity of 22RV1 and PC-3 cells to
   olaparib and niraparib following DNPH1 knockdown or overexpression.
   (G-H) Comet assays were used to assess DNA damage levels in 22RV1 and
   PC-3 cells with DNPH1 knockdown or overexpression, as well as in cells
   treated with the DNPH1 inhibitor N6-benzyladenosine. Data are presented
   as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, ns,
   no significance

   Given the established role of DNPH1 in maintaining nucleotide pool
   balance and facilitating DNA repair, we investigated whether it
   contributes to resistance against PARPi, a class of drugs increasingly
   used in CRPC therapy. IC₅₀ assays showed that DNPH1 knockdown
   significantly increased cellular sensitivity to olaparib and niraparib,
   whereas its overexpression conferred resistance to both agents
   (Fig. [110]5E, F), suggesting that DNPH1 may function as a modulator of
   therapeutic response to PARPi. To further elucidate whether DNPH1
   influences DNA damage levels, we performed alkaline comet assays. DNPH1
   depletion significantly increased DNA strand breaks, while its
   overexpression reduced DNA damage under the same conditions.
   Importantly, pharmacologic inhibition of DNPH1 using N6-benzyladenosine
   phenocopied the effects of gene knockdown, leading to increased DNA
   damage accumulation (Fig. [111]5G, H), and further supporting its role
   in DNA repair facilitation.

   Together, our results demonstrate that DNPH1 promotes PCa cell
   proliferation and plays a critical role in modulating sensitivity to
   PARP inhibition by attenuating DNA damage. These findings provide
   mechanistic insight into how DNPH1 may contribute to therapeutic
   resistance in CRPC and highlight it as a potential predictive biomarker
   and therapeutic target for combinatorial strategies with PARPis.

Hsa_circ_0038737 induces PARP inhibitor resistance through a DNPH1-Dependent
mechanism

   To directly test the hypothesis that hsa_circ_0038737 modulates
   sensitivity to PARPis, we conducted a series of in vitro drug response
   experiments in 22RV1 and PC-3 prostate cancer cell lines. IC₅₀ assays
   revealed that knockdown of hsa_circ_0038737 significantly enhanced
   cellular sensitivity to PARPis, including olaparib and niraparib,
   whereas overexpression of hsa_circ_0038737 conferred marked resistance
   to both agents (Fig. [112]6A, B). Additionally, co-treatment with DNPH1
   inhibitor (DNPH1i) reversed the PARPi resistance induced by
   hsa_circ_0038737 overexpression. DNPH1 knockdown in two prostate cancer
   cells abolished the effect of DNPH1i on proliferation (Figure S4C, D),
   indicating that the compound acts primarily through DNPH1 inhibition
   rather than off-target mechanisms.

Fig. 6.

   [113]Fig. 6
   [114]Open in a new tab

   Hsa_circ_0038737 Confers Resistance to PARP Inhibitors and Can Be
   Functionally Antagonized by DNPH1 Inhibition. (A-B) IC50 assays were
   performed to evaluate the sensitivity of 22RV1 and PC-3 cells to
   olaparib and niraparib following hsa_circ_0038737 knockdown or
   overexpression, with or without co-treatment with DNPH1 inhibitor
   (DNPH1i). (C-D) Comet assays were conducted to assess DNA damage levels
   in 22RV1 and PC-3 cells subjected to hsa_circ_0038737 knockdown or
   overexpression, with or without DNPH1i co-treatment. (E-F) γ-H2AX IF
   assays were carried out to evaluate DNA damage in 22RV1 and PC-3 cells
   with hsa_circ_0038737 knockdown or overexpression, with or without
   DNPH1i co-treatment. Data are presented as mean ± SD. *P < 0.05,
   **P < 0.01, ***P < 0.001, ****P < 0.0001, ns, no significance

   Given our earlier findings that hsa_circ_0038737 stabilizes DNPH1 mRNA
   via IGF2BP3, and the role of DNPH1 in DNA repair, we next examined
   whether pharmacologic inhibition of DNPH1 could rescue the
   drug-resistant phenotype induced by hsa_circ_0038737 overexpression.
   Strikingly, co-treatment with a DNPH1 inhibitor (DNPH1i,
   N6-benzyladenosine) effectively reversed the resistance to PARPi
   conferred by hsa_circ_0038737 overexpression, restoring drug
   sensitivity to near-baseline levels. To investigate the mechanistic
   basis of this effect, we performed alkaline comet assays to quantify
   DNA strand breaks. Knockdown of hsa_circ_0038737 led to a pronounced
   increase in DNA damage, while its overexpression significantly reduced
   DNA damage accumulation. Notably, co-treatment with DNPH1i in
   hsa_circ_0038737-overexpressing cells restored DNA damage levels,
   confirming that DNPH1 activity is functionally required for the DNA
   repair advantage conferred by hsa_circ_0038737 (Fig. [115]6C, D).

   These findings were further validated by γ-H2AX immunofluorescence
   staining, a marker of double-strand DNA breaks. Consistent with the
   comet assay results, γ-H2AX signal intensity increased significantly
   upon hsa_circ_0038737 knockdown, and was reduced upon overexpression.
   Importantly, DNPH1 inhibition in hsa_circ_0038737-overexpressing cells
   reactivated DNA damage, as evidenced by elevated γ-H2AX foci formation
   (Fig. [116]6E, F). In circRNA-overexpressing 22RV1 and PC-3 cells,
   combined DNPH1i/PARPi treatment significantly increased PARP1 retention
   in the chromatin-bound fraction, mirroring the effect of PARPi
   monotherapy in circRNA-knockdown cells (Figure S5A, B). Concurrent PLA
   analysis revealed pronounced γ-H2AX/PARP1 co-localization foci in both:
   (1) circRNA-overexpressing cells treated with DNPH1i + PARPi and (2)
   circRNA-depleted cells receiving PARPi alone (Figure S5C, D),
   indicating amplified DNA damage signaling.

   Collectively, these results reveal that hsa_circ_0038737 plays a
   critical role in modulating the DNA damage response and confers
   cellular resistance to PARP inhibition via a DNPH1-dependent mechanism.
   Therapeutically, this resistance can be effectively overcome through
   pharmacologic targeting of DNPH1, highlighting a potential
   combinatorial strategy to resensitize resistant tumors to PARPi and
   improve treatment outcomes in castration-resistant prostate cancer.

BRCA mutation enhances DNPH1i-Mediated cytotoxicity in PARPi-Resistant cells

   To further investigate the therapeutic potential of DNPH1 inhibition in
   BRCA-mutated contexts, we established BRCA-knockdown models in 22RV1
   and PC-3 cell lines. Western blot analysis confirmed that knock down
   either BRCA1 or BRCA2 did not alter DNPH1 protein expression (Figure
   S6A, B), demonstrating that DNPH1 regulation occurs independently of
   BRCA status. Functionally, in both BRCA1- and BRCA2-deficient models,
   combined PARPi and DNPH1i treatment significantly enhanced suppression
   of cell proliferation compared to PARPi monotherapy (Figure S6C, D).
   This increased sensitivity was consistently observed in both prostate
   cancer cell lines irrespective of whether BRCA1 or BRCA2 was depleted,
   indicating that DNPH1 inhibition can potentiate PARPi efficacy even in
   BRCA-deficient settings. Collectively, these results demonstrate that
   the PARPi/DNPH1i combination remains effective in both BRCA-mutated and
   wild-type models, with DNPH1 modulation operating independently of BRCA
   status, supporting the broader therapeutic implications of our
   findings.

   Hsa_circ_0038737 Promotes Tumor Growth and Mediates PARP Inhibitor
   Resistance via DNPH1 in Vivo.

   To validate the oncogenic role of hsa_circ_0038737 and its impact on
   PARPi sensitivity in vivo, we established subcutaneous xenograft models
   using 4-week-old immunodeficient nude mice. Mice were inoculated with
   22RV1 prostate cancer cells stably expressing either hsa_circ_0038737
   overexpression, knockdown constructs, or control vectors. Ten days
   post-implantation, mice were randomized to receive vehicle control,
   olaparib monotherapy, or olaparib combined with the DNPH1i.

   Tumor growth analysis revealed that overexpression of hsa_circ_0038737
   significantly accelerated tumor progression, whereas circRNA knockdown
   markedly suppressed tumor growth, with no observable difference in body
   weight across groups, indicating minimal systemic toxicity
   (Fig. [117]7A-D). Notably, hsa_circ_0038737-overexpressing tumors
   exhibited resistance to olaparib, while co-treatment with DNPH1i
   restored drug sensitivity, resulting in significant tumor growth
   inhibition. Conversely, xenografts with hsa_circ_0038737 knockdown were
   more responsive to olaparib, further supporting the role of this
   circRNA in modulating PARPi efficacy.

   IHC staining of tumor tissues demonstrated that tumors in the
   overexpression group exhibited elevated levels of DNPH1 and Ki-67,
   consistent with enhanced proliferation and DNA repair capacity. In
   contrast, tumors in the knockdown group displayed markedly reduced
   expression of these markers (Fig. [118]7E-G). Furthermore, to assess
   the clinical relevance of this regulatory axis, we analyzed human
   prostate cancer tissues and found a strong positive correlation between
   IGF2BP3 and DNPH1 expression (R²= 0.5679, P < 0.0001), supporting the
   existence of a conserved regulatory mechanism in patient samples
   (Fig. [119]7H, I).

   To further validate the translational significance of hsa_circ_0038737
   in therapeutic resistance, we utilized patient-derived organoid (PDO)
   models generated from prostate cancer tissues with varying
   hsa_circ_0038737 expression levels (Figure S7 A, B). Organoids derived
   from high hsa_circ_0038737 expressing tumors were resistant to
   olaparib, but this resistance was substantially reversed by
   co-treatment with DNPH1i. In contrast, PDOs from low hsa_circ_0038737
   expressing tumors displayed robust sensitivity to olaparib monotherapy
   (Fig. [120]7J, K). Furthermore, we performed IHC on tumor specimens
   from patients with high versus low hsa_circ_0038737 expression (Figure
   S8A) and observed concordant upregulation or downregulation of HNRNPDL,
   IGF2BP3, and DNPH1, consistent with our proposed regulatory axis
   (Figure S8B-D).

   Taken together, these in vivo and ex vivo findings robustly demonstrate
   that hsa_circ_0038737 promotes prostate tumor growth and mediates
   resistance to PARP inhibition through DNPH1 stabilization. Importantly,
   this resistance is pharmacologically reversible via DNPH1 inhibition,
   offering a promising therapeutic avenue for overcoming circRNA-driven
   treatment resistance in advanced prostate cancer.

Alu elements and HNRNPDL cooperatively mediate the circularization and
biogenesis of hsa_circ_0038737

   The biogenesis of circRNAs is often facilitated by
   reverse-complementary Alu elements located within flanking intronic
   regions, which promote back-splicing through the formation of RNA
   duplex structures [[121]24, [122]25]. To explore whether this mechanism
   underlies the formation of hsa_circ_0038737, we examined the genomic
   architecture of its parental gene GTF3C1 using the UCSC Genome Browser.
   We identified two highly complementary Alu repeats (AluSp and AluSq2)
   positioned in the intronic regions flanking exons 35 and 36, the exonic
   components of hsa_circ_0038737 (Figure S9A).

   To functionally assess the contribution of these Alu elements to
   circRNA formation, we designed and constructed four minigene plasmid
   sequences: #1 Exon 35–36 and full-length flanking introns containing
   both AluSp and AluSq2; #2 Deletion of AluSp only; #3 Deletion of AluSq2
   only; #4 Deletion of both Alu elements (Figure S9B). These constructs
   were transfected into 22RV1 and PC-3 cells, and the circularization
   efficiency of hsa_circ_0038737 was assessed via qRT-PCR. The results
   revealed that deletion of either Alu element significantly impaired
   circRNA formation, and removal of both elements resulted in the most
   pronounced reduction (Figure S9C), confirming AluSp and AluSq2 are
   essential cis-regulatory elements for the efficient circularization of
   hsa_circ_0038737.

   Given emerging evidence that RBPs can facilitate circRNA biogenesis by
   interacting with Alu elements [[123]25, [124]26], we utilized RBPmap to
   predict candidate RBPs that may bind to the Alu-containing intronic
   regions of hsa_circ_0038737. Among the candidates, HNRNPDL emerged as
   the only RBP predicted to bind both AluSp and AluSq2 simultaneously
   (Figure S9D). To test its functional involvement, we transfected PCa
   cells with shRNA targeting HNRNPDL, and observed a significant
   reduction in hsa_circ_0038737 expression, indicating that HNRNPDL
   likely facilitates its biogenesis (Figure S9E, F). Interestingly, DNPH1
   expression was also downregulated upon HNRNPDL silencing, suggesting a
   regulatory cascade whereby HNRNPDL promotes circRNA formation, which in
   turn stabilizes DNPH1 mRNA via IGF2BP3.

   To further support this regulatory link, we analyzed RNA-seq data from
   multiple prostate cancer cohorts. A statistically significant positive
   correlation between HNRNPDL and DNPH1 mRNA levels was observed (Figure
   [125]S9G), consistent with the hypothesis that HNRNPDL indirectly
   controls DNPH1 expression through modulation of circRNA biogenesis.
   Overall, our data identify AluSp and AluSq2 as critical intronic
   elements and HNRNPDL as a key trans-acting factor that cooperatively
   facilitate the circularization of hsa_circ_0038737. This, in turn,
   impacts downstream gene regulation via the
   hsa_circ_0038737-IGF2BP3-DNPH1 axis, highlighting an additional layer
   of post-transcriptional control with therapeutic implications in
   prostate cancer.

Comprehensive validation of MYC-Mediated transcriptional control of HNRNPDL

   Through integrative bioinformatic analyses utilizing six independent
   platforms (FIMO_JASPAR, PWMEnrich_JASPAR, ChIP-Atlas, ENCODE, CHEA, and
   GTRD), we identified MYC as the highest-ranking transcription factor
   candidate regulating HNRNPDL expression (Figure S10A). Cross-platform
   validation identified evolutionarily conserved MYC binding sites within
   the proximal promoter region of HNRNPDL (Figure S10B), with the
   canonical MYC motif < CACGTG > detected as the predominant binding
   signature (Figure [126]S10C). To experimentally validate this
   regulatory interaction, we performed chromatin immunoprecipitation
   quantitative PCR (ChIP-qPCR), which confirmed a significant enrichment
   of MYC occupancy at the HNRNPDL promoter compared to IgG controls in
   both 22Rv1 and PC-3 cell lines (Figure S10D). Functionally, MYC
   knockdown led to a Marked reduction in HNRNPDL protein levels in 22Rv1
   and PC-3 cells (Figure S10E, F), while MYC overexpression resulted in a
   corresponding increase in HNRNPDL expression. These findings establish
   a bidirectional regulatory relationship, demonstrating that MYC acts as
   a direct transcriptional activator of HNRNPDL, thereby contributing to
   prostate cancer pathogenesis.

Discussion

   Our study reveals hsa_circ_0038737 as a critical oncogenic circRNA in
   prostate cancer that exerts its effects by forming a functional complex
   with IGF2BP3 to stabilize DNPH1 mRNA, thereby enhancing DNA repair
   capacity and promoting resistance to PARP inhibitors. The biogenesis of
   hsa_circ_0038737 is tightly controlled by Alu element-mediated
   circularization and the RNA-binding protein HNRNPDL, adding a new layer
   to our understanding of circRNA regulation. Importantly, the
   pharmacologic inhibition of DNPH1 effectively reverses
   hsa_circ_0038737-induced PARPi resistance in both cell line-based
   models and patient-derived organoids model (Fig. [127]8). Our findings
   uncover a novel circRNA/RBP-mediated regulatory axis,
   hsa_circ_0038737/IGF2BP3/DNPH1, that promotes CRPC progression and
   therapeutic resistance, offering new insights into a rationale for
   combinatorial treatment strategies and identifying a potential
   combinatorial therapeutic strategy to overcome PARPi resistance.

   Recent studies have increasingly emphasized the functional versatility
   of circRNAs in cancer progression, particularly their ability to act as
   scaffolds for RBPs to regulate mRNA stability and translation
   [[128]18]. For instance, circITGB6 was recently shown to enhance
   cisplatin resistance in ovarian cancer by binding to IGF2BP2 and
   stabilizing FGF9 mRNA [[129]27]. Similarly, our work demonstrates that
   hsa_circ_0038737 interacts with IGF2BP3 to stabilize DNPH1, a
   nucleotide sanitizer critical for maintaining genomic integrity under
   PARPi-induced DNA damage. These parallels underscore a conserved
   mechanism by which circRNAs exploit RBP partnerships to drive therapy
   resistance across malignancies.

   First, we identified hsa_circ_0038737 as a circRNA significantly
   upregulated in PCa tissues and associated with aggressive
   clinicopathological features, including higher T stage and Gleason
   scores. Its nuclear localization and resistance to RNase R digestion
   align with previous reports demonstrating the stability and functional
   versatility of circRNAs in cancer progression [[130]28]. Notably,
   hsa_circ_0038737 promoted PCa cell proliferation and G2/M phase
   progression, independent of its linear counterpart GTF3C1. These
   observations reinforce the emerging paradigm that circRNAs can exert
   oncogenic effects distinct from their parental genes [[131]29]. Recent
   work by Cheng et al. further supports this notion, showing that
   METTL3-mediated m6A modification of circGLIS3 drives prostate cancer
   progression by sequestering miR-661 and activating the MDM2-p53 axis
   [[132]19]. Our findings expand this landscape by linking circRNA
   activity to PARPi resistance through DNPH1-mediated DNA repair, a
   pathway not previously associated with circRNA regulation.

   Mechanistically, we revealed that hsa_circ_0038737 interacts with
   IGF2BP3, an RBP known to enhance mRNA stability [[133]21]. This
   interaction stabilizes DNPH1 mRNA, a nucleotide sanitizer implicated in
   DNA repair and PARPi resistance [[134]22]. Our RNA-seq and functional
   assays demonstrated that DNPH1 upregulation reduces DNA damage and
   confers resistance to olaparib and niraparib, consistent with prior
   studies linking DNPH1 to PARPi sensitivity [[135]23]. Importantly, the
   hsa_circ_0038737/IGF2BP3 axis represents a previously unrecognized
   regulatory layer in PARPi resistance, expanding the repertoire of
   circRNA-mediated mechanisms in therapy evasion [[136]14, [137]30]. The
   reversal of PARPi resistance by DNPH1 inhibition underscores the
   therapeutic potential of targeting this pathway. Recent advancements in
   DNPH1 inhibitor development, such as N6-benzyladenosine derivatives,
   have shown promise in sensitizing BRCA-deficient tumors to PARPi in
   preclinical models [[138]22], aligning with our findings and suggesting
   translational relevance for combinatorial therapies in CRPC.

   Furthermore, we identified Alu elements flanking hsa_circ_0038737 as
   critical drivers of its biogenesis, with HNRNPDL facilitating this
   process. This aligns with established models of circRNA formation
   involving complementary Alu sequences and RBPs [[139]31, [140]32]. The
   correlation between HNRNPDL and DNPH1 expression in clinical cohorts
   suggests a broader regulatory network Linking circRNA biogenesis to DNA
   repair pathways. These findings deepen our understanding of circRNA
   generation and its functional implications in cancer. A 2023 study by
   Rzechorzek et al. further elucidated the enzymatic mechanism of DNPH1
   in hydrolyzing damaged nucleotides, providing structural insights into
   its role in mitigating PARPi-induced DNA lesions [[141]22]. Our work
   complements these discoveries by positioning DNPH1 within a
   circRNA-driven regulatory circuit, highlighting the interplay between
   non-coding RNA biogenesis and enzymatic DNA repair processes.

   Clinically, our in vivo xenograft and patient-derived organoid models
   demonstrated that hsa_circ_0038737 overexpression promotes tumor growth
   and PARPi resistance, while DNPH1 inhibition restores drug sensitivity.
   The positive correlation between IGF2BP3 and DNPH1 in patient tissues
   further supports the translational relevance of this axis. These data
   suggest that hsa_circ_0038737 expression could serve as a biomarker for
   predicting PARPi response, while combinatorial targeting of DNPH1 may
   enhance PARPi efficacy in CRPC patients. Recent clinical trials, such
   as the PROfound study [[142]33], have established PARPi as a standard
   therapy for HRR-deficient mCRPC; however, resistance remains a major
   limitation. Our findings propose a strategy to extend PARPi utility by
   co-targeting DNPH1, particularly in tumors with elevated
   hsa_circ_0038737 expression.

   However, this study has limitations. The sample size of clinical
   tissues and organoids remains modest, necessitating validation in
   larger cohorts. Additionally, the functional role of hsa_circ_0038737
   in other types of cancer and its potential as a universal biomarker for
   cancer progression need to be explored in future studies. Recent
   advances in spatial transcriptomics and single-cell sequencing could
   further resolve the heterogeneity of circRNA expression within tumor
   microenvironments [[143]18], offering insights into context-dependent
   resistance mechanisms. Finally, while DNPH1 inhibitors showed efficacy
   in preclinical models, their pharmacokinetic and safety profiles in
   humans require further evaluation. These limitations, however, do not
   diminish the core findings but highlight directions for future
   research.

Conclusion

   In conclusion, this study identifies a novel
   hsa_circ_0038737–IGF2BP3–DNPH1 regulatory axis as a critical driver of
   PARP inhibitor resistance in castration-resistant prostate cancer
   (CRPC). By elucidating how circRNA-mediated stabilization of DNA
   repair–related transcripts contribute to therapeutic escape, our
   findings expand the current understanding of non-coding RNA function in
   drug resistance. The demonstrated reversal of resistance via DNPH1
   inhibition highlights a promising avenue for combination therapies
   targeting circRNA-protein interactions. These insights lay the
   groundwork for future translational studies aimed at optimizing
   DNPH1-targeted compounds and validating their clinical utility in
   enhancing PARPi efficacy in advanced prostate cancer.

Supplementary Information

   [144]Supplementary Material 1.^ (192.5KB, pdf)
   [145]12943_2025_2447_MOESM2_ESM.pdf^ (135.8KB, pdf)

   Supplementary Material 2: Mapping of the binding domains of IGF2BP3 for
   hsa_circ_0038737. (A-B) Schematic diagrams illustrate the constructed
   plasmids containing different fragments of FLAG-tagged IGF2BP3. (C) RIP
   experiments were performed using FLAG antibody to identify the domains
   of IGF2BP3 that interact with hsa_circ_0038737 in 22RV1 and PC-3 cells.
   Data are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001,
   ****P < 0.0001, ns, no significance.
   [146]12943_2025_2447_MOESM3_ESM.pdf^ (342.2KB, pdf)

   Supplementary Material 3: IGF2BP3 function and prognosis in prostate
   cancer. (A-B) IGF2BP3 was knocked down and overexpressed in 22RV1 and
   PC-3 cell lines. (C-D) CCK-8 assays were performed to evaluate the
   effects of IGF2BP3 modulation on cell proliferation in 22RV1 and PC-3
   cells. (E-F) Colony formation assays were conducted to assess the
   impact of IGF2BP3 modulation on clonogenic survival in 22RV1 and PC-3
   cells, with corresponding quantification of colony numbers. (G-H)
   Kaplan-Meier survival analysis was conducted using data from the SU2C
   and MSKCC cohorts to assess the relationship between IGF2BP3 expression
   levels and overall survival (OS) or disease-free survival (DFS). (I)
   Western Blotting was conducted to verify whether the expression changes
   of hsa_circ_0038737 affect the protein levels of IGF2BP3
   [147]Supplementary Material 4.^ (140.3KB, pdf)
   [148]12943_2025_2447_MOESM5_ESM.pdf^ (302.3KB, pdf)

   Supplementary Material 5 : IGF2BP3 function and prognosis in prostate
   cancer. (A-B) IGF2BP3 was knocked down and overexpressed in 22RV1 and
   PC-3 cell lines. (C-D) CCK-8 assays were performed to evaluate the
   effects of IGF2BP3 modulation on cell proliferation in 22RV1 and PC-3
   cells. (E-F) Colony formation assays were conducted to assess the
   impact of IGF2BP3 modulation on clonogenic survival in 22RV1 and PC-3
   cells, with corresponding quantification of colony numbers. (G-H)
   Kaplan-Meier survival analysis was conducted using data from the SU2C
   and MSKCC cohorts to assess the relationship between IGF2BP3 expression
   levels and overall survival (OS) or disease-free survival (DFS). (I)
   Western Blotting was conducted to verify whether the expression changes
   of hsa_circ_0038737 affect the protein levels of IGF2BP3IGF2BP3
   function and prognosis in prostate cancer. (A-B) IGF2BP3 was knocked
   down and overexpressed in 22RV1 and PC-3 cell lines. (C-D) CCK-8 assays
   were performed to evaluate the effects of IGF2BP3 modulation on cell
   proliferation in 22RV1 and PC-3 cells. (E-F) Colony formation assays
   were conducted to assess the impact of IGF2BP3 modulation on clonogenic
   survival in 22RV1 and PC-3 cells, with corresponding quantification of
   colony numbers. (G-H) Kaplan-Meier survival analysis was conducted
   using data from the SU2C and MSKCC cohorts to assess the relationship
   between IGF2BP3 expression levels and overall survival (OS) or
   disease-free survival (DFS). (I) Western Blotting was conducted to
   verify whether the expression changes of hsa_circ_0038737 affect the
   protein levels of IGF2BP3
   [149]12943_2025_2447_MOESM6_ESM.pdf^ (1MB, pdf)

   Supplementary Material 6: Molecular and functional analyses in
   BRCA-deficient prostate cancer models. (A-B) Western blot analysis of
   DNPH1 protein expression in (A) BRCA1-knockdown and (B) BRCA2-knockdown
   22RV1 and PC-3 cells using two independent shRNAs (ShBRCA1-1/ShBRCA1-2;
   ShBRCA2-1/ShBRCA2-2) with non-targeting shRNA control (NC); GAPDH
   serves as loading control. (C-D) Cell proliferation assay in
   BRCA1-knockdown 22RV1 and PC-3 cells under four conditions: NC,
   BRCA1-targeting shRNA (ShBRCA1), ShBRCA1 with PARP inhibitor (PARPi),
   and ShBRCA1 with both PARPi and DNPH1 inhibitor (DNPH1i)
   [150]12943_2025_2447_MOESM7_ESM.pdf^ (5.5MB, pdf)

   Supplementary Material 7: Immunohistochemical analysis of AR and PSMA
   in patient-derived organoids from CRPC patients. (A-B)
   Immunohistochemical staining was performed to detect AR and PSMA
   expression in patient-derived organoids established from CRPC tissues
   [151]12943_2025_2447_MOESM8_ESM.pdf^ (3.5MB, pdf)

   Supplementary Material 8: Immunohistochemical assessment of protein
   expression in prostate cancer tissues stratified by Hsa_circ_0038737
   levels. (A) Representative immunohistochemistry staining images of
   DNPH1, HNRNPDL, and IGF2BP3 in CRPC tissues with low or high
   Hsa_circ_0038737 expression. (B-D) Immunoreactive scores of (B)
   HNRNPDL, (C) IGF2BP3, and (D) DNPH1 protein expression. Data are
   presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001,
   ****P < 0.0001, ns, no significance
   [152]12943_2025_2447_MOESM9_ESM.pdf^ (1MB, pdf)

   Supplementary Material 9: Role of Alu elements and HNRNPDL in
   hsa_circ_0038737 biogenesis. (A) The UCSC Genome Browser was used to
   identify two highly reverse-complementary Alu elements (AluSp and
   AluSq2) in the flanking introns of hsa_circ_0038737. (B) Four plasmid
   constructs were generated: #1 with both Alu elements, #2 lacking AluSp,
   #3 lacking AluSq2, and #4 lacking both elements, with an empty vector
   as a control. (C) qRT-PCR analysis was performed after transfecting
   these plasmids into 22RV1 and PC-3 cells to measure the circularization
   efficiency of hsa_circ_0038737. (D) RBPmap prediction was conducted to
   identify RBPs binding to both Alu elements. (E-F) HNRNPDL was knocked
   down using shRNA, and the effects on hsa_circ_0038737 and DNPH1 levels
   were assessed. (G) RNA-seq data from multiple prostate cancer cohorts
   were analyzed to examine the correlation between HNRNPDL and DNPH1 mRNA
   levels. Data are presented as mean ± SD. *P < 0.05, **P < 0.01,
   ***P < 0.001, ****P < 0.0001
   [153]12943_2025_2447_MOESM10_ESM.pdf^ (880.8KB, pdf)

   Supplementary Material 10: Analysis of MYC-mediated transcriptional
   regulation of HNRNPDL. (A) Bioinformatics prediction of MYC binding
   sites within the HNRNPDL promoter region using integrated public
   ChIP-seq datasets. (B) Publicly available MYC ChIP-seq peak profile at
   the HNRNPDL promoter region. (C) Consensus MYC binding motif identified
   from ChIP-seq data. (D) Chromatin immunoprecipitation quantitative PCR
   validation of MYC occupancy at the HNRNPDL promoter in 22RV1 and PC-3
   prostate cancer cells; data presented as fold enrichment relative to
   input DNA. (E-F) Western blot analysis of HNRNPDL protein expression in
   (E) 22RV1 and (F) PC-3 cells under MYC modulation: non-targeting shRNA
   control (NC), MYC knockdown with two independent shRNAs (MYC Sh1/Sh2),
   empty vector control (Vector), and MYC overexpression (MYC OE); GAPDH
   serves as loading control
   [154]Supplementary Material 11.^ (11.5KB, docx)
   [155]Supplementary Material 12.^ (23.5KB, xlsx)
   [156]Supplementary Material 13.^ (2.2MB, xlsx)

Acknowledgements