PDIA1 promotes androgen receptor activation and prostate cancer cell survival through enhancing HMMR stability

Abstract

To sustain rapid proliferation, cancer cells increase protein synthesis, intensifying reliance on protein disulfide isomerase A1 (PDIA1). It is largely unknown whether disulfide bond formation of PDIA1 substrates is driven by one or both CGHC motifs. Using active‐site trapping mutants in prostate cancer cells combined with mass spectrometry, we identified 29 proteins uniquely bound to the C₅₃GHC₅₆ domain and 20 proteins uniquely bound to the C₃₉₇GHC₄₀₀ domain. Hyaluronan‐mediated motility receptor (HMMR) was validated as a PDIA1 C₃₉₇GHC₄₀₀‐specific substrate, with PDIA1 catalysing disulfide bond formation between Cys242 and Cys293. PDIA1 knockdown induced HMMR ubiquitination, blocked androgen receptor nuclear translocation, and suppressed prostate cancer cell growth, survival, and migration. These findings reveal a previously unknown role of PDIA1 in prostate cancer biology.

Using active‐site trapping mutants in prostate cancer cells, we identified HMMR as a specific substrate of the PDIA1 C₃₉₇GHC₄₀₀ domain. PDIA1 catalyses disulfide bond formation between Cys242 and Cys293 of HMMR, preventing its ubiquitination. This interaction promotes androgen receptor nuclear translocation and supports prostate cancer cell proliferation, survival, and migration, revealing a novel PDIA1 function.

Abbreviations

4‐MU, 4‐methylumbelliferone

ADT, androgen deprivation therapy

AR, androgen receptor

CRPC, castration‐resistant prostate cancer

DHT, dihydrotestosterone

ER, endoplasmic reticulum

HMMR, hyaluronan‐mediated motility receptor

PDIA1, protein disulfide isomerase A1

UPR, unfolded protein response

Prostate cancer is one of the most common cancers in men worldwide. In 2022, there were an estimated 1.46 million new cases and approximately 394 200 deaths globally [1]. While most cases are androgen‐dependent at diagnosis and initially respond to androgen deprivation therapy (ADT), about 20–30% progress to castration‐resistant prostate cancer (CRPC), which is associated with a significantly poorer prognosis [2]. The androgen receptor (AR) is a central driver of disease progression. Binding of androgens like testosterone triggers AR nuclear translocation, dimerization, and regulation of target genes controlling proliferation [3, 4, 5]. AR signalling remains active in nearly all prostate cancers, including CRPC. Cellular alterations in prostate cancer, such as PTEN loss or MYC overexpression, can increase protein synthesis, leading to endoplasmic reticulum (ER) stress and subsequently, unfolded protein response (UPR). Cancer cells rely on UPR to cope with any proteostatic imbalance [6], and a key aspect of this adaptation is the induction of ER chaperone proteins.

Protein disulfide isomerases (PDI) are a group of ER chaperones that facilitate the oxidation, reduction, or isomerization of thiol groups on cysteine residues within unfolded or misfolded polypeptides [7]. Deficiency of PDI or PDI activity may lead to the accumulation of misfolded proteins in the ER, which are then retro‐translocated to the cytosol and subsequently degraded by the proteasome or the autophagosome systems, resulting in the activation of UPR and ultimately cell death [8]. Dysregulation of PDI expression or their functions plays important roles in the onset and progression of various human illnesses, such as cardiovascular diseases [9, 10], diabetes [11, 12] and cancer [13, 14].

The PDI family has 21 family members and each one interacts with different substrates to assist their folding. PDIA1, one of the most abundant PDIs in human cancer cells [6, 15], has been associated with tumour growth, metastasis, therapy resistance and poor prognosis in cancer patients [13, 14]. Hence, the development of PDIA1 inhibitors for cancer treatment has become a hot topic of attention in the past two decades [16, 17, 18]. The common thread among all PDIs is the presence of at least one thioredoxin‐like domain, either active (undergoing shuffling of disulfide bonds) or inactive [19]; the conventional consensus for the redox active site is comprised of cysteine‐X‐X‐cysteine (CXXC, where X can be any amino acid residue), such as C₅₃GHC₅₆ and C₃₉₇GHC₄₀₀ in PDIA1. The N‐terminal cysteine within the CGHC motif plays a role in forming disulfide bonds with the corresponding protein substrate; whereas the C‐terminal cysteine acts to release of the substrate. During the reduction of disulfide bonds, a mixed disulfide must form between the substrate and the PDI. Therefore, PDIA1 mutants that are missing only the resolving C‐terminal cysteines at the CGHC sites (cysteine (C) mutated to serine (S), i.e., CGHS) have been used as ‘trapping’ mutants to discover PDIA1 client substrates [6, 20, 21].

Although a great effort has been dedicated to elucidating potential PDIA1 substrates, it is yet largely unknown whether the disulfide bond formation of these substrates is mainly driven by one of the CGHC motifs or both. In this study, using PDIA1 substrate‐trapping mutants (‐C56S and ‐C400S) and immunoprecipitation‐coupled mass spectrometry (IP‐MS, Table S1), we identified various C₅₃GHC₅₆‐ or/and C₃₉₇GHC₄₀₀‐specific protein clients, such as hyaluronan‐mediated motility receptor (HMMR). Knocking down or pharmacological inhibition of PDIA1 destabilized HMMR, prevented AR nuclear import and hence activation in response to stimuli, which further supported our recent notion that PDIA1 plays a pivotal role in prostate cancer cell growth and viability via regulating the stability of its protein substrates [6].

Materials and methods

Cell culture

All cell lines were obtained from the American Type Culture Collection (LNCaP, RRID:CVCL_0395; C4‐2B, RRID:CVCL_4784; 22RV1, RRID:CVCL_1045; 293T, RRID:CVCL_0063). LNCaP, C4‐2B and 22RV1 cells were maintained in RPMI‐1640 medium (Gibco 6125565) supplemented with 10% (v/v) fetal bovine serum (FBS) and 1× penicillin/streptomycin (Gibco 15140122). 293T cells were cultured in Dulbecco's Modified Eagle Medium (DMEM Gibco 6125458) containing 10% (v/v) FBS and 1× penicillin/streptomycin. For androgen deprivation experiments, the medium was replaced with phenol red‐free RPMI‐1640 containing 10% (v/v) dextran‐coated charcoal (DCC)‐stripped serum (cat. no.: CSF15‐0212; Moocow, Guangzhou, Guangdong, China). All experiments were conducted with cells within 20 passages of thawing. Regular mycoplasma testing was performed using a PCR‐based method.

Cell growth and apoptosis assays

Cell growth and viability were assessed for LNCaP, 22RV1 and C4‐2B cell lines. Cells were seeded in biological triplicates into 24‐well plates at a density of 7 × 10⁴ cells per well. Viable cell counts were determined by manual Trypan blue exclusion (cat. no.: E607320; Sangon Biotech, Shanghai, China). Cell apoptosis was measured by the Incucyte S3 live‐cell imaging and analysis platform using SYTOX™ Green dye (cat. no.: S7020; Thermo Fisher, Waltham, MA, USA).

Cell line transfection

For siRNA‐mediated knockdown, PDIA1 was targeted using siRNAs from Merck Sigma‐Aldrich (Table S3). A nontargeting siRNA (siNC, cat no.: A101225; Sangon Biotech) was served as control. siRNA amounts were scaled as follows: 450 pmol for a 10 cm dish (3 × 10⁶ cells), 150 pmol for a 6‐well plate (3.5 × 10⁵ cells), and 40 pmol for a 24‐well plate (7 × 10⁴ cells).

pcDNA3.1‐PDIA1‐V5 constructs were kindly provided by Dr. Sonam Parakh and Dr. Julie Atkin (Macquarie University, Australia). Plasmids encoding full length and truncated HMMR were synthesized by GENTLEGEN in the pcDNA3.1 backbone. All plasmid transfections were carried out using GP‐transfect‐Mate (cat. no.: G04008; GenPharma, Shanghai, China).

Chemicals and reagents

All chemicals and drugs were purchased from MedChem Express (Monmouth Junction, NJ, USA) unless otherwise stated. Stock solutions were prepared in DMSO.

Cell lysis and protein quantification

Following treatments, cells were lysed on ice with a buffer containing 1% (v/v) Triton X‐100, 50 mm HEPES (pH 7.5), 150 mm NaCl, 10 mm CaCl₂, 20 mm N‐Ethylmaleimide and protease inhibitor cocktail (cat no.: RM02916; ABclonal, Wuhan, Hubei, China). Lysates were centrifuged at 14000 × g for 10 min at 4 °C. The resulting supernatants were collected, and total protein concentration was determined using the BCA assay (cat. no.: ZJ101; Epizyme, Shanghai, China).

Cell migration assay

C4‐2B and 22RV1 cells were seeded into 6‐well plates at a density of 8 × 10⁵ cells per well. After 24 h of incubation, a linear wound was created in the confluent monolayer by scratching with a sterile 200‐μL pipette tip. The cells were then washed gently to remove detached cells and subsequently treated with either vehicle control (DMSO) or 0.4 mm 4‐MU in fresh culture medium containing 0.5% (v/v) serum. Wound closure was monitored at the same designated areas and photographed every 24 h postscratching using a microscope equipped with a 4× objective lens.

Immunoprecipitation

Protein A/G immune precipitation magnetic beads (cat. no.: B23201; Selleck Chem, Houston, TX, USA) were conjugated with the respective antibody (0.1 μg per sample) by incubation for 1 h at room temperature with constant rotation. Antibody‐bound beads were then incubated with clarified cell lysates for 2 h at 4 °C with rotation. Beads were washed twice with lysis buffer, and bound proteins were eluted in 2 × Laemmli sample buffer for subsequent analysis by western blotting or mass spectrometry.

Immunoprecipitation‐ mass spectrometry (MS)

Sample preparation

Immunoprecipitated samples on beads were washed thrice with phosphate‐buffered saline (PBS) and digested with trypsin in 100 mm ammonium bicarbonate at 25 °C for 2 h. Samples were subsequently reduced (5 mm dithiothreitol, 30 min), alkylated (10 mM iodoacetamide, 30 min in darkness), and subjected to a second tryptic digestion for 1 h at 25 °C. The digestion was terminated by acidification with trifluoroacetic acid, and peptides were desalted using homemade C18 StageTips.

LC–MS/MS analysis

Peptides were separated on a 250 mm × 75 μm Acclaim PepMap100 C18 column using a 90‐min gradient with mobile phases A (0.1% formic acid in water) and B (0.1% formic acid in 80% acetonitrile) at a flow rate of 300 nL/min. Analysis was performed on an Orbitrap Eclipse mass spectrometer coupled to a Vanquish Neo UHPLC system, operating in data‐dependent acquisition mode with FAIMS compensation voltages of −45 V and −65 V. MS1 spectra were acquired at 60000 resolution, and MS2 spectra at 15000 resolution.

Data processing

Raw data were processed using Proteome Discoverer (v2.5) and searched against the UniProt human database (version October 2022) with SEQUEST HT. Search parameters included as follows: carbamidomethylation as a fixed modification; methionine oxidation and N‐terminal acetylation as variable modifications; precursor mass tolerance of 10 ppm; fragment mass tolerance of 0.02 Da; and a maximum of two missed cleavages. A 1% false discovery rate threshold was applied at peptide and protein levels. Differential expression analysis was performed with the Deseq2 package. Gene Ontology (GO) enrichment and protein–protein interaction analyses were conducted using clusterProfiler and STRING, respectively.

Western blot analysis

Proteins were denatured in Laemmli sample buffer at 100 °C for 3 min, separated by SDS/PAGE, and transferred to nitrocellulose membranes. Membranes were blocked with 5% (w/v) skimmed milk in Tris‐buffered saline with Tween‐20 for 1 h at room temperature, followed by incubation with primary antibodies overnight at 4 °C. After washing, membranes were incubated with fluorescently labelled secondary antibodies for 1 h at room temperature. Signals were detected using an Odyssey CLx imaging system. Antibody specific information includes AR (cat no.: 5153 T, RRID:AB_10691711; Cell Signalling Technology, Danvers, MA, USA), prostate‐specific antigen (PSA, cat. no.: 5365, RRID:AB_2797609; Cell Signalling Technology), HMMR (Proteintech Group, Wuhan, Hubei, China), PDIA1 (cat. no.: A19239; RRID:AB_3719405; ABclonal), β‐actin (cat. no.: 66009‐1‐Ig; RRID:AB_2687938; Proteintech Group), FLAG (cat. no.: 80801‐2‐RR; RRID:AB_3670488; Proteintech Group), V5 (cat. no.: AE089; RRID:AB_3719406; ABclonal), Ubiquitin (cat. no.: A19686; RRID:AB_2862735; ABclonal) and Lamin B1 (cat. no.: 66095‐1‐Ig; RRID:AB_11232208; Proteintech Group).

Site‐directed mutagenesis

Point mutations were introduced into plasmid constructs using PCR‐based mutagenesis with Pfu polymerase.

Immunofluorescence

LNCaP cells were plated in 20 mm diameter confocal dishes at 5 × 10⁵ cells per well and cultured in DCC‐stripped serum medium for 72 h. Cells were then further treated with 5 μm CCF642 or 0.4 mm 4‐methlumbelliferone (4‐MU) for 24 h, followed by stimulation with 10 nm dihydrotestosterone (DHT) or vehicle for 4 h. After removal of culture medium, cells were washed thrice with PBS. 200 μL of immunostaining fixation solution (cat. no.: P0098; Beyotime, Shanghai, China) was added to each well and incubated at room temperature for 15 min. Following fixation, the cells were washed thrice with PBS and permeabilized with 0.1% (v/v) Triton X‐100 in PBS for 10 min at room temperature. After another three PBS washes, the cells were blocked with 3% (m/v) BSA for 1 h at room temperature. The anti‐AR antibody (cat. no.: 5153 T, RRID:AB_10691711; Cell Signalling Technology) was then applied at a dilution of 1:2000 (v/v) and incubated overnight at 4 °C. On the next day, the primary antibody was removed, and cells were washed thrice with PBS. A secondary antibody, CoraLite594‐conjugated Goat Anti‐Rabbit IgG (H + L) (cat. no.: SA00013‐4; RRID:AB_2810984; Proteintech Group), diluted at 1:200 (v/v), was then added and the samples were further incubated for 1 h at room temperature. Following three additional PBS washes, the nuclei were stained with DAPI for 10 min at room temperature. Slides were imaged on a NIKON AXE N‐STORM, and protein abundance was quantified using Image J version 1.8.0 (RRID:SCR_003070).

Statistical analyses

All experiments were repeated independently for at least three times. Data were analysed using graphpad prism version 10.4.1 (RRID:SCR_002798). Specific statistical tests and measures of variation are indicated in the figure legends. False discovery rate <0.05 was defined as statistically significant for the MS data.

Results

Identification of PDIA1 C₅₃GHC₅₆ ‐ or C₃₉₇GHC₄₀₀ ‐specific protein substrates

In a recent study, we reported that PDIA1 and PDIA5 regulate protein stability of the AR in prostate cancer cells and hence promote cancer cell growth [6]. To fully unravel PDIA1 substrates in prostate cancer cells, we first transfected LNCaP cells, an androgen‐responsive prostate cancer cell line model, with pcDNA3.1 empty vector, or the same vector but expressing wild‐type (WT), C56S or C400S mutants of the C‐terminal V5‐tagged PDIA1 (PDIA1‐V5; Fig. 1A,B). We then immunoprecipitated PDIA1‐V5 from the cell lysates with a V5‐specific antibody and analysed the immunoprecipitates with liquid chromatography tandem mass spectrometry (LC‐MS/MS). MS data were normalized prior to conducting subsequent analyses (Fig. S1). IP‐MS discovered more than 3000 proteins associated with PDIA1, most of which bound to all three (WT, C56S and C400S) versions of PDIA1 (Fig. 1C–E). Proteins bound to PDIA1 are mainly implicated in the dynamic changes of DNA, mRNA translation and protein folding (Fig. 1F–H). Notably, we identified 29 proteins that were specifically bound to PDIA1^C56S‐V5; 20 proteins that were specifically bound to PDIA1^C400S‐V5 and 37 proteins that were associated with both PDIA1 ^C56S‐V5 and PDIA1^C400S‐V5 (Fig. 1I, Table S2). Leveraging our prior dataset, we found that PDIA1 knockdown in LNCaP cells significantly reduced the expression of several proteins enriched in PDIA1 trapping mutant immunoprecipitates (Fig. S2), which likely reflects the fact that the stability of newly synthesized polypeptides highly relies on protein disulfide bond formation.

Fig. 1.

Identification of PDIA1‐C₅₃GHC₅₆‐ and C₃₉₇GHC₄₀₀‐specific substrates. (A) Schematic workflow of the substrate‐trapping strategy using PDIA1 catalytic mutants. (B) LNCaP cells were treated with empty vector (NC: negative control) or C‐terminal V5‐tagged PDIA1‐WT, PDIA1‐C56S, PDIA1‐C400S for 72 h, followed by immunoprecipitation (IP) with a V5 antibody and MS analysis of the immunoprecipitates. (C–E) Volcano plots displaying proteins enriched in the PDIA1‐WT (C), PDIA1‐C56S (D) and PDIA1‐C400S (E) pull‐downs compared with NC. (F–H) ‘Over‐representation analysis’ of the top 10 significantly enriched Gene Ontology (GO) pathways in C (F), D (G) and E (H), respectively. (I) Venn diagram illustrates the number of identified proteins whose abundances differ significantly from other groups.

HMMR is a PDIA1‐specific protein substrate

Following the identification of substrates specific to each domain, we directed our attention towards the binding substrates specifically associated with the C₃₉₇GHC₄₀₀ domain. Notably, HMMR exhibited the strongest interaction with PDIA1‐C₃₉₇GHC₄₀₀ in comparison to other client proteins (Fig. 2A,B). We also confirmed the binding of HMMR to PDIA1‐V5^C400S in 293T and LNCaP cells (Fig. 2C,D), indicating that the association between PDIA1 and HMMR was not prostate cancer cell‐specific. In consistent with our previous findings, knocking down PDIA1 by siRNA reduced PSA protein expression in LNCaP cells [6], suggesting a decrease in AR activity (Fig. 2E). Knockdown of PDIA1 also led to a decrease in HMMR protein levels, concomitant with an increase in its ubiquitination (Fig. 2F,G). Of interest, the proteasome degradation inhibitor MG‐132, but not the autophagic degradation inhibitor Bafilomycin A1 (Baf‐A1), modestly restored the decrease of HMMR protein levels caused by PDIA1 knockdown (Figs 2H and S3), implying that the absence of PDIA1 triggered proteasomal degradation of HMMR, which, at least in part, led to a reduction in HMMR protein expression. These results collectively suggest that HMMR is one of the PDIA1 substrates, and PDIA1 plays a pivotal role in the maintenance of HMMR protein stability.

Fig. 2.

HMMR is a substrate of PDIA1. (A) Heatmap illustration for the top 10 proteins specifically bound to the PDIA1^C400S trapping mutant from the immunoprecipitation‐mass spectrometry (IP‐MS) experiment in Fig. 1A. (B) HMMR protein abundance in A. Groups were compared using one‐way ANOVA with Dunnett's multiple comparisons tests. (C, D) 293T cells and LNCaP cells were transfected with an empty vector or PDIA1^C400S‐V5 for 72 h, cell lysates were subjected to IP with an HMMR antibody. (E, F) LNCaP (E) and 293T (F) cells were treated with either siNC or siPDIA1 for 72 h. LNCaP lysates were subjected to immunoblotting analysis against HMMR, PDIA1, PSA and β‐Actin. 293T lysates were subjected to immunoblotting analysis against HMMR, PDIA1 and β‐actin. Groups were compared using unpaired two‐tailed Student's t‐tests. (G) Immunoblotting analysis of protein lysates from LNCaP cells treated with indicated siRNA before IP with the HMMR antibody. Ub: ubiquitin. (H) LNCaP cells were treated with indicated siRNA for 48 h before MG‐132 (10 μm) treatment for a further 16 h. Lysates were subjected to immunoblotting analysis against HMMR and β‐Actin. Groups were compared using two‐way ANOVA with Tukey's multiple comparison tests. Data are shown as means ± S.E.M., *, 0.01 ≤ P < 0. 05; **, 0.001 ≤ P < 0.01; ***, 0.0001 ≤ P < 0.001; ****, P < 0.0001.

PDIA1 catalyses disulfide bond formation between Cys242 and Cys293 of HMMR

Having established that HMMR serves as a PDIA1 substrate, we next sought to identify the specific binding region and critical residues mediating this interaction. Interaction mapping using a series of HMMR truncation mutants revealed that its N‐terminal domain (residues 2–300) was responsible for its binding to PDIA1 (Fig. 3A). This region contains three cysteine residues: Cys17, Cys242 and Cys293. Site‐directed mutagenesis and immunoprecipitation assays revealed that HMMR‐C242S and HMMR‐C293S, but not HMMR‐C17S, abolished binding to PDIA1‐C400S (Fig. 3B), indicating that Cys242 and Cys293 are essential for this interaction. Functional rescue experiments in PDIA1‐knockdown LNCaP cells confirmed that WT HMMR and HMMR‐C17S restored cell survival, whereas HMMR‐C242S and HMMR‐C293S failed to rescue PDIA1 deficiency‐induced cell death (Fig. 3C,D). These findings demonstrate that PDIA1‐mediated disulfide bond formation at Cys242 and Cys293 of HMMR is required for prostate cancer cell viability.

Fig. 3.

PDIA1 catalyses disulfide bond formation between Cys242 and Cys293 of HMMR. (A) LNCaP cells were co‐transfected with N‐terminal V5‐tagged PDIA5‐C400S along with N‐terminal FLAG‐tagged full length HMMR (HMMR‐FL) or HMMR‐2‐300, and the interaction between HMMR and PDIA1 was analysed by IP using anti‐FLAG primary antibodies. (B) Wild‐type HMMR (HMMR‐WT) or its mutants (HMMR‐C17S, HMMR‐C242S and HMMR‐C293S) were co‐transfected into 293T cells with V5‐tagged PDIA1‐C400S. IP was performed with anti‐V5 primary antibodies. * indicates nonspecific bands. (C, D) siNC or siPDIA1 was transfected into control or HMMR‐overexpressing LNCaP cells for 72 h, followed by trypan blue staining to evaluate cell count and viability. Groups were compared using one‐way ANOVA with Tukey's multiple comparison tests. Data are shown as means ± S.E.M., ****, P < 0.0001.

HMMR is highly expressed in prostate cancer and is associated with poor prognosis

According to several prostate cancer transcriptomic datasets [22, 23, 24], the mRNA level of HMMR was upregulated in primary prostate tumours and metastases compared with nonmalignant tissues (Fig. 4A–C). We also observed a significant upregulation in HMMR protein levels in metastatic CRPC compared with primary tumours in a prostate cancer proteomic dataset [25] (Fig. 4D). HMMR mRNA expression was reduced by ADT at a bulk tumour level [26] (Fig. 4E). mRNA levels of HMMR were positively correlated with AR in primary tumours [22] (Fig. 4F). Survival analysis of two independent prostate cancer patient cohorts [22, 27] revealed that high HMMR expression was associated with a poorer progression‐free survival (Fig. 4G). Therefore, HMMR expression is increased in prostate cancer and dictates unfavourable prognosis.

Fig. 4.

HMMR is highly expressed in prostate cancer. (A, B) Expression level of HMMR in benign tissues and primary cancer. Data are from refs. 22 (n = 51) and 23 (n = 127). Groups were compared using unpaired two‐tailed Student's t‐tests. NM: Non‐Malignant. (C) Expression level of HMMR in benign tissues, primary and metastatic (met) prostate cancer. Data are from ref. 24 (NM n = 29; cancer n = 131; met n = 19). Groups were compared using one‐way ANOVA followed by Tukey's multiple comparison tests. (D) Violin plot of HMMR protein expression in benign prostate tissues versus primary and metastatic CRPC tissues. Data are from ref. 25. (NM n = 8, cancer n = 28, met n = 16). Groups were compared using one‐way ANOVA by Tukey's multiple comparison tests. (E) HMMR mRNA expression level in patient prostate tissues biopsies pre‐ and post‐ADT (26) (n = 6). Groups were compared using paired two‐tailed Student's t‐tests. (F) Correlation plots of HMMR and AR mRNA levels in primary prostate cancer (22). (G) HMMR mRNA expression was associated with shorter progression‐free survival in the TCGA (primary prostate cancer) and DKFZ (early‐onset prostate cancer) cohorts (27). Data are shown as means ± S.E.M., **, 0.001 ≤ P < 0.01; ***, 0.0001 ≤ P < 0.001; ****, P < 0.0001.

Discovery of a new PDIA1‐HMMR‐AR signalling axis critical for AR activation

It has been established that HMMR inhibition blocks AR nuclear translocation in prostate cancer cells [28]. To examine whether the absence of PDIA1 activity would affect AR nuclear import, we treated LNCaP cells with the PDIA1 inhibitor CCF642 [17] or the HMMR inhibitor 4‐MU [29] prior to DHT stimulation. Similar to PDIA1 knockdown (Fig. 2), CCF642 treatment indeed also reduced HMMR and AR levels without affecting PDIA1 protein expression in these cells (Fig. 5A,B). Immunofluorescence analysis showed that both inhibitors abolished DHT‐induced AR nuclear accumulation (Fig. 5C). We previously reported that PDIA1 inhibition impairs prostate cancer cell viability by promoting AR degradation [6], therefore, the induction of prostate cancer cell death evoked by PDIA1 inhibition may also involve HMMR degradation. Indeed, restoring HMMR expression partially rescued the cell death induced by PDIA1 knockdown in both LNCaP and C4‐2B cells (Fig. 5D–G). In summary, our findings ([6] and data shown in this study) have demonstrated that the cytotoxic effects of PDIA1 inhibition in prostate cancer cells could be mediated through affecting both AR stability and nuclear import; hence activation, underscoring the complexity of PDIA1's role in prostate tumour cell survival.

Fig. 5.

Prostate cancer cell cytotoxicity induced by PDIA1 inhibition is HMMR‐dependent. (A, B) LNCaP cells were treated with either vehicle or CCF642 (5 μm) for 96 h, before subjected to immunoblotting analysis of HMMR, AR, PDIA1 and β‐actin. n = 3 biological replicates. Groups were compared using unpaired two‐tailed Student's t‐tests. (C) LNCaP cells were treated with CCF642 (5 μm) for 96 h and 4‐MU (0.4 mm) for 24 h before ± DHT (10 nm) treatment for 4 h, the localization of AR (red) within the cell nucleus decreased in response to CCF642 or 4‐MU, as observed through fluorescence microscopy. The cell nucleus was stained with DAPI (blue). Data were analysed using two‐way ANOVA with Tukey's multiple comparison tests. The scale bar represents 0.2 micron/pixels. (D–G) siNC or siPDIA1 were transfected for 72 h in control or HMMR‐overexpressing LNCaP and C42B cells, trypan blue counting (D, E) or SYTOX Green staining plus IncuCyte live imaging (F, G), respectively, were performed to assess cell viability. Images were acquired with a 10x objective. Groups were compared using one‐way ANOVA followed by Tukey's multiple comparison tests. Data are shown as means ± S.E.M., *, 0.01 ≤ P < 0.05; **, 0.001 ≤ P < 0.01; ***, 0.0001 ≤ P < 0.001; ****, P < 0.0001.

Inhibition of HMMR impairs prostate cancer cell migration

Increasing evidence suggests that HMMR is associated with cell migration and invasion in cancers [30, 31, 32]. To investigate whether HMMR impacts on cell mobility in prostate cancer cells, we first assessed the effect of HMMR inhibition on cell viability. Treatment with 0.4 mm 4‐MU did not significantly affect cell confluence or viability in C4‐2B and 22RV1 cells (Fig. 6A,B), indicating that HMMR inhibition does not induce overt cytotoxicity under these conditions. We next examined whether HMMR inhibition influences cell migratory capacity. LNCaP cells were excluded from this assay due to their sensitivity to 4‐MU‐induced cell death upon prolonged treatment (data not shown). Wound healing assays revealed that 4‐MU treatment markedly attenuated wound closure in both C4‐2B and 22RV1 cells compared with vehicle controls (Fig. 6C,D). These results demonstrate that HMMR activity is required for cell migration in prostate cancer cells.

Fig. 6.

Inhibition of HMMR impairs prostate cancer cell migration. (A, B) C4‐2B and 22RV1 cells were treated with DMSO (control) or 0.4 mm 4‐MU. Cell confluence was tracked over time via live‐cell imaging to evaluate viability. Groups were compared using two‐way ANOVA with Tukey's multiple comparison tests. (C, D) C4‐2B and 22RV1 cells were seeded at 8 × 10⁵ cells per well in 6‐well plates. After cell attachment, a scratch wound was created, and cells were treated with 0.4 mm 4‐MU. Images were captured every 24 h, and wound area was quantified using Image J. The scale bar represents 2 mm. Groups were compared using unpaired two‐tailed Student's t‐tests. Data are shown as means ± S.E.M., *, 0.01 ≤ P < 0. 05; **, 0.001 ≤ P < 0.01.

Discussion

Prostate cancer pathogenesis is critically driven by the AR signalling axis. Consequently, elucidating the molecular mechanisms governing AR transcriptional activity remains a central focus. Beyond canonical ligand‐dependent regulation, increasing evidence highlights the importance of post‐translational modifications and protein–protein interactions in fine‐tuning AR stability, localization, and function. PDIA1, a key ER chaperone, has recently emerged as a modulator of AR stability and hence prostate cancer progression. PDIA1 catalyses disulfide bond formation, reduction, and isomerization to support proper protein folding. It is notably abundant in prostate cancer cells and has been implicated in promoting tumour growth. PDIA1 contains two catalytic thioredoxin domains with characteristic CXXC motifs. Although several PDIA1 substrates have been identified, the specific functional contribution of each catalytic domain to substrate recognition and binding remains systematically underexplored. To identify domain‐specific substrates, we leveraged PDIA1 substrate‐trapping mutants (PDIA1^C56S‐V5 and PDIA1^C400S‐V5) and IP‐MS to reveal client proteins that were enriched in the association with specific PDIA1 CXXC domains, during which process we identified HMMR as a new PDIA1^C400‐binding partner.

HMMR is known to serve as a receptor for hyaluronic acid and bind to microtubules to regulate protein transport and spindle assembly before mitosis. In prostate cancer, the microtubule network is particularly important as it directly promotes the nuclear translocation of AR and the transcriptional activation of downstream AR signalling [28]. Our study establishes HMMR as a bona fide substrate of PDIA1, specifically interacting with the C₃₉₇GHC₄₀₀ domain through Cys242 and Cys293 within its N‐terminal region. We demonstrated that PDIA1 is essential for maintaining HMMR stability in prostate cancer cells, as PDIA1 depletion promoted HMMR ubiquitination and subsequent proteasomal degradation. Beyond PDIA1's direct effect on AR protein stability [6], we have now identified an indirect PDIA1‐related mechanism affecting AR activation in response to stimuli in prostate cancer cells, wherein PDIA1 inhibition impaired HMMR‐mediated AR nuclear translocation, an effect phenocopied by direct HMMR inhibition. Crucially, HMMR reconstitution partially rescued cell death induced by PDIA1 knockdown, confirming a newly identified PDIA1‐HMMR‐AR signalling cascade which is of great importance to prostate cancer cell viability. In addition, HMMR was also found to be critical for prostate cancer cell migration, as HMMR inhibition significantly suppressed cell migratory capacity.

Collectively, our findings delineated a novel AR regulatory axis that aids prostate cancer cell survival. The sustained expression of HMMR, even under PDIA1 suppression, conferred a persistent survival advantage, highlighting its key position in this pathway. We thus propose that PDIA1 inhibition can facilitate prostate cancer cell death through governing HMMR stability and function. The current study thus expands our mechanistic understanding of PDIA1 in tumour biology and has important implications in the designing of future therapeutic strategies against prostate cancer.

Author contributions

JX conceived the idea. SY, ZK and JX designed and performed most of the experiments. PY designed and performed immnofluorescence imaging studies. JX provided supervision and oversaw the project. All authors analysed the data. SY, ZK and JX wrote the paper.

Supporting information

Fig. S1. Normalization of the IP‐MS data.

Fig. S2. Client proteins associated with the two PDIA1 CXXC domains trapping mutants exhibited a reduction in abundance following PDIA1 knockdown. Protein expression analysis was performed using a previously published mass spectrometry data where PDIA1 was knocked down by siRNA in LNCaP cells (6). The datasets for the PDIA1^C56S‐V5 and PDIA1^C400S‐V5 mutants, along with proteins identified as binding substrates for either or both mutants, were extracted and analysed to determine their expression levels following PDIA1 knockdown.

Fig. S3. HMMR degradation was unrelated to autophagy. LNCaP cells were transfected with either siNC or siPDIA1. 72 h post‐transfection, the cells were treated with either DMSO or Baf‐A1 for 16 h. Subsequently, cell lysates were collected and subjected to SDS/PAGE and western blotting analysis. The expression levels of HMMR and PDIA1 were analysed by immunoblotting, with β‐actin as the loading control. Data (n = 3) are shown as means ± S.E.M. Groups were compared using unpaired t‐tests: *, 0.01 ≤ P < 0. 05; **, 0.001 ≤ P < 0.01; ***, 0.0001 ≤ P < 0.001; ****, P < 0.0001.

Table S1. MS raw data of the IP experiment.

Table S2. List of proteins specifically binding to PDIA1‐WT, PDIA1‐C56S, and PDIA1‐C400S.

Table S3. List of siRNAs used in this study.

Table S4. Key reagents or resources used in this study.

Data accessibility

Original data generated and analysed during this study are included in this published article or in the data repositories listed in References. IP‐MS raw data are available as Table S1. See also Table S4 for a full description of key reagents or resources used in this study. Further information and requests for resources and reagents should be directed to and will be fulfilled by the corresponding author Jianling Xie (jianlingxie@scut.edu.cn).