High SRD5A2 expression is associated with poor prognosis and promotes tumor cell progression in colorectal cancer

Introduction

Colorectal cancer (CRC) constitutes a significant global health burden and is one of the leading gastrointestinal malignancies. As reported by the Global Cancer Observatory (GLOBOCAN) 2022 database, there were about 1,926,118 new CRC cases and about 903,859 deaths globally in 2022, accounting for 9.6% of all new cancer cases and 9.3% of the death of the disease (1). The incidence of CRC is particularly high in developed countries, and the incidence rate has steadily ranked third in the cancer spectrum. Among them, the incidence among younger adults has shown an annual increase of 1–4% (2), and the mortality rate ranks second (1). Although screening and comprehensive treatment have benefited some patients, there are still about 20% of patients who have local advanced stage or distant metastasis at initial diagnosis, and approximately 25% of non-metastatic patients eventually progress to metastatic disease (3,4). The prognosis for these patients remains unfavorable in the long term, and recurrence and metastasis are still the main cause of death. At present, although targeted drugs such as bevacizumab and anti-epidermal growth factor receptor (EGFR) monoclonal antibody have achieved certain results in CRC treatment and improved the survival of some patients (3,5), CRC is prone to recurrence, and the prognosis is still not ideal. Consequently, there is an urgent need to identify robust molecular biomarkers and therapeutic targets to improve early detection, prognostic stratification, and precision treatment in CRC (6,7).

Accumulating evidence supports a significant association between sex hormones and CRC (8). Steroid 5α-reductase (SRD5A), a key enzyme in steroid metabolism, comprises three isoenzymes encoded by SRD5A1, SRD5A2, and SRD5A3, corresponding to types I, II, and III. These isoenzymes have garnered increasing attention in various diseases, especially hormone-related cancers (9,10). SRD5A1 and SRD5A2 exhibit distinct tissue-specific expression patterns. SRD5A1 is predominantly localized to non-reproductive tissues including the liver and skin. In contrast, SRD5A2 is primarily expressed in reproductive organs such as the genitals and prostate. Critically, both isoenzymes share the essential function of converting testosterone (T) to the more potent androgen dihydrotestosterone (DHT), a reaction integral to androgen metabolism (11). Through the androgen receptor (AR), androgens can activate a series of downstream pro-tumorigenic signaling pathways, thereby contributing to several tumor-promoting processes, including cancer cell proliferation, invasion, and angiogenesis (12). In contrast, emerging evidence indicates that relatively late expression of SRD5A3 is overexpressed in human fetal liver, endometrial cancer (EC), and other diseases (13). Finasteride is a 5α-reductase inhibitor, and SRD5A2 is its target (14). And Prostate Cancer Prevention Trial (PCPT) demonstrated that long-term use of finasteride can reduce the risk of prostate cancer (PC) by about 25% (15). In addition, accumulating evidence indicates that SRD5A2 and related pathways are upregulated in hormone-dependent malignancies, including PC and breast cancer (BC), and are associated with tumor aggressiveness and prognosis (16,17).

However, the currently publicly reported studies of SRD5A2 in CRC are extremely limited. Previous studies mainly focused on the role of SRD5A1 in CRC, and there were almost no reports related to SRD5A2 (5). At present, studies have shown that there are gender differences in the development of CRC (18). In view of the potential relationship between CRC and hormone metabolism and the research in hormone-sensitive tumors such as PC, this study has designed a systematic analysis. We assessed SRD5A2 expression patterns and their prognostic relevance using tissue microarray (TMA) based immunohistochemistry (IHC) with QuPath quantification and follow-up data. We further investigated its functional role in CRC cells through small interfering RNA (siRNA)-mediated knockdown assays and performed preliminary mechanistic analyses focusing on EMT-related molecular changes, classical oncogenic signaling pathways, and metabolic homeostasis (19,20). This study aimed to clarify the clinical and biological significance of SRD5A2 in CRC and to evaluate its potential as a prognostic biomarker and therapeutic target. We present this article in accordance with the MDAR and REMARK reporting checklists (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-0658/rc).

Methods

Human colorectal tissues and patient information

This study is a retrospective cohort study, including 180 patients who underwent radical surgical treatment in Wuxi Second People’s Hospital between 2010 and 2011 and were pathologically diagnosed with CRC. Radical surgical resection was selected as the primary treatment for all patients according to the National Comprehensive Cancer Network (NCCN) and Chinese Society of Clinical Oncology (CSCO) guidelines for resectable CRC, and treatment allocation was based on clinical indications and guidelines, not randomized. Clinicopathological variables were collected, including age, gender, tumor site, gross type, depth of invasion, lymph node metastasis, distant metastasis, differentiation, Ki-67 index levels, and elevated carcinoembryonic antigen (CEA). Exclusion criteria were as follows: (I) patients who received neoadjuvant radiotherapy or chemotherapy prior to surgery; (II) patients with a history of other malignancies; (III) patients with insufficient tumor tissue or inadequate tissue quality for TMA construction; and (IV) patients with incomplete clinicopathological or follow-up data. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Ethics Committee of Wuxi Second People’s Hospital (Ethics Review No. [Y-200] of 2024). Written informed consent was obtained from all patients prior to surgery.

Follow-up and outcome definition

Follow-up was conducted every 6 months, and the first follow-up time was 6 months after surgery. The follow-up method included inpatient or outpatient medical record consultation and telephone follow-up. The follow-up data were current as of July 2025, with a maximum follow-up duration of 15 years. The median follow-up time was 180 months (range, 12–180 months). The follow-up result was recorded as death, loss to follow-up, or survival at the end of the follow-up. Overall survival (OS), defined as the duration stretching from surgical intervention to death, served as the primary endpoint. Patients who were alive or lost to follow-up were censored at the date of last confirmed survival.

TMA construction

All tissue samples were fixed in 10% neutral buffered formalin for 24–48 hours, then embedded in paraffin and stored at room temperature until sectioning. TMA was used to construct an IHC detection cohort. For each patient, two cores were taken from representative tumor areas, and one core was obtained from paired distant normal colonic mucosa. The core diameter was 1.3 mm. All TMA blocks were constructed by the Department of Pathology at Wuxi Second People’s Hospital.

IHC staining

The thickness of TMA wax block slice was 4 µm. After conventional deparaffinization and rehydration of the slice, citrate buffer (10 mM, pH 6.0) was used for antigen repair, high-pressure thermal repair treatment, and the experiment was continued after cooling to room temperature. Endogenous peroxidase was incubated at 3% hydrogen peroxide (H₂O₂) at room temperature for 10 minutes, and then closed with 5% bovine serum albumin (BSA) for 30 minutes. The sections were then incubated with the primary antibody against SRD5A2 (Bioss, Beijing, China; bsm-62269R) overnight at 4 ℃. Next, we applied the corresponding horseradish peroxidase (HRP)-linked secondary antibody (Beijing Zhongshan Jinqiao, Beijing, China; ZB2301/ZB2305) to the sections. Incubation was carried out at room temperature for 60 minutes. Following diaminobenzidine (DAB) development and hematoxylin counterstaining, we performed dehydration and clearing of the sections prior to final mounting. Negative controls and known positive tissue samples were included as positive controls for each batch of staining.

IHC scoring and QuPath analysis

The IHC TMA slides were digitally scanned and exported to .svs format. The image pixel resolution was 0.2466 µm and then quantitatively analyzed in QuPath (v0.6.3). The cells were categorized into four grades: 0 (negative), 1+ (weak), 2+ (moderate), and 3+ (strong) according to the staining intensity. An H-score was subsequently determined for each sample using the formula: 1 × (% of 1+ cells) + 2 × (% of 2+ cells) + 3 × (% of 3+ cells), resulting in a potential range of 0 to 300.

The expression level of tumor SRD5A2 in each patient was taken as the average of 2 tumor core H-scores. Based on the median tumor H-score (cutoff =214.9), patients were dichotomized into high and low SRD5A2 expression groups for comparative analysis of clinical features and survival. Pairing distal normal mucosal coring was used for pairing and comparison with tumor tissue.

Cell culture

This study used a normal colonic epithelial cell line [NCM460; American Type Culture Collection (ATCC), Manassas, VA, USA] and a panel of five CRC cell lines [Caco-2, HCT-116, LoVo, SW480, and RKO; Chinese Tissue Culture Collections (CTCC, Jinhua, China); HCL-0065, HCL-0067, HCL-0069, HCL-0072, and HCL-0071]. All cell lines were cultured in a humidified incubator. The temperature and CO₂ level were set at 37 ℃ and 5%, respectively. NCM460 cells were cultured in RPMI 1640 medium (Hyclone, Logan, UT, USA; SH30011.03), HCT-116 cells in McCoy’s 5A medium (Procell, Wuhan, China; PM150710), LoVo cells in F-12K medium (Solarbio, Beijing, China; LA1320), and Caco-2, SW480, and RKO cells in high-glucose Dulbecco’s modified Eagle medium (DMEM) (Hyclone; SH30243.01). Cells were cultured under standard conditions, which involved the use of media containing 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin (P/S). Medium changes were performed every 2–3 days, guided by the observed growth status of the cells.

siRNA transfection

In order to silence SRD5A2, three siRNAs targeting SRD5A2 (si-SRD5A2-648, si-SRD5A2-478, si-SRD5A2-281) and negative control siRNA (si-NC) were designed and synthesized. Use Lipo8000™ transfecting reagent (Beyotime, Shanghai, China; C0533) for transfecting. The cells were inoculated to the six-hole plate, and transfected when the cell fusion reached 80%. Forty-eight hours after transfection, reverse transcription quantitative polymerase chain reaction (RT-qPCR) and western blotting (WB) were carried out to verify the knockdown efficiency, and the siRNA with the best knockdown effect was selected for subsequent functional experiments.

RT-qPCR

A commercial RNA extraction kit (Wuxi Bohe, Wuxi, China; M004) was employed to isolate total RNA from all experimental cell groups. RNA quantification and purity were measured with a microvolume spectrophotometer (Bio-Rad, Hercules, CA, USA; SmartSpec Plus) to confirm that the A260/A280 ratio was between 1.8 and 2.0. 1.0 µg of total RNA was used for reverse transcription. TRUEScript H Minus M-MuLV Reverse Transcriptase (Beijing Aidley, Beijing, China; PC1703) was used for reverse transcription reaction, and the reverse transcription system contained dNTP Mixture (Beijing Aidley; PC2403), RNasin RNA enzyme inhibitor (Beijing Aidley; RN3501), and random primer (see Table S1 for the specific reaction system). Subsequently, 2× SYBR Green qPCR Mix (Beijing Aidley; PC3302) was used for amplification detection on the ABI 7500 RT-qPCR instrument. The primer sequence of SRD5A2 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was found in Table S2. The reaction system was 20 µL, which contained specific primers and cDNA templates. The amplification protocol included an initial 10-minute denaturation at 95 ℃, with 40 subsequent cycles performed under the following conditions: 20 seconds of denaturation at 95 ℃, 20 seconds of annealing at 55 ℃, and 20 seconds of extension at 72 ℃. Following amplification, the melting curve was generated to assess product specificity. Then, relative expression of SRD5A2 was quantified using the 2^−ΔΔCt method, with GAPDH serving as the endogenous control for normalization.

WB

The cells were collected after 48 hours of transfection, and the pre-cooled radioimmunoprecipitation assay (RIPA) lysate (Beyotime; P0013B) containing 1% phenylmethylsulfonyl fluoride (PMSF) (RUIBIO, Beijing, China; BP2655) was lysed on ice for 30 minutes. Protein concentrations were measured by bicinchoninic acid (BCA) (Bioshar P, Hefei, China; BL521A). Equivalent amounts were then subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene fluoride (PVDF) membranes (Millipore, Billerica, MA, USA; IPVH00010). After blocking with Tris-buffered saline with Tween 20 (TBST) containing 5% BSA for 1 hour at room temperature, the membranes were incubated overnight at 4 ℃ with primary antibodies against SRD5A2, GAPDH (Proteintech, Rosemont, IL, USA; 60004-1-Ig), E-cadherin (Proteintech; 20874-1-AP), N-cadherin (Proteintech; 22018-1-AP), Vimentin (Proteintech; 10366-1-AP), Snail (Proteintech; 12129-1-AP), ZEB1 (Proteintech; 21544-1-AP), MMP-2 (Proteintech; 10373-2-AP), ERK1/2 (Affinity, Cincinnati, OH, USA; BF8004), p-ERK1/2 (Affinity; AF1015), JNK (Affinity; AF6318), p-JNK (Affinity; AF3318), IKKα/β (Affinity; AF6014), P65 (Proteintech; 80979-1-RR), p-P65 (Proteintech; 82335-1-RR), AKT (Proteintech; 10176-2-AP), p-AKT (Proteintech; 66444-1-Ig), mTOR (Proteintech; 66888-1-Ig), and p-mTOR (Proteintech; 67778-1-Ig). The next day, membranes were incubated with HRP-conjugated goat anti-rabbit secondary antibody (ZSGB, Beijing, China; ZB2301) or HRP-conjugated goat anti-mouse secondary antibody (ZSGB; ZB2305), followed by ECL chemiluminescence detection (Dingguo, Beijing, China; ECL-0011) and imaging using a ChemiScope 5300 Pro system (CLINX, Shanghai, China). Band intensities were quantified using ImageJ. Total protein expression levels were normalized to GAPDH, and phosphorylated protein levels were expressed relative to their corresponding total proteins where appropriate.

Cell viability assay

After transfection, the cells were inoculated in a 96-hole plate (5×10⁴/mL). At 0 and 48 hours post-seeding, 20 µL of Cell Counting Kit-8 (CCK-8) (CTCC; M006) reagent was applied to each well. Once incubated for 2 hours at 37 ℃, the enzyme label detector (Thermo Fisher Scientific, Waltham, MA, USA; MK3) was used in 450 nm waves to measure the absorbance of each hole. Wells with medium and CCK-8 reagent without cells were used for blanks. The proliferative activity of treated cells was derived from their absorbance relative to the untreated control.

Wound healing assay

Following transfection, cells were seeded in six-well plates and cultured to confluence. A linear wound was generated across the monolayer with a sterile tip. Following this, the wells were gently washed twice with phosphate-buffered saline (PBS) to remove shredded cells, followed by replacement with serum-free medium. At 0 and 48 hours, images of the same field were captured to record changes in the scratch width. Wound healing rate was calculated using ImageJ as follows: %wound closure = (W0 − Wt)/W0 × 100%. Data are from at least three independent replicates.

Transwell invasion assay

Transwell chambers (FALCON, Franklin Lakes, NJ, USA; 353097) were used for the invasion experiment. Matrigel was pre-laid in the upper chamber (Matrigel: serum-free culture medium =1:3; v/v; BD, Franklin Lakes, NJ, USA; 356234). After gelation at 37 ℃, the cells that had been inoculated and transfected were seeded in the serum-free culture medium in the upper chamber at a density of 1×10⁵ cells/mL. Medium containing 10% FBS served as the chemoattractant in the lower chamber. Following 48 hours of incubation, the upper chamber cells were removed by swabbing. The migrated cells adherent to the lower membrane were fixed. Then, they were ready for subsequent staining with crystal violet. For quantification, Quantification was performed by enumerating cells from five randomly chosen microscopic fields per well. Data are from at least three independent replicates.

Flow cytometry analysis of apoptosis

After harvesting and a PBS wash, cells were subjected to apoptosis staining. The Annexin V-FITC/PI kit (Beyotime; 401006) was used strictly following the manufacturer’s instructions. Samples were immediately analyzed on a flow cytometer (BD; FACSVerse). Early and late apoptotic cell fractions were determined from the cytometry plots, and their sum was denoted as the total apoptotic rate. Data are from at least three independent replicates.

Glucose consumption assay

HCT-116 cells were transfected with si-NC or si-SRD5A2. After 48 hours, the medium was replaced with fresh McCoy’s 5A medium (Procell; PM150710). After 24 hours, conditioned media were collected and centrifuged (1,000 ×g, 10 min). Glucose concentration in the supernatant was measured using a glucose assay kit (Beyotime; S0201S) according to the manufacturer’s instructions. Briefly, 5 µL of sample was mixed with 185 µL of assay reagent, heated at 95 ℃ for 8 min in a thermal cycler (Bio-Rad; T100), and then cooled to 4 ℃. Absorbance was read at 630 nm using a microplate reader (Thermo Fisher Scientific; MK3). Glucose consumption was calculated by subtracting the residual glucose concentration from that of blank medium and normalized to total protein content.

ATP assay

HCT-116 cells were transfected with si-NC or si-SRD5A2. After 48 hours, cells were collected and lysed using an adenosine triphosphate (ATP) assay kit (Shanghai Yubo Biotechnology, Shanghai, China; YB0305) according to the manufacturer’s instructions. Briefly, cells were homogenized in the provided extraction buffer by sonication on ice (200 W, 2 s on/1 s off for 1 min). The lysate was centrifuged at 10,000 ×g for 10 min at 4 ℃, and the supernatant was collected. Then, 20 µL of each sample or standard was mixed with 128 µL of Reagent I and 52 µL of working solution in a 96-well ultraviolet (UV) plate. The mixture was incubated at 37 ℃ for 3 min, and the absorbance was measured at 340 nm using a microplate reader (Thermo Fisher Scientific; MK3). ATP levels were calculated based on a standard curve and normalized to total protein content.

Intracellular reactive oxygen species (ROS) assay

HCT-116 cells were transfected with si-NC or si-SRD5A2. After 48 hours, cells were collected and resuspended in serum-free medium containing 10 µmol/L DCFH-DA (Beyotime; S0033S) at a density of 1–2×10⁶ cells/mL. The cells were incubated at 37 ℃ for 20 min in a CO₂ incubator, with gentle mixing every 5 min. After incubation, the cells were washed three times with serum-free medium to remove extracellular probe. ROS levels were immediately analyzed using a flow cytometer (BD; FACSVerse). Data were processed with FlowJo V10 software, and the percentage of ROS-positive cells was used to evaluate intracellular ROS levels.

Statistical analysis

Statistical analyses were performed using SPSS 29.0 and GraphPad Prism 10. Continuous variables were expressed as mean ± standard deviation (SD) or median [interquartile range (IQR)], as appropriate, and compared using Student’s t-test or the Mann-Whitney U test. Categorical variables were compared using the Pearson χ² test or Fisher’s exact test. Paired H-scores between tumor tissues and distal normal mucosa were compared using the Wilcoxon signed-rank test. OS was analyzed using the Kaplan-Meier method and compared by the log-rank test. Univariate and multivariate Cox proportional hazards regression analyses were performed to evaluate the prognostic value of SRD5A2 and clinicopathological variables. Subgroup Kaplan-Meier analyses were further conducted according to T stage, lymph node status, and distant metastasis status. For in vitro experiments, data from at least three independent experiments were presented as mean ± SD and analyzed using an unpaired Student’s t test or one-way/two-way analysis of variance (ANOVA), as appropriate. All tests were two-sided, and P<0.05 was considered statistically significant.

Results

SRD5A2 is upregulated in CRC tissues

In order to evaluate the expression of SRD5A2 in CRC tissues, we collected the tumor tissue of 180 CRC patients and their paired distal normal mucosal tissue and used IHC for detection. SRD5A2 expression was detected in CRC tumor tissues using IHC staining (Figure 1A). The corresponding IHC images were further quantitatively analyzed using QuPath, and cells were classified according to staining intensity to calculate the H-score (Figure 1B). SRD5A2 expression was also examined in paired distant normal colonic mucosal tissues (Figure 1C), and the corresponding QuPath-based quantitative analysis was performed (Figure 1D). Statistical analysis showed that a substantial elevation of SRD5A2 was observed in tumor tissues compared with matched distant normal tissues (P<0.001; Figure 1E). These results suggested that SRD5A2 showed significantly upregulated expression in CRC tumor tissue.

Figure 1 Upregulation of SRD5A2 expression in CRC tissues and QuPath quantitative analysis. (A) Representative IHC image of SRD5A2 expression in CRC tumor tissues using DAB chromogenic staining with hematoxylin counterstaining, with the scale bar as indicated in the figure. (B) Corresponding QuPath pseudocolor overlay of (A), cells were detected and classified based on staining intensity (0: negative; 1+: weak positive; 2+: moderate positive; 3+: strong positive), with the color bar below representing the intensity levels. (C) Representative IHC image of SRD5A2 expression in paired distant normal colonic mucosa using DAB chromogenic staining with hematoxylin counterstaining. (D) Corresponding QuPath pseudocolor overlay of (C). (E) Comparison of SRD5A2 H-scores between tumor tissues and paired distant normal tissues, presented as a box plot. ****, P<0.001. CRC, colorectal cancer; DAB, diaminobenzidine; IHC, immunohistochemistry; SRD5A2, steroid 5α-reductase type II.

High SRD5A2 expression correlates with clinicopathological features

According to the median of the average H-score of the IHC of two tumor tissues per patient, 180 CRC patients were categorized into low- and high-expression groups, each constituting 90 patients (Table 1). Regarding age and gender, the difference in the distribution of the two groups did not reach statistical significance (P>0.05). A statistically significant difference was observed in the tumor site distribution between groups (P=0.01): the high-expression group was more common in the right colon (47.8% vs. 34.4%), while the proportion of the rectum was lower (23.3% vs. 44.4%). In terms of invasion-related indicators, high SRD5A2 expression correlated with deeper infiltration depth (P=0.004), and its T3–T4 proportion was higher (86.7% vs. 68.9%). At the same time, the high-expression group was more prone to lymph node metastasis (55.6% vs. 38.9%). No significant difference was observed in the rate of distant metastasis between the two groups (P=0.35). With regard to tumor differentiation, there was a statistically significant difference in the distribution of differentiation grades between the two groups (P<0.001). Compared with the low-expression group, the low differentiation rate of the SRD5A2 high-expression group increased significantly (27.8% vs. 10.0%), while the high differentiation rate decreased (7.8% vs. 21.1%); the medium differentiation ratio was similar to the two groups (64.4% vs. 68.9%), suggesting high expression of SRD5A2 was associated with adverse differentiation tendency. Furthermore, the high-expression group was related with higher Ki-67 levels (P=0.01), with a higher proportion of Ki-67 “+++” (53.3% vs. 40.0%) and a lower proportion of “+” (5.6% vs. 21.1%). In terms of serology, the CEA level in the high-expression group was significantly increased (77.8% vs. 37.8%). In conclusion, high SRD5A2 expression correlated significantly with several clinicopathological characteristics such as tumor site, infiltration depth, lymph node metastasis, degree of differentiation, Ki-67 proliferation level, and elevated CEA (Table 1).

Elevated SRD5A2 correlates with poorer OS in CRC

A prognostic analysis was conducted to evaluate SRD5A2. Using the median IHC H-score as the cutoff, patients were dichotomized into high- and low-expression cohorts. Kaplan-Meier analysis was performed to compare their OS (Figure 2). OS was significantly shorter in patients with high SRD5A2 expression relative to those with low expression (log-rank test: χ²=23.52, P<0.001). In the high-expression group, the median OS was 168 months, while the low-expression group did not reach median OS during the follow-up period [not reported (NR)]. In summary, we conclude that elevated SRD5A2 was a significant indicator of poor OS, which had potential prognostic stratification value.

Figure 2 Kaplan-Meier OS curves stratified by SRD5A2 expression. A total of 69 death events occurred (51 in the high-expression group, 18 in the low-expression group). P<0.001. OS, overall survival; SRD5A2, steroid 5α-reductase type II.

SRD5A2 is an independent prognostic factor for OS in CRC

To further clarify the prognostic value of SRD5A2, univariate and multivariate Cox regression analyses were performed for OS. In univariate analysis, high SRD5A2 expression, older age, deeper invasion, lymph node metastasis, and elevated CEA were significantly associated with poorer OS. After adjustment for major clinicopathological variables, high SRD5A2 expression remained independently associated with shorter OS [hazard ratio (HR) =2.812; 95% confidence interval (CI): 1.510–5.234; P=0.001]. Age, depth of invasion, and lymph node metastasis also remained significant in the multivariate model (Table 2).

Subgroup Kaplan-Meier analyses support the prognostic stability of SRD5A2 in different clinical strata

To further assess the prognostic stability of SRD5A2 in different clinical strata, subgroup Kaplan-Meier survival analyses were performed according to T stage, lymph node status, and distant metastasis status. High SRD5A2 expression remained significantly associated with poorer OS in the T1–T2 subgroup (χ²=10.72, P=0.001), T3–T4 subgroup (χ²=11.22, P=0.001), N1 subgroup (χ²=20.40, P<0.001), M0 subgroup (χ²=20.79, P<0.001), and M1 subgroup (χ²=3.853, P=0.050). In the N0 subgroup, a similar trend was observed, although the difference did not reach statistical significance (χ²=2.926, P=0.09) (Figure 3).

Figure 3 Subgroup Kaplan-Meier survival analyses according to SRD5A2 expression in CRC patients. (A) T1–T2 subgroup. (B) T3–T4 subgroup. (C) N0 subgroup. (D) N1 subgroup. (E) M0 subgroup. (F) M1 subgroup. OS was estimated using the Kaplan-Meier method and compared by the log-rank test. M, metastasis; N, node. OS, overall survival; SRD5A2, steroid 5α-reductase type II; T, tumor.

Selection of cell line and validation of SRD5A2 knockdown

This study aimed to investigate the role of SRD5A2 in CRC cells. A cell model that could be used for in vitro functional research was established, and high-efficiency siRNA was screened for transient knockdown. The endogenous expression levels of SRD5A2 were first assessed across a panel of CRC cell lines using both RT-qPCR and WB. It was found that compared with the normal colonic epithelial cell line, SRD5A2 showed high expression in multiple CRC cell lines, among which the expression levels of HCT-116 cells were the highest at both the messenger RNA (mRNA) and protein levels (Figure 4A,4B), so it was selected as a cell model for subsequent functional studies. Subsequently, in order to achieve an efficient gene silencing, we transfected three distinct siRNAs targeting SRD5A2 (SRD5A2-648, SRD5A2-478, SRD5A2-281) in HCT-116 cells. After 48 hours of transfection, RT-qPCR and WB analysis showed that all three siRNAs could significantly reduce the expression of SRD5A2. Among them, the SRD5A2-281 sequence showed the strongest interference effect at both the mRNA and protein levels (Figure 4C,4D). Based on this result, all subsequent functional experiments were carried out using SRD5A2-281 (hereafter referred to as si-SRD5A2).

Figure 4 Expression profile of SRD5A2 across cell lines and screening of siRNA knockdown efficiency. (A) RT-qPCR analysis of SRD5A2 mRNA expression in the normal colonic epithelial cell line NCM460 and multiple CRC cell lines. (B) WB analysis of SRD5A2 protein expression; representative bands were shown on the right (GAPDH served as the loading control). (C) SRD5A2 mRNA levels in HCT-116 cells at 48 hours after transfection with siRNAs targeting SRD5A2 (si-SRD5A2-648, si-SRD5A2-478, and si-SRD5A2-281). (D) Corresponding SRD5A2 protein levels and representative WB bands. ****, P<0.001. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; mRNA, messenger RNA; RT-qPCR, reverse transcription quantitative polymerase chain reaction; si-NC, negative control siRNA; siRNA, small interfering RNA; SRD5A2, steroid 5α-reductase type II; WB, western blotting.

SRD5A2 knockdown inhibits cell viability

After confirming the effective knockdown of SRD5A2, we first evaluated its effect on cell proliferation. Knockdown of SRD5A2 markedly suppressed the proliferation of HCT-116 cells relative to the si-NC control, as demonstrated by the CCK-8 assay (Figure 5A,5B). These results proved that SRD5A2 was a critical factor in sustaining the ability of CRC cell proliferation.

Figure 5 SRD5A2 knockdown suppresses malignant phenotypes and promotes apoptosis in CRC cells. (A) Representative images of the CCK-8 assay in HCT-116 cells after transfection with si-NC or si-SRD5A2 (×100). (B) Quantification of cell viability. (C) Representative images of the wound-healing assay at 0 and 48 hours (×100). (D) Quantification of wound closure. (E) Representative images of crystal violet-stained invaded cells (×200). (F) Quantification of invaded cells. (G) Representative flow cytometric dot plots of apoptosis. (H) Quantification of total apoptotic rate. ****, P<0.001. AV-FITC, Annexin V-fluorescein isothiocyanate; CCK-8, Cell Counting Kit-8; CRC, colorectal cancer; PI, propidium iodide; si-NC, negative control siRNA; si-SRD5A2, siRNAs targeting SRD5A2; siRNA, small interfering RNA; SRD5A2, steroid 5α-reductase type II.

SRD5A2 depletion impairs cell migration and invasion

In view of the fact that SRD5A2 was related to tumor metastasis in clinical samples, we further explored its impact on the migratory and invasive potential of cells. The results of Wound-healing and Transwell indicated that SRD5A2 knockdown significantly delayed scratch closure (Figure 5C,5D) and greatly decreased the number of invading cells (Figure 5E,5F). These data together showed that SRD5A2 could enhance the migration and invasion of CRC cells.

Silencing SRD5A2 promotes apoptosis

Finally, the effect of SRD5A2 on cell apoptosis was examined. Flow cytometric analysis showed that SRD5A2 knockdown significantly increased the apoptotic rate of HCT-116 cells (Figure 5G,5H). These findings indicated that SRD5A2 may facilitate malignant tumor progression, at least in part, by suppressing apoptosis.

SRD5A2 knockdown reverses epithelial-mesenchymal transition (EMT)-related molecular changes

Given that SRD5A2 depletion significantly impaired the migratory and invasive abilities of HCT-116 cells, we next examined whether EMT-related proteins were altered after SRD5A2 silencing. WB analysis showed that SRD5A2 knockdown increased the expression of E-cadherin, whereas the expression levels of N-cadherin, Vimentin, Snail, ZEB1, and MMP-2 were markedly decreased compared with the si-NC group (Figure 6A,6B). These findings suggest that SRD5A2 may promote the malignant phenotype of CRC cells, at least in part, by facilitating EMT-related molecular programs.

Figure 6 SRD5A2 knockdown reverses EMT-related molecular changes and suppresses classical oncogenic signaling pathways in HCT-116 cells. (A) Representative WB bands of EMT-related proteins, including E-cadherin, N-cadherin, Vimentin, Snail, ZEB1, and MMP-2. (B) Quantification of EMT-related protein expression. (C) Representative WB bands of signaling pathway-related proteins, including ERK1/2, p-ERK1/2, JNK, p-JNK, IKKα/β, P65, p-P65, AKT, p-AKT, mTOR, and p-mTOR. (D) Quantification of signaling pathway-related protein expression. ****, P<0.001. EMT, epithelial-mesenchymal transition; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; si-NC, negative control siRNA; si-SRD5A2, siRNAs targeting SRD5A2; siRNA, small interfering RNA; SRD5A2, steroid 5α-reductase type II; WB, western blotting.

SRD5A2 knockdown suppresses classical oncogenic signaling pathways

To further investigate the molecular mechanisms underlying the pro-tumorigenic role of SRD5A2, we examined several classical signaling pathways associated with CRC progression. WB analysis demonstrated that SRD5A2 silencing reduced the levels of ERK1/2, p-ERK1/2, JNK, p-JNK, IKKα/β, P65, p-P65, AKT, p-AKT, mTOR, and p-mTOR. In particular, the phosphorylation levels of ERK1/2, JNK, P65, AKT, and mTOR were markedly decreased after SRD5A2 knockdown (Figure 6C,6D). These findings indicate that SRD5A2 may contribute to CRC progression by sustaining multiple oncogenic signaling pathways, including the MAPK/ERK, JNK, NF-κB, and AKT/mTOR pathways.

SRD5A2 knockdown impairs metabolic homeostasis and increases oxidative stress

Because metabolic reprogramming is closely associated with CRC progression, we further evaluated whether SRD5A2 silencing affected metabolic homeostasis in HCT-116 cells. Compared with the si-NC group, SRD5A2 knockdown reduced glucose consumption and intracellular ATP levels, while intracellular ROS levels were significantly increased (Figure 7A-7C). These findings suggest that SRD5A2 may participate in maintaining metabolic activity and redox balance in CRC cells.

Figure 7 SRD5A2 knockdown impairs metabolic homeostasis and increases oxidative stress in HCT-116 cells. (A) Glucose consumption in HCT-116 cells after transfection with si-NC or si-SRD5A2. (B) Intracellular ATP levels. (C) Intracellular ROS levels. ****, P<0.001. ATP, adenosine triphosphate; ROS, reactive oxygen species; si-NC, negative control siRNA; si-SRD5A2, siRNAs targeting SRD5A2; siRNA, small interfering RNA; SRD5A2, steroid 5α-reductase type II.

Discussion

CRC is a highly heterogeneous gastrointestinal malignancy (21), whose initiation and progression are driven by multiple factors, including molecular subtypes, the tumor microenvironment, and metabolic reprogramming (22). Clinically, even after curative resection, patients may still experience recurrence or develop distant metastases during follow-up. Moreover, substantial inter-individual differences in survival outcomes are observed even within the same tumor-node-metastasis (TNM) stage (22), indicating that reliance on conventional pathological staging and a limited set of serum biomarkers remains insufficient for precise risk stratification and individualized management. In this context, using a surgical cohort with comprehensive clinicopathological and follow-up data, the present study employed TMA and a standardized follow-up endpoint to evaluate the clinical value of a prognostically relevant molecular biomarker, thereby providing evidence to refine prognostic stratification in CRC.

Given that SRD5A2 is a key enzyme for androgen metabolism, it is necessary for us to discuss its potential impact in CRC from the perspective of sex hormone action. Research increasingly indicates that the occurrence and development of CRC may be affected by the sex hormone axis. Epidemiological data show that the incidence and mortality of CRC in men are higher than those of women, and the prognosis of female patients of the same age is also relatively good. The risk of CRC in premenopausal women is significantly lower than that of men, and the incidence and mortality rate of CRC in women who receive hormone replacement therapy after menopause are also lower (23). And some studies have shown that estrogen may have a protective effect on CRC tumors, while androgens may promote the occurrence of intestinal tumors. These phenomena suggest that sex hormones may play an important role in CRC (24,25).

To our knowledge, this study is among the first to show that SRD5A2 is highly expressed in CRC tissues, and patients in the high expression group were often accompanied by poor clinicopathological features and poorer OS. Specifically, elevated SRD5A2 expression was significantly associated with advanced T stage, lymph node metastasis, a higher proliferative index, and increased serum CEA levels, which are all indicators of tumor invasion and metastasis risk (26,27). Among them, high Ki-67 expression usually means that tumor proliferation is active and the prognosis is poor (28), and CEA >5 µg/L has been recognized as a sign of poor CRC prognosis (27). The concomitant increases in Ki-67 and CEA observed in the SRD5A2 high-expression group further support the association between SRD5A2 and an aggressive phenotype. Moreover, multivariate Cox regression analysis demonstrated that high SRD5A2 expression remained independently associated with poorer OS after adjustment for major clinicopathological variables. Subgroup Kaplan-Meier analyses further supported the relative prognostic stability of SRD5A2 across different T-, N-, and M-based strata, although the association did not reach statistical significance in the N0 subgroup.

In vitro experimental results further support that SRD5A2 may play a pro-tumorigenic role. After we knocked down SRD5A2 in HCT-116 cells, cell proliferative vitality decreased, migration and invasion abilities weakened, and the level of apoptosis increased, suggesting that SRD5A2 may be involved in maintaining the malignant phenotype of CRC cells. We further found that SRD5A2 knockdown reversed EMT-related molecular changes, as evidenced by increased E-cadherin and decreased N-cadherin, Vimentin, Snail, ZEB1, and MMP-2 expression. Moreover, SRD5A2 silencing suppressed several classical oncogenic signaling pathways, including MAPK/ERK, JNK, NF-κB, and AKT/mTOR (19). In addition, reduced glucose consumption and ATP levels together with increased intracellular ROS suggested that SRD5A2 may also participate in maintaining metabolic homeostasis and redox balance in CRC cells (29).

Recent evidence suggests that sex steroid receptor signaling may influence CRC progression and may offer potential opportunities for prognostic stratification and personalized therapeutic intervention in CRC (12). Moreover, sex-based differences in CRC epidemiology, tumor biology, and clinical outcomes further support the relevance of hormone-related pathways in this disease (30). Also, evidence indicates that high expression of AR correlates with poorer prognosis in CRC tissues (26). Our data support a similar pro-tumorigenic role of SRD5A2 in CRC and further suggest that SRD5A2 may represent a potential biomarker and therapeutic target in this disease. In addition, SRD5A3 has also been confirmed to be highly expressed in a variety of tumors as a homozygic isomerase and has a pro-tumorigenic effect (13), suggesting that the enzyme family may be generally involved in tumorigenesis.

Several limitations should be considered of this study. To begin with, the study cohort come from a single center, the sample size was limited, and the retrospective design may be biased. There was no in-depth analysis of the differences in SRD5A2 expression of different sexes. In the future, it is necessary to examine the influence of sex hormone levels or gender on the effect of SRD5A2. Second, although multivariable Cox and subgroup analyses were performed in the study, these prognostic findings still require validation in larger multicenter cohorts. And the prognostic analyses in the present study were mainly based on OS and a median H-score cut-off, and further validation of the optimal cut-off and additional survival endpoints such as DFS will be needed in future studies. In addition, the functional research and verification were mainly based on a single cell line, which was only completed in the in vitro cell model, so future work should further validate its applicability and biological significance in more CRC cell lines and in vivo models. Finally, although we performed preliminary mechanistic analyses and found that SRD5A2 knockdown was associated with EMT reversal, suppression of several classical oncogenic signaling pathways, and impaired metabolic homeostasis, the direct upstream interaction network and precise molecular targets of SRD5A2 remain to be clarified. Therefore, future research should combine animal models and more molecular biological approaches to further define its direct signaling axis and therapeutic potential in CRC. In summary, this study established a link between high SRD5A2 expression and poor prognosis in CRC, and preliminarily confirmed its carcinogenic function, laying the foundation for the future use of SRD5A2 as a diagnostic or prognostic marker and therapeutic target.

Conclusions

In conclusion, this study demonstrates that SRD5A2 is significantly upregulated in CRC tissues and that its high expression is associated with aggressive clinicopathological features and poorer OS. Multivariable survival analysis further suggests that SRD5A2 may serve as an independent prognostic factor in CRC, and subgroup analyses further support its relative prognostic stability across different clinical strata. Functional experiments further indicate that SRD5A2 contributes to the malignant phenotype of CRC cells by promoting proliferation, migration, and invasion while inhibiting apoptosis. Preliminary mechanistic analyses further suggest that SRD5A2 may exert its pro-tumorigenic effects through EMT-related molecular changes, activation of classical oncogenic signaling pathways, and maintenance of metabolic homeostasis. Further multicenter studies and in-depth mechanistic investigations are warranted to clarify its clinical utility and biological role.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the MDAR and REMARK reporting checklists. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-0658/rc

Data Sharing Statement: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-0658/dss

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-0658/prf

Funding: This work was supported by “Light of Taihu Lake” science and technology research project sponsored by Science and Technology Bureau of Wuxi City (No. Y20232022), Medical Key Discipline Program of Wuxi Health Commission (No. Z202415), and Top Talent Support Program for young and middle-aged people of Wuxi Health Committee (No. HB2023027).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-0658/coif). All authors report that this work was supported by research grants from “Light of Taihu Lake” Science and Technology Research Project sponsored by Science and Technology Bureau of Wuxi City (No. Y20232022), Medical Key Discipline Program of Wuxi Health Commission (No. Z202415), and Top Talent Support Program for young and middle-aged people of Wuxi Health Committee (No. HB2023027). The authors have no other conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Wuxi Second People’s Hospital (Ethics Review No. [Y-200] of 2024). Written informed consent was obtained from all patients prior to surgery.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63. [Crossref] [PubMed]
2. Siegel RL, Torre LA, Soerjomataram I, et al. Global patterns and trends in colorectal cancer incidence in young adults. Gut 2019;68:2179-85. [Crossref] [PubMed]
3. Underwood PW, Ruff SM, Pawlik TM. Update on Targeted Therapy and Immunotherapy for Metastatic Colorectal Cancer. Cells 2024;13:245. [Crossref] [PubMed]
4. Biller LH, Schrag D. Diagnosis and Treatment of Metastatic Colorectal Cancer: A Review. JAMA 2021;325:669-85. [Crossref] [PubMed]
5. Wei R, Zhong S, Qiao L, et al. Steroid 5α-Reductase Type I Induces Cell Viability and Migration via Nuclear Factor-κB/Vascular Endothelial Growth Factor Signaling Pathway in Colorectal Cancer. Front Oncol 2020;10:1501. [Crossref] [PubMed]
6. Zhang J, Zhu H, Liu W, et al. Prognostic and predictive molecular biomarkers in colorectal cancer. Front Oncol 2025;15:1532924. Erratum in: Front Oncol 2025;15:1725836. [Crossref] [PubMed]
7. Chilimoniuk Z, Gładysz K, Moniczewska N, et al. The Role of Epigenetic Biomarkers as Diagnostic, Predictive and Prognostic Factors in Colorectal Cancer. Cancers (Basel) 2025;17:2632. [Crossref] [PubMed]
8. Wu J, Bai Y, Lu Y, et al. Role of sex steroids in colorectal cancer: pathomechanisms and medical applications. Am J Cancer Res 2024;14:3200-21. [Crossref] [PubMed]
9. Han Y, Zhuang Q, Sun B, et al. Crystal structure of steroid reductase SRD5A reveals conserved steroid reduction mechanism. Nat Commun 2021;12:449. [Crossref] [PubMed]
10. Chen E, Yi J, Ren Q, et al. Overexpression of SRD5A3 in Hepatocellular Carcinoma and Its Molecular Mechanism: A Study of Bioinformatics Exploration Analysis with Experimental Verification. Evid Based Complement Alternat Med 2022;2022:7853168. [Crossref] [PubMed]
11. SharkeyCGuBLongXStromal SRD5A2 promotes prostate growth through WNT5A-LEF1-IGF1 signaling in benign prostatic hyperplasia.bioRxiv 2025;2025.05.22.655540.
12. Beniwal SS, Janu Y, Cummings K, et al. Sex steroid receptors in colorectal cancer: implications for tumor progression, therapeutic opportunities, and sex-specific outcomes. Ann Med Surg (Lond) 2026;88:1427-36. [Crossref] [PubMed]
13. Mai Q, Sheng D, Chen C, et al. Steroid 5 alpha-reductase 3 (SRD5A3) promotes tumor growth and predicts poor survival of human hepatocellular carcinoma (HCC). Aging (Albany NY) 2020;12:25395-411. [Crossref] [PubMed]
14. Xiao Q, Wang L, Supekar S, et al. Structure of human steroid 5α-reductase 2 with anti-androgen drug finasteride. Nat Commun 2020;11:5430. [Crossref] [PubMed]
15. Thompson IM, Goodman PJ, Tangen CM, et al. The influence of finasteride on the development of prostate cancer. N Engl J Med 2003;349:215-24. [Crossref] [PubMed]
16. Neslund-Dudas C, Bock CH, Monaghan K, et al. SRD5A2 and HSD3B2 polymorphisms are associated with prostate cancer risk and aggressiveness. Prostate 2007;67:1654-63. [Crossref] [PubMed]
17. Durand S, Trillet K, Uguen A, et al. A transcriptome-based protein network that identifies new therapeutic targets in colorectal cancer. BMC Genomics 2017;18:758. [Crossref] [PubMed]
18. Choi Y, Kim N. Sex Difference of Colon Adenoma Pathway and Colorectal Carcinogenesis. World J Mens Health 2024;42:256-82. [Crossref] [PubMed]
19. Li Q, Geng S, Luo H, et al. Signaling pathways involved in colorectal cancer: pathogenesis and targeted therapy. Signal Transduct Target Ther 2024;9:266. [Crossref] [PubMed]
20. Pang B, Wu H. Metabolic reprogramming in colorectal cancer: a review of aerobic glycolysis and its therapeutic implications for targeted treatment strategies. Cell Death Discov 2025;11:321. [Crossref] [PubMed]
21. Dunne PD, Arends MJ. Molecular pathological classification of colorectal cancer-an update. Virchows Arch 2024;484:273-85. [Crossref] [PubMed]
22. Cao YD, Sun HW, Li TC, et al. Validated nomograms for non-metastatic colorectal cancer prognosis prediction: a population-based study. Front Oncol 2025;15:1691693. [Crossref] [PubMed]
23. Refaat B, Aslam A, Idris S, et al. Profiling estrogen, progesterone, and androgen receptors in colorectal cancer in relation to gender, menopausal status, clinical stage, and tumour sidedness. Front Endocrinol (Lausanne) 2023;14:1187259. [Crossref] [PubMed]
24. Rodríguez-Santiago Y, Garay-Canales CA, Nava-Castro KE, et al. Sexual dimorphism in colorectal cancer: molecular mechanisms and treatment strategies. Biol Sex Differ 2024;15:48. [Crossref] [PubMed]
25. Xia T, Sun H, Huang H, et al. Androgen receptor gene methylation related to colorectal cancer risk. Endocr Connect 2019;8:979-87. [Crossref] [PubMed]
26. Albasri AM, Elkablawy MA. Clinicopathological and prognostic significance of androgen receptor overexpression in colorectal cancer. Experience from Al-Madinah Al-Munawarah, Saudi Arabia. Saudi Med J 2019;40:893-900. [Crossref] [PubMed]
27. Ozawa T, Matsuda K, Ishihara S, et al. The robust performance of carcinoembryonic antigen levels after adjuvant chemotherapy for the recurrence risk stratification in patients with colorectal cancer. J Surg Oncol 2021;124:97-105. [Crossref] [PubMed]
28. Liu Q, Ran D, Wang L, et al. Association between Ki67 expression and therapeutic outcome in colon cancer. Oncol Lett 2023;25:272. [Crossref] [PubMed]
29. Qin R, Fan X, Huang Y, et al. Role of glucose metabolic reprogramming in colorectal cancer progression and drug resistance. Transl Oncol 2024;50:102156. [Crossref] [PubMed]
30. Tsokkou S, Konstantinidis I, Papakonstantinou M, et al. Sex Differences in Colorectal Cancer: Epidemiology, Risk Factors, and Clinical Outcomes. J Clin Med 2025;14:5539. [Crossref] [PubMed]