A total of 24 male Sprague-Dawley (SD) rats (12-weeks old; weight, ~200 g) were obtained from the Experimental Animal Center of Guizhou Medical University. The animal protocols were approved by the Animal Ethics Committee of Guizhou Medical University (approval no. 2402989; Guiyang, China). All rats were acclimatized for 1 week under controlled temperature conditions at 22-25°C, 55-70% humidity, and received unrestricted access to food and water. They were divided into one Control group and three CRD groups randomly, 6 rats in each group. Two groups of CRD rats were treated with MT (cat. no. M5250; MilliporeSigma) administered intraperitoneally with the dosages of 5 and 10 mg/kg/d, as the low dosage group (CRD + MT-LD) and high dosage group (CRD + MT-HD), while the other groups were treated with vehicle (equivalent volume of normal saline). To explore the association between CRD and erectile function, a CRD model was constructed by altering daily light-dark cycles in rats (24-26). Rats were housed under a 12/12-h light/dark cycle (LD 12:12, lights on at Zeitgeber time (ZT) 0 and lights off at ZT12). The lights for CRD group were switched to LD 2:2 for 4 weeks, the light source was a white, fluorescent lamp (500 lux at cage level), and then weekly body weight monitoring was conducted.

MT administration protocol and solution preparation: MT (MilliporeSigma) was administered intraperitoneally (i.p.) once daily for 4 consecutive weeks, starting from the first day of the CRD modeling protocol. To maintain consistency with the circadian rhythm experiment, MT was injected at a fixed Zeitgeber time (ZT1), i.e., 1 h after lights on (9:00 AM under a 12:12 light-dark cycle). The temporal relationship between MT injection and CRD modeling was as follows: each day, the light cycle was first changed to the CRD schedule (LD 2:2), and MT was administered intraperitoneally within 30 min thereafter. MT is a lipophilic compound with poor solubility in aqueous vehicles. Therefore, a stock solution was first prepared by dissolving 100 mg MT in 4.305 ml dimethyl sulfoxide (DMSO) to obtain a 100 mM stock solution. The stock solution was aliquoted into amber tubes and stored at −20°C protected from light. On each day of administration, the stock solution was freshly diluted with sterile normal saline (0.9% NaCl) to achieve the desired working concentrations (5 mg/kg for low dose, 10 mg/kg for high dose). The final concentration of DMSO in the working solution was ≤0.1% (v/v), a concentration previously shown to have no adverse effects on animal physiology or erectile function. The solution was vortexed and, if necessary, briefly sonicated to ensure complete dissolution before injection. Control animals received an equivalent volume of vehicle (sterile normal saline containing 0.1% DMSO).

The immortalized human umbilical vein endothelial cell line HUVEC-SV40 (SUNNCELL; https://www.app17.com/c163183/products/b3383_p1.html) was used in the present study. HUVECs were maintained in RPMI-1640 medium supplemented with 10% fetal bovine serum (both from Gibco; Thermo Fisher Scientific, Inc.) and 1% penicillin/streptomycin at 37°C under 5% CO₂. To assess dose and time-dependent effects of MT and LPS, HUVECs were seeded in 96-well plates at 1×10⁴ cells/well and cultured for 24 h. Cells were then treated with specified concentrations of MT or LPS. Post-incubation, cell viability was measured using a Cell Counting Kit-8 (CCK-8) kit (cat. no. AR1160; Wuhan Boster Biological Technology, Ltd.) according to the manufacturer's protocol. Briefly, 10 μl of CCK-8 solution was added to each well containing 100 μl of culture medium. The plates were then incubated at 37°C for 2 h in a humidified incubator with 5% CO₂. Absorbance was quantified at 450 nm using a Multiskan FC microplate reader (Thermo Fisher Scientific, Inc.). HUVECs were categorized into seven distinct experimental groups as follows: Control, LPS (1 mg/l), LPS + MT (800 μM), LPS + MT + ML385 (20 μM), LPS + MCC950 (20 μM), LPS + MT + BMS986299 (20 μM), and LPS + NAC (10 mM).

Erectile function was measured from 10:00 to 12:00 a.m, and the penile tissue was subsequently dissected. Moreover, erectile function was assessed by recording the maximum intracavernous pressure (mICP) and mICP/mean arterial pressure (MAP) ratio as previously described (27). Rats were anesthetized by intraperitoneal injection of 3% pentobarbital sodium (30 mg/kg), the carotid was carefully exposed and cannulated with a heparinized detaining venipuncture (26 G) needle to monitor the arterial pressure through a pressure transducer. The cavernous nerve (CN) was carefully separated with a low midline abdominal incision, then a heparinized scalp acupuncture needle was inserted into the penile crus to record ICP through another pressure transducer. When the CN was electrically stimulated (using a voltage of 5 V at frequency of 20 Hz, pulse width of 5 msec, and sustained for 60 sec) (27), the ICP and arterial pressure were simultaneously recorded using a BL420 bio-function experiment system (Chengdu TME Technology Co., Ltd.). The mICP and MAP were analyzed, and the final mICP/MAP ratio and total ICP (area under curve, AUC-ICP) were calculated. At the end of the experiment, all rats were euthanized by intraperitoneal injection of excessive pentobarbital sodium (150 mg/kg), and death was confirmed by cardiac arrest, respiratory arrest and corneal reflex loss.

Penile corpus cavernosum tissues were obtained from rats in the Control, CRD, and CRD + MT-HD groups. Total RNA was isolated with TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc.). Total RNA was extracted, and sequencing libraries were prepared using TruSeq Stranded mRNA Library Prep Kit. The concentration of the final libraries was measured using a Qubit 4.0 Fluorometer with the Qubit dsDNA HS Assay Kit, and the fragment size distribution was assessed using an Agilent 2100 Bioanalyzer with the High Sensitivity DNA Kit. Each library was loaded at a final concentration of 10 nM onto the Illumina NovaSeq 6000 platform (Illumina, Inc.) according to the manufacturer's instructions for paired-end sequencing (2×150 bp). Sufficient double-stranded complementary DNA (dsDNA) for library preparation was generated through SMART pre-amplification (cat. no. 634925/634926; Takara Bio USA, Inc.). The dsDNA was then fragmented using dsDNA fragmentase, and fragments of the desired size were selected with sample purification beads. The resulting ligation product was amplified by PCR to construct the sequencing library. Raw reads obtained from sequencing included adapter sequences and low-quality bases. These were filtered using Cutadapt (version: cutadapt-1.9) (28) to generate high-quality clean reads. The clean data were then aligned to the rat reference genome for gene expression quantification and subsequent bioinformatic analyses.

The predicted protein structures were generated using AlphaFold (https://deepmind.google/technologies/alphafold/). Both structures were subsequently prepared with AutoDockTools-1.5.7 (https://autodock.scripps.edu), including manual removal of water molecules, addition of hydrogens, and other structural refinements. Protein-protein docking was then carried out using the GRAMM web server (29-31). The resulting complex was further optimized in AutoDockTools-1.5.7. Finally, protein-protein interaction (PPI) analysis and visualization were performed using PyMOL (pymol.org).

Penile corpus cavernosum was fixed in 4% paraformaldehyde, paraffin-embedded, and sectioned at 5 μm. Histological analysis included Masson trichrome staining per standard protocols. Smooth muscle and collagen content in penile corpus cavernosum were assessed by calculating their area ratio using Image-Pro Plus software (ver. 6.0; Media Cybernetics, Inc.).

The paraffin-embedded sections of penile corpus cavernosum tissues and HUVECs slides were used for IF to explore the expression of target proteins. The primary antibodies used for incubating sections overnight at 4°C were rabbit anti-eNOS (1:100; cat. no. AF0096; Affinity Biosciences), anti-Nrf2 (1:200; cat. no. 33123-1-AP; Proteintech Group, Inc.), anti-HO-1 (1:100; cat. no. 10701-1-AP; Proteintech Group, Inc.), anti-α-SMA (1:100; cat. no. 14395-1-AP; Proteintech Group, Inc.) and anti-NLRP3 (1:100; cat. no. DF7438; Affinity Biosciences). Then, appropriate secondary antibodies [Alexa Fluor 488-conjugated goat anti-rabbit IgG (H+L); 1:200; cat. no. A32731; Invitrogen; Thermo Fisher Scientific, Inc.] were selected for further incubation. A total of five random fields per group were imaged. Semi-quantitative analysis of relative fluorescence intensity was conducted using Image-Pro Plus software (v6.0).

To detect pyroptotic cells, TUNEL staining was performed using a commercial kit according to the manufacturer's instructions, with modifications for paraffin-embedded sections. Briefly, penile corpus cavernosum tissue sections were dewaxed and rehydrated. After washing with PBS (three times, 5 min each), sections were incubated with proteinase K working solution (1 μl of 100X proteinase K in 99 μl PBS) for 20 min at 37°C to permeabilize the tissue. Following three additional PBS washes, sections were incubated with TdT equilibration buffer for 20 min at 37°C. The buffer was then removed, and labeling working solution (35 μl TdT equilibration buffer+10 μl labeling solution + 5 μl TdT enzyme) was added to each section. Sections were incubated in a humidified chamber for 60 min at 37°C in the dark. After three PBS washes, nuclei were counterstained with DAPI working solution [4 μl DAPI reagent (25 μg/ml) in 96 μl PBS] for 5 min at room temperature in the dark. Sections were washed four times with PBS, mounted with anti-fade mounting medium, and immediately observed under a fluorescence microscope.

For simultaneous detection of Caspase-1 and TUNEL, sections were first processed for TUNEL staining as aforementioned, followed by IF staining for Caspase-1 using rabbit anti-Caspase-1 primary antibody (1:100; cat. no. 31020-1-AP; Proteintech Group, Inc.) and appropriate Alexa Fluor-conjugated secondary antibody. Nuclei were counterstained with DAPI. Co-localization of Caspase-1 and TUNEL signals was assessed to identify pyroptotic cells (18,32). The percentage of Caspase-1⁺/TUNEL⁺ double-positive cells was calculated from at least five random fields per section.

The paraffin-embedded sections of penile corpus cavernosum tissues were fixed in 4% paraformaldehyde at 4°C for 24 h, embedded in paraffin, and cut into sections as aforementioned. Sections were incubated with rabbit anti-eNOS (1:100; Affinity Biosciences), anti-Nrf2 (1:200; Proteintech Group, Inc.), anti-HO-1 (1:100; Proteintech Group, Inc.), anti-Collagen-I (1:200; Abcam) and anti-Collagen-III (1:200; Abcam) overnight at 4°C. Then the sections were washed by PBS and incubated by relevant HRP-conjugated goat anti-rabbit IgG (H+L) secondary antibodies (1:500; cat. no. 31460; Thermo Fisher Scientific, Inc.) for 1 h at room temperature. Five random fields per group were imaged. Semi-quantitative analysis of relative fluorescence intensity was conducted using Image-Pro Plus software (v6.0).

Fresh penile corpus cavernosum tissues were dissected into ≤1 mm³ fragments and fixed in 2.5% glutaraldehyde for 4 h. Samples were post-fixed in 1% osmium tetroxide for 2 h at room temperature. After dehydration through a graded ethanol series and transition in propylene oxide, tissues were embedded in resin and sectioned into 60-80 nm ultrathin slices. Sections were dual-stained with uranyl acetate and lead citrate, then observed and imaged using a SEM.

NO and cGMP levels were quantified in fresh penile corpus cavernosum tissues and HUVECs. NO Assay Kit (cat. no. A013-2-1; Nanjing Jiancheng Bioengineering Research Institute) was applied to detect the level of NO. cGMP ELISA Kit (cat. no. JM-01434R2; JINGMEI; http://jsjmsw.com/) was used to measure the concentration of cGMP. They were performed according to the manufacturer's protocols. The levels of NO and cGMP were normalized to the protein concentration.

Total RNA was extracted from rat penile corpus cavernosum tissues and HUVECs using a Rapid RNA Isolation Kit (cat. no. G3013; Wuhan Servicebio Technology Co., Ltd.). cDNA was synthesized with a PrimeScript™ RT Reagent Kit (cat. no. RR047A; Takara Bio, Inc.) according to the manufacturer's instructions. RT-qPCR was performed on a CFX96 Real-Time System (Bio-Rad Laboratories, Inc.) using TB Green Premix Ex Taq II (cat. no. RR820A; Takara Bio, Inc.). The thermocycling conditions were as follows: Initial denaturation at 95°C for 30 sec, followed by 40 cycles of denaturation at 95°C for 5 sec and annealing/extension at 60°C for 30 sec. A melting curve analysis was performed after amplification to verify the specificity of the PCR products (95°C for 15 sec, 60°C for 1 min, then gradually increasing to 95°C). The relative expression levels of target genes were calculated using the 2^-ΔΔCq method (33), where ΔCt=Ct(target)-Ct(GAPDH) and ΔΔCt=ΔCt(treatme nt)-ΔCt(control). The specific primers used in RT-qPCR are shown in Tables SI and SII.

Penile corpus cavernosum tissues and HUVECs were lysed in RIPA buffer (cat. no. R0010; Beijing Solarbio Science & Technology Co., Ltd.) containing protease inhibitor cocktail (cat. no. HY-K0010; MedChemExpress), followed by sonication and centrifugation (12,000 × g, 15 min, 4°C) to collect supernatants. Total protein concentration was determined by BCA assay (cat. no. AR0146; Boster Biological Technology). Protein samples (30 μg) were separated by 10% SDS-PAGE and transferred to PVDF membranes. Membranes were blocked with 5% BSA (http://www.genenode.com)/TBST (0.1% Tween-20) for 1 h at room temperature, then incubated overnight at 4°C with primary antibodies: Rabbit anti-Nrf2 (1:2,000) and anti-HO-1 (1:1,000). After primary incubation, membranes were probed with species-matched HRP-conjugated secondary antibodies (1:5,000) for 1 h at room temperature. Band intensities were quantified using ImageJ software (version 1.53t; National Institutes of Health) and normalized to GAPDH.

Cellular ROS and mitochondrial ROS (mtROS) levels were assessed using the fluorescent probes H2DCFH-DA (cat. no. D6470; Beijing Solarbio Science & Technology Co., Ltd.) and MitoSOX™ Red Mitochondrial Superoxide Indicator (cat. no. HY-D1055; MedChemExpress), respectively.

For tissue sample preparation, penile corpus cavernosum tissue fragments were enzymatically digested at 37°C for 30 min, filtered through a 70-μm nylon mesh, and centrifuged at 500 × g. The pellet was washed twice with PBS and subsequently incubated with 10 μM H2DCFH-DA at 37°C for 1 h. After incubation, the samples were centrifuged at 1,000 × g and washed again twice with PBS. Fluorescence intensity was measured using a microplate reader at Ex/Em=488/525 nm.

For the cell-based assay, HUVECs cultured on adherent slides were incubated with 10 μM H2DCFH-DA at 37°C for 30 min under light-protected conditions, following the manufacturer's instructions. Fluorescence images were acquired immediately using a fluorescence microscope, with five random fields captured per sample. Semi-quantitative analysis of fluorescence intensity was performed using Image-Pro Plus software (v6.0). For mtROS detection, the MitoSOX™ Red probe was used according to the manufacturer's protocol, and fluorescence was detected at Ex/Em=510/580 nm under light-protected microscopy.

MDA, GSH, SOD and T-AOC levels were quantified in rat CC tissue homogenates and HUVECs lysates using commercial kits: MDA (cat. no. BC0025), GSH (cat. no. BC1175), SOD (cat. no. BC5165) and T-AOC (cat. no. BC1315; all from Beijing Solarbio Science & Technology Co., Ltd.). Assays were performed strictly according to manufacturer's protocols. All values were normalized to total protein concentration.

MMP was assessed using JC-1 staining (cat. no. E-CK-A301; Elabscience Biotechnology, Inc.) and analyzed via fluorescence microscopy.

All results were presented as mean ± standard deviation and analyzed with GraphPad Prism software (version 9.5.0; Dotmatics). Shapiro-Wilk test was used for normality and Levene's test was used for homogeneity of variance before applying parametric tests. For data that did not meet normality or equal variance assumptions, Kruskal-Wallis H test followed by Dunnett's multiple comparisons test was utilized. It was clarified that Dunnett's test was used for comparisons against the control group (primary aim), while Tukey's or Dunnett's tests were applied for all pairwise comparisons or non-parametric data, respectively. P<0.05 was considered to indicate a statistically significant difference. The number of independent biological replicates were as follows: n=6 per group for animal experiments; n=3 independent culture batches for cell experiments, each with technical triplicates.

As was shown, the CRD group (44.91±7.51 mmHg) showed significantly lower mICP than Control group (76.82±3.81 mmHg, P<0.0001); however, the CRD + MT-LD (57.03±2.03 mmHg, P=0.0008) and CRD + MT-HD (66.15±2.49 mmHg, P<0.0001) groups significantly increased mICP than the CRD group (Fig. 1A and C); the total ICP also showed a similar trend (Fig. 1A and D). Meanwhile, the CRD group significantly reduced mICP/MAP ratio (0.38±0.07) than Control group (0.64±0.03, P<0.0001), while the CRD + MT-LD (0.50±0.02, P=0.0013) and CRD + MT-HD (0.59±0.05, P<0.0001) significantly increased it (Fig. 1A and F). Moreover, the bICP and MAP were consistent among the four groups (P>0.05) (Fig. 1A, B and E). These results suggested that CRD impaired rats' normal erectile function, while MT demonstrated a protective effect for CRD-induced ED. In addition, these procedures had no impact on rats' mean body weight (P>0.05) (Fig. S1), which excluded potential confounding such as systemic metabolic effects.

The IF and IHC staining of penile corpus cavernosum revealed that CRD significantly reduced eNOS expression, which was partially preserved by MT-LD and MT-HD (Fig. 1G-J). Meanwhile, the penile corpus cavernosum of CRD group also demonstrated significantly decreased NO and cGMP concentrations, which was also inhibited by MT-LD and MT-HD (Fig. 1K and L). These in vivo results also suggested that MT could preserve the CRD-induced ED.

To further confirm these findings, LPS-treated HUVECs were selected to imitate CRD stimulation in vitro (34,35) to verify the underlying mechanism, while the specific concentrations and exposure times of LPS and MT for HUVECs were provided in the following statement. Firstly, HUVECs were treated with different concentrations of LPS and MT for different time intervals to evaluate their cytotoxicity. It was found that the concentration of LPS up to 0.5 mg/l and MT up to 1,600 μM had significant effects on the viability of HUVECs for 24 h (Fig. S3A and C). However, LPS with 1.0 mg/l for 24 h reduced HUVECs viability to 52.03% than the control group, indicating a cytotoxic effect with an approximate IC₅₀ of 1.0 mg/l, whereas MT with 800 μM was well-tolerated with no significant effect on viability even after 48 h (Fig. S3B and D). Thus, HUVECs treated with LPS at 1.0 mg/l were used to establish a CRD stimulation model in vivo. Therefore, cells were incubated with 400 or 800 μM MT for 2 h before exposure to 1.0 mg/l LPS. The protective effect of 800 μM MT was significantly more pronounced, a finding consistently corroborated by both CCK-8 and phase-contrast microscopy assay (Fig. S3E and F). As was shown, LPS could significantly reduce eNOS expression while decreased NO and cGMP levels, confirming that CRD could impair HUVECs; however, these damages were reversed by MT (Fig. 1M-P).

Collectively, these in vivo and in vitro results suggested that CRD impaired rats' erectile function and injured HUVECs by disrupting eNOS-NO-cGMP pathway, however, MT treatment could reduce these alterations to ameliorate CRD-induced ED.

To confirm that photoperiodic manipulation successfully induced circadian rhythm disorders, the expression of core clock genes was first analyzed in the corpus cavernosum tissue of a model rat by sequencing data. As revealed in Fig. S4, firstly, RNA quality was evaluated using the Affy package in R software (version 4.2.1; https://www.R-project.org) and gene expression levels were normalized using limma package (https://bioconductor.org/packages/limma/) and StringTie software. (https://github.com/gpertea/stringtie). Visual analysis was conducted using a box plot of relative logarithmic expression levels, and the results demonstrated favorable consistency among samples within each group (Fig. S4A). Secondly, the limma package and Wilcoxon test were used to screen differentially expressed genes (DEGs), with a screening threshold set at |logFC|>0.5 and FDR<0.05. The results showed that a total of 872 DEGs were identified between the Control group and the CRD group, of which 389 genes were significantly downregulated, and 483 genes were significantly upregulated (Fig. S4B). Further differential expression analysis was conducted on the core clock genes in the selected DEGs. The heatmap results identified that 12 core clock genes, including Per1/2/3, Npas2, Csnk1e, Cry1/2, Nr1d1/2, Clock, Rorα and Timeless, were generally upregulated in the CRD group, indicating the successful construction of a CRD model by changing the light cycle (Fig. S4C). Secondly, through IF, IHC, WB and RT-qPCR experiments, it was found that compared with the control group, the protein and mRNA expression levels of the core circadian clock gene Per1 in the corpus cavernosum tissue of CRD group rats were significantly increased (Fig. S5A-G). This result is consistent with RNA-seq data, and a CRD model was successfully established at the functional and molecular levels.

Haspel et al (34) demonstrated that LPS exposure in vitro and in vivo leads to reprogramming of the circadian clock, including changes in the expression of core clock genes such as Per1, Per2 and Bmal1. Ryzhikov et al (35) further demonstrated that LPS induces a CRD-like state in cultured cells, characterized by loss of rhythmic expression of clock control genes. In the authors' preliminary experiments, LPS treatment significantly upregulated the expression of Per1 in HUVECs (Fig. S6), simulating the molecular clock disruption observed in the corpus cavernosum of rats with CRD. This validates the use of LPS treated HUVEC as a relevant in vitro model for studying the mechanism of CRD-induced endothelial injury.

For the microstructural changes of penile corpus cavernosum, Masson trichrome staining showed that SMC/Collagen ratio in CRD group was significantly reduced, which was increased in CRD + MT-LD and CRD + MT-HD groups (Fig. 2A and B). IF staining revealed that the reduced α-SMA in CRD group was ameliorated in CRD + MT-LD and CRD + MT-HD groups (Fig. 2C and D), while IHC staining presented that CRD + MT-LD and CRD + MT-HD significantly attenuated Collagen-I and III deposition than CRD group (Fig. 2E-H). For the SEM of penile corpus cavernosum, CRD group presented that collagen fibers (→) in the outer layers of tunica albuginea were irregularly lined and sinusoidal (S) spaces were significantly narrowed; however, these collagen fibers (→) were regularly lined and sinusoidal (S) spaces were widened in CRD + MT-HD group (Fig. 2I and J). All these demonstrated that CRD impaired the microstructural integrity of penile corpus cavernosum, which were partially reversed through MT intervention.

To elucidate the therapeutic mechanisms of MT on CRD-induced ED, a systematic bioinformatics analysis was performed. Firstly, bioinformatic screening across multiple databases [including GeneCards (https://www.genecards.org), STRING (https://string-db.org), DAVID, PubChem (https://pubchem.ncbi.nlm.nih.gov), CTD (https://ctdbase.org) and OMIM (https://www.omim.org)] identified 1,500 ED-associated genes and 825 predicted MT target genes, with 284 overlapping targets selected for further analysis (Fig. 3A). PPI network construction and analysis of these overlapping targets (Fig. 3B and C) identified 63 core targets (Fig. 3D), which notably included oxidative stress-related genes (OSRGs). These findings establish oxidative stress as a core pathway through which MT exerts its therapeutic effects in ED.

The oxidative stress level of rats' penile corpus cavernosum was thus evaluated. The oxidative stress markers (ROS and MDA) in CRD group were significantly increased and antioxidant parameters (GSH, SOD and T-AOC) were decreased, all of which were partially reversed by MT-LD and MT-HD (Fig. 3E-I). All these inferred that oxidative imbalance is an essential pathological process for CRD-induced ED, which is also a core therapeutical mechanism for MT.

Sequencing for penile corpus cavernosum was further performed, which identified 1,320 DEGs between CRD and CRD + MT-HD groups (Fig. 4A); all the top 20 OSRGs exhibited significantly differential expression. Notably, the core Nrf2 and HO-1 of OSRGs in CRD + MT-HD group were significantly upregulated compared with CRD group (Fig. 4B), suggesting their crucial role in mediating the therapeutic effect of MT. To further investigate whether MT directly interacts with Nrf2 and HO-1, molecular docking was performed. The results showed binding energies of -5.7 kcal/mol for Nrf2 and -6.4 kcal/mol for HO-1, indicating stable binding between MT and both proteins (Fig. 4C and D). Given the core role of Nrf2 (36) and HO-1 (37) in oxidative stress response, it was hypothesized that MT primarily exerted antioxidant effects through activating Nrf2/HO-1 pathway.

In vivo analyses (IF and IHC of penile corpus cavernosum) revealed that Nrf2 and HO-1 protein expression levels were slightly increased in CRD rats, whereas MT intervention significantly upregulated their levels (Fig. 4E-L); this pattern was corroborated in BCNI rat model, that the disease itself did not alter Nrf2 or HO-1 expression, yet antioxidant treatment elevated them (38). To validate these findings, the LPS-treated HUVECs were selected to imitate CRD stimulation in vitro (34,35). Consistent with the in vivo data, western blotting and RT-qPCR analysis demonstrated that the levels of Nrf2 and HO-1 in the LPS + MT group were significantly higher than LPS group, while Nrf2 inhibitor (ML385) significantly reduced Nrf2 and HO-1 expression compared with LPS + MT group (Fig. 4M-R).

It has been previously proved by the authors that oxidative stress was involved in the therapeutic effects of MT on CRD-induced ED; whether this occurs through Nrf2/HO-1 pathway remains unclear. The present in vitro experiments demonstrated that MT reduced the LPS-induced elevation of oxidative stress markers (ROS, mtROS and MDA) while rescued reduction of antioxidant parameters (MMP, GSH, SOD and T-AOC); notably, these beneficial effects were partially reversed by Nrf2 inhibitor (ML385), indicating that MT exerted antioxidant effects through activating Nrf2/HO-1 pathway (Fig. 5A-J).

To further explore the downstream mechanism of how MT regulates CRD-induced ED through oxidative stress, sequencing analysis was subsequently performed on penile corpus cavernosum tissues from CRD rat model groups. A total of 37 pyroptosis-related DEGs were identified by intersecting RNA-seq data from CRD and CRD + MT-HD groups with pyroptosis-related genes from GeneCard (Fig. 6A). PPI network analysis was constructed using STRING and visualized in Cytoscape (version 3.10.3; https://web.cytoscape.org). Functional enrichment analysis (DAVID; https://davidbioinformatics.nih.gov) showed significant involvement of the 'NOD-like receptor signaling pathway' (Fig. 6B). Based on these findings, along with the established link between NLRP3 and pyroptosis (39) (Fig. 6C), it was therefore hypothesized that NLRP3-mediated pyroptosis may contribute to the development of CRD-induced ED.

The IF staining of penile corpus cavernosum was performed to evaluate the core protein of pyroptosis (NLRP3, Caspase-1 and GSDMD). As was demonstrated, NLRP3 protein expression in rats with CRD was significantly increased, which was markedly attenuated following CRD + MT-LD and CRD + MT-HD groups (Fig. 6D and E). Meanwhile, both the caspase-1⁺/TUNEL⁺ index [identified as pyroptosis (18,32)] (Fig. 6F and G) and GSDMD expression (Fig. 6H and I) revealed consistent alterations of NLRP3, rising under CRD and declining after MT. The in vitro LPS-treated HUVECs also presented similar consequences. LPS significantly elevated the key pyroptosis protein levels (NLRP3, caspase-1 and GSDMD); however, all these alterations were partially reversed by MT treatment (Fig. 6J-O). These demonstrated that CRD promoted pyroptosis, whereas MT treatment alleviated this process.

To further elucidate the mechanism by which MT mitigates LPS-induced pyroptosis in HUVECs, a dual-intervention strategy was employed using the NLRP3 inhibitor (MCC950) and agonist (BMS986299). IF analysis demonstrated that both the NLRP3 inhibitor (MCC950) and MT similarly suppressed the LPS-induced upregulation of key pyroptosis markers (NLRP3, Caspase-1 and GSDMD). By contrast, the NLRP3 agonist (BMS986299) attenuated the inhibitory effect of MT on these LPS-induced pyroptosis markers (Fig. 6J-O). The present findings suggested that CRD could lead ED through inducing pyroptosis, whereas MT counteracted this effect through inhibiting NLRP3 activation.

To further explore whether MT alleviates NLRP3-mediated pyroptosis through oxidative stress, we incorporated the Nrf2 inhibitor (ML385) and NLRP3 inhibitor/agonist (MCC950/BMS986299) to enable a systematic comparison across these targeted interventions. IF analysis revealed that MT significantly suppressed the LPS-induced up-regulation of pyroptosis markers, an effect that was also found by NLRP3 inhibitor (MCC950). Meanwhile, Nrf2 inhibitor (ML385) completely abolished the protective effect of MT, an effect that was also found by NLRP3 agonist (BMS986299) (Fig. 6J-O). All these proved the core mechanism that MT alleviates NLRP3-mediated pyroptosis through suppression of oxidative stress.

N-Acetyl-L-cysteine (NAC), an amino acid derivative of glutathione and cysteine, serves as a potent scavenger of ROS (40) and plays a crucial role in cellular antioxidant stress (41). To evaluate the inhibitory effect of NAC on LPS-induced NLRP3 activation, 10 mM NAC was pretreated for 3 h before LPS intervention in HUVECs. Initially, the impact of NAC on LPS-treated HUVECs was assessed. LPS induced a significant redox imbalance, that significantly elevated oxidative stress markers (ROS, mtROS and MDA) and decreased antioxidant parameters (MMP, GSH, SOD and T-AOC), all of which were partially reversed by NAC treatment (Fig. 7A-J), suggesting the ability of anti-oxidative stress.

Subsequently, it was explored whether the anti-oxidative effect of NAC could influence CRD-induced pyroptosis. As expected, LPS-treated HUVECs presented elevated pyroptosis activation, as indicated by increased expression of NLRP3, caspase-1 and GSDMD; however, this activation was partially reversed by NAC (Fig. 7K-P), suggesting that NAC reduced CRD-induced pyroptosis by inhibiting oxidative stress.

Furthermore, the endothelial function was also evaluated, that the endothelial function indicators of NO and cGMP in LPS-treated HUVECs were significantly decreased, however, these reductions were partially reversed by NAC (Fig. 7Q and R).

These data confirmed that NAC mitigated LPS-induced endothelial impairment by reducing oxidative stress to curb NLRP3-mediated pyroptosis, highlighting the central role of the oxidative stress-pyroptosis axis. In conclusion, all these proved that CRD-induced ED by inhibiting oxidative stress mediated pyroptosis via Nrf2/HO-1 axis, which was also the core therapeutic mechanism of MT.

The circadian rhythm represents a fundamental intrinsic timekeeping system (42). Epidemiological and clinical studies have consistently shown that CRD exposures, such as shift work (43), frequent time-zone changes (44) and chronic sleep loss (45), are strongly associated with an increased risk of ED, a finding also confirmed by the authors' previous studies (9,10). Based on direct in vivo and in vitro evidence, the present study establishes CRD as an independent risk factor for ED, thereby providing direct implications for understanding the high prevalence of ED among shift workers and other high-risk populations. The physiological process of erection critically depends on the NO-mediated vasodilatory signaling pathway within penile endothelial cells (11). The current investigation reveals that CRD substantially impairs the activity of this central eNOS-NO-cGMP pathway, which constitutes the direct molecular foundation of ED pathogenesis (46). Notably, beyond the disruption of this crucial signaling axis, the deeper pathophysiological consequences of CRD extend to substantial tissue structural damage and programmed cell death. These observations prompted a crucial investigation into the specific downstream pathways through which CRD mediates its damaging effects.

The circadian rhythm system and cellular redox state maintain a sophisticated bidirectional interplay (47). CRD disrupts this coordination, resulting in both overproduction and inefficient clearance of ROS (48). At the molecular level, core circadian components directly regulate the expression of numerous genes involved in antioxidant defense; however, CRD leads to the dysregulation of these genes, thereby increasing susceptibility to oxidative stress (49). Concurrently, oxidative stress products (for example, ROS) can feed back and impair the function of circadian rhythm proteins (for example, PER and EZH2), forming a vicious cycle (50). At the physiological level, CRD disrupts metabolism under the control of the circadian rhythm (for example, glycolysis and lipophagy). This metabolic reprogramming elevates oxidative byproducts (for example, ROS), while the capacity for their clearance is compromised due to CRD, collectively promoting oxidative stress (51). In the present study, penile corpus cavernosum tissue from rats with CRD displayed clear signs of oxidative stress; the current data strongly support a model in which oxidative stress acts as a critical link between CRD and downstream pathological injury. Furthermore, having identified oxidative stress as the central pathway in the therapeutic effect of MT on CRD-induced ED through transcriptomic analysis, it was next sought to elucidate how MT mediates its antioxidant protection at the molecular level.

Nrf2 is a master transcriptional regulator of cellular antioxidant responses (52). Under physiological conditions, Nrf2 is constitutively bound to Keap1, localizes in the cytoplasm, and undergoes continuous proteasomal degradation (53). During oxidative stress, Nrf2 dissociates from Keap1 and translocates to the nucleus (54), where it binds to antioxidant response elements (AREs) (55) and initiates the transcription of antioxidant genes, including the critical cytoprotective enzyme HO-1 (56). HO-1 exerts potent antioxidant and cytoprotective effects by catalyzing the degradation of heme into biliverdin, carbon monoxide and free iron (57). Core circadian rhythm proteins regulate the expression and activity of Nrf2, generating circadian oscillations in cellular antioxidant capacity (58); Nrf2 knockout attenuates the rhythmicity of circadian genes while inducing aberrant oscillations in stress-response genes (59). Moreover, CRD impairs the time-giving function of the circadian rhythm, preventing pre-activation of antioxidant defenses prior to peak oxidative stress, thereby compromising defensive capacity when it is most needed (58). Ultimately, a vicious cycle emerges: CRD causes Nrf2 dysfunction and impairs antioxidant capacity, while accumulated ROS further disrupts circadian rhythm operation (60). Exogenous MT has been shown to activate the Nrf2 pathway and reduce oxidative stress in various models (61-63). The current data demonstrate that MT mediates its antioxidant activity by activating the Nrf2/HO-1 axis, the central transcriptional regulator of endogenous antioxidant responses, thus rebuilding cellular antioxidant capacity and counteracting CRD-induced oxidative damage.

If oxidative stress acts as an upstream event, what is its downstream effector target? In recent years, pyroptosis has been recognized as a key mechanism in the pathogenesis of multiple organ injuries (64). Pyroptosis is a novel form of inflammatory programmed cell death. The present study is the first, to the best of our knowledge to show the activation of NLRP3-mediated pyroptosis in a CRD-induced ED model. The mechanisms to establish the causal relationship between oxidative stress and pyroptosis in the present model were systematically analyzed. First, it was observed that MT alleviates pyroptosis by suppressing oxidative stress. This suggests an association between these processes. When Nrf2 was inhibited, the protection was abolished. This confirmed that Nrf2-dependent antioxidant activity is indispensable for preventing pyroptosis. The present data support that MT inhibits NLRP3-mediated pyroptosis mainly by activating Nrf2/HO-1 and reducing oxidative stress. This is shown by the reversal of MT's effects by the ROS scavenger NAC. However, Nrf2 may also directly suppress NLRP3 inflammasome activation through non-antioxidant mechanisms. Some studies have shown that Nrf2 can downregulate NLRP3 transcriptionally by binding to AREs in its promoter (65-67).

Nrf2 has also been reported to interact physically with NLRP3 or its upstream regulators (for example, TXNIP) and inhibit inflammasome assembly (68,69). In the present study, the possibility that MT-activated Nrf2 directly represses NLRP3 activation cannot be excluded. Future studies with Nrf2 mutants that keep transcriptional activity but lack antioxidant capacity, or chromatin immunoprecipitation assays to examine Nrf2 binding to the NLRP3 promoter, would clarify the roles of antioxidant-dependent and direct pathways.

The core innovations of the present study are reflected in the following three aspects: i) First report of the therapeutic effect of MT on CRD-induced ED: Although MT has shown protective effects in diabetic ED and neurogenic ED models, its effect on CRD-related ED has never been investigated. The present study is the first to demonstrate that MT significantly improves erectile function in CRD rats, restoring the ICP/MAP ratio and preserving the eNOS-NO-cGMP pathway. i) First revelation of the 'oxidative stress-pyroptosis' axis in CRD-related ED: Previous mechanistic studies on ED have not linked CRD with the oxidative stress-pyroptosis axis. The present study is the first to establish that CRD triggers NLRP3-mediated pyroptosis through oxidative stress, and that MT interrupts this cascade via Nrf2/HO-1 activation, representing a completely new molecular pathway in iii) CRD-induced ED. Clinical translational value: CRD-related ED is a modern lifestyle-associated condition affecting shift workers, frequent travelers, and individuals with chronic sleep deprivation. These patients often have contraindications to PDE5is (for example, due to nitrate use for cardiovascular diseases) or suboptimal responses because of endothelial dysfunction. As an endogenous, low-toxicity, multi-target agent, MT represents a promising alternative or adjunctive therapy. The present study provides a strong preclinical rationale for future clinical trials evaluating MT supplementation in CRD-exposed populations with ED.

There are certain limitations to the present study. First, the in vitro experiments were performed using HUVECs rather than corpus cavernosum endothelial cells (CCECs). These two cell types differ in tissue origin, phenotype and physiological function; HUVECs may not fully recapitulate the in vivo pathophysiological state of the corpus cavernosum. Future studies should use primary CCECs for validation.

Second, while the present study measured mitochondrial ROS and membrane potential, key indicators of mitochondrial oxidative stress, it is recognized that a more comprehensive assessment of mitochondrial homeostasis would strengthen the mechanistic understanding. Mitochondrial quality control involves not only redox balance but also dynamic changes in morphology (fusion/fission), biogenesis and mitophagy. Future studies should examine mitochondrial ultrastructure using transmission electron microscopy to visualize cristae integrity and swelling. Additionally, markers of mitochondrial biogenesis (PGC-1α, NRF1 and TFAM) and mitophagy (Parkin, PINK1 and LC3B) would help determine whether MT promotes the removal of damaged mitochondria or enhances the generation of healthy ones. It has been previously demonstrated that MT preserves mitochondrial function in various oxidative stress models by promoting PINK1/Parkin-dependent mitophagy (70). It would be of great interest to test whether similar mechanisms operate in CRD-induced ED. The current findings provide a strong foundation for such future investigations, which has been added as a key direction in the authors' ongoing studies.

Third, although programmed cell death pathways (pyroptosis, apoptosis and necroptosis) dynamically interact and may lead to integrated processes such as PANoptosis, pyroptosis was only assessed in endothelial cells. Markers of apoptosis (caspase-3), ferroptosis (GPX4), or necroptosis (RIPK3) were not systematically detected. Therefore, a comprehensive understanding of the individual contributions and crosstalk among different cell death modalities under pathological conditions is lacking.

Fourth, regarding the detection of pyroptosis, total GSDMD protein expression and caspase-1/TUNEL double staining were only assessed, without measuring the gold-standard indicators, including the cleaved N-terminal fragment of GSDMD (GSDMD-N). Thus, the conclusions are based primarily on NLRP3 and caspase-1 activation and morphological changes. Future studies should include these additional endpoints to provide more definitive evidence of pyroptosis.

In conclusion, it was demonstrated that CRD-induced ED by triggering an oxidative stress-pyroptosis cascade. Conversely, MT treatment effectively counteracts this pathology by activating the Nrf2/HO-1 pathway to suppress oxidative stress, thereby attenuating NLRP3-mediated pyroptosis and ultimately restoring erectile function. These results provide the first systematic evidence for the central role of the oxidative stress-pyroptosis axis in CRD-induced ED, establishing a solid theoretical foundation for MT as a promising therapeutic strategy for CRD-related ED.