The human TNBC cell line MDA-MB-231 (obtained from Nawah Scientific, Egypt, which supplies cells from ATCC) was cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Biowest, France) supplemented with 10% fetal bovine serum (FBS) (Biowest, France), 4.5 g/L glucose, 1% penicillin‒streptomycin, and L-glutamine. The cells were kept at 5% CO₂ and 37 °C (Schmittgen and Livak, 2008). In accordance with the experimental protocol, MDA-MB-231 cells were seeded in 96-, 24-, or 6-well plates and treated with various concentrations of chemotherapeutic agents (doxorubicin, capecitabine, or paclitaxel; MedChemExpress, United States) and/or the 3-MST inhibitor HMPSNE (Molport, Latvia) (Li et al., 2017). Doxorubicin and capecitabine were dissolved in DEPC water, and paclitaxel and HMPSNE were dissolved in DMSO. The concentration of DMSO did not exceed 0.001%. This concentration was tested, and no toxicity was observed.

MDA-MB-231 TNBC cells (20,000 cells/well) were seeded in 96-well plates and treated the next day according to the experimental treatment groups as follows: the control group, drug-only treatment group, drug- and HMPSNE-treated group, and HMPSNE-only treatment group.

The cell culture medium was removed after 48 h. MTT powder (0.1 g) was dissolved in 20 mL of phosphate-buffered saline (PBS) to prepare the MTT stock solution. For each experiment, the stock solution was diluted 1:10 in DMEM supplemented with 10% FBS to obtain the MTT working solution. A volume of 200 µL of the MTT working solution was then added to each well containing MDA-MB-231 cells. The mixture was then incubated for 4 h. Following incubation, the formazan crystals were dissolved in 200 µL of a DMSO/ethanol solution (1:1, v/v) prior to absorbance measurement. Absorbance was measured at 595 nm (Li et al., 2018). The experiments were performed in triplicate and independently repeated three times.

In black 96-well plates, MDA-MB-231 cells (20,000 cells/well) were seeded and subjected to the previously mentioned treatments (see the MTT cell viability assay Section 2.2). Hydrogen sulfide levels were measured using 7-azido-4-methylcoumarin (AzMC; Sigma-Aldrich, Germany). AzMC was initially dissolved in DMSO and subsequently diluted in HBSS to achieve a final working concentration of 100 μM, with a final DMSO concentration of 0.1%.

A volume of 100 µL of the AzMC working solution was added to the cells, followed by incubation for 1 h at 37 °C.

The fluorescence was measured via a Victor multilabel plate reader at 365 and 450 nm for excitation and emission, respectively (Santos et al., 2023). The experiments were performed in triplicate and independently repeated three times.

In 24-well plates, MDA-MB-231 cells (60,000 cells/well) were seeded, subjected to treatment (see the MTT cell viability Section 2.2), and then incubated for 48 h. To allow colonization, 1,000 cells from each treatment group were trypsinized from the initial cell culture and then, plated per well on a 6-well plate and cultured for 7–10 days. After fixation with 6% glutaraldehyde, the colonies were stained with 0.5% crystal violet. The colonies were manually counted (Youness et al., 2021). The experiments were performed in triplicate and independently repeated three times.

The in vitro migration capacity of the cells was assessed via a modified Boyden chamber assay (BD Biosciences, Bedford, MA, United States). MDA-MB-231 cells were initially seeded in 24-well plates at a density of 60,000 cells/well and treated with HMPSNE, doxorubicin, or their combination according to the experimental design. After 48 h of treatment, the cells were trypsinized, counted, and resuspended in low-serum medium (DMEM supplemented with 1% FBS). Transwell inserts with a 5 µm pore size were placed into 24-well plates. A cell suspension containing 50,000 cells in 300 µL of low-serum medium was added to the upper chamber of the Transwell for each treatment group. The lower chamber was filled with 700 µL of high-serum medium (DMEM supplemented with 20% FBS) to serve as a chemoattractant. After 20–24 h of incubation, non-migrated cells were removed from the upper surface of the membrane using a cotton swab. The migrated cells on the lower surface were fixed and stained with 1% crystal violet for 30 min. The membranes were visualized via an inverted brightfield microscope. Following image acquisition, the crystal violet was eluted by transferring the inserts into 700 µL of 2% ethanol. The optical density was quantified by transferring 200 µL of the eluted dye in triplicate to a 96-well plate, and the absorbance was measured at 595 nm via a Victor 1420 multilabel plate reader. The experiments were performed in triplicate and independently repeated three times.

Total RNA was extracted from MDA-MB-231 cells via the Biozol reagent (Invitrogen, United States). Using a RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific, United States), reverse transcription was carried out according to the manufacturer’s instructions (Bullock et al., 2024).

The relative quantification method was used to analyze the expression of CD44 via the StepOne™ Plus Real-Time (RT) PCR system (Applied Biosystems, United States). The following TaqMan® real-time PCR assay was used: CD44 (Thermo Fisher Scientific, Hs01075864_m1).

The housekeeping gene used was β-actin, which was labeled with VIC reporter dye (Hs01060665_g1). Relative quantification was performed via the 2^−ΔΔCT method (Schmittgen and Livak, 2008).

After being seeded in 6-well plates at 150000 cells per well, the MDA-MB-231 cells were treated in the same manner as described in Section 2.2 and incubated for 48 h. Following trypsinization, the cells underwent two rounds of washing before being resuspended in PBS containing 1% FBS (10⁶ cells/100 µL). The cells were then treated with a rat anti-CD44 FITC-conjugated antibody (Sigma Aldrich, United States) for 30 min at 4 °C. After washing, the labeled cells were resuspended in PBS containing 1% FBS. FACS analysis was carried out via Beckman Coulter’s CytoFLEX. CytExpert software was used to analyze the data (Alkaraki et al., 2020).

The cells were scraped from the well plate surface and lysed via RIPA lysis buffer (Thermo Fisher Scientific) supplemented with 1x Halt™ Protease and Phosphatase Inhibitor Cocktail (Thermo Fisher Scientific). The lysates were sonicated for 15 s in an ultrasonic bath (XUBA3, Grant, United Kingdom). Protein concentrations were measured via a BCA protein assay kit (Thermo Fisher Scientific) on a SpectraMax iD5 (Molecular Devices, CA, United States). Proteins were reduced and denatured by boiling in lithium dodecyl sulfate (LDS) sample buffer (Bolt™ LDS Sample Buffer, Invitrogen, Thermo Fisher Scientific) or (NuPAGE™ LDS Sample Buffer, Invitrogen) at 95 °C for 5 min. Equivalent amounts of protein were then loaded and separated on Bolt™ 4%–12% Bis-Tris Plus Gels (Invitrogen) or NuPAGE™ 3%–8% Tris-acetate Gels under a constant voltage of 100 V. NuPAGE™ 3%–8% Tris-Acetate gels were used only for detecting the mTORC signals as this gel system is specifically designed for high-molecular-weight proteins and allowed efficient separation and detection of mTOR. However, all other proteins were detected using Bolt™ 4%–12% Bis-Tris gels as it is the preferred for broad-range and smaller proteins.

Proteins were transferred to PVDF membranes via dry transfer via the iBlot™ 2 System (Invitrogen). The membranes were blocked with 5% milk in TBS/0.1% Tween (TBST) for 1 h at RT. Primary antibodies diluted in 5% BSA-TBST or 5% milk-TBST were incubated for 1 h at RT or overnight at 4 °C. Following incubation, the membranes were washed three times with TBST and then incubated for 1 h at RT with either anti-rabbit IgG or anti-mouse IgG HRP-conjugated secondary antibodies (1:5,000 dilution in 5% milk-TBST). After three additional washes with TBST, the blots were developed via Radiance Plus chemiluminescence (AC2103, Azure Biosystems, CA, United States) on an Azure Imaging System 300 (Azure Biosystems). Signal intensities were quantified via ImageJ software (NIH, Bethesda, MD, United States), with band intensities normalized to those of β-actin as the loading control. Representative blots from at least four independent experiments are shown.

We used the following antibodies: rabbit monoclonal anti-CBS (D8F2P), anti-AKT (pan) (C67E7), anti-p-Akt (Ser473) (D9E), anti-GSK-3β (D5C5Z), mTOR (7C10) and anti-mouse IgG HRP-linked antibodies, which were purchased from Cell Signaling Technology (Danvers, MA, United States). Mouse monoclonal anti-β-actin (AC-15) was obtained from Sigma‒Aldrich (Burlington, MA, United States). Anti-rabbit IgG (H + L) cross-adsorbed secondary antibody-HRP was purchased from Invitrogen (Thermo Fisher Scientific). Rabbit polyclonal anti-3-MST (ab224043), anti-CSE (ab151769), anti-TST (ab231248) and rabbit monoclonal anti-ETHE1 (ab174302) were acquired from Abcam (Cambridge, England).

The normality test was performed via the Shapiro‒Wilk test. When two independent groups were compared, an unpaired Student’s t-test was employed for all normally distributed data. For data from more than two groups, ordinary one-way ANOVA was used for all normally distributed data. The means ± standard errors of the means (SEMs) are used to express the data. Statistical significance was defined as * = p < 0.05 and ** = p < 0.01. GraphPad Prism 8.0 software was used for the statistical analysis.

MDA-MB-231 cells were exposed to different concentrations of doxorubicin, capecitabine, and paclitaxel for 48 h (Augsburger et al., 2020; Ascencao et al., 2021; Hellmich et al., 2015; Kajsik et al., 2022; Stokes et al., 2018; Szabo et al., 2014; Vogel and Marcotte, 2012; Zhang et al., 2024). The percentage of viable cells was calculated relative to the control. MDA-MB-231 cells were resistant to paclitaxel; only at 30 nM did this drug result in a slight decrease in cell viability (Figure 2A). Capecitabine was found to reduce the viability of the cells, and this reduction was statistically significant at concentrations ranging from 30 to 1000 µM (Figure 2B). Compared with other chemotherapeutic agents, doxorubicin decreased cell viability in a concentration range of 0.1–10 µM in a concentration-dependent manner (Figure 2C).

Next, we tested the effects of inhibiting 3-MST via HMPSNE (100 µM) on the cytotoxic effects of the three chemotherapeutic agents used in our study. The addition of HMPSNE to paclitaxel (10 nM) significantly decreased the percentage of viable cells. (Figure 3A), compared with paclitaxel alone. The addition of HMPSNE (100 µM) to 10 µM capecitabine did not have a significant effect (Figure 3B). The addition of HMPSNE (100 µM) increased the cytotoxic effect of 1 µM doxorubicin (Figure 3C). However, the HMPSNE alone tended to decrease in viability.

HMPSNE (100 µM) only slightly but nonsignificantly affected H₂S levels in MDA-MB-231 TNBC cells. Compared with the addition of paclitaxel alone, the addition of HMPSNE (100 µM) to 10 nM paclitaxel (Figure 4A) caused a significant decrease in H₂S levels. This effect was not observed with capecitabine cotreatment (Figure 4B). In contrast, compared with doxorubicin or HMPSNE single treatments, dual treatment with HMPSNE and doxorubicin significantly reduced H₂S levels. Notably, doxorubicin alone also decreased H₂S levels and reached statistical significance relative to those in the control untreated group (Figure 4C).

However, doxorubicin treatment significantly increased the protein expression of the H₂S-producing enzymes CBS, CSE, and 3-MST (by approximately 1.5-fold, 1.5-fold, and 1.3-fold, respectively) (Figures 5A–C, F), this was not accompanied by an increase in total H₂S production. Notably, doxorubicin also markedly upregulated the H₂S-catabolizing enzyme TST and, to a lesser extent, ETHE1 (by approximately 3-fold and 1.3-fold, respectively) (Figures 5D–F), which may contribute to enhanced mitochondrial H₂S clearance.

In contrast, HMPSNE treatment reduced 3-MST protein expression by approximately 1.3-fold decrease, indicating partial suppression of the mitochondrial H₂S-producing pathway (Figure 5C).

Compared with treatment with either doxorubicin or HMPSNE alone, combined treatment of MDA-MB-231 cells with HMPSNE or doxorubicin led to an almost complete decrease in the number of formed colonies (Figures 6A,B) and a significant decrease in cellular migration (Figures 6C,D). However, compared with control cells, HMPSNE-alone-treated cells tended to have decreased clonogenicity (Figure 6A) and significantly decreased migration (Figure 6C).

CSCs, which contribute to the stemness of tumor cells and consequently the resistance of TNBC cells, have been shown to be affected by H₂S signaling (Li et al., 2017; Li et al., 2018). Therefore, we investigated whether doxorubicin, on its own or in combination with 3-MST inhibition, affects the formation of the CSC surface marker CD44. Figure 7A shows the gene expression profile of CD44 after the treatment of MDA-MB-231 cells with HMPSNE and doxorubicin (1 µM). Compared with doxorubicin treatment alone, the combination of HMPSNE (100 µM) and doxorubicin (1 µM) markedly decreased the quantity of CD44 mRNA transcripts. HMPSNE (100 µM), on its own, was also associated with low expression of CD44 mRNA transcripts. When the protein expression of CD44 was investigated via flow cytometry analysis, doxorubicin treatment clearly increased CD44 expression. CD44 expression in HMPSNE-treated cells was slightly reduced. Compared with treatment with doxorubicin alone, the combination of doxorubicin and HMPSNE caused a slight decrease in CD44 protein expression. CD44 expression in HMPSNE-treated cells was low. However, compared with doxorubicin alone, the combination of doxorubicin and HMPSNE caused only a slight decrease in CD44 protein expression (Figure 7B).

One of the pathways most commonly dysregulated in TNBC is the PI3K/AKT/mTOR pathway. Mutations or alterations in this pathway occur in almost 25% of TNBC patients (Bullock et al., 2024; Zhang et al., 2024), promoting chemoresistance, in addition to being implicated in H₂S signaling (Steelman et al., 2008; Liu et al., 2021). Therefore, we quantified the levels of AKT, GSK-3β and mTOR in TNBC cells treated with HMPSNE or doxorubicin. Figures 8A,B,F show the effects of HMPSNE or doxorubicin alone or in combination on the expression levels of AKT and p-AKT. Doxorubicin significantly increased p-AKT levels (Figures 8B,F). The addition of HMPSNE (100 µM) to doxorubicin slightly reduced p-AKT expression (Figures 8B,F). Total AKT protein levels were not altered in any of the treated groups (Figures 8A,F), suggesting an increase in the phosphorylation of existing AKT protein without a change in total AKT protein levels. GSK-3β protein levels were significantly suppressed upon doxorubicin treatment (Figures 8C,F), which may be related to increased p-AKT levels via the inhibition of GSK-3β, which in turn tends to increase mTOR levels, as shown in Figures 8D–F. The inhibition of 3-MST with HMPSNE did not significantly affect the doxorubicin-induced changes in pAKT or GSK-3β expression (GSK-3β) (Figures 8B,C,F). Although combined treatment with doxorubicin and HMPSNE significantly decreased the level of GSK-3β, similar to that of doxorubicin (Figures 8C,F), the level of mTOR tended to decrease compared with that of doxorubicin alone (Figures 8D,F). These observations can be explained by the slight decrease in p-AKT levels (compared with those in doxorubicin alone), which may subsequently reduce mTOR levels.