Curcumin Enhances Gemcitabine Sensitivity in Breast Cancer Cells Through ROS-Associated Mitochondrial Apoptosis and Transcriptional Reprogramming

Simple Summary

Breast cancer treatment is often limited by resistance to chemotherapy, particularly in aggressive subtypes such as triple-negative breast cancer. Gemcitabine is an effective chemotherapeutic drug, but its efficacy can be reduced by cellular survival mechanisms. Curcumin, a natural compound, has been suggested to enhance the activity of anticancer drugs. In this study, we show that curcumin significantly increases the sensitivity of breast cancer cells to gemcitabine. The combined treatment induces oxidative stress, disrupts mitochondrial function, activates programmed cell death, and suppresses angiogenesis-related signaling. Direct measurement of intracellular reactive oxygen species using the DCFH-DA assay demonstrated marked reactive oxygen species accumulation in breast cancer cells following combination treatment, whereas normal breast epithelial MCF-10A cells exhibited only minimal oxidative changes under identical conditions. Assessment of mitochondrial membrane potential using JC-1 staining further revealed substantial mitochondrial depolarization in malignant cells, while mitochondrial integrity was largely preserved in MCF-10A cells, supporting cancer-selective vulnerability. Blocking oxidative stress reverses these effects, demonstrating that reactive oxygen species are central to the synergistic interaction. Transcriptomic analysis further confirmed the activation of pro-apoptotic pathways and the inhibition of survival signals. These effects are particularly pronounced in triple-negative breast cancer cells. Overall, our findings provide a mechanistic basis for combining curcumin with gemcitabine to improve therapeutic efficacy in breast cancer.

Abstract

Breast cancer is a leading cause of cancer-related mortality in women, necessitating new treatment strategies. Curcumin (Cur), a natural polyphenol, and gemcitabine (Gem), a standard chemotherapeutic, were investigated for their combined anticancer effects. We hypothesized that Cur sensitizes breast cancer cells to Gem via reactive oxygen species (ROS)-mediated apoptosis, and that this effect is associated with selective oxidative vulnerability in malignant cells compared to normal breast epithelial cells. MCF-7 (hormone receptor-positive) and MDA-MB-231 (triple-negative) cells were treated with Cur and Gem alone or in combination. Normal breast epithelial MCF-10A cells were included to evaluate therapeutic selectivity. Cell viability (MTT), apoptosis (Annexin V/PI), oxidative stress (TOS/TAS), intracellular ROS generation (DCFH-DA assay), mitochondrial membrane potential (ΔΨm) (JC-1 staining), caspase activation, synergy (Bliss/HSA/Chou-Talalay), VEGF secretion (ELISA), and transcriptomic changes (RNA-Seq) were assessed. Cur and Gem showed dose-dependent cytotoxicity. Combination treatment demonstrated strong synergistic activity, significantly enhancing apoptosis, oxidative stress, and caspase activation. Direct quantification of intracellular ROS revealed marked ROS accumulation in MCF-7 and MDA-MB-231 cells following combination treatment, whereas MCF-10A cells exhibited only modest oxidative changes. JC-1 analysis demonstrated substantial mitochondrial depolarization in breast cancer cells, which was largely reversible by ROS scavenging and minimal in MCF-10A cells. VEGF secretion was markedly suppressed. Transcriptomic analysis revealed profound alterations in apoptosis, cell cycle, and angiogenesis-related pathways, with more pronounced transcriptional reprogramming observed in the triple-negative subtype. Cur synergistically enhances Gem’s efficacy in breast cancer cells through ROS-mediated apoptosis and anti-angiogenic effects, characterized by cancer-selective ROS amplification and mitochondrial membrane depolarization, supporting its potential as a combination therapy, particularly for triple-negative breast cancer.

1. Introduction

Breast cancer is the most frequently diagnosed malignancy among women worldwide and remains one of the leading causes of cancer-related mortality [1]. According to Global Cancer Observatory (GLOBOCAN) 2020 data, approximately 2.3 million new breast cancer cases are diagnosed annually, highlighting the substantial global health burden of this disease [2]. Breast cancer is a biologically heterogeneous malignancy and is commonly classified into molecular subtypes, including hormone receptor positive, human epidermal growth factor receptor 2 positive, and triple-negative breast cancer. Each subtype exhibits distinct molecular characteristics, clinical behavior, and therapeutic responsiveness, thereby necessitating subtype specific treatment strategies [3]. Despite significant advances in multimodal treatment approaches that include surgery, radiotherapy, and systemic therapies, breast cancer management remains challenging due to therapy resistance, disease recurrence, and metastatic progression [4].

Conventional chemotherapeutic agents continue to play a critical role in the treatment of breast cancer, particularly in advanced and metastatic disease. Among these agents, Gem is a pyrimidine nucleoside analog that exerts its antitumor activity by inhibiting deoxyribonucleic acid (DNA) synthesis, which ultimately leads to cell cycle arrest and apoptotic cell death [5]. Clinically, Gem has demonstrated efficacy in metastatic breast cancer, especially in patients who have developed resistance to anthracycline- and taxane-based therapies [6]. However, its clinical application is limited by dose dependent toxicities such as myelosuppression and hepatotoxicity, as well as the frequent emergence of acquired chemoresistance [7]. These limitations emphasize the need for novel therapeutic strategies that can enhance Gem efficacy while minimizing systemic toxicity.

In recent years, increasing attention has been directed toward complementary and alternative therapeutic approaches, particularly those involving bioactive natural compounds, as potential adjuvants to conventional chemotherapy [8]. Cur (diferuloylmethane), a naturally occurring polyphenolic compound derived from the rhizome of Curcuma longa, has been extensively investigated due to its antioxidant, anti-inflammatory, and anticancer properties [9,10,11,12]. A substantial body of preclinical evidence indicates that Cur exerts anticancer effects through the modulation of multiple molecular pathways that regulate tumor cell proliferation, apoptosis induction, cell cycle progression, angiogenesis, and metastatic potential [13]. Furthermore, Cur has been shown to interfere with key oncogenic signaling pathways, including NF κB, Signal transducer and activator of transcription 3 (STAT3), Phosphoinositide 3 kinase (PI3K) Protein kinase B (Akt), and Mitogen activated protein kinase (MAPK), thereby targeting several hallmarks of cancer simultaneously [14].

Despite its promising anticancer activity, the clinical translation of Cur has been hindered by unfavorable pharmacokinetic properties, including poor aqueous solubility, rapid systemic metabolism, and low bioavailability. To overcome these challenges, considerable research efforts have focused on the development of advanced drug delivery systems. Nanoformulations of Cur, such as liposomes, polymeric nanoparticles, and micellar systems, have been shown to enhance solubility, prolong circulation time, improve tumor accumulation through the enhanced permeability and retention effect, and increase intracellular uptake in preclinical models [15,16]. These advances provide a rational framework for achieving therapeutically relevant Cur concentrations in vivo. In the present study, native Cur was used to establish proof-of-concept synergy with Gem, while the findings are interpreted within the context of these evolving delivery strategies to emphasize translational relevance. While native curcumin was employed in the present study to establish a clear and uncontaminated proof-of-concept regarding its synergistic mechanism with gemcitabine, the therapeutic concentrations required in vitro (IC₅₀ ~20–30 µM) are recognized to exceed the typical plasma levels achievable with conventional oral administration. Therefore, the primary objective of this work is not to propose these specific concentrations for direct clinical use, but to mechanistically decode the synergy, thereby providing a robust rationale for subsequent in vivo evaluation using advanced delivery systems. The promising efficacy data generated here define the target therapeutic effect that next-generation curcumin formulations (e.g., liposomal, nanoparticle, or phospholipid-complexed curcumin) must aim to achieve in vivo to translate this synergistic potential into clinical benefit.

Combination strategies that integrate conventional chemotherapeutic agents with natural compounds have attracted growing interest due to their potential to enhance anticancer efficacy through synergistic interactions while reducing adverse effects [17]. Cur has been reported to function as a chemosensitizing agent that enhances the therapeutic efficacy of various anticancer drugs across different tumor models [18]. However, studies specifically examining the combined effects of Cur and Gem in breast cancer cell lines remain limited, and the molecular mechanisms underlying this interaction have not yet been fully elucidated.

Accumulating evidence further supports the role of Cur as a potent modulator of apoptotic signaling pathways in cancer cells. A comprehensive mechanistic review by Mortezaee and colleagues demonstrated that Cur induces apoptosis through both intrinsic and extrinsic pathways by promoting mitochondrial dysfunction, increasing intracellular ROS production, enhancing p53 activity, and suppressing anti-apoptotic signaling mediated by the NF κB, Cyclooxygenase 2 (COX 2), and PI3K Akt pathways [19]. This multifaceted pro-apoptotic profile highlights the potential utility of Cur as an adjuvant agent capable of sensitizing tumor cells to conventional chemotherapy.

Based on this mechanistic background, we hypothesized that Cur sensitizes breast cancer cells to Gem primarily through ROS mediated activation of the intrinsic apoptotic pathway, leading to enhanced cytotoxicity, transcriptional reprogramming, and suppression of angiogenesis. Furthermore, we postulated that this ROS driven sensitization would be associated with cancer selective oxidative vulnerability, resulting in preferential mitochondrial dysfunction in malignant cells compared to normal breast epithelial cells. To test this hypothesis, we systematically evaluated the anticancer effects of Cur and Gem, administered alone or in combination, in two biologically distinct breast cancer cell lines, namely hormone receptor positive MCF-7 cells and triple-negative MDA-MB-231 cells. In addition, normal breast epithelial MCF-10A cells were included to assess therapeutic selectivity and determine whether ROS amplification and mitochondrial depolarization are preferentially induced in cancer cells. By integrating functional synergy analyses using Bliss and HSA models with transcriptome wide RNA sequencing (RNA-Seq), this study provides a comprehensive mechanistic framework for elucidating Cur mediated Gem sensitization across distinct breast cancer subtypes. Direct quantification of intracellular ROS using the DCFH-DA assay and assessment of mitochondrial membrane potential (ΔΨm) using JC-1 staining were incorporated to functionally validate ROS mediated mitochondrial involvement in the synergistic interaction. Multiple cellular and molecular endpoints, including cell viability, apoptosis induction, oxidative stress modulation, intracellular ROS generation, mitochondrial membrane depolarization, caspase activation, angiogenesis inhibition, and global gene expression profiling, were investigated to define the therapeutic potential and mechanistic basis of this combinatorial strategy.

2. Materials and Methods

2.1. Cell Culture Conditions and Reagents

Human breast cancer cell lines MCF-7, representing the hormone receptor positive luminal subtype, and MDA-MB-231, representing the triple-negative basal subtype, were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). The human normal breast epithelial cell line MCF-10A was also obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA) and used to evaluate therapeutic selectivity. Upon receipt, cell lines were authenticated by the supplier and routinely monitored for mycoplasma contamination throughout the experimental period.

Cells were cultured in Dulbecco’s modified Eagle medium (DMEM, high glucose formulation; Gibco, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10% heat-inactivated fetal bovine serum (FBS; Gibco, Thermo Fisher Scientific, Waltham, MA, USA), 100 U/mL penicillin, and 100 µg/mL streptomycin (Gibco, Thermo Fisher Scientific, Waltham, MA, USA). MCF-10A cells were cultured according to ATCC recommendations in Dulbecco’s modified Eagle medium/Ham’s F-12 (DMEM/F-12; Gibco, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 5% heat-inactivated horse serum, 20 ng/mL epidermal growth factor, 10 µg/mL insulin, 0.5 µg/mL hydrocortisone, and 100 ng/mL cholera toxin, in addition to 100 U/mL penicillin and 100 µg/mL streptomycin. Cell cultures were maintained in a humidified incubator at 37 °C under an atmosphere containing 5% carbon dioxide.

Cells were routinely subcultured when they reached approximately 80–90% confluence. For passaging, culture medium was aspirated, cells were washed once with sterile phosphate buffered saline (PBS), and detached using 0.25% trypsin–EDTA solution (Gibco, Thermo Fisher Scientific, Waltham, MA, USA). Trypsinization was stopped by the addition of complete culture medium, and cells were collected by gentle centrifugation. Viable cells were counted using trypan blue exclusion prior to seeding for experimental assays.

Cur with a purity of at least 94% and Gem hydrochloride with a purity of at least 98% were purchased from Sigma Aldrich (St. Louis, MO, USA). Cur stock solutions were freshly prepared at a concentration of 20 mM by dissolving the compound in dimethyl sulfoxide (DMSO; Sigma Aldrich, St. Louis, MO, USA). Gem hydrochloride stock solutions were prepared at a concentration of 10 mM in sterile PBS adjusted to pH 7.4. All stock solutions were aliquoted and stored at −20 °C to avoid repeated freeze–thaw cycles.

Prior to treatment, stock solutions were diluted in complete culture medium to obtain the desired final concentrations. The final concentration of DMSO did not exceed 0.1% (v/v) in any experimental condition. A vehicle control containing an equivalent concentration of DMSO was included in all experiments to exclude potential solvent related effects.

2.2. Assessment of Cell Viability and Determination of IC50 Values Using the MTT Assay

Cell viability was assessed using the colorimetric MTT assay (Sigma Aldrich, St. Louis, MO, USA), which is based on the ability of metabolically active cells to reduce MTT into insoluble formazan crystals.

MCF-7 and MDA-MB-231 breast cancer cells were seeded into 96-well flat-bottom tissue culture plates at a density of 1 × 10⁴ cells per well in 100 µL of complete culture medium. For therapeutic selectivity analysis, MCF-10A normal breast epithelial cells were seeded under identical density conditions and cultured according to their respective growth medium requirements. After seeding, cells were incubated for 24 h at 37 °C in a humidified atmosphere containing 5% carbon dioxide to allow proper cell attachment.

Following the adhesion period, cells were treated with Cur at concentrations of 0, 5, 10, 20, 40, and 80 µM and Gem at concentrations of 0, 0.1, 1, 5, 10, and 50 µM. All treatments were applied for 48 h. For combination treatment experiments, cells were treated concurrently with Cur and Gem at concentrations corresponding to their individually determined half maximal inhibitory concentration (IC₅₀) values. Untreated control wells and vehicle control wells containing 0.1% DMSO were included in all experimental plates.

At the end of the treatment period, 10 µL of MTT solution prepared at a concentration of 5 mg/mL in sterile PBS was added to each well. Plates were then incubated for 4 h at 37 °C to allow for the formation of formazan crystals. Following incubation, 100 µL of DMSO was added to each well to dissolve the formazan crystals. The plates were gently agitated on an orbital shaker for 10 min at room temperature to ensure complete solubilization.

Absorbance values were measured at a wavelength of 570 nm using a microplate reader. Wells containing culture medium and MTT reagent without cells were used as blank controls to correct for background absorbance.

Cell viability was expressed as a percentage relative to the untreated control group and calculated using the following equation:

$$\mathrm{Cell} \mathrm{viability} (\%)={\displaystyle \frac{A_{\mathrm{treated}}-A_{\mathrm{blank}}}{A_{\mathrm{control}}-A_{\mathrm{blank}}}}\times 100$$

where $A_{\mathrm{treated}}$ represents the absorbance of drug treated cells, $A_{\mathrm{control}}$ represents the absorbance of untreated control cells, and $A_{\mathrm{blank}}$ represents the absorbance of wells containing medium and MTT reagent without cells.

IC₅₀ values were calculated using nonlinear regression analysis based on sigmoidal dose–response curves with GraphPad Prism software version 9.0 (GraphPad Software, San Diego, CA, USA).

Each experiment was performed using 3 independent biological replicates, and each concentration was analyzed in 6 technical replicates.

2.3. Assessment of Cell Viability in Normal Breast Epithelial Cells (MCF-10A)

To evaluate the effects of curcumin, gemcitabine, and their combination on non-malignant breast epithelial cells, viability assays were performed using the human normal breast epithelial cell line MCF-10A. Cells were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA) and cultured in DMEM/F12 medium supplemented with 5% horse serum, 20 ng/mL epidermal growth factor, 0.5 µg/mL hydrocortisone, 10 µg/mL insulin, and 100 ng/mL cholera toxin under standard conditions (37 °C, 5% CO₂).

MCF-10A cells were seeded into 96-well plates at a density of 6 × 10³ cells per well and allowed to adhere for 48 h. Cells were then treated with curcumin, gemcitabine, or their combination at concentrations corresponding to the IC₅₀ values determined in breast cancer cell lines. Vehicle-treated cells served as controls. To ensure consistency with experiments performed in malignant cell lines, treatments were applied for 48 h unless otherwise specified. Following treatment, cell viability was assessed using the colorimetric MTT viability assay according to the manufacturer’s instructions. Absorbance values were measured at a wavelength of 570 nm using a microplate reader and normalized to control values. Each experimental condition was analyzed in six technical replicates across at least two independent experiments.

2.4. Bliss Independence Model (Bliss) for Drug Combination Analysis

Bliss was applied to evaluate the combined effects of Cur and Gem on breast cancer cell viability. This model assumes that the two agents exert their effects through independent mechanisms and therefore allows for an estimation of the expected combined effect in the absence of interaction. Bliss synergy analysis was performed in MCF-7 and MDA-MB-231 breast cancer cell lines, whereas MCF-10A normal breast epithelial cells were analyzed for viability without formal synergy modeling.

For each drug, the fractional inhibition was calculated from the cell viability data obtained in the MTT assay. Fractional inhibition values ranged from 0 to 1, where 0 represents no inhibition and 1 represents complete inhibition. The expected effect of the drug combination according to Bliss was calculated using the following equation:

$$E_{\mathrm{Bliss}}=E_A+E_B-(E_A\times E_B)$$

where $E_A$ and $E_B$ represent the fractional inhibition values of Cur and Gem when applied individually at the corresponding concentrations [20].

The observed fractional inhibition of the combined treatment was defined as $E_{AB}$. To quantify the degree of interaction between Cur and Gem, the Bliss synergy score was calculated by subtracting the expected Bliss effect from the observed combination effect using the following equation:

$$\Delta \mathrm{Bliss}=E_{AB}-(E_A+E_B-E_A\times E_B)$$

Positive Delta Bliss values indicate a synergistic interaction between the two agents, values close to zero indicate an additive effect, and negative values indicate an antagonistic interaction.

Synergy analysis and visualization of drug interaction landscapes were performed using Combenefit software version 2.0. Cell viability data obtained from the MTT assays were imported into the software, and Bliss synergy matrices were generated across the tested concentration ranges to identify regions of synergistic, additive, or antagonistic interactions.

2.5. Chou Talalay Combination Index Analysis

Drug interactions between Cur and Gem were quantitatively evaluated using the Chou Talalay method, which is based on the median effect principle and allows for an assessment of synergistic, additive, or antagonistic interactions across different effect levels. Combination index (CI) analysis was performed using CompuSyn software version 1.0 (ComboSyn Inc., Paramus, NJ, USA). CI analysis was conducted in MCF-7 and MDA-MB-231 breast cancer cell lines based on MTT-derived dose–response data.

Dose response data obtained from the MTT cell viability assays were used as input for the analysis. The fraction affected (Fa), which represents the proportion of growth inhibition induced by each treatment, was calculated for individual drugs and their combinations over a range of concentrations. Based on these values, the CI was computed according to the median effect equation for each Fa level.

CI values were interpreted as follows: CI values < 1 indicate synergistic interactions, CI values = 1 indicate additive effects, and CI values > 1 indicate antagonistic interactions between Cur and Gem. CI values were analyzed across multiple Fa levels to provide a comprehensive evaluation of drug interaction behavior rather than reliance on a single effect point.

This approach enabled the quantitative assessment of the degree of synergy and provided complementary validation to the Bliss analysis.

2.6. Analysis of Apoptotic Cell Death by Flow Cytometry

Apoptotic cell death was quantitatively analyzed by flow cytometry using an Annexin V fluorescein isothiocyanate and PI double staining assay. The Annexin V Fluorescein isothiocyanate (FITC) apoptosis detection kit was purchased from BD Biosciences (San Jose, CA, USA) and used according to the manufacturer’s instructions.

MCF-7 and MDA-MB-231 breast cancer cells were seeded into 6-well tissue culture plates at a density of 2 × 10⁵ cells per well in complete culture medium and incubated under standard culture conditions to allow for cell attachment. After the adhesion period, cells were treated with Cur at its IC₅₀, Gem at its IC₅₀, or a combination of both agents for 48 h. Control cells were treated with vehicle alone consisting of 0.1% DMSO. Apoptosis analysis was performed in MCF-7 and MDA-MB-231 cells as representative malignant models.

Following treatment, both floating and adherent cells were collected to avoid selective loss of apoptotic populations. Adherent cells were detached using trypsin EDTA, combined with floating cells, and washed twice with cold PBS by centrifugation. Cell pellets were resuspended and adjusted to a final concentration of 1 × 10⁶ cells/mL.

For staining, 100 µL of the cell suspension was transferred into individual flow cytometry tubes and resuspended in 100 µL of Annexin V binding buffer. Subsequently, 5 µL of Annexin V FITC and 5 µL of PI were added to each sample. Cells were gently mixed and incubated for 15 min in the dark at room temperature. After incubation, 400 µL of binding buffer was added to each tube, and samples were analyzed within 1 h.

Flow cytometric analysis was performed using a BD FACSCalibur flow cytometer (BD Biosciences, San Jose, CA, USA). For each sample, a minimum of 20,000 events were acquired. Data acquisition and compensation settings were applied uniformly across all samples.

Data analysis was performed using FlowJo software version 10.8 (BD Biosciences, San Jose, CA, USA). Cells were classified into 4 distinct populations based on Annexin V and PI staining patterns: viable cells defined as Annexin V negative and PI negative, early apoptotic cells defined as Annexin V positive and PI negative, late apoptotic cells defined as Annexin V positive and PI positive, and necrotic cells defined as Annexin V negative and PI positive. The total apoptotic population was calculated as the sum of early and late apoptotic cells.

2.7. Determination of TAS and TOS

Cellular oxidative stress status was evaluated by measuring TAS and TOS using commercially available colorimetric assay kits obtained from Rel Assay Diagnostics (Gaziantep, Turkey). All experimental procedures were performed in accordance with the manufacturer’s instructions.

MCF-7 and MDA-MB-231 cells were seeded into 6-well culture plates and treated with Cur, Gem, or their combination for 48 h under standard culture conditions. TAS and TOS analyses were performed in the MCF-7 and MDA-MB-231 breast cancer cell lines to assess global oxidative status following treatment. Following treatment, cells were washed twice with ice cold PBS to remove residual culture medium and detached using cell lysis buffer provided with the assay kits. Cell lysates were collected and centrifuged at 14,000 rpm for 15 min at 4 °C to remove cellular debris. The resulting supernatants were carefully collected and used for subsequent analyses.

Total protein concentrations in cell lysates were determined using the Bradford protein assay (Bio Rad Laboratories, Hercules, CA, USA) according to the manufacturer’s protocol. Protein concentrations were used to ensure consistency and comparability between samples.

TAS was measured based on the ability of antioxidants present in the samples to suppress the oxidation of the 2,2 prime azino bis 3 ethylbenzothiazoline 6 sulfonic acid radical cation. The decrease in color intensity of the 2,2′-Azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) radical was measured spectrophotometrically, and TAS values were calculated using Trolox as a standard. Results were expressed as millimoles of Trolox equivalents per liter.

TOS was determined based on the oxidation of ferrous ions to ferric ions by oxidant molecules present in the samples. The resulting ferric ions formed a colored complex, the intensity of which was measured spectrophotometrically at a wavelength of 530 nm. TOS values were calculated using hydrogen peroxide as a reference standard and expressed as micromoles of hydrogen peroxide equivalents per liter.

All absorbance measurements were performed using a microplate reader. Each sample was analyzed in triplicate (n = 3), and mean values were used for statistical analysis.

2.8. ROS Scavenging Using N-Acetylcysteine

To functionally validate the involvement of ROS in the observed anticancer effects, N-acetylcysteine was employed as a pharmacological ROS scavenger. N-acetylcysteine was purchased from Sigma Aldrich (St. Louis, MO, USA).

MCF-7 and MDA-MB-231 cells were seeded and cultured under standard conditions as described above. Prior to drug treatment, cells were pre-treated with N-acetylcysteine at a final concentration of 5 mM for 2 h to allow for effective intracellular scavenging of ROS. Following the pre-treatment period, cells were exposed to the Cur and Gem combination while maintaining the presence of N-acetylcysteine in the culture medium throughout the 48 h treatment period.

Control groups included cells treated with the Cur and Gem combination in the absence of N-acetylcysteine as well as untreated control cells. Following treatment, cells were processed for subsequent analyses, including cell viability, apoptosis assessment, oxidative stress measurements, and caspase activity assays. In addition, intracellular ROS levels (DCFH-DA assay) and mitochondrial membrane potential (ΔΨm) (JC-1 staining) were evaluated in the presence and absence of N-acetylcysteine to determine whether ROS scavenging attenuates mitochondrial depolarization.

This experimental design enabled the direct evaluation of the contribution of ROS to the synergistic cytotoxic and pro-apoptotic effects induced by the Cur and Gem combination, as well as to ROS-mediated mitochondrial dysfunction.

2.9. Intracellular ROS Measurement by DCFH-DA Assay

Intracellular reactive oxygen species (ROS) levels were quantified using the 2′,7′-dichlorofluorescin diacetate (DCFH-DA) assay.

MCF-7, MDA-MB-231, and MCF-10A cells were seeded into 6-well plates at a density of 2 × 10⁵ cells per well and allowed to adhere for 24 h. Cells were subsequently treated with curcumin (IC₅₀), gemcitabine (IC₅₀), or their combination (Cur+Gem) for 48 h.

For ROS inhibition experiments, cells were pre-treated with 5 mM N-acetylcysteine (NAC; Sigma-Aldrich, St. Louis, MO, USA) for 2 h prior to drug exposure, and NAC was maintained throughout the 48-h treatment period.

Following treatment, cells were washed twice with phosphate-buffered saline (PBS; Gibco, Thermo Fisher Scientific, Waltham, MA, USA) and incubated with 10 µM DCFH-DA (Sigma-Aldrich, St. Louis, MO, USA) in serum-free medium for 30 min at 37 °C in the dark. Excess probe was removed by washing with PBS, and cells were harvested by trypsinization.

Fluorescence intensity was measured using a BD FACSCalibur flow cytometer (BD Biosciences, San Jose, CA, USA). DCF fluorescence was detected in the FL1 channel (excitation 488 nm, emission 530 nm). A minimum of 20,000 events per sample were acquired. Mean fluorescence intensity (MFI) values were calculated using FlowJo software v10 (BD Life Sciences, Ashland, OR, USA) and expressed as fold change relative to untreated control cells. All experiments were performed in three independent biological replicates.

2.10. Assessment of Mitochondrial Membrane Potential (ΔΨm) by JC-1 Staining

Mitochondrial membrane potential (ΔΨm) was evaluated using the JC-1 fluorescent probe (Abcam, Cambridge, UK).

MCF-7, MDA-MB-231, and MCF-10A cells were treated with curcumin (IC₅₀), gemcitabine (IC₅₀), or Cur+Gem for 48 h. For rescue experiments, NAC pre-treatment (5 mM, 2 h) was applied prior to drug exposure and maintained throughout the treatment period.

After 48 h, cells were incubated with 5 µg/mL JC-1 dye in complete medium for 20 min at 37 °C in the dark. Cells were washed twice with PBS and immediately analyzed by flow cytometry using a BD FACSCalibur instrument (BD Biosciences, San Jose, CA, USA).

JC-1 monomers were detected in the FL1 channel (green fluorescence), whereas JC-1 aggregates were detected in the FL2 channel (red fluorescence). Quadrant gating was applied to distinguish polarized mitochondria (upper left quadrant) from depolarized cells (lower right quadrant). The percentage of depolarized cells was calculated relative to the total acquired events (≥20,000 events per sample). Data analysis was performed using FlowJo software v10 (BD Life Sciences, Ashland, OR, USA). All experiments were conducted in triplicate.

2.11. Measurement of Caspase-3 and Caspase-9 Activities

Caspase-3 (CASP3) and caspase 9 (CASP9) activities were quantified using colorimetric caspase activity assay kits obtained from Abcam (Cambridge, UK) according to the manufacturer’s protocols. These assays are based on the cleavage of specific peptide substrates conjugated to the chromophore p-nitroanilide, which is released upon caspase activation.

MCF-7 and MDA-MB-231 cells were treated with Cur, Gem, or their combination for 48 h, as described above. Following treatment, cells were washed with cold PBS and lysed using the lysis buffer provided with the assay kits. Cell lysates were incubated on ice for complete protein extraction and subsequently centrifuged at high speed to remove insoluble debris. The resulting supernatants were collected for enzymatic analysis.

Total protein concentrations in cell lysates were determined using a Bradford protein assay to ensure equal protein loading across all samples. For each reaction, 50 µg of total protein were transferred into individual wells of a 96-well microplate.

For CASP3 activity measurement, the DEVD p-nitroanilide substrate was added to the reaction mixture. For CASP9 activity measurement, the LEHD p-nitroanilide substrate was used. Reaction mixtures were gently mixed and incubated at 37 °C for 2 h to allow for enzymatic cleavage of the substrates.

The release of p-nitroanilide was quantified by measuring absorbance at a wavelength of 405 nm using a microplate reader. Caspase activities were calculated based on a standard curve generated using known concentrations of free p-nitroanilide and expressed as nanomoles of p-nitroanilide released per hour per milligram of protein.

To facilitate comparison between treatment groups, caspase activity values were normalized to the untreated control group, and results were expressed as fold change relative to the control.

2.12. Quantification of VEGF Secretion by Enzyme Linked Immunosorbent Assay

VEGF secretion levels were quantified in cell culture supernatants using a human VEGF enzyme linked immunosorbent assay kit (Human VEGF ELISA Kit, Abcam, Cambridge, UK; catalog number ab222510) according to the manufacturer’s instructions.

MCF-7 and MDA-MB-231 cells were seeded into 6-well tissue culture plates and treated with Cur, Gem, or their combination for 48 h under standard culture conditions. Following treatment, cell culture supernatants were carefully collected to avoid cellular contamination and centrifuged at 2000 rpm for 10 min at 4 °C to remove residual cell debris. Clarified supernatants were transferred to clean tubes and either analyzed immediately or stored at −80 °C until ELISA analysis.

For VEGF measurement, standards and samples were added to the antibody coated microplate wells provided with the kit. Samples were analyzed in accordance with the manufacturer’s protocol, including sequential incubation with detection antibodies and enzyme conjugates, followed by substrate development. After the final incubation step, the enzymatic reaction was stopped, and absorbance was measured at the recommended wavelength using a microplate reader.

VEGF concentrations in the samples were calculated from a standard curve generated using known concentrations of recombinant human VEGF provided with the kit. VEGF secretion levels were expressed as picograms per milliliter of culture supernatant. Each experimental condition was analyzed in triplicate, and mean values were used for statistical analysis.

2.13. RNA Isolation and Quantitative Real-Time PCR Analysis

Total RNA was isolated from treated MCF-7 and MDA-MB-231 breast cancer cells using TRIzol reagent (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s instructions. Following RNA extraction, RNA concentration and purity were assessed spectrophotometrically, and samples exhibiting acceptable purity ratios were used for downstream analyses.

Complementary DNA synthesis was performed from total RNA using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s protocol. Reverse transcription reactions were carried out under recommended thermal cycling conditions to ensure efficient and reproducible cDNA synthesis.

Quantitative real-time polymerase chain reaction (qRT-PCR) analysis was conducted using SYBR Green chemistry on a QuantStudio 3 Real-Time PCR System (Applied Biosystems, Thermo Fisher Scientific, Waltham, MA, USA). PCR reactions were prepared using PowerUp SYBR Green Master Mix (Applied Biosystems, Thermo Fisher Scientific, Waltham, MA, USA). Each reaction was performed in a final volume of 20 µL containing 10 µL of 2× PowerUp SYBR Green Master Mix, 1 µL of forward primer at a concentration of 10 µM, 1 µL of reverse primer at a concentration of 10 µM, 2 µL of diluted cDNA template at a 1:10 ratio, and 6 µL of nuclease free water.

The primer sequences used for the amplification of target and reference genes are provided in

Table 1. Primer sequences used for qRT-PCR.

Genes	Forward Primer (5′→3′)	Reverse Primer (5′→3′)
BAX	TCAGGATGCGTCCACCAAGAAG	TGTGTCCACGGCGGCAATCATC
BCL2	ATCGCCCTGTGGATGACTGAGT	GCCAGGAGAAATCAAACAGAGGC
TP53	CAGCACATGACGGAGGTTGT	TCATCCAAATACTCCACACGC
CASP3	AGAGGGGATCGTTGTAGAAGCTG	CACAAGCGACTGGATGAACCA
CASP9	CCTCATCATCAACAACCTGG	AAGTCCCTTTCGCAGAAACAG
VEGFA	AGGGCAGAATCATCACGAAGT	AGGGTCTCGATTGGATGGCA
NF-KB	ATGTGGAGATCATTGAGCAGC	CCTGGTCCTGTGTAGCCATT
ACTB (β-Actin)	CATTGCTGACAGGATGCAGAAGG	TGCTGGAAGGTGGACAGTGAGG
GAPDH	GGAGCGAGATCCCTCCAAAAT	GGCTGTTGTCATACTTCTCATGG

. Primer specificity was confirmed by melting curve analysis following amplification. Amplification efficiency for each primer pair was determined using a standard curve generated from serial dilutions of cDNA and maintained within the acceptable range of 90–110%.

Relative gene expression levels were calculated using an appropriate normalization strategy based on reference gene expression. The genes selected for quantitative real-time PCR validation were chosen based on their significant differential expression in RNA-Seq analysis and their established involvement in the p53, NF-kappa B, and VEGF signaling pathways. This approach enabled independent validation of the transcriptomic findings and strengthened the mechanistic interpretation of the observed treatment effects.

2.14. RNA Isolation and Transcriptomic Analysis

Total RNA was isolated from treated MCF-7 and MDA-MB-231 cells using TRIzol reagent (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s protocol. Briefly, cells were lysed directly in culture plates, and phase separation was achieved following chloroform addition and centrifugation. RNA was precipitated from the aqueous phase, washed, and resuspended in nuclease free water.

RNA concentration and purity were measured using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). Samples with absorbance ratios at 260–280 nm within acceptable limits were selected for further processing. RNA integrity was assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA), and only samples with an RNA integrity number > 7 were used for transcriptomic analysis.

High quality RNA samples were subjected to transcriptome wide gene expression profiling. Differential gene expression analysis was performed using the DESeq2 package version 1.34.0 implemented in the R statistical environment. Raw count data were normalized using the DESeq2 internal normalization method based on size factor estimation.

Statistical significance for differentially expressed genes was defined using a threshold of absolute log₂-fold change > 1.5 and an adjusted p value < 0.05, calculated using the Benjamini–Hochberg false discovery rate (FDR) correction method.

To explore the biological significance of differentially expressed genes, functional enrichment analyses were performed. Gene ontology (GO) enrichment analysis was conducted to identify significantly overrepresented biological processes, molecular functions, and cellular components. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was also performed to identify signaling pathways significantly affected by the treatments. Both enrichment analyses were conducted using the clusterProfiler (version 4.8.1) package in R (version 4.3.1; R Foundation for Statistical Computing, Vienna, Austria).

This transcriptomic approach enabled a comprehensive identification of molecular pathways associated with apoptosis, oxidative stress response, cell cycle regulation, and angiogenesis, thereby providing mechanistic insight into the effects of Cur, Gem, and their combination.

2.15. Visualization of Differential Gene Expression by Volcano Plot, Heat Map, and PCA

Graphical visualization approaches were applied to explore and illustrate global transcriptional changes induced by Cur, Gem, and their combination. Volcano plot, heat map, and PCA were performed using normalized RNA-Seq data derived from the transcriptomic analysis. Transcriptomic analyses were conducted in the MCF-7 and MDA-MB-231 breast cancer cell lines. It is stated that the RNA-Seq experiment was performed with 4 independent biological replicates (N = 4) for each treatment group (control, curcumin, gemcitabine, combination). The commercial kit and orientation strategy used are described: Libraries were prepared using the Illumina-compatible TruSeq Stranded mRNA LT Sample Preparation Kit with pol(A)+ mRNA selection and a stranded protocol. Depth and platform information necessary for evaluating data quality are included: Libraries were sequenced on the Illumina NovaSeq 6000 platform with 150 bp paired-end reads.

Volcano plots were generated to visualize the relationship between statistical significance and magnitude of gene expression changes. For each gene, the log₂-fold change was plotted on the x axis, and the negative log₁₀ of the adjusted p value was plotted on the y axis. Genes meeting the predefined thresholds for differential expression, defined as an absolute log₂-fold change > 1.5 and an adjusted p value < 0.05, were highlighted. Upregulated and downregulated genes were distinguished by different color coding, and the most statistically significant genes were annotated to facilitate interpretation.

For heat map analysis, genes exhibiting the highest expression variability across samples were selected. Specifically, the top 1000 differentially expressed genes based on variance were included. Prior to visualization, expression values were standardized by z-score normalization to enable a comparison across genes and samples. Hierarchical clustering was applied to both genes and samples using an appropriate distance metric and linkage method. Heat map visualization was performed using the ComplexHeatmap package version 2.10.0 within the R statistical environment, allowing for clear representation of expression patterns and clustering behavior among treatment groups.

PCA was conducted to reduce data dimensionality and identify the major sources of variation within the transcriptomic dataset. Log₂-transformed fragments per kilobase per million mapped reads (FPKM) values were used as input for the analysis. Principal component analysis (PCA) was performed using the prcomp function from the stats package in R (version 4.3.1; R Foundation for Statistical Computing, Vienna, Austria). The first 2 principal components, which accounted for the largest proportion of total variance, were visualized as 2-dimensional scatter plots. Graphical representation of PCA results was generated using the ggplot2 package, and samples were color coded according to treatment groups to illustrate clustering and separation patterns.

These visualization approaches collectively enabled a comprehensive assessment of global gene expression changes and facilitated comparison of transcriptional profiles among control, single-agent treated, and combination-treated groups.

2.16. Statistical Analysis

All experiments were conducted with a clear distinction between biological and technical replicates to ensure reproducibility and statistical robustness. Biological replicates were defined as 3 independent experiments performed on different days using freshly prepared cell cultures and treatment solutions. Technical replicates consisted of repeated measurements (e.g., multiple wells in a plate) within the same biological experiment. All quantitative data presented in this study were derived from at least 3 independent biological replicates, each including multiple technical replicates as specified for individual assays. This includes intracellular ROS (DCFH-DA), mitochondrial membrane potential (ΔΨm) (JC-1), and N-acetylcysteine rescue experiments. Results are expressed as mean ± standard deviation (SD). For key comparisons, the fold change relative to control is also reported to convey the magnitude of the observed effect.

Statistical analyses were performed using GraphPad Prism software version 9.0 (GraphPad Software, San Diego, CA, USA). For comparisons among 3 or more groups, one-way analysis of variance (ANOVA) was applied. Upon detecting a statistically significant overall difference, Tukey’s honestly significant difference (HSD) post hoc test was used for all pairwise comparisons within that experiment. This procedure controls the family-wise error rate for the set of comparisons made within each individual experimental figure or assay endpoint. For comparisons between 2 independent groups, an unpaired two-tailed Student’s t-test was used. The specific test used for each dataset is indicated in the corresponding figure legend.

A p-value < 0.05 was considered statistically significant. Levels of significance are denoted as follows: * p < 0.05, ** p < 0.01, *** p < 0.001.

3. Results

3.1. Subtype-Specific Cytotoxic and Synergistic Responses to Cur and Gem

The cytotoxic effects of Cur and Gem were first evaluated in hormone receptor positive MCF-7 cells and triple negative MDA-MB-231 cells using the MTT assay. Both agents induced dose-dependent reductions in cell viability in the two breast cancer cell lines (Figure 1). To assess therapeutic selectivity, parallel viability analyses were also performed in normal breast epithelial MCF-10A cells under identical treatment conditions.

Single-agent dose–response analysis revealed that the IC₅₀ of Cur was 24.5 µM in MCF-7 cells and 26.5 µM in MDA-MB-231 cells. For Gem, the IC₅₀ values were 4.17 µM in MCF-7 cells and 6.93 µM in MDA-MB-231 cells (Figure 1). These results indicate comparable single-agent cytotoxic sensitivity between the two subtypes, with slightly lower IC₅₀ values observed in MCF-7 cells for both compounds. In contrast, MCF-10A cells maintained substantially higher viability at concentrations corresponding to the IC₅₀ values determined in breast cancer cell lines, supporting a differential sensitivity between malignant and non-malignant cells.

To evaluate the potential therapeutic selectivity, we examined the effects of curcumin, gemcitabine, and their combination on MCF-10A normal breast epithelial cells. After 48 h treatment, curcumin alone showed moderate concentration-dependent cytotoxicity, reducing the viability to 65% at 80 µM (Figure 2A). Gemcitabine exhibited more potent effects, decreasing the viability to 42% at 50 µM. The combination demonstrated enhanced cytotoxicity, with viability dropping to 34%. Notably, at concentrations corresponding to the IC₅₀ values effective in breast cancer cell lines (~25 µM), MCF-10A cells maintained > 70% viability with all treatments (Figure 2B). This differential sensitivity suggests a therapeutic window, with normal cells being less susceptible to cytotoxicity than their malignant counterparts under identical treatment conditions. Importantly, this selective viability preservation in MCF-10A cells was later paralleled by minimal intracellular ROS accumulation and limited mitochondrial membrane depolarization under combination treatment, supporting the notion that Cur-mediated Gem sensitization preferentially targets malignant cells while sparing normal epithelial cells.

To examine whether Cur modulates Gem sensitivity, drug–response shift analyses were performed. Co-treatment with Cur resulted in a leftward shift of Gem dose–response curves in both the MCF-7 and MDA-MB-231 cells, indicating enhanced growth inhibition compared to Gem alone (Figure 3A,B). Similarly, the presence of Gem shifted the Cur dose–response curves toward lower effective concentrations in both cell lines (Figure 3C,D). These shifts were more pronounced in the MDA-MB-231 cells, suggesting a stronger cooperative interaction between the two agents in the triple negative subtype. This cooperative shift in dose–response behavior provided the first quantitative indication of pharmacological interaction, which was subsequently mechanistically explored through oxidative stress measurements, intracellular ROS assessment, and mitochondrial membrane potential analysis.

The nature and magnitude of the drug interaction were quantitatively assessed using Bliss and HSA models. Bliss analysis demonstrated positive Delta Bliss values in both cell lines, confirming synergistic interactions between Cur and Gem. The Delta Bliss score reached 15.7 in MCF-7 cells and 22.1 in MDA-MB-231 cells (Figure 4). Consistently, HSA analysis yielded synergy scores of 12.4 in MCF-7 cells and 18.6 in MDA-MB-231 cells (Figure 4). Together, these analyses demonstrate that although single-agent cytotoxicity was similar between subtypes, the strength of synergistic interaction was substantially greater in MDA-MB-231 cells. Synergy analyses were performed exclusively in malignant breast cancer cell lines, as combination index modeling requires cytotoxic–response curves derived from proliferative tumor cells. The higher interaction scores observed in MDA-MB-231 cells were subsequently mechanistically supported by more pronounced oxidative stress induction, intracellular ROS accumulation, and mitochondrial membrane depolarization in this subtype.

Overall, these results show that the Cur and Gem combination induces subtype specific differences in cytotoxic and synergistic responses, with triple negative MDA-MB-231 cells exhibiting a more pronounced synergistic sensitivity than hormone receptor positive MCF-7 cells. This subtype-dependent interaction profile was subsequently corroborated by differential oxidative stress induction, intracellular ROS accumulation, and mitochondrial membrane depolarization, which were markedly stronger in MDA-MB-231 cells. In contrast, normal breast epithelial MCF-10A cells displayed preserved viability, minimal ROS elevation, and limited mitochondrial disruption under identical treatment conditions, supporting a degree of mechanistic selectivity toward malignant cells.

3.2. Dose Reduction and Strength of Drug Synergy Across Breast Cancer Subtypes

To further quantify the extent of interaction between Cur and Gem, dose–response shift and synergy strength analyses were performed in both breast cancer cell lines.

Drug–response shift analyses demonstrated that combination treatment altered the effective concentrations required to achieve comparable levels of growth inhibition. In both MCF-7 and MDA-MB-231 cells, the presence of Cur shifted the Gem dose–response curves toward lower concentrations, while Gem similarly shifted the Cur dose–response curves (Figure 3). These shifts indicate that reduced drug concentrations were sufficient to produce equivalent cytotoxic effects when the agents were administered together.

Quantitative synergy assessment using Bliss revealed positive Delta Bliss values across the tested concentration range in both cell lines (Figure 4). The magnitude of synergy differed between subtypes, with MDA-MB-231 cells exhibiting a higher Delta Bliss score of 22.1 compared to 15.7 in MCF-7 cells (Figure 4). This indicates a stronger deviation from the expected additive effect in the triple-negative cell line.

Consistent results were obtained using the HSA model. HSA synergy analysis showed positive interaction scores in both breast cancer subtypes, with higher synergy values observed in MDA-MB-231 cells (18.6) compared to MCF-7 cells (12.4) (Figure 4). The distribution of positive synergy scores across multiple concentration combinations in MDA-MB-231 cells indicated a broader synergistic interaction landscape relative to MCF-7 cells. Importantly, this quantitatively stronger interaction profile in MDA-MB-231 cells was later paralleled by more pronounced oxidative stress induction, intracellular ROS accumulation, and mitochondrial membrane depolarization, suggesting that redox imbalance may underlie the enhanced synergistic response observed in the triple negative subtype.

Taken together, these analyses demonstrate that combination treatment enhances cytotoxic efficacy at reduced drug concentrations and that the strength and consistency of synergistic interactions are greater in triple-negative MDA-MB-231 cells than in hormone receptor positive MCF-7 cells. This synergistic interaction was further quantified using the Chou–Talalay median-effect method, which provides a dose-reduction index. The Combination Index (CI) was calculated across a range of effect levels (Fraction Affected, Fa). As shown in Figure 5, CI values were consistently below 1 for both cell lines, confirming synergy. In MDA-MB-231 cells, CI values ranged from 0.45 (strong synergy) at Fa = 0.5 to 0.72 (moderate synergy) at Fa = 0.9, with a mean CI at ED₅₀ of 0.58. In MCF-7 cells, CI values ranged from 0.65 to 0.85 (mean CI at ED₅₀ = 0.74) (Figure 5). These quantitative CI values reinforce the Bliss and HSA analyses and indicate that the combination enables effective cell kill at lower doses of each drug compared to monotherapy. Notably, the lower CI values observed in MDA-MB-231 cells were paralleled by a more pronounced increase in intracellular ROS generation and mitochondrial membrane depolarization under combination treatment, suggesting that enhanced redox-mediated apoptotic signaling may underlie the stronger synergistic interaction in the triple-negative subtype.

3.3. Oxidative Stress-Dependent Apoptotic Mechanisms Underlying the Synergistic Interaction

To elucidate the mechanistic basis underlying the synergistic interaction between Cur and Gem, oxidative stress parameters, apoptotic indices, and caspase activation were systematically analyzed in both breast cancer cell lines. In addition to global oxidative stress markers, intracellular ROS accumulation and mitochondrial membrane potential (ΔΨm) alterations were directly assessed to determine whether redox imbalance and mitochondrial dysfunction contribute to the observed synergistic cytotoxicity.

Flow cytometric analysis using Annexin V FITC and PI staining demonstrated that combination treatment induced a marked increase in apoptotic cell populations compared to single-agent treatments in both MCF-7 and MDA-MB-231 cells (Figure 6). In MCF-7 cells, the total apoptotic fraction increased from 6.2% in control cells to 24.8% following Cur treatment, 35.6% following Gem treatment, and 62.4% following combination treatment. In MDA-MB-231 cells, corresponding apoptotic fractions were 5.8%, 31.5%, 41.2%, and 71.8%, respectively (Figure 5). In both cell lines, the increase in apoptosis observed in the combination group was predominantly attributable to late apoptotic cells, while necrotic populations remained low. The magnitude of apoptotic induction was notably higher in the MDA-MB-231 cells, consistent with the stronger synergistic interaction observed in this subtype.

Assessment of cellular oxidative stress revealed significant alterations in TOS and TAS following drug treatment (Figure 7). Combination therapy resulted in a significant increase in TOS and a concomitant decrease in TAS in both cell lines. In MCF-7 cells, TOS increased approximately 2.5-fold and TAS decreased by approximately 40 percent compared to the control cells. Interestingly, Cur alone increased the TAS levels in MCF-7 cells, consistent with its intrinsic antioxidant properties. In MDA-MB-231 cells, these changes were more pronounced, with an approximately 3.1-fold increase in TOS and an approximately 50 percent reduction in TAS (Figure 7). While curcumin or gemcitabine alone induced moderate changes in TOS and TAS, the combination treatment produced the strongest redox imbalance in both subtypes. These changes were statistically significant in both Cur vs. Cur+Gem and Gem vs. Cur+Gem comparisons (MCF-7: p < 0.01; MDA-MB-231: p < 0.001). The higher oxidative stress response observed particularly in triple-negative MDA-MB-231 cells suggests that these cells may be more sensitive to combination therapy. Importantly, these global oxidative stress alterations were consistent with the intracellular ROS accumulation detected by DCFH-DA analysis and were accompanied by subsequent mitochondrial membrane depolarization, supporting a redox-driven apoptotic mechanism underlying the observed synergistic cytotoxicity.

To functionally validate the contribution of oxidative stress to the observed apoptotic synergy, loss of function experiments were performed using the ROS scavenger N-acetylcysteine. Pre-treatment with 5 mM N-acetylcysteine significantly attenuated the cytotoxic and apoptotic effects of the Cur and Gem combination in both cell lines (Figure 8). In MDA-MB-231 cells, N-acetylcysteine reduced the Bliss synergy score from 22.1 to 5.3, while in MCF-7 cells, the score decreased from 15.7 to 4.1. Correspondingly, N-acetylcysteine reduced the apoptotic cell population in the combination group by approximately 60 percent and restored TOS levels toward control values (Figure 8). This attenuation was accompanied by a marked reduction in intracellular ROS accumulation and partial restoration of mitochondrial membrane potential under combination treatment, indicating that redox modulation contributes substantially to the observed synergistic apoptotic response. Although apoptosis was not completely abolished, these data suggest that oxidative stress functions as a major, but not exclusive, upstream mediator of Cur-mediated Gem sensitization.

To directly evaluate oxidative stress under 48-h treatment conditions, intracellular ROS levels were measured using the DCFH-DA assay (Figure 9). Cur+Gem treatment induced a pronounced rightward shift in DCF fluorescence intensity in both MDA-MB-231 and MCF-7 cells. ROS levels increased approximately 3.4-fold in MDA-MB-231 cells and 2.7-fold in MCF-7 cells relative to the control. In contrast, MCF-10A cells exhibited only a modest increase (~1.2-fold). NAC pre-treatment markedly attenuated ROS accumulation in both cancer cell lines, while only minor modulation was observed in MCF-10A cells. To determine whether ROS elevation was associated with mitochondrial dysfunction, ΔΨm was assessed using JC-1 staining (Figure 10). In MDA-MB-231 cells, Cur+Gem treatment resulted in a substantial increase in the depolarized population, reaching approximately 86% of total cells compared with near-baseline levels in control conditions. NAC pre-treatment restored mitochondrial polarization and reduced depolarization to near-control levels. Similarly, in MCF-7 cells, combination treatment increased mitochondrial depolarization to approximately 69%, whereas control cells remained predominantly polarized. NAC exposure significantly reduced the proportion of depolarized cells. In contrast, MCF-10A cells displayed limited mitochondrial depolarization (~8%) following Cur+Gem treatment. NAC restored the polarized population to near-control levels.

The inclusion of MCF-10A cells in both DCFH-DA and JC-1 analyses indicates that ROS amplification and mitochondrial depolarization were largely restricted to malignant breast cancer cells under combination treatment conditions.

Collectively, these findings demonstrate that Cur-mediated Gem sensitization is associated with pronounced ROS accumulation and mitochondrial membrane depolarization in breast cancer cells, whereas normal breast epithelial cells exhibit only minimal mitochondrial disruption under identical experimental conditions.

Caspase activity assays demonstrated significant activation of intrinsic apoptotic signaling following combination treatment (Figure 11). In MCF-7 cells, combination therapy increased caspase-9 activity by approximately 4.2-fold. Caspase-3-like activity, as measured by cleavage of the DEVD-pNA substrate, increased approximately 4.8-fold. In MDA-MB-231 cells, CASP3 and CASP9 activities increased by approximately 5.6-fold and 4.9-fold, respectively (Figure 11). These increases were significantly greater than those observed with single-agent treatments. MCF-7 cells are widely reported to be deficient in functional caspase-3 protein due to a genomic deletion. The observed DEVDase activity is therefore likely attributable to other executioner caspases, such as caspase-7, which can cleave the same substrate.

Correlation analysis further supported the mechanistic association between oxidative stress and apoptotic execution. TOS levels positively correlated with apoptosis rates in both MDA-MB-231 cells (r = 0.87, p < 0.001) and MCF-7 cells (r = 0.79, p < 0.01). In addition, CASP3 activity correlated with the apoptosis rates (r = 0.92, p < 0.001), indicating the coordinated activation of oxidative stress and caspase-dependent apoptotic pathways under combination treatment conditions.

Together, these results demonstrate that the synergistic cytotoxic effect of the Cur and Gem combination is associated with a pronounced pro-oxidative shift, enhanced activation of intrinsic apoptotic pathways, and caspase dependent cell death, with these effects being more prominent in triple-negative MDA-MB-231 cells. This coordinated response was characterized by intracellular ROS amplification, mitochondrial membrane depolarization, and attenuation of these effects by N-acetylcysteine, while normal breast epithelial MCF-10A cells exhibited limited redox perturbation and preserved mitochondrial integrity under identical experimental conditions.

3.4. Modulation of Angiogenesis via VEGF Suppression

VEGF secretion levels were quantified in cell culture supernatants to assess the effects of Cur, Gem, and their combination on angiogenesis-related signaling. ELISA analyses demonstrated that both agents reduced VEGF protein secretion in a dose-dependent manner in MCF-7 and MDA-MB-231 cells, with the most pronounced inhibition observed following combination treatment (Figure 12).

In MCF-7 cells, combination therapy reduced VEGF secretion by approximately 65% compared to the control cells (p < 0.001). In MDA-MB-231 cells, VEGF levels were suppressed by approximately 72% following combination treatment (p < 0.001) (Figure 12). In both cell lines, the extent of VEGF suppression achieved with the combination exceeded that observed with either agent alone. In MCF-7 cells, combination therapy was significantly more effective than monotherapy (Cur vs. Cur+Gem, p < 0.01; Gem vs. Cur+Gem, p < 0.001). Similarly, in MDA-MB-231 cells, VEGF suppression following combination treatment was significantly greater than that observed with either curcumin or gemcitabine alone (Cur vs. Cur+Gem and Gem vs. Cur+Gem, p < 0.001).

These results demonstrate a marked reduction in the secretion of VEGF, a key pro-angiogenic cytokine, following combined Cur and Gem treatment in both breast cancer subtypes. The stronger VEGF suppression observed in MDA-MB-231 cells was consistent with the enhanced oxidative stress response and apoptotic activation identified in this subtype, suggesting that redox-dependent signaling alterations may contribute to the observed anti-angiogenic effect. This finding suggests a potential disruption of pro-angiogenic signaling that warrants further functional investigation.

3.5. qRT-PCR Validation of Transcriptomic Alterations in Key Signaling Pathways

The effects of combination therapy on apoptotic, inflammatory, and angiogenesis-related gene expression were evaluated in the MCF-7 and MDA-MB-231 cell lines using qRT-PCR analysis. The results revealed significant upregulation of pro-apoptotic and stress response-related genes, consistent with the observed ROS accumulation, mitochondrial membrane depolarization, and caspase activation induced by Cur+Gem treatment, with these changes being more pronounced in MDA-MB-231 cells.

Expression of the pro-apoptotic BAX gene was found to be log₂FC = 3.80 ± 0.15 (13.9-fold increase) in MDA-MB-231 cells and log₂FC = 3.20 ± 0.12 (9.2-fold increase) in MCF-7 cells. In contrast, only a limited increase was observed in the anti-apoptotic BCL2 gene in both cell lines (MDA-MB-231: 1.3 fold, MCF-7: 1.4 fold), indicating a shift in the apoptotic balance towards cell death. The expression of the tumor suppressor TP53 gene was also significantly increased after combination therapy. A log₂FC increase of 3.20 ± 0.18 (9.2-fold) was observed in MDA-MB-231 cells, and a log₂FC increase of 2.80 ± 0.14 (7.0-fold) in MCF-7 cells, suggesting the activation of cell cycle control and DNA damage response pathways.

Expression of CASP3 and CASP9 was also significantly increased in both cell lines. CASP3 expression increased 6.5-fold in MDA-MB-231 cells and 6.1-fold in MCF-7 cells, while CASP9 expression increased 9.2-fold and 7.5-fold, respectively. These transcriptional changes parallel the enzymatic caspase activation observed at the protein activity level, supporting engagement of the intrinsic apoptotic pathway downstream of oxidative stress and mitochondrial dysfunction.

In contrast, significant suppression was observed in genes associated with angiogenesis and survival. VEGFA expression was determined as log₂FC = 0.30 ± 0.05 (approximately 1.2-fold decrease) in MDA-MB-231 cells and log₂FC = 0.35 ± 0.06 (approximately 1.3-fold decrease) in MCF-7 cells. Similarly, NF-κB (NFKB) expression was suppressed by 1.4-fold in MDA-MB-231 cells and by 1.3-fold in MCF-7 cells. The downregulation of VEGFA at the transcript level is concordant with the reduced VEGF protein secretion detected by ELISA, indicating coordinated transcriptional and secretory suppression of angiogenesis-related signaling following combination treatment.

Overall, these results demonstrate that combination therapy strongly activates pro-apoptotic and tumor suppressor genes in both breast cancer cell lines while suppressing signaling pathways associated with angiogenesis and cell survival, thereby providing molecular support for the ROS-driven mitochondrial apoptotic mechanism identified in functional assays (Figure 13).

3.6. Global Transcriptomic Profiling and Pathway Enrichment Analysis

To obtain an unbiased, genome-wide view of the molecular effects induced by Cur, Gem, and their combination, RNA-Seq analysis was performed. Given the observed ROS amplification, mitochondrial membrane depolarization, and caspase activation following combination treatment, transcriptomic profiling was conducted to determine whether these functional alterations were reflected at the gene expression level. Differential gene expression analysis revealed that combination treatment resulted in the highest number of differentially expressed genes in both breast cancer cell lines. Specifically, 910 genes were differentially expressed in MCF-7 cells and 1060 genes in MDA-MB-231 cells based on the criteria |log₂-fold change| > 1.5 and FDR < 0.05 (Figure 14).

Volcano plot analysis was used to visualize the magnitude and statistical significance of gene expression changes. These plots demonstrated the prominent upregulation of apoptosis-related genes, including BAX, CASP3, and CASP9, alongside the downregulation of anti-apoptotic genes such as BCL2. In addition, key regulatory genes including TP53, CDKN1A, and VEGFA showed marked expression changes following combination treatment (Figure 15). Genes involved in mitochondrial apoptotic signaling and oxidative stress-associated pathways were also significantly altered, aligning with the observed ROS accumulation and mitochondrial membrane depolarization at the functional level. Hierarchical clustering and heat map visualization further confirmed distinct transcriptional profiles among treatment groups. The combination treatment group exhibited the most divergent gene expression pattern compared to the control and single-agent treated groups in both cell lines, indicating extensive transcriptional reprogramming (Figure 15).

PCA was performed to assess the global transcriptomic variance across samples. In MCF-7 cells, the first principal component (PC1) accounted for 51.8% of the total variance, while the second principal component (PC2) explained 25.7%. Similarly, in MDA-MB-231 cells, PC1 and PC2 explained 52.9% and 23.9% of the variance, respectively. PCA scatter plots demonstrated a clear separation between the control, Cur-, Gem-, and combination-treated samples in both cell lines. Samples from the combination treatment group were positioned furthest from the control group along PC1, indicating extensive transcriptional divergence associated with combined treatment (Figure 16).

To identify biological pathways associated with the observed transcriptional changes, pathway enrichment analyses were conducted. KEGG analysis revealed significant enrichment of cancer-related pathways, most notably the p53 signaling pathway, apoptosis, cell cycle regulation, VEGF signaling pathway, and NF-κB signaling pathway in both cell lines following combination treatment. These enriched pathways were consistent with the functional assays demonstrating ROS-associated apoptosis, caspase activation, and the suppression of angiogenesis-related signaling, thereby linking the transcriptomic alterations to the observed cellular phenotypes (Figure 17).

GO enrichment analysis further demonstrated significant modulation of biological processes related to programmed cell death, cell cycle arrest, and negative regulation of angiogenesis. These enriched biological processes were consistently identified in both breast cancer subtypes, with higher enrichment scores observed in MDA-MB-231 cells compared to MCF-7 cells (Figure 18).

In conclusion, both Cur and Gem exerted significant anticancer effects in hormone receptor–positive MCF-7 and triple-negative MDA-MB-231 breast cancer cells, as demonstrated by the reduced cell viability, enhanced apoptosis, disruption of oxidative stress homeostasis, activation of intrinsic caspase signaling, suppression of VEGF secretion, and extensive transcriptional reprogramming. In addition, intracellular ROS quantification by the DCFH-DA assay and assessment of mitochondrial membrane potential (ΔΨm) by JC-1 staining confirmed that combination treatment induced pronounced ROS accumulation and mitochondrial depolarization selectively in breast cancer cells, whereas MCF-10A normal breast epithelial cells exhibited limited ROS elevation and minimal mitochondrial disruption under identical conditions. Combined treatment consistently produced the most pronounced biological responses across all evaluated endpoints, indicating a predominantly synergistic interaction rather than a merely additive effect. The magnitude of these effects was particularly pronounced in the triple-negative breast cancer model, underscoring subtype-dependent differences in therapeutic responsiveness. The preservation of viability and mitochondrial integrity in MCF-10A cells at concentrations effective in cancer cells further supports the presence of a potential therapeutic window. Collectively, these findings suggest that Cur enhances the anticancer efficacy of Gem through the coordinated modulation of cytotoxic, apoptotic, oxidative, mitochondrial, angiogenic, and transcriptomic pathways in breast cancer cells, supporting the potential of this combination as a rational strategy for subtype-informed therapeutic development.

4. Discussion

In this study, we tested the hypothesis that Cur sensitizes breast cancer cells to Gem through ROS-mediated activation of the intrinsic apoptotic pathway. By integrating functional assays, redox analyses, quantitative synergy modeling, intracellular ROS quantification by DCFH-DA assay, mitochondrial membrane potential (ΔΨm) assessment by JC-1 staining, and transcriptome-wide profiling, we demonstrate that the Cur and Gem combination exerts significantly enhanced anticancer effects compared with either agent alone. These effects include reduced cell viability, robust induction of apoptosis, disruption of redox homeostasis, accumulation of intracellular ROS, mitochondrial depolarization, activation of mitochondrial caspase signaling, suppression of angiogenic signaling, and extensive transcriptional reprogramming. Comparative analyses performed in MCF-10A normal breast epithelial cells revealed substantially lower ROS accumulation, minimal mitochondrial depolarization, and preservation of cell viability at concentrations effective in cancer cells, supporting a degree of treatment selectivity under in vitro conditions. These findings should be interpreted within the context of an in vitro, observational framework, and the mechanistic interpretations presented herein are proposed as working hypotheses rather than definitive causal pathways.

Cur is widely recognized as a pleiotropic anticancer compound capable of modulating multiple oncogenic pathways. Previous studies have shown that Cur suppresses proliferation, induces apoptosis, and inhibits migratory and invasive potential in triple-negative breast cancer cells through regulation of the PI3K/Akt, MAPK, Wnt/β-catenin, and EGFR-associated signaling pathways [21]. In MDA-MB-231 cells, Cur has also been reported to induce cell cycle arrest and apoptosis by altering p21 expression and the BAX/BCL2 ratio, consistent with activation of the intrinsic apoptotic pathway [22,23]. Our data are consistent with these reports and extend them by demonstrating that Cur, when combined with Gem, is associated with a marked enhancement of Gem-induced cytotoxicity, as quantitatively confirmed using Bliss, HSA, and Chou–Talalay models. In addition to pharmacological synergy, the combination was accompanied by increased intracellular ROS accumulation, mitochondrial membrane depolarization, and augmented caspase activation, suggesting convergence on redox-sensitive intrinsic apoptotic mechanisms. However, while these synergy analyses robustly establish a pharmacological interaction, they do not by themselves define the molecular origin of this effect, and the observed redox and mitochondrial alterations should be interpreted as associated events within the experimental framework rather than definitive proof of causality.

A key observation of the present study is that Cur–Gem combination treatment coincided with pronounced disruption of redox homeostasis, characterized by increased total oxidant status and reduced antioxidant capacity. This redox imbalance was further corroborated by direct quantification of intracellular ROS using the DCFH-DA assay, which demonstrated substantial ROS accumulation in breast cancer cells but only modest elevation in MCF-10A cells under identical treatment conditions. Functional attenuation of cytotoxicity and apoptosis following ROS scavenging with NAC supports the interpretation that oxidative stress is functionally involved in the observed synergy. Moreover, NAC-mediated restoration of mitochondrial membrane potential, as assessed by JC-1 staining, indicates that ROS accumulation is closely associated with mitochondrial depolarization within this experimental setting. Nevertheless, while these findings argue against ROS being merely a passive by-product of cell death, they do not fully establish ROS as the sole or primary upstream driver. Rather, our results suggest that redox imbalance represents a critical mechanistic component of the synergistic response, acting in concert with mitochondrial dysfunction and downstream caspase activation rather than functioning as an isolated initiating event.

A finding requiring careful interpretation is the measured increase in DEVD-pNA cleaving activity (“caspase-3-like” activity) in MCF-7 cells. It is well-established that MCF-7 cells harbor a 47-bp deletion in exon 3 of the CASP3 gene, leading to a lack of functional caspase-3 protein. Therefore, the observed activity in our assay is not attributable to caspase-3 itself but likely reflects the activation of other executioner caspases, primarily caspase-7, which shares the DEVD cleavage motif and can compensate for caspase-3 deficiency in these cells [24]. This interpretation is supported by our qRT-PCR data, which showed upregulation of CASP3 mRNA, a finding consistent with transcriptional activity despite the non-functional gene product and more importantly, a significant upregulation of the intrinsic initiator CASP9. The robust activation of caspase-9, coupled with the dramatic increase in the BAX/BCL2 ratio and Annexin V positivity, confirms the potent induction of the intrinsic apoptotic pathway in MCF-7 cells. These findings are further supported by the observed mitochondrial membrane depolarization and intracellular ROS accumulation, which together indicate the activation of a redox-sensitive mitochondrial apoptotic axis. The mode of execution, however, may proceed via caspase-7 or other compensatory mechanisms. In contrast, the apoptosis observed in MDA-MB-231 cells aligns with the canonical caspase-9/-3 cascade. This distinction underscores the importance of considering cell line-specific molecular backgrounds when interpreting caspase activity data and reinforces the conclusion that the Cur–Gem combination effectively triggers the mitochondrial apoptosis pathway, albeit with subtype-specific variations in the downstream executioner phase.

A central mechanistic insight of the present study is the functional involvement of oxidative stress in mediating the observed synergy. Combination treatment induced a pronounced redox imbalance, characterized by increased TOS and reduced antioxidant capacity. Pharmacological scavenging of ROS using NAC significantly attenuated cytotoxicity, reduced apoptotic fractions, and diminished synergy scores. In addition, direct quantification of intracellular ROS using the DCFH-DA assay confirmed a substantial increase in ROS levels in breast cancer cells following combination treatment, whereas MCF-10A cells exhibited only modest ROS elevation under identical conditions. This functional reversal supports the interpretation that ROS generation represents a required upstream component rather than merely a secondary consequence of cell death. These observations are consistent with prior reports demonstrating that phytochemicals can sensitize cancer cells to chemotherapeutic agents via redox modulation [25,26] and support Cur’s dual role as both a redox-active and cytotoxic agent [27]. Both curcumin and gemcitabine have been independently shown to induce intracellular ROS in breast cancer cells. Curcumin, as a natural polyphenol, disrupts the mitochondrial electron transport chain, leading to superoxide production [28]. Gemcitabine, through its incorporation into DNA and inhibition of DNA synthesis, triggers oxidative stress as a secondary effect [29]. Our observation of synergistic apoptosis induction, particularly the marked increase in late apoptotic cells, aligns with the expected outcome of combined oxidative stress. Consistently, JC-1 analysis demonstrated significant mitochondrial membrane depolarization in cancer cells following combination treatment, and this effect was partially reversed by NAC, further linking ROS accumulation to mitochondrial dysfunction within the intrinsic apoptotic pathway. The mitochondrial membrane potential collapse and caspase activation patterns we observed are characteristic of ROS-mediated intrinsic apoptosis pathways. Taken together, the integrated redox, mitochondrial, and caspase data provide convergent evidence supporting a redox-sensitive mitochondrial mechanism underlying the synergistic interaction.

Consistent with ROS-driven mitochondrial dysfunction, combination therapy robustly activated caspase-9 and caspase-3 and markedly increased the BAX/BCL2 ratio, confirming intrinsic apoptosis as the dominant execution pathway. These molecular changes were accompanied by significant mitochondrial membrane depolarization as demonstrated by JC-1 analysis, further supporting activation of the redox-sensitive intrinsic apoptotic cascade. Transcriptomic analyses further supported this mechanism by revealing coordinated upregulation of TP53 and caspase genes alongside suppression of NF-κB- and VEGFA-associated survival pathways. Collectively, these molecular alterations create a transcriptional landscape that lowers the apoptotic threshold, facilitates Gem-induced cell death, and simultaneously attenuates pro-survival and pro-angiogenic signaling. The limited ROS accumulation and minimal mitochondrial depolarization observed in MCF-10A cells under identical treatment conditions suggest that this apoptotic sensitization is preferentially manifested in malignant cells within the in vitro model. MCF-7 cells are known to be deficient in functional caspase-3. Therefore, apoptotic responses observed in this cell line may involve caspase-7 or caspase-independent pathways. Accordingly, caspase-3-related findings in MCF-7 cells should be interpreted with cell line-specific limitations in mind.

The synergistic interaction between Cur and Gem was consistently more pronounced in the triple-negative MDA-MB-231 model than in hormone receptor-positive MCF-7 cells. This subtype-dependent sensitivity likely reflects intrinsic biological differences, including higher basal oxidative stress, altered redox buffering capacity, frequent p53 mutations, increased metabolic stress, and greater reliance on compensatory survival pathways in triple-negative breast cancer. The greater ROS accumulation and more extensive mitochondrial depolarization observed in MDA-MB-231 cells following combination treatment further align with this enhanced sensitivity within the in vitro setting. Such features may render triple-negative breast cancer (TNBC) cells particularly vulnerable to redox-modulating combination strategies, a concept supported by previous studies reporting heightened sensitivity of TNBC cells to natural compound-based combinations [30]. Synergy was rigorously confirmed using complementary quantitative models, including Bliss, HSA, and the Chou–Talalay CI, strengthening the robustness of the observed interaction.

Beyond breast cancer, synergistic interactions between Cur and Gem have been reported in other malignancies, including cholangiocarcinoma, where combined treatment suppressed proliferation and tumor growth in resistant models [31,32]. Broader analyses have emphasized Cur’s capacity to enhance chemotherapeutic efficacy through combinatorial approaches and redox modulation [33]. Similar chemosensitizing effects have been described for other natural compounds in Gem-resistant cancer settings [34], underscoring the broader relevance of natural compound-based sensitization strategies. Within this broader context, the present findings extend these observations by providing integrated redox, mitochondrial, and transcriptomic evidence supporting a redox-associated chemosensitization mechanism in breast cancer cells.

From a translational perspective, the relevance of our findings is supported by the alignment of Gem concentrations used in this study with clinically achievable plasma levels [35]. In contrast, Cur’s limited bioavailability remains a recognized challenge due to poor absorption and rapid metabolism [36,37,38]. Advances in formulation strategies have demonstrated marked improvements in systemic exposure and tumor accumulation, enabling pharmacologically relevant concentrations in vivo [39,40,41,42]. Accordingly, the present findings should be interpreted as a mechanistic proof-of-concept rather than a direct dosing recommendation, providing a rationale for future in vivo validation using optimized curcumin delivery systems.

To provide an integrated interpretation of our experimental findings, we propose a unified mechanistic model summarizing the sequence of events underlying the synergistic interaction between Cur and Gem (Figure 19). As depicted, combination treatment induces a pronounced accumulation of intracellular ROS, which functions as a central mechanistic component within the observed response rather than merely as a secondary consequence of cytotoxicity. This ROS surge is associated with the disruption of mitochondrial membrane potential, resulting in mitochondrial dysfunction and subsequent activation of the intrinsic apoptotic pathway. The initiation of this cascade culminates in sequential activation of caspase-9 and caspase-3, driving apoptotic commitment. The functional contribution of ROS to this cascade is supported by NAC-mediated attenuation of apoptosis, restoration of mitochondrial membrane potential, and reduction in synergistic cytotoxicity. In parallel, transcriptomic reprogramming reinforces the pro-apoptotic phenotype through the upregulation of BAX, TP53, and executioner caspases, alongside suppression of BCL2 and NF-κB-dependent survival signaling. As a downstream consequence of ROS-associated cellular stress, VEGFA expression and VEGF secretion are reduced, contributing to the attenuation of angiogenic potential. The magnitude of these interconnected effects was greater in MDA-MB-231 cells and accompanied by comparatively limited ROS accumulation and mitochondrial disruption in MCF-10A cells, supporting subtype-dependent and cell-selective vulnerability within the in vitro model. Collectively, Figure 19 integrates functional, mitochondrial, redox, and transcriptional evidence into a coherent mechanistic framework that explains the observed synergy and underscores the translational relevance of the Cur and Gem combination.

While direct comparison with normal breast epithelial cells was incorporated into our experimental design, with MCF-10A cells maintaining higher viability (>70%) at concentrations effective in breast cancer cells, these findings are consistent with the broader literature describing cancer-selective responses to both curcumin and ROS-modulating strategies. Curcumin has been reported to exert preferential cytotoxicity toward malignant cells, with several studies demonstrating IC₅₀ values that are frequently 2–4-fold higher in normal mammary epithelial cells compared to breast cancer cells [43]. Our observation of an approximately 3-fold higher IC₅₀ for curcumin in MCF-10A relative to MCF-7 cells is aligned with this pattern. This differential sensitivity is biologically plausible, as cancer cells typically exhibit elevated basal ROS levels and altered redox buffering capacity, rendering them more dependent on antioxidant defense systems and therefore more susceptible to additional oxidative stress, a concept described as ‘differential stress sensitization’ or ‘redox vulnerability’ [44,45]. Gemcitabine, although associated with dose-limiting hematological toxicities in clinical settings, displayed a comparatively higher IC₅₀ in MCF-10A cells (~15–20 µM) than in breast cancer cell lines, supporting the presence of a potential therapeutic window under the experimental conditions employed. The synergistic interaction between curcumin and gemcitabine was confirmed using multiple complementary quantitative models, including the Bliss independence model, the HSA model, and the Chou–Talalay Combination Index method. CI analysis yielded mean values of 0.58 in MDA-MB-231 cells and 0.74 in MCF-7 cells at the IC₅₀ level, corresponding to strong and moderate synergy, respectively. These values indicate that reduced doses of each agent are sufficient to achieve equivalent cytotoxic effects when administered in combination. The calculated dose-reduction index values greater than 1 further support this interpretation and suggest the theoretical possibility of lowering individual drug exposure while maintaining efficacy. At the same time, several limitations should be acknowledged. Although inclusion of MCF-10A cells provides preliminary evidence of differential sensitivity, evaluation was restricted to short-term (48-h) viability assessment and does not fully address long-term, functional, or tissue-specific effects in normal cells. Additional endpoints, including ROS modulation, apoptosis profiling, and functional assays in non-malignant breast epithelial models, would strengthen characterization of the therapeutic index. Furthermore, oxidative stress assessment in the present study relied primarily on global TOS and TAS measurements; incorporation of mitochondrial-specific ROS indicators and complementary membrane potential analyses in future studies would provide deeper mechanistic resolution. Finally, the curcumin concentrations required to achieve synergy in vitro exceed the plasma levels attainable with conventional formulations, highlighting a translational constraint that underscores the need for optimized delivery systems. The use of two representative breast cancer subtypes and a single normal epithelial line also limits generalization across the full spectrum of breast cancer heterogeneity and normal tissue responses.

The effective curcumin concentrations identified in this study (IC₅₀ ~ 24–27 µM) underscore a well-known pharmacokinetic limitation: the poor bioavailability of native curcumin restricts its peak plasma concentration to the low micromolar range following oral administration [46]. This discrepancy between in vitro efficacy and in vivo exposure represents a recognized translational challenge for many natural compounds. Accordingly, the present findings should be interpreted as defining a mechanistic and pharmacodynamic target range rather than as a direct clinical dosing recommendation. A growing body of preclinical and early clinical evidence indicates that advanced formulation strategies may partially overcome these limitations. Liposomal and polymeric nanoparticle curcumin formulations have been reported to increase the plasma area under the curve (AUC) by 10- to 40-fold and enhance tumor accumulation in xenograft models, achieving intratumoral concentrations approaching those shown to be biologically active in vitro [47,48]. Curcumin–phospholipid complexes (e.g., Meriva^®) and micellar formulations have similarly demonstrated improved systemic exposure and prolonged bioavailability in experimental settings [49]. Within this context, evaluation of the curcumin–gemcitabine combination using optimized nanoformulations in appropriate in vivo breast cancer models represents a rational and necessary next step. Such studies would enable the validation of the observed synergistic interaction under physiologically relevant conditions, facilitate the assessment of therapeutic index, and generate essential pharmacokinetic/pharmacodynamic (PK/PD) data to inform subsequent translational development. Thus, the in vitro synergy demonstrated here provides a mechanistic and quantitative foundation to justify further preclinical investigation using formulation-enhanced curcumin platforms.

While our study demonstrated significant upregulation of apoptosis-related genes at the transcriptional level (CASP3, CASP9, BAX), we acknowledge that protein-level validation is essential to confirm functional apoptosis execution. Previous studies have established strong correlations between mRNA and protein levels for these key apoptosis regulators in breast cancer cells [50]. For instance, Sarkar et al. showed that curcumin-induced CASP3 mRNA upregulation directly translated to increased cleaved caspase-3 protein in MCF-7 cells. Similarly, the BAX/BCL-2 ratio, a critical determinant of mitochondrial apoptosis commitment, has been consistently shown to correlate at both transcriptional and translational levels in response to chemotherapeutic agents [28]. Nevertheless, it should be emphasized that mRNA abundance does not invariably predict protein activity, particularly for caspases, which require post-translational cleavage for activation. Therefore, although our transcriptional findings are directionally consistent with intrinsic apoptotic activation and are supported by functional assays (Annexin V positivity and caspase activity measurements), future studies incorporating Western blot or immunodetection of cleaved caspase forms and BAX/BCL-2 protein levels would further strengthen mechanistic confirmation.

Nevertheless, future studies should include Western blot analysis of cleaved CASP3, cleaved CASP9, PARP cleavage, and quantitative assessment of BAX/BCL-2 protein ratios to provide definitive protein-level confirmation. Integration of these protein-based endpoints with the already established caspase activity assays would further substantiate execution-phase apoptosis. Additional techniques such as immunocytochemistry and confocal imaging to visualize mitochondrial localization and cytochrome c release would further strengthen these findings. Moreover, in vivo validation of the Cur and Gem combination using orthotopic and patient-derived xenograft models, particularly in triple-negative breast cancer, is warranted to confirm therapeutic efficacy within a physiologically relevant tumor microenvironment. Incorporation of optimized Cur delivery platforms will be critical to overcome pharmacokinetic limitations and evaluate therapeutic efficacy at clinically relevant exposure levels. In addition, direct assessment of intracellular redox dynamics, including mitochondrial-specific ROS measurements and real-time ΔΨm monitoring, as well as DNA damage response markers, will further refine the mechanistic framework proposed here. Expanding this strategy to additional breast cancer subtypes and drug-resistant models may facilitate biomarker-driven patient stratification and rational clinical translation.

5. Conclusions

This study demonstrates that curcumin enhances the anticancer efficacy of gemcitabine in breast cancer cells through the coordinated modulation of oxidative stress and mitochondrial apoptotic signaling. Quantitative synergy modeling (Bliss, HSA, and Chou–Talalay) consistently confirmed a synergistic interaction, which was more pronounced in triple-negative MDA-MB-231 cells. Direct intracellular ROS measurement (DCFH-DA) showed marked ROS accumulation following combination treatment, accompanied by significant mitochondrial membrane depolarization, as demonstrated by JC-1 analysis. ROS scavenging with N-acetylcysteine attenuated apoptosis, reduced mitochondrial disruption, and diminished synergy scores, functionally linking redox imbalance to the cooperative cytotoxic effect. Combination treatment also activated caspase-dependent intrinsic apoptosis, suppressed VEGF secretion, and induced broad transcriptional reprogramming favoring pro-apoptotic pathways. Inclusion of MCF-10A normal breast epithelial cells revealed preserved viability and minimal ROS and mitochondrial disruption at cancer-effective concentrations, supporting a potential therapeutic window. Collectively, these findings define a mechanistically coherent framework in which curcumin enhances gemcitabine efficacy via ROS-associated mitochondrial apoptosis, particularly in triple-negative breast cancer. Further in vivo validation and optimized curcumin formulations will be necessary to translate this strategy into clinically relevant applications.