Drug resistance (DR) in cancer is developed, regulated, and balanced by multiple determinants that include environmental and intrinsic factors besides treatment pressures. 1 Molecular mechanisms by which these determinants promote cancer DR are diverse and span from drug-receptor Breast cancer (BC) chemotherapy frequently ends up with DR which may confer a more aggressive nature to patient tumors. Doxorubicin (DOX; Adriamycin) is a natural anthracycline used to treat solid and blood tumors. 2 DOX is broadly considered the most effective single drug available for the treatment of BC. 23 The compound damages cancer cells by intercalating into DNA, thereby disrupting topoisomerase-II (TOP2)-mediated DNA repair, and destroying cellular membranes by generating free radicals. 4 A combination of anthracyclines such as DOX and taxanes has produced the most desirable disease-free and/or overall survival in BC patients. 5 DOX therapy, on the other hand, is capable of inducing DR that leads to tumor growth and eventually poor patient prognosis and survival. 6 Despite several efforts to remedy it, DOX resistance remains a major unresolved issue in BC therapy. Multiple mechanisms, including reduced intracellular DOX concentrations caused by overexpressed membrane efflux pumps, increased drug-metabolizing enzymes, reduced concentrations and activity of TOP2A, and deregulated intracellular apoptotic pathways, are the main causes of DOX resistance. 67 Some reports are available that outline modeling DOX resistance in BC and other cancer types in vitro and in vivo. They include cell lines of breast and colorectal cancers 891011 and a murine model of BC. 6 These studies indicate profound morphological, molecular, and behavioral changes in resistant tumor models that need to be taken into consideration when searching for effective solutions to DOX resistance.

We have been studying DR in BC for some time and have created trastuzumab-resistant models of BC tumor xenografts for molecular analyses. 12 In the current study, we generated MCF-7 cell sublines resistant to various doses of DOX and examined some changes that occur in the course of resistance acquisition. We then tested our notion that a combination of the anti-tumorigenic compound shikonin (SHKN) and anti-diabetic metformin (MTFN) could effectively remedy resistance and induce apoptotic cell death.

Human BC cell line MCF-7 was obtained from the National Cell Bank of Pasture Institute (Tehran, Iran), cultured in Dulbecco's modified eagle's medium (DMEM; Gibco, 12430054) supplemented with 10% fetal calf serum (FBS; Gibco, A4766801), and maintained at 37 °C and 5% CO2. Shikonin was purchased from Sigma (54952-43-10) and dissolved in dimethyl sulfoxide (DMSO; Thermo Fisher, 20688). Metformin was gifted by Dr. Abidi Co (Tehran) and dissolved in PBS to form treatment stocks.

Doxorubicin Hydrochloride (DOX.Hcl) was dissolved in DMSO ( 2mg/ml) to prepare initial stocks. MCF-7 cells were split 24 hrs before setting up any experiment involving the cells. About 15000 cells/well were seeded in a 96-well plate. After 24 hrs of growth, the cells were treated with serial concentrations of DOX stocks prepared beforehand from 1 to 2500 nanomolar (nM). The cells were reincubated for another 24 hrs before adding the MTT solution and being subjected to the viability assay for IC 50 determination. This experiment was repeated three times and, parallel to the drug, the same volumes of DMSO, a DOX solvent, were applied to the treatment/viability assay.

MTT viability was done as a standard method. Briefly, MCF-7 cells were treated with serial concentrations of DOX in 96-well plates. DMSO, PBS (drugs' solvents), and untreated cells were used in parallel as our base controls. After a 24-hr incubation period, we exposed the cells to the MTT (Sigma, M5655) solution to measure viability. In the end, we read the absorbance at 580 nm using an ELISA reader. The calculated inhibitory concentrations (IC50s) represent the treatment concentrations that inhibit 50% of cells' growth versus controls.

The strategy to generate DOX-resistant sublines is outlined in the Results section. To make sure of resistance continuity, we treated the cells with both specific doses of DOX and serial concentrations of the drug, before examining viability.

The scratch assay was used to assess migration. Wild-type MCF-7 cells and resistant sublines were seeded in 6-well plates (4×10 5 cells per well) to become 80% confluent in 24 hrs and ready for drug treatment. The next day, a sterile razor blade was used to scrape cells off the culture plate, leaving a denuded area and a sharp visible demarcation line at the wound edge. The wounded monolayers were gently washed with PBS twice and inspected under phase-contrast microscopy immediately after wounding. Then, sections of the wounds were selected according to the criteria and numbered. The photographs collected from cell samples were trimmed and edited so that they would more closely represent the migration process. Migrated cells were counted in sections 500 μm in length, allowing a 20 μm space from the scratch line to minimize the possible physical effects of cell movement resulting from cell proliferation. Statistical analysis was calculated by averaging a mean of six sections per test for each experiment. The number of migrated cells was expressed as mean±SEM. We repeated these experiments trice in each group to ensure reproducibility.

Cell co-staining and live-cell count Acridine orange (AO) co-applied with ethidium bromide (EB) stains nuclear DNA green in live cells and orange in dead cells. We followed our standard AO/EB co-staining method to measure the percentage of viable treated MCF-7 cells. Briefly, 50-µL suspension of each cell group was mixed with 50 µL of the pre-made stain stock and subjected to counting under fluorescent light. Six random microscopic fields per well were selected to count an average of 120 cells for each cell group in triplicates (three wells per group), as we have reported. 12

Suspensions of untreated cell samples and those treated for 24 hrs were incubated with fluorochromeconjugated antibodies (1:1) against human CD24 (PE-conjugated) (Thermo Fisher, 12-0242-82) and CD44 (FITC-conjugated) (Thermo Fisher, 11-0441-82) at 4°C in dark for 30-40 min and detected in the FL2/PE channel. The labeled cells were analyzed by flow cytometry.

RNA extraction and cDNA synthesis were done as we have described. 12 Each reaction included template cDNA, gene-specific primers, and SYBRGreen I PCR Master Mix. The human GAPDH was used as an internal control. All data were normalized to the expressed human GAPDH gene and alterations of expression were measured using the delta-delta Ct method. 13 Table 1 shows the primers used for our reactions.

The structure of DOX and shikonin were taken, respectively, from these sites: www.zinc15.docking.rg, and https://pubchem.ncbi.nlm.nih.gov/. Autodock Vina software was used to examine molecular interactions in an in-silico environment. The software was also applied for final analyses and binding energy measurement, as we have reported. 14 Finally, Discovery Studio Visualization (DSV) was applied to visualize 2D and 3D binding structures.

The results are presented as mean ± SE. Statistical comparisons between groups were carried out using one-way ANOVA followed by the Student's t-test. We used the data of triplicates or more repetitions for each set of experiments, and comparisons were carried out only between relevant pairs of data. We considered a value of P<0.05 as statistically significant, and P<0.01 or P<0.001 as highly significant.

First, we treated WT MCF-7 cells with serial concentrations of DOX and measured cell viability ( Figure 1A). We determined 1 µM as the IC 50 of DOX. To begin with our strategy, we split the WT cells and, 24 hrs after, treated them using 5% of the DOX IC50 for a few weeks until they adapted to the drug ( Figure 1B). We split the cells again, preserved a fraction as our first MCF-7 subline named 5%, and maintained the second half in a drug-free medium for one week. This time, we split these cells and treated them with 5% of the drug IC50 for 24 hrs, followed by a second dose (10%) for a few weeks for adaptation and generation of the second subline that was named 10%. We repeated these cycles of treatment, adaptation, and drug-free growth until sublines of 15%, 20%, and 25% were generated. Morphologically, there were no major differences between various sublines and their ancestral WT cells when we observed them under the phase-contrast microscope ( Figure 1C). We found the sublines at lower doses (5-15%) unstable in the long term as they started decreasing in number and perishing after several times of splitting. In comparison, the 20% and particularly 25% sublines showed long-term stability that provided a reasonable time window for us to subject them to further experimentation. In addition, our attempt to generate 30% and beyond faced failure as the cells did not survive to adapt to higher DOX concentrations. Because of their reasonable stability, the 25% subline was most relied upon for our subsequent analyses. Each column represents the average of three independent experiments that included triplicate wells of cells in 96-well plates for each DOX concentration. As a control, the same volumes of DMSO as the DOX solvent were applied. These data were applied to normalize the data of DOX-and shikonin-treated cells in all subsequent graphs of cell growth. Symbol *, **, *** represent statistical differences between each DOX treatment with its DMSO control. b) A strategy to generate resistant sublines. c) Morphology of a number of isolated resistant sublines.

To test if our sublines maintain their DOXresistant property, we first grew them in a medium that contained their dose of DOX for a few days. Next, we split the cells and seeded them in equal numbers in 96-well plates, incubated them with and without DOX for 24 hrs, and tested their viability using MTT. In our control (no drug) group, both 20% and 25% resistant sublines showed comparable viability to our WT cells (Figure 2A). Treatment with 20% of the drug IC50 slowed down viability in WT cells to some extent (non-significant) but failed to do so in our 20% and 25% sublines. The 25% dose of the IC50 slightly reduced viability in WT and the 20% subline but not in our 25% subline (Figure 2A). These observations indicate that the sublines maintained their growth and stability when exposed to their dose of drug concentrations. Next, we measured cell viability by treating our sublines with serial concentrations of DOX. We found that while there is a sharp decline in WT cell survival, this decline is slowed down in our resistant sublines ( Figure 2B). Our last test was on the migrating capacity of the two sublines (Fig. 2C). Numerical measurement based on the counting of the number of cells crossing the scratch line indicated our 20% and 25% sublines preserve their potential to migrate, and, the 25% even does significantly better than both WT control and the 20% subline ( Figure 2D; P<0.05).

Once we ensured our generated sublines maintained their DOX resistance capacity, we monitored three classes of molecular changes as postresistance acquisition in our 25% subline. Firstly, real-time experiments revealed upregulated genes that included cell cycle gene CCND1, growth signaling genes PI3K and AKT1, ABCB1, and ABCC1, and CYP1A1, anti-apoptosis gene BCL-2, pro-migration genes SNAIL, STAT3 and MMP1 ( Figure 3A and Table 2). CYP1A1 showed the most significant upregulation (by 105 folds; P<0.01, compared to the WT MCF-7), while others had a 3 to 14 folds increase in their mRNA expression. In contrast, these genes were downregulated: proapoptosis genes BAX, p53, anti-migration gene CDH1, anti-chemoresistant gene TOP2A, and tumor suppressors BRCA2 and p21. The most notable reduction was observed in the expression of TOP2A (by 102 folds; P<0.01), followed by CDH1 (21 folds; P<0.01). Our second class of molecular studies examined virtual binding between DOX and a number of these molecules. In particular, DOX binding to CYP1A1 and TOP2A was evaluated ( Figure 3B). We detected strong binding energies: -8.4 between DOX and TOP2A and -7.5 between DOX and CYP1A1. We have already published the images of binding between DOX and key molecules of intracellular signaling pathways. 14 The energies of these bindings are listed in Table 3.

Due to the importance of CSCs in cancer DR, we devoted our third class of molecular studies to changes in stem cell markers in our resistant 25% subline. Figure 3.C shows over 120% increase in the CD44/CD24 ratio in our resistant subline compared to its WT ancestor (P<0.05). Each column represents an average of three independent experiments. Symbols * and *** represent, respectively, significant and highly significant differences in gene expression between the 25% resistant subline the and WT MCF-7 cells. b) A diagram representing binding between doxorubicin and key molecules TOP2A, and CYP101, generated by the molecular docking software, as outlined in our previous study. 14 c) The ratio of stem cell markers CD44:CD24 in resistant subline compared to WT MCF-7.

Based on our previous observations on the impact of SHKN-MTFN synergy on the inhibition of cancer cell migration and reversal of EMT, 14 we examined the impact of the combination on our resistant sublines. Applying a similar range of doses, we determined IC50 of SHKN to be 10 µM for WT MCF-7, 12 µM for our 20% subline, and 15 µM for our 25% subline ( Table 4). The figures for MTFN stood, respectively, at 60, 68, and 75 mM. Based on our previous report, 14 we used the RSM method and determined these IC50 doses for SHKN-MTFN cotreatment: 7 µM /40 mM for WT MCF-7, 9 µM /45 mM for our 20% subline, and 11 µM /52 mM for our 25% subline (see Table 4). Figure 4A shows that, compared to the untreated WT MCF-7 cells, increasing concentrations of SHKN and MTFN in separation significantly reduce cell survival. This trend emerged in our resistant 25% subline as well. When we used the two compounds in their sublethal doses, similar or lower levels of cell viability were detected in both WT and resistant cell groups. In particular, WT cells showed significantly reduced levels of viability when treated with DOX IC50 (P<0.05). However, the resistant subline was not affected by 25% of the IC50 dose, indicating its ongoing adaptation to the drug. Our SHKN-MTFN co-treated samples showed similar levels of viability to their single-treated counterparts, which were further and significantly reduced when higher doses of the combination drugs were applied ( Figure 4A). We also examined the impact of SHKN-MTFN synergy on cell apoptosis ( Figure 4B) and found that the combined treatment significantly induces cell death compared to WT cells ( Figure 4C; P<0.01) and single treatments (P<0.05).

We then tested the capacity of SHKN-MTFN synergy to inhibit subline migration. This potential of our 25% subline was reduced dramatically by single treatments (Figure 5A,B). Numerical counting of the cells crossing the scratch line determined more reduction in cell number with increasing SHKN or MTFN dosage and in 24 hrs post-scratch compared to 12 hrs ( Figure 5C). Co-treatment with both compounds ( Figure 5C) significantly reduced the number of migrating cells compared to both WT and single-treated cell groups ( Figure 5D; P<0.01).

We examined molecular changes that might occur upon treatment and co-treatment of our resistant subline with SHKN and MTFN. We looked at the changes in CD44/CD24 ratios in treated samples. Table 5 summarizes these changes. As reported before, 14 treatments with SHKN, and to some extent MTFN, significantly reduced the CD44/CD24 ratio in WT MCF-7 (P<0.01). Likewise, the increasing concentrations of either drug significantly reduced the ratio in our resistant 25% subline (Table 5; SHKN P<0.01; MTFN P<0.05). Similar to the synergistic impact of SHKN-MTFN co-treatment on MCF-7 cell tumorigenicity, 14 co-treatment using various sublethal concentrations of the compounds also reduced the ratio in our 25% subline. These reductions were highly significant compared to our single treatments (see Table 5 for comparisons and significance of changes). In our previous study, we showed a strong binding between SHKN and several key molecules involved in BC cell growth and survival. 14 Based on the role of SHKN in inhibiting DR, migration, and EMT observed in the current study, we further examined SHKN binding to two key more molecules that showed profound changes in their expression: CYP1A1 and TOP2A. The binding energy of SHKN was determined to be -6.4 KCal/Mol for CYP101 and -6.7 KCal/Mol for TOP2A. There is no fixed baseline for ligand-receptor binding, but based on different investigators' comments on the world wide web, we took -6 Kcal/Mol and below as a sign of strong binding. Table 3 lists the binding energies.

Doxorubicin is currently the most effective single chemo-drug for BC therapy. However, treated tumors often face DOX resistance that ends up in tumor recurrence and poor patient prognosis. In general, intra-tumoral heterogeneity and clonal evolution that stem from somatic mutations and genomic plasticity during tumor development form the foundations of cancer DR including DOX resistance. 15 These changes are fueled by interactions between signaling pathways that induce proliferation, cell cycle progression and apoptosis inhibition. 16 Our study was aimed at dissecting some details of mechanisms behind DOX resistance and examining the notion that shikonin-metformin synergy can restore cancer cell responsiveness.

In order to generate our resistant sublines, we designed and adopted a strategy based on which DOX doses beginning with 5% of IC50 and increasing in increments were applied for cell treatment. We also applied phases of treatment omission in order to break continuous drug pressure on cells and minimize their stressful time periods. Other studies also report the establishment of their DOX-resistant MCF-7 sublines by sequentially treating the cells with concentrations of the drug that ranged from 0.1 to 1 µM. 111617 Christowitz et al. further used their resistant sublines to develop a xenograft model. 6 DOX-resistant BC tumors have also been generated in vivo by injecting mouse models with the drug at intervals until the tumors grew to a certain volume. 6 Our generated sublines demonstrated their capacity to maintain resistance to DOX over time. Preliminary analyses indicated that they have a higher growth rate and less apoptosis rate than their DOXsensitive ancestors. Also, the 25% subline displayed significantly more capacity to migrate compared to control cells. Molecular analyses showed expression alterations in key gene candidates. We found that two of these molecules were bound by DOX with high propensity. Lastly, our 25% subline showed significantly increased CD44:CD24 ratio compared to the control cells, indicating DR and invasive nature of the subline.

The altered expression of multiple gene groups in our resistant subline that we examined significantly supports DOX resistance. The most notable was over 105-fold increase in the mRNA levels of drug metabolizing gene CYP1A1 whose product may support DOX resistance by binding to and inactivating the drug, thereby increasing the rate of cell survival.

ABCB1 and ABCC1, two representatives of ATPdependent drug transporter genes, were also significantly upregulated in our resistant subline as an indication of reduced intracellular DOX concentration. Mechanistically, the passive diffusion of the hydrophobic DOX molecules and their active efflux from the cell membrane via these transporters prevents nuclear accumulation and activity of DOX, 18 thereby inhibiting DOX-DNA adduct formation and its downstream events that transduce the DNA damage signal into apoptosis. 19 Expression of efflux transporters causes multidrug resistance (MDR) in various cancer types. P-gp encoded by ABCB1 is an important efflux pump that contributes to MDR development. 20 ABCC1 encodes Multidrug Resistance-associated Protein 1 (MRP1) which is involved in MDR by transporting many chemotherapeutic drugs out of cells. ABCC1 overexpression indicates a reduced response to chemotherapeutic drugs and a lower rate of survival. 21 DOX is a substrate for ABCC1, and ABCC1 has already been linked to DOX resistance in MDR cancer cells. 2223 Lu et al. report upregulation of ABCC1 expression at both mRNA and protein levels in their DOX-resistant subline. 20 They also show that over-expression of miR-134 significantly downregulates ABCC1.

The PI3K and AKT1 genes are important components of cell growth signaling pathways and showed overexpression in our resistant subline. While their upregulation supports DOX resistance, 16 inhibition of the components of this signaling pathway using molecular or pharmacological procedures can re-sensitize resistant cancer cells to the drug. 172425 We observed several folds increase in antiapoptotic BCL-2 levels and significant reduction in pro-apoptotic BAX levels in our 25% subline. Upregulation of BCL-2 induces DOX resistance, whereas its inhibition restores drug sensitivity. 26272829 Expansion of cancer cell interactions with its extracellular matrix strengthens DOX resistance in correlation with BCL-2 and BCL-XL upregulation. 2630 In our resistant subline, pro-EMT genes SNAIL, STAT3, MMP1 were significantly upregulated, in support of DOX resistance. 31 Downregulation or inactivation of SNAILl and STAT3 breaks cancer cell resistance, and when followed by DOX treatment, prevents xenograft tumor growth. 3233 In contrast, the 21-folds downregulation of the tumor suppressor gene CDH1 was notable. CDH1encodes E-cadherin which acts as an invasion suppressor. 34 E-cadherin is indeed a crucial cell-cell adhesion molecule that tightens the connection of epithelial cells together. Loss of Ecadherin expression induces EMT and metastasis by triggering the expression of pro-EMT transcription factors (TFs). E-cadherin levels change in EMT and MET and, therefore, the molecule is considered an important switch in EMT.

EMT is an important mechanism of DOX resistance in BC including ER + BC. ERα overexpression supports DOX action in E-cadherin downregulation towards EMT and DOX resistance. 35 Therefore, due to ERα presence in MCF-7 cells, frequent DOX treatment of our subline likely downregulated E-cadherin in favor of EMT and increased resistance. Cross-comparison between and ERα + and ERα -cell lines suggested that inhibition of ER activity enhances DOX responsiveness in resistant ERα + BC. 35 Both TOP2A and TP53 are involved in DNA processing. Over a 100-fold TOP2A downregulation in our resistant subline and a similar report by another study 11 indicates the importance of this enzyme as the prime target of DOX. Overexpression of mutated p53 (TP53) detected in our resistant cells, with modified protein conformation and impairment of p53 activity upon changes in its regulating proteins contributes to DOX resistance. 36 Our resistant subline expressed several folds higher CCND1 but slightly reduced p21 than did the WT MCF-7. DOX treatment enhances CCND1 expression, 37 whereas suppression of CCND1-encoded cyclin D1 by CDK4 inhibitor reverses DOX resistance and induces apoptosis in MCF-7 cells. 38 Cyclin D1 downregulation by knocking down mTOR/p70S6K in leukemia cells results in cell cycle arrest and apoptosis induction by DOX. 39 DOX activates the expression of cell cycle inhibitors including p21. 40 We presume that p21 downregulation in our cells could have occurred during the acquisition of DOX resistance in our resistant subline.

Our subline expressed several folds less tumor suppressor gene BRCA2 than did the parental MCF-7, similar to a previous report. 12 BRCA2 expression is induced by BRCA1 which supports the response to DOX-induced stress in prostate cancer. 39 Among the gene groups studied, two candidates displayed profound changes in their expression levels in our resistant subline: CYP1A1 with 106 folds upregulation, and TOP2A with 102 folds downregulation. We simulated DOX binding to either of these two molecules and found that the drug has a strong affinity to bind either molecule, regardless of its net suppressive or inductive effect. According to the literature, some of these molecular changes may have been gained during resistance acquisition and cell adaptation to pressures of drug treatment, whereas others must be considered the outcome of the resistance phenomenon. Given that intracellular pathways are heavily interconnected, it appears from the available reports and our own observations that the candidate genes extensively cross-talk within a network of interconnections, so that early events such as cell survival and growth are shaped by TFs and tumor suppressors before they promote the late events that include DR, EMT, and migration. 41 SHKN-MTFN co-treatment had important impacts on our DOX-resistant subline. The first change included reduced survival and induced apoptosis. When we increased doses of SHKN and MTFN, survival was diminished in all groups. More importantly, in co-treated groups exposed to DOX, no significant differences were detected in survival between WT cells and our resistant subline implying that DOX and SHKN+MTFN combination synergized to reduce cell survival to a minimum in both WT and resistant cells. Our co-staining experiments indicated that cell death was due to apoptosis and the SHKN-MTFN combination more significantly induced the apoptotic process than did single treatments. Reduced capacity to migrate was the second change. Treatment of cell groups with SHKN stopped cell migration to some extent and, unlike MTFN, the SHKN-treated cells did not show significant movements from 12 hrs to 24 hrs postscratch. More importantly, SHKN-MTFN cotreatment likely synergized with DOX so that cell migration almost stopped at higher doses of the compounds. The third observation was a reduced CD44/CD24 ratio. The data we generated using flow cytometry indicates that this ratio significantly rose from the drug-sensitive parental cells to our resistant subline. Treatment with SHKN or MTFN significantly reduced the ratio and this reduction was further accelerated in co-treatment samples. Our data indicate that, compared to the untreated resistant control, CD24 populations increased in treated and SHKN-MTFN-co-treated samples. Treatment of wildtype MCF-7 with either drug and particularly with both had significant reducing impact on the CD44/CD24 ratio, as we recently reported too. 12 Despite a significant increase in the ratio observed during resistance acquisition, single and dual treatments significantly reduced it. In particular, using 11 µM SHKN and 52 mM MTFN (determined as their IC50 when used in combination in our 25% subline), we co-treated our resistant subline populations with three varying doses of the two compounds from sublethal to lethal and achieved between 10 to 16 folds reduction in CD44/CD24 ratios.

The last change we monitored was in the energy of SHKN binding to growth signaling molecules. In our previous study, binding between SHKN, on the one hand, and a number of molecules that contribute to cell signaling pathways, on the other, was simulated and we found favorable binding energy in each case. 12 Table 5 lists these binding energies. SHKN affinity to these molecules is compatible with our observations on cell survival, migration and CSCs markers outlined above. In particular, we made observations similar to DOX, i.e., the energy of SHKN binding to CYP1A1 and TOP2A was found to be high, as it was in the case of DOX. As an antiresistance mechanism of action, we presumed that SHKN could compete with DOX in binding CYP1A1 towards its downregulation and to TOP2A toward upregulation of this molecule.

In conclusion, our current study was carried out to elucidate some of the uncovered cancer cell-drug interactions leading to the process of DOX resistance acquisition and subsequent changes. These findings can be consolidated and improved using stably-made resistant sublines and animal models of various BC subtypes each having distinct characteristics of receptor expression, invasiveness and metastasis. Dissection of such targeted DR models by genome editing and other versatile tools could unravel many mysteries behind DOX resistance. Mechanistic investigations on DR may include the role of chromatin accessibility alterations, TFs and noncoding RNA molecules, followed by downstream developments in gene expression and intracellular tumorigenic pathways. The findings of these steps will be important for efficient re-sensitization of resistant cells by specific DR-targeting measures such as combination of epigenetic therapy and SHKN-MTFN synergy.