α‐Mangostin induces oxidative damage, mitochondrial dysfunction, and apoptosis in a triple‐negative breast cancer model

Triple‐negative breast cancer (TNBC) does not express estrogen receptor, progesterone receptor, and human epidermal growth factor receptor; therefore, TNBC lacks targeted therapy, and chemotherapy is the only available treatment for this illness but causes side effects. A putative strategy for the treatment of TNBC could be the use of the polyphenols such as α‐Mangostin (α‐M), which has shown anticancerogenic effects in different cancer models and can modulate the inflammatory and prooxidant state in several pathological models. The redox state, oxidative stress (OS), and oxidative damage are highly related to cancer development and its treatment. Thus, this study aimed to evaluate the effects of α‐M on redox state, mitochondrial metabolism, and apoptosis in 4T1 mammary carcinoma cells. We found that α‐M decreases both protein levels and enzymatic activity of catalase, and increases reactive oxygen species, oxidized proteins and glutathione disulfide, which demonstrates that α‐M induces oxidative damage. We also found that α‐M promotes mitochondrial dysfunction by abating basal respiration, the respiration ligated to oxidative phosphorylation (OXPHOS), and the rate control of whole 4T1 cells. Additionally, α‐M also decreases the levels of OXPHOS subunits of mitochondrial complexes I, II, III, and adenosine triphosphate synthase, the activity of mitochondrial complex I as well as the levels of peroxisome proliferator‐activated receptor‐gamma co‐activator 1α, showing a mitochondrial mass reduction. Then, oxidative damage and mitochondrial dysfunction induced by α‐M induce apoptosis of 4T1 cells, which is evidenced by B cell lymphoma 2 decrease and caspase 3 cleavage. Taken together, our results suggest that α‐M induces OS and mitochondrial dysfunction, resulting in 4T1 cell death through apoptotic mechanisms.


| INTRODUCTION
Breast cancer is the most common cancer worldwide in women. Triplenegative breast cancer (TNBC) is the type of breast cancer with the worst prognosis at all types of cancer (Ismail-Khan & Bui, 2010). Because TNBC does not express estrogen receptor (ER), progesterone receptor (PgR), or human epidermal growth factor receptor (HER), which are present in ER+ and HER2+ breast tumors, this type of breast cancer lacks targeted therapy. Due to the latter, chemotherapy is the only available therapy for TNBC, which could cause chemoresistance and side effects (Anders & La, 2008). A possible strategy for the treatment of TNBC is the use of the polyphenols such as α-Mangostin (α-M). α-M is a xanthone from Mangosteen (Garcinia mangostana L.) that possess antimicrobial, antiviral, and anticancer activity in vivo and in vitro models.
Concerning anticancer activity, α-M induces inhibition of cell proliferation and cell death via apoptosis in models of melanoma, leukemia, colon, lung and cervical carcinoma models (Rojas-Ochoa et al., 2021;Zhang et al., 2017). Interestingly, it has been shown that α-M did not exhibit deleterious effects on the kidney and liver tissues of normal murine models (Ibrahim et al., 2015) and did not affect the cell viability of normal human osteoblasts (Yang et al., 2021) or non-malignant cells such as human peripheral blood mononuclear and WI-38 cells when are treated with α-M (Zhang et al., 2018). These data suggest that α-M-mediated anticancer effects might be specific for malignant cells.
The redox state, oxidative stress (OS), and oxidative damage are very related to cancer development and its treatment. For instance, it is found that OS is present in the three stages of cancer: initiation, promotion, and progression. In the initiation stage, high levels of reactive oxygen species (ROS) induce deoxyribonucleic acid (DNA) damage and mutation; in promotion stage activate growth factors, up-regulating cell proliferation, meanwhile induces chromosome instability in the progression stage. However, it should be noted that high levels of ROS also induce mitochondrial dysfunction, which enhances additional ROS production, having positive feedback that increases OS (Kudryavtseva et al., 2016;Zorov et al., 2014).
Mitochondrial dysfunction is characterized by decreased mitochondrial respiration, which compromises cell bioenergetics leading to apoptotic cell death (Kudryavtseva et al., 2016). According to the latter, α-M might have the ability to induce ROS overproduction, producing OS and cell death in cancer cells; however, the role of α-M-induced apoptotic cell death associated with OS, oxidative damage and mitochondrial dysfunction in 4T1 breast cancer cells is poorly studied. Moreover, there are no reports of the anticancer effects of this polyphenol in vivo or in vitro models of TNBC associated with OS or mitochondrial dysfunction. Therefore, the aim of this study was to examine the redox state, mitochondrial metabolism, and proapoptotic effects of α-M in TNBC cell line 4T1.

| General and mitochondrial ROS evaluation
ROS production was measured using the fluorescent probe, MitoSOX,

| Antioxidant enzyme activities
The antioxidant enzyme activities were evaluated, as previously described by Cruz-Gregorio et al. (2018). Briefly, superoxide dismutase (SOD) activity was spectrophotometrically measured at 560 nm based on NBT reduction to formazan. The level of protein that inhibits NBT reduction to 50% was defined as one unit of SOD activity, including the activity of SOD1 and SOD2 isoforms. CAT activity was assayed at 240 nm by a method based on the decomposition of H 2 O 2 .
Enzymatic activity was expressed as k/mg protein, as previously described by Aebi (1984). Glutathione peroxidase (GPx) activity was evaluated by the disappearance of NADPH at 340 nm in a coupled reaction with GR, as previously reported by Lawrence and Burk (1976). GR activity was evaluated by measuring the disappearance of NADPH at 340 nm in presence of GSSG, according to Carlberg and Mannervik (1975). Glutathione S-transferase (GST) activity was determined, as previously reported by Habig and Jakoby (1981), by the formation of 1-chloro-2,4-dinitrobenzene-GSH complex and was monitored spectrophotometrically at 340 nm.

| Oxidized proteins
Protein carbonyl content was determined by the method of Reznick and Packer (1994). Cell lysates, obtained from cells seeded onto p100 plates at a density of 1.8 million cells/well and incubated in a medium containing 80 μM α-M for 24 h, were sonicated and after incubated with 2.5% streptomycin sulfate overnight to remove nucleic acids. Cell lysates were treated with 10 mM DNPH (in 2.5 M HCl), and the protein pellet was dissolved in 0.5 mL of 6 M guanidine hydrochloride. Assessment of carbonyl formation was done based on the formation of protein hydrazone by reaction with DNPH. The absorbance was measured at 370 nm. Protein carbonyl content was expressed as nmol of carbonyl/mg protein.

| Glutathione quantification (GSH/GSSG ratio)
Total glutathione (GSH + GSSG) was evaluated by the enzymatic recycling method described by Rahman et al. (2006), in which GSH is oxidized by DTNB to 5-thio-2-nitrobenzoic acid (TNB, detectable at λ = 412 nm) and glutathione-TNB adducts (GSH-TNB). Both GSH-TNB and GSSG are reduced by GR in the presence of NADPH, to GSH, which in turns is oxidized by DTNB to TNB. In this manner, the amount of total glutathione calculated in this first step represents the sum of GSH and GSSG.
Next, GSSG was determined by the enzymatic recycling method mentioned above, where samples were previously treated with 2-VP. 2-VP, that can covalently associate with GSH, to remove GSH, leaving GSSG as the only measurable substrate of the assay. Finally, GSH was calculated by subtracting GSSG from total glutathione (GSH + GSSG).
Briefly, the 4T1 cell extracts were diluted with 120 μL of 0.1 M K 2 HPO 4 , 5 mM disodium EDTA buffer, pH 7.5. Then, two separate samples of 20 μL each were used to measure either GSH + GSSG or GSSG (these samples were previously treated with 2-VP), mixed with 2.5 mM DTNB and 250 U/mL GR. Finally, β-NADPH was added and the absorbance at λ = 412 nm was measured at intervals of 60 s, for 2 min. The rate of change in absorbance for each experiment was compared with GSH or GSSG standards.

| Statistical analysis
All the experiments were done in triplicate, and the data were analyzed as the mean ± standard error of the mean (SEM). One-way analysis of variance, Tukey and t-student test were used to determine the statistical significance of the experimental condition versus the control (DMSO). Experiments were not randomized/blinded.  We also measured CAT protein levels to discard that its activity reduction was due to the decrease in CAT protein levels. As observed in Figure 4F,G, CAT protein levels are also reduced, resulting in the reduction of CAT function and proteins levels due to 18 μM αM.
3.5 | α-M increases protein carbonyl content and GSSG, decreasing GSH/GSSG ratio, inducing oxidative damage High levels of ROS could induce OS, which is evidenced by changes in markers such as GSH, GSSG, and protein carbonyls. We measured GSH total, GSH, and GSSG. We also calculated the ratio GSH/GSSG, and we estimated the levels of oxidized proteins. We did not observe significant changes in GSH + GSSG and GSH during α-M treatment with respect to the DMSO ( Figure 5A,B). However, we found that α-M increases GSSG content 2-fold, and consequently, the GSH/GSSG ratio also decreased 2-fold ( Figure 5C,D). We also observed that α-M significantly increases 2-fold protein carbonyl content ( Figure 5E). Therefore, the above results strengthen the evidence that α-M induces OS and oxidative damage in 4T1 cells.

| α-M decreases basal respiration, P respiration, and respiratory control (RC) in 4T1 cells
It has been reported that OS is one of the principal factor that induces mitochondrial bioenergetics dysfunction, inducing mitochondrial decoupling (Guo et al., 2013). Thus, we investigated if OS induced by α-M reduced basal cell respiration, P, leak, uncoupled and RC respiration parameters in 4T1 cells. Figure

F I G U R E 3 α-Mangostin (α-M) increases reactive oxygen species (ROS) in whole 4T1 cells. (a) Representative images and (b) quantitative data of ROS generated by α-M treatment.
Quantitative data were obtained from 4T1 cells compared with control (DMSO). The mean intensity of DHE (ethidium, Et) and dichlorofluorescein (DCF) fluorescence was measured using Gen5™ 3.0 software for image acquisition and quantification. The fluorescence intensity is expressed as the mean ± SEM. tstudent test, ***p versus control (DMSO), n = 3 per group. DMSO, dimethyl sulfoxide.
control (DMSO) treatment. In fact, we found that α-M reduced basal cellular respiration by 2-fold, P by 2.7-fold, and RC by 2.6-fold ( Figure 6B,C,F). We also found a tendency in the increase of leak parameter ( Figure 5D) and a decrease of uncoupled parameter ( Figure 5E). Taken together, our results demonstrated that α-M reduces mitochondrial basal, OXPHOS and RC, inducing mitochondrial dysfunction, which could be related to OS.

| α-M decreases OXPHOS protein levels
Since respiration is associated with a decrease in the OXPHOS complex capacity, we measured the levels of OXPHOS complexes by immunoblot, finding that α-M decreases the protein expression of ATP synthase by 9-fold, that of CIII by 3-fold, that of CII by 3-fold and that of CI by 9-fold compared to the control ( Figure 7A,B,C,E,F). We did not find changes in the protein levels of CIV ( Figure 7A,D). Thus, our results show that α-M reduces OXPHOS protein levels.

| α-M decreases CI-linked respiration in 4T1 cells
Also, we wondered if OXPHOS capacity reduction was linked to reduced complex activity. We found that the activity of CI-linked respiration is reduced nearly 2-fold when 4T1 cells are treated with α-M in comparison with the control (DMSO, Figures 8A,B). In contrast, CIV-associated respiration was unchanged by α-M ( Figure 8C,D).
Thus, these results show that α-M reduces the activity of CI-linked respiration.

| α-M decreases PGC1α
Since OXPHOS protein level depend on mitochondrial biogenesis, we wanted to measure the levels of proliferator-activated receptorgamma coactivator 1α (PGC1α), a marker transcription factor that induces mitochondrial biogenesis. We found that α-M downregulates 4-fold the expression of PGC1α ( Figure 9A,B). Thus, our results which functional group of α-M is the one that induces ROS overexpression has not been studied, which deserves to be studied and therefore has certainty about the mechanism by which α-M produces cytotoxicity by overproduction of ROS, something contrary to the first discovered mechanism of α-M, that is, an antioxidant effect. One of the limitations of our study is lack of positive control.
The aim of this study was to determine the cytotoxic effect of We wanted to study if ROS overproduction is due to an antioxidant activity decrease. Indeed, we found that CAT antioxidant activity decreases, which correlated with ROS overproduction. CAT is a key enzyme and very efficient in removing H 2 O 2 produced during cell metabolism and thus would prevent ROS formation such as hydroxyl radicals, a highly reactive ROS produced by Fenton or Haber-Weiss reactions (Valko et al., 2007). Moreover, we found in 4T1 cells an increase in oxidative damage markers such as GSSG and oxidized proteins during α-M treatment. Our results agree with those obtained by Rojas-Ochoa et al. (2021), who demonstrated that α-M increases GSSG and oxidized proteins in a medulloblastoma cell model. Since OS could affect mitochondrial function, we studied the effect of α-M on cellular respiration and mitochondrial metabolism. We found that α-M decreases basal respiration, P, and RC, which correlates with reduced levels of ATP synthase, CI, CII, and CIII. Furthermore, we found that the activity of CI-linked respiration was also decreased along with the expression of PGC1α, suggesting a reduction of mitochondrial biogenesis. To our knowledge, this is the first work showing that α-M decreases basal respiration, P, RC, OXPHOS protein levels, and mitochondrial biogenesis in 4T1 cancer cells; however, it is important mentioning that α-M has already been reported to induce mitochondrial membrane depolarization, leading to apoptotic cell death in PC12 cells (Sato et al., 2004). Moreover, it has been shown that α-M dramatically reduced mitochondrial respiration and inhibited the activ- as decreased Bcl2 levels associated with mitochondrial dysfunction (Shimizu et al., 1998), provide information that α-M-induced apoptosis in 4T1 cells is via mitochondrial apoptosis (or intrinsic pathway).
Moreover, our results agree with the results of Lee et al., (2017) Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a loading control. Data are expressed as the mean ± SEM. t-student test, n = 3. DMSO, dimethyl sulfoxide was used as a control.
F I G U R E 9 α-Mangostin (α-M) decreases the protein levels of peroxisome proliferator-activated receptor-gamma co-activator (PGC1α). (a) Representatives immunoblot and (b) densitometric analysis of PGC1α in 4T1 cells treated with DMSO and α-M. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a loading control. Data are expressed as the mean ± SEM. t-student test, n = 3. DMSO = dimethyl sulfoxide was used as a control.

| CONCLUSION AND CONCLUDING REMARKS
TNBC differs from other types of invasive breast cancer in that it grows and spreads faster and has limited treatment options. This has led to the development of new and more effective therapies. Among these therapies is chemotherapy. Unfortunately, these treatments can cause side effects and chemoresistance. One of the potential strategies that could work without side effects in the treatment of TNBC is the use of the polyphenol α-M. Our results show that α-M induces OS, and mitochondrial bioenergetics dysfunction, which is related to apoptosis cell death of a TNBC cell model. Therefore, our data support the potential use of α-M as an anticancer molecule in the treatment of TNBC, information that reinforces and encourages the studies related to the anticancer activities of α-M in in vivo models or even in clinical trials.