Epigenetic-based combinatorial resveratrol and pterostilbene alters DNA damage response by affecting SIRT1 and DNMT enzyme expression, including SIRT1-dependent γ-H2AX and telomerase regulation in triple-negative breast cancer

Breast cancer is among the most common cancers of women in the United States, second only to skin cancer. After lung cancer, it is the second leading cause of cancer death in women. Currently, an estimate of 231,840 new cases of invasive breast cancer are expected to be diagnosed and about 40,290 breast cancer deaths in women are expected in the year 2015 [1]. Out of the total breast cancers, around 10–20 % of breast cancer are designated as triple-negative breast cancers (TNBC). They lack potential biomarkers such as endocrine receptors: estrogen and progesterone receptors (ER and PR, respectively) and HER2 protein. They are typically associated with a poor prognosis and are recalcitrant to conventional hormonal therapies, in contrast to triple-positive breast cancers and HER2-positive breast cancers. Epidemiological studies have indicated that the incidence rate of cancers is higher in affluent nations where the diet and the environment contribute to genomic instabilities [2, 3]. In many developing countries, such as those in Asia where fruits and vegetables are the staple diet, incidences of cancer are remarkably lower [4]. Cancer has been viewed as a set of diseases that are driven by progressive genetic abnormalities as well as epigenetic alterations [5].

Diet-or environment-induced aberrations of genes may develop over time but are often reversible, indicating that only gene expression has changed while the actual DNA sequence remains unchanged. Such reversible gene-based alterations are mediated by a process referred to as epigenetics. Epigenetic changes are believed to occur at a higher frequency than sequence-based genetic changes. The common epigenetic modifications that occur are DNA methylation, histone modifications such as acetylation, methylation and phosphorylation, and chromatin remodeling. DNA methylation and histone modification play important roles in the transcriptional regulation of genes involved in cell cycle progression, proliferation, apoptosis, and cell death. Alterations in the epigenetic-controlled expression of these genes may mediate carcinogenic processes as well. This paper will focus on these two key players in regulating the epigenetic machinery and will further address down-stream effects of SIRT1 (Type III HDAC) down-regulation [6‐8].

Recent studies have indicated that bioactive dietary components are strong epigenetic modulators that appear to play a part in prevention of breast cancer [8‐10]. A growing group of studies has indicated that nutritional factors play an important role in many human diseases and have been identified and evaluated for their activities against cancer [11, 12]. One such food source is berries, which are rich sources of a wide variety of antioxidant phytochemicals such as stilbenes. Stilbenes are natural phenolic compounds, consisting of resveratrol and pterostilbene. Resveratrol (3, 4′, 5 trihydroxystilbene) is a natural phytoalexin, synthesized by plants such as grapevines, berries and peanuts in response to an injury [12].

Both of these key polyphenols have been speculated to mimic caloric restriction at the molecular level [12] and are proposed to affect DNA methyltransferase (DNMT) enzyme activity [13, 14]. However, for the first time we are reporting the combinatorial effects of these two compounds on inhibiting the activity of SIRT1, a type III histone deacetylase (HDAC). HDACs play a major role in various cellular processes by preventing chromatin accessibility [11]. Recent studies have shown the importance of HDAC inhibition and its possible implication in cancer therapy [8, 10, 11, 15].

SIRT1 is a mammalian homologue of the yeast silent information regulatory Sir2, which requires nicotinamide adenine dinucleotide (NAD+) as a cofactor for its action. It is a class III HDAC, which is responsible for modifying histones as well as some non-histone proteins through deacetylation and thereby regulating cell growth, apoptosis, stress response, adaptation to calorie restriction, metabolism, cellular senescence and tumorigenesis [16‐18]. Recent studies have highlighted a unique feature of SIRT1 in regulating DNA damage and repair as well as its role in maintaining telomere length and genomic stability [11, 19, 20]. Moreover, previous studies have shown that SIRT1 deficiency impairs the formation of repair foci which can result in DNA damage, thereby protecting against tumors by inducting apoptosis [21, 22]. Interestingly, epigenetic patterns can be reversed, and epigenetic processes have in recent years become candidates for drug-mediated therapies.

In addition to the genetic changes caused by DNA damage, several tumors often contain epigenetically silenced protective genes with aberrant promoter region CpG island DNA hypermethylation [11]. Errors in double stranded break (DSB) repair can also cause mutations and chromosome instability that lead to cancer instead of cell death. SIRT1, when localized to the promoter of a gene, can also induce gene silencing by increasing the methylation of the promoter region of that gene by possibly engaging DNMTs to the damage site [21, 23]. Previous studies have shown involvement of SIRT1 in upregulating DNMT1 and DNMT3B presence, specifically at the DNA damage site [23]. Interestingly, SIRT1 has been found to localize at the promoters of these methylated-silenced tumor suppressor genes in some cancer cells, but not to promoters of the same tumor suppressor genes were they are normally in an active state [24].

In normal cells, SIRT1 has been linked to the role of protecting cells against potential carcinogenic agents and environmental stress. Whereas in malignant growth, it provides a strong stimulus which involves aberrant methylation as well as deacetylation of the promoter region of various protective genes and contributes to gene silencing and initiation and/or maintenance of cancer [18, 21, 23]. This could also be one of the potential mechanisms for overproduction of the telomerase enzyme, which enables cancer cells to replicate indefinitely. For the first time, we are reporting a relationship between SIRT1 and the human telomerase reverse transcriptase gene (hTERT) in human TNBC cells, thus opening a new area which requires further investigation. These findings suggest that DNA damage may directly contribute to the large number of epigenetically silenced genes in tumors due in part from hypermethylation [25‐27] and histone deacetylation [10, 15] across the damage region [28] . Human telomerase reverse transcriptase (hTERT) encodes the catalytic subunit of telomerase and is an epigenetically-regulated gene. hTERT is over-expressed in more than ~90 % of human cancers but not in normal somatic cells. In recent years, hTERT has emerged as a promising target for cancer therapeutics and is an important marker for the diagnosis of malignancy [10, 29]. We have found that combinatorial resveratrol and pterostilbene resulted in down-regulation of hTERT at both the gene and enzymatic activity level.

Most compounds used in chemotherapy are synthetic or analogs of compounds that are present in dietary food sources. The application of these compounds in treatment often requires high dosage and prolonged exposure for therapy-related effectiveness. Unfortunately, such treatments often result in problems such as non-specific tissue-cell cytotoxic effects and multi-drug resistance tumors. To avoid these issues, it is imperative that the chemicals used in the treatment should be safe, easily available and cost-effective. Most dietary epigenetic phytochemicals meet these criteria and can be administered as dietary supplements, thereby offering promising new options for the development of more effective chemopreventive and chemotherapeutic strategies [10]. Resveratrol and pterostilbene are two phytoalexins produced by plants in response to an infection. Ideally, they are the plants’ own defense system. Once the mechanistic pathways of phytochemicals are determined, modification of these compounds to increase their stability without compromising their therapeutic efficacy can be achieved [9]. The combination of two biomolecules can synergistically or additively improve the therapeutic impact on tumors at lower doses than when used alone.

In the present study, resveratrol or pterostilbene single treatment, as well as combination treatment, was administered to breast cancer and control breast cells. The concentrations we used in this study model are close to a normal physiologically achievable range [33, 34]. Depending on the expression of the receptors, breast cancers can be classified into various groups. Breast cancers which are positive for estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2) represent discrete biological entities with distinct clinical profiles and are often associated with better prognosis and can be treated with hormone therapy. In contrast, women with triple-negative breast cancers [TNBC (i.e., ER-, PR-, and HER2-)] are typically associated with less favorable prognosis [35, 36] and account for about 10–20 % of breast cancers. Breast cancers which are ER-, PR- and HER2-negative cannot be successfully treated with hormonal therapies or medications that work by blocking the receptor and are often fatal. We undertook this study in part to address this important challenge. TNBC used in this study were MDA-MB-157 and HCC1806 cells. MCF10A breast non-tumorigenic epithelial cells were used as a control cell line to determine the toxicity and efficacy of our combinatorial dietary regimen.

Cell viability analysis was performed for 24 h and 72 h using various concentrations of resveratrol and pterostilbene to determine time- as well as dose-dependency in all breast cancer cell lines. The combination effect was also analyzed on breast control (MCF10A) cells for 72 h to detect any evidence of toxicity. Resveratrol at 15 μM with pterostilbene at 5 μM after 72 h was found to be the most effective combination and was highly significant when compared with DMSO and 24 h of combination treatments. This combination in both the cell lines was found to possess synergism when compared with different combination doses after 72 h of treatments. Combination index (CI) values at 72 h of treatment for HCC1806 and MDA-MB-157 breast cancer cells were found to be the lowest with respect to other CI values (See Additional files 1 and 2). The CompuSyn software used to calculate synergy generates different values of combination index where CI <1, synergism; CI = 1, additive effects; CI >1, antagonism, and have been employed for drug combination and general dose effect analysis in various studies [30, 37]. Our analysis using CompuSyn software indicated that combination treatments lower than 15 μM resveratrol and 5 μM pterostilbene at 72 h showed less effectiveness and qualitatively displayed a degree of antagonism, addictiveness and/or moderate synergism in both types of breast cancer cells (see Additional files 1 and 2). Collectively, both resveratrol and pterostilbene were found to exhibit dose- and time-dependent inhibitory effects on TNBC cells. Interestingly, the same concentration was used to treat breast control MCF10A cells and did not result in any significant inhibition. To further confirm the effects over a long term, colony forming assays were performed in the tested cell lines (Table 1). Depending on the growth rate and colony forming potential of the cells, MDA-MB-157 breast cancer cells were analyzed for 10 days and MCF10A control cells were analyzed for 14 days. The combination of two polyphenols displayed a decreased in the colony forming potential of MDA-MB-157 cells with no significant decrease in MCF10A control cell. Ironically, HCC1806 breast cancer cells did not form successful colonies (≥50 cells). Cells were plated at an increased seeding density from 200 cells/plate to 1000 cells/plate from 7 days to 15 days, none of them formed successful colonies (data not shown). One explanation for this uncertainly could be due to critical density phenomenon in which cells won’t grow if they are too widely dissociated and makes too few contacts with other cells. The goal of this study was to determine the effects of resveratrol and pterostilbene combination treatment on both of these TNBC cell lines and to reveal the potential mechanism responsible for the effect. Our results indicated for the first time that resveratrol and pterostilbene can synergistically inhibit cellular viability in combination and can further impact cellular proliferation through cell cycle arrest and apoptosis induction in TNBC cells. Interestingly, the combination treatment in HCC1806 cells resulted in a tremendous apoptosis induction in comparison to DMSO and individual treatments. Moreover, treatment of MDA-MB-157 cells also resulted in apoptosis although the rate of apoptosis in comparison to HCC1806 cells was less but nevertheless significant.

Progress through each phase of the cell cycle is regulated carefully to avoid proliferation. Cells can be arrested in G1, S and G2/M phases to prevent any replicative errors. Cancer cells are characterized by unchecked cell cycles, often resulting in uncontrolled replicative errors and DNA damage. Previous studies have shown the role of p53 in arresting cells in G1 phase but not in S and/or G2/M phases [31]. Our experiments with resveratrol and pterostilbene combination treatment in HCC1806 and MDA-MB-157 cells resulted in predominant G2/M phase and S phase arrest, respectively. These observations have been found to be consistent with previous studies involving defective p53 status in various cancer cell lines [8, 31]. Resveratrol and pterostilbene combination at close to physiologically achievable concentration are shown in this investigation to inhibit cancer cell growth by inhibiting cell cycle progression and by the induction of apoptosis in these TNBC cells, with no significant effects in MCF10A control cells.

SIRT1, a class III histone deacetylase (HDAC) has been linked in normal cells to a role in protecting cells against potential carcinogenic agents and environmental stress. Whereas, in malignant growth it provides a strong stimulus, involving aberrant methylation and deacetylation of the promoter region of various genes and contributes to gene silencing, resulting in initiation and/or maintenance of cancer [17‐19, 25]. SIRT1 knockdown also results in cell cycle arrest and apoptosis induction. SIRT1 is one of the key epigenetic-modifying enzymes which, by itself or in complex, implements various cellular events such as DNA damage response (DDR), regulating stress, stabilizing the damage DNA along with γ-H2AX and DNMT’s and stabilizing eroded telomeres [11, 18, 21, 28, 38, 39]. For the first time, we have reported a significant down-regulation of SIRT1 with this combinatorial dietary approach. This down-regulation was confirmed at the transcriptional and translational as well as enzymatic levels. There was no significant change observed at the enzyme activity in MCF10A breast epithelial control cells, further suggesting the effectiveness and utility of this combinatorial approach.

DNA double-strand breaks are serious lesions which can lead to genomic instability and to cancer. A key marker and a way to monitor the DNA damage and repair mechanism is Phospho-H2AX [38]. H2AX is a member of the histone H2A family, one of the five histones that form chromatin along with eukaryotic DNA. SIRT1, along with γH2AX and ataxia telangiectasia mutated (ATM), is recruited to the DNA damage site and participates in the formation of repair foci [18, 21] and plays a role in stabilizing repair foci containing DNA repair factors. This mark is maintained at the break site until the break is repaired [18, 21]. Consistent with previous studies, cell lines with defective p53 show a longer persistence of γH2AX, as an indicator of stress [32]. Resveratrol and pterostilbene in combination showed a significant down-regulation of γH2AX (72 h) in both tested cancer cell lines with no effects on MCF10A control cells, further demonstrating the effectiveness of this dietary regimen in affecting DDR. This could also account for the increased apoptosis in both the cancer cell lines with this combination treatment. This γH2AX down-regulation in HCC1806 breast cancer cells was also confirmed using IF microscopy and interestingly was found to be consistent with our western blot data.

Recent studies demonstrated the recruitment of DNA methyltransferases (DNMTs) enzymes along with SIRT1 to the site of DNA damage as a part of the repair mechanism, which might be defective in cancer cells, resulting in hypermethylation of CpG-island promoters across the double stranded break [17, 23]. DNA methyltransferases and histone modifying enzymes are no longer considered independent epigenetic regulators, but are supposed to work in tandem or in cohesion to alter the epigenome. HDACs and DNMTs have been shown to assist with the DNA damage response by recruiting various other repair proteins to the site of damage. HDAC inhibitors such as butyrate and trichostatin A (TSA) have been shown to affect the repair machinery [18]. Double-stranded breaks can cause the development of cancer and also result in the induction of a hypermethylated state with the aid of silencing factor, SIRT1. Thus, in cancer cells we may observe increased expression of SIRT1, DNMTs and γ-H2AX proteins, as they are the key markers for genomic stress, DNA damage and repair foci formation [18, 23, 32, 38]. These two dietary polyphenols when administered (72 h) in HCC1806 breast cancer cells resulted in down-regulation of DNMT1, 3A and 3B at the gene level as well as at the enzymatic activity level in HCC1806 breast cancer cells. Interestingly, there was no significant change in overall DNMT enzyme activity in MCF10A control cells after 72 h of treatment. Thus, the observed molecular and cellular effects in this investigation could be due to a combined effect of this dietary regimen on SIRT1 and DNMT enzymes levels.

With every cell division eroded-critically short telomeres reveal double-strand breaks resulting in cellular senescence in normal mammalian cells [40], whereas under abnormal circumstances, they transform into cancer cells in lieu of senescence and/or apoptosis. The catalytic subunit of the enzyme telomerase (hTERT) is highly expressed in ~90 % of human cancers [10, 31]. Previous studies with normal human umbilical cord fibroblasts and stem cells have reported the role of SIRT1 in increasing the transcription of hTERT either directly by the involvement of FOXO3a and/or indirectly by affecting c-MYC [20, 39]. For the first time, we have reported down-regulation of hTERT with the use of resveratrol and pterostilbene in combination at close to physiological relevant doses in HCC1806 breast cancer cells. This down-regulation was evidenced by both real-time PCR and telomerase activity (TRAP) assay. Consistent with the previous studies, a time-dependent inhibition of telomerase enzyme was observed because of the stable half-life of telomerase enzyme [41, 42]. To further understand the mechanism of hTERT down-regulation, SIRT1 knockdown was performed in HCC1806 breast cancer cells. It was discovered that SIRT1 knockdown in human breast cancer cells resulted in a decrease in hTERT mRNA expression at the third day and enzyme activity at the fifth day of treatment as evidenced by RT-PCR and TRAP activity, respectively. One explanation for this hTERT down-regulation, which could account for growth inhibition in the breast cancer cells, could be due to a decrease in the downstream target of SIRT1, FOXO3a and c-MYC, which has been shown to increase hTERT expression in a SIRT1-dependent manner [20, 39]. Alternatively, many proteins involved in DDR also play a key role in telomere maintenance. One such protein is SIRT1, along with γ-H2AX and DNMTs, which are documented to maintain and stabilize the repair site and are significantly down-regulated in the present study [11, 28, 39]. Overall the effect of this dietary regimen was also analyzed on MCF10A control cells in order to determine the effectiveness of this approach. At the fifth day of the treatment there was a slight increase in telomerase activity in resveratrol (15 μM) and combination treatments when compared with the DMSO set, but this increase in activity was significantly below the positive control TRAP readings.

In the future, this dietary regimen may be used to provide a safe and effective treatment for TNBCs. It is documented that targeting γ-H2AX may also enhance the cytotoxic effects of irradiation therapy (IR), while circumventing adverse effects on unirradiated cells [43]. This down-regulation is also documented to sensitize cells to DNA damaging agent such as cisplatin [44]. Hence, combining DNA damaging agents (such as IR and cisplatin) with the current dietary regimen might provide a better treatment option. SIRT1, a key silencing factor known to be involved in a variety of cellular processes including DDR response and γ-H2AX expression, is shown in our study to participate in regulating hTERT expression, thereby opening future investigation which includes promoter-specific analysis of hTERT gene.