Background
Colorectal cancer is the third most common cancer among both men and women in the United States. It is also the second most common cause of cancer-related deaths in men and women combined [
1]. With regular screening, colon cancer can be detected early, when treatment is most effective; however, in the majority of cases colon cancer is detected late. Over 95 % of colon cancer cases are considered sporadic thus placing environmental factors as the major cause [
1]. The most important environmental factors among them are diet and lifestyle. The highest incidence rates of colon cancer are in developed nations including the U.S. [
2]. Diets rich in refined starch, sugar, and saturated and trans-fatty acids but poor in fruits, vegetables and whole grains (prevalent in developed nations), have been shown to be closely associated with an increased risk of colon cancer [
2‐
4]. A meta-analysis of case–control studies suggests that fruit consumption was associated with a 13 % decrease in colon cancer risk. The benefits from consuming a diet rich in fruits and vegetables could be attributed to the plethora of bioactive compounds present in them [
5].
Grapes are consumed around the world and are a rich source of many bioactive compounds. Red grapes are rich in resveratrol (RSV), a stilbene that has shown anti-cancer properties in a variety of models, including human studies [
6]. We previously reported that RSV suppressed proliferation and induced apoptosis via p53 activation in HT-29 and SW-480 human colon cancer cell lines, however, it was effective only at higher concentrations (75–100 μM) [
7]. Grape seed extract (GSE) is a popular dietary supplement rich in proanthocyanidins and has been reported to have anti-colon cancer properties in a variety of in vitro and in vivo models [
8]. As bioactive compounds exist in a complex mixture in fruits and vegetables, laboratory assessment of their biological activity in combination is more relevant to human exposure. In addition, because these compounds have pleiotropic effects, there is the potential that they will exert additive or synergistic chemopreventive actions. A recent study that compared GSE induced anti-cancer effects to the effects of its individual components found that GSE was more potent in growth inhibition compared to its individual constituents epigallocatechin, procyanidins and their association [
9]. Our previous studies also support this notion, as we demonstrated using a well-established combination index method that a RSV (~25 μM) and GSE (35–50 μg/ml) mixture was potent in suppressing proliferation and elevating apoptosis in the HCT-116 human colon cancer cell line at lower concentrations compared to RSV or GSE alone [
10,
11]. Combination index methods is based on the classic isobologram equation CI = D1/d1 + D2/d2. D1 and D2 are the doses of RSV and GSE respectively in the combination system where as d1 and d2 are the doses of RSV and GSE alone for the same fractional inhibition, respectively. In addition, RSV potentiated GSE-induced p53-dependent apoptosis via mitochondrial apoptotic signaling, and demonstrated specificity to cancer cells, as it was non-toxic in the normal colonic epithelial cell line CRL-1831 [
10]. Our preliminary results led us to believe that a combinatorial approach towards colon cancer chemoprevention using bioactive compounds is a feasible strategy.
Historically, colon tumorigenesis has been viewed as a stochastic model where wide populations of abnormal colonocytes have an equal propensity to initiate tumor growth [
12]. However, the cancer stem cell (CSC) theory suggests that most, if not all, cancerous tumors are driven by CSCs, probably through dysregulation of self-renewal pathways [
13]. CSCs are capable of self-renewal, cellular differentiation, and maintain their stem cell-like characteristics even after invasion and metastasis [
14]. Furthermore, they are resistant to standard therapies and thus are thought to be responsible for cancer relapse. The Wnt/β-catenin signaling pathway plays a critical role in maintenance of stemness, and survival/proliferation of CSCs [
15], and as such, targeting the Wnt/β-catenin signaling is a good strategy for cancer prevention. Aberrant Wnt signaling in colon cancer is typically followed by mutation in the K-ras gene and loss of the tumor suppressor p53. It is estimated that p53 is abnormal in 50 to 75 % of colorectal cancer cases, and that this change marks the transition from noninvasive to invasive disease [
16,
17]. Thus, treatments, which act independent of p53 status, are desirable.
Sulindac, a nonsteroidal anti-inflammatory drug has shown promising results in treatment of colon cancer. Recently it was shown to induce SMAC (a mitochondrial apoptogenic protein)-dependent apoptosis in cells with nuclear β-catenin (an indicator of colon CSCs) and decrease polyp numbers in the
APC
Min/+
mouse that is highly susceptible to spontaneous intestinal adenoma formation [
18]. Chemotherapeutic drugs like sulindac are effective against certain types of cancers but can also have unexpected adverse effects such as gastrointestinal bleeding, hepatotoxicity [
19,
20] and in some cases chronic inflammation promoted colon cancer [
21]. This has stimulated active pursuit of new approaches and/or combination strategies for cancer chemoprevention. Based on our preliminary data, we hypothesized that the combination of RSV and GSE suppresses proliferation and induces apoptosis in colon CSCs. Azoxymethane (AOM), a DNA alkylating agent, induced mouse colon cancer model is a well-established and reproducible model of sporadic colon carcinogenesis to predict chemopreventive efficacy [
22]. Intraperitoneal injection of AOM in A/J mice, a breed that is susceptible to chemically-induced carcinogenesis, for six weeks resulted in tumor formation within six weeks of the last injection [
23]. Thus, we determined the efficacy of the RSV-GSE combination using A/J mice with six week AOM injection regimen and compared its effects to those of sulindac. Furthermore, we examined the possible molecular mechanisms that underlie the anti-cancer activity of RSV-GSE (and compared to sulindac) using colon CSCs, positive for CD 44, CD 133 and ALDH1b1 markers, isolated from primary human colon cancer tumors.
Methods
Chemicals
Grape seed extract (GSE, ORAC value 9000–13,000 μmole Trolox equivalents/g, total phenolic content > 85 % gallic acid equivalents) was a generous gift from San Joaquin Valley Concentrates (Fresno, CA). We had previously characterized the GSE used in this study using UPLC-MS and we detected presence of (+)-catechin and (−)-epicatechin monomers and their oligomers, and their gallate derivatives similar to other published papers [
24,
25]. The GSE used in this study lacks resveratrol (RSV) as described earlier [
10]. BrdU Cell Proliferation Assay Kit was obtained from Cell Signaling Technology (Danvers, MA). Antibodies for PARP and cleaved PARP, p53, pGSK3β, Bax, Bcl-2, β-actin, β-catenin, cyclin D1, c-Myc, COX-2 and topoisomerase-2β (Topo ii b) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Cytochrome C was obtained from Cell Signaling Technology (Beverly, MA). All other chemicals including RSV were obtained from Sigma (St. Louis, MO).
Animal study
A/J male mice (six weeks old; n = 13 per group) purchased from the Jackson Laboratories (Bar Harbor, ME) were housed in stainless steel wire cages (three or four per cage) with a 12 h light/dark cycle. Mice were allowed access to laboratory rodent chow and water ad libitum. After two weeks of acclimatization, all mice were randomly assigned to four groups and fed AIN-93G diets obtained from Harlan Laboratories (Indianapolis, IN).
Azoxymethane carcinogen injection
All mice except saline controls received six weekly subcutaneous injections of azoxymethane (AOM, Sigma) in saline for colon carcinogenesis at 5 mg/kg starting at eight weeks of age.
Experimental diets
At 16 weeks of age i.e., two weeks following the last AOM injection, the animals were assigned to the following diets – AIN-93G control, AIN-93G supplemented with RSV-GSE (0.03 and 0.12 % w/w, respectively) or AIN-93G supplemented with sulindac (0.06 % w/w). RSV and GSE concentrations were chosen based on the earlier human (
n = 32) study that showed a decrease in serum oxidative stress markers in obese subjects orally supplemented with RSV and GSE separately [
26]. Sulindac concentration was chosen based on previous clinical trial study in humans (
n = 12) with familial adenomatous polyposis where administration of sulindac resulted in significant reduction of polyp number [
27]. The saline control animals received AIN-93G control diets. All animals had free access to food and water.
Colon tissue collection
After one week of dietary intervention, five animals from each group were euthanized using isoflurane. The remaining animals (n = 8/group) were euthanized after four weeks of dietary intervention. The colon was resected and washed with RNAse free PBS and observed under a dissection microscope for counting tumors. Tumor number was recorded for each animal. At the end of the study the tumor number was averaged for each treatment group and represented as means ± S.D. For immunohistochemistry and immunofluorescence analysis, about 1 cm of the colon tissue was collected and fixed with 10 % buffered formalin. Specimens were then paraffin-embedded and orthogonally sectioned. The tissue was sectioned at four microns thickness and mounted on positively charged slides.
Immunofluorescence staining
Pre-treatment of slides
Prior to staining, the paraffin was softened and the tissue specimens fixed additionally by baking the slides in an oven at 55 °C for 20 min. Deparaffinization was performed with Fisherbrand (Pittsburg, PA) clearing agent citrisolv twice for five minutes and hydrated with decreasing concentrations of ethanol (100-100-95-70 v/v). For target retrieval, the slides were incubated in citrate buffer at pH 6 (9 mM citrate, 1 mM citric acid) at 95 °C for 20 min. To quench auto fluorescence from formalin residues, slides were pretreated with sodium borohydride (1 mg/mL) for five minutes. Mouse sections were blocked with mouse IgG serum from the M.O.M kit and avidin/biotin obtained from Vector Labs (Burlingame, CA) as per the manufacturer’s protocol.
β-catenin staining
β-catenin staining was performed at 4 °C overnight using a Abcam rabbit anti-β-catenin antibody (Cambridge, MA). Biotinylated secondary antibody in combination with streptavidin fluorescein (Vector Labs) was used for visualization. Mounting media with DAPI (Vector Labs) was used as a counterstain. All images were taken in Olympus BX-63 microscope with the help of Cell Sens software from Olympus America (Center Valley, PA).
TUNEL staining
TUNEL staining was performed using a cell death detection kit from Roche Diagnostics (Indianapolis, IN) according to the manufacturer’s protocol for formalin fixed paraffin embedded tissues.
Cancer stem cells
Isolated human colon CSCs positive for cancer stem cell markers CD133, CD44, CD34, aldehyde dehydrogenase, telomerase, Sox2, cKit, and Lin28, were obtained from Celprogen Inc. (San Pedro, CA). To maintain the cells in their undifferentiated state, colon CSCs maintenance media and specially coated cell culture flasks obtained from Celprogen were used. Cells were maintained in incubation at 37 °C and 5 % CO2. Cell cultures at approximately 80 % confluence were used for all in vitro experimental procedures. For all experiments low passage number (less than 10) cells were used (not more than three weeks after resuscitation). The authentication information for the cell line obtained from Celprogen is available under supplemental information.
Lentiviral particles encoding shRNA targeting p53 obtained from Santa Cruz Biotechnology were used to attenuate p53 expression in colon CSCs as described earlier [
28].
Cell proliferation
Cell viability was assessed by BrdU (5-bromo-2′-deoxyuridine) assay kit from Cell Signaling Technology (Danvers, MA). Briefly, cells were plated at a density of 1 × 10
5 per well in 12-well plates. Media was replaced after 24 h with serum-free colon CSCs media (Celprogen) and dosed with RSV-GSE and/or sulindac. For all in vitro experiments sulindac sulfide, the active form of sulindac was used. Preliminary experiments revealed that lower concentrations of RSV-GSE were potent in suppressing proliferation of colon CSCs compared to the concentrations used in our earlier study using HCT-116 early colon cancer cells [
10]. Interestingly, other researchers also showed that dietary bioactive compounds are more potent against highly proliferating or advanced cancer cells that are distinctly different from normal cells [
29]. Hence, for this study doses of RSV were kept constant at 9 μM, whereas GSE doses in the combination varied (6.25, 12.5 and 25 μg/mL). Sulindac was dosed at 6.25, 12.5 and 25 μg/mL. After 24 h, BrdU incorporation was assayed as described in manufacturer’s protocol. The experiment was carried out in triplicate, and results were expressed as the means ± S.E.
TUNEL assay
Apoptosis was quantified by using fluorescein labeled nucleotide and terminal deoxynucleotidyl transferase (TdT) to identify DNA fragmentation (characteristic of apoptosis). Briefly, cells (9 × 104) were seeded in four-chambered glass slides, and after treatment with RSV-GSE or sulindac for 12 h, the in situ cell death detection kit from Roche Diagnostics was used for quantifying apoptosis based on the manufacturer protocol. The experiment was carried out in triplicate, and results were expressed as means ± S.E.
Briefly, colon CSCs (10,000 cells per well) were cultured in stem cell specific serum free media in an ultra-low attachment six-well plates. The cells were maintained in similar conditions as mentioned earlier under the cancer stem cells section. RSV-GSE or sulindac was added six hours after the cells were added to the six-well plates. At the end of ten days, the number of spheres was assayed using a phase contrast microscope.
Western blot
Cells were plated in six-well plates at a concentration of 3.0 × 10
5 cells per well in colon CSCs media. After 24 h, cells were transferred to serum free medium for 18 h. Protein was extracted according to our previously published protocols [
10,
30]. The blots were incubated with primary antibodies overnight at 4 °C at a dilution of 1:500. Subsequently, the blots were incubated with secondary antibodies for two hours at room temperature at a dilution of 1:10,000. Blots were imaged and quantified using the Odyssey infrared imaging system and software (Lincoln, NE) and normalized to β-actin, a loading control for cytoplasmic proteins and topoisomerase-2β as a loading control for nuclear proteins. Each treatment was carried out in triplicate, and results were expressed as means ± S.E.
Statistical analysis
Data are expressed as means ± S.E. for all the data. Significance was determined by one-way ANOVA with post hoc Tukey analysis for in vitro data (SPSS v21, IBM, Armonk, NY). For animal studies, analysis of data was done using mixed procedure in SAS v9.4 software (Cary, NC). The p values < 0.05 were considered statistically significant.
Discussion
The objective of present study was to evaluate the anti-cancer efficacy of RSV-GSE in a mouse model of colon cancer, and to determine the mechanisms of action using human colon CSCs in vitro. Our results present the first evidence of in vivo anti-colon cancer efficacy of a combination of the grape bioactive components RSV and GSE in the mouse model with AOM-induced colon carcinogenesis. Our data in vitro in colon CSCs demonstrate suppression of nuclear translocation of β-catenin (Wnt/β-catenin signaling pathway) and induction of mitochondrial-mediated apoptosis.
Our results indicate that RSV-GSE, bioactive components from grapes, suppress tumor incidence in a mouse model with AOM-induced colon carcinogenesis (Fig.
1a). Furthermore, RSV-GSE consumption had reduced toxicity compared to sulindac, suggesting specific targeting of cancer cells (Fig.
1b). Indeed, clinical trials in humans have shown that RSV is quite safe [
31], similar results have been observed for GSE [
32].
Accumulated experimental evidence has suggested that most cancers, including colon cancer, have a hierarchal organization regulated by a small number of self-renewing cancer cells, called CSCs [
33]. CSCs including colon CSCs have shown to be resistant to conventional chemotherapeutic regimens that target homogeneous populations of rapidly proliferating differentiated tumor cells. For e.g., CD133-positive colon CSCs were shown to be resistant to the conventional cytotoxic drug 5-florouracil and the resistance was shown to be dependent on Wnt signaling [
34]. The proliferation and the acquisition of the stem cell fates is coordinated by a small number of highly evolutionarily conserved signaling pathways, including the Wnt/β-catenin signaling pathway, which is commonly deregulated in most colon cancers [
35]. Nuclear accumulation of β-catenin is implicated in the transformation of stem cells to oncogenic stem cells in the colon [
15]. Although, it has been observed that nuclear β-catenin accumulation is also seen in normal colonic stem cells and progenitor cells which are located at the bottom proliferative compartment of the intestinal crypts [
36,
37], a recent study has shown that it is observed in less than 0.01 % of crypts in wild-type mice [
18]. Hence, we [
38] and others [
18] considered increased number of crypts with colonic stem cells with nuclear β-catenin accumulation (Additional file
1: Figure S1) as a hallmark of colon carcinogenesis and a signature feature of elevated oncogenic stem cells. Qiu et al. reported that one week of sulindac treatment resulted in a 75 % reduction in the number of crypts containing cells with nuclear β-catenin. Most importantly, a vast majority (98 %) of identifiable stem cells with accumulated nuclear β-catenin in sulindac-treated
APCMin/+ mice were TUNEL-positive at this time point [
18]. In the current study, where AOM (a well-known colon specific carcinogen) was used to induce colon carcinogenesis, RSV-GSE consuming animals had 62 % reduction in number of crypts containing cells that have accumulated nuclear β-catenin (Fig.
2a). Additionally, our data suggest that this could be due to induction of apoptosis (Fig.
2b). Efficacy of RSV-GSE was comparable or better than sulindac. The in vivo data was supported by our in vitro observations where we noticed that RSV-GSE at physiologically relevant doses suppressed proliferation and induced apoptosis as well as suppressed sphere formation in colon CSCs (Fig.
3a–d).
There is evidence that nuclear accumulation of β-catenin results in accelerated tumor cell proliferation and tumor progression through the transcriptional activation of target genes including c-Myc, cyclin D1 and COX-2 [
39]. Mechanistic data in vitro confirmed our in vivo observations as RSV-GSE suppressed nuclear β-catenin accumulation in colon CSCs (Fig.
4a). RSV-GSE also suppressed cytoplasmic levels of pGSK3β (Fig.
4b) shown to induce nuclear β-catenin translocation and down-regulated nuclear levels of proteins downstream of Wnt/β-catenin pathway, c-Myc and cyclin D1 (Fig.
4c, d). c-Myc and cyclin D1 are the key signatory genes of Wnt signaling and both function in the stimulation of cell proliferation and in preventing apoptosis. Coordination of c-Myc with cyclin D1 or its upstream activators not only accelerates tumor formation, but also may drive tumor progression to a more aggressive phenotype [
40]. Although, previously published research has shown that RSV and GSE suppressed nuclear β-catenin translocation, to our knowledge, this is the first study to show such an effect in colon CSCs.
Alterations in Wnt/β-catenin signaling might also explain why RSV-GSE also suppressed sphere formation ability in vitro (Fig.
3d). Because c-Myc and cyclin D1 also play a role in stemness [
41], our data showing suppressed c-Myc and cyclin D1 only by RSV-GSE treatment might explain its higher potency compared to sulindac. Our data is in line with recent research showing that dietary compounds including grape seed extract, curcumin, lycopene and resveratrol are promising chemopreventive agents against various types of cancers owing to their direct and indirect effects on CSC self-renewal pathways, such as Wnt/β-catenin signaling pathway [
42‐
45].
P53 plays a critical role in tumor suppression by inducing growth arrest, apoptosis, and senescence, as well as by blocking angiogenesis. Consistent with the role of p53 as a cell stress-associated transcription factor [
46,
47], we observed increased expression of p53 (Fig.
4e) and p53-responsive Bax (and Bax/Bcl-2 ratio) (Fig.
4g) in colon CSCs with RSV-GSE treatment. This indicates RSV-GSE induced intrinsic apoptotic signaling pathway by Bax-induced increased permeation of mitochondrial membrane, resulting in release of cytochrome C and activation of caspases. Whether similar pathway of apoptosis is activated in p53 knockout cells remains to be seen, although cytochrome C was elevated by RSV-GSE treatment in colon CSCs with shRNA attenuated p53. Mutational inactivation of p53 is one of the most frequent events found in over 50–75 % of colon cancer cases, and marks transition to metastasis [
17,
48‐
50]. Our results showing that RSV-GSE exerts its biological efficacy, both anti-proliferative and pro-apoptotic, in colon CSCs independent of their p53 status (Fig.
5a–e) confers an advantage to the use of RSV-GSE for primary and secondary chemoprevention.
Preclinical and clinical studies suggest that COX-2 is involved in chronic inflammation and its activation may be involved in inflammation-mediated stem cell proliferation/differentiation [
51]. Our data suggests that RSV-GSE was more effective compared to sulindac in suppressing nuclear COX-2 levels (Fig.
6a, b) in colon CSCs and colon CSCs with shRNA attenuated p53 and might further explain higher potency of RSV-GSE combination compared to sulindac. Further, NSAIDs like sulindac can suppress both COX-1 and COX-2 thereby deplete prostaglandin in tissues, which mediate mucosal bicarbonate production, mucus secretion, and maintenance of blood flow [
52] and thus mucosal healing [
53]. This explains the increased gastrointestinal toxicity (stomach ulcers and loss of adipose tissue deposits) in mice fed with sulindac. Unlike, sulindac, RSV and GSE (proanthocyanidins) have minimal effects on COX-1/PGE2 [
54] and thus explains the lack of stomach ulcers and adipose tissue loss. Thus, in the future studies, it is critical to explore whether sulindac eliminates normal colon stem cells along with colon cancer stem cells whereas RSV-GSE is selective against only colon cancer stem cells. This aspect could not be assessed in the current study, as we did not include normal control animals consuming RSV-GSE or sulindac.
Acknowledgments
We would like to thank members of Laboratory Animal Resources at Colorado State University for their assistance with mice study. We would like to thank San Joaquin Valley Concentrates (Fresno, CA) for supplying GSE powder. We would also like to thank Laura Markham for her assistance in collecting colon tissues and Abigail Sido for her help with editing the manuscript.