Novel Evidence for Curcumin and Boswellic Acid–Induced Chemoprevention through Regulation of miR-34a and miR-27a in Colorectal Cancer

Colorectal cancer is one of the leading causes of cancer-related mortality in the United States (1). There is growing evidence suggesting that dietary and lifestyle modifications can influence risk prevention for colorectal cancer (2). Increasing focus on chemopreventive properties of natural compounds from traditional medicine has led to discovery of potential therapeutic agents targeting colorectal cancer and other cancers. Curcumin, the principle curcuminoid and a derivative of the spice turmeric curcuma longa, is used as a naturally occurring medicine to treat variety of inflammatory disorders, and human cancers (3). In terms of its well-established antitumorigenic properties, curcumin has been shown to interfere with various stages of colorectal tumorigenesis, including tumor initiation, promotion, and progression (4–7). Although curcumin is known to regulate multiple signaling pathways involved in carcinogenesis, the mechanism by which curcumin simultaneously interferes with these pathways remains unclear. One probable explanation is that some botanicals, including curcumin, have the capacity to regulate microRNAs (miRNA) associated with cancer (8). miRNAs are small noncoding RNAs that regulate gene expression via interaction with the 3′-untranslated regions of target mRNAs, thereby regulating up to several hundred genes posttranscriptionally (9). Although miRNAs are known for their involvement in many physiologic processes, dysregulation of miRNA expression in cancer is well established (10). Curcumin has been shown to downregulate putative oncogenic miRNAs, such as miR-21 (11, 12), while it upregulates key tumor-suppressor miRNAs, including miR-200 family, let-7 family, miR-185b, and miR-22 (13–15). Therefore, regulation of miRNAs appears to be a crucial antitumorigenic mechanism of curcumin and warrants further systematic and comprehensive investigation.

Boswellic acid, an organic acid derived from the plant boswellia serrata, is another botanical used traditionally for the treatment of various inflammatory diseases, including colitis and arthritis. Although boswellia serrata is composed of various derivatives, including acetyl-β-boswellic acid, 11-keto-β-boswellic acid, and 3 acetyl-11-keto-β-boswellic acid (AKBA), AKBA is identified as the most potent anti-inflammatory constituent of boswellic acid (16, 17). Similar to curcumin, AKBA exerts its antitumorigenic effects through regulation of multiple cancer signaling pathways (16, 18–22). Interestingly, we recently demonstrated that AKBA upregulates key putative tumor-suppressive miRNAs in colorectal cancer and the expression of these miRNAs inversely corresponded with tumor size and volume in a xenograft animal model (23).

Despite lack of preclinical studies on combined treatment with curcumin and AKBA together, curcumin has been used in combination approaches with other dietary components. Treatment with curcumin and green tea catechins attenuated aberrant crypt formation in a carcinogen-induced colorectal cancer mouse model (24), while a combination of curcumin and resveratrol synergistically suppressed tumor proliferation in a mouse xenograft model (25). Although further investigations are required to fully understand the antitumorigenic properties of these compounds individually and in combination, these studies highlight the enormous therapeutic potential of using these safe and cost-effective botanicals together to help prevent and possibly treat colorectal cancer. Here, we identified key molecular mechanisms by which curcumin and AKBA both individually and in combination affect specific miRNAs and their downstream target genes involved in the cell-cycle regulation of colorectal cancer cell lines. Furthermore, we confirmed these antitumorigenic properties of curcumin and AKBA, both alone and together, in a mouse xenograft model.

Human colorectal cancer cell lines, HCT116, RKO, SW480, SW620, HT29, and Caco2 were purchased from the American Type Culture Collection. All cell lines were routinely authenticated by analyzing a panel of specific genetic and epigenetic biomarkers. The HCT116p53^−/− cell line was a generous gift from Bert Vogelstein (Johns Hopkins Medical Institute, Baltimore, MD). All cells were grown in Iscove Modified Dulbecco's medium (IMDM; Invitrogen) with 10% fetal bovine serum and 1% penicillin and streptomycin and maintained at 37°C in a humidified incubator (5% CO₂). Both curcumin (BCM-95) and AKBA (Bospure) were provided by Dolcas Biotech. These botanicals were dissolved in DMSO and diluted to appropriate experimental concentrations with tissue culture medium. In addition, aspirin (Acetylsalicylic acid) was purchased from Sigma-Aldrich and dissolved in DMSO and diluted to appropriate experimental concentration with tissue culture medium.

Cellular cytotoxicity was determined by the 3-(4, 5-dimethylthiazole-2-yl_2,5-diphenyl tetrazolium bromide; MTT) assay as described previously (23). In brief, approximately 4,000 cells were seeded in each well and treated with various concentrations of curcumin and/or AKBA for 72 hours. Optical density was determined using Tecan Infinite 200 Pro multireader and i-control 1.10 software (Tecan Group Ltd.). The Chou–Talalay equation (26) was used to calculate the combination index (CI) to determine the interaction between curcumin and AKBA treatments. Cell-cycle analysis was conducted using a Cell-cycle assay kit (MCH100106; Millipore) and apoptotic cell fraction was measured using Annexin V and Dead Cell Assay Kit (Millipore) according to the manufacturer's instructions using the Muse Cell Analyzer (Millipore). For clonogenic assays, approximately 500 cells were seeded in each well of a 6-well plate, and were treated with various concentrations of curcumin and/or AKBA. After 8 days, colonies were stained with crystal violet (Sigma-Aldrich) and dried overnight at room temperature. The number of colonies with >50 cells were counted using GeneTools (Syngene). All experiments were conducted in replicates and at least three independent experiments.

Microarray gene-expression analysis was performed in HCT116 and SW480 cell lines treated with curcumin and/or AKBA- and DMSO-treated controls according to the manufacturer's instructions and a method described previously (27). Detailed methodologies are supplied in Supplementary Materials and Methods.

Total RNA was extracted from curcumin and AKBA-treated colorectal cancer cell lines (10 μmol/L curcumin, 30 μmol/L AKBA, or the combination of both botanicals for 24 hours) and from the xenograft tumor tissue samples using the miRNeasy Mini Kit (Qiagen) following the manufacturer's instructions. For analysis of the mRNA expression, 1 μg of total RNA was reverse transcribed to complementary DNA using the Advantage RT PCR kit (Clontech Laboratories Inc.). Power SYBR Green (Applied Biosystems) real-time PCR was performed using StepOnePlus system (Applied Biosystems). For specific primer sequences, refer to Supplementary Figures (Supplementary Table S1). All qRT-PCR target genes expression was normalized to the expression of glyceraldehydes-3-phosphate dehydrogenase (GAPDH) and analyzed using the ΔΔC_t method. The expression of miRNAs was analyzed using TaqMan real-time PCR assay kit (Applied Biosystems). All miRNA primers used in this study were purchased from Ambion. Ten nanograms of RNA from each sample was reverse transcribed using TaqMan microRNA Reverse Transcription kit (Applied Biosystems) and 6 ng of complementary DNA was used for real-time qRT-PCR. All data were analyzed using the ΔΔC_t method and normalized to RNU6B.

A total of 1 × 10⁵ cells were seeded and transfected with pre-miR-34a, anti-miR-27a, pre-miR-negative-control 1, or anti-miR-negative-control 1 (Ambion) at a final concentration of 50 nmol/L using siPort NeoFX (Applied Biosystems) and Opti-MEM (Gibco) according to the manufacturer's instructions. The relative levels of miRNAs were measured to confirm the efficiency of miRNA transfection and functional analyses were conducted as described above.

Western immunoblotting experiments were performed as described previously (28). In brief, cells were treated with curcumin and/or AKBA for 24 hours, thereafter lysed using 100 μL of 1× SDS sample buffer containing β-mercaptoethanol. All antibodies are listed in Supplementary Table S2. 1. All samples were compared against β-actin as a reference (Sigma-Aldrich). The bands were visualized using GeneTools (Syngene) and images were captured using Syngene GBox.

Apoptotic cells were evaluated using 4′,6-diamidino-2-phenylindole (DAPI) staining as described previously (29). In brief, cells were treated with different concentrations of curcumin and/or AKBA for 24 hours. The cells were then fixed with methanol for 30 minutes at 4°C and incubated in DAPI solution for 1 hour in the dark. The number of apoptotic cells was evaluated under fluorescence microscope (Leica).

Three-dimensional (3D) cultures were generated as described previously (30). In brief, approximately 1 × 10⁶ cells were seeded on a cellulose filter on top of a steel mesh bridge. Cells were then treated with curcumin (10 μmol/L), AKBA (30 μmol/L), or the combination of both for 1, 3, 7, or 10 days. The growth of 3D cultures was assessed by determining the diameter of the 3D cultures using light microscopy.

The 5-week-old male athymic nude mice were purchased from Harlan Laboratories and kept under controlled conditions of light (12-hour cycles), fed ad libitum diet, and had free access to water. Xenograft tumors were generated by subcutaneous injection of 5 × 10⁶ HCT116 cells. Tumor growth was determined by measuring the length (L) and width (W) of the tumor every other day using calipers and the tumor volume was calculated according to the following formula: 1/2LW². Once average tumor size reached 50 mm³, animals were randomly divided into four groups with 10 animals in each group: (i) control vehicle (DMSO), (ii) curcumin (25 mg/kg), (iii) AKBA (75 mg/kg), or (iv) curcumin and AKBA. Treatments or vehicle were injected intraperitoneally daily for 3 weeks. The tumor samples were fixed in RNAlater (Sigma-Aldrich) and then stored in a −80°C freezer for subsequent analyses. For the xenograft tumor growth experiment, the interaction between curcumin and AKBA was evaluated as described previously (31). The fraction of tumor volume (FTV) affected by curcumin and/or AKBA was calculated individually and together as a ratio to control. The ratio of expected FTV (Exp FTV) and observed FTV (Obs FTV) was calculated for the combined treatment (A ratio >1 indicates synergistic effect, <1 indicates an additive effect). The animal protocol was approved by Institutional Animal Care and Use Committee at the Baylor Research Institute (Dallas, TX).

All analyses were performed using GraphPad Prism Ver. 6.0 (GraphPad Software Inc.). All data were expressed as mean ± SEM with statistical significance indicated when P < 0.05. Statistical comparisons between control and treatment groups were determined using an unpaired t test or one-way ANOVA with the Tukey post-hoc tests.

The average IC₅₀ of four cell lines (HCT116, RKO, SW480, and SW620; Supplementary Fig. S1A) was used to determine the potency of both botanicals and the ratio between these botanicals was calculated (Curcumin:AKBA, 1:2.76; Supplementary Fig. S1B). On the basis of these data, we selected the concentration of AKBA as three times the dose of curcumin for this study. We then examined the cytotoxicity of curcumin and/or AKBA on two primary colorectal cancer cell lines (HCT116 and SW480). Both curcumin and AKBA treatment increased the cytotoxicity in a dose-dependent manner (Fig. 1A). The combined treatment further enhanced cytotoxicity in both cell lines. Next, we determined whether this enhanced cytotoxicity by the combined treatment was synergistic using the Chou–Talalay CI (26). Interestingly, while the cytotoxicity of the both treatments together was additive in HCT116 cell line (CI = 1.09), combined treatment of curcumin and AKBA in the SW480 cell line resulted in synergism (CI = 0.768; Fig. 1A, insets). Considering tumor-suppressor p53 is mutated in the SW480 cell line, but not in the HCT116 cell line, we next investigated whether p53 dysregulation contributed to the discrepancy of AKBA and curcumin-induced cytotoxicity between these cell lines using p53^−/− HCT116 cells (HCT116p53^−/−). Although treatment of curcumin or AKBA resulted in a dose-dependent increase in cytotoxicity in HCT116p53^−/− cell line, the combination of curcumin and AKBA showed synergistic enhancement in cytotoxicity (CI < 1; Fig. 1A, inset), suggesting that AKBA complements cytotoxicity of curcumin in cell lines lacking active p53. In addition, to ensure that the combined botanicals exert synergistic enhancement in cytotoxicity in p53-mutated cell lines, we determined synergism between curcumin and AKBA in RKO (p53 WT), Caco3, and HT29 (both p53-mutant cell lines; Supplementary Fig. S2A). Although RKO cells showed no synergistic cytotoxicity between curcumin and AKBA (CI = 1.21), the combined botanicals exerted synergistic cytotoxicity in both HT29 and Caco2 cell lines (CI = 0.989 and 0.971, respectively). Next, we evaluated the effects of these botanicals on cell survival and proliferation using the clonogenic assays (Fig. 1B and Supplementary Fig. S2B). To our surprise, even at a low dose of 2.5 μmol/L, curcumin significantly suppressed colony formation, while over a 10-fold higher concentration of AKBA was required to exert a similar effect (Fig. 1B), indicating that curcumin has superior inhibition of cell survival and proliferation compared with AKBA.

To identify cancer signaling pathways modulated by curcumin and AKBA in colorectal cancer, we used microarray analysis to examine gene-expression alterations caused by curcumin and AKBA individually, and in combination in the two colorectal cancer cell lines. Hierarchical clustering analysis showed that a markedly different cluster of genes was regulated by curcumin and AKBA in HCT116 and SW480 cell lines (Fig. 2A). Furthermore, gene-expression profiles of the combined treatments revealed that a large cluster of genes unaffected by curcumin treatment alone was regulated by the combined treatment with curcumin and AKBA (Fig. 2B), suggesting that these botanicals in combination modulate additional cluster of genes compared with individual treatments. We then assessed all genes regulated by the dual treatments in both colorectal cancer cell lines and identified 355 genes with overlapping expression alterations (Fig. 2C). Next, we used the Ingenuity Pathway Analysis (IPA) to determine potential molecular signaling pathways regulated by the combined curcumin and AKBA treatment (Fig 2D). Top cancer-related molecular functions common to both cell lines were cell death and survival, cellular growth and proliferation, and cell cycle (Fig. 2D). Because cell cycle is related to both cell death and proliferation, we next focused on systematic evaluation of genes involved in cell-cycle regulation among the subset of 355 differentially expressed genes. Subsequently, we identified MYC, CCND1, CCNE1, and CDK6 as potential target genes of the combined curcumin and AKBA treatment (Fig. 2C).

We next investigated the mechanism of cell-cycle regulation by curcumin and AKBA alone and in combination in colorectal cancer cells. Curcumin treatment induced G₂–M arrest in HCT116 and HCT116p53^−/− cell lines, but not in the SW480 cell line, while AKBA induced G₁ arrest in HCT116 and SW480 cell lines. When cells were treated with curcumin and AKBA together, G₁ arrest was induced in HCT116 and SW480 cell lines, while G₂–M arrest was predominantly observed in the HCT116p53^−/− cells (Fig. 3A). These data indicate that combinatorial treatment with curcumin and AKBA exerts antitumorigenic effects through cell-cycle regulation. Next, we determined whether curcumin and AKBA enhanced apoptosis in colorectal cancer cell lines. Both curcumin and AKBA treatment increased cellular apoptosis in all three cell lines, while the two botanicals together further elevated apoptotic cell population regardless of the cell line analyzed (Fig. 3B). We then used the DAPI nuclear staining to confirm that both curcumin and AKBA enhanced cellular apoptosis (Fig. 3C). In somatic cells, the Bax:Bcl ratio serves as a regulator for cell susceptibility to apoptosis (32); hence, we first assessed whether curcumin and AKBA regulate cell survival proteins, Bcl-2 and Bcl-x_L. Although Bcl-2 and Bcl-x_L were downregulated by both curcumin and AKBA alone, the effect of the dual treatment augmented Bcl-2 and Bcl-x_L inhibition in all three colorectal cancer cell lines (Fig. 3D). We then showed that both curcumin and AKBA upregulated proapoptotic protein Bax, while the two treatments together further elevated the expression of Bax (Fig. 3D). In addition, poly (ADP-ribose) polymerase 1 (PARP1), a DNA single-strand breaks repair protein and a suppressor of the caspase pathway, was downregulated by curcumin and AKBA, both individually and in combination. Collectively, in line with microarray data, we confirmed that the combination of curcumin and AKBA results in increased cell-cycle arrest and apoptosis. In addition, to confirm the antitumorigenic properties of these botanicals, we used 3D cultures to determine whether curcumin and/or AKBA could interfere with colorectal cancer tumor growth (Supplementary Fig. S3). Although both curcumin and AKBA inhibited cellular growth individually, combined treatment with curcumin and AKBA further inhibited the growth in HCT116 cells. Together, we demonstrated that curcumin and AKBA exert antitumorigenic effects through cell-cycle arrest and apoptosis, while suppressing cellular growth in both 2D and 3D in vitro models.

Using the genes identified by the microarray analysis, we used TargetScan (Ver 6.2) and miRTarBase to determine potential miRNAs regulated by curcumin and AKBA. We initially selected miR-34a, miR-145, and miR-16 as potential candidates based on their predicted target genes. miR-34a expression was upregulated by curcumin and further elevated by concurrent treatment with curcumin and AKBA in HCT116 cells (Fig. 4A), but had no significant effect in both SW480 and HCT116p53^−/− cell lines. Considering that miR-34a has been shown to be primarily activated by p53 (33), these data indicate that curcumin and AKBA modulate miR-34a expression through p53. This finding was consistent in RKO (p53 wild-type), Caco2, and HT29 (both p53-mutated) cell lines as well (Supplementary Fig. S4A). Next, we used Western blot analysis and qRT-PCR analyses to determine whether these botanicals suppressed expression of miR-34a target genes. Both curcumin and AKBA downregulated cMyc, CDK4, CDK6, and cyclin D1 expression, whereas the combined treatment further increased the inhibitory effects in the HCT116 cell line (Fig. 4B). To our surprise, miR-34a target gene expression was also downregulated in cell lines with p53 mutation/deficiency, indicating that these botanicals exert antitumorigenic effects only, in part, through miR-34a upregulation in colorectal cancer cells. Next, we investigated the expression of miR-145 and miR-16, both putative tumor-suppressive miRNA involved in modulation of cell cycle (34, 35) in colorectal cancer cell lines. Although miR-145 expression was increased by AKBA alone and with the combined treatment in the HCT116 cell line, curcumin treatment alone did not alter miR-145 expression in any of the colorectal cancer cell lines (Supplementary Fig. S4B). The combination of curcumin and AKBA treatment resulted in a significant increase in miR-16 levels in the HCT116 p53^−/− cells, while curcumin or AKBA alone did not significantly influence miR-16 expression in any cell line (Supplementary Fig. S4B). Therefore, these data collectively indicate that not all tumor-suppressive miRNAs are regulated coordinately by curcumin, and that AKBA and curcumin target different miRNAs to exert their antitumorigenic effects.

Because our initial focus was on the identification of putative tumor-suppressive miRNAs, we investigated whether curcumin and AKBA can also regulate the expression of oncogenic miRNAs. Recent studies demonstrated that miR-27a, an oncogenic miRNA, was suppressed by curcumin in colorectal cancer cell lines (36, 37). Although we confirmed that miR-27a was downregulated by curcumin treatment alone in all three colorectal cancer cell lines, its expression was further inhibited by the combined treatment. These data indicate that unlike miR-34a, curcumin and AKBA inhibit miR-27a independently of p53 activation (Fig. 4A). Furthermore, we confirmed downregulation of miR-27a expression by the combined curcumin and AKBA treatment in RKO, HT29, and Caco2 cell lines (Supplementary Fig. S4A). We then investigated whether inhibition of miR-27a by curcumin and AKBA resulted in alteration of FBXW7, a tumor suppressor and a direct target of miR-27a (38). Although curcumin treatment alone upregulated FBXW7 expression, the combined treatment further enhanced FBXW7 expression (Fig. 4B; Supplementary Fig. S5). In addition, we demonstrated that both curcumin and AKBA individually and in combination suppressed cyclin D1 and cMyc, as well as downstream targets of FBXW7 (Fig. 4B; Supplementary Fig. S5).

To confirm that miR-34a upregulation contributes to the antitumorigenic effects of botanicals, we transiently overexpressed miR-34a and investigated its molecular function in the three colorectal cancer cell lines (Supplementary Fig. S6A). Overexpression of miR-34a resulted in an enhanced apoptotic cell population in all three cell lines (Supplementary Fig. S6B); however, miR-34a–induced apoptosis appeared to be less prominent in p53-mutated/deficient cell lines. These results support the idea that there is a positive feedback loop between miR-34a and p53 (33). Furthermore, miR-34a overexpression resulted in G₁ arrest in all cell lines (Supplementary Fig. S6C), while it resulted in suppressed cell proliferation and colony formation (Supplementary Fig. S6D and S6E). Similarly, we investigated antitumorigenic activity of miR-27a inhibition in colorectal cancer cell lines. Transient knockdown of miR-27a enhanced cellular apoptosis in all three colorectal cancer cell lines (Supplementary Fig. S7A and S7B), while cell-cycle analysis showed that miR-27a knockdown resulted in an increase in G₂–M phase arrest in HCT116 and HCT116 p53^−/− cell lines, while prolonged S-phase fraction in the SW480 cell line (Supplementary Fig. S7C). miR-27a knockdown resulted in suppression of cell growth and colony formation in all three cell lines (Supplementary Fig. S7D and S7E).

To confirm the therapeutic potential of curcumin and AKBA in vivo, we generated mouse xenograft tumors by injecting HCT116 cells subcutaneously into nude athymic mice. We subsequently treated the animals with curcumin (25 mg/kg body weight) and/or AKBA (75 mg/kg b/w) for 20 days (Fig. 5A). Although previous studies used 100 mg/kg curcumin to attenuate tumor growth in the mouse xenograft model (39, 40), we purposely chose a lower dose of curcumin to mimic a more reasonable and physiologically achievable clinical dose of curcumin. The body weight of the animals was unaffected by the treatment during the experiment (Fig. 5B). The progressive growth in tumor volume showed that treatment with curcumin or AKBA suppressed tumor growth at a similar rate (Fig. 5C). Interestingly, treatment with botanicals appeared to influence tumor growth immediately, because the average tumor volume of treatment groups deviated from the control group 4 days from the initiation of treatment. The combined curcumin and AKBA treatment resulted in further inhibition in the growth of the xenograft tumors. Confirming our measurement for tumor volume, tumor weight positively correlated with changes the tumor volume (Fig. 5D). To determine whether combined treatments resulted in an enhanced suppression of tumor growth, we calculated the expected tumor volume using FTV of curcumin and AKBA (Fig. 5C). Interestingly, throughout the treatment period, curcumin and AKBA together resulted in synergistic tumor growth suppression (ratio of expected:observed FTV > 1). In addition, we wanted to ascertain whether these in vivo antitumorigenic effects of curcumin and AKBA corresponded with the expression alterations of miR-27a and miR-34a, as witnessed in cell culture studies. The expression of miR-27a was downregulated in tumors from mice treated with AKBA or curcumin and AKBA (Fig. 5E), while tumors from mice treated with curcumin and/or AKBA had increased miR-34a (Fig. 5F). In addition, we measured the expression of target genes of these miRNAs in the xenograft tumor tissues. We confirmed that FBXW7, cMyc, CDK6, and cyclin E1 were modulated by the combined treatment (Supplementary Fig. S8). Collectively, these data indicate that changes in expression of these two miRNAs in tumors, at least in part, contributed to tumor growth inhibition.

Finally, to compare the antitumorigenic properties of both botanicals with a well-established chemopreventive drug, we used the MTT and colony formation assays to compare the potency of curcumin and AKBA with aspirin, a nonsteroidal anti-inflammatory drug (NSAID). A previous study showed that aspirin is more effective in PI3KCA-mutated cancers (41); hence, we compared the effects of these botanicals to aspirin in two cell lines HCT116 (PI3KCA-mutated) and SW480 (PI3KCA wild-type). Although both curcumin and AKBA treatments significantly inhibited colony formation at 2.5 μmol/L, we found that 5,000 μmol/L aspirin was required to moderately inhibit colony formation in both HCT116 and SW480 cell lines, indicating that curcumin was significantly more effective in inhibiting colony formation than aspirin (Fig. 6A). Similarly, results of the MTT assay indicated that 20,000 μmol/L aspirin was needed to exert a similar cellular cytotoxicity to the equivalent of the combined curcumin and AKBA treatment of 10 μmol/L for HCT116 cell line, while almost 30,000 μmol/L aspirin was needed for SW480 cell line (Fig. 6B). In addition, while PI3KA wild-type SW480 cell line appeared to be more resistant to aspirin than HCT116 cell line, the combined botanicals were just as potent in both cell lines. Although there are other parameters contributing to drug efficacy, such as bioavailability and drug retention time, these comparisons highlight the potency of these botanicals against colorectal cancer cells.

Herein, we demonstrate for the first time that curcumin and AKBA can act synergistically to exert antitumorigenic effects in colorectal cancer cells, in both in vitro and in vivo experimental models. Analysis of gene-expression microarrays identified cell-cycle regulation, cellular apoptosis, and proliferation as the primary protective mechanisms by which curcumin and AKBA act in colorectal cancer cell lines, which were subsequently validated in a series of molecular and functional analyses during the course of this study. On the basis of data obtained from gene-expression microarrays, we illustrate that curcumin and AKBA regulated the expression of miR-34a and miR-27a, two miRNAs involved in cell-cycle regulation in cancer and modulated their downstream target genes. Finally, we used a mouse xenograft model to confirm our in vitro data and demonstrated that the combination of these botanicals resulted in a synergistic suppression of tumor growth.

miRNAs are a class of small noncoding RNAs that play critical roles in the regulation of gene expression. miR-34a is a well-established tumor suppressor that regulates both cell cycle and apoptosis pathways. Therefore, it is not surprising that miR-34a is primarily regulated by p53, a key cell cycle–regulating tumor suppressor (33). We demonstrated that curcumin and AKBA only upregulated miR-34a in the HCT116 cell line and not HCT116p53^−/− and SW480 cell lines, suggesting that these botanicals control miR-34a expression through p53 activation. miR-34 is one of the miRNAs identified for its chemotherapeutic use and it is currently being clinically analyzed with patients who have unresectable primary liver cancer and advanced or metastatic liver cancer (NCT01829971). We overexpressed miR-34a in colorectal cancer cell lines that resulted in inhibition of cell proliferation and colony formation; however, it was noted that the effect was less pronounced in p53-mutated/deficient cell lines. These results suggest that interference in the miR-34a and p53-positive feedback loop may have contributed to the lack of efficacy of miR-34a–induced inhibition of cell proliferation and colony formation (33). Considering p53 mutation occurs in the latter stages of colorectal tumorigenesis (42), the role of botanicals in upregulating miR-34a during initiation and progression stages of colorectal cancer is of importance. With curcumin and AKBA suppressing validated miR-34a target genes, including cMyc, cyclin D1, CDK4, CDK6, and Bcl-2, our data imply that these botanicals could be used therapeutically to upregulate miR-34a in colonic epithelial cells.

In contrast, miR-27a has been identified recently as a putative oncogene in colorectal cancer (38). miR-27a facilitates proliferation through Sp protein repressor ZBTB10 and FBXW7 (38, 43) and this study confirmed the oncogeneic role of miR-27a in colorectal cancer through series of knockdown experiments. Previous studies have demonstrated that curcumin has been shown to downregulate miR-27a in colorectal cancer cell lines (36, 37). Although we confirmed these findings, we also demonstrated that curcumin and AKBA in combination can further suppress miR-27a expression in colorectal cancer cell lines. One of the oncogenic functions of miR-27a in colorectal cancer is through direct inhibition of the tumor-suppressor FBXW7 (38). FBXW7 is a member of F-box family of proteins that interacts with Skp1 and CUL1 to form a ubiquitin ligase complex (44) that play a pivotal role in the regulating the cell cycle by degrading key proteins in cell division and determination of cell fate, including cMyc, cyclin E1, c-Jun, and Notch (45). This study demonstrated that treatment of colorectal cancer cells with curcumin and AKBA upregulated FBXW7, resulting in downregulation of its downstream target genes. Collectively, our data indicate that these botanicals play an important role in regulation of the miR-27a–FBXW7 axis. On the basis of the findings from this study, we propose a mechanism by which curcumin and ABKA control cell-cycle regulation in colorectal cancer cells (Fig. 6C). Curcumin and AKBA can modulate miRNAs, such as miR-34a and miR-27a, in the presence of p53 to control cell-cycle regulation. However, these botanicals can still maintain cell-cycle regulation through miRNA regulation in the absence of p53. Although additional investigations are required to fully understand the complex mechanisms involved in miRNA regulation by curcumin and AKBA, this study highlights the therapeutic potential of these botanicals to coordinate modulation of cancer-related miRNAs.

Mutations in p53 are found in nearly all types of cancers and p53 is the most frequently mutated gene in colorectal cancer (46). p53 is known for its role in maintaining genomic integrity by regulating cell cycle and cell death pathways (47). This study has shown that the combination of curcumin and AKBA induced synergistic cytotoxicity in p53-deficient/mutated cell lines, but not p53 wild-type cell lines. Although curcumin has been shown to induce apoptosis through p53 independently, it primarily induces apoptosis through stabilization of p53 through inhibition of MDM2 (48). We suspect that the synergism we observed between curcumin and AKBA in this study is a result of AKBA compensating for lack of curcumin's cytotoxicity in these cell lines. Considering all cancers are heterogeneous, combining antitumorigenic agents may increase effectiveness of these agents in various human tumors. Although further investigation is needed to clarify how these botanicals interact to overcome mutations in colorectal cancer, these compounds have a promising potential to provide safe and cost-effective adjunctive treatment alongside conventional therapies.

To confirm the antitumorigenic properties of curcumin and AKBA in vivo, we generated xenograft tumors using human colorectal cancer cells in mice and treated these mice with curcumin and/or AKBA. As expected, treatment of mice with curcumin or AKBA resulted in decreased tumor proliferation. However, to our surprise, the combined treatment resulted in synergistic suppression of tumor growth. Although the synergistic tumor-suppressive effect observed for the combined treatment was modest, this finding is important considering that consumption of these two botanicals concurrently is extremely safe. Given that curcumin and AKBA are both readily available and affordable, there is a significant potential for use of these botanicals to prevent colorectal cancer development.

Finally, to highlight the potency of these botanicals with more conventional chemopreventive drugs, we compared the antitumorigenic properties of curcumin and AKBA to aspirin. Recently, a clinical observational study concluded that daily consumption of aspirin, a NSAID, lowers the risk of colorectal cancer (49). The cytotoxicity of curcumin was over 2,000 times more than aspirin in vitro, while it suppressed cellular proliferation at a significantly lower dose than aspirin demonstrating that antitumorigenic potential of these botanicals far exceeded that of aspirin. Although our data along with many preclinical studies confirmed the antitumorigenic effects of botanicals, the poor bioavailability associated with both curcumin and AKBA has been perceived as a potential hindrance for their use in prevention trials (50, 51). However, more recently, there is increasing research focus on improving bioavailability of botanicals, including curcumin, through development of new formulations (52). In addition, in vivo studies have demonstrated that curcumin exerts antitumorigenic effects when consumed orally (25, 53). Similarly, we have also demonstrated previously that gavaging of AKBA resulted in attenuation of tumor growth in a mouse xenograft model (23). Although this study focused on antitumorigenic synergism between these botanicals through intraperitoneal administration, our study provides strong evidence for these two botanicals as potential chemotherapeutic agents. Furthermore, unlike aspirin, high levels of curcumin can be consumed safely without increasing the risk of gastrointestinal bleeding and peptic ulcer—the risks associated with continuous administration of aspirin (49). In addition, daily consumption of these botanicals can provide added health benefits as curcumin is effective against other diseases such as rheumatoid arthritis and depression (54, 55).

In summary, we for the first time demonstrate that individual and combined treatment with curcumin and AKBA exerted their antitumorigenic effects, at least in part, by regulation of miRNAs and their downstream target genes involved in cell-cycle regulation. Furthermore, using a mouse xenograft model, we confirmed that curcumin or AKBA suppressed tumor growth, while the combination resulted in synergistic tumor suppression. Collectively, this study highlights the novel therapeutic potential of using the combined treatment of two botanicals, curcumin and AKBA, to inhibit colorectal cancer tumor growth.

No potential conflicts of interest were disclosed.

Conception and design: S. Toden, A. Goel

Development of methodology: S. Toden, M. Shakibaei

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S. Toden, Y. Okugawa, C. Buhrmann, E. Anguiano, M. Shakibaei

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S. Toden, Y. Okugawa, D. Nattamai, N. Baldwin, M. Shakibaei, A. Goel

Writing, review, and/or revision of the manuscript: S. Toden, Y. Okugawa, D. Nattamai, C.R. Boland, A. Goel

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): E. Anguiano, A. Goel

Study supervision: C.R. Boland, A. Goel

Other (funded the project and supervised the laboratory): C.R. Boland

The authors thank Ms. Patricia Kraehe for technical assistance. The authors also thank Margaret Hinshelwood, manager of the Office of Scientific Publications, Baylor Charles A. Sammons Cancer Center, Dallas, for critical suggestions and editing to further improve the quality of this article.

This study was supported by the grants R01 CA72851, CA 181572, and U01 CA187956 from the National Cancer Institute, NIH, and funds from the Baylor Research Institute to A. Goel.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.