Chemopreventive Activity of Plant Flavonoid Isorhamnetin in Colorectal Cancer Is Mediated by Oncogenic Src and β-Catenin

Colorectal cancer is the fourth most commonly diagnosed type of cancer in the United States, and it is the second leading cause of cancer-related deaths (1). Chronic inflammation, such as ulcerative colitis and Crohn disease, are associated with increased risk of colorectal cancer (2–5).

The Src family of tyrosine kinases (SFK) are nonreceptor protein tyrosine kinases that are activated in multiple cancers, including colorectal cancer (6). Increased Src activity in primary colorectal cancer is an indicator of poor prognosis (7). C-terminal Src kinase, also known as c-Src kinase (CSK), negatively regulates SFK by phosphorylation of a C-terminal tyrosine (Y530 in c-Src; ref. 8). C-terminal phosphorylation of SFK stabilizes the protein in an inhibitory confirmation that prevents autophosphorylation of the activation loop tyrosine (Y419 in c-Src). The receptor-like tyrosine phosphatase CD45 and similar phosphatases impose a reciprocal regulation of Src by removing the C-terminal tyrosine phosphate (9).

Activated Src can regulate many downstream pathways including phosphoinositide 3-kinase (PI3K), Ras–Mek–extracellular signal–regulated kinase (ERK), STAT3, and p130 to increase survival, proliferation, angiogenesis, motility, and invasion (6). Activated Src can also phosphorylate β-catenin, causing its release from E-cadherin sequestration at the plasma membrane and enhancing its nuclear localization (10). Colorectal cancer seems to be sensitive to loss of Src regulation. Both the loss of expression of the negative regulator CSK and the overexpression of Src have been implicated in colorectal carcinogenesis (8).

The cost of treating colorectal cancer is estimated to be $6.5 billion per year (11). Dietary change and use of dietary supplements, both feasible and safe, represent a viable and important strategy for preventing colorectal cancer (12). Recent analysis of the Polyp Prevention Trial (PPT), a clinical trial that investigated the role of diet modulation in the prevention of colorectal adenoma recurrence, suggested that consuming an isorhamnetin-rich diet was associated with a decreased risk of advanced adenoma recurrence (13).

Recently, it was shown that isorhamnetin can suppress skin cancer by binding to and inhibiting MAP (mitogen-activated protein)/ERK kinase (MEK) 1 and PI3K (14). To further investigate the potential role of flavonols in colorectal cancer prevention, we assessed the effects of four flavonols commonly consumed by humans, isorhamnetin, quercetin, rutin, and myricetin, in a mouse model for colorectal cancer. We found that supplementing the diet with isorhamnetin significantly reduced colorectal tumorigenesis in these mice. The isorhamnetin diet seemed to resolve the dextran sodium sulfate (DSS)–induced inflammation as measured by MRI, histopathology, and immunohistochemical analysis (IHC). Isorhamnetin also inhibited Src activity and nuclear localization of β-catenin and increased expression of CSK both in vivo and in colorectal carcinoma cells. Finally, we showed that inhibition of Src and nuclear β-catenin was dependent on the upregulation of CSK in colon cancer cells and associated with CSK expression in isorhamnetin-fed mice.

All mouse experiments were agreed to and regulated by the Animal Care and Use Committee of the National Cancer Institute (Frederick, MD). Pathogen-free male FVB/N mice were purchased from the NCI-Frederick Animal Production Area at 5 weeks of age and were exposed to a 12:12–hour light/dark cycle. Mice were given an AIN-96G–purified diet from Harlan Teklad and drinking water ad libitum. At 6 weeks of age (day 0; Fig. 1), mice were injected intraperitoneally with azoxymethane (Sigma) at a dose of 10 mg/kg body weight in 0.1 mL of saline (Fig. 1). Seven days later, mice were treated with 2% DSS, 36,000 to 50,000 kDa (MP Biomedicals LL), dissolved in normal drinking water (reverse osmosis–purified water) for 7 days and then switched back to normal drinking water. Body weights were measured during the DSS treatment. Three days after DSS treatment ended, the mice were evenly sorted by change in body weight into five diet groups to ensure that effects of the DSS were evenly represented in each of the different diets. Mice were single caged and started on the indicated diet (day 17; Supplementary Table S1). Isorhamnetin and myricetin were obtained from Jinan Haohua Industry Co., Ltd. and quercetin and rutin were obtained from Sigma-Aldrich Co., LLC. Purified diets were prepared by Harlan Laboratories. Mice were allowed to eat ad libitum. One day before euthanasia (days 20, 25, and 33), 12 mice per group and time point were imaged by MRI as described previously (15). Mice were euthanized at indicated times and colon tissue was collected.

Colonic tissue was harvested from mice at various times after azoxymethane or AOM/DSS treatment, fixed in 4% formaldehyde overnight, and stored in 70% ethanol. Fixed sections of colonic tissues were embedded in paraffin, cut into 5-μm sections, and put onto microscopic slides by Histoserv Inc. Slides were either stained with hematoxylin and eosin (H&E) for histologic analysis via light microscopy or deparaffinized in xylene and rehydrated in graded ethanol for IHC. For histopathology analysis, a section of approximately 5 mm in length of the distal colon was evaluated for each animal. The tissues were blinded and scored by the pathologist semiquantitatively from 0 to 15 (Supplementary Table S2). Colons were evaluated beginning at the colorectal junction and proceeding adorally 5 mm to the middle colon. Briefly, the severity of the leukocytic infiltrate in the mucosa was subjectively assessed as mild, moderate, or severe (1, 2, and 3, respectively); the distribution was evaluated and denoted as focal/locally extensive, multifocal, or diffuse (1, 2, and 3, respectively); and the depth was evaluated and denoted as mucosa only, extends to submucosa, or extends to muscularis mucosa or serosa (1, 2, and 3, respectively). The distribution of atypical glandular hyperplasia (AGH)/squamous (sq) metaplasia was assessed as focal, multifocal, or diffuse (1, 2, and 3, respectively). The distribution of ulceration was assessed as focal, multifocal, or diffuse (1, 2, and 3, respectively). If necrosis was present it was subjectively assessed as mild, moderate, or severe and scored accordingly (1, 2, and 3, respectively). A score of 0 was assigned for each criterion not represented in the section, which was the case for the majority of mice for ulceration and necrosis. Total disease score per mouse was calculated by summation of the six parameters for each mouse.

For IHC, the sections were placed in low-pH–modified citrate buffer (Dako) and subjected to antigen retrieval via pressure chamber. The sections were preincubated in 3% H₂O₂ for 20 minutes and blocked in normal goat serum for 1 hour. Sections were incubated overnight in primary antibody solution diluted as follows: Ki-67, phospho-β-catenin⁶⁵⁴ (Abcam Laboratories) 1:50, CD45 (BD Biosciences) 1:50, pSrc^Y417 and β-catenin (Cell Signaling Technology) 1:100 and 1:1,000, respectively. Slides were incubated for 1 hour in a biotin-conjugated secondary antibody diluted 1:200. The sections were treated with a Vectorstain ABC kit (Vector Laboratories); the resulting activity was detected with diaminobenzidine (Sigma) and counterstained in hematoxylin. For CSK (Abcam Laboratories), tissues were incubated 1:50 for 1 hour and detected with a Ventana iView DAB detection kit (Ventana Medical Systems) on a Ventana BenchMark.

Human colon cancer HT29 cell line was obtained from American Type Culture Collection and cultured as described previously (16).

HT29 cells were seeded on 100-mm plates and treated with either dimethyl sulfoxide (DMSO) vehicle (<1%) or isorhamnetin (10, 20, and 40 μmol/L) for 24 hours. The media was removed and cells were washed with cold PBS. Cells were lysed with complete radioimmunoprecipitation assay (RIPA) lysis buffer (Santa Cruz Biotechnology). Protein concentration was calculated using standard bovine serum albumin (BSA) curve. Approximately, 30 to 80 μg of protein was loaded and separated by SDS-PAGE, transferred to nitrocellulose membrane, and probed with the following antibodies at indicated dilutions: CSK (1:1,000), pSRC⁴¹⁷ (1:1,000), pSRC⁵²⁹ (1:1,000), β-catenin (1:2,000), pAKT (1:1,000), pERK (1:2,000), E-cadherin (1:1,000), pGSK3αβ (1:1,000; Cell Signaling Technology), and pβ-catenin⁶⁵⁴ (Abcam Laboratories) overnight at 4°C, followed by incubation in horseradish peroxidase–labeled secondary antibodies for 1 hour. Cells were developed with SuperSignal West Femto Chemiluminescent Substrate (Thermo Scientific). Each membrane was probed with β-actin (1:5,000; Sigma) to ensure consistent protein loading.

Total RNA from HT29 cultured cells was isolated using TRIzol RNA extraction method. Briefly, chloroform was added to the sample, and after shaking and centrifugation the aqueous layer was isolated. Isopropanol was next added to the sample to precipitate the RNA. Following centrifugation, the RNA pellet was washed with 70% ethanol. RNA was further purified using an RNeasy kit (Qiagen). The cDNA was synthesized using Bio-Rad iScript reverse transcriptase, and the PCR reactions were carried out using Bio-Rad SYBR Green Master Mix carried out in Bio-Rad iCycler (Bio-Rad Laboratories). The human csk splicing variant 1 and 2 were purchased from IDT and sequence is as follows: sense, 5′-TCCGGCCCCGTCTCTCTTGG and antisense, 5′-ACCCTCACGGGCAGGACAGG. The PCR included nontemplate control and glyceraldehyde-3-phosphate dehydrogenase (gapdh) as control primer set. The csk transcript activity level was calculated using |$2^{- {\rm \Delta \Delta}C_{\rm T}}$|⁠, where ΔC_T = C_T csk − C_T gapdh and Δ (ΔC_{T stimulated} − ΔC_T control).

HT29 cells were stably transfected with three unique 27mer siRNA duplexes—2 nmol each against csk siRNA (Trilencer-27; Origene Technologies) or a scrambled control siRNA using the manufacturer's transfection reagent and recommendations. After transfection, cells were incubated for 36 hours. The total protein was isolated and subjected to Western blot analysis.

HT29 cells were seeded onto 100-mm plates and pretreated with either DMSO vehicle (<1%) or isorhamnetin (10, 20, and 40 μmol/L) for 24 hours. Cells were trypsinized and seeded at 30,000 cells in 2× Dulbecco's Modified Eagle Medium (DMEM). Cell suspension was added 1:1 with 0.5% agarose (2 mL/well in a 6-well plate) and constitutes the top layer. The bottom layer consisted of 2 mL of 1.2% agarose. The cells were maintained in an incubator for 14 days and the colonies were scanned and counted with GelCount (Oxford Optronix Ltd.).

Statistical analyses were done using SAS, version 9.2 (SAS, Inc.) software. A Student unpaired t test was used to compare treatment group averages for histopathology and cell culture data. Generalized linear models were used to compare treatment group averages for mouse morbidity (binomial; PROC GLIMMIX) and tumor data (PROC GLM). In addition to diet, the number of DSS rounds was also included in the linear models. To achieve normality, tumor number was natural log–transformed and tumor burden and size were twice natural log–transformed, after adding a 1 to prevent values below zero. Histopathology and cell culture data are shown as raw mean and their SEs; tumor data are shown as least-squares mean and their SEs, which were back transformed to their original scale. All statistical tests were two-sided. The P values were not adjusted for multiple comparisons. Significance was declared at P ≤ 0.05 and a tendency at 0.05 to 0.10.

We previously showed that a high intake of flavonol-enriched foods and beverages, specifically isorhamnetin, kaempferol, and quercetin, was associated with decreased risk of advanced adenoma recurrence (13, 17). To verify the chemoprotective effects of individual flavonols on colorectal tumorigenesis, we fed FVB/N male mice previously treated with azoxymethane and DSS, diets enriched with four primary flavonols consumed by humans (Table 1 and Supplementary Table S1). FVB/N mice are sensitive to the AOM/DSS and only require one cycle of DSS to promote tumorigenesis. Three days after the DSS treatment, mice were started on an AIN93-G control diet (18) or one supplemented with the indicated flavonol. Sensitive strains of mice treated with azoxymethane and 1 week of DSS will develop colon dysplastic lesions, adenomas, and adenocarcinomas; unfortunately some of these mice will have to be euthanized prematurely due to the large tumor burden and colorectal prolapse (15, 19). In our study, feeding mice isorhamnetin after the DSS exposure produced a significantly higher survival rate (80%) compared with the mice fed the control diet (48%, P = 0.02; Table 1). The isorhamnetin-fed mice also developed 35% (P = 0.03) fewer tumors and a 59% (P = 0.04) smaller tumor burden than mice fed the control diet (Table 1). Because the decreased mortality and tumor number in the quercetin intervention was not significant and we saw no beneficial effects with rutin or myricetin diets, we conducted an in-depth study of the isorhamnetin intervention. A repeat of the isorhamnetin intervention showed 63% decrease in tumor number (P = 0.002) and an 83% decrease in tumor burden (P = 0.02; Fig. 1). In this study, tumors did not progress to adenocarcinoma on either of the diet interventions. These data indicate that a diet supplemented with isorhamnetin prevents colorectal tumorigenesis in a mouse model of colitis-associated colorectal cancer.

FVB/N mice exposed to one cycle of DSS will develop colitis that can be detected by MRI (15). Mice treated with AOM/DSS were imaged on day 20, 25, and 33 (Fig. 1) using MRI before and after i.v. administration of 0.2 mmol/kg gadolinium diethylenetriamine-pentaacid (Gd-DTPA) contrast agent (Fig. 2). In normal colons, little to no contrast was detected in the colonic epithelium (15), whereas in DSS-treated mice, the contrast agent clearly reached the epithelial lining of the colon, indicating colitis at day 20 in both diet groups (Fig. 2A, left). By day 25, colitis increased in mice fed both diets; however, colitis was less severe in mice fed isorhamnetin (Fig. 2A, middle). By day 33, isorhamnetin-fed mice had less uptake of contrast agent in the colonic epithelium as compared with mice fed the control diet, indicating colitis had begun to resolve in these mice (Fig. 2A, right).

Histopathology analysis of the colonic tissue collected 1 day after MRI (days 21, 26, and 34) confirmed the results of the imaging (Fig. 2B). Regardless of day of sacrifice (no time × diet interaction at P < 0.20), dietary isorhamnetin decreased leukocyte distribution (P = 0.003), depth (P = 0.07), and severity (P = 0.11) of inflammation, and the distribution of AGH or squamous metaplasia (P = 0.02), resulting in an overall lower histopathology score for isorhamnetin-fed mice (P = 0.02). The severity and distribution of inflammation was similar in the mice on days 21 and 26, irrespective of the diet. By day 34, however, the mice fed the isorhamnetin diet showed only localized or multifocal distribution of leukocytes with mild to moderate severity, whereas the mice fed the control diet still exhibited moderate to severe inflammation, which was distributed diffusely within the section of colon evaluated (Fig. 2B). Colons collected at days 21 and 26 from mice fed both the control and isorhamnetin diet showed AGH and squamous metaplasia confined to the colorectal area in most sections. By day 34, the colons of the mice that were fed the control diet had a higher rate of microadenoma formation and a greater degree of hyperplasia and squamous metaplasia, which are considered premalignant lesions, as compared with the mice fed the isorhamnetin diet. Unlike in our previous study (15), no adenocarcinoma was detected in this study in either group. These results suggest that the mice on the isorhamnetin diet recovered from colitis faster than the mice on the control diet and further suggest a chemopreventive effect of isorhamnetin on colorectal tumorigenesis.

The DSS-induced inflammation resulted in infiltration of the colon by leukocytes, which was quantified by expression of the cell surface marker of leukocytes, CD45. The leukocyte distribution in the colonic mucosa and submucosal tissue was similar at days 21 and 26, irrespective of diet (data not shown). By day 34, however, dietary isorhamnetin decreased infiltration of CD45-positive cells compared with mice on the control diet (Fig. 2C). The expression of CD45 was predominantly localized to the gut-associated-lymphoid tissue (GALT) in mice fed isorhamnetin, similar to what is seen in mice not exposed to DSS. The GALT provides immune protection to the gut (20). These results suggest that isorhamnetin prevents the release of CD45-positive leukocytes. Consistent with the histopathology and MRI data (Fig. 2A and B), the decrease in CD45-positive leukocyte infiltration indicates that mice on the isorhamnetin diet recover more quickly from the colitis than do mice fed the control diet.

To assess the effects of the isorhamnetin diet on cell proliferation, Ki-67 expression was evaluated by IHC on day 34 in the colon tissue of AOM/DSS–treated mice. Mice fed the isorhamnetin diet had a 16% crypt proliferation fraction compared with mice on the control diet with a crypt proliferation fraction of 63% (P < 0.001; Fig. 3). Furthermore, Ki-67 expression in the isorhamnetin-treated mice was predominantly in the basal portion of the colon crypt that contains multipotent stem cells (21), whereas in mice fed the control diet, Ki-67 expression showed aberrant cell proliferation throughout the crypts. Similar results were seen at day 59 (Fig. 3), indicating that the mice on the control diet never completely recovered from the DSS-induced proliferation. The decrease in proliferation suggests that the mice on the isorhamnetin diet recover more quickly from the inflammation-induced cell proliferation response in the colon.

To assess the effects of isorhamnetin in a human relevant model, human colorectal carcinoma cell lines were treated with varying concentrations of isorhamnetin. Treatment of HT29 cells showed no effect on cell viability with up to 40 μmol/L isorhamnetin and very little effect with up to 100 μmol/L. HCT116 colorectal cell viability was more sensitive to isorhamnetin than viability of HT29 cells (Supplementary Fig. S2). At 40 μmol/L, isorhamnetin significantly inhibited anchorage-independent growth of HT29 cells by 62% (P = 0.001) compared with solvent control (Supplementary Fig. S2). These results suggest that isorhamnetin at 40 μmol/L can inhibit tumorigenesis without significantly affecting cell viability.

Previous studies have shown that isorhamnetin inhibits MEK and PI3K in squamous cell carcinoma skin cancer (14). Here, we show a similar dose-dependent inhibition of phosphorylation of Akt, a PI3K target, and of ERK1/2, a MEK target in HT29 colorectal cancer cells (Fig. 4A). Interestingly, we did not see inhibition of S6kinase phosphorylation, a target of Akt. The tyrosine phosphatase CD45 dephosphorylates Src tyrosine 529, leading to autophosphorylation at tyrosine 417 and activation of Src. Here, we show a dose-dependent inhibition of Src Y417 phosphorylation by isorhamnetin (Fig. 4A). Conversely, a dose-dependent increase was shown in Src Y529 phosphorylation. Src activation was inhibited by isorhamnetin in a dose-dependent manner with a 34% inhibition (increase in Y529 phosphorylation) at 40 μmol/L.

Because tyrosine 529 on Src is phosphorylated by CSK, we examined the effect of isorhamnetin on CSK. Isorhamnetin induced a dose-dependent increase in CSK mRNA (2.2-fold change at 40 μmol/L) and protein (2.6-fold change at 40 μmol/L; Fig. 4B). Isorhamnetin at 40 μmol/L also induced CSK expression in the breast cancer cell line MCF7 (data not shown). These results indicate that isorhamnetin activates gene expression of csk, leading to inhibition of Src activity.

To determine if the isorhamnetin inhibition of Src is dependent on CSK expression, we knocked down csk expression with siRNA and assessed Src activity. Transfection with the control siRNA (siCt) did not affect the dose-dependent increase of CSK protein in the isorhamnetin-treated cells, whereas knockdown of the csk gene blocked the isorhamnetin-induced increase in CSK protein (Fig. 4C). Knockdown of csk expression reversed the inhibition of Src activation by isorhamnetin. Transfection with control siRNA had no effect on the isorhamnetin-dependent decrease in Src Y417 phosphorylation. Conversely, transfection of HT29 cells with siRNA against csk reversed the isorhamnetin-induced decrease in Y417 phosphorylation on Src. In fact, siRNA knockdown of csk resulted in a slight increase in Y417 phosphorylation (Fig. 4C), showing that inhibitory effects of isorhamnetin on Src is dependent on CSK in HT29 cells.

Transfection of HT29 cells with siCSK also reversed the isorhamnetin inhibition of Akt and ERK activation seen in cells transfected with the control siRNA (Fig. 4C), indicating that activation of the MAPK and PI3K pathways is in part regulated by Src or an Src family kinase in HT29 cells. Src can activate both the ERK and the PI3K pathways (22–24). These observations show that in HT29 colorectal cancer cells isorhamnetin induces CSK expression, leading to inactivation of Src by phosphorylation at Y527 and decrease in ERK1/2 and AKT activity.

The Wnt/β-catenin pathway is activated in most colorectal cancer. β-Catenin can complex with APC and GSK3 in the cytoplasm. Perturbation of this complex releases a stabilized β-catenin. β-Catenin can also be tethered to the plasma membrane by binding with α-catenin to E-cadherin (10). We found a dose-dependent reduction in nuclear β-catenin with a corresponding accumulation of cytoplasmic β-catenin in HT29 cells treated with isorhamnetin (Fig. 5A). Interestingly, isorhamnetin did not change GSK3 activity or E-cadherin levels (Fig. 5C). Alternatively, isorhamnetin inhibited tyrosine phosphorylation at Y654 on β-catenin (Fig. 5A, nuclear). Tyrosine 654 is phosphorylated by Src, which results in release of β-catenin from its complex with E-cadherin allowing nuclear localization (10). The corresponding increase in cytoplasmic β-catenin did not show an increase in phosphorylation at Y654, consistent with a decrease in Src-regulated nuclear translocation of β-catenin. To assess whether nuclear translocation of β-catenin was Src-dependent, we transfected HT29 cells with csk siRNA. Knockdown of csk expression reversed the inhibitory effects of isorhamnetin treatment on Y654 phosphorylation and nuclear translocation of β-catenin (Fig. 5B). These results indicate that nuclear localization of β-catenin in HT29 colorectal cancer cells is a consequence of Src phosphorylation of Y654 on β-catenin and that isorhamnetin induces expression of CSK leads to inhibition of Src phosphorylation of Y654 on β-catenin, blocking its release from E-cadherin and its nuclear localization.

To determine if the effects of isorhamnetin on Src activation and β-catenin nuclear localization seen in HT29 cells can be translated to the mouse colon, we measured CSK levels, Src activity, and β-catenin levels by IHC in AOM/DSS–treated mice. Similar to the HT29 cells, mice consuming isorhamnetin showed higher levels of CSK in both the DSS-treated and untreated groups of mice compared with the mice fed the control diet (Fig. 6A). CSK expression was higher in the lamina propria surrounding the crypts compared with the epithelial tissue. CSK was also detected, although at lower levels, in normal human colon tissue. Furthermore, activated Src (phosphorylated at Y417) was higher in the crypts of AOM/DSS–treated mice fed the control diet at both day 34 and 59 (Fig. 6B). Conversely, inactivated Src (phosphorylated at Y529) was higher in crypts of the isorhamnetin-fed mice (Fig. 6B). These results indicate that isorhamnetin induces expression of CSK and inhibits Src activation in vivo in the colons of AOM/DSS–treated mice.

Nuclear β-catenin levels were increased in AOM/DSS–treated mice on the control diet, consistent with activation of Src (Fig. 6C, middle). By day 34, AOM/DSS–treated mice on the control diet showed a diffuse staining for β-catenin throughout the cells, indicating both cytoplasmic and nuclear accumulation of β-catenin. Furthermore, IHC staining with the Y654 phospho-specific β-catenin antibody indicated a high level of Src-induced phosphorylation. In contrast, in the isorhamnetin-fed mice, β-catenin remained predominantly in the cytoplasm (Fig. 6C, bottom), with lower amounts of Y654 phosphorylated β-catenin and decreased nuclear localization. These results suggest that Src regulated phosphorylation of β-catenin and that its nuclear translocation can be inhibited with an isorhamnetin-supplemented diet.

In a mouse model of colorectal cancer, we evaluated the effects of dietary isorhamnetin for colorectal cancer prevention. In this model, FVB mice treated with azoxymethane and DSS develop tumors that will progress to adenoma and adenocarcinomas (Table 1; refs. 15, 19). Dietary isorhamnetin reduced inflammation, neutrophil infiltration, cell proliferation, tumor burden, and mortality associated with the AOM/DSS treatment. Src activation and β-catenin nuclear localization induced by AOM/DSS were also reduced in the isorhamnetin-fed mice and in HT29 colon cancer cells treated with isorhamnetin. Isorhamnetin induced the expression of CSK, a negative regulator of Src. In HT29 cells isorhamnetin-induced inhibition of Src activity, and nuclear localization of β-catenin was dependent on CSK expression (Fig. 4). Isorhamnetin did not affect the expression of E-cadherin, activation of GSK3, or activation of S6Kinase. These results suggest that the anti-inflammatory and anticancer activities of isorhamnetin are linked to inhibition of oncogenic Src activity, which can phosphorylate β-catenin at Y654, leading to its dissociation from the membrane and its nuclear localization.

The Wnt/β-catenin pathway is activated in most colorectal cancer (25, 26). This pathway is also activated in the AOM/DSS mouse model (19, 27). β-Catenin can complex with α-catenin and E-cadherin at the cytoplasmic membrane, providing adherin junction communication (10). Activated Src can phosphorylate β-catenin at Y654, releasing it and α-catenin from the E-cadherin complex. Dissociation of β-catenin from the cadherin complex will cause dysregulation of tyrosine kinase signaling, will affect cell–cell communication, will affect β-catenin–regulated gene expression, and can lead to transformation and survival of β-catenin–driven cancer (10).

Members of the SFK family are nonreceptor tyrosine kinases that are recruited to the membrane by integrin or receptor tyrosine kinase–induced phosphorylation of focal adhesion kinase (FAK). Recruitment of SFK to the membrane provides a molecular switch that is important for regulating proliferation, differentiation, cell adhesion, and cell mobility (28). Activated Src in primary colorectal cancer is an indicator of poor prognosis (7), and elevated Src activity can be detected in the majority of human colon cancer (29). Inhibition of Src or SFK can enhance cell to cell adhesion and can suppress migration and invasion in vitro and metastasis in vivo, suggesting an anti-invasive role for Src inhibitors (10). Inhibitors of Src are currently being tested in clinical trials (10).

The CSK is a nonreceptor tyrosine kinase that serves as a negative regulator of Src and SFK. CSK phosphorylates the C-terminal regulatory site of SFK, resulting in a conformation change and inactivation of the kinase activity (30–33). A reduction in CSK mRNA, protein, and kinase activity in colorectal carcinoma has been shown to be correlated with an increase in Src activity, suggesting that a loss in CSK may influence transformation of colorectal carcinoma (8). Rengifo-Cam and colleagues have shown that CSK regulates signaling from integrin–SFK–mediated cell adhesion, which can influence the metastasis of cancer cells (34). Recruitment to the membrane by scaffolding proteins such as CSK-binding protein (Cbp) is required for inactivation of Src by CSK and is crucial for preventing tumorigenesis (31, 35). The fact that none of the other flavonols tested, including mycetin, quercetin, and rutin, induced expression of csk in HT29 cells (results not shown) and were not effective for inhibiting carcinogenesis in vivo is consistent with our conclusion that the chemoprotective effects of isorhamnetin in colon cancer are linked to its anti-inflammatory activities and its inhibition of oncogenic Src activity and consequential loss of nuclear β-catenin, activities that are dependent on csk expression. Although csk expression can be regulated translationally (36), very little is known about how csk is regulated transcriptionally. Further investigation is needed to learn how csk expression is lost in colorectal cancer and how isorhamnetin restores csk expression.

Isorhamnetin, quercetin, kaempferol, and myricetin are flavonols that are present in a wide variety of fruits and vegetables and have anticancer activity (37). Computational and binding assays have shown that isorhamnetin can bind directly to MEK1 and to PI3K and quercetin can bind to RSK2 (14, 38). Lee and colleagues (39) reported that kaempferol can bind to Src, and Jung and colleagues (40) showed that myricetin can bind to the SFK Fyn. Our findings show that isorhamnetin can inhibit Src activity, but that this inhibition is dependent on CSK expression, suggesting that unlike the structurally similar flavonols, kaempferol, and myricetin, isorhamnetin does not bind directly to Src. The mechanism of how isorhamnetin is upregulating CSK expression is currently under investigation. Our results and those of others (8, 34, 35) show the importance of CSK as a negative regulator of SFK and as a tumor suppressor, suggesting that preventing the loss of or restoring the expression of CSK would be beneficial for preventing tumorigenesis, tumor progression, and tumor metastasis.

No potential conflicts of interest were disclosed.

Conception and design: M.R. Young, N.H. Colburn, Y.S. Kim, G. Bobe

Development of methodology: M.R. Young, M.O. Evbuomwan, G. Bobe

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S.M. Saud, M.R. Young, Y.L. Jones-Hall, L. Ileva, M.O. Evbuomwan, G. Bobe

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S.M. Saud, M.R. Young, G. Bobe

Writing, review, and/or revision of the manuscript: S.M. Saud, M.R. Young, N.H. Colburn, G. Bobe

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S.M. Saud, M.R. Young, J. Wise, Y.S. Kim, G. Bobe

Study supervision: S.M. Saud, M.R. Young, N.H. Colburn, Y.S. Kim, G. Bobe

The authors thank Craig Driver of the Laboratory Animal Sciences Program of SAIC-Frederick, Darlene Green and Tammy Beachley of the Pathology/Histotechnology Laboratory of SIAC-Frederick, and Thomas G. McCloud from the Natural Products Support Group of Applied Developmental Research Program of SAIC-Frederick.

This study was funded by the Office of Complementary and Alternative Medicine, Office of Dietary Supplements, the Division of Cancer Prevention and the Intramural Research Program, National Cancer Institute, NIH, Department of Health and Human Services (DHHS), Bethesda, MD.

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.