Myricetin inhibits matrix metalloproteinase 2 protein expression and enzyme activity in colorectal carcinoma cells

The extracellular matrix is a complex structure supporting cells in mammalian tissues, and there are several biomolecules in the extracellular matrix including collage and casein involved in maintaining the three-dimensional structure of the body (1). During carcinogenesis, degradation of the extracellular matrix occurs in the process of metastasis of malignant tumors and is involved in the migration and invasion of malignant tumor cells (2, 3). Matrix metalloproteinases (MMP) are proteins for extracellular matrix breakdown, and activation of these protein activities has been detected in malignant tumors including colon carcinoma, lung cancer, and hepatoma (4, 5). Both MMP-2 and MMP-9 have been shown to participate in the tumor metastatic process, and the more malignant the tumors the more potent MMP-2 and MMP-9 activity is in stimulating metastasis of tumors (6, 7). Therefore, an agent capable of inhibiting the activity of MMPs, especially MMP-2 and MMP-9, may potentially be an antitumor metastatic drug.

12-O-tetradecanoylphorbol-13-acetate (TPA), a potent tumor promoter, induces a variety of cellular responses including differentiation, proliferation, and apoptosis. Protein kinase C (PKC) activation by TPA via translocation of PKC proteins from cytoplasm to membrane has been shown to be an essential step in TPA-induced tumor promotion (8). At least five isoforms of PKC including α, β, γ, δ, and ε have been isolated from different human tissues, and activation of PKCα is involved in the formation of colorectal carcinoma. After PKC activation, a series of downstream genes such as mitogen-activated protein kinases (MAPK), c-Jun, and c-Fos is stimulated. Our previous study indicated that induction of extracellular signal–regulated kinase (ERK) 1/2 phosphorylation, ornithine decarboxylasse, c-Jun, and cyclooxygenase 2 proteins is involved in TPA-induced transformation in NIH3T3 cells. Zeliadt et al. (9) indicated that TPA induced MMP-13 gene expression via an ERK-dependent pathway in mouse keratinocytes. In human cervical cancers, TPA induced MMP-9 activity via activation of PKCδ and c-jun-NH₂-kinase (JNK). Arnott et al. (10) suggested that TPA was able to up-regulate MMP-9 expression that was closely correlated with tumor development. However, the exact mechanism of TPA regulation of MMPs in colorectal carcinoma is still unclear.

Colorectal cancer is the third leading cause of cancer mortality in the Western world and increases strikingly. Early diagnosis and surgery to remove primary tumor have increased the survival rate of patients with colorectal cancer. Several studies indicated that metastasis is a major factor that causes death in patients with colorectal cancer (11). Therefore, investigators keep continuing to search for effective agents to inhibit the metastasis of colorectal cancer.

Flavonoids are benzo-γ-pyrone derivatives with different numbers of hydroxyl substitutions in the structures. Beneficial effects of flavonoids include antioxidant, antitumor, and anti-inflammation effects (12, 13). Results of structure-activity relationship studies have indicated that OH substitutions may affect the biological functions of flavonoids, and increasing OH substitution seems to enhance the antioxidant activity of flavonoids. Our recent studies also showed that flavonoids such as quercetin and wogonin induced apoptosis of tumor cells, and addition of glycoside attenuated the apoptosis-inducing activity of flavonoids (14, 15). In addition, flavones with one or no OH substitution significantly inhibit epidermal growth factor–induced proliferation in A431 (16). These data suggest that chemical structure, especially OH substitution, plays important roles in the biological activities of flavonoids. However, the structure-activity relationship of flavonoids on inhibiting MMPs activity is still undelineated. In the present study, 36 structure-related flavonoids are used to investigate the inhibitory activities on MMP activity in colorectal carcinoma cells. The results suggest that myricetin is the most potent inhibitor on MMP-2 activity and TPA-induced cellular responses in COLO 205 cells. The inhibitory mechanism by myricetin via inhibiting MMP-2 gene expression and enzyme activity, and structure-activity relationship of flavonoids on MMP-2 inhibition were elucidated.

The human colorectal carcinoma cells, including COLO 205, COLO 320HSR, COLO 320 DM, and HT 29 cells, were obtained from American Type Culture Collection (Rockville, MD). COLO 205-X cells were primary cultured from COLO 205–xenografted nude mice. These cells were grown in RPMI 1640 containing 10% heat-inactivated fetal bovine serum and maintained at 37°C in a humidified incubator containing 5% CO₂. All culture reagents were purchased from Life Technologies, Inc. (Gaithersburg, MD).

Flavone, 3-OH flavone, 5-OH flavone, 6-OH flavone, 7-OH flavone, 2′-CH₃O flavone, 3-CH₃O flavone, 5-CH₃O flavone, 6-CH₃O flavone, flavanone, 2′-OH flavanone, 4′-OH flavanone, 6-OH flavanone, 7-OH flavanone, 4′-CH₃O flavanone, 5-CH₃O flavanone, 6-CH₃O flavanone, 7-CH₃O flavanone, kaempferol, quercetin, quercitrin, myricetin, myricitrin, baicalein, baicalin, morin, wogonin, fisetin, luteolin, rutin, taxifoin, naringenin, naringin, hesperidin, hesperetin, (±)catechin, (+)catechin, (−)catechin, and H-7 (isoquinoline-5-sulfonic 2-methyl-1-piperazide) were purchased from Sigma Chemical Co. (St, Louis, MO). GF-109203X was purchased from Calbiochem Co. (La Jolla, CA). Gelatin from porcine skin and MMP-2 proteins from human fibroblast for direct enzyme inhibition assay were purchased from Sigma. Antibodies for MMP-1, MMP-2, MMP-3, and MMP-9 protein detection in Western blotting were obtained from Oncogene Research Products (Cambridge, MA). EHS-Matrigel for in vitro invasion assay was purchased from BD Biosciences (Bedford, MA).

Cell viability was assessed by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) staining as described previously (15). Briefly, COLO 205 cells were plated at a density of 10⁵ cells per well into 24-well plates. After overnight growth, cells were treated with a different concentration of flavonoids described above for 12 hours. At the end of treatment, 30 μL tetrazolium compound MTT and 270 μL fresh RPMI medium were added. The supernatant was remove and formazon crystals were dissolved in DMSO. After incubation for 4 hours at 37°C, 200 μL of 0.1 N HCl in 2-propanol was replaced per well to dissolve the tetrazolium crystals. At the end, absorbance at wavelength 600 nm was recorded using an ELISA plate reader.

Total cellular extracts (30 μg) were prepared, separated on 8% SDS-polyacrylamide minigels for poly(ADP-ribose) polymerase detection and on 12% SDS-polyacrylamide minigels for MMP-1, MMP-2, MMP-3, and MMP-9 detection, and transferred to Immobilon polyvinylidenedifluoride membranes (Millipore, Bedford, MA). The membrane was incubated overnight at 4°C with 1% bovine serum albumin and at room temperature for 1 hour and then incubated with indicated antibodies for a further 3 hours at room temperature followed by incubation with alkaline phosphatase–conjugated anti-mouse IgG antibody for 1 hour. Protein was visualized by incubating with the colorimetric substrate nitroblue tetrazolium and 5-bromo-4-chloro-3-indolyl-phosphate as described in our previous article (14).

COLO 205 cells at logarithmic growth phase were treated with 100 or 200 μmol/L of myricetin and myricitrin in serum-free RPMI medium for 24 hours. The conditioned media were collected and analyzed for MMPs using gelatin zymography (17). In brief, samples were mixed with SDS sample buffer without heating or reduction and applied to 10% polyacrylamide gels copolymerized with 0.1% gelatin (from porcine skin). The loading volume of each conditioned medium sample was normalized according to the cell number. After electrophoresis, gels were washed for 1 hour at room temperature in buffer containing 2.5% (v/v) Triton X-100 in 50 mmol/L Tris-Cl (pH 7.4) to remove SDS. Gels were then incubated at 37°C in reaction buffer (50 mmol/L Tris-Cl, pH 7.4, with 5 mmol/L CaCl₂, 1 μmol/L ZnCl₂, and 0.02% NaN₃) for 18 hours. Then the gels were stained with Coomassie brilliant blue R-250 (0.25%) in 10% acetic acid/20% ethanol and destained in the same solution without dye. The zones of gelatinolytic activity were detected as clear bands against a blue background and were quantified using a densitometer.

MMP-2 (400 ng) from human fibroblasts were incubated with 0, 25, 50, 100, and 200 μmol/L of myricetin in 50 μL Tris buffer (50 mmol/L, pH 7.4) at 37°C for 30 minutes, respectively. Twenty microliters of incubated buffer were then analyzed for gelatinolytic activity as described above. The zones of gelatinolytic activity were detected as clear bands against a blue background and were quantified using a densitometer.

In vitro invasion assay was carried out by the method of Huang et al. (18). Briefly, 24-well Transwell units with 8 μmol/L porosity polycarbonate filters (Becton Dickinson, Franklin Lakes, NJ) were coated with 0.1 mL of 0.8 mg/mL Englebreth-Holm-Swarm sarcoma tumor extract (EHS Matrigel) at room temperature for 1 hour to form a genuine reconstituted basement membrane. COLO 205 cells (2 × 10⁵ cells per 0.4 mL RPMI) were placed in the upper compartment, and TPA (25 ng per 0.4 mL RPMI) was added in the lower compartment. The indicated Transwell plates were then incubated at 37°C for 48 hours in a humidified atmosphere with 5% CO₂. The cells migrated on the lower surface were quantified by MTT assay and observed under light microscope by Giemsa staining. Each treatment was assayed in triplicate, and two independent experiments were done.

Values are expressed as the mean ± SE The significance of the difference from the respective controls for each experimental test condition was assayed by using Student's t test for each paired experiment. A P value < 0.01 or < 0.05 was regarded as indicating a significant difference.

Thirty-six structure-related flavonoids including flavones, flavanones, and catechins were used to investigate the inhibitory effect on MMP-2 enzyme activity in human colorectal carcinoma cells COLO 205. These compounds possess a basic benzo-γ-pyrone in structure with different substitutions such as OH, OCH3, and glycosides at different locations. Among them, flavones contain a double bond at C2-C3 of flavanones, and catechin possesses five OH substitutions at C3′, C4′, C3, C5, and C7 without an oxo group at C4 of flavanone. To investigate the inhibitory effect of tested compounds on MMP activity, release of MMPs in the medium was done by gelatin zymography. In COLO 205 cells, a detectable endogenous MMP-2, but not MMP-9, enzyme activity in the medium was examined in the absence of tested compounds. As elucidated in Table 1, myricetin at 200 μmol/L exhibited 90% inhibition of activity of MMP-2 enzyme released from COLO 205 cells. Kaempferol (200 μmol/L) also showed 47% inhibition of released MMP-2 enzyme activity in the medium. Other flavonoids examined in the present study exhibited <20% inhibition of released MMP-2 enzyme activity.

Results of MTT assay showed that flavone, 3-OH flavone, 2′-OCH₃ flavone, 2′-OH flavanone, 4′-OH flavanone, 6-OH flavanone, quercetin, wogonin, luteolin, and baicalein, at 200 μmol/L, showed significant cytotoxicity (>50%) in COLO 205 cells. Myricetin and kaempferol at 200 μmol/L exhibited 20% and 30% reduction in the viability of COLO 205 cells, respectively. Morphologic observations and DNA integrity assay showed that myricetin did not change the morphology of COLO 205 cells and no DNA laddering effect was detected in myricetin-treated COLO 205 cells (data not shown). These data suggest that myricetin is the most effective inhibitor among 36 flavonoids examined on the activity of MMP-2 released from COLO 205 cells.

Previous data suggested that myricetin exhibited MMP-2 inhibitory effect without severe cytotoxicity in COLO 205 cells. Therefore, several colorectal carcinoma cells including COLO 320HSR, COLO 320DM, HT-29, and COLO 205-X cells were used to investigate the MMP-2 inhibition by myricetin. Figure 1A shows the structures of myricetin and its related glycosylated compound, myricitrin. Myricitrin contains a rhamnose at C3 of myricetin. Myricetin, but not myricitrin, dose-dependently inhibits MMP-2 enzyme activity released from COLO 205, COLO 320HSR, COLO 320DM, HT-29, and COLO 205-X cells (Fig. 1B). The IC₅₀ values of myricetin on MMP-2 activity in COLO 205, COLO 320HSR, COLO 320DM, HT-29, and COLO 205-X cells are 7.82, 11.18, 11.56, 13.25, and 23.51 μmol/L, respectively. MTT assay showed that myricetin only at 200 μmol/L exhibited slight but significant reduction on the viability of COLO 205 cells. These data suggest that myricetin inhibition of MMP-2 enzyme activity was not due to its cytotoxic effect on cells.

Two possibilities might be involved in myricetin inhibition of activity of MMP-2 enzyme released from COLO 205 cells: one is to block MMP-2 protein expression and the other is to directly inhibit MMP-2 enzyme. Results of Western blotting indicated that myricetin inhibited endogenous MMP-2 protein expression in COLO 205 cells without affecting that of MMP-1, MMP-3, and MMP-9 proteins. In contrast, myricitrin did not affect any MMP protein expression in COLO 205 cells (Fig. 2A). Furthermore, whether myricetin or myricitrin might directly affect MMP-2 enzyme was examined by direct incubation of different concentrations of myricetin or myricitrin with the medium taken from that culturing COLO 205 at 37°C for 30 minutes, and MMP-2 activity was examined by gelatin zymography. It indicated that myricetin, but not myricitrin, directly inhibited MMP-2 enzyme activity in vitro (Fig. 2B). These data suggested myricetin inhibition of activity of MMP-2 enzyme released from COLO 205 cells might be through both decreasing MMP-2 protein expression and/or suppressing MMP-2 enzyme activity.

To elucidate if myricetin directly blocked MMP-2 activity, a purified MMP-2 protein isolated from human fibroblasts was incubated with or without indicated doses (25, 50, 100, 200 μmol/L) of myricetin or myricitrin at 37°C for 30 minutes, and MMP-2 enzyme activity was analyzed by gelatin zymography. Results shown in Fig. 2C and D indicated that purified MMP-2 protein was able to digest gelatin at three clear zones, and myricetin, but not myricitrin, concentration-dependently inhibited MMP-2 enzyme activity, as indicated by a reduction in gelatin digestion in the gel. Results from Western blot analysis using a specific MMP-2 antibody indicated that neither myricetin nor myricitrin caused a decrease in MMP-2 protein level. These data provide direct evidence suggesting that myricetin directly inhibits MMP-2 activity.

We further investigated if myricetin blocked TPA–induced MMP-2 enzyme activity. TPA is a potent tumor promoter in carcinogenesis. As shown in Fig. 3A, TPA time-dependently induced both MMP-2 protein expression and enzyme activity released from COLO 205 cells by Western blotting and gelatin zymography. An increase in c-Jun, but not MMP-1, MMP-3, and MMP-9, protein was observed in a time-dependent manner under TPA treatment. Furthermore, PKCα translocation from cytosol to membrane and activation of ERK1/2, but not p38 and JNKs, were detected in TPA-treated COLO 205 cells by Western blotting. These data revealed events including induction of MMP-2 protein expression and enzyme activity and increases in phosphorylated ERK1/2 protein and c-Jun protein expression with translocation of PKCα from cytosol to membrane in TPA-treated COLO 205 cells.

To elucidate if phosphorylation of ERK1/2, translocation of PKCα from cytosol to membrane, and induction of c-Jun protein are essential events in TPA-induced MMP-2 enzyme activity, pharmacologic studies using specific inhibitors including PD98059 (mitogen-activated protein/ERK kinase inhibitor), SB203580 (p38 inhibitor), SP600120 (JNK inhibitor), H-7 (PKA and PKC inhibitor), and GF-109203X (specific PKC inhibitor) were done. The results indicated that TPA treatment induced significant morphologic changes in COLO 205 cells by microscopic observation with (Fig. 4, bottom) or without (Fig. 4, top) Giemsa staining. Addition of PD98059, H-7, and GF-109203X, but not SB203580 and SP600120, showed a significant inhibition of TPA-induced morphologic changes in COLO 205 cells (Fig. 4A). In addition, PD98059, GF-109203X, and H 7, but not SB203580 or SP600120, inhibit TPA-induced MMP-2 protein expression with a decrease in activity of MMP-2 enzyme released from COLO 205 cells (Fig. 4B). Moreover, PD98059, GF-109203X, and H 7 significantly blocked TPA-induced ERK1/2 protein phosphorylation and c-Jun protein expression (Fig. 4C). These data suggest that activation of ERK1/2 located downstream of PKC activation was followed by inducing c-Jun protein expression, and these events occurred upstream of MMP-2 activation induced by TPA.

The possibility that myricetin inhibited TPA-induced MMP-2 activity was further examined. Results shown in Fig. 5A indicated that myricetin significantly inhibited TPA-induced activity of MMP-2 enzyme released from COLO 205 cells. In addition, TPA-induced PKCα translocation, ERK1/2 protein phosphorylation, and c-Jun and MMP-2 protein expression were significantly blocked by myricetin, but not by myricitrin (Fig. 5B). Addition of low concentrations of PD98059 (1, 2, and 4 μmol/L) or GF-109203X (0.5, 1, and 2 μmol/L) enhanced the MMP-2 inhibition of myricetin (5 μmol/L) under TPA treatment (Fig. 5C). These data provided additional evidence that myricetin inhibited TPA-induced MMP-2 activity via blocking intracellular signaling transduction.

We further investigated if myricetin blocked TPA-induced invasion in COLO 205 cells. In vitro invasion assay using a Transwell chamber coated with a reconstituted basement membrane (EHS Matrigel) was done to examine the anti-invasive effect of myricetin. In this assay system, COLO 205 cells were cultured on the upper chamber in the presence or absence of myricetin or myricitrin (100 and 200 μmol/L), and TPA (100 ng/mL) was added to the lower chamber. After 24 hours incubation, COLO 205 cells in the lower chamber were observed microscopically with (Fig. 6A, bottom) or without (Fig. 6A, top) Giemsa staining, and the number of viable cells in the lower chamber was quantified by MTT assay (Fig. 6B). Results shown in Fig. 6A indicated that COLO 205 cells did not penetrate the EHS-coated filter in the absence of TPA in the lower chamber (control group). Addition of TPA induced invasion of COLO 205 cells from the upper to the lower chambers in the absence of myricetin or myricitrin. Addition of myricetin, but not myricitrin, significantly reduced TPA-induced invasion in COLO 205 cells. Results of quantification of cells in the lower chambers showed that myricetin but not myricitrin inhibited TPA-induced invasion in COLO 205 cells (Fig. 6B).

Results of the present study show that myricetin significantly inhibits MMP-2 activity and tumor invasion in colorectal carcinoma cells. Myricetin inhibition of MMP-2 activity was via a decrease in MMP-2 protein expression and a direct inhibition of MMP-2 enzyme activity. In the presence of TPA, myricetin suppressed PKC activation, ERK1/2 protein phosphorylation, and c-Jun protein expression together with a reduction in MMP-2 protein expression and enzyme activity. Inhibition of TPA-induced tumor invasion by myricetin was shown by in vitro invasion assay in the present study. These results suggest that myricetin, a compound capable of blocking both MMP-2 protein expression and enzyme activity, is a potentially effective antitumor agent.

Flavonoids are known to possess several biological functions including antitumor, antioxidant, and anti-inflammation activities. Structure-activity analysis has shown that the number and location of OH substitutions are essential for the biological effects of flavonoids (19, 20). Results of the present study indicated that myricetin significantly inhibited MMP-2 activity. However, quercetin, which is a potent antioxidant and an apoptotic inducer, did not suppress MMP-2 activity. In contrast to quercetin, myricetin contains an additional OH substitution at C4′ of the quercetin structure. In addition, an OH group at C4′ is also found in kaempferol and luteolin, and both showed about 50% and 13% inhibition of MMP-2 enzyme activity, respectively. These results suggest that OH at C4′ of flavone is important for MMP-2 inhibition.

In addition, glycoside has been shown to affect the biological functions of flavonoids via increasing their hydrophilic properties. Our previous studies showed that rutinoside and rhamnoglucoside attenuated apoptosis-inducing activity and nitric oxide synthesis inhibition of flavonoids (14, 15, 21). In the present study, all the tested glycosylated compounds including quercitrin, myricitrin, baicalin, rutin, naringin, and hesperedin did not inhibit activity of MMP-2 enzyme released from COLO 205 cells. A significant difference was found in myricetin and its glycoside myricitrin. Myricitrin, which contains a rhamnoside at C3 of myricetin, did not inhibit endogenous and TPA-induced MMP-2 enzyme activity and invasion. These data support the notion that glycosides are negative moieties on MMP-2 inhibition and invasion in colorectal carcinoma cells.

It has been shown that flavonoids containing more hydroxyl substitutions exhibit stronger antioxidant activity (22, 23). Wang and Joseph (24) indicated that OH groups at C3′ and C4′ of the B ring and a 2,3 double bond in conjugation with a 4-oxo group in the C ring, along with the polyphenolic structures, were crucial for the protection from H₂O₂-induced oxidative stress. Cos et al. (25) indicated the OH groups at C5 and C7 and a double bond between C2 and C3 were important for the inhibitory activity of flavonoids on xanthine oxidase activity. In contrast to OH substitutions, methyoxyl (OCH₃) addition inactivated both antioxidant and prooxidant activities of the flavonoids, and hydroxylation but not methyoxylation at C2 of flavone showed high affinity on benzodiazepine receptors (26, 27). To elucidate the relationship between the antioxidant activity and MMP-2 inhibition of flavonoids, anti-1,1-diphenyl-2-picrylhydrazyl radical activity of indicated flavonoids was examined. It indicated that flavonoids with more than two OH substitutions, such as quercetin, kaempferol, baicalein, myricetin, myricitrin, and toxifolin, exhibited potent anti-1,1-diphenyl-2-picrylhydrazyl radical activity (data not shown). However, results of the present study indicated that myricetin but not others exhibited significant inhibition on MMP-2 enzyme activity. This suggests that myricetin inhibition of invasion and MMP-2 enzyme activity is not only due to its antioxidant activity but also to its suppression of MMP-2 protein expression and enzyme activity via blocking intracellular signaling transduction processes.

TPA, a tumor promoter, has been used extensively in induction of tumor progression in vitro and in vivo. Activation of PKCs via translocation of PKCs from cytosol to membrane has been identified under TPA treatment (28). In addition to PKCs, phosphorylation of MAPKs, including ERK1/2, p38, and JNK, and induction of c-Jun gene expression were detected in TPA-induced proliferation (8, 29). Hah and Lee (30) reported that TPA induced the invasion of the hepatocellular carcinoma cells through MMP-9 but not MMP-2 secretion. However, the effect of TPA on MMP-2 enzyme activity and invasion in colorectal carcinoma cells is still undefined. Results of the present study indicated that TPA induced invasion and MMP-2 protein expression in COLO 205 cells. Sequential induction of PKCα translocations, ERK1/2 protein phosphorylation, and c-Jun protein expression were involved. Myricetin inhibited TPA-induced MMP-2 enzyme activity and invasion via blocking activated signaling cascades. These data suggested that addition of TPA was able to induce invasion via activation of MMP-2 enzyme activity in colorectal carcinoma cells, which was blocked by myricetin (Fig. 7).

Several natural products such as curcumin, epigallocatechin, and resveratrol have been found to inhibit MMP activity and invasion in different assay systems (31–33). The mechanism of inhibition was attributed to blockade of MMP gene expression but not enzyme activity. In the present study, myricetin inhibited both endogenous and TPA-induced MMP-2 enzyme activity and invasion in colorectal carcinoma cells. The most important event was that myricetin inhibition of MMP-2 released from cells was through blocking both MMP-2 protein expression and enzyme activity. The inhibition by myricetin of MMP-2 protein expression may be due to interruption of signaling transduction cascades such as PKC translocations and ERK1/2 proteins phosphorylation. However, the reason for myricetin-inhibited MMP-2 enzyme activity is still unclear. Several previous studies showed the metal-chelating activity of flavonoids (34, 35). Therefore, it is suggested that a direct inhibitory effect of myricetin on MMP-2 enzyme activity may be through chelating metals such as zinc and calcium, which are essential for the activity of MMP-2.

In conclusion, we have provided evidence demonstrating that myricetin inhibits invasion and both MMP-2 protein expression and enzyme activity in colorectal carcinoma cells in both unstimulated and TPA-stimulated conditions. Therefore, myricetin is potentially a useful anti-invasive agent in the treatment of colorectal carcinoma.