Polyphenols are distributed widely in the plant kingdom, and are structurally diverse (Freidman and Jugens, 2000). They are plentiful in certain plants but not in others (Paganqa et al, 1999; Leontowicz et al, 2002; Mouria et al, 2002). For example, tannic acid is not found in tea but is found in apples (Graham, 1992; Kanda et al, 1998). The polyphenol content and composition are affected by various factors, such as the plant variety, growth conditions, and manufacturing processes (Graham, 1992; Astill et al, 2001; van der Sluis et al, 2001). In this respect, green tea is rich in catechins, whereas black tea is rich in theaflavins (Graham, 1992).

The most studied of these compounds are the tea polyphenols, particularly catechins, and there are many reports regarding the antitumor activities of epigallocatechin-3-gallate (Blanko et al 2003; Gupta et al, 2003). Other polyphenols, such as curcumin, rutin, quercetin, and trans-reveratrol, have also been reported to be chemopreventive through their anti-proliferative, anti-metastatic, and/or anti-invasive properties (Menon et al, 1995; Maeda-Yamamoto et al, 1999; Menon et al, 1999; Caltagirone et al, 2000; Mouria et al, 2002).

Immature apples have been reported to contain large amounts of several types of polyphenol, which include chlorogenic acid, catechin, epicatechin, rutin, and condensed tannins (Kanda et al, 1998). Although the biological activities of these apple-derived compounds, especially as they pertain to the control of tumor progression, remain to be determined, a crude polyphenol fraction has been reported to inhibit histamine release from RBL-2H3 cells and rat mast cells (Kanda et al, 1998).

The apple polyphenol crude fraction (5% in solution) was kindly supplied by Nikka Whisky Distilling Co. Ltd. (Chiba, Japan). Dulbecco's modified Eagle's medium mixture F-12 Ham (DME/F-12), RPMI-1640, insulin, transferrin, and gelatin were obtained from Sigma Chemical Co. (St. Louis, MO, USA). Trypsin and LPS (E. coli 055:B5) were obtained from DIFCO Laboratories (Detroit, MI, USA). Fetal bovine serum (FBS), bovine fibronectin, gelatinase zymography standards (human MMP-2 and -9), recombinant mouse IFN-g , Perfect Protein^TM Markers, ISOGEN^®, RNA PCR kit, AmpliTaq Gold with Gene Amp 10´ PCR Gold Buffer, agarose-LE, and the DIG Oligonucleotide Tailing Kit were purchased from JRH Biosciences (Lenaxa, KS, USA), Biomedical Technologies Inc. (Stoughton, MA, USA), Chemicon International Inc. (Temecula, CA, USA), Novagen Inc. (Madison, WI, USA), Genzyme Corp. (Cambridge, MA, USA), Nippon Gene (Tokyo, Japan), Takara Biochemicals, Inc. (Tokyo, Japan), Applied Biosystems (Tokyo, Japan), Nacalai Tesque (Kyoto, Japan), and Roche Diagnostics Co. Ltd. (Tokyo, Japan), respectively. All of the other reagents were purchased from Wako Pure Chemical Industries Ltd. (Osaka, Japan).

The degree of r/m HM-SFME-1 cell metastasis in the lungs was estimated by quantifying the levels of the human c-Ha-ras 1 gene in the cells, as described previously (Matano et al, 1995), with minor modifications. In brief, the DNA samples were amplified for 35 cycles of 94^oC for 30 s, 60^oC for 30 s, and 72^oC for 1min. The 72^oC incubation period was extended to 10 min in the last cycle. The PCR products were electrophoresed in a 3% agarose-LE gel, and transferred to a positively charged nylon membrane (Roche Diagnostics). The transferred PCR products were hybridized with probes that had been labeled with the DIG Oligonucleotide Tailing Kit. The membrane was washed and exposed in the Luminescence Image Analyzer (Fujifilm LAS-1000; Fuji Photo Film Co. Ltd., Tokyo, Japan), and the intensities of the PCR products were measured. The 123-bp PCR product of the human c-Ha-ras 1 gene was used as a probe for this gene, since the total region rather than partial regions of the gene was required to detect the PCR products. In the case of the human c-Ha-ras 1 gene, the upstream (sense) and downstream (antisense) primers were 5'-ATgACggAATATAAgCTggT-3' and 5'-CgCTAggCTCACCTCTATA-3', which correspond to nucleotide positions 1670-1689 and 1773-1792, respectively. The standard used for estimation of the metastasized tumor cell numbers was derived from the DNA samples from normal lungs that contained 10² to 10⁶ r/m HM-SFME-1 cells.

The invasiveness of r/m HM-SFME-1 cells was assayed using 8.0-m m pore size polyvinylpyrrolidone-free polycarbonate filter chambers (Chemotaxicell^®, Kurabo Industries, Ltd. Osaka, Japan). The filter had been previously coated on the reverse side with fibronectin (10 m g/filter) and dried at 37^oC for 2 days. Gelatin (25 m g/filter) was applied to the front side of the filter, which was then dried at 37^oC for 1 day. The gelatin coating was applied twice, and fibronectin (10 m g/filter) was applied subsequently. The cell suspension (100 m l of 1 ´ 10⁶ cells/ml) in F/D medium containing 5% FBS and 100 m l apple polyphenol solution (0-50 m g/ml) was plated onto the Chemotaxicells, which were placed in 24-well microplates, and 0.6 ml of F/D medium containing 5% FBS and the apple polyphenol solution (0-50 m g/ml) was immediately added to the outer well of the microplates. The chambers were incubated at 37^oC. The culture medium was removed after 3 days, and the cells on the gelatin-coated (front) side of the filter were removed by wiping with a cotton swab. The cells on the fibronectin-coated (reverse) side of the filter were collected by rinsing with phosphate-buffered saline that contained 0.1% trypsin, and were counted with a Coulter counter (model Z_BI; Coulter Electronics Inc., Hialeah, FL, USA).

After a 2 h pre-incubation of the cells (10⁶) in 6-cm-diameter culture dishes that contained 2 ml of culture medium, 10 m l of the apple polyphenol crude fraction (0.1-10 mg/ml) was added, and the dishes were incubated for 24 h at 37^oC. Cell-free culture supernatants were prepared. The culture supernatants of the r/m HM-SFME-1 cells, but not those of the NIH3T3 and J774.1 cells, were concentrated 30-fold using an Ultrafree-CL concentrator (Amicon, UFC4LCC25; Millipore Corp., Bedford, MA, USA). Aliquots of the culture supernatants were electrophoresed in a 10% polyacrylamide gel that contained sodium dodecyl sulfate and 0.1% gelatin. After electrophoresis, the gel was washed with 2.5% Triton-X100 and incubated at 37^oC for 20 h in 100 mM Tris-HCl (pH 8.0) that contained 5 mM CaCl₂, 0.005% polyoxyethylene lauryl, and 0.001% sodium azide. The gel was then stained with Quick-CBB^® that contained 1% Coomassie brilliant blue R. Perfect^TM Protein Markers were used as the molecular mass standards, and gelatinase zymography standards for human MMP-2 and MMP-9 were used as the positive controls for MMP-2 and MMP-9, respectively.

The cDNA was prepared as described previously (Ryoyama et al, 2003). Aliquots of the cDNA samples were subjected to PCR using AmpliTaq Gold with the Gene Amp 10´ PCR Gold Buffer that contained 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 2.5 mM MgCl₂, 0.2 m M of each primer (sense and antisense), 0.2 mM of each dNTP, and 25 units/ml Taq polymerase.

The primers for mouse MMP-2, MMP-9, TIMP-1, and VEGF were synthesized according to the GenBank sequences. The primers for MMP-2 were derived from the sequence with accession no. M84324: forward primer, 5'-ggCCATgCCATggggCTg-3' (nucleotides 1259-1276) and reverse primer, 5'-CCAgTCTgATTTgATgCTTC-3' (nucleotides 2001-2020). The MMP-9 primers were derived from the sequence with accession no. D12712: forward primer, 5'-gggCAACTCggCAggAgAgC-3' (nucleotides 1044-1063) and reverse primer, 5'-CCAggTgACgggCTgCTTgT-3' (nucleotides 1513-1532). The TIMP-1 primers were derived from the sequence with accession no. X04684: forward primer, 5'-CTgTgCCCCACCCCACCCAC-3' (nucleotides 184-203) and reverse primer, 5'-AAggCTTCAggTCATCgggC-3' (nucleotides 716-735). The VEGF primers were derived from the sequence with accession no. NM009505: forward primer, 5'-CACgACAgAAggAgAgCAgAAgTC-3' (nucleotides 79-119) and reverse primer, 5'-gCCATCATCgTCACCgTTgA-3' (nucleotides 742-761). The mouse beta-actin gene was used as the internal control with the forward primer, 5'-gTgggCCgCTCTAggCACCAA-3' (nucleotides 25-45) and the reverse primer, 5'-CTCTTTgATgTCACgCACgATTTC-3' (nucleotides 541-564).

The levels of NO in the culture supernatants were determined with the Griess reagent, as described previously (Ryoyama et al, 1993).

The effects of the crude fraction of the apple polyphenol on spontaneous metastasis of r/m HM-SFME-1 cells in the lung were assayed according to the method of Matano et al. (Matano et al, 1995) with minor modifications. At 39 days after tumor implantation, the number of tumor cells in the lungs was estimated to be between 1.5 ´ 10⁴ and 3.0 ´ 10⁶ cells/lung (Figure 1A). Oral administration of the apple polyphenol fraction significantly inhibited tumor metastasis; the estimated number of tumor cells after treatment was between 1.0 ´ 10² and 3.9 ´ 10⁵ cells/lung. A lower dose of the fraction (0.05%) also tended to inhibit lung metastasis, although this level of inhibition was not statistically significant (data not shown). In addition, since the mice drank 4.4 ± 0.5 ml/day of fluid, each mouse had an approximate intake of 22 ± 2.5 mg/day of the apple polyphenol crude fraction.

Since the apple polyphenol fraction scarcely affected the growth of r/m HM-SFME-1 cells, its effect on the in vitro invasion and migration of these cells was examined. It has been reported that r/m HM-SFME-1 cells are highly metastatic in the lungs of mice (Matano et al, 1995). Indeed, r/m HM-SFME-1 cells could be detected in the lungs seven days after the injection of 10⁵ to10⁶ cells into the footpad (unpublished data). This result suggests that these cells are highly invasive in the in vitro invasion model. Thus, we attempted to clarify the invasive activities of the r/m HM-SFME-1 cells. A preliminary experiment showed that 10% to 50% of the input cells had invaded three days after incubation. Since the rates of invasion varied from experiment to experiment, the effects of the apple polyphenol crude fraction on invasion are expressed as percentages of the mean number of invading cells in the control group for each experiment. Figure 2A shows the effect of the crude fraction on r/m HM-SFME-1 invasion. The fraction dose-dependently inhibited cell invasion, and the level of inhibition at 50 m g/ml crude fraction was significant, at 40-50% of the control level of invasion. This inhibition was not due to cytotoxicity, since the fraction was not cytotoxic at these concentrations (data

Our preliminary experiments showed that r/m HM-SFME-1 cells produced MMP-9 at low levels, but did not produce MMP-2, and that LPS augmented MMP-9 gene expression. Moreover, Wang et al. (Wang et al, 2002) have reported that fibroblasts promote breast cancer cell invasion by stimulating MMP-9 synthesis. Therefore, we examined the effects of the apple polyphenol crude fraction on MMP-9 activity and mRNA expression in r/m HM-SFME-1 cells, in the presence and absence of LPS and in the conditioned medium from a culture of the fibroblast cell line NIH3T3. Figure 3A shows that the fraction scarcely affected the MMP-9 activities of the tumor cells. The MMP-9 activity in the presence of the conditioned medium was not tested because the conditioned medium itself had potent MMP-9 activity.

Since tumor growth sites appear to consist of both tumor cells and inflammatory cells, increases in footpad thickness reflect not only tumor growth but also inflammation. Many reports have indicated a close correlation between tumor progression and in situ inflammation; tumor invasion and metastasis appear to be affected by soluble factors and free radicals that are produced locally by host tissue cells (fibroblasts, endothelial cells, macrophages) (Lala and Chakraborty, 2001; Coussens and Werb, 2002; Wang et al, 2002). Furthermore, in the majority of human and experimental tumors, NO appears to stimulate tumor growth and metastasis by enhancing the invasive, angiogenic, and migratory capacities of the tumor cells (Lala and Chakraborty, 2001). In fact, the spontaneous metastasis of r/m HM-SFME-1 in the lung is blocked by inhibitors of NO production (manuscript in preparation). The apple polyphenol crude fraction inhibited slightly but not significantly the increase in thickness of the footpads into which r/m HM-SFME-1 cells had been implanted (Figure 1B), and did not inhibit in vitro tumor growth (data not shown). Furthermore, the fraction inhibited the augmentation of MMP-9 gene expression caused by the addition of fibroblast-conditioned medium (Figure 3B). These results suggest that the apple polyphenol crude fraction inhibits inflammation.

Plasminogen activation systems have been implicated in extracellular matrix degradation (Sidenius and Blasi, 2003). One of these systems, the urokinase-type plasminogen activator (uPA), has been clearly implicated in cancer progression, particularly invasion and metastasis (Andreasen et al, 2000; Rabbani and Mazar, 2001). Therefore, we examined the effect of the apple polyphenol crude fraction on uPA production by the r/m HM-SFME-1 cells. The addition of 0.5-50 m g/ml of the fraction did not affect uPA production (data not shown). On the other hand, the crude fraction augmented significantly the uPA activity of J774.1 cells, whereas that of the NIH 3T3 cells was scarcely affected (data not shown).