Digestive Diseases and Sciences, Vol. 36, No. 11 (November 1991), pp. 1606-1610

Effects of Fat and Fiber on Human Colon Cancer Xenografted to Athymic Nude Mice THOMAS J. McGARRITY, MD, LAURIE P. PEIFFER, BS, SCOTT T. KRAMER, BA, and JILL P. SMITH, MD

The effects o f unsaturated fat and fiber (cellulose) on the growth o f human colon cancer explanted to athymic nude mice was evaluated. Eighty-seven male nude mice bearing xenografts of human HT29 or WiDr colon cancer were divided into three groups o f equal weight and tumor volume. Each group was fed one o f three diets: normal fat~no fiber (N/N), high fat~no fiber (H/N) or high fat~high fiber (H/H). To equalize caloric intake, animals in the H/N group received 4 g of food per day and the other animals were f e d 5 g o f food per day. At sacrifice tumor volume and weight was recorded, and tumors were analyzed for protein and DATA content and ornithine decarboxylase activity. Tumor volume, weight, and protein were greater in the H/N group compared to the N / N group for both colon cancer cell lines. Tumor DNA content was greater in the HT29 H / N group compared to the N / N group (P < 0.05) and tumor ornithine decarboxylase activity in the WiDr H/N group was greater than the N / N animals (P < 0.002). The tumor growthpromoting effects of the high unsaturated fat diet were attenuated by the addition o f fiber. Animal weight was higher in the H/N group compared to the N / N and H/H groups. This study suggested that a high-fat diet stimulated and fiber decreased the growth o f human colon cancer explanted to athymic nude mice. The growth-promoting effects o f a high-fat diet in colorectal cancer may be due in part to a circulating trophic factor since these tumors were remote from the large intestine. KEY WORDS: unsaturated fat; fiber; human colon cancer; growth.

Epidemiological data have provided clues to the etiology of colorectal cancer, which is endemic in Western society. Analysis of inter- and intracountry differences in the incidence of colorectal cancer implicate dietary determinations of this disease. A diet high in fat and low in fiber is believed to act as a promoter of large bowel carcinogenesis in suscepManuscript received October 18, 1990; revised manuscript received May 1, 1991, accepted May 23, 1991. From the Department of Medicine, University Hospital, Milton S. Hershey Medical Center, Penn State University, Hershey, Pennsylvania. The study was support in part' by grants IR29 CA45468(TJM) and CA50303(JPS) from the National Institutes of Health, and by the Research services the Veteran's Administration. Address for reprint requests Dr. Thomas J. McGarrity, Department of Medicine, Division of Gastroenterology, The Milton S. Hershey Medical Center, Penn State University, Hershey, Pennsylvania 17003.

1606

tible individuals (1, 2). A high-fat diet has been shown to increase the incidence of carcinogeninduced large bowel tumors in rodents (3-6). Several mechanisms for the tumor-promoting effects of fat have been proposed. These include alterations in fecal cholesterol metabolites and free fatty acids toxic to colonic epithelium and changes in gut flora and increased prostaglandin content (3-6). A high-fat diet also increased colonic crypt cell proliferation as measured by increased ornithine decarboxylase activity, the major enzyme of polyamine biosynthesis (7). Dietary fiber is believed to be protective against the development of colorectal cancer by increasing fecal bulk, thereby diluting and also binding potential carcinogens. Increased dietary fiber also deDigestive Diseases and Sciences, Vol. 36, No. lJ (November 1991)

0163-2116/91/I100-1606506.50/09 1991 Plenum PublishingCorporation

FAT, FIBER, AND CANCER creases transit time, thereby reducing the exposure time of colonic epithelium to toxic agents. A protective role o f fiber has been shown in some animal models of large bowel carcinogens but not in others (8-11). The study o f effects of diet on large bowel carcinogenesis has concentrated on the changes in luminal contents with different diets and their effects on the colonic epithelium. A high-fat diet also has been implicated as a risk factor for other malignancies outside the gastrointestinal tract, such as breast (12). We examined the effects of unsaturated fat and fiber (cellulose) on the growth o f two human colon cancers, HT29 and WiDr xenografted subcutaneously to nude mice. MATERIALS AND METHODS Male athymic nude mice, ages 5-6 weeks were obtained from the National Cancer Institute. Four to six animals per cage were housed in sterile flexible film isolation with autoclaved bedding, food, and water. HT29 and Wi.Dr human colon cancer ceils were obtained from American Type Culture Collection (RockviUe, Maryland). HT29 ceils were grown in 25-cm2 Coming tissue culture flasks Containing McCoy's medium supplemented with 10% fetal calf serum, and penicillin (100 units/ml) and streptomycin (100 ~g/ml) at 37 ~ in humidified air containing 5% CO2. WiDr cells were grown in Leibovitz L-15 medium supplemented with 10% fetal calf serum, penicillin (100 units/ml), streptomycin (100 ~g/ml) and L-glutamine ( 2 raM). Cells were harvested with a solution containing 0.25% (w/v) trypsin and 0.1% (w/v) ethylenediaminetetraacetic acid, and a single cell suspension was made in medium. Fifty-eight mice received a single subcutaneous dorsal injection of 10 x 106 viable HT29 ceils and 60 mice received 5 x 106 WiDr ceils in a volume of 0.5 ml, Until tumors became palpable, animals were fed standard autoclaved Wayne Blox rodent chow (Allied Mills Laboratories). By the twelfth day, 51 of 58 (80%) mice had palpable and measurable HT29 xenografts and on day 14, 45 of 60 (75%) mice had palpable WiDr tumors. These animals were divided into three groups of equal weight and tumor volume. For HT29 xenografted animals the three groups were fed different ~-irradiated diets (BioServ, Frenchtown, New Jersey) as shown in Table 1. Animals with HT29 tumors fed the above irradiated normal fat/no fiber (N/N) diet had poor intake. Therefore, for animals with explanted WiDr tumors, the standard autoclaved Wayne Blox chow (Allied Laboratories) was used for this group (N/N). After autoclaving, the diet contains 6% fat, 3.6% fiber, 23.5% protein and 4.25 kcal/g. The H/N and H/H group of WiDr animals received the identical diet as outlined above for the HT29 experiment. To equalize caloric intake, the H/N group was given 4 g food/animal/day and the other two groups received 5 g food/animal/day. This amount of food per animal was determined from our previous study of panDigestive Diseases and Sciences, Vol. 36, No. 11 (November 1991)

TABLE 1. COMPOSITIONOF DIETS* IN HT29 EXPERIMENT

Fat (%)t Fiber (cellulose) (%) Protein (%) COH (%) kcal/g

N/N

H/N

H/H

6.5 0.0 21 59.2 3.98

25 0.0 19.4 42.3 5.03

25 16.2 19.7 25.8 3.98

*The mineral and vitamin profiles of the three diets were approximately equal. The exact composition of the diets was supplied to the editor. tThe source of fat in the three diets was from corn oil. creatic cancer ceils explanted to nude mice (13). The HT29 experiment was ended on day 21 of the special diet treatment. Animals with explanted WiDr tumors were maintained on the special diets for 35 days. Animals were reweighed and tumor volumes were calculated on length x (width)2 x 0.5 according to the method of Osieka et al (14). At sacrifice, animal weight and tumor volume were recorded. The tumor was then excised and weighed. Tumors were homogenized in buffer (5 x vol/weigh0 containing 5 mM NaI-I2PO4, 0.1 mM EDTA, 2 mM dithiothreitol (pH 7.4), and homogenized with a Polytron tissue homogenizer (Brinkman). Protein content was determined by the method of Lowry et al (15) using bovine serum albumin as a standard. Deoxyribonucleic acid content was determined by the procedure of Burton (16) using calf thymus DNA as the standard. Ornithine decarboxylase activity was measured as the amount of ~4CO2 liberated from [14C]ornithine (17). Statistical Analysis. The data are expressed as the mean _ standard error for the treatment groups. Differences in weekly tumor volume between diet treatment groups was measured by repeated measure analysis of variance. Statistical analysis between animal weights, tumor weight, protein, DNA and ornithine decarboxylase activity was done using unpaired t test. RESULTS H T 2 9 N u d e M o u s e Expe_riment. Weekly tumor v01umes are shown in Figure 1. At day 21 tumor volume of the H / N was 70% larger than the N/N group (798.3 --- 235 vs 459.5 +-- 67.4 m m s, respectively). The addition of a high-cellulose diet attenuated the effect o f a high-fat diet (group H/H, tumor volume 454.6 --- 51.4 m m 3) (P < 0.08 b y repeated, measure ANOVA). Similarly, final tumor weight in the H / N group (498 ___ 100.2 mg) was 50% greater than the N/N (326 -+ 050 mg) and H / H group (357.5 --- 76.7 mg). These differences were not statistically significant, however. T u m o r protein content, D N A content, and ornithine decarboxylase activity are shown in Table 2. All t h r e e parameters are larger in the H/N group compared to N / N and H / H group. T u m o r ODC activity in the H / N g r o u p ' i s significantly greater

1607

McGARRITY ET AL TUMOR VOLUME COMPARISON 800 ' 9

N/N

[] H/N 600'

i

9

H/H

400 9

200

DAY 7

DAY 0

DAY ~t

DAY 14

Tumor.Volume (ram=) + S.EM , Treatment

Day 0

Day 7

Day 14

Day 21

N/N

69.94• 1,42

124.91--22.56 248,94+_36.11

459.53+_67.4

H/N

72,18_+14.65 143,12~-_28.76 357.35Z68.46 *798.26+_235.04

H/H

73,06+7.67

136.31--26.45 270.56+38,14 454.62•

WiDr Nude Mouse Experiment. As noted above, 45 nude mice with measurable tumors were separated into three groups and maintained on the three special diets for 35 days. Similar to the HT29 experiment, tumor volume and tumor weight was higher in the H/N group compared to the N/N group, whereas the H/H group was intermediate between the H/N and N/N groups (Table 2). These differences were not statistically significant (Table 3). Tumor ODC activity was statistically higher in the H/N group compared to the N/N (P < 0.008) and H/H group (P < 0.002) (Table 2). Likewise, tumor protein and DNA content was highest in the H/N group compared to the N/N and H/N group, but the differences were not statistically significant (Table 3). At sacrifice, animal weights for the N/N group (22.3 - 0.5 g) were not different from the H/H group (22.1 -+ 0.4 g). Animal weights for the H/N group (24.9 - 0.6 g) were greater than the N/N group (P < 0.005) and H/H group (P < 0.0001).

*p 0.08 vs N/N and H/H by repeated measures A N O V A

Fig 1. Each column represents mean tumor volume of explanted HT29 tumors for each diet group with the standard error of the mean. Tumor volume in the high fat/no fiber (H/N) was larger at sacrifice than the normal fat/no fiber (N/N) and high fat/high fiber (H/H) groups. This difference was intermediate in significance (P < 0.08 by repeated measures ANOVA).

than the H/H group. Tumor DNA of the H/N and H/H group is significantly larger than the N/N group. Animals in the H/N group consistently ate all food given. Food intake in the N/N group was sporadic and decreased progressively during the experiment. At day 21 animal weights in the N/N group (21.1 --- 0.6 g) were statistically less than the H/N (25.3 --- 0.6 gm) and H/H group (22.51 -4- 0.5), P < 0.05.

DISCUSSION Our results suggest that a diet high in unsaturated fat stimulates the growth of human colon cancer cells, HT29 and WiDr, explanted to athymic nude mice. HT29 tumor DNA content was statistically greater in the H/N group compared to the N/N group. Tumor volume increased sixfold in the animals consuming a low-fat, no-fiber (N/N) diet whereas the higher-fat, no-fiber (H/N) group experienced an l 1-fold increase in tumor volume. In concert, tumor weight, protein, and ornithine decarboxylase activity was increased in the H/N group compared to the N/N group. These changes, however, fell short of statistical significance. This may be due to the large variation noted in each

TABLE 2. EFFECTS OF DIETARY FAT AND FIBER ON TUMOR VOLUME, WEIGHT, PROTEIN, AND DNA CONTENT AND ORNITHINE DECARBOXYLASEACTIVITY (ODC) OF HT29 TUMORS EXPLANTED TO ATHYMIC NUDE MICE N/N

Tumor Tumor Tumor Tumor Tumor

volume (mm 3) weight (mg) protein(rag) DNA (mg) ODC*

460 326 37.4 2.00 29.9

- 67 - 50 • 5.9 *-. 0.3 -+ 3.3

H/N

798 498 47.3 3.37 37.3

--- 46 -+ 100 +- 9.1 0.82t --. 3.4r

H/H

455 357 42.1 3.07 27.2

+- 51 - 76 +- 9.4 -+ 0.62t - 4.3

*Measure in pmole/mg protein hour. t P < 0.05 vs N/N. *P < 0.05 vs H/H.

1608

Digestive Diseases and Sciences, Vol. 36, No. I1 (November 1991)

FAT, FIBER, AND CANCER TABLE 3. EFFECTS OF DIETARY FAT AND FIBER ON TUMOR VOLUME, WEIGHT, PROTEIN, AND DNA CONTENT AND ORNITHINE DECARBOXYLASE (ODC) ACTIVITY OF WlDR TUMOR EXPLANTED TO ATI-IYMIC NUDE MICE N/N

Tumor Tumor Tumor Tumor Tumor

volume (mm3) weight (rag) protein(mg) DNA(mg) ODC*

921 573 106 94.5 43

--- 270 • 160 • 34 • 20 -+ 6.7

H/N

1602 1101 133 180 79.9

--- 435 --- 307 • 37 • 8.0t • 6.1:~w

H/H

1295 832 102 126 53

• 384 • 203 • 26 • 42 • 4.7

*pmole/mg protein hr. t P < 0.05 < P < 0.10 compared to N/N. ~:P < 0.008 vs N/N. w < 0.002 vs H/H.

group, which was a normal phenomenon in this tumor model (18). Explanted WiDr tumor cells showed a similar response to the addition of fat to a low fiber diet. Tumor ODC activity was statistically greater in the H/N group compared to the N/N group. In general, the addition of fiber (cellulose) attenuated the growth-promoting effects of a high-fat diet for explanted HT29 and WiDr tumor cells. The mechanism for the effects of fat and fiber on explanted tumor growth were not investigated in this study. In most animal models of primary colon carcinogenesis, increased dietary fat promotes tumor growth and incidence. The effects of fat are believed to be due to enhanced cellular proliferation induced by fatty acids and bile salts, increased prostaglandin production by unsaturated fats and increased fluidity of cell membranes, which increases the permeability of dietary carcinogens (37). These studies have focused on the changes in luminal contents induced by diet on in situ colon tumors. Epidemiological studies have demonstrated that a high-fat diet increases the incidence of extraluminal gastrointestinal tumors such as pancreas as well as breast cancer (11). We have shown previously that a high unsaturated fat diet increased the growth of human pancreatic cancer xenografted to nude mice (13). This effect of fat on tumor growth was inhibited by the specific CCK-receptor antagonist, L364,718 (13). Whether the growth-promoting effects of dietary fat on explanted human colon cancer is secondary to circulating gastrointestinal peptides remains to be tested. We have demonstrated previously that exogenously administered gastrin stimulated the growth of human colon cancer xenografts to nude mice (19). The current study suggests that the growth-promoting effects of fat on explanted tumors is attenuated by high fiber diet. Sinkeldan et al (20) noted an enhancing effect of a high-fat diet on tumor number and incidence in Digestive Diseases and Sciences, Vol. 36, No. 11 (November 1991)

chemically induced colon cancers in rats, whereas fat had no effect when dietary fiber content was high. Nutter et al (21) studied the effects of diet on growth of a chemically induced colon cancer cell line explanted to BALB/c mice. Animals fed a milk-based low-protein diet had smaller tumor volumes and greater cell-mediated immunity, whereas the level or source of dietary fat had no effect. Similar to our results, Reisser et al (18) noted that the tumor volume of dimethylhydrazine-induced colon cancer cells explanted to inbred BDIX rats was greater but did not reach statistical significance. Other parameters of tumor growth, such as weight, protein, or DNA content were not measured. In Reisser's study, a high-fat diet did not influence cytotoxic activity of peritoneal macrophages. In our initial study (HT29), the caloric intake of animals in the low fat/no fiber(N/N) group was sporadic and less compared to the other two groups. We felt that this was due to poor palatability of an irradiated low-fat diet. For the WiDr study, autoclaved standard lab chow was substituted for the N/N group. It appears that the fat content affected palatability to a greater amount than the irradiation. The differences we noted in HT29 tumor growth parameters may be due to caloric intake in the N/N group. In the WiDr experiments the final weights of the N/N (22.3 -+0.5 g) and H/H (22.1 -+ 0.4 g) groups were similar but decreased compared to the H/N group (24.9 -+ 0.6 g). Prior studies have shown that increased and restricted colonic intake decreased colon tumor formation in animal models of carcinogenesis (22, 23). It also should be noted that although the mineral and vitamin composition of the diets were very similar in the three groups, the H/N group received less food (4 g/day) and therefore less micronutrients thari the other two 1609

McGARRITY ET A L

groups. These subtle differences may have influenced tumor growth in this model. In conclusion, our studies suggest that the growth of human colon cancer xenografts is enhanced by a diet high in unsaturated fat. The mechanism of this growth-promoting effect has not been established. Since several human colon cancers have been shown to have different hormone receptors, such as gastrin, and we have shown that the growth of human colon cancer in vivo is promoted by gastrin and inhibited by somatostatin (19), we suggest that dietary fat promotes growth by increasing trophic gastrointestinal peptide secretion. Dietary fiber may have an adjuvant role in decreasing fatstimulated growth in cancer-preventative diets. Whether dietary manipulation affects hormonestimulated growth o f colon cancer in vivo remains to be tested. ACKNOWLEDGMENTS The authors acknowledge Dr. Saeed Bagheri of the VA Medical Center, Lebanon, Pennsylvania, for providing some of the experimental animals, and Sue Huntzinger for her expert secretarial assistance.

REFERENCES 1. Barkett DP: Epidemiology of cancer of the colon and rectum. Cancer 28:3-13, 1971 2. Bjelke E: Epidemiologic studies of cancer of the stomach, colon and rectum with special emphasis on the role of diet. Scand J Gastroenterol Suppl 9:1-235, 1974 3. Weisburger JH, Reddy BS, Barnes WS, Wynder EL: Bile acids, but not neutral sterols, are tumor promoters in the colon in man and rodents. Environ Health Perspect 50:101107, 1983 4. Reddy BS, Maeura Y: Tumor promotion by dietary fat in azoxymethane-induced colon carcinogenesis in female F344 rats: Influence of amount and source of dietary fat. J Nail Cancer Inst 72:745-750, 1984 5. Sakaguehi M, Hiramatsu Y, Takada H, Yamamura M, Hioki K, Saito K, Yamarnoto M: Effect of dietary unsaturated and saturated fats on azoxymethane-induced colon carcinogenesis in rats. Cancer Res 44:1472-1477, 1984 6. Bull AW, Soullier BK, Wilson PS, Hayden MT, Nigro ND: Promotion of azoxymethane-induced intestinal cancer by high fat in rats. Cancer Res 39:4956-4959, 1979 7. Rozhin J, Wilson PS, Bull AW, Nigro ND: Orrtithine decarboxylase activity in the rat and human colon. Cancer Res 44:3226-3230, 1984

1610

8. Lipkin M: Dietary, environmental, and hereditary factors in the development of colorectal cancer. CA-Cancer J Clin 29:291-299, 1979 9. Kritchevsky D: Fiber, steroids, and cancer. Cancer Res 43:2491s-2495s, 1983 10. Jacobs LR, Lupton JR: Relationship between colonic luminal pH, cell proliferation, and colon carcinogenesis in 1,2dimethylhydrazine treated rats fed high fiber diets. Cancer Res 46:1727-1734, 1986 11. Armstrong B, Doll R: Environmental factors and cancer incidence and mortality in different countries, with special reference to dietary practices. Int J Cancer 15:617-631, 1975 12. Rogers AE, Nanss KM: Rodent models for carcinoma of the colon. Dig Dis Sci 30:87S-102S, 1985 13. Smith JP, Kramer S, Bagheri S: Effects of a high-fat diet and L364,718 on growth of human pancreas cancer. Dig Dis Sci 35:726-732, 1990 14. Osieka R, Houchess DP, Goldin A, Johnson RF: Chemotherapy of human colon cancer xenografts in athymic nude mice. Cancei"40:2640-2650, 1977 15. Lowry OH, Rosebrough NJ, Farr AL, Randall RS: Protein determination with the Folin phenol reagent. J Biol Chem 193:265-275, 1951 16. Burton K: A study of the condition and mechanisms of the diphenylamine reaction for the calorimetric determination of the deoxyribonucleic acid. Biochem J 62:315-323, 1956 17. Pegg AE, Seely JC: Ornithine decarboxylase (mouse kidney). Methods Enzymol. 94:158-166, 1983 18. Reisser D, Lagadee P, Pelletier H, Fady C, Olsson WO, Jeannin JF: Lack of effect of a high polyunsaturated fat diet in the growth of .transplantable colon tumors and on the cytolytic activity of macrophages in rats. Reprod Nutr Dev 27:1013-1021, 1987 19. Smith JP, Solomon TE: Effects of gastrin, proglumide and somatostatin on growth of human colon cancer. Gastroenterology 95:1541-1548, 1988 20. Sinkeldan EJ, Kuper CF, Bosland MC, Hollander VM, Vedder DM: Interactive effects of dietary wheat bran and lard on N-methyl-Nl-nitro-N-nitrosoquanidine-induced colon carcinogenesis in rats. Cancer Res 50:1092-1096, 1990 21. Nutter RL, Gridiey DS, Kettering JD, Andres MC, Aprecio RM, Slater JM: Modification of a transplantable colon tumor and immune responses in mice fed different sources of protein, fat and carbohydrate. Cancer Lett 18:49-62, 1983 22. Reddy BS, Tanaka T, Simi B: Effect of different levels of dietary trans fat or corn oil on azoxymethane-induced colon carcinogenesis in F344 rats. J Natl Cancer Inst 75:791-798, 1985 23. Reddy BS, Chung-Xiou W, Marayama H: Effect of restricted caloric intake on azoxymethane-induced colon tumor incidence in male F344 rats. Cancer Res 47:1226-1228, 1987

Digestive Diseases and Sciences, Vol. 36, No. 11 (November 1991)