The ideas and opinions expressed herein are those of the authors and endorsement by the State of California, Department of Health Services, the National Cancer Institute, the Centers for Disease Control and Prevention, and/or the Genetic Epidemiology Research Institute of the University of California, Irvine is not intended nor should be inferred. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Cancer Institute, National Institutes of Health, or the Lon V. Smith Foundations.
Risk and risk reduction involving arginine intake and meat consumption in colorectal tumorigenesis and survival
Version of Record online: 9 NOV 2006
Copyright © 2006 Wiley-Liss, Inc.
International Journal of Cancer
Volume 120, Issue 3, pages 459–468, 1 February 2007
How to Cite
Zell, J. A., Ignatenko, N. A., Yerushalmi, H. F., Ziogas, A., Besselsen, D. G., Gerner, E. W. and Anton-Culver, H. (2007), Risk and risk reduction involving arginine intake and meat consumption in colorectal tumorigenesis and survival. Int. J. Cancer, 120: 459–468. doi: 10.1002/ijc.22311
- Issue online: 29 NOV 2006
- Version of Record online: 9 NOV 2006
- Manuscript Accepted: 31 JUL 2006
- Manuscript Received: 24 APR 2006
- NIH. Grant Number: CA72008
- UC Irvine Divisions of Hematology/Oncology and Epidemiology
- Public Health Institute, State of California. Grant Number: 050N-8707-S1527
- California Department of Health Services. Grant Numbers: 103875, 103885
- National Cancer Institute's Surveillance, Epidemiology and End Results Program
- Centers for Disease Control and Prevention National Program of Cancer Registries
- National Cancer Institute, National Institutes of Health, Department of Health and Human Services. Grant Numbers: 5R01CA63706, 5R01CA67151
- Lon V. Smith Foundations. Grant Number: LVS-18840
- National Institutes of Health. Grant Number: CA82450 (RFH)
- familial colorectal cancer;
- colon cancer;
- rectal cancer;
Elevated polyamine and nitric oxide levels (both derived from arginine) promote tumorigenesis, whereas non-steroidal anti-inflammatory drugs (NSAIDs) inhibit colorectal cancer (CRC) incidence in experimental and epidemiologic studies. We investigated dietary arginine-induced intestinal tumorigenesis and NSAID-inhibitory effects in ApcMin/+ mice differentially expressing nitric oxide synthase-2 (Nos2). We also studied effects of estimated arginine exposures through meat consumption on tumor characteristics and survival in human CRC cases. Dietary arginine increased high-grade colon adenoma incidence in ApcMin/+Nos2+/+ mice, but not in Nos2 knockout mice. Additionally, celecoxib suppressed intestinal steady state ornithine decarboxylase RNA levels (p < 0.001), induced steady state spermidine/spermine N1-acetyltransferase RNA levels (p = 0.002), decreased putrescine levels (p = 0.04) and decreased tumor number in the small intestines of ApcMin/+Nos2+/+ mice (p = 0.0003). Five hundred and eleven cases from our NCI-supported CRC gene-environment study were analyzed based on self-reported meat (as a surrogate for arginine) consumption. Familial CRC cases (n = 144) in the highest meat consumption quartile (Q4) had no statistically significant differences in tumor grade compared to cases in Q1–Q3 (p = 0.32); however, they were observed to have decreased overall survival (OS) (10-year OS = 42% vs. 65%; p = 0.017), and increased risk of death in an adjusted analysis (hazards ratio [HR] = 2.24; p = 0.007). No differences in tumor grade, OS or adjusted HR were observed for sporadic CRC cases (n = 367) based on meat consumption. Our results suggest important roles for arginine and meat consumption in CRC pathogenesis, and have implications for CRC prevention. © 2006 Wiley-Liss, Inc.
Extensive epidemiologic studies have investigated the relationship between meat consumption and colorectal cancer (CRC) risk. Meat consumption was shown not to be associated with CRC in a large meta-analysis of 34 case–control and 14 cohort studies published during 1973–1999, while red meat and processed meat intake were associated with moderate increases in CRC risk.1 A large cohort study has recently confirmed an increased risk of developing CRC associated with consumption of red meat and processed meats, and demonstrated a protective effect for fish and poultry.2 While many epidemiologic studies have assessed the risk of meat consumption on developing CRC,1, 3 there is a paucity of data addressing the risk of meat consumption on outcomes in CRC patients. Meat consumption is the major source of dietary arginine in humans, with high quantities of arginine found not only in red meat but also in pork, fish and chicken.4 Meat and meat products accounted for 37–38% of total daily arginine in two European studies, where subjects were estimated to have similar daily arginine intake to U.S. men (i.e. 4.2–5.0 g/day).5, 6 In cell culture, normal cells become quiescent in an arginine-deficient diet, whereas most transformed or malignant cells die.7, 8 Arginine supplementation,8 particularly L-arginine,9 has been shown to increase tumor growth in rodents whereas D-arginine may have tumor-inhibitory properties.9 In human breast cancer tissue, dietary arginine has been shown to stimulate tumor growth.10
Importantly, arginine is the key substrate for two competing metabolic pathways believed to be involved in carcinogenesis: the nitric oxide (NO) synthase pathway and polyamine synthesis. Arginine is catabolized by NO synthase 2 (NOS2) and other NO synthases to form nitric oxide. Inducible isoforms of NOS2 are abundant in human colorectal adenomas.11 NOS2 is also overexpressed in rat colon carcinoma tissues after azoxymethane treatment,12, 13 an effect that is inhibited by aspirin.12ApcMin/+Nos2 knockout mice provided with dietary arginine were noted to have a decreased number of intestinal and colonic adenomas compared to ApcMin/+Nos2+/+ mice.14, 15 Thus it has been suggested that NOS2-selective inhibitors may have a future role in CRC chemoprevention.
Arginine is catabolized by the enzyme arginase to form ornithine, the substrate for putrescine synthesis. This is the first step in polyamine biosynthesis and is catalyzed by the enzyme ornithine decarboxylase (ODC).16 Multiple abnormalities in the control of polyamine content results in increased polyamine levels that can promote tumorigenesis.17 APC, a gene that is mutated in FAP and in the majority of sporadic colon cancers, regulates ODC and therefore polyamine synthesis.18 In accordance, loss of APC increases ODC expression, polyamine content and small intestinal tumor formation.19 Treatment with the chemopreventive agent α-difluoromethylornithine (DFMO), a suicide inhibitor of ODC, has been shown to decrease polyamine levels and suppress tumorigenesis in the small intestine and colon of ApcMin/+ mice (an FAP mouse model),19 and decrease colorectal tissue polyamine levels in human patients with colon polyps.20
Together, DFMO and non-steroidal anti-inflammatory drugs (NSAIDs) inhibit intestinal cancer development in mouse models.21 Sulindac22 and aspirin23 are NSAIDs that have been demonstrated to activate the polyamine catabolic pathway via increased expression of spermine spermidine acetyltranferase (SSAT), polyamine acetylation and subsequent export. It is not known whether this mechanism exists for the selective cyclooxygenase-2 (COX-2) inhibitors.
In the experimental portion of this investigation, we examine the role of Nos2 and dietary arginine in murine carcinogenesis via gene knockout mouse model systems. We also investigate the effects of celecoxib on Odc and Ssat, and test whether celecoxib treatment can inhibit intestinal tumorigenesis in FAP mouse models. In the epidemiologic portion of this study, we investigate the effects of dietary meat consumption on tumor characteristics at presentation and survival in CRC cases. We hypothesized that any differential effects of meat consumption on these clinical outcomes might be mediated through gene–environment interactions, and thus manifest predominantly in cases with multifactorial inherited susceptibility to CRC (i.e. familial CRC cases).
Material and methods
Mouse experimental procedures
Animals and experimental conditions.
C57BL/6J-ApcMin mice were acquired at 6 weeks of age from the Jackson Laboratory (Bar Harbor, ME) and bred in The University of Arizona's Animal Care Facility. Affected ApcMin/+ mice were obtained by crossing C57BL/6J-ApcMin males with C57BL/6J females. For assessment of dietary arginine effects, C57BL/6J-ApcMin+/-Nos2−/− mice (ApcMin+/-Nos2−/−) were produced by crosses between C57BL/6J-ApcMin male and C57BL/6J-Nos2tm1Lau/j female mice for more than 10 generations. Offspring were genotyped using allele-specific polymerase chain reaction (PCR) as described in genotyping protocols from the Jackson Labs website (http://jaxmice.jax.org/). All animals were raised in cages under nonsterile microisolator conditions and were put on the various treatments at age 5–6 weeks. Animals were fed with an AIN-93G diet (Harlan Teklad), a defined diet that includes 1.08 g arginine/kg mouse body weight per day. Mice were sacrificed at age of 110 days by CO2 inhalation and were analyzed by tumor scoring, tumor grade evolution and Northern blotting. Throughout the experiment, animals had access to drinking water and food ad libitum.
Assessment of dietary arginine effects on tumor number and grade in ApcMin/+ Nos2+/+ mice.
For assessment of dietary arginine effects on tumor grade, the diet was supplemented with arginine in the drinking water (0–0.2%, Sigma-Aldrich, St. Louis, MO). Mice have a 7.5 times larger metabolic rate compared to humans.24 Therefore, mouse dietary arginine concentrations (1.08–1.38 g arginine/kg mouse body weight per day, as described previously15) were adjusted for differences in metabolic rates, i.e. concentrations in the upper range of that consumed by human (121 mg/kg per day, as described6). After termination, whole small intestine and colon segments were removed, flushed with buffered saline, opened longitudinally and spread out with the mucosal surface up. A Swiss gut roll of small intestine, colon and all grossly evident colon tumors were fixed in 10% buffered formalin. Processed, paraffin-embedded 5-μm sections were stained with hematoxylin and eosin prior to evaluation by a veterinary pathologist. The grade of dysplasia for each adenoma examined histologically was determined on the basis of criteria outlined previously.25
Celecoxib effects in intestinal tissue of ApcMin/+ mice.
To determine the effect of celecoxib on gene expression and tumor number, beginning at the age of 5 weeks, C57BL/6J-ApcMin mice were fed the semi-purified AIN-93G diet containing 150 or 500 ppm celecoxib.
Odc and Ssat expression:Odc and Ssat expression in small intestinal tissue of ApcMin/+ mice was measured after 60 days of treatment via Northern blot analysis. After termination, the intestinal ileum was removed, opened and flushed with ice-cold saline. The 1 cm tumor-free portions of the ileum were placed in TRIzol® reagent (Gibco, BLR) homogenized using a Polytron homogenizer, and processed for the isolation of total cellular RNA.26 Twenty microgram of total RNA was separated on a 1% agarose/formaldehyde gel and transferred to a nylon membrane (Hybon-N, Amersham). The cDNA probes for Northern blotting were prepared using [32P]-dCTP and a random priming technique (Roche Molecular Biochemicals, Indianapolis, IN). The probes were purified with G-50 Sepharose columns (Roche Molecular Biochemicals, Indianapolis, IN) and quantitated using a scintillation counter. Membranes were hybridized with 32P-labeled cDNA encoding for mouse Odc (1.38 kb EcoRI-EcoRI fragment) and human Ssat (0.67 kb EcoRI-EcoRI fragment). Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (0.75 kb PstI-XbaI fragment) was used to control for variations in RNA loading. Northern blot autoradiograms were quantitated by densitometric analysis (ImageQuant, Molecular Dynamics, Sunnyvale, CA).
Polyamine analysis: We measured polyamine content in the intestinal tissue of animals treated with 500 ppm celecoxib, as this represents a valuable endpoint of drug effect on the polyamine pathway. The 1 cm tumor-free portions of the ileum were collected, flushed with ice-cold saline and stored frozen at −80°C. Samples were processed and assayed for polyamine (putrescine, spermidine, spermine and N1-acetylspermine) content by reverse-phase high performance liquid chromatography with 1,7-diaminoheptane as an internal standard.27 Protein content in each sample was determined by the BCA protein assay kit (Pierce, Rockford, IL). Data are expressed as nmol polyamine per mg protein.
Combined effects of DFMO and celecoxib on tumorigenesis in ApcMin/+ mice.
To assess the combined effect of celecoxib and DFMO on intestinal tumorigenesis, DFMO was administered alone, or in combination with celecoxib (500 ppm in the diet) at two different doses of 0.5% and 2% in the drinking water. The independent effects of celecoxib (500 ppm in the diet) on intestinal tumorigenesis were also tested. After termination, whole small intestine and colon segments were removed, flushed with buffered saline, opened longitudinally and spread out with the mucosal surface up. Tissues were fixed in 10% neutral buffered formalin for 24 hr and then placed in 70% ethanol. Tumor number per mouse was evaluated using a LSGA epi-illuminator dissecting microscope (Olympus Optical Co., Ltd, Tokyo, Japan) at 20× magnification. Tumors larger than 0.5 mm were counted and their diameter was measured. All assessments were done by the same observer who was blinded to the mouse treatment group. Animal weight at the time of sacrifice was recorded to determine if they absorbed similar amounts of food and water after the aforementioned treatments.
Epidemiologic CRC gene–environment study
For the present study, all incident CRC cases diagnosed between 1994 and 1996 were obtained from a previously described study investigating gene–environment interactions in CRC.28 Family history of cancer was ascertained via telephone interview. Familial (non-HNPCC non-FAP) CRC cases were identified as those having at least one first degree relative with CRC. Amsterdam criteria were used to define HNPCC families.29 HNPCC and FAP cases, familial cases and sporadic CRC cases were identified using available family history data. During the defined period 1,256 cases were consented for the study, including 1,007 sporadic CRC cases, 228 familial CRC cases and 21 HNPCC cases, as described elsewhere.30 All HNPCC cases were subsequently excluded from the analysis. Five hundred and eleven of the sporadic and familial CRC cases completed a food frequency questionnaire (FFQ).
Clinical and demographic data.
Clinical and demographic data from consented cases were obtained from the Cancer Surveillance Programs of Orange County, Imperial County and San Diego County, California (CSPOC/SANDIOCC databases).28 Recorded data included demographic information (age, gender and ethnicity), histology, tumor site, tumor grade, stage at presentation, survival status and treatment with surgery, radiation therapy or chemotherapy. These data were abstracted from medical and laboratory records by trained tumor registrars according to Cancer Reporting in California: Vol. 1, Abstracting and Coding Procedures for Hospitals.31 Tumor site and histology were coded according to criteria specified by the World Health Organization in International Classification of Diseases for Oncology (ICD-O).32 Primary site code was searched using the SEER site code for colon (21041, 21043-21048) and rectum (21051-21052). Appendiceal cancers were excluded. Histology codes included invasive cases of adenocarcinoma (8140, 8144, 8210, 8260–8263, 8380, 8490), mucinous adenocarcinoma (8470, 8480, 8481), carcinoma (8010, 8020, 8070, 8071, 8124, 8240) and not otherwise specified (8041, 8042, 8120, 8130, 8243, 8246, 8560, 8570, 8722). Staging was grouped into three broad categories that could be classified from clinical and pathologic records, and defined according to SEER summary staging as localized disease, regional disease and remote disease (localized or regional disease with distant spread). Hospital registrars contacted cases annually, and CSPOC/SANDIOCC staff annually reviewed state death certificates to identify deceased registry cases. Follow-up data through December 2003 were available for analysis. The last date of follow-up was either the date of death or the last date the individual was contacted.
Meat and other food consumption.
Food consumption was self-reported via a validated 100-item NCI-Block FFQ, where patients were asked to report their usual eating habits during the 1 year prior to diagnosis of CRC.33, 34 Micronutrient data (i.e. calcium, iron), total daily fiber intake, total daily energy intake and body mass index were calculated from the self-reported FFQ responses. The number of medium-sized servings consumed per week for each meat item was calculated by multiplying the number of servings consumed per week by the estimated serving size (i.e., 0.5 for small, 1.0 for medium and 1.5 for large). Meat consumption quartiles were subsequently calculated. Cases in the highest quartile of meat consumption (Q4) were compared to those in all of the other meat consumption quartiles (Q1–Q3). Arginine intake for each FFQ food item was calculated by estimating the quantity of arginine per serving from widely available computer-based food composition tables,4 and multiplying this value by the consumption frequency and serving size for each food item. As an internal control measure, manual chart review was conducted on all cases reporting total daily energy intake of less than 500 kcal (n = 15) or greater than 5,000 kcal (n = 6) to ensure valid FFQ reporting, including one patient reporting extremely high meat intake (61.5 medium-size serving equivalents per week). All such cases were determined to contain complete and valid data and were included in the analysis.
A nonparametric test (Wilcoxon rank sum or Mann–Whitney) was performed for the analysis of Odc and Ssat gene expression. For the difference in the number of tumors between control and treated mice, the Poisson regression analysis was used. Logistic regression was used to analyze the difference in low and high grade adenoma incidence.
Comparisons of demographic, clinical and pathologic variables between cases were performed using Pearson χ2 statistic or Fisher's exact test for nominal variables and Student t- test for continuous variables. For the analysis of continuous variables by meat consumption quartile, a nonparametric one-way analysis of variance using Wilcoxan (Kruskal-Wallis) test was performed. Linear regression analysis was used to assess dependent relationships between continuous variables. Univariate survival rate analyses were estimated using the Kaplan and Meier method, with comparisons made between groups by the log rank test. Cox proportional hazards modeling using time since diagnosis were performed. Each variable in the model was coded using dummy variables. All statistical analyses were conducted using SAS 9.1 statistical software (SAS Institute, Cary, NC). Statistical significance was assumed for a two-tailed p value less than 0.05.
Cases signed a consent form allowing for release of medical information, including pathology reports, tissue blocks and linkage to the aforementioned Cancer Surveillance Programs. This study has been approved by the University of California, Irvine Institutional Review Board (No. 1993-257).
Mouse experimental studies—The role of arginine and its catabolic product, polyamines, in intestinal tumorigenesis
Supplemental dietary arginine effect on incidence of high-grade adenomas in ApcMin/+ Nos2+/+ mice.
We examined the effect of dietary arginine on colonic tumor incidence in mice. ApcMin/+Nos2+/+ mice or ApcMin/+Nos2 knockout mice were fed an AIN-93G diet supplemented with dietary arginine in the drinking water (0, 0.02 and 0.2%). Supplementing the diet with arginine did not affect animal weight. All mice weighed 20–23 g, i.e. they absorbed similar amounts of food and water in the different treatment groups. ApcMin/+ mice expressing Nos2 showed an arginine-dependent increase in high-grade adenoma incidence in the colon (Fig. 1). Supplementing the diet with 0.2% arginine resulted in a more than 2-fold increase in tumor incidence (p = 0.042). In contrast, low-grade adenoma incidence in mice expressing Nos2 was not affected by dietary arginine. In Nos2 knockout mice, adenoma incidence was not affected by arginine; therefore, the increase in colon tumor grade with dietary arginine was Nos2-dependent.
Celecoxib effect on polyamine metabolic enzymes in intestinal tissue of ApcMin/+ mice.
We tested whether celecoxib affects polyamine metabolism in the intestinal tract of Min mice. After 60 days of treatment, celecoxib suppresses the steady state RNA levels of Odc, and induces steady state RNA levels of Ssat in a dose dependent manner (Figs. 2a and 2b). Celecoxib effects on Odc and Ssat expression were observed in the small intestine but not in the colon, and had statistically significant values for Odc at both 150 and 500 ppm concentrations (p < 0.001) and for Ssat at the concentration of 500 ppm (p = 0.002). We found a statistically significant decrease in the concentration of putrescine, the product of ODC activity, in the small intestine of mice treated with 500 ppm of celecoxib (p = 0.04) (Table I). We found also a marginal decrease in tissue levels of N1-acetylspermidine (the product of the SSAT enzyme activity, p = 0.07). This decrease was small and not statistically significant possibly due to luminal secretion of acetylated spermidine.
|Polyamine content, nmol/mg protein (Mean + SD)|
|Control (n = 6)||0.54 ± 0.25||8.14 ± 1.41||5.55 ± 0.86||0.22 ± 0.06|
|Celecoxib, 500 ppm (n = 3)||0.18 ± 0.12*||6.58 ± 1.12||6.98 ± 1.16||0.14 ± 0.03|
Combination of DFMO and celecoxib: Effect on tumorigenesis in ApcMin/+ mice.
DFMO at a concentration of 0.5% did not affect tumor number in ApcMin/+ mice fed with a defined AIN93G diet (Fig. 2c). A higher dose of DFMO (2%) reduced the tumor number per mouse by about 50%. We did not observe a significant tumor location dependency with the DFMO treatment (data not shown). Celecoxib (500 ppm) reduced tumor number by about 50% with little location dependency (p = 0.0003). The combination of 0.5% DFMO and 500 ppm of celecoxib reduced tumor number by over 80% (p < 0.0001) and was more effective in suppressing distal intestinal tumor number when compared to proximal intestinal tumor number (data not shown). The combination of a higher dose of DFMO with celecoxib did not improve the efficacy of these agents to suppress tumorigenesis (p < 0.0001). We found a statistically significant increase in body weights of mice treated with celecoxib (p = 0.02, data not shown). Mice treated with 2% DFMO alone or in combination with celecoxib had a statistically significant decrease in body weights compared to control animals (p = 0.005 and p = 0.04, respectively). No effect on body weight was found in mice treated with 0.5% DFMO in combination with celecoxib, implying that the water consumption and the caloric intake of the treatment mice was similar to those of the control mice.
Epidemiologic studies—Effect of meat and arginine intake on clinical outcomes
CRC study population.
The 511 incident CRC cases identified included 367 sporadic CRC cases and 144 familial CRC cases. Demographic comparisons for these groups (Table II) reveal significant baseline differences. Familial patients were older at presentation, and had a greater proportion of localized tumors and tumors in the proximal colon. While the proportion of cases treated surgically or with radiation was similar for familial and sporadic CRC cases, fewer familial CRC cases received chemotherapy. However, this may be related to the (expected) finding that chemotherapy treatment was strongly associated with increased stage at presentation (i.e. recorded use of chemotherapy was 16, 68 and 81% among local, regional and remote stage cases, respectively; p < 0.0001). Types of meat and corresponding medium serving sizes as listed in the FFQ are depicted in Table III, along with estimated arginine contents. Because of the baseline differences between sporadic and familial CRC cases, meat consumption quartiles were constructed separately for each group: familial CRC cases – Q1 (0–4.3 servings per week), Q2 (4.4–7.0 servings per week), Q3 (7.1–10.7 servings per week), Q4 (10.8–27.4 servings per week); sporadic CRC cases – Q1 (0–5.3 servings per week), Q2 (5.4–8.2 servings per week), Q3 (8.3–12.5 servings per week), Q4 (12.6–61.5 servings per week). Comparisons for each of the investigated variables across all four meat consumption quartiles are presented in Table IV for familial and sporadic CRC patients.
|Familial CRC (n = 1 44)||Sporadic CRC (n = 367)||p|
|Median age (in years, with range)||64 (32–90)||57 (22–95)||<0.0001|
|Male||72 (50%)||190 (52%)||0.72|
|Female||72 (50%)||177 (48%)|
|Caucasian||130 (90%)||306 (83%)||0.16|
|African-American||1 (<1%)||2 (<1%)|
|Hispanic||9 (6%)||31 (8%)|
|Asian||3 (2%)||27 (7%)|
|Other||1 (<1%)||1 (<1%)|
|Stage at presentation|
|Local||71 (50%)||139 (38%)||0.049|
|Regional||53 (38%)||172 (48%)|
|Remote||17 (12%)||51 (14%)|
|Proximal and transverse||64 (44%)||106 (29%)||0.002|
|Descending||8 (6%)||14 (4%)|
|Sigmoid||29 (20%)||104 (29%)|
|Rectosigmoid||22 (15%)||46 (13%)|
|Large intestine-Other||0 (0%)||10 (3%)|
|Adenocarcinoma||128 (89%)||336 (92%)||0.30|
|Mucinous adenocarcinoma||11 (8%)||18 (5%)|
|Carcinoma||5 (4%)||8 (2%)|
|Not otherwise specified||0 (0%)||4 (1%)|
|Grade 1||25 (19%)||43 (13%)||0.36|
|Grade 2||87 (66%)||237 (71%)|
|Grade 3||20 (15%)||55 (16%)|
|Grade 4||0 (0%)||1 (<1%)|
|First course of treatment*|
|Surgery||136/144 (94%)||339/366 (93%)||0.46|
|Radiation therapy||16/144 (11%)||62/366 (17%)||0.10|
|Chemotherapy||50/137 (37%)||182/350 (52%)||0.002|
|Fiber intake (mean g/day)||11.7 ± 0.5 SE||12.2 ± 0.4 SE||0.41|
|Calcium intake (mean mg/day)||710 ± 265E||0.96|
|Iron intake (mean mg/day)||10.7 ± 0.4 SE||12.0 ± 0.3 SE||0.026|
|Energy intake (mean kcal/day)||1477 ± 70 SE||1695 ± 49 SE||0.015|
|Body mass index (mean BMI, kg/m2)||26.0 ± 0.4 SE||26.4 ± 0.3 SE||0.42|
|Meat consumption (mean number of medium servings per week)||8.1 ± 0.5 SE||9.5 ± 0.3 SE||0.022|
|Mean weekly arginine intake (g)||21.1 ± 10 SE||24.0 ± 0.7 SE||0.022|
|Mean weekly meat-derived arginine intake (g)||10.1 ± 06 SE||12.4 ± 0.4 SE||0.006|
|Food item||Medium serving size||Estimated arginine content (per medium serving) (g)|
|Sausage||2 patties or links||0.34|
|Beef (includes steaks, roasts, beef sandwiches)||4 ounces||1.94|
|Beef stew or pot pie||1 cup||0.68|
|Hamburger (includes cheeseburger, meatloaf, beef burritos, tacos)||1 medium sandwich or 4 ounces||1.01|
|Hot dogs||2 hot dogs||0.64|
|Liver (includes chicken liver)||4 ounces||2.28|
|Ham (includes bologna, salami, other lunch meats)||2 slices or 2 ounces||0.58|
|Pork (includes pork chops, roasts)||2 chops or 4 ounces||1.87|
|Chicken or Turkey (includes sandwiches, also roasted/stewed/broiled)||2 small or 1 large piece||1.85|
|Fried chicken||2 small or 1 large piece||1.39|
|Fried fish or fish sandwich||4 ounces or 1 sandwich||1.12|
|Tuna (includes tuna salad, tuna casserole)||1/2 cup||1.26|
|Oysters||5 pieces, 1/4 cup or 3 ounces||0.36|
|Shellfish (includes shrimp, crab, lobster, etc.)||5 pieces, 1/4 cup or 3 ounces||1.58|
|Other fish (includes broiled or baked fish)||2 pieces or 4 ounces||1.63|
|Familial CRC patients (n = 144)||p||Sporadic CRC patients (n = 367)||p|
|Meat Q1 (n = 36)||Meat Q2 (n = 36)||Meat Q3 (n = 36)||Meat Q4 (n = 36)||Meat Q1 (n = 92)||Meat Q2 (n = 92)||Meat Q3 (n = 92)||Meat Q4 (n = 91)|
|Median age (years) (range)||68 (40–90)||66 (32–82)||63.5 (37–87)||62 (37–89)||0.08||59 (22–95)||57 (30–93)||58 (31–77)||53 (29–64))||0.0005|
|Male||9 (25%)||20 (56%)||20 (56%)||23 (64%)||0.005||35(38%)||40 (43%)||47 (51%)||68 (75%)||<0.0001|
|Female||27 (75%)||16 (44%)||16 (44%)||13 (36%)||57 (82%)||52 (57%)||45 (49%)||23 (25%)|
|Caucasian||33 (92%)||34 (94%)||35 (97%)||28 (78%)||0.22||71 (77%)||80 (87%)||81 (88%)||74 (81%)||0.34|
|All others||3 (8%)||2 (8%)||1 (3%)||8 (22%)||21 (23%)||12 (13%)||11 (12%)||17 (19%)|
|Local||19 (54%)||18 (51%)||21(58%)||13 (37%)||0.48||40 (44%)||40 (43%)||36 (40%)||23 (26%)||0.18|
|Regional||11 (31%)||14 (40%)||10 (28%)||18 (51%)||31 (41%)||41 (45%)||42 (46%)||52 (58%)|
|Remote||5 (14%)||3 (9%)||5 (14%)||4 (11%)||13 (14%)||11 (12%)||13 (14%)||14 (16%)|
|Colon||24 (87%)||28 (78%)||22 (61%)||27 (75%)||0.39||55 (60%)||57 (62%)||57 (62%)||55 (80%)||0.99|
|Rectum||12 (33%)||8 (22%)||14 (39%)||9 (25%)||34 (37%)||32 (35%)||33 (36%)||34 (37%)|
|NOS||–||–||–||–||3(3%)||3 (3%)||2 (2%)||2 (2%)|
|Adenoca||35 (97%)||32 (89%)||32 (89%)||29 (81%)||0.23||84 (91%)||85 (92%)||84 (92%)||83 (91%)||0.74|
|Other||1 (3%)||4 (11%)||4 (11%)||7 (19%)||8 (9%)||7 (8%)||7 (8%)||8 (9%)|
|Grade 1||7 (21%)||4 (13%)||7 (21%)||7 (21%)||0.32||14 (16%)||11 (13%)||11 (13%)||7 (8%)||0.75|
|Grade 2||21 (84%)||25 (78%)||23 (70%)||18 (53%)||60 (70%)||61 (71%)||57 (70%)||59 (72%)|
|Grade 3/4||5 (15%)||3 (9%)||3 (9%)||9 (26%)||12 (14%)||14 (16%)||14 (17%)||16 (20%)|
|Surgery||34 (94%)||35 (97%)||33 (92%)||34 (94%)||0.79||86 (94%)||88 (96%)||83 (91%)||82 (90%)|
|Radiation||2 (8%)||3 (8%)||6 (17%)||5 (14%)||0.42||11 (12%)||15 (16%)||17 (19%)||19 (21%)|
|Chemo||11 (34%)||11 (31%)||13 (37%)||15 (44%)||0.69||42 (49%||35 (38%)||52 (60%)||53 (81%)|
|Fiber (g/day)||9.0 ± 0.9||9.5 ± 0.7||12.3 ± 0.9||15.9 ± 1.2||<0.0001||9.5 ± 0.6||11.0 ± 0.5||123 ± 0.5||16.3 ± 1.1||<0.0001|
|Calcium (mg/day)||448 ± 48||542 ± 53||867 ± 86||972 ± 169||<0.0001||524 ± 45||660 ± 42||724 ± 52||934 ± 62||<0.0001|
|Iron (mg/day)||6.8 ± 0.6||8.8 ± 0.6||11.9 ± 0.6||15.1 ± 0.9||<0.0001||8.1 ± 0.4||9.8 ± 0.4||12.1 ± 0.4||18.1 ± 0.9||<0.0001|
|Energy (kcal/day)||853 ± 72||1161 ± 83||27.0 ± 0.8||1712 ± 92||<0.0001||1079 ± 52||25.8 ± 0.5||1365 ± 45||25.6 ± 0.6||<0.0001|
|BMI (kg/m2)||25.4 ± 0.8||26.6 ± 0.9||2183 ± 174||24.9 ± 0.7||0.31||1709 ± 56||26.3 ± 0.5||2634 ± 123||27.9 ± 0.7||0.046|
|Total arginine (g/week)||10.1 ± 0.9||15.3 ± 0.9||23.5 ± 0.9||34.8 ± 2.1||<0.0001||12.6 ± 0.7||19.3 ± 0.5||25.5 ± 0.7||38.6 ± 1.5||<0.0001|
|Meat-derived arginine (g/week)||3.2 ± 0.3||7.1 ± 0.3||11.0 ± 0.3||19.2 ± 0.9||<0.0001||4.0 ±0.2||9.1 ± 0.2||137 ± 0.3||22.8 ± 1.0||<0.0001|
Relationship between meat consumption and total daily arginine intake.
The relationship between meat consumption and total daily arginine intake was investigated analytically. For the overall group (n = 511), mean arginine intake from all foods was estimated at 23.2 g/week (i.e. 3.3 g/day). The median meat-derived arginine intake was estimated at 11.7 g/week (1.7 g/day). Forty-nine percent of total daily arginine intake in this study was due to meat consumption. Meat consumption was revealed to be a predictor of total daily arginine intake in linear regression analysis (R2 = 0.77, p < 0.0001) (Fig. 3).
Effects of meat intake on clinical outcomes.
For familial cases, no differences in tumor grade (p = 0.32) or stage at presentation (p = 0.48) were observed based on meat consumption quartile. Similarly, among sporadic CRC cases, no differences in tumor grade (p = 0.75) or stage at presentation (p = 0.18) were noted based on meat consumption quartile (Table IV). Among familial CRC cases, overall survival (OS) was significantly decreased for cases in the highest meat consumption quartile (n = 36) vs. cases in all other meat consumption quartiles (n = 108) (quartile 4, 10-year OS = 42%; Q1–Q3, 10-year OS = 65%; p = 0.017) (Fig. 4). No statistically significant differences in OS were detected for cases in meat consumption quartiles Q1–Q3 (p = 0.34). Subset analysis revealed that among the 101 colon cancer cases, borderline significant decreased OS was observed for cases in the highest meat consumption quartile (n = 74) vs. cases in all other meat consumption quartiles (n = 27) (Q4, 10-year OS = 44%; Q1–Q3, 10-year OS = 70%; p = 0.052). No statistically significant differences were noted for the 43 familial rectal cancer cases based on meat consumption group (Q4, 10-year OS = 33%; Q1–Q3, 10-year OS = 60%; p = 0.19); however, these subset analyses were underpowered as there were only 9 rectal cancer patients in meat consumption Q4. Among sporadic CRC cases, OS was similar for cases in the highest meat consumption quartile (n = 91) compared to cases in all other meat consumption quartiles (n = 276) (Q4, 10-year OS = 62%; Q1–Q3, 10-year OS = 62%; p = 0.89.
Variables known to predict survival in CRC were included into the multivariate survival model, including age, gender and stage at presentation. The following variables were highly correlated with meat consumption quartile but were not significantly associated with survival: daily fiber intake, daily calcium intake, daily iron intake and total daily energy intake—thus they were not included in the final multivariate analysis. For familial CRC cases, the highest meat consumption quartile was associated with an increased hazard ratio (HR) of 2.24, 95% C.I. 1.25–4.03, (p = 0.007) compared to familial CRC cases in all other meat consumption quartiles after adjustment for age, gender and stage at presentation (Table V). The trend test for meat consumption quartile among familial CRC cases was not statistically significant (HR = 1.08, 95% CI 0.84–1.39; p = 0.57). For sporadic CRC cases, no statistically significant increased hazard was noted for the highest quartile of meat consumption group (HR = 1.01, 95% CI 0.67–1.51; p = 0.98) compared to sporadic CRC cases in all other meat consumption quartiles after adjustment for age, gender and stage at presentation (Table V).
|Familial CRC (n = 144; deaths = 6l)||Sporadic CRC patients (n = 367; deaths = l42)|
|HR||95% HR confidence limits||p||HR||95% HR confidence Limits||p|
|Meat Consumption Quartile|
|Quartile 4 (highest)||2.24||(1.25–4.03)||0.007||1.01||(0.67–1.51)||0.98|
The results presented in this study demonstrate that meat and its major component, arginine, increase tumor grade in mice and are associated with adverse outcomes among familial CRC cases. Among familial CRC cases, those with the highest quartile of meat consumption prior to diagnosis (i.e. 10.8–27.4 servings per week) had significantly decreased survival compared to cases in all other meat consumption quartiles in univariate and adjusted analyses. The different pathways involved in arginine metabolism are presented in Figure 5. As depicted, arginine can be converted directly to polyamines by intestinal bacteria,6 but this route is not essential, as the hepatic and extra-hepatic arginases catabolize arginine to ornithine, which is itself converted by ODC to form polyamines. The arginase enzymes compete with NOS, which converts arginine to nitric oxide and citrulline.
Arginine supplementation increased high grade intestinal tumor incidence in ApcMin/+Nos2+/+ mice, but not in Nos2 knockout mice. Thus, Nos2 is involved in the arginine-dependent increase in colon adenoma grade. In accordance, a separate study from our laboratory described dietary arginine-induced increases in overall colon tumor incidence and number that were Nos2-dependent.15 In that study, dietary arginine was found to primarily affect colonic tumor initiation in ApcMin/+ mice. Similarly, our present findings suggest that supplemental arginine-induced tumorigenesis is mediated through NO synthase. Assessment of polymorphisms in the NOS genes and promoters could potentially help explain our epidemiologic findings, which implicate gene–environment interactions between meat intake and improved survival for familial CRC cases but not sporadic CRC cases. Currently, Nos2 single nucleotide gene polymorphisms with functional promoter activity have been described,35, 36, 37 but have not been reported in CRC patients.
As shown in Table III, fish, poultry and pork contain comparable amounts of arginine to red meat. We have demonstrated in this study that meat consumption represents a reasonable estimator of total arginine intake (Fig. 3). Others have shown that arginine estimates from nutrient composition tables and short-term food intake records correlate with serum arginine levels.6 Even so, the poor outcomes associated with the highest quartile of meat consumption noted in our study may reflect the effects meat constituents other than arginine, or different cooking methods that could not be accounted for in this analysis. Heterocyclic amines (HCAs) formed in meat during cooking,38, 39, 40, 41 polycyclic aromatic hydrocarbons (PAHs),42 and N-nitroso compounds (NOCs, i.e. from processed meat)43 are known carcinogens that have been implicated in colorectal carcinogenesis, and could account for the observed survival effects of the present study. Red meat intake and also intake of heme iron has been associated with high NOC levels, thus implicating a role for iron in carcinogenesis.44 The FFQ used in our study is not adequately designed to assess cooking methods or level of doneness, thus we were unable to assess the effects of HCAs, PAHs or NOCs. However, a FFQ model has recently been developed by others to assess intake of these carcinogenic compounds, and future outcomes-based studies using this type of model may reveal the effect of these compounds on survival in CRC.45 Genotypic differences in UDP-glucuronosyltransferase-1A7 have been reported to modify the risk of colon cancer to dietary carcinogen exposures,46 and single nucleotide polymorphisms in N-acetyltransferase-2 have been reported to modify the metabolic activation of heterocyclic amine carcinogens.47 Therefore, the observed survival differences based on reported meat consumption among familial CRC cases may be related to different distributions of such genes involving heterocyclic amine metabolism.
Researchers from our group have shown that arginine increases intestinal polyamine levels in ApcMin/+ mice, which can be inhibited by treatment with DFMO.48 Additionally, DFMO inhibits the arginine-induced increases in colonic tumor burden and tumor grade in ApcMin/+ mice. DFMO has been shown previously in clinical trials to decrease colorectal mucosa polyamine content in humans,20 and such findings serve as the basis for ongoing CRC clinical trials using DFMO as a chemopreventive agent.49 Our experimental data revealed that celecoxib induced expression of Ssat, lowered Odc levels and suppressed putrescine levels in the intestines of ApcMin/+ mice. Additionally, celecoxib significantly decreased the intestinal tumor number in ApcMin/+ mice–an effect augmented by DFMO. Ssat induction by celecoxib promotes polyamine catabolism, and Odc-repression by celecoxib and DFMO leads to decreases in polyamine synthesis, as depicted in Figure 5, with a resultant decreased tumor incidence. An important finding in this study is that the celecoxib- and DFMO-inhibitory effects on tumorigenesis in ApcMin/+ mice operate in an additive manner. This reflects the inhibitory effects of DFMO and celecoxib on separate pathways involved in polyamine synthesis. This finding correlates well with reports on NSAIDs and DFMO in experimental models of tumorigenesis, and with the design of phase II chemoprevention clinical trials incorporating the use of NSAIDs and DFMO for CRC prevention.22, 49
The exact mechanisms for celecoxib-induced transcriptional activation of Ssat are unknown. However, transcription of Ssat has been demonstrated to occur by COX-dependent and COX-independent mechanisms.22, 23 This may involve activation via peroxisome-proliferator-activated receptor-γ (PPARγ), decreased polyamine levels and apoptosis in colon cancer cells.22 Alternatively, celecoxib treatment may result in activation of NFκB signaling and subsequent binding to Ssat promoter sites.50
Our findings implicate gene–environment interactions between meat consumption, and presently unknown genotypes among affected individuals with multifactorial inherited susceptibility to CRC. Our in vivo results showing the additive benefits of DFMO and celecoxib in mouse intestinal tumorigenesis models combined with epidemiologic data demonstrating poor survival for familial CRC cases that were the highest consumers of meat support further molecular epidemiologic investigations into genotypic analysis of CRC cases. Furthermore, these findings add to the growing support for development of therapeutic CRC clinical trials involving arginine depletion via dietary modification, NOS2 inhibition using targeted molecular therapies and multi-targeted polyamine inhibition in select patient populations.
The authors thank Karen A. Blohm-Mangone, Peggy Jane Criswell and Jose L. Padilla-Torres for assistance with breeding and genotyping mice. The authors also thank Frank Meyskens for his critical reading of the manuscript and thoughtful comments about potential mechanisms underlying our findings.
- 4US Department of Agriculture, Agricultural Research Service. USDA National Nutrient Database for Standard Reference, Release 18, 2005. Nutrient Data Laboratory Home Page, http://www.nal.usda.gov/fnic/foodcomp.
- 24Formula for laboratory animals, 3rd edn. Ames, IA: Blackwell, 2005., , .
- 30Wine consumption prior to diagnosis improves survival in familial colorectal cancer patients. Presented at AACR colorectal cancer: molecular pathways and therapies—Conference Proceedings. October 21, 2005, Dana Point, CA. 2005. Abstract B21., , , , , .
- 31Cancer reporting in California, vol. 1.Abstracting and coding procedures for hospitals. Emeryville, CA: California Tumor Registry, Department of Health Services, Cancer Surveillance Section, 1986.
- 32International classification of diseases for oncology, 3rd edn. Geneva: World Health Organization, 2000., , , , , , .