![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
,, **,, **,
* Department of Poultry Science, Department of Animal Science and ** Graduate Faculty of Nutrition,
Texas A&M University, College Station, TX 77843-2472
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
FOOTNOTES
LITERATURE CITED
ABSTRACT 
An analysis of viable bacterial populations enumerated on carbohydrate selective media was used to simulate the colonic environment in vitro and determine if differential media could detect significant microbial shifts due to dietary fiber source, dietary fat source, and carcinogen. Male Sprague-Dawley rats were provided with either pectin or cellulose as a fiber source, either corn or fish oil as a source of fatty acids, and injected with either azoxymethane (AOM), a gastrointestinal carcinogen, or saline in a 2 × 2 × 2 factorial design. At 6 and 10 mo of age, fresh feces were collected, homogenized in anaerobic buffer and anaerobically plated onto differential media. Diets containing pectin supported more anaerobes at 6 mo of age (P < 0.01) than diets containing cellulose. Rats injected with AOM and consuming either pectin or corn oil supported more anaerobes at 10 mo of age (P < 0.05) than rats injected with saline and consuming the same diets. Rats consuming cellulose and receiving AOM but not expressing tumors possessed larger anaerobic populations at 10 mo of age (P < 0.05) than rats consuming cellulose, injected with AOM and expressing tumors. These effects show that gastrointestinal bacterial populations, as measured by carbohydrate specific media, respond to dietary changes such as dietary fiber source, and thus may play a key role in the etiology of colon cancer.
Key words: azoxymethane, bacteria, colon, fiber, rats, cancer.INTRODUCTION
In 1995, approximately 138,000 new cases of colon and rectum cancer were reported, and 55,000 people died of colon and rectum cancer (Wingo et al. 1995
). Between 80 and 90% of colon cancers may be attributed to environmental factors (Drasar and Hill 1972). For this reason, researchers have concentrated mainly on factors involved in the promotion of colon cancer, with an emphasis on prevention through diet modification.
Drasar and Hill (1972) hypothesized that the resident colonic bacterial population may modify diet components or gastrointestinal secretions to form carcinogens or cocarcinogens, or activate procarcinogens (Moore and Moore 1995). In response to this hypothesis, intensive research has been conducted in an attempt to define the colonic ecosystem and, in particular, the species associated with carcinogen or cocarcinogen production (Finegold et al. 1974, 1975 and 1977, Moore and Holdeman 1974). Microbial mass was found to comprise more than 30% of the wet volume of feces (Moore and Holdeman 1974) or between 1011 and 1012 bacteria/g dry weight colonic contents (Cummings and MacFarlane 1991). Finegold et al. (1974) isolated over 220 distinct species, with numerous additional isolates that could not be identified to the species level, and detected several diet-based microfloral differences (Finegold et al. 1977). However, a National Institutes of Health study section concluded that a data set involving more than 200 species of colonic bacteria was too complex to interpret (Moore and Moore 1995).
In vitro studies of selected strains of colonic microbes revealed that, despite the taxonomic complexity indicated from isolation studies, the functional metabolism of the colonic microbial population is reproducible and predictable. Freter et al. (1983b) reported an ability to coculture 37 specific strict anaerobes in continuous coculture and demonstrated a strong similarity between the ecology within the mouse large intestine and the in vitro culture vessel. From these studies, Freter et al. (1983a) hypothesized that bacterial species in the colon utilize an affinity for specific substrates to survive and suggested that environmental factors such as high hydrogen sulfide concentrations, low pH, and strict anaerobiosis (Rumney and Rowland 1992) may also contribute to an ecological balance. Thus, diet-induced changes in the substrates reaching the colon could shift colonic total microfloral numbers or population characteristics (Monsma et al. 1992).
Differential media containing dietary carbohydrates may detect microbial shifts not observed using clinical media. A differential medium supports only those microbes able to use the substrates present in the medium (Leedle and Hespell 1980). Therefore, it is possible to characterize and estimate population sizes based on substrate preference using media containing only those carbohydrates present in vivo. Rumen
bacteria (Dehority and Grubb 1976, Leedle and Hespell 1980), cecal bacteria in pigs (Allison et al. 1979) and cecal microflora from chickens (Fan et al. 1995) have all been successfully characterized using this approach. Furthermore, because the predominant species in the colon are strict anaerobes (Rumney and Rowland 1992), anaerobic techniques will permit greater sensitivity for detection of dietary influences on colonic populations. The objective of this study, therefore, was to detect potential microfloral population shifts due to diet and carcinogen by a noninvasive procedure utilizing differential media to simulate the colonic luminal environment.
MATERIALS AND METHODS
). Rats were given free access to diets and water, and all animal handling procedures were approved by the Texas A&M University Laboratory Animal Care Committee.
|
Table 1. Composition of experimental diets |
). Citrus pectin (polygalacturonic acid methyl ester, Grinstead, Industrial Airport, KS) was washed with two changes of ethanol (700 mL/L) to remove impurities and detached pectin subunits and dried at 37°C (Leedle and Hespell 1980). Vitamin stocks were stored at 
20°C until use, and the purified pectin and reductants were stored at room temperature.
Media preparation.
Ingredients were mixed and boiled to melt agar and reduce dissolved oxygen levels (Table 2). All media components were reagent grade and were acquired from Sigma Chemical Co. (St. Louis, MO) except for yeast extract, peptone, agar (Difco Laboratories, Detroit, MI), trypticase (Becton Dickinson Microbiology Systems, Cockeyville, MD) and pectin. Immediately after autoclaving for 30 min, vitamin and reductant stocks were aseptically added. Dyes and oils were aseptically added as filter-sterilized solutions to media after autoclaving.
|
Table 2. Composition of media used for the enumeration of bacterial populations in rat feces |
). Components were mixed, autoclaved, and allowed to cool under a stream of CO2 before the addition (10 mL/L) of a sterile 25 g/L solution of cysteine hydrochloride. The buffer was placed in the anaerobic chamber and allowed to reduce overnight, indicated by a loss of pinkish tint. The diluent was aseptically dispensed into 4.5 mL aliquots.
Sample preparation.
Fecal pellets were collected from a subsample of rats at 6 and 10 mo of age. The n from each group at 6 and 10 mo of age, respectively, were as follows: CO-CEL-AOM, 15, 9; CO-PEC-AOM, 9, 8; FO-CEL-AOM, 8, 10; FO-PEC-AOM, 7, 9; CO-CEL-saline, 11, 13; CO-PEC-saline, 7, 9; FO-CEL-saline, 7, 9; and FO-PEC-saline, 5, 11. On each day of sampling, fecal samples were collected from between 4 and 8 rats within 2 h of defecation. The samples were placed into sterile, preweighed, polyvinyl chloride tubes using sterile tweezers and immediately transported on ice to the lab. Fecal pellets were weighed and moved into an anaerobic chamber. The collected feces were placed in 3-mL aliquots of anaerobic phosphate buffer described above containing 10 glass beads and left at room temperature for 1 h to loosen pellets. The sample was vortexed until completely suspended, and a 1:10 dilution series was prepared by standard methods (Swanson et al. 1992). Three dilutions were chosen from preliminary experiments, which produced anaerobes between 9.6 × 108 and 1.2 × 109 anaerobe colony forming units (CFU)/g wet feces (data not shown). Dilutions for differential populations were chosen assuming carbohydrate selective anaerobe numbers to be lower than those on a complete carbohydrate (CC) medium, with an inoculation volume of 0.05 mL. Two replicates of each media were inoculated from each dilution, for a total of six plates/rat. The inoculant was spread across the plates by shaking with glass beads. The plates were inverted and incubated anaerobically at 37°C for 72 h.
Congo red assay.
After enumeration of total colony numbers, carboxymethylcellulose (CMC) differential plates were flooded with 1 g Congo red/L and left at room temperature for 15 min to determine which of the colonies present were true cellulolytics vs. those that utilized the by-products of cellulose fermentation (Teather and Wood 1982). The dye was decanted, and the plates were destained with 1 mol NaCl/L at room temperature for 15 min. Colonies that degraded CMC were surrounded by clear zones on an opaque, red plate.
Cetavlon assay.
Preliminary experiments revealed that an incubation for 20-30 min (Jayasankar and Graham 1970) was not sufficient to produce clear zones of lysis. Therefore, pectin differential plates were flooded with cetavlon (cetrimonium bromide) solution, 10 g/L, after enumeration and left at room temperature for 2-3 h to
determine how many of the total number were true pectinolytics. The plates were shaken, and the dye was decanted. Colonies that degraded washed pectin were indicated by clear zones on an opaque, white plate.
Calculations.
Pectin is a highly fermentable fiber compared to cellulose which is only slightly fermented in the large intestine of rats (Gazzaniga and Lupton 1987). Because of the resulting differences in material presented to the lower colon, there are large differences in fecal excretion patterns. Thus, we performed total fecal collections for 48 h, determined microbial numbers on a wet fecal weight basis, and expressed anaerobe numbers/gram wet feces over the 48-h period. This adjustment prevented biasing the data because of differences in microbial excretion at any one time point. The resulting population estimates, in units of CFU/48 h, were converted using a log10 function for statistical analysis.
Statistical analyses.
Because rats were used for other terminal procedures after fecal collection, the two sampling times represent separate rats and, therefore, the data were analyzed as two distinct sets rather than as a repeated measures design (GLM procedure of SAS, version 6.08, Cary, NC). Overall model adequacy was considered significant at P < 0.05. If interactions were not significant (P 
0.25), they were removed from the model. The effect of tumors was analyzed by dividing the AOM injected rats into two separate groups, those that possessed tumors at the time of kill and those that did not. The data set was then reanalyzed as a 2 × 2 × 3 factorial (2 fibers; 2 fats; with or without tumors with carcinogen, without carcinogen). If tumor incidence had no effect on the data (P > 0.25), then the data was analyzed as a 2 × 2 × 2 factorial design. Differences in least squares means were determined using a Fisher's protected least significant difference test (Steel and Torrie 1980).
RESULTS
0.05). Rats injected with AOM had twice the number of true pectinolytics than rats injected with saline (P < 0.02) as detected by applying the cetavlon assay to PEC differential medium.
Ten months of age. Due to the lack of dietary or carcinogen effects on CMC differential medium at 6 mo of age, hydrolytic activity was also investigated using the Congo red assay at the 10 mo sampling. CMC differential medium, PEC differential medium, the cetavlon assay and the Congo red assay did not detect any significant differences due to diet or carcinogen injection in populations at 10 mo of age (Figure 2, P
0.05). However, PEC-fed rats receiving an AOM injection had nearly 5 times the number of CC anaerobes compared to CEL-fed rats receiving an AOM injection and PEC-fed rats receiving a saline injection (Figure 3a, P < 0.05). Rats ingesting CEL and receiving a saline injection had intermediate numbers of CC anaerobes. Rats receiving CO and an AOM injection possessed CC populations which were 3.8 times that of CO consuming rats injected with saline (Figure 3b, P < 0.05). Rats consuming FO,
regardless of injection group, had intermediate numbers of CC anaerobes compared to rats ingesting CO.
Rats with tumors at ten months of age. The carcinogen group was divided into two groups, those that possessed colon tumors and those that did not, and reanalyzed. None of the sampled rats receiving a saline injection developed colon tumors. The presence of tumors resulted in interactions between tumor incidence and either fat and fiber sources with respect to CC numbers. The n of each group were as follows: CEL-saline, 22, CEL-AOM-tumors, 12, CEL-AOM-no tumors, 7, PEC-saline, 20, PEC-AOM-tumors, 11, PEC-AOM-no tumors, 6, CO-saline, 22, CO-AOM-tumors, 13, CO-AOM-no tumors, 4, FO-saline, 20, FO-AOM-tumors, 10, FO-AOM-no tumors, 9. No difference was detected in CC numbers of saline injected rats consuming either PEC or CEL (Figure 4a). Rats consuming PEC and injected with AOM had between 4.8 and 7.1 times more CC anaerobes than rats consuming PEC and injected with saline (P < 0.05). However, CEL rats injected with AOM had diverse responses, depending on whether tumors were present. When tumors developed, CC numbers were 72% less than in saline injected rats (P < 0.05), whereas when tumors did not develop, CC numbers were numerically greater than, but similar to, those in saline injected CEL rats. Anaerobe numbers on the CC plates were in the range of 2.88 to 4.90 × 109 CFU/48 h for all rats consuming FO (Figure 4b). Similar numbers occurred in CO rats, except for the non-tumor-bearing, AOM-injected rats, which had greater numbers compared to all other groups (P < 0.05).
DISCUSSION
and 1975, Macy et al. 1982, Rumney and Rowland 1992). However, the numbers are consistent with the mean total cecal population of male Sprague-Dawley rats of 9 × 109 CFU/g reported by Macy et al. (1982). The differences may reflect microfloral differences between animal species, as bacterial
numbers may be lower in the rat model than are isolated from fecal samples from human subjects (Rumney and Rowland 1992). Bacterial viability may have also been reduced by fecal sampling. Monsma and Marlett (1995) reported that the collection site of rat microflora affects the fermentation patterns of in vitro fermentation. Enumerations on basal media without carbohydrates yielded populations (1.4-1.9 × 108 CFU/48 h,
data not shown) between 1 and 10% of CC anaerobic populations for each rat sampled. This basal population was consistent across dietary treatment and was probably due to the use of noncarbohydrate substrates such as the cysteine reductant, yeast extract, and trypticase (Dehority and Grubb 1976).
Leedle and Hespell (1980) also observed that basal medium without carbohydrate substrates supported numerous pinpoint colonies. For this reason, any colonies under 1 mm in diameter were not included in the determination of basal and differential populations in this study.
). However, no significant trends were noted in this study using CMC differential medium. Several factors may account for this. Microbes present in a fecal pellet high in cellulose may remain attached to cellulose particles, underestimating enumeration by colony growth on CMC media (Weimer 1992). Strict anaerobes on the surface of the fecal pellet may have become nonviable due to exposure to oxygen. It is also possible that, as cellulose is a poorly digested substrate, the microflora may have been driven to utilize more endogenous sources of nutrients. This attached population would probably be present in the feces in lower numbers, except for dead or dying cells.
), and subunits released from pectin fermentation would provide a more available source of energy to a broader range of bacterial groups than cellulose. Fermentation of pectin in the colon promoted bacterial growth relative to cellulose, demonstrating that a change in diet can effectively alter the colonic flora. Six-month-old rats injected with AOM had somewhat fewer CC numbers (P > 0.1) and greater true pectinolytic numbers. The reduction in CC bacterial numbers with AOM injection, though not statistically significant, parallels the reduction in total short chain fatty acids, acetate, and butyrate concentrations we have observed in rats treated with the same dietary and injection protocols (Zoran et al. 1996). In addition, other work from our lab has determined that AOM depresses metabolism of glucose and butyrate by colonocytes from AOM-injected rats at 4 mo of age (Zhang et al. 1996), suggesting a potential change in colonocyte metabolic set points due to depressed substrate supply.
previously demonstrated that butyrate and acetate concentrations in the colon luminal contents of rats consuming FO are greater than those consuming CO. Even though there is a trend towards greater populations on pectin differential media with PEC diets vs. CEL diets (P < 0.08), the difference may be due to bacterial populations which benefit from and utilize the products of pectinolytic metabolism, as suggested by Osborne and Dehority (1989). Thus, bacterial populations shifted with a change in fiber source.
reported that rats consuming CO and injected with AOM had greater total short-chain fatty acids and propionate than saline injected CO consuming rats, which may be a result of the greater populations observed in this study. When the data were evaluated with respect to whether tumors were present in AOM injected rats, we found that non-tumor bearing rats consuming CEL or CO had greater CC numbers than tumor-bearing rats consuming the same diets. Zhang et al. (1996) reported a depression in colonocyte metabolism of glucose, butyrate, and glutamine from non-tumor bearing rats compared to saline injected rats, whereas colonocyte metabolism of these nutrients by AOM injected, tumor-bearing rats was intermediate. Bacterial numbers on CC
media were the same in AOM injected rats with tumors or saline injected rats regardless of the oil source consumed. However, fewer numbers were detected with CEL consumption in comparison to rats consuming PEC.
). Pectin as a fiber source has been shown to promote topological damage and partial brush border loss in the large intestines of rats (Cassidy et al. 1981). Pectin also has been reported to promote cell proliferation in carcinogen-exposed rats (Jacobs and Lupton 1986) and exfoliated cell loss from the rat colon (Davidson et al. 1995). Mucosal cell loss, potentially due to exposure to carcinogens or tumor growth, may provide numerous substrates to the microflora.
FOOTNOTES
Manuscript received 16 September 1996. Initial reviews completed 15 October 1996. Revision accepted 14 November 1996.
LITERATURE CITED
This article has been cited by other articles:
|
|
 |
|
  K. C. Crim, L. M. Sanders, M. Y. Hong, S. S. Taddeo, N. D. Turner, R. S. Chapkin, and J. R. Lupton Upregulation of p21Waf1/Cip1 expression in vivo by butyrate administration can be chemoprotective or chemopromotive depending on the lipid component of the diet Carcinogenesis, July 1, 2008; 29(7): 1415 - 1420. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. M. Sanders, C. E. Henderson, M. Y. Hong, R. Barhoumi, R. C. Burghardt, N. Wang, C. M. Spinka, R. J. Carroll, N. D. Turner, R. S. Chapkin, et al. An Increase in Reactive Oxygen Species by Dietary Fish Oil Coupled with the Attenuation of Antioxidant Defenses by Dietary Pectin Enhances Rat Colonocyte Apoptosis J. Nutr., December 1, 2004; 134(12): 3233 - 3238. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. J. Coleman, E. K. Landstrom, P. J. Royle, A. R. Bird, and G. H. McIntosh A Diet Containing {alpha}-Cellulose and Fish Oil Reduces Aberrant Crypt Foci Formation and Modulates Other Possible Markers for Colon Cancer Risk in Azoxymethane-Treated Rats J. Nutr., August 1, 2002; 132(8): 2312 - 2318. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. L. Topping and P. M. Clifton Short-Chain Fatty Acids and Human Colonic Function: Roles of Resistant Starch and Nonstarch Polysaccharides Physiol Rev, July 1, 2001; 81(3): 1031 - 1064. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P Perrin, F Pierre, Y Patry, M Champ, M Berreur, G Pradal, F Bornet, K Meflah, and J Menanteau Only fibres promoting a stable butyrate producing colonic ecosystem decrease the rate of aberrant crypt foci in rats Gut, January 1, 2001; 48(1): 53 - 61. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Zhang, G. Wu, R. S. Chapkin, and J. R. Lupton Energy Metabolism of Rat Colonocytes Changes during the Tumorigenic Process and Is Dependent on Diet and Carcinogen J. Nutr., August 1, 1998; 128(8): 1262 - 1269. [Abstract] [Full Text]
|
|
|
|
|
|
W.-C. L. Chang, R. S. Chapkin, and J. R. Lupton Fish Oil Blocks Azoxymethane-Induced Rat Colon Tumorigenesis by Increasing Cell Differentiation and Apoptosis Rather Than Decreasing Cell Proliferation J. Nutr., March 1, 1998; 128(3): 491 - 497. [Abstract] [Full Text]
|
|
|
|
|
|
D. L. Zoran, N. D. Turner, S. S. Taddeo, R. S. Chapkin, and J. R. Lupton Wheat Bran Diet Reduces Tumor Incidence in a Rat Model of Colon Cancer Independent of Effects on Distal Luminal Butyrate Concentrations J. Nutr., November 1, 1997; 127(11): 2217 - 2225. [Abstract] [Full Text]
|
|
|
|
|
|
Y.-H. Jiang, J. R. Lupton, and R. S. Chapkin Dietary Fat and Fiber Modulate the Effect of Carcinogen on Colonic Protein Kinase C lambda Expression in Rats J. Nutr., October 1, 1997; 127(10): 1938 - 1943. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
