The Journal of Nutrition Vol. 128 No. 5 May 1998, pp. 843-847

Stimulation of Epithelial Cell Proliferation of Isolated Distal Colon of Rats by Continuous Colonic Infusion of Ammonia or Short-Chain Fatty Acids Is Nonadditive¹

Hirofumi Ichikawa^* and Takashi Sakata^{, 2}

^* The Second Department of Surgery, Tohoku University School of Medicine, Sendai 980, Japan and  Department of Basic Sciences, Ishinomaki Senshu University, Ishinomaki 986-8580, Japan

	ABSTRACT

Abstract Introduction Methods Results Discussion References

Dietary fibers accelerate colonic epithelial cell proliferation at least in part by modulating bacterial metabolism in the large intestine. Ammonia and short-chain fatty acids (SCFA) are major metabolites of hindgut bacteria and are believed to affect epithelial cell kinetics of the colon. However, the effect of luminal ammonia itself and the possible interaction of ammonia with SCFA on colonic epithelial cell proliferation have not yet been studied. The colon of rats was surgically isolated and continuously administered infusates with saline, ammonia, SCFA or both into the isolated colon for 7 d in a two-way factorial design. On d 7, vincrystine sulfate was administered intravenously to cause metaphase arrest. The activity of epithelial cell proliferation in the distal colon was estimated by using a stathmokinetic method and by histologic examination. The crypt size was significantly larger in rats given infusates containing SCFA than in rats given infusates without SCFA. Infusion of ammonia or SCFA significantly stimulated colonic epithelial cell proliferation compared with the saline infusion. Infusion of both ammonia and SCFA resulted in accumulated mitoses per crypt that did not differ from the other three infusions although the value tended to be lower than when SCFA alone were infused. Thus, stimulation of epithelial cell proliferation by ammonia and SCFA is not additive, and the interaction between them should be considered when the effects of dietary fibers on gut epithelial proliferation are investigated.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

Bacterial degradation of nitrogenous substrates such as protein, nucleic acid and urea in the hindgut produces ammonia³ (Macfarlane et al. 1986, Visek 1972, Wrong et al. 1985). The concentration of ammonia in the hindgut reaches 70 mmol/L in rats (Lin and Visek 1991a) and 60 mmol/kg of luminal contents in humans (Macfarlane et al. 1986). Many dietary fibers provide energy to hindgut bacteria for their growth and metabolism. This results in the release of short-chain fatty acids (SCFA) in the hindgut lumen and often reduces the luminal concentration of ammonia by increasing utilization of ammonia by the bacteria.

SCFA such as acetic, propionic and n-butyric acids stimulate epithelial cell proliferation of the large intestine in rats (Kripke et al. 1989, Sakata 1987) and humans (Scheppach et al. 1992). On the other hand, ammonia is potentially toxic to cells; it shortens cell life span and alters DNA synthesis in various tissues including the ileum and colon (Visek 1972 and 1978). Ammonia is also thought to promote colon carcinogenesis (Clinton et al. 1988, Topping and Visek 1976). Moreover, acute perfusion of ammonium salts injures colonic mucosa in rats (Lin and Visek 1991b). However, the direct influence of physiological concentrations of luminal ammonia on the colonic epithelial cell proliferation remains obscure.

SCFA and ammonia co-exist in the hindgut lumen and affect the absorption (Bödecker 1996) and metabolism by colonocytes of one another (Darcy-Vrillion et al. 1996). Such an interaction of different luminal trophic factors should be taken into account when we consider realistic in vivo conditions in the hindgut lumen.

Accordingly, the aim of this study was to investigate the individual effects and a possible interactive effect of luminal ammonia and SCFA on colonic epithelial cell proliferation. For this purpose, we infused ammonia, a SCFA mixture or both at their physiological concentrations into surgically isolated rat colons for 7 d. A two-way factorial design (ammonia × SCFA) was used.

	MATERIALS AND METHODS

Abstract Introduction Methods Results Discussion References

Animals.  Twenty-four male Wistar rats, 8 wk of age and weighing ~250 g, were purchased from Funabashi Farm (Funabashi, Japan). They were housed in stainless steel wire mesh-bottomed cages (3 or 4 rats per cage) and maintained under a 12-h light:dark cycle. Rats were allowed free access to a nonpurified diet⁴ (CA-1, Clea Japan, Tokyo, Japan) and water for 7-10 d before surgical operation. The animal care and the following experimental procedures were performed according to the guidelines of the Institute for Experimental Animals, Tohoku University School of Medicine.

Surgical operation.  Rats were deprived of food overnight before surgery. After an intraperitoneal injection of sodium pentobarbital (30 mg/kg body weight), we performed a midline laparotomy.

The terminal ileum was sectioned 1 cm proximal to the ileocecal junction (Fig. 1). Ileocecal vessels were ligated and severed close to the cecocolonic junction. The cecum was then resected. The cut end of the proximal colon was closed with five or six interrupted sutures.

View larger version (29K): [in this window] [in a new window]   Fig 1. Schematic presentation of the surgical operation performed on rats. We isolated the entire colon from the stream of digesta, resected the cecum, reconstructed bowel continuity, inserted a tube for continuous infusion into the proximal colon and made an opening for the anal cut-end of the distal colon onto the abdominal skin.

Thereafter, the distal colon was cut 1 cm from the rectovesical pouch. The aboral cut end of the distal colon and the oral cut end of the terminal ileum were anastomosed together with six to eight interrupted sutures to re-establish continuity of the bowel. The oral cut end of the distal colon was pulled through the abdominal wall onto a skin opening at the left of the lower abdomen and fixed in an open condition with four sutures between the colonic seromuscular layer and abdominal wall (Fig. 1).

A silicon tube (Silascon Medical Grade Tube, Kaneka Medics, Osaka, Japan; i.d., 0.5 mm; o.d., 1.0 mm) was inserted into the proximal colon through a small opening near the closed cut end and fixed with a purse string suture. The tube was pulled out through the laparotomized wound, fixed onto the rectus abdominis muscle with a suture and tunneled out subcutaneously to the skin opening in the midscapular region. The external end of the silicon tube was ligated and fixed subcutaneously for later use. Then, the skin opening in the midscapular region was closed. We used 6-0 nylon monofilament thread mounted on half-curved atraumatic needles (Iwashiya, Sendai, Japan) for all gut sutures.

After surgery, rats were housed individually in stainless steel cages and allowed free access to nonpurified diet and water.

Colonic infusion of experimental solutions.  We anesthetized rats with diethylether and reopened the skin in the midscapular region between 0800 and 1000 h on the 10th postoperative day (POD) . We excavated the end of the silicon tube previously embedded under the skin and connected the tube to another silicon tube using a 23-gauge needle. The tube was flushed with 1.5 mL of 150 mmol/L NaCl solution to confirm flow through the stoma. We attached a harness mounted with a spring coil to the rats. The externalized tube was pulled through the coil and further connected to a swivel. We transferred rats to wire mesh-bottomed metabolic cages. The swivel was connected to a continuous infusion system including a peristaltic pump (Watson Marrow, Cornwall, England).

The operated rats were randomly placed into four groups of four or five animals to receive one of the four colonic infusates using a 2 × 2 two-way factorial design (ammonia × SCFA) (Table 1). Solutions were infused into the isolated colon at 1 mL/h for 7 d.

Sampling of gut tissues.  Rats were anesthetized with diethylether between 0600 and 0800 h on POD 17 (d 7 of colonic infusion). The right external jugular vein was exposed to administer 1 mg/kg body weight of vincrystine sulfate (0.5 mg/mL in 150 mmol/L NaCl) for metaphase arrest. The skin wound was closed with 4-0 silk sutures. Vincrystine sulfate was administered between 0600 and 0800 h.

Rats were re-anesthetized with diethylether, laparotomized and killed with an overdose of sodium pentobarbital (more than 100 mg/kg body weight) into the inferior vena cava 3 h after the administration of vincrystine sulfate. The proximal and distal colons were immediately removed and fixed in a mixture of acetic acid and ethyl alcohol (1:3, v/v) for 3 h. The tissue samples were transferred to 70% ethanol and stored until crypt dissection.

Estimation of epithelial cell proliferation.  Fixed tissue pieces (~3 × 3 mm) of the distal colon located 1 cm distally from the middle colic artery were stained with Feulgen reaction en bloc. Crypts were dissected under a stereomicroscope and squashed onto a glass slide (Wimber and Lamerton 1973).

Metaphase figures per crypt were counted in twenty crypts per specimen. The mean metaphase frequency in twenty crypts was arbitrarily termed the "accumulated mitoses per crypt" in this study. This parameter roughly reflected the rate of epithelial cell production in a crypt.

Histological analysis.  The segment of the distal colon just orad to the portion used for crypt dissection was embedded in paraffin. A 2.5-µm thick cross section was prepared from each specimen and stained with hematoxylin and eosin.

We counted the number of epithelial cells on the right side of axial crypt sections (number of cells per crypt column) and measured the linear distance from the base to the mouth of the crypt in 20 randomly chosen complete axial crypt sections per rat. We also calculated the "accumulated mitotic index" as the number of metaphase figures divided by the total number of epithelial cells in 20 randomly chosen axial crypt sections.

Statistical analysis.  The results were expressed as means ± SEM. The differences among means of experimental groups were tested by two-way ANOVA by using the Stat View 4.0 program (Abacus Concept, Berkeley, CA). Logarithmic transformation of the data (Sokal and Rohlf 1995) was performed to stabilize variance when accumulated mitoses per crypt was analyzed, because the variance of this parameter varied proportionally to the mean value. When an interaction effect was significant, Scheffé's post hoc comparison test (Sokal and Rohlf 1995) was performed. Differences between means were considered significant at P < 0.05.

	RESULTS

Abstract Introduction Methods Results Discussion References

Body weight.  Four rats died during the recovery period. Three rats were removed from the study due to stoma obstruction. The remaining 17 rats were used for colonic infusion. Some suffered from diarrhea for several days after the operation. However, their stools became solid by the end of the recovery period. Accordingly, these rats were considered sufficiently recovered from surgical stress to include in the study.

The body weight on the day of operation, POD 10 (after the recovery period) and POD 17 (after the infusion period) did not differ among the four infusate groups (270.1 ± 3.0, 288.3 ± 4.3 and 308.0 ± 4.6 g, respectively) with reasonable weight gain during the recovery period (18.2 ± 1.8 g). The rats used for the ammonia infusion gained more weight than those not given ammonia during the recovery period (22.4 ± 2.3 g vs. 14.7 ± 2.2 g, P < 0.05). However, the weight gain during the infusion period and total weight gain during the 17 d of experimental period did not differ among groups (20.6 ± 2.2 and 38.8 ± 2.8 g, respectively).

Epithelial cell proliferation of the distal colon.  There was a significant interaction between ammonia and SCFA on the accumulated mitoses per crypt of the distal colon (Table 2). Infusion of ammonia or SCFA, when given individually, but not simultaneously, resulted in a greater number of accumulated mitoses per crypt than the infusion with saline alone. The value when SCFA alone were infused tended to be greater than when both SCFA and ammonia were infused. The accumulated mitotic index did not differ among rats infused with ammonia, SCFA or both, and the reaction between ammonia and SCFA was significant. Values in these three groups were significantly greater than in rats given saline alone (Table 2).

Histological analysis.  Both the number of epithelial cells per crypt column (38.1 ± 1.1 vs. 31.6 ± 1.0, P < 0.001) and crypt depth (273 ± 9 vs. 226 ± 9 µm, P < 0.01) were greater in rats given SCFA than in those not given SCFA (Table 2). Ammonia had no effect on the number of epithelial cells per crypt column or on crypt depth. There was no interaction between these variables.

	DISCUSSION

Abstract Introduction Methods Results Discussion References

These results clearly demonstrate that both ammonia and SCFA stimulate colonic epithelial cell proliferation. However, the proliferative activities of all infusate groups were lower than that previously measured in intact rats fed fiber-containing food (Sakata 1988), very likely due to the reduced production of SCFA and ammonia because of colonic isolation. Such a decrease in crypt cell production rate was observed in rats after hindgut by-pass surgery (Sakata 1988). Accordingly, the increase in mitotic activity due to SCFA or ammonia in this study might be considered not as a stimulation exceeding the normal level, but as a kind of recovery by the compensatory administration of SCFA or ammonia.

Infusion of test solutions into the isolated colon in this study had the methodological advantage of minimizing possible influences of previously existing luminal constituents and the influx of endogenous nitrogen from the upper digestive organs. We adjusted the pH of infusates to 7.2, the upper limit of the physiological pH range in the colonic lumen, because luminal alkalization may increase the effect of ammonia (Lin and Visek 1991a) as a result of an increase in non-ionic diffusion of ammonia across the apical membrane of colonocytes (Cohen et al. 1988).

Previous results concerning the effect of ammonia on cell proliferation are not consistent. Experimental rodents intoxicated with ammonia have shortened cell life spans and altered DNA synthesis in various tissues (Visek 1972 and 1978). On the other hand, reduction of ammonia release into the gut in mice immunized against urease resulted in reduced DNA synthesis in the colonic epithelium (Visek 1972). Colonic epithelial cell proliferation increased in rats fed a high protein, high fat diet and accompanied high luminal concentrations of ammonia (Lin and Visek 1991a). A high protein diet or intrarectal infusion of ammonium acetate increased colonic tumorigenesis in chemical carcinogen-induced rats (Clinton et al. 1988, Topping and Visek 1976). These studies indirectly suggest that luminal ammonia might affect colonic epithelial cell proliferation.

This study demonstrated that 75 mmol/L (an upper limit of the physiological concentration range) of sodium ammonium solution infused into surgically isolated rat colon stimulated colonic epithelial proliferation without changing crypt size (Table 2). Thus, it is likely that ammonia accelerates rates of both cell production and cell loss. In other words, ammonia is not sufficient to sustain normal crypt size of the colon.

Colonic infusion of 145 mmol/L SCFA for 7 d increased the crypt cell production rate in this study (Table 2). This was consistent with previous studies in which bolus instillation (Sakata 1987) or continuous infusion (Kripke et al. 1989) of SCFA into the cecum or colon stimulated colonic epithelial cell proliferation in rats with depressed gut epithelial proliferation. In contrast to ammonia, the trophic effect of SCFA was accompanied by greater crypt size. This suggests that SCFA are necessary to maintain normal size of the colonic crypts and possibly normal functions of the colon (Roediger 1980, von Engelhardt 1995).

Ammonia and SCFA individually stimulated colonic epithelial cell proliferation in this study. However, their different effects on crypt size suggest that the underlying mechanisms of the trophic effects might be different for ammonia and SCFA. It is also noteworthy that enhanced epithelial cell proliferation did not always accompany crypt enlargement. This suggests that mass or content of protein or DNA of the tissue is a poor indication of epithelial cell proliferation and vice versa.

It is very interesting that the effects of ammonia and SCFA were not additive and that the presence of ammonia tended to reduce the effect of SCFA on epithelial cell proliferation (Table 2). This effect of ammonia was not due to the titration effect, because we adjusted the pH of all infusates to 7.2, although we did not measure the pH of the colonic lumen or that of the effluent.

The coexistence of ammonia and SCFA increases absorption of each substance, possibly by increasing the proportion of the nondissociated form of these solutes (Bödecker 1996). Darcy-Vrillion et al. (1996) also reported that when ammonium and n-butyrate co-exist in the medium, the ammonium reduces n-butyrate utilization by colonocytes in vitro. Such a modulation of absorption or intracellular metabolism might influence their trophic effects by providing more energy to colonocytes.

Bartram et al. (1993) did not observe any differences between ammonium and sodium ions when they incubated human colonic mucosa with 10 mmol/L n-butyrate in vitro. The differences in experimental settings (in vivo vs. in vitro, chronic vs. short term, different concentrations of SCFA and ammonia) could explain the discrepancy between the present results and those of Bartrum et al.

The trophic effects of dietary fibers seem to depend at least in part on their fermentability and resultant SCFA production. However, in some studies there was not a positive correlation between fermentability of dietary fibers or luminal concentration of SCFA and colonic epithelial proliferation (Boffa et al. 1992, Butler et al. 1992, Fleming et al. 1992, Whiteley et al. 1996a). Further, Whiteley et al. (1996b) reported that the concentration of SCFA, bile acids or ammonia had no consistent correlation with the colonic mucosal volume. Apart from the fact that lumen concentration is a poor indicator of SCFA production, these discrepancies might be due to interactions among luminal constituents.

Lupton and Marchant (1989) reported that protein levels and fiber types in diets interactively affect ammonia concentrations in the large intestine. Dietary fibers alter luminal concentration of ammonia and bile acids in addition to that of SCFA (Whiteley 1996b). Thus, the above discrepancies among studies on the proliferative effects of dietary fibers and luminal constituents could be explained by the interaction among luminal constituents, such as that between ammonia and SCFA shown in this study. All of these results suggest that when we consider effects of fermentable fibers we should keep in mind that the effects of SCFA can be modified by nitrogen influx to the large intestine and by microbial ammonia production.

In conclusion, ammonia and SCFA individually stimulated the epithelial cell proliferation of isolated rat colon in vivo. SCFA, but not ammonia, increased crypt size, and the effects of ammonia and SCFA on colonic epithelial cell proliferation were not additive. The significant interactions between ammonia and SCFA on the accumulated mitoses per crypt and mitotic index should be considered when the effects of dietary fibers on gut epithelial cell proliferation are investigated.

	ACKNOWLEDGMENTS

We thank Yasuko Furukawa for the excellent preparation of histological sections. We also thank Akiko Inagaki for the technical assistance with crypt microdissection.

	FOOTNOTES

Manuscript received 2 September 1997. Initial reviews completed 1 October 1997. Revision accepted 16 January 1998.

	LITERATURE CITED

Abstract Introduction Methods Results Discussion References

• Bartram P., Scheppach W., Schmid H., Dusel G., Richter F., Richter A., Kasper H. Proliferation of human colonic mucosa as an intermidiate biomarker of carcinogenesis: effects of butyrate, deoxycholate, calcium, ammonia and pH. Cancer Res. 1993; 53:3283-3288 ^{[Abstract/Free Full Text]}
• Boffa L. C., Lupton J. R., Mariani M. R., Ceppi M., Newmark H. L., Scalmati A., Lipkin M. Modulation of colonic epithelial cell proliferation, histon acetylation, and luminal short chain fatty acids by variation of dietary fiber (wheat bran) in rats. Cancer Res. 1992; 52:5906-5912 [Abstract/Free Full Text]
• Bödeker, D. (1996) Einflu von kurzkettigen Fettsäuren auf den Ammoniaktransport durch die Colonschleimhaut von Pferden. Europäische Konferenz über die Ernährung des Pferdes 147-149.
• Butler R. N., Bruhn B., Pascoe V., Fettman M. J., Roberts-Thomson I. C. Regional factors affecting proliferation in the large intestine of the rat. Proc. Soc. Exp. Biol. Med. 1992; 200:133-137 [Medline]
• Clinton S. K., Bostwick D. G., Olson L. M., Mangian H. J., Visek W. J. Effects of ammonium acetate and sodium cholate on N-methyl-N-nitro-N-nitrosoguanidine-induced colon carcinogenesis of rats. Cancer Res. 1988; 48:3035-3039 [Abstract/Free Full Text]
• Cohen R. M., Stephenson R. L., Feldman G. M. Bicarbonate secretion modulates ammonia absorption in rat distal colon in vivo. Am. J. Physiol. 1988; 254:F657-F667 [Abstract/Free Full Text]
• Darcy-Vrillon B., Cherbuy C., Morel M.-T., Durand M., Duée P.-H. Short chain fatty acid and glucose metabolism in isolated pig colonocytes: modulation by NH⁺₄. Mol. Cell. Biochem. 1996; 156:145-151 [Medline]
• Fleming S. E., Fitch M. D., de Vries S. The influence of dietary fiber on proliferation of intestinal mucosal cells in miniature swine may not be mediated primarily by fermentation. J. Nutr. 1992; 122:906-916
• Kripke S. A., Fox A. D., Berman J. M., Settle R. G., Rombeau J. L. Stimulation of intestinal mucosal growth with intracolonic infusion of short-chain fatty acids. J. Parent. Enteral Nutr. 1989; 13:109-116 [Abstract/Free Full Text]
• Lin H. C., Visek W. J. Large intestinal pH and ammonia in rats: dietary fat and protein interactions. J. Nutr. 1991a; 121:832-843
• Lin H. C., Visek W. J. Colon mucosal cell damage by ammonia in rats. J. Nutr. 1991b; 121:887-893
• Lupton J. R., Marchant L. J. Independent effects of fiber and protein on colonic luminal ammonia concentration. J. Nutr. 1989; 119:235-241
• Macfarlane G. T., Cummings J. H., Allison C. Protein degradation by human intestinal bacteria. J. General Microbiol. 1986; 132:1647-1656
• Roediger W.E.W. Role of anaerobic bacteria in the metabolic welfare of the colonic mucosa in man. Gut 1980; 21:793-798 ^{[Abstract/Free Full Text]}
• Sakata T. Stimulatory effect of short-chain fatty acids on epithelial cell proliferation in the rat intestine: a possible explanation for trophic effects of fermentable fibre, gut microbes and luminal trophic factors. Br. J. Nutr. 1987; 58:95-103 [Medline]
• Sakata, T. (1988) Depressed intestinal epithelial cell production rate by hindgut bypass in rats. Scand. J. Gastroenterol. 23, 1200-1202.
• Scheppach, W., Bartram, P., Richter, A., Liepord, H., Dusel, G., Hofstetter, G., Rüthlein, J. & Kasper, H. (1992) Effect of short-chain fatty acids on the human colonic mucosa in vitro. J. Parent. Enteral Nutr.16: 43-48.
• Sokal, R. R. & Rohlf, F. J. (1995) Biometry, 3rd ed. W. H. Freeman and Company, New York, NY.
• Topping D. C., Visek W. J. Nitrogen intake and tumorigenesis in rats injected with 1,2-dimethylhydrazine. J. Nutr. 1976; 106:1583-1590
• Visek W. J. Effects of urea hydrolysis on cell life-span and metabolism. Fed. Proc. 1972; 31:1178-1193 [Medline]
• Visek W. J. Diet and cell growth modulation by ammonia. Am. J. Clin. Nutr. 1978; 31:S216-S220 [Abstract]
• von Engelhardt, W. (1995) Absorption of short-chain fatty acids from the large intestine. In: Physiological and Clinical Aspects of Short-Chain Fatty Acids (Cummings, J. H., Rombeau, J. L. & Sakata, T., eds.), pp. 149-170. Cambridge University Press, Cambridge, UK.
• Whiteley L. O., Higgins J. M., Purdon M. P., Ridder G. M., Bertram T. A. Evaluation in rats of the dose-response relationship among colonic mucosal growth, colonic fermentation, and dietary fiber. Dig. Dis. Sci. 1996a; 41:1458-1467 [Medline]
• Whiteley L. O., Purdon M. P., Ridder G. M., Bertram T. A. The interactions of diet and colonic microflora in regulating colonic mucosal growth. Toxicol. Pathol. 1996b; 24:305-314 [Abstract/Free Full Text]
• Wimber D. R., Lamerton L. F. Cell population studies on the intestine of continuously irradiated rats. Radiat. Res. 1973; 18:137-146
• Wrong O. M., Vince A. J., Waterlow J. C. The contribution of endogenous urea to faecal ammonia in man, determined by ¹⁵N labelling of plasma urea. Clin. Sci. 1985; 68:193-199 [Medline]

0022-3166/98 $3.00 ©1998 American Society for Nutritional Sciences