Volatile Fatty Acids and Aerobic Flora in the Gastrointestinal

Transcription

1 INFECTION AND IMMUNITY, Mar. 1979, p /79/ /05$02.00/0 Vol. 23, No. 3 Volatile Fatty Acids and Aerobic Flora in the Gastrointestinal Tract of Mice Under Various Conditions B. M. BYRNE AND J. DANKERT* Laboratory for Medical Microbiology and Department of Hospital Epidemiology, University Hospital, Groningen, The Netherlands Received for publication December 1978 Volatile fatty acids are reported to exert a repressive effect upon Enterobacteriaceae and Pseudomonas species in vitro and in vivo in young mice. The mean total volatile fatty acid concentration in the cecal samples of conventional mice fed ad libitum was 81.7 pmol/g (wet weight), which is antibacterial in vitro, and in the rectal samples it was 41.1 umol/g (wet weight). The mean count of Enterobacteriaceae in the cecum was only 102/g, whereas in the rectum it was 10'/g. Volatile fatty acid levels were influenced by food intake and increased to peak levels approximately 6 to 10 h after eating and then declined. In mice fasted for 17 h, the butyric acid concentration was considerably lower and the number of cecal samples positive for Enterobacteriaceae increased. When fasted for 4 days, mice had extremely low cecal and rectal volatile fatty acid concentrations and the Enterobacteriaceae and enterococci counts increased to mean of 2 x 106/ g and 3 x 106/g, respectively, in the cecum and to means of 107 and 5 X 106/g in the rectum. We conclude that volatile fatty acids are probably one of the many interference mechanisms which are involved with control of the levels of Enterobacteriaceae (and enterococci) in the large intestine of mice. The gastrointestinal flora is regulated by both direct and indirect interference mechanisms which have been reviewed recently by Savage (14). One of the former mechanisms includes the production of toxic microbial metabolites such as volatile fatty acids (VFA) of which the undissociated state (7, 8) is antagonistic for Enterobacteriaceae species and Pseudomonas aeruginosa. The inhibitory activity of these acids has been thoroughly investigated in vitro (1, 2, 3, 7, 8, 9). The VFA are metabolic end products of the anaerobes in particular (10) and are produced mainly in the cecum. In vivo experiments in mice show that the decrease in VFA levels after antibiotic treatment is associated with a lowered resistance to infection by Salmonella species (2, 3, 8) and by Staphylococcus aureus (9) and with a strong increase in the numbers of coliform bacilli (6). The latter authors have also observed that in young mice the increase in VFA results in a decrease in coliform numbers. However, Freter and Abrams (5) find that VFA levels in the gastrointestinal tract of mice do not always correlate with the Escherichia coli population levels. VFA levels are constantly changing due to variations in the bacterial metabolic activity on the one hand and due to absorption and peristalsis on the other hand. With this in mind, we have tried to investigate the relationship between VFA concentration and the intestinal aerobic flora in mice fed ad libitum and under restricted feeding conditions. MATERIALS AND METHODS Mice. Female Swiss random mice of strain Cpb.SE (TNO, Zeist, The Netherlands), 8 to 12 weeks old, were maintained individually in Macrolan cages containing autoclaved wood shavings. The mice were fed autoclaved food pellets (Hope Farms, Woerden, The Netherlands) and received sterilized water ad libiturm, except where otherwise stated. For collection of fresh fecal samples, mice were placed in cages containing wired grids. Samples were collected immediately after they were passed to prevent VFA loss. VFA and aerobic bacterial levels in the cecal and fecal samples of conventional mice. A total of 40 mice were screened daily for 7 days for fecal VFA to determine VFA levels and the day-to-day variation. Bacteriological culturing was also performed. A total of 21 mice were sacrificed to determine VFA concentrations and levels of aerobic bacteria in the large intestine (group A). Contents of cecum and rectum were removed and divided for VFA analysis and bacteriological culturing. Experimental procedures. To assess the influence of food intake upon VFA levels in the feces, four mice were fed at fixed times, i.e., at 8.00 to 9.00, to 13.00, and to h for 5 days. On the fifth day, every fecal stool passed was collected individually from to h and analyzed for VFA. To study the influence of food restriction upon VFA and bac- 559

2 560 BYRNE AND DANKERT terial levels, 17 mice were sacrificed 15 to 17 h after feeding (group B). Contents of cecum and rectum were removed for VFA analysis and culturing. Lastly, a fasting experiment was performed involving 10 mice (group C). After 4 days of fasting (with sterilized water ad libitum), the contents of the cecum and rectum were collected aseptically and divided for VFA analysis and bacteriological culturing. In some experiments there was insufficient material for both determinations. VFA extraction and analysis. Immediately after collection, samples were stored at -20 C awaiting analysis. The pretreatment and analysis of specimens is similar to that previously described (18) involving vacuum distillation and gas chromatography. To ensure that this method could be applied to minute samples, the detection limit was first determined. Samples were suspended in 1.5 ml of distilled water containing 500 pl of 2-methylvaleric acid as internal standard and 11 ml of H2SO4 per liter and then directly distilled. Quantitative aerobic culturing technique. After collection, samples were stored at 4 C to prevent microbial growth. Culturing was performed, in most cases, within 30 min, but in any case within 20 h. No differences were found when culturing was done immediately or after 20 h. Specimens were weighed, homogenized in nutrient broth (Oxoid Ltd., London) supplemented with 1% (wt/vol) glucose and serially diluted up to 10'9 in the same medium. Tenfold dilutions were prepared in trays (Thovadec, Nieuwkoop, The Netherlands), using calibrated pipettes, and were incubated at 37 C for 18 to 24 h. Subinoculations were prepared on Sheep blood agar, MacConkey agar (Merck, Darmstad, Federal Republic of Germany) selective for gram-negative organisms, Mannitol Salt agar (Oxoid Ltd., London) selective for staphylococci, Aesculine Azide Bile agar (Merck) selective for enterococci, acetamide (111%, wt/vol) agar (Merck) selective for Pseudomonas species, and Sabouraud dextrose agar (ph 5.2, Merck) selective for yeasts. All inoculated media were incubated at 37 C for 24 h, except Sabouraud dextrose agar which was incubated for 72 h at 23 C. The highest dilution from which growth was obtained, in each of the specimens, represented the concentration of microorganisms. Ten to 20 colonies from MacConkey medium, presumed to be gram-negative bacteria, were identified by using the API-system (API 20 Enteric, Analytical Products, Inc., New York, N.Y.). Enterococci cultured on aesculine azide bile agar appear as black colonies. When less than 100 organisms were present, no growth was obtained and this was consequently denoted as <10i in the results. RESULTS VFA and aerobic bacterial levels in fecal samples from conventional mice. By using our vacuum distillation method to determine the VFA concentrations, without concentrating the distillate, it was found that VFA could accurately be analyzed from at least 0.04 g of cecal and rectal samples. Major VFA were acetic acid (C2), propionic acid (03), and n-butyric acid (04). INFECT. IMMUN. Isobutyric, valeric, and isovaleric acids accounted for less than 2% of the total amount of VFA in the samples. Fecal samples were taken daily for 1 week from 40 mice to assess the day-to-day variation in VFA levels and the total number of aerobic bacteria, gram-negative bacteria, and enterococci. In Table 1, the mean levels of the major VFA per day are given. The average daily values were remarkably constant, with the exception of C3. However, standard deviations (SD) were high in all cases. The mean total number of bacteria was 108 (range 106 to 109) per g (wet weight), whereas the mean level of gram negatives was 2 x 105 (range 102 to 106). The mean enterococcus level was 5 x 107 (range 105 to 109). There was no apparent relationship between VFA levels and bacterial counts, when examined for each mouse individually. Influence of food intake upon fecal VFA. VFA were determined in every fecal stool passed, beginning at h (3 h after the first feed) until h. On average, 26 stools were analyzed from each mouse (four mice, total). The VFA (02, C3 and C4) pattern shows a rise in concentration approximately 6 to 8 h after the first and second feeding respectively (Fig. 1). From to h and from to h, fecal VFA levels were rather stable. Mean levels of C2 were (SD 9.85) and of C3 and C4 were 1.40 (SD 0.89) and 1.17 (SD 0.75), respectively. At h and at h, C2 levels in the fecal pellets were highest, but peak levels remained during the following 2 h when C3 and C4 levels also increased. Mean peak levels of C2 were (SD 9.61) and (SD 11.88), those of C3 were 2.27 (SD 0.81) and 1.89 (SD 1.69), and those of C4 were 4.46 (SD 2.81) and 8.00 (SD 6.82), respectively, for the first and second peak. Unfortunately, culturing could not be performed due to the minute amount of sample available. Mean VFA and aerobic bacterial levels in cecal and rectal samples from mice fed ad libitum. VFA levels were determined in 21 con- TABLE 1. Mean daily concentration of the three major VFA in rectal samples from 40 mice Mean concn (Iumol/g) of:a Day Propionic acid Butyric acid Acetic acid (C2) (03) (04) (11.60) 1.46 (0.68) 4.88 (5.44) (14.20) 1.24 (0.65) 3.70 (3.66) (10.14) 1.37 (0.46) 4.20 (3.19) (7.93) 3.38 (3.35) 4.16 (3.33) (10.49) 4.34 (2.83) 4.19 (3.40) (15.40) 2.55 (0.92) 4.03 (2.93) (12.98) 2.05 (0.87) 3.68 (3.00) a Values in parentheses indicate SD.

3 VOL. 23, 1979 ventional mice (group A). The mean concentrations of C2, C3, and C4 (umol/g, wet weight) were (SD 22.55), 4.36 (SD 2.28) and (SD 8.92) in the cecum and (SD 14.38), 2.36 (SD 1.15), and 5.06 (SD 3.10), in the rectum (Table 2). VFA levels in the cecum were two to three times higher than in the rectum in 70% of the mice. The ratio of the three major VFA was 13:1:6 in the cecum and 16:1:2 in the rectum. The percent decrease in the C2, C3, and C4 levels between the cecum and the rectum was 37, 46, and 77, respectively. Mean total counts of aerobic bacteria in the cecum and the rectum were 1.5 x 106 (range 103 to 106) and 2.0 x 107 (range 10W to 108) per g (wet weight), respectively E 'a 10 VFA AND AEROBES AFTER FEEDING AND FASTING aoo L.00 < 30 > 20 Time (hi The VFA pattern in the fecal contents after FiG. feeding. Each fecal pellet from four mice was analyzed for VFA. Each point represents a mean VFA value from fecal samples collected 15 min before and after the denoted time. Feeding times are indicated by striped blocks. Symbols: 0, acetic acid (C2); *, propionic acid (C3); A, butyric acid (C4). tterobacteriaceae species (mainly E. coli),ed a mean level of 1.0 x 10' (range <10' to n the cecum and 1.0 x 10' (range 10' to 106) e rectum. Similarly, enterococcus levels in- ;ed from a mean of 10' (range 10' to 104) in ecum to a mean of 10' (range 103 to 106) in *ectum. Yeasts, staphylococci, and Pseudoas species were not cultured from the cecum he rectum. More than half of the mice red a complete absence of gram-negative eria in the cecum, whereas in the rectum of e mice Enterobacteriaceae species levels ed from 102 to 106 (Fig. 2a). Results suggest ationship between VFA levels and the gram- Ltive count. fect of fasting upon VFA and bacterial Js. Seventeen mice were sacrificed 15 to 17 ter the last feeding time (group B). Mean L levels and bacterial counts were within the nal range (Table 2). VFA concentrations ooo tolgo Enterobacteriocsat ou t/grom G. 2. Individual VFA concentrations and Enterteriaceae levels from cecal (0) and rectal (A) les of mice fed ad libitum (a), fasted for 17 h (b), fasted for 4 days (c). TABLE 2. Mean major VFA concentrations and aerobic bacterial levels in cecal and rectal samples from mice fed ad libitum (group A), fasted for 15 to 17 h (group B), and fasted for 4 days (group C) Concn (jmol/g, wet wt) of:a Bacterial level (counts/g)b Group Specimen Propiomc acid Butyric acid Total Enterobac- Enterococci Acetic acid (C2 Eneoci (03) (C4) aerobes teriaceae A (n = 21) Cecum (22.55) 4.36 (2.28) (8.92) 1.5 X X x 10' (103_106) (1w-1M~) (102_104) Rectum (14.38) 2.36 (1.15) 5.06 (3.10) 2.0 x X x 105 (104_101s) (102_1o63) (103_1o6) B (n = 17) Cecum (25.99) 4.05 (2.80) (5.59) 1 X x 103 ND (104_107) (<102_104) Rectum (10.64) 1.91 (1.46) 1.56 (0.85) 7 x x 105 ND ( 105_107) ( 103_107) C (n = 10) Cecum (5.32) 3.14 (1.18) 3.07 (2.00) 2 x x x 106 (104_1C8) ( 104_107) ( 103_107) Rectum (4.88) 1.86 (1.00) 0.87 (0)C 2 x 10' 1 X X 106 ( ) (104_1081) (104_107) a Values in parentheses indicate SD. b Values in parentheses indicate range. ND, Not done. 0 C4 was present in only one mouse. 561

4 562 BYRNE AND DANKERT in the rectum were 58, 55, and 82% lower for C2, C3, and C4, respectively, than in the cecum, whereas the Enterobacteriaceae levels were 100- to 1,000-fold higher. In Fig. 2b, the relationship between VFA concentrations and gramnegative bacterial counts in the cecum and the rectum for each mouse individually is presented. High VFA concentrations and low Enterobacteriaceae levels are characteristic of the cecum and the reverse is true for the rectum. In all ten mice sacrificed on the fifth day, after fasting for 4 days (group C), at least 104 Enterobacteriaceae per g of cecal content were cultured, whereas VFA values only reached maximum concentrations of 40 pmol/g (wet weight) (Fig. 2c). The mean VFA concentration in the cecum was approximately 2.5 times lower than in mice fed ad libitum (Table 2), whereas the mean Enterobacteriaceae level was higher by a factor of log 4. Likewise, the enterococcus level increased by a factor of log 3. In the rectum, VFA levels were also about 2.5 times lower, whereas the Enterobacteriaceae count reached 108 bacteria per gram (wet weight). The concentrations of C2, C3, and C4 in the rectum were, respectively, 33, 59, and 88% lower than in the cecum. Similar reduction percentages were found in mice in the previous experiments. DISCUSSION Our findings suggest a possible relationship between VFA levels and Enterobacteriaceae and enterococcus counts in the large intestine of mice under various feeding conditions. In general, a low VFA level was associated with high numbers of Enterobacteriaceae and enterococci and high VFA levels with low numbers of these microorganisms in rectal and cecal samples. The total concentration of the major VFA (acetic, propionic, and butyric acids) in the cecum showed a wide variation under normal feeding conditions (23 to 116,umol/g, wet weight) which demonstrates that the VFA level is subject to rather strong fluctuations. The majority of the mice had cecal levels of the major VFA which were higher than concentrations found to be toxic for Enterobacteriaceae species and P. aeruginosa in vitro (2, 7, 8). However, some cecal samples (19%) showed VFA concentrations below the toxic level, whereas gram negatives were not cultured. Similarly, 28% of the mice had high VFA levels and gram negatives were present at up to a 10'/g cecal content. In the rectal samples, VFA were below in vitro toxic concentrations, high numbers of Enterobacteriaceae species and enterococci were cultured, but by examining the individual results (Fig. 2), no correlation could be INFECT. IMMUN. made between the level of VFA and the Enterobacteriaceae count. For a given in vivo VFA concentration, various numbers of bacteria were found. In samples collected for 1 week, the same was true and wide fluctuations in the VFA levels occurred. These findings are in agreement with those of Freter and Abrams (5). VFA are end products of the fermentation of soluble carbohydrates and other nutrients by anaerobes in particular. Anaerobic microorganisms are present in high numbers in the cecum especially. Feeding influences VFA concentrations in the rumen of cattle (4) and in the cecum of rabbits (12) and rats (17). Since the mice in our first experiment received food ad libitum, a wide range in VFA levels can be expected, influenced by food intake, the moment of sampling, and the composition and metabolic activity of the anaerobic flora. In the study concerning the effect of food intake upon the VFA levels in the feces, a significant increase in the major VFA occurred 6 to 8 h and 8 to 10 h after feeding times. Since these are fecal values, one can imagine that some hours previously the level of the VFA in the cecum was considerably higher and therefore toxic. However, VFA require a certain contact time (8) to exert an antibacterial effect at the ph present in the cecum (7). When VFA levels are measured 17 h after feeding, total amounts are slightly lower, but this is largely due to the decrease in the butyric acid levels (Table 2). The mean count of Enterobacteriaceae species 17 h after feeding was not reduced in comparison with mice fed ad libitum. However, 21% of the cecal samples showed an absence of gram-negative bacteria (Fig. 2b), whereas in ad libitum-fed mice this figure was 67% (Fig. 2a). In mice fasted for 2 days (results not included) or for 4 days, extremely low cecal VFA levels were found. On the contrary, Enterobacteriaceae levels strongly increased by a factor of log 3 to 4. The number of enterococci increased by a factor of log 3. The VFA levels and the mean Enterobacteriaceae count in the cecum of fasted mice were comparable to those found in the rectum of mice fed ad libitum. The percent decrease in VFA levels between cecal and rectal samples was similar in fasted and ad libitum-fed mice. In both cases, a noticeable drop in butyric acid levels was observed, probably due to rapid absorption and metabolic turnover (11, 13). In rats, food deprivation also results in a fall of more than 60% in the cecal VFA levels after 30 to 48 h, whereas the ph increases from 6.2 to 7.6 (13). Bacteriological culturing was not performed in conjunction with the study. However, Tannock and Savage (15) report the influence of

5 VOL. 23, 1979 dietary manipulations upon the gastrointestinal microflora of mice. They find a strong increase in E. coli and enterococci in the cecum after 48 h of fasting. Anaerobic counts remain the same, but the VFA level, as an indicator of the metabolic activity of this section of the flora, was not measured. Our findings from fasted mice demonstrate a reduction in VFA levels, indicating a decrease in the metabolic activity of the anaerobic flora. Van der Waaij et al. (16) conclude that the anaerobic flora are an important constituent of the so-called colonization resistance in the gut. The increase in the gram negatives and enterococci may indeed be due to the decreased metabolic activity of the anaerobes. However, impairment of other host defense mechanisms may also occur during fasting (15). We have attempted to study in vivo the effect of VFA upon the colonization of the aerobic bacteria in the gut, using orally antibiotictreated (freshly decontaminated) conventional mice by administering VFA in solution intragastrically or via a cecal cannula and subsequently challenging with aerobic gram-negative and gram-positive bacteria. However, due to the extremely rapid VFA absorption and technical difficulties, this model was unsuccessful. ACKNOWLEDGMENTS We are grateful to the members of the Central Animal Laboratory, State University of Groningen, The Netherlands for their kind cooperation and to Yolanda Schoenmaker and Anne Schaafsma for their contributions. VFA AND AEROBES AFTER FEEDING AND FASTING 563 LITERATURE CITED 1. Bergeim, Toxicity of intestinal volatile fatty acids for yeasts and Esch. con. J. Infect. Dis. 66: Bohnhoff, M., C. P. Miller, and W. R. Martin Resistance of the mouse's intestinal tract to experimental Salmonella infection. 1. Factors which interfere with the initiation of infection by oral inoculation. J. Exp. Med. 120: Bohnhoff, M., C. P. Miller, and W. R. Martin Resistance of the mouse's intestinal tract to experimental Salmonella infection. 2. Factors responsible for its loss following streptomycin treatment. J. Exp. Med. 120: Brownie, L E., and F. H. Grau Effect of food intake on growth and survival of salmonellas and Escherichia coli in the bovine rumen. J. Gen. Microbiol. 46: Freter, R., and G. D. Abrams Function of various intestinal bacteria in converting gernfree mice to the normal state. Infect. Immune. 6: Lee, A., and E. Ge meli Changes in the mouse intestinal microflora during weaning: role of volatile fatty acids. Infect. Immun. 5: Levison, M. E Effect of colon flora and short-chain fatty acids on growth in vitro of Pseudomonas aeruginosa and Enterobacteriaceae. Infect. Immun. 8: Meyneli, G. G Antibacterial mechanisms of the mouse gut. 2. The role of Eh and volatile fatty acids in the normal gut. Br. J. Exp. Pathol. 44: Meynel, G. G Antibacterial mechanism of the mouse gut. 3. The fate of Staphylococcus aureus in normal and streptomycin-treated mice. Br. J. Exp. Pathol. 4: Moore, W. E. C., E. P. Cato, and L V. Holdeman Anaerobic bacteria of the gastrointestinal flora and their occurrence in clinical infections. J. Infect. Dis. 119: Mottaz, P., and J. F. Worbe Transfert des acides gras volatils dans la paroi du caecum isoli de rat. C.R. Soc. Biol. (Paris) 171: Parker, D. S., and R. T. McMillan The determination of volatile fatty acids in the cecum of the conscious rabbit. Br. J. Nutr. 35: Remesy, C., and C. Demigne Partition and absorption of volatile fatty acids in the alimentary canal of the rat. Ann. Rech. Veter. 7: Savage, D. C Microbial ecology of the gastrointestinal tract. Annu. Rev. Microbiol. 31: Tannock, G. W., and D. C. Savage Influences of dietary and environmental stress on microbial populations in the murine gastrointestinal tract. Infect. Immun. 9: Van der Waaij, D., J. M. Berghuis-de Vries, and J. E. C. Lekkerkerk-van der Wees Colonization resistance of the digestive tract in conventional and antibiotic-treated mice. J. Hyg. 69: Yang, M. G., K. Manoharan, and 0. Mickelson Nutritional contribution of volatile fatty acids from the cecum of rats. J. Nutr. 100: Zijlstra, J. B., J. Beukema, B. G. Wolthers, B. M. Byrne, A. Groen, and J. Dankert Pretreatment methods prior to gas chromatographic analysis of volatile fatty acids from fecal samples. Clin. Chim.Acta 78: