Inhibition of Shigella sonnei and Enterotoxigenic

MICROBIAL ECOLOGY ZN HEALTH AND DISEASE VOL. 2 153-161 (1989) Inhibition of Shigella sonnei and Enterotoxigenic Escherichia coli by Volatile Fatty Acids in Mice P. PONGPECH* and D. J. HENTGES Department of Microbiology, Texas Tech University, Health Sciences Center, Lubbock, Texas 79430, USA. Received 24 August 1988; revised 22 December 1988 The role of volatile fatty acids (VFA) in resistance against the colonisation of the intestinal tract ofmice with two enteric pathogens, Shigella sonnei 3SR and enterotoxigenic E. coli 2SR (ETEC 2SR) was determined. The implantation dose 50 (IDSO) of S. sonnei 3SR strain for untreated mice was greater than 5.0 x lo9 viable cells and was 1.8 x lo6 viable cells for ETEC 2SR. Administration of streptomycin to the mice prior to challenge lowered the ID50 to fewer than one viable cell for both organisms. The antibiotic caused an increase in the ph of caecal contents from 6.66 in untreated animals to 6.94 in treated animals and a decrease in total VFA concentrations from 109.1 1 microequivalent per ml in untreated animals to 78.3 1 microequivalent per ml in treated animals. The discontinuation of streptomycin treatment resulted in a gradual restoration of colonisation resistance accompanied by an increase in total VFA concentration and a decrease in ph of caecal contents. These values returned to pre-treatment levels by day 7. VFA were added to nutrient broth in concentrations present in the caeca of either untreated or streptomycin treated mice and the ph levels were adjusted accordingly. S. sonnei 3SR multiplied in broth adjusted to simulate conditions present in the caeca of treated mice but failed to multiply in broth adjusted to simulate conditions present in the caeca of untreated mice. ETEC 2SR, on the other hand, multiplied in both types of broth but its total populations after 24 h incubation were significantly smaller in broth adjusted to simulate conditions present in the caeca of untreated mice. The results demonstrate that the multiplication of both pathogens is inhibited by the presence of VFA. They indicate that VFA operating at the ph level present in the intestinal tract of conventional mice represent important factors responsible for resistance against colonisation with S. sonnei and ETEC. KEY WORDS-Shigella sonnei; Enterotoxigenic E. coli; Volatile fatty acids; Mice. INTRODUCTION Several mechanisms have been proposed to account for the inhibition of pathogens by indigenous intestinal flora component^.^^^^^*^*^*^^.'^ Of the various mechanisms, VFA production appears to play a very important role. q5 VFA which are present throughout the intestinal tract of mammals are produced as a result of the fermentation of soluble carbohydrates and other nutrients by components of the intestinal flora. MeynellI3 and Bohnhoff et al.3,4 demonstrated that the protection against Salmonella enteritidis infection in mice was the result of VFA production by intestinal flora and a concomitant decrease in the oxidation-reduction potential and ph of the intestinal contents. Maier et al. found that caecal contents obtained from germfree mice, which had high Eh and ph values and contained little VFA, supported the multiplication of Shigella *Author to whom correspondence should be addressed. 0891-060X/89/0301S3-09 $05.00 0 1989 by John Wiley & Sons, Ltd. JEexneri. When the contents were adjusted to simulate conditions found in conventional mice (low Eh and ph values and high VFA concentrations), the growth of S. JEexneri was inhibited to a degree similar to that observed in contents from conventional mice. Hentges et a1. found that the ph of caecal contents of untreated mice was significantly lower and the VFA concentrations were significantly higher than those of streptomycin treated mice. More recently Que and Hentges and Que et a1. demonstrated that Salmonella typhimurium easily colonised streptomycin treated mice but failed to colonise untreated controls suggesting a role of VFA and ph in the inhibitory process. Recently we reported differential sensitivities of intestinal pathogens to VFA in broth culture.16 Among the strains studied, S. sonnei 3SR was most sensitive to VFA and ETEC 2SR the most resistant. These two strains were therefore used to examine more fully the role of VFA in colonisation

154 P. PONGPECH AND D. J. HENTGES *1 T T T T T 1 2 4 7 14 Days After Dlscontlnuatlon of Treatment Figure 1. Effect of discontinuation of streptomycin treatment on the population levels of Shigellu sonnei 3SR in mouse caeca. Results are expressed as the mean log,, of values obtained from three separate experiments. T-bars indicate standard errors of the mean. (m =treated, H =treatment discontinued, I =untreated) Challenge dose = 4.3 x lo4 viable cells. resistance. The results of this study indicate that the presence of high concentrations of VFA in the intestinal tract is essential for protection against colonisation of mice by both S. sonnei 3SR and ETEC 2SR. MATERIALS AND METHODS Cultures S. sonnei was obtained from a lyophilised stock culture in our laboratory. ETEC was kindly provided by Dr Donald K. Winsor Jr, University of Texas Health Science Center, Houston, Texas. Spontaneous mutants of both strains resistant to 1 mg/ml streptomycin sulphate were selected by the method of Stocker (described by Joys, 1961)." The mutants were indistinguishable from the parent strains morphologically, biochemically and in growth rate kinetics. Mice Outbred Swiss white mice (Cox variety, Laboratory Supply Co., Indianapolis, IN) weighing 25-30gm were used. Animals were housed in groups of five in cages with wire mesh bottoms (after challenge, mice were housed individually in each cage) and were fed Purina Laboratory rodent Chow 500 (Ralston Purina Co., St Louis, MO) and water ad libitum. Only mice free of infection were used. This was established prior to experimentation by screening faecal samples for various pathogens. Implantation Dose 50 The inoculum required to colonise the intestinal tract of 50 per cent of the mice (termed implantation dose 50; ID50) was determined for streptomycin treated and untreated mice. Streptomycin was administered in drinking water at a concentration of

COLONISATION RESISTANCE TO S. SONNEI AND ETEC lo 1 155 8-6- 4-2- 0' 1 2 4 7 14 Days After Dlscontlnuatlon of Treatment Figure 2. Effect of discontinuation of streptomycin treatment on the population levels of Enterotoxigenic E. coli 2SR in mouse caeca. Results are expressed as the mean log,, of values obtained from three separate experiments. T-bars indicate standard errors of the mean. (W =treated, =treatment discontinued, H =untreated) Challenge dose = 1.8 x lo4 viable cells 5 mg/ml. On the seventh day of streptomycin treatment both the treated and untreated control mice were challenged orogastrically with graded numbers of S. sonnei 3SR or ETEC 2SR suspended in 0.1 ml of Brain Heart Infusion (BHI) broth (Difco). Three days after challenge, two faecal pellets collected from each mouse, were emulsified in 1 ml sterile saline and seeded on MacConkey agar (Difco) containing 1 mg/ml streptomycin sulphate (S-Mac agar). The presence or absence of bacteria was recorded after 24 h incubation at 37 C. The ID50 was calculated by using the method of Reed and Muench. Discontinuation of Streptomycin Treatment The re-establishment of resistance to colonisation with S. sonnei 3SR and ETEC 2SR in the mice was examined after discontinuation of streptomycin treatment. Thirty mice were housed individually in cages with wire mesh bottoms. Twenty of the mice were given water containing 5 mg/ml streptomycin sulphate ad libitum for a period of 1 wk and ten control mice were given distilled water. One week after initiation of antibiotic treatment, streptomycin was withdrawn from ten of the 20 mice on treatment and these animals given distilled water for the remainder of the experiment. On days 1,2,4, 7 and 14 after discontinuation of antibiotic, two mice from each group (control, treated, and treatment discontinued) were challenged orogastrically with 0-1 ml of a lo5 CFU/ml culture of either S. sonnei 3SR or ETEC 2SR. Forty-eight hours after challenge, mice were sacrificed by cervical dislocation, caeca were removed and bacterial counts/gm caecal content determined by seeding serial dilutions on S-Mac agar. Plates were incubated for 24 h at 37 C after which colonies were counted.

156 P. PONGPECH AND D. J. HENTGES 7.00 T T T 6.75 6.50 6.25 6.00 -i 1 2 4 7 14 Days After Discontinuation of Treatment Figure 3. Effect of discontinuation of streptomycin treatment on the ph of mouse caecal contents. Results are expressed as mean values obtained from three separate experiments. T-bars indicate standard errors of the mean. (W =treated, S =treatment discontinued, =untreated) VFA Concentrations and ph levels in Caecal Contents The VFA concentrations in caecal contents from the three groups of mice described in the previous section were determined as described by Rolfe. Briefly, the caecal contents were diluted in 1 ml of distilled water, the homogenate acidified with 0.1 ml of a 50 per cent solution of sulphuric acid per gram of content, and the acidified sample allowed to stand at 4 C overnight, after which an equal volume of anhydrous ethyl ether was added for extraction of VFA. The ether-caecal content homogenate was mixed by inverting the tubes 20 times and venting occasionally to release pressure. The mixture was spun 2500 rpm for 15 min using an IEC clinical centrifuge (International Equipment Co., Needham Heights, MS) to break the emulsion. Fourteen microliters of the ether extract were injected into a gas chromatograph equipped with a thermal conductivity detector (series 550; Gow-Mac Instrument Co., Bridgewater, NJ). The chromatograph and its settings were described previously. Peaks of the VFA were identified and quantitated by reference to tracings of different concentrations of VFA standards. VFA concentrations were reported as microequivalents per gram of caecal content (wet weight). The ph of caecal contents from the three groups of mice was measured in situ by inserting the tip of a micro combination ph electrode (Microelectrodes, Inc., Londonberry, NH) into the lumen of the caecum of a mouse within 10 min of sacrifice. Care was taken not to touch the mucosa with the electrode which causes fluctuation in ph readings. The ph electrode was standardised at room temperature against ph 4 and 7 buffers (American Scientific Products, McGaw Park, IL). Inhibitory Activity of Volatile Fatty Acids andph on S. sonnei 3SR and ETEC 2SR in Nutrient Broth VFA, in concentrations corresponding to those observed in caecal contents of streptomycin treated and untreated mice (day 7 levels), were added to nutrient broth in an anaerobic glove box isolator

COLONISATION RESISTANCE TO S. SONNEI AND ETEC 157 rn I L C a c.- 3 100 50 -l 1 2 4 7 14 Days After Discontinuation of Treatment Figure 4. Effect of discontinuation of streptomycin treatment on the total volatile fatty acids concentrations in mouse caecal contents. Results are expressed as the sum of mean values of individual acid concentration obtained from three separate experiments. (W =treated, F2 =treatment discontinued, =untreated) (model 1024; Forma Scientific, Marietta, OH). The ph levels of the broth were then adjusted to 6.94 and 6.66 respectively. The broth, which had been prereduced in the isolator for 48 h, was inoculated with lo5 viable cells of either S. sonnei 3SR or ETEC 2SR and incubated anaerobically at 37 C. At the time of incubation and at 6, 12, 24, 48 and 72 h intervals thereafter, serial 10-fold dilutions of the broth culture were prepared. The dilutions were seeded on S-Mac agar. Plates were incubated at 37 C for 24 h after which colonies were counted. Counts were expressed as viable organisms per ml culture. Statistical Analyses Statistical evaluations of the significance of differences in ph levels, VFA concentrations, log,,, bacterial counts in broth and in caecal contents of streptomycin treated and untreated mice were performed using the Fisher s least significant difference test at the 95 per cent confidence interval level. RESULTS ID50 Determinations ID50 determinations of the two strains for mice were done first. The number of S. sonnei 3SR required to colonise the intestinal tract of 50 per cent of untreated mice was 1.2 x lo9 viable organisms while the number for streptomycin treated mice was fewer than 1.2 viable organisms. The ID50 of ETEC 2SR in untreated mice was 6.8 x lo6 viable organisms but was fewer than 6.8 viable organisms in streptomycin treated mice. Efect of Discontinuation of Streptomycin Treatment on the Population Levels of S. sonnei 3SR and Enterotoxigenic ETEC 2SR in Mouse Caeca S. sonnei 3SR failed to colonise the caeca of untreated mice except on day 1 when four of ten

P. PONGPECH AND D. J. HENTGES 1 " I I I I I I 0 12 24 36 48 60 72 Hours Incubation Figure 5. Growth of Shigella sonnei 3SR in nutrient broth (NB). Each point represents the mean log,, of viable bacteria per ml culture obtained from three separate experiments. (-li-=control NB, ph 6.8, -*-=treated conditions, NB+78.31 UeqVFA ph 6.94, -E-= untreated conditions, NB + 109.11 UeqVFA ph 6.66) mice were colonised with an average of approximately 103/gm viable S. sonnei 3SR (Figure 1). By contrast, in streptomycin treated animals, population levels of S. sonnei 3SR remained at high levels between lo7 and lo8 viable organisms/gm throughout the 14 day experimental period. When streptomycin treatment was discontinued, there was a gradual decrease in the population levels of the bacteria with time. On day two after discontinuation of treatment, population levels of lo6 viable S. sonnei 3SR/gm were observed. By day seven, no S. sonnei 3SR were isolated. In untreated mice, the population levels of ETEC 2SR were approximately lo2 viable cells/gm except for day seven when no ETEC 2SR were recovered from these animals (Figure 2). The population levels of ETEC 2SR in streptomycin treated mice were between lo8 and lo9 viable organisms/gm. When streptomycin treatment was discontinued, the population levels of the bacteria decreased gradually with time from approximately lo7 viable cells/gm on day one after antibiotic discontinuation to approximately lo2 viable cells/gm on day four. Populations remained at a low level from day four after discontinuation through day 14. Determination of ph and VFA Concentrations in Caecal contents after Discontinuation of Streptomycin Treatment The ph level of caecal contents was found to be higher in streptomycin treated mice than in untreated mice (Figure 3). After discontinuation of streptomycin treatment, the ph gradually decreased from the treated level of approximately 6.94 to an approximated level of 6.66. Three major short chain volatile fatty acids were identified in the caecal contents, acetic, n-butyric and propionic acids. Acetic acid was present in the highest concentration. There was a fluctuation in total concentrations of VFA in the caeca of both treated and untreated mice over the 14 day period (Figure 4). However, the acid concentrations in the caeca of untreated mice were consistently higher than those in the caeca of treated mice (pgo.05) over this time period. The figure also shows that

COLONISATION RESISTANCE TO S. SONNEIAND ETEC 159 2 1 I I I I I I 0 12 24 36 48 60 72 Hours Incubation Figure 6. Growth of Enterotoxigenic E. coli2sr in nutrient broth (NB). Each point represents the mean log,, of viable bacteria per ml culture obtained from three separate experiments. (-m-=control NB, ph 6.8, -+-=treated conditions, NB+78.31 UeqVFA ph 6.94, -0-=untreated conditions, NB+ 109.11 UeqVFA ph 6.66) total VFA concentrations in the caeca of mice in which streptomycin treatment was discontinued increased at a steady rate with time and reached the untreated level within 7 days. Inhibitory Activity of VFA andph on S. sonnei 3SR and ETEC 2SR in Nutrient Broth S. sonnei 3SR multiplied in nutrient broth containing a total of 78.31 Ueq VFA adjusted to ph 6.94, simulating conditions present in caecal contents of streptomycin treated mice and attained a population of approximately 5-0 x lo7 viable cells per ml after 24 h incubation under anaerobic conditions (Figure 5). By contrast, the organism failed to multiply in nutrient broth containing a total of 109-1 1 Ueq VFA adjusted to ph 6.66, simulating conditions present in caecal contents of untreated animals. Under the latter conditions, the population declined to approximately ten viable cells per ml. ETEC 2SR was less sensitive to the presence of VFA in nutrient broth than S. sonnei 3SR (Figure 6). The organism was able to multiply in broth adjusted to simulate conditions present in untreated mice and reached a population level of approximately 7.0 x lo6 viable cells per ml after 48 h incubation. However, after 48 h in nutrient broth simulating conditions in treated mice, the ETEC 2SR population was significantly greater (approximately 7.0 x lo7 viable cells per ml) than that in broth simulating conditions in untreated mice ( p < 0.05). In nutrient broth adjusted to ph levels observed in caecal contents from streptomycin treated and untreated mice (ph 6.94 and 6.66, respectively) without added VFA, the population levels of both S. sonnei 3SR and ETEC 2SR were the same as in unadjusted control broth at ph 6.80. This demonstrated that ph variation in itself had no effect on the multiplication of the pathogens. DISCUSSION Short chain volatile fatty acids (VFA) have long been thought to play an important role in excluding pathogens from the intestinal tract. Meynell' and

160 Bohnhoff et ~ 1. demonstrated ~ 3 ~ that the multiplication of Salmonella enteritidis was inhibited in vitro by suspensions of intestinal contents from conventional mice. VFA were recovered from the intestinal contents of the mice in concentrations that inhibited S. enteritidis multiplication at the ph (6.1) and oxidation-reduction potential (Eh - 200 mv) of the caecum. Oral administration of streptomycin resulted in a decrease in total VFA of intestinal contents and an increase in oxidation-reduction potential and ph, providing conditions that favoured multiplication of S. enteritidis. Sometime later, Maier et al. l2 performed experiments to determine if VFA are responsible for the inhibition of Shigellajexneri in the intestinal tract of mice. Comparative studies done with conventional and germfree mice showed that s. jexneri multiplied more rapidly and attained greater populations in germfree than conventional mice. The oxidation-reduction potential and ph of caecal contents of conventional mice were lower and VFA concentrations higher than in caecal contents of germfree mice. Although the role of VFA as control factors in microbial population development is well established, there is controversy about their importance as determinants of colonisation resistance in the intestinal tract. Studies by Freter and Abrams indicate that E. coli populations are not influenced by VFA concentrations in the intestine. Germfree mice were associated with either whole flora obtained from conventional mice or with various mixtures of facultative and anaerobic intestinal bacteria and were then implanted with an E. coli strain. The concentrations of the VFA associated with the different floras in the caeca did not correlate with the E. coli population levels. Similar experiments of Koopman et al. l1 confirmed these findings. Thus, the evidence is conflicting. Experiments reported in this paper, however, provide additional information to support the theory that VFA, generated by the indigenous flora, are important colonisation resistance factors. First, a positive correlation was observed between the VFA resistance of S. sonnei 3SR and ETEC 2SR and the ability of these organisms to colonise the caeca of the mice. The number of VFA sensitive S. sonnei 3SR required for colonisation of 50 per cent of untreated mice was more than 1000- fold greater than the number of more VFA tolerant ETEC 2SR (Figures 1 and 2). This might be explained by the results from broth culture.studies which showed that S. sonnei 3SR was far more sensitive to the effects of VFA at ph levels present in untreated mice than was ETEC 2SR (compare P. PONGPECH AND D. J. HENTGES Figures 5 and 6). The important resistance mechanism, therefore, appeared to be the relatively low ph and high concentration of VFA in the intestinal tract of these animals. Streptomycin administration lowered VFA concentrations and increased the ph of caecal contents of mice and greatly diminished colonisation resistance against s. sonnei 3SR and ETEC 2SR. When treatment with this antibiotic was discontinued, VFA concentrations and ph levels of caecal contents returned to pre-treatment levels within seven days (Figures 3 and 4). This produced conditions which inhibited multiplication of pathogens to a degree similar to that observed in untreated animals. It is reasonable to conclude, therefore, that the relatively high VFA concentrations and low ph levels present in the caecal contents of conventional mice are important factors responsible for colonisation resistance. Since ETEC 2SR is more resistant to the inhibitory effects of VFA than S. sonnei 3SR, the mice are more susceptible to colonisation with this pathogen than with S. sonnei 3SR. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. Arnold WJ. (1974). Inhibition of Shigella species by environmental factors in the mouse intestine. M.S. thesis, University of Missouri, Columbia. Bohnhoff M, Miller CP. (1962). Enhanced susceptibility to Salmonella infection in streptomycin-treated mice. Journaloflnfectious Disease 111,117-12 7. Bohnhoff M, Miller PC, Martin WR. (1964). Resistance of the mouse s intestinal tract to experimental Salmonella infection. I. Factors which interfere with interaction of infection by oral inoculation. Journal of Experimental Medicine 120,805-8 16. Bohnhoff M, Miller PC, Martin WR. (1964). Resistance of the mouse s intestinal tract to experimental Salmonella infection. 11. Factors responsible for its loss following streptomycin treatment. Journal of Experimental Medicine 120,817-828. Byrne BM, Dankert J. (1979). Volatile fatty acids and aerobic flora in the gastrointestinal tract of mice under various conditions. Infection and Immunity 23,559-563. Floch MH, Binder JJ, Filburn B, Gershengoren W. (1972). The effect of bile acids on intestinal microflora. American Journal of Clinical Nutrition 25, 1418-1426. Freter R, Abrams GD. (1972). Function of various intestinal bacteria in converting germfree mice to the normal state. Infection and Immunity 6, 119-1 26. Hentges DJ, Maier BR. (1970). Inhibition of Shigellaflexneri by the normal intestinal flora. 111. Interactions with Bacteroides fragilis strains in vitro. Infection and Immunity 2,364-370.

COLONISATION RESISTANCE TO S. SONNEI AND ETEC 161 9. 10. 11. 12. 13. 14. Hentges DJ, Que JU, Casey SW, Stein AJ. (1985). Protective role of the intestinal flora against infection with Pseudomonas aeruginosa in mice: influence of antibiotics on colonization resistance. Infection and Immunity 47,118-1 22. Joys TM. (1961). Mutation of flagella antigen i in Salmonella typhimurium. Ph.D. thesis. University of London, London, England. 17. Koopman JP, Welling GW, Huybregts AWM, Mullink JWMA, Prins RA. (1981). Association of germ-free mice with intestinal microfloras. Zeitschriyt Fur versuchstierkunde 23, 145-1 54. Maier BR, Onderdonk AB, Baskett RC, Hentges DJ. (1972). Shigella-indigenous flora interactions in mice. American Journal of Clinical Nutrition 25, 1433-1 440. 19. Meynell GG. (1963). Antibacterial mechanisms of the mouse gut. 11. The role of Eh and volatile fatty acids in the normal gut. British Journal of Experimental Pathology 44,209-219. Miller CE, Feely JC. (1975). Competitive effects of intestinal microflora on Vibrio chlorae in gnotobiotic mice. Laboratory Animal Science 25,454458. 5. 6. 18. 20. Onderdonk AB. (1973). Mechanism of Shigella inhibition by the mouse intestinal microflora. Ph.D. dissertation, University of Missouri, Columbia. Pongpech P, Hentges DJ, Casey SW. (1987). Role of volatile fatty acids in colonization resistance against enteric pathogens. Abstracts Annual Meeting American Society of Microbiology, p. 56. Que JU, Hentges DJ. (1985). Effect of streptomycin administration on colonization resistance to Salmonella typhimurium in mice. Infection and Immunity 48,169-174. Que JU, Casey SW, Hentges DJ. (1986). Factors responsible for increased susceptibility of mice to intestinal colonization after treatment with streptomycin. Infection and Immunity 53,116123. Reed LJ, Muench H. (1938). A simple method of estimatingfifty percent end points. American Journal of Hygiene 27,493-497. Rolfe RD. (1984). Role of volatile fatty acids in colonization resistance to Clostridium dificile. Infection and Immunity 45, 185-191.