Prolonged Impact of Antibiotics on Intestinal Microbial Ecology and Susceptibility to Enteric Salmonella Infection

Infect Immun. 2009 July; 77(7): 2741–2753. Published online 2009 April 20. doi: 10.1128/IAI.00006-09 Copyright © 2009, American Society for Microbiology Amy Croswell,1 Elad Amir,1 Paul Teggatz,1 Melissa Barman,1 and Nita H. Salzman1,2* Abstract The impact of antibiotics on the host's protective microbiota and the resulting increased susceptibility to mucosal infection are poorly understood. In this study, antibiotic regimens commonly applied to murine enteritis models are used to examine the impact of antibiotics on the intestinal microbiota, the time course of recovery of the biota, and the resulting susceptibility to enteric Salmonella infection. Molecular analysis of the microbiota showed that antibiotic treatment has an impact on the colonization of the murine gut that is site and antibiotic dependent. While combinations of antibiotics were able to eliminate culturable bacteria, none of the antibiotic treatments were effective at sterilizing the intestinal tract. Recovery of total bacterial numbers occurs within 1 week after antibiotic withdrawal, but alterations in specific bacterial groups persist for several weeks. Increased Salmonella translocation associated with antibiotic pretreatment corrects rapidly in association with the recovery of the most dominant bacterial group, which parallels the recovery of total bacterial numbers. However, susceptibility to intestinal colonization and mucosal inflammation persists when mice are infected several weeks after withdrawal of antibiotics, correlating with subtle alterations in the intestinal microbiome involving alterations of specific bacterial groups. These results show that the colonizing microbiotas are integral to mucosal host protection, that specific features of the microbiome impact different aspects of enteric Salmonella pathogenesis, and that antibiotics can have prolonged deleterious effects on intestinal colonization resistance. DISCUSSION Antibiotic use carries both benefit and risk. Antibiotic therapy is critical for the treatment of life-threatening infection, but misuse of antibiotics leads to the development of antibiotic resistance in common pathogens. Even routine and appropriate use of antibiotics may have a detrimental impact on the host microbial ecosystem, which is important for host mucosal protection In this study, we investigated how the use of oral antibiotics perturbs the intestinal microbial ecosystem in mice and the impact of that disruption on host susceptibility to a common enteric pathogen, Salmonella enterica serovar Typhimurium. Using molecular methods to identify and quantify bacterial numbers, a more-complete understanding of the impact of antibiotics on the intestinal microbial ecosystem can be determined. While antibiotic treatment resulted in the loss of culturable bacteria, none of the antibiotic combinations tested was able to sterilize the gut. Since antibiotic treatment was limited to 1 week, it is reasonable to consider that the treatments would not be able to eliminate the slower-growing organisms. Nevertheless, complete elimination of the bacterial component of the intestinal microbiota by antibiotics is difficult to achieve and cannot be determined by culture-based methods. The various antibiotic regimens used resulted in changes in the abundance and composition of the intestinal microbiome that were antibiotic specific. The firmicute class of bacteria (including the E. rectale-C. coccoides group, Lactobacillus sp., and SFB) appears to be the most susceptible to all of the antibiotics used. This class of bacteria is also the most susceptible to disruption by diarrheal illness. After antibiotic treatment, the intestinal biome gravitates over time to that of untreated mice. Using bacterial culture analysis, after the initial elimination of culturable bacteria, anaerobes recover to the levels found in untreated mice within 3 days. Levels of aerobic bacteria expand dramatically but by 3 weeks closely approximate the levels found in untreated mice. Molecular analysis reveals that total bacterial numbers rapidly recover, driven by the swift recovery of the most dominant bacterial group, the E. rectale-C. coccoides group, members of the Firmicutes. Another member of the Firmicutes, the lactobacilli, also rapidly recovers, while other groups represented at a lower abundance, such as SFB, do not. If increased susceptibility to Salmonella infection was correlated with the quantity of colonizing bacteria eliminated by antibiotics, one would predict that streptomycin-treated mice would show less translocation and intestinal colonization by Salmonella and milder enteritis than would streptomycin-bacitracin- or AVNM-treated mice, but this was not evident in this study. Oral Salmonella challenge of antibiotic-treated mice resulted in comparable increases in intestinal Salmonella colonization, enteritis, and invasion irrespective of the antibiotic combinations used. Antibiotic recovery experiments allowed a more-complete dissection of specific aspects of Salmonella enteritis, including susceptibility to pathogen invasion, intestinal colonization, and mucosal inflammation. Interestingly, over the time course of biome recovery from antibiotics, mice regained resistance to Salmonella translocation rapidly, but susceptibility to increased Salmonella colonization and local mucosal inflammation persisted. This suggests that invasion is mediated by different pathogen-commensal-host interactions than colonization. The prevention of Salmonella translocation is associated with a recovery of total bacterial numbers, which is driven by the recovery of the dominant firmicute populations (the E. rectale-C. coccoides group and Lactobacillus sp.). The presence of an intact commensal biota could be preventing Salmonella invasion by effectively competing with Salmonella for attachment sites and nutrition, reducing the numbers of luminal Salmonella cells available for invasion. This would account for the association between biota recovery and resistance to translocation. However, commensal interactions with the host may also have an important role in the prevention of translocation. Several innate antimicrobial effectors produced by the intestinal epithelium are induced by intestinal colonization, including angiogenin 4 (19), RegIIIγ (9), and intestinal alkaline phosphatase (7), all of which have dual roles in intestinal homeostasis and host defense. The reduction in levels of intestinal bacteria by antibiotic treatment can result in decreased levels of expression of these effectors (44), allowing increased pathogen translocation. Specific bacterial species may also play a role in host protection. SFB, for example, are highly susceptible to both exogenous antibiotics, as shown here, and endogenous antimicrobial peptide activity (our unpublished observations). This organism is both immune stimulatory (38, 42) and highly immune responsive (20), with increased numbers found in immunoglobulin A-deficient mice (41). The absence of SFB has been associated with increased susceptibility to enteric pathogens (15) and enteritis. While SFB abundance does not correlate with pathogen translocation, it may be critical for the increased susceptibility of animals to Salmonella colonization and enteritis. It is possible that SFB have a directly protective effect by interacting with Salmonella and preventing pathogen interactions with the mucosa. SFB may also be acting through the stimulation of the mucosally associated lymphoid tissue, resulting in more-effective mucosal host responses to the invading pathogen. Additional work, using gnotobiotic mice and controlling for the presence of this unculturable bacterial group, is needed to address this issue. The finding that both extremely minimal disruptions in the intestinal biome, like those present 3 weeks after antibiotic withdrawal, and extremely profound disruption, such as that with AVNM treatment, result in similar luminal colonization by Salmonella was both unexpected and intriguing. It is unlikely that the intestinal biota is responsible for restricting the growth of Salmonella in these circumstances since it varies so profoundly between treatments. One explanation is that the limitation in Salmonella colonization is a result of innate host immune responses. Another possibility is that this is a colonization set point due to quorum sensing by Salmonella. Additional work will be required to distinguish these processes and determine whether this is specific to Salmonella or common to other enteric pathogens. It is evident that the intestinal microbial ecosystem serves an important but incompletely defined role in mucosal protection. In this study, we have demonstrated that although antibiotics cannot sterilize the intestinal tract, they can have a profound impact on intestinal colonization. The murine intestinal microbiome recovers from antibiotic disruption to closely recapitulate that of untreated mice, supporting the hypothesis that attributes of the host select for a core microbiome, as previously suggested using germfree models. Despite the rapid recovery of several measurable parameters of the biome, residual subtle alterations in bacterial composition can persist and result in profoundly enhanced susceptibility to bacterial enteritis. In conclusion, the host drives the selection of a core biome in which bacterial quantity and composition contribute to intestinal colonization resistance. Minimal disruption of this complex balance by antibiotics can result in prolonged harmful effects on the ability of the host to resist infection.