KeywordsFecal Microbiota Human Microbiome Project Metagenomic Approach Secondary Bile Acid Primary Bile Acid
Antibiotic-associated diarrhea is diarrhea (abnormally frequent intestinal evacuation with more or less fluid stools) that develops within a few hours following antibiotic treatment up to 8 weeks after antibiotic discontinuation.
The discovery of antibiotics is one of the most significant scientific achievements to date. Although antibiotics help treat most bacterial infections, they can have short- and/or long-lasting effects on the gastrointestinal tract microbiota depending on the class of antibiotic and the course of treatment. Antibiotic-associated diarrhea (AAD) is a common problem that results in extended hospital stays. The lack of treatment options for patients with refractory and recurrent disease further increases the morbidity and cost of this condition. The incidence of AAD is hard to pinpoint, but it is estimated that between 5 % and 25 % of patients taking antibiotics will develop diarrhea. AAD can be persistent, with symptoms continuing long after the discontinuation of antibiotic therapy (Bartlett 2002; Beaugerie and Petit 2004; McFarland 2008). Patients with AAD present with a broad spectrum of clinical syndromes ranging from uncomplicated mild diarrhea or “nuisance diarrhea” to severe colitis including toxic megacolon and in rare cases death. Antibiotics associated with a higher rate of AAD generally have a broader spectrum of antimicrobial activity and include clindamycin, cephalosporins, and ampicillin/amoxicillin (McFarland 2008). Onset and the degree of severity of AAD are dependent on many factors including the type of antibiotic, the health of the host, and exposure to other pathogens including Clostridium difficile. C. difficile is responsible for an estimated 10–20 % of AAD cases.
With the advent of the field of metagenomics, the genomic analysis of uncultured microorganisms, researchers are in a better position to define the structure and the function of the gastrointestinal tract after antibiotic treatment and prior to infection with pathogens. Presently, no metagenomic approaches have been applied to study AAD. Researchers have relied on microbial ecology and bacterial community analysis to study the gut microbiota before and after antibiotic treatment. Antibiotics alter the gut bacterial community structure, and this can suggest changes at the functional level as well. The objective of this review is to demonstrate how studies using both microbial ecology and metagenomic approaches have contributed to understanding the etiology of AAD and how future studies are needed to develop prevention and treatment.
The Indigenous Gut Microbiota
The human microbiota (the total bacterial population on the human body) is made up of 1014 cells, which is an order of magnitude higher than the number of host cells found on the human body. Over 70 % of this bacterial population resides within the gastrointestinal tract, with the majority of this community in the colon. It is estimated that the density of the gut microbiota is 1 × 1012 organisms per gram (dry weight) of feces.
Early studies of the gut microbiota were dependent on culture-based techniques to isolate anaerobic organisms, and it was thought to encompass 400–500 distinct species. Presently, it is estimated that the majority of gut bacteria cannot be cultivated. Since the advent of the International Human Microbiome Consortium in 2005 in conjunction with the National Institute of Health’s Human Microbiome Project, researchers from around the world are currently characterizing bacterial populations found on the human body. Researchers are just starting to understand what type and number of bacterial communities populate the human body and how they correlate with health and disease. With the use of culture-independent technologies focusing on the highly conserved and variable regions of the 16S rRNA gene, researchers are finding that initial estimates of bacteria found in the gut were highly underestimated. Advances in DNA sequence technology in the field of metagenomics have allowed bacterial genomes of the gastrointestinal population to be surveyed both structurally and functionally.
Disturbances in the gut microbiota have been associated with many diseases including obesity, diabetes, ulcerative colitis, and inflammatory bowel disease (Turnbaugh et al. 2006; Greenblum et al. 2012). Within the past 5 years, there have been seminal studies trying to define the variation between the gastrointestinal microbiota from one person to another and ultimately how perturbation of this community affects health outcomes. One of the first studies by Gill et al. looked at the intestinal metagenome of two healthy adults. The bacterial genes identified were important for energy metabolism and production of short-chain fatty acids (SCFAs), which provide energy to the intestine. The two bacterial phyla that made up the majority of the gut bacterial population were the Firmicutes and Bacteroidetes, while a lesser abundant population consisted of Proteobacteria, Actinobacteria, Fusobacteria, TM7, and Verrucomicrobia (Gill et al. 2006). Within the highly abundant Firmicutes phylum, the family of Lachnospiraceae made up the most of the population, specifically members of the Clostridium XIVa and IV groups. In a larger study in 2010, Qin et al. used a metagenomic approach to define the fecal microbiota of 124 patients who were defined as healthy, overweight, or obese or had irritable bowel disease, IBD. Within all patients surveyed, 40 % of all fecal microbial genomes were shared or contained a “core microbiome” with a predominance of the Firmicutes and Bacteroidetes phyla. Not only did the patients share a core microbiome, but they shared a minimal gut metagenome as well. A core group of bacterial functional genes were conserved throughout all samples including essential housekeeping genes for sugar metabolism and biosynthesis of molecules like SCFAs, amino acids, and vitamins (Qin et al. 2010). Another group surveying 154 adult patients also concluded there was a shared core microbiome although more at the functional gene level, not the organismal lineage level. Variation from the core microbiome was dependent on different physiologic states. It is still unclear and somewhat controversial if there is a “core microbiome” shared between all of us. It is becoming clearer that many things can cause an imbalance in the gut microbiota including antibiotic use, changes in diet, medications, immune system or inflammation, and pathogens. The field of metagenomics will help define how these perturbations affect the gut microbial community and ultimately how it affects human health.
Antibiotics and Structural Changes to the Gut Microbiota
Recent work has examined how antibiotics alter the gut microbiota and to what extent it causes diarrhea and decreases colonization resistance, making us more susceptible to colonization of pathogens. Dethlefsen et al. defined the gut microbiota in three healthy adults after a 5-day course of ciprofloxacin. They reported that ciprofloxacin administration decreased richness and diversity of the gut bacterial community. Four weeks after the end of antibiotic treatment, the microbiota for one returned to the levels it was prior to antibiotic treatment, and for another, it did not return for up to 6 months later. This raises the possibility that the gut microbiota may never completely return to the baseline state after antibiotic treatment. The consequences of these persistent changes are still not known (Dethlefsen and Relman 2011).
To understand how antibiotics perturb the gut in a controlled experiment, many researchers use murine models because they can control the host genetic background, the microbiota, and the feeding environment. Treatment of mice with an antibiotic cocktail consisting of ampicillin, gentamicin, metronidazole, neomycin, and vancomycin resulted in a tenfold reduction in fecal bacterial density (Hill et al. 2010). Antibiotic treatment was associated with significant temporal affects on the gut community including decreased abundance of the Firmicutes phylum and persistence of Bacteroidetes and Proteobacteria.
In another study, Antonopoulos et al. demonstrated that treatment of mice with a cocktail of antibiotics (amoxicillin, metronidazole, and bismuth or AMB) altered the gut microbiota with a persistent decrease in overall diversity. In control animals, the prevalence of Firmicutes and Bacteroidetes was very high, with a small percentage belonging to Proteobacteria, with those belonging to family Enterobacteriaceae. In antibiotic-treated animals, there was a shift, with Proteobacteria becoming the dominant phylum and Bacteroidetes and Firmicutes only making up a small portion of the total population (Antonopoulos et al. 2009). After AMB-treated animals were given time to recover off of antibiotics for 2 weeks, the gut microbiota was restored to baseline or prior to antibiotic treatment, whereas mice treated with a broader-spectrum antibiotic, cefoperazone, showed longer-lasting alterations up to 6 weeks of recovery.
Another study looked specifically at the effect of vancomycin on the murine gut microbiota and found that again Proteobacteria, specifically from the family Enterobacteriaceae, was predominant after treatment. When challenged with Enterococcus, vancomycin-treated mice were susceptible to infection (Ubeda et al. 2010). Alternatively, in another study, when mice were treated with vancomycin and then given 3 weeks to recover, the microbiota returned to baseline levels (Robinson and Young 2010). It is important to address that there are differences between the baseline murine gut communities in these studies that seem to be driven by the environment. Different housing and husbandry environments can play a role in shaping the mouse gut microbiota, and the baseline may differ from place to place.
Metagenomic studies examining the gastrointestinal tract are limited at this time. It is only through bacterial community analysis that the scientific community is starting to appreciate the damage that antibiotics can cause to the intestinal gut microbiota, and the implication of these changes is still being investigated.
Antibiotics and Functional Changes to the Gut Microbiota
In 2004, Young and Schmidt demonstrated that alteration of the gastrointestinal bacterial community structure following administration of the antibiotic amoxicillin-clavulanic acid was associated with the development of AAD. Antibiotic treatment resulted in a marked reduction of members of the Clostridiaceae family, which includes many of the butyrate-producing bacteria that are essential for colonic health (Young and Schmidt 2004). After a recovery period off of antibiotics, resolution of AAD was seen with a return of members from the Clostridiaceae family prior to antibiotic therapy.
When antibiotics are administered, both structure and function of the gastrointestinal microbiota are altered. Each day, 70 g of undigested carbohydrates makes its way to the colon. The colon contains a large anaerobic bacterial population that is essential for fermentation of complex carbohydrates and amino acids into SCFAs, primarily acetate, propionate, and butyrate. SCFAs, especially butyrate, are important for colonic health and have been shown to contribute energy to the colonic mucosa and aid in the regulation of gene expression, inflammation, differentiation, and apoptosis of host cells. Antibiotics that target these anaerobic bacterial populations leave the colon with an increased load of undigested carbohydrate, which can ultimately lead to osmotic diarrhea.
Another repercussion of an altered gut microbiota by antibiotics is a decreased metabolism of bile acids. Primary and secondary bile acids are important for dietary breakdown of fat and regulating cholesterol levels in the host. Gram-positive anaerobic bacteria that make up the intestinal microbiota are able to derive secondary bile acids from primary bile acids by two enzymatic reactions: deconjugation and 7α-dehydroxylation. Antibiotics alter the gut bacterial community depleting the population of 7α-dehydroxylating bacteria, therefore allowing a buildup of primary bile acids in the gastrointestinal tract. The buildup of specific non-dehydroxylated bile acids can cause electrolyte and water secretion leading to diarrhea (Beaugerie and Petit 2004).
The alteration of the gut microbiota by antibiotics also decreases colonization resistance to pathogens, which subsequently leads to an infectious diarrhea. Pathogens that are associated with AAD include C. difficile, Clostridium perfringens, Staphylococcus aureus, Klebsiella oxytoca, Candida species, and Salmonella species. The most common pathogen associated with AAD is C. difficile, which is responsible for an estimated 10–20 % of AAD cases (Bartlett 2002). A large area of study now focuses on how the intestinal microbiome contributes to host susceptibility to C. difficile.
Clostridium difficile AAD
C. difficile is the leading cause of antibiotic-associated colitis worldwide. In the USA alone, it causes an estimated 500,000 cases of diarrhea and colitis per year. The total annual excess healthcare costs due to this nosocomial infection have reached $3.2 B. It has been shown that antibiotic use disrupts the indigenous gut microbiota leading to a loss of colonization resistance and subsequent C. difficile infection (CDI). The mechanisms that mediate colonization resistance against C. difficile are still unknown.
C. difficile is an anaerobic, spore-forming, Gram-positive bacillus that was first isolated in 1935. In 1977, C. difficile was identified as the causative agent of AAD and in more severe cases lead to pseudomembranous colitis (PMC). Bartlett et al. developed the first rodent model using Syrian hamsters to fulfill Koch’s postulate; when hamsters were challenged with clindamycin and then C. difficile, they developed PMC (Bartlett et al. 2004). Although the disease used to be referred to as clindamycin colitis, many antibiotic treatments can be risk factors for CDI, but the highest risk can be found by using clindamycin, penicillins, cephalosporins, and fluoroquinolones. Other major risk factors for this disease include hospitalization, advanced age, and gastroenterology procedures.
C. difficile is the leading cause of hospital-acquired infections next to MRSA, methicillin-resistant S. aureus. It accounts for 20 % of all cases of diarrhea in hospitals and virtually all cases of PMC. The changing epidemiology of CDI has been associated with the emergence of hyperendemic, hypervirulent strains attributed with increased toxin production and sporulation ability, which also leads to ease of transmission (O’Connor et al. 2009). The main virulence factors associated with C. difficile include toxins TcdA and TcdB, which are glucosyltransferases. Although much attention has been placed on these microbial bacterial virulence factors, far less is known about how the intestinal microbiome contributes to host susceptibility to these emerging C. difficile strains.
When a patient reports symptoms of CDI, treatment regularly includes taking them off of antibiotics that are likely causing the AAD and starting them on oral metronidazole or vancomycin. After successful treatment, there are an increasing number of patients who experience one or more relapses of disease. In 2008, Chang et al. found that patients with recurrent C. difficile-associated diarrhea (CDAD) had decreased diversity of the fecal microbiota with highly variable composition, suggesting this could be a factor in colonization resistance (Chang et al. 2008). In another study looking at 599 patients after 72 h of admission to a Montreal hospital, the fecal microbiota of patients with CDAD showed an increased abundance of Firmicutes, Proteobacteria, and Actinobacteria with lower loads of Tenericutes and Bacteroidetes. They also found an association between CDAD patients and increased levels of Lactobacillaceae and Enterococcaceae (Manges et al. 2010).
Antibiotic-treated mice are susceptible to C. difficile challenge, and this allows researchers the opportunity to examine the role of the gastrointestinal microbiota in colonization resistance in an experimental murine model. Previous work using this model demonstrated that antibiotic pretreatment altered the gut microbiota by decreasing the relative abundance of Firmicutes and Bacteroidetes phyla, with an increase in Proteobacteria from family Enterobacteriaceae (Reeves et al. 2011). A broad-spectrum cephalosporin, cefoperazone, was also used to make mice susceptible to C. difficile infection, although cefoperazone showed significant and longer-lasting alterations to the mouse gut microbiota (Antonopoulos et al. 2009; Reeves et al. 2011). Most recently, Buffie et al. demonstrated that by giving mice clindamycin alone, it decreased microbial diversity and had long-lasting effects on the gut microbiota, ultimately making mice susceptible to C. difficile (Buffie et al. 2012). Specific changes to the indigenous gut microbiota by antibiotic use were associated with the loss of colonization resistance against C. difficile.
Future studies are needed to better understand the etiology of antibiotic-associated diarrhea. With the advent of “omics” technology including the field of metagenomics, researchers will be able to define both the structural and functional changes in the gastrointestinal tract after antibiotic administration.
Clearer bacterial community profiles will be generated, which will help define the role the gastrointestinal microbiota plays in antibiotic-associated diarrhea and in resistance to pathogens. Identifying specific bacterial populations and/or small molecules that are important for the health of the gut microbiota may prove important for future development of new classes of preventive or therapeutic agents.
The field of metagenomics has allowed researchers to define the bacterial populations that make up the gastrointestinal tract and determine what functional role they might play. Perturbation of the gut microbiota with antibiotics alters the structure of the gastrointestinal bacterial community, and this can take a toll on the bacterial metabolism of carbohydrates, SCFAs, and bile acids. These changes can have a profound effect on the host including onset of antibiotic-associated diarrhea and/or decreasing colonization resistance to pathogens. Using a metagenomic approach to study the gut microbiota will allow us to better understand the mechanism of AAD and ultimately aid in the development of new treatments.
- Bartlett JG, Onderdonk AB, Cisneros RL, et al. Commentary: Bartlett JG, Onderdonk AB, Cisneros RL, Kasper DL. Clindamycin-associated colitis due to a toxin-producing species of Clostridium in hamsters. J Infect Dis 1977; 136:701. J Infect Dis. 2004;190(1):202–9Google Scholar