Accumulation of data on the human microbiome

The tremendous expansion of information collected on the human microbiome in recent years is highlighted by data generated through several large-scale endeavors to characterize the human microbiome, namely the European Metagenomics of the Human Intestinal Tract (MetaHIT) and the NIH-funded Human Microbiome Project (HMP) (6, 7). In 2010, the initial MetaHIT consortium study reported sequencing 3.3 million non-redundant fecal microbial genes, representing almost 200 times the quantity of microbial DNA sequences reported in all previous studies (7). In July 2014, a combined set of metagonomic sequencing data from 1267 gut metagenomes from 1070 individuals, including 760 European samples from MetaHIT, 139 American samples from HMP and 368 Chinese samples from a large diabetes study, was published with a non-redundant gene catalog of 9.8 million microbial genes (*8). Each sample contained about 750,000 genes or about 30 times the number of genes in the human genome, and less than 300,000 genes were shared by greater than 50% of individuals. The majority of the new genes identified in this latest study were relatively rare, found in less than 1% of individuals. This collection is thought to contain nearly a complete set of genes for most human gut bacteria and illustrates the quantity and variability of the human microbiome.

Structure and dynamics of the healthy adult microbiota

Characterization of the microbiome in healthy individuals is an important initial step in understanding the role of the microbiome in contributing to health and disease. Healthy adult humans each typically harbor more than 1000 species of bacteria belonging to a relatively few known bacterial phyla with Bacteroidetes and Firmicutes being the dominant phyla (9). The microbiota of the gut is quite diverse (Table 1) compared to other body sites, and there is considerable variation in the constituents of the gut microbiota among apparently healthy individuals (10). As a way of accounting for the microbial variability among healthy individuals, researchers have tried to identify certain stable patterns of microbial populations in the human population (11). Data from the HMP was used to identify community types at different body sites based on statistical analysis of the configuration of multiple bacterial taxa (*12). Four distinct communities were found in the stool, and metadata factors associated with community types included breastfeeding, gender and education. Interestingly, community types in the oral cavity were predictive of those in the stool though the specific constituents were different. In a recent study of 37 healthy American adults not taking antibiotics, greater than 70% of fecal bacterial species within an individual were stable over 1 year, and few additional changes were measured out to 5 years (*13). Calculations showed that species were likely stable over decades if not for an individual’s entire lifetime as evidenced by species shared with adult family members but not with unrelated individuals. While we have gained a deep understanding of the structure and dynamics of the microbiome in healthy humans, this effort is complicated by significant variability in the population, modest variability over time within an individual, and uncertainty over the most meaningful ways to characterize the microbiota.

Metabolic functions of the microbiota

After characterizing microbial community membership and dynamics, it is critical to understand the functional activities that ultimately affect host physiology. Gut microbiota are integral to host digestion and nutrition, and they can generate nutrients from substrates that are otherwise indigestible by the host. For instance, xyloglucans are commonly found in dietary vegetables such as lettuce and onions, and the capacity for microbial digestion of xyloglucans was recently mapped to a single locus in a certain species of Bacteroides (*14). The ability to digest xyloglucans was shown to be a relatively rare trait in members of the phylum Bacteroidetes, and the importance of this capability to the human host was demonstrated by analysis of a public metagenome database showing that 92% of individuals contained at least one of these rare Bacteroides species capable of digesting xyloglucans. These findings illustrate how humans have cultivated mutually beneficial relationships with gut microbiota with implications for diet and nutrition.

Microbes liberate short chain fatty acids (SCFA) from indigestible dietary fibers, and SCFA are an important energy source for intestinal mucosa and critical for modulating immune responses and tumorigenesis in the gut. The role of butyrate, an abundant bioactive SCFA in the gut, plays a complex role in colon cancer that seems to be concentration and context dependent as illustrated by two recent preclinical studies. Butyrate was reported to promote tumorigenesis in transgenic mice with combined tumor suppressor gene (APC) mutation and mismatch repair gene (MSH2) deficiency, because tumor formation was decreased by antibiotic treatment or low carbohydrate diet, both of which decrease butyrate levels, and increased by feeding butyrate to antibiotic-treated mice (*15). Conversely, butyrate was reported to inhibit tumorigenesis, because mice deficient in Grp109a, a receptor for butyrate, had increased tumorigenesis promoted by inflammatory stimuli or APC mutation and signaling through Grp109a inhibited tumorigenesis induced by these stimuli (*16). Further investigations into the role of butyrate produced by microbiota in colitis and colorectal cancer are awaited. The studies discussed in this section demonstrate the need to assess the function of the microbiota in order to better understand its role in health and disease.

Host-microbe interactions on the immune system

Interactions between the microbiota and the host immune system are numerous, complex, and bidirectional. The immune system must learn to tolerate the commensal microbiota and respond appropriately to pathogens, and in turn the microbiota is integral to educating the immune system to function properly. Here we highlight studies that describe how members of the microbial community promote the differentiation of anti-inflammatory regulatory T cells (Treg) to illustrate how the microbiota can influence immune homeostasis. A series of experiments showed that collections of nonpathogenic species of Clostridia from clusters IV, XIVa and XVIII, isolated after application of a series of non-specific selection steps, were capable of inducing colonic Treg, and one mechanism may involve the production of butyrate that affects epigenetic control of the Foxp3 promoter that controls Treg development (17, **18, 19-22). In germ-free mice that do not contain endogenous microbiota, another group also devised a novel method to screen human fecal samples for bacterial strains able to promote Treg development, and they noted this functional capacity in more strains than anticipated (*23). While not discussed here, there is evidence detailing host-microbe interactions that influence immune functions at all levels from the initial innate defenses to the complex acquired responses discussed in this section (24). There is great interest in elucidating how the microbiota can influence immune homeostasis inside and outside the gut, as this process has important implications for the pathogenesis and treatment of inflammatory disorders and a growing list of diseases linked to inflammation.

The role of the microbiota in specific diseases and conditions

The above sections have described some of the many ways that the microbiota can influence human physiology, and it is no surprise that there is great interest in studying microbiota changes associated with diseased states, often referred to as dysbiosis (Table 1). However, the relationship between dysbiosis and disease pathogenesis is uncertain in the majority of examples at this time. It is often not clear what microbiota changes associated with disease are meaningful and distinguishing between cause and effect is inherently challenging. While it is intriguing to speculate that dysbiosis may cause disease as we learn more about how the microbiota can influence the host, it is also noted that the diseased state can lead to changes to the microbiota through various mechanisms, including changes in eating habits and bowel function as well as through the addition of medications such as antibiotics. In this section, we highlight a few of the recent findings on the role of the microbiota in particular diseases or conditions, but we cannot touch on all of the emerging findings in a multitude of other diseases both inside and outside the gut, including but not limited to rheumatoid arthritis (25), colorectal cancer (26), obesity (27), and diabetes (28).

Cardiovascular disease

There is growing interest in a link between microbiota and cardiovascular disease based on data showing microbial metabolism of dietary phosphatidylcholine into the proatherosclerotic metabolite trimethylamine-N-oxide (TMAO) (29). A recent study of healthy patients challenged with dietary phosphatidylcholine showed increased plasma levels of TMAO that were suppressed by prior treatment with antibiotics. They also found that plasma TMAO levels were associated with increased risk for cardiovascular events in patients with cardiovascular disease risk factors (*30). In another study, the same group showed that healthy human volunteers who maintained a vegan diet, as opposed to those on omnivorous diets, did not demonstrate increased plasma TMAO levels after dietary phosphatidylcholine challenge, and this trait was associated with specific fecal microbiota compositional states (*31). Thus, there is much interest in this microbiota-dependent pathway that may offer diagnostic and therapeutic potential for cardiovascular disease.

Irritable bowel disease and the microbiota-gut-brain axis

A role for the microbiota in irritable bowel syndrome (IBS) is suspected, though unproven, and therapies that alter the microbiota, including dietary changes, probiotics, and antibiotics, have shown encouraging, though inconsistent, results (32). In two reports describing the results of dietary intervention with a low FODMAP (fermentable oligosaccharides, disaccharides, monosaccharides and polyols) diet consisting of restricted intake of certain fermentable substrates compared to a typical Australian diet in a small number of Australian patients with IBS, the low FODMAP diet improved symptoms and resulted in changes in the gut microbiota, including reductions in putatively healthy bacteria, such as those in butyrate-producing Clostridium cluster XIVa (33, *34). One proposed pathway involved in IBS is through a microbiota-gut-brain axis, linking changes in the gut to symptom perception in the central nervous system. An interesting recent report demonstrated that the intake of a probiotic-rich fermented milk product resulted in alterations in brain activity in response to visual emotional stimuli as measured by functional magnetic resonance imaging as compared to the intake of a control product (35). The study of IBS is challenging due to the lack of specific diagnostic tests and the possibility of heterogeneous etiologies. There may be a subset of patients where microbiota changes are particularly important and for whom therapies to affect microbiota composition and function may be beneficial.

Clostridium difficile infection

Clostridium difficile infection (CDI) is the prime example of a human disease that develops as a result of critical changes to the gut microbiota and is effectively treated by microbiota-based therapy (36). A meta-analysis to review the use of fecal microbiota transplant (FMT) for prevention of recurrent CDI identified 11 studies with 273 patients through 2012; the overall resulting efficacy was about 90% and no substantial FMT-related adverse events were reported (37). In a prospective clinical trial, patients with recurrent CDI were randomly assigned to one of three treatment groups, standard vancomycin therapy, vancomycin therapy followed by bowel lavage, and vancomycin therapy followed by bowel lavage and subsequent infusion of donor stool into the duodenum (**38). The study was stopped early after interim analysis due to the superiority of FMT. Analysis of the fecal microbiota of patients who were treated with FMT for recurrent CDI demonstrated that the post-transplant microbiota of the recipients becomes more similar to that of the donor. After FMT the recipient microbiota was characterized by increased diversity, increased abundance of various Firmicutes and Bacteroidetes, and decreased abundance of Proteobacteria (*39). In a proof-of-principle study, two patients with CDI refractory to antibiotics were successfully treated with a stool substitute consisting of 33 strains of bacteria isolated and cultured from a healthy donor (*40). In this study, the stool substitute was delivered with colonoscopy, but it does provide promise for using a selected population of bacteria that could be prepared in a laboratory. Collectively, these studies strongly support the treatment of recurrent CDI with microbiota-based therapies.

While susceptibility to CDI after antibiotic use is associated with decreased microbiota diversity, little is known about the functional difference in the microenvironment that permits CDI. In preclinical studies, the microbiota changes that occurred after antibiotic treatment in mice susceptible to C. difficile were accompanied by changes in the metabolome that supported C. difficile germination and growth (*41). In the antibiotic-treated mice, primary bile acids that support C. difficile germination and certain carbohydrates that support C. difficile growth were present at significantly increased levels compared to control C. difficile-resistant mice. This study, and others, are helping to unravel the mechanism by which antibiotic-induced changes to the microbiota contribute to CDI, potentially leading to novel therapies for this burdensome disease.

• None

Disclosure of funding: This work was funded by the NIH: DK094775 (ABS), DK087708 (JYK), DK083993 (VBY) and AI090871 (VBY).