Translating the microbiome – what’s the target?

Two celebrated scientific movements, precision medicine and microbiome science, promise to converge on clinical medicine. Both are relevant to a diversity of clinical disciplines and capture the attention of increasingly expectant patients seeking individualized treatments based on exquisitely precise diagnostics. Clinicians responding to requests for a microbiome test often have no understanding of how such a test might be interpreted. Other than known infectious agents, what can a doctor find in a single isolated fecal microbiome analysis that is therapeutically actionable? The most successful translation of microbiome science to clinical medicine has been in the setting of specific identifiable microbial targets, such as infection with Helicobacter pylori and overgrowth of Clostridioides difficile. Despite this, microbiome science is attracting huge commercial interest, with hundreds of start-up companies exploring microbiome-based therapies and diagnostics—already with some spectacular failures. Not until 2022 was the first microbiome-based therapeutic product approved by the US Food and Drug Administration; this is a refinement of an old treatment for recurrent C difficile and a reminder that microbiome science is still catching up with established clinical truths. The most important obstacle to clinical translation of microbiome science is the unresolved question: Are microbiome alterations in human diseases causative or consequential and/or contributory to disease modification? In this Commentary, we summarize evidence favoring a consequential role for the reported microbiome alterations in many chronic human disorders and propose that the focus of clinical microbiome research should shift from blind pursuit of microbiome modulation toward identifying specific microbes and their metabolites influencing the host response and disease severity. A significant impediment to establishing the equivalent of Koch’s postulates for microbiome-related disease is the reliability of animal models for representing complex, multifactorial disease; despite this challenge, a 2020 review found that 95% of published studies using human to rodent microbiome transfer reported a transfer of pathologic symptoms.1Walter J. Armet A.M. Finlay B.B. et al.Establishing or exaggerating causality for the gut microbiome: lessons from human microbiota-associated rodents.Cell. 2020; 180: 221-232Abstract Full Text Full Text PDF PubMed Scopus (241) Google Scholar This figure is implausible against a backdrop of different physiology and different native microbiomes in humans and rodents. The challenge is particularly acute for establishing models for human behavioral and cognitive functions, for which gut–brain links have been proposed. New statistical approaches will improve the interpretation of existing cohort data2Corander J. Hanage W.P. Pensar J. Causal discovery for the microbiome.Lancet Microbe. 2022; 3: e881-e887Abstract Full Text Full Text PDF PubMed Scopus (4) Google Scholar but cannot overcome weak or absent evidence for cause and effect. Interpreting variance of the microbiome in human disease is problematic because a normal or healthy microbiome has not been defined,3Shanahan F. Ghosh T.S. O'Toole P.W. The healthy microbiome—what is the definition of a healthy gut microbiome?.Gastroenterology. 2021; 160: 483-494Abstract Full Text Full Text PDF PubMed Scopus (116) Google Scholar most of the variance have not been explained, and confounders and the clinical significance of interindividual microbiome variation are not adequately understood. Microbiomes differ with the age and lifestyle of the host and vary considerably between different ethnic groups, geographic regions, and even different individuals in the same community.4Claesson M.J. Jeffery I.B. Conde S. et al.Gut microbiota composition correlates with diet and health in the elderly.Nature. 2012; 488: 178-185Crossref PubMed Scopus (2235) Google Scholar Moreover, most of our understanding of the human microbiome is based on studies of affluent people living in industrialized countries.3Shanahan F. Ghosh T.S. O'Toole P.W. The healthy microbiome—what is the definition of a healthy gut microbiome?.Gastroenterology. 2021; 160: 483-494Abstract Full Text Full Text PDF PubMed Scopus (116) Google Scholar Not surprisingly, for most diseases where a disturbed microbiome has been implicated, no mechanisms have been conclusively identified. Studies in human disorders need to offer more incisive evidence than the loss of microbial diversity that has been described for many disorders in which the microbiome has been implicated. Diversity analysis alone is an inadequate, nonactionable, misleading, and simplistic metric when it is agnostic of the microbial taxa that are gained or lost. Of greater concern is that few disease-specific microbial signatures have been consistently identified. Instead, commonalities or overlapping microbial changes have been found repeatedly, especially the loss of commensal species associated with healthy subjects and the outgrowth of a general subset of disease-associated commensals termed “pathobionts.”3Shanahan F. Ghosh T.S. O'Toole P.W. The healthy microbiome—what is the definition of a healthy gut microbiome?.Gastroenterology. 2021; 160: 483-494Abstract Full Text Full Text PDF PubMed Scopus (116) Google Scholar Duvallet et al5Duvallet C. Gibbons S.M. Gurry T. et al.Meta-analysis of gut microbiome studies identifies disease-specific and shared responses.Nat Commun. 2017; 8: 1784Crossref PubMed Scopus (509) Google Scholar studied microbiome changes across 28 case-control gut microbiome studies spanning 10 diseases and noted that half the microbiome changes were shared across the diseases studied, and indeed there is sufficient commonality across diseases to derive a predictive index for health status based on such shared taxa6Gupta V.K. Kim M. Bakshi U. et al.A predictive index for health status using species-level gut microbiome profiling.Nat Commun. 2020; 11: 4635Crossref PubMed Scopus (74) Google Scholar and to diagnose multiple diseases.7Su Q. Liu Q. Lau R.I. et al.Faecal microbiome-based machine learning for multi-class disease diagnosis.Nat Commun. 2022; 13: 6818Crossref PubMed Scopus (7) Google Scholar Many of these taxa are also those we reported as more abundant in unhealthy aging.8Ghosh T.S. Das M. Jeffery I.B. et al.Adjusting for age improves identification of gut microbiome alterations in multiple diseases.eLife. 2020; 9 (e50240)Crossref Scopus (82) Google Scholar Some of the commensals repressed in multiple diseases overlap with (known or suspected) anti-inflammatory microbes first proposed as being depleted in inflammatory bowel disease, whereas pathobionts are often associated with inflammation in multiple diseases.9Chow J. Tang H. Mazmanian S.K. Pathobionts of the gastrointestinal microbiota and inflammatory disease.Curr Opin Immunol. 2011; 23: 473-480Crossref PubMed Scopus (295) Google Scholar This raises the likelihood that these microbiome changes are consequential rather than causative, but it is not possible to be definitive because of the observational study design and the fact that most studies report relative abundance rather than absolute. However, some microbiome changes, such as a bloom of Proteobacteria commonly found in inflammatory bowel disease and other disorders, are indeed consequential and predictable because of increases in mucosal oxygen tension in the presence of inflammation.10Zeng M.Y. Inohara N. Nunez G. Mechanisms of inflammation-driven bacterial dysbiosis in the gut.Mucosal Immunol. 2017; 10: 18-26Abstract Full Text Full Text PDF PubMed Scopus (421) Google Scholar Moreover, comparative analyses of paired biopsies from inflamed and noninflamed colonic mucosa and longitudinal studies during relapse and remission provided no circumstantial support for disease causality of the microbiome.11Clooney A.G. Eckenberger J. Laserna-Mendieta E. et al.Ranking microbiome variance in inflammatory bowel disease: a large longitudinal intercontinental study.Gut. 2021; 70: 499-510Crossref PubMed Scopus (86) Google Scholar In other chronic conditions where the microbiome has been implicated, the microbial disturbances have since been shown at least in part to be due to drug treatment, as in type 2 diabetes,12Forslund K. Hildebrand F. Nielsen T. et al.Disentangling type 2 diabetes and metformin treatment signatures in the human gut microbiota.Nature. 2015; 528: 262-266Crossref PubMed Scopus (1329) Google Scholar or, in the case of autism, to result from dietary preferences.13Yap C.X. Henders A.K. Alvares G.A. et al.Autism-related dietary preferences mediate autism-gut microbiome associations.Cell. 2021; 184: 5916-5931Abstract Full Text Full Text PDF PubMed Scopus (95) Google Scholar Another feature confounding the interpretation of apparent microbiome abnormalities in several human diseases is the translocation of oral microbes to the distal gut. This is particularly evident in the inflammatory bowel diseases Crohn’s disease and ulcerative colitis and in sporadic colorectal cancer and liver disease.14Read E. Curtis M.A. Neves J.F. The role of oral bacteria in inflammatory bowel disease.Nat Rev Gastroenterol Hepatol. 2021; 18: 731-742Crossref PubMed Scopus (49) Google Scholar Oral microbes are detected in the gut more frequently in people with gastrointestinal disease and cancer than in healthy subjects. As early as 2016, Pasolli et al15Pasolli E. Truong D.T. Malik F. et al.Machine Learning meta-analysis of large metagenomic datasets: tools and biological insights.PLoS Comput Biol. 2016; 12e1004977Crossref PubMed Scopus (299) Google Scholar (including one of us, Nicola Segata) noted from a study of over 2000 gut metagenomes that the presence of some oral microbes was a general feature of diseases involving dysbiosis including liver cirrhosis, where Veillonella spp, Streptococcus spp, and Haemophilus parainfluenzae were signature taxa. In addition, many of the disease-associated microbes identified by Duvallet et al,5Duvallet C. Gibbons S.M. Gurry T. et al.Meta-analysis of gut microbiome studies identifies disease-specific and shared responses.Nat Commun. 2017; 8: 1784Crossref PubMed Scopus (509) Google Scholar such as Porphyromonas and Fusobacterium, were likely of oral origin. Oral microbes, especially Fusobacterium nucleatum, were not only detected in fecal samples in large colon cancer meta-analyses16Wirbel J. Pyl P.T. Kartal E. et al.Meta-analysis of fecal metagenomes reveals global microbial signatures that are specific for colorectal cancer.Nat Med. 2019; 25: 679-689Crossref PubMed Scopus (485) Google Scholar but are consistently found in colon cancer biopsies, often as polymicrobial biofilm-forming communities. Oral microbes are more abundant in the fecal microbiota of patients with ulcerative colitis than healthy control subjects, and ectopic integration of Klebsiella strains from patients with inflammatory bowel disease drove gut inflammation in a gnotobiotic model.17Atarashi K. Suda W. Luo C. et al.Ectopic colonization of oral bacteria in the intestine drives T(H)1 cell induction and inflammation.Science. 2017; 358: 359-365Crossref PubMed Scopus (450) Google Scholar However, although colonization of the gut by oral microbes was confirmed for patients with colorectal cancer and rheumatoid arthritis, it also occurred at high frequency in healthy control subjects.18Schmidt T.S. Hayward M.R. Coelho L.P. et al.Extensive transmission of microbes along the gastrointestinal tract.eLife. 2019; 8 (e42693)Crossref Scopus (222) Google Scholar Of note, oral microbes are highly transmissible between individuals sharing the same household, with an almost 40% median strain-sharing rate between adult partners living together for several years against close to a rate of 0% between unrelated individuals.19Valles-Colomer M. Blanco-Miguez A. Manghi P. et al.The person-to-person transmission landscape of the gut and oral microbiomes.Nature. 2023; 614: 125-135Crossref PubMed Scopus (21) Google Scholar This transmission rate is, as expected, higher than that of the gut microbiomes and could mean that the likelihood of acquiring oral microbes with a role in gut inflammation is partially dependent on social and direct interactions. Increasing availability of large metagenomic cohorts is further fueling the possibility of inferring reproducible links between the presence of oral microbes in the gut and host conditions. In an analysis of 30 metagenomic datasets (4646 total samples) spanning 15 distinct host diseases, an increased abundance of oral microbes in the gut was observed in all but 1 dataset and proved significant in half of them with a very significant overall meta-analysis association (Manghi et al, unpublished data). With respect to aging, in our 2020 study of 2500 subjects,8Ghosh T.S. Das M. Jeffery I.B. et al.Adjusting for age improves identification of gut microbiome alterations in multiple diseases.eLife. 2020; 9 (e50240)Crossref Scopus (82) Google Scholar approximately half of the taxa associated with dysbiosis-type diseases in young and middle-aged subjects were classifiable as oral microbes, but only 1 of 9 taxa that were more abundant in older subjects with disease, namely Streptococcus parasanguinis, was an oral taxon. In a later meta-analysis,20Ghosh T.S. Shanahan F. O'Toole P.W. The gut microbiome as a modulator of healthy ageing.Nat Rev Gastroenterol Hepatol. 2022; 19: 565-584Crossref PubMed Scopus (64) Google Scholar we identified 19 taxa associated with unhealthy aging, 6 of which were clearly oral microbes (Streptococcus sp, Actinomyces sp, Desulfovibrio, Campylobacter, Atopobiaceae, and Veillonella). In our most recent study of 21,000 microbiome datasets,21Ghosh T.S. Shanahan F. O'Toole P.W. Toward an improved definition of a healthy microbiome for healthy aging.Nature Aging. 2022; 2: 1054-1069Crossref PubMed Scopus (8) Google Scholar a newly defined index, Kendall Uniqueness, showed the strongest association (negatively, for this metric) with retention of health during aging. Twelve of 22 taxa that associated positively with Kendall Uniqueness (indicating unhealthy aging) were oral microbes, with the remaining taxa being clearly intestinal microbes. Interestingly, 8 of these oral microbes formed a separate co-abundance network (see Figure 5 in Ghosh et al21Ghosh T.S. Shanahan F. O'Toole P.W. Toward an improved definition of a healthy microbiome for healthy aging.Nature Aging. 2022; 2: 1054-1069Crossref PubMed Scopus (8) Google Scholar), possibly reflecting the tendency of oral microbes to form mixed biofilms. Although there is a selection bias in published studies because young adults that are sick are less likely to survive and be surveyed as older subjects, we conclude that introgression of oral microbes is indicative of, but not sufficient for, loss of health in aging. There is a robust case for using microbiome analysis for disease detection or diagnosis such as for colon cancer,22Liang J.Q. Li T. Nakatsu G. et al.A novel faecal Lachnoclostridium marker for the non-invasive diagnosis of colorectal adenoma and cancer.Gut. 2020; 69: 1248-1257Crossref PubMed Scopus (127) Google Scholar potentially autism23Wan Y. Zuo T. Xu Z. et al.Underdevelopment of the gut microbiota and bacteria species as non-invasive markers of prediction in children with autism spectrum disorder.Gut. 2022; 71: 910-918Crossref PubMed Scopus (38) Google Scholar recognizing dietary influences discussed above,13Yap C.X. Henders A.K. Alvares G.A. et al.Autism-related dietary preferences mediate autism-gut microbiome associations.Cell. 2021; 184: 5916-5931Abstract Full Text Full Text PDF PubMed Scopus (95) Google Scholar and other diseases.7Su Q. Liu Q. Lau R.I. et al.Faecal microbiome-based machine learning for multi-class disease diagnosis.Nat Commun. 2022; 13: 6818Crossref PubMed Scopus (7) Google Scholar However, the quest for a causal link with the microbiome for many chronic human disorders is undermined by the common occurrence of nonspecific, overlapping, and expected microbiome disturbances. Targeting secondary microbiome changes is unlikely to yield sustained disease remission, but significant therapeutic improvement can still be achieved by addressing specific disturbances of the microbiota that exacerbate disease severity. An insightful example of this strategy is the recent clarification of the role of the microbiome in the inherited human disorder, familial dysautonomia. This monogenic disorder is characterized by progressive degeneration of the peripheral and central nervous systems with deficits of innervation in many organs, including the gastrointestinal tract. This leads to disturbances of the microbiome and metabolome, which likely exacerbate the neurodegeneration.24Cheney A.M. Costello S.M. Pinkham N.V. et al.Gut microbiome dysbiosis drives metabolic dysfunction in familial dysautonomia.Nat Commun. 2023; 14: 218Crossref PubMed Scopus (4) Google Scholar Similarly, with more common and more complex conditions such as chronic kidney disease, there is a role for the gut microbiota modulating disease severity, regardless of the cause.25Pluznick J.L. The gut microbiota in kidney disease.Science. 2020; 369: 1426-1467Crossref PubMed Google Scholar Improved responses to existing treatments and dietary regimes may also be achievable with microbiome modulation as shown by encouraging evidence linking the microbiome with responsiveness to cancer immunotherapy. It is also possible that part of the missing causal links can be unraveled by multiomic studies or by mechanistic insights that have to be explored at the level of profiling the functional and metabolic capabilities of individual strains, but these tasks are limited by challenges related to high-dimensional statistics and stubbornly prohibitive costs of high-depth metagenome sequencing, especially for research groups in low- or middle-income countries. Thus, overall, the promise of microbiome-based therapeutics is alive, but more likely to be fulfilled with identification of specific disease-modifying microbial metabolites.