The Micro­biome

For most of the history of medicine, microorganisms were enemies. The germ theory of disease, confirmed beyond doubt by Louis Pasteur and Robert Koch in the second half of the 19th century, established that specific microorganisms caused specific diseases, and the subsequent century of medical progress in antibiotics, vaccines, and sanitation was built on the premise that the relationship between humans and microbes was fundamentally adversarial. Microbes made people sick. The goal was to eliminate them.

This picture was not wrong. It was incomplete in a way that took most of the 20th century to fully appreciate. The same human body that is killed by certain bacteria is inhabited, continuously and necessarily, by approximately 38 trillion microbial cells: bacteria, archaea, fungi, viruses, and single-celled eukaryotes living in the gut, on the skin, in the respiratory tract, in the urogenital system, and in essentially every body cavity that communicates with the external environment. These organisms are not invaders that the immune system has failed to clear. They are residents, most of them permanent, many of them essential, collectively encoded by a genetic repertoire approximately 150 times larger than the human genome itself.

The microbiome is the collective term for these microbial communities and their genomes. The microbiota refers specifically to the organisms. A human being is not an organism that occasionally encounters microbes. A human being is an ecosystem in which microbial life is an integral and functional component, one that has co-evolved with the host for millions of years and whose disruption has measurable and sometimes severe consequences for health.

The figure of 38 trillion microbial cells in the human body, compared to approximately 37 trillion human cells, was established by Ron Sender, Shai Fuchs, and Ron Milo at the Weizmann Institute in a 2016 paper that corrected a widely cited earlier estimate of a 10:1 microbial to human cell ratio. The revised figure of approximately 1.3:1 is more accurate. An organism whose microbial cells approximately equal its human cells in number is a fundamentally different kind of entity than the classical image of a human body as a singular biological system.

The vast majority of the microbiota, in terms of cell numbers, is concentrated in the large intestine (colon), where microbial densities reach approximately 10¹¹ cells per millilitre of gut content, among the highest microbial cell densities in any environment on Earth. The stomach and small intestine have far lower densities because of the acidic and antimicrobial conditions there. The skin carries approximately 10¹² microbial cells in total, distributed across its surface. The oral cavity harbours approximately 700 bacterial species in various niches, each niche supporting its own distinct community.

The microbial composition of the gut is dominated by two bacterial phyla: Firmicutes and Bacteroidetes, which together account for approximately 90 percent of the gut microbiota in most adults. No two individuals have the same gut microbiota composition. Even identical twins, who share their genome and typically their early environment, show substantial differences in their gut microbial communities by adulthood. The microbiome is more individually variable than the genome.

The Human Microbiome Project, launched by the National Institutes of Health in 2007 and producing its first major results in 2012, characterised the microbiomes of 242 healthy adults across 18 body sites using 16S rRNA gene sequencing and shotgun metagenomic sequencing. A central finding was that microbial communities at different body sites are highly site-specific and more similar between individuals at the same site than between different sites within the same individual. The gut microbiome of any healthy adult resembles the gut microbiome of another healthy adult more than it resembles that person's own skin microbiome.

The commensal microbiota performs functions that the human organism cannot perform without them. These are not marginal or supplementary functions. They include capabilities that are central to metabolism, immune development, and protection against pathogens.

The gut microbiota ferments dietary fibre, polysaccharides from plant cell walls that the human genome encodes no enzymes to digest, into short-chain fatty acids (SCFAs), principally butyrate, propionate, and acetate. Butyrate is the primary energy source for the colonocytes, the epithelial cells lining the colon. Without butyrate from microbial fermentation, colonocytes are deprived of their main fuel. Propionate is transported to the liver and contributes to hepatic metabolism. Acetate circulates systemically and reaches the brain. The gut microbiota also synthesises vitamins the human genome cannot produce in sufficient quantities, including vitamin K and several B vitamins, and modifies bile acids secreted by the liver in ways that affect lipid absorption and glucose metabolism. Approximately 10 percent of an individual's daily caloric intake from complex carbohydrates is extracted not by human enzymes but by microbial fermentation.

The immune system does not develop properly in the absence of microbial colonisation. Germ-free animals raised under sterile conditions from birth develop severely abnormal immune systems: reduced Peyer's patches, fewer intestinal IgA-producing B cells, impaired regulatory T cell populations, and an inflammatory bias that predisposes them to exaggerated responses to antigens they should tolerate. The microbiota provides the antigenic stimulation and molecular signals required to calibrate the immune system appropriately during early development. Some of this calibration window appears to be narrow: microbial exposure in the first days and weeks of life has effects on immune development that cannot be fully compensated by later colonisation.

A dense, diverse, established commensal microbiota physically and chemically resists colonisation by pathogens, by competing for nutrients and attachment sites, by producing antimicrobial substances called bacteriocins, and by stimulating immune responses that hold potential pathogens below pathogenic thresholds. This is colonisation resistance, and its importance becomes starkly apparent when it is disrupted. Broad-spectrum antibiotic treatment creates conditions in which Clostridioides difficile, an opportunistic pathogen normally held below dangerous levels by the commensal community, can expand catastrophically, causing severe and potentially fatal colitis. C. difficile infection is almost exclusively a disease of antibiotic-treated individuals.

The composition of the gut microbiota is not static. It changes with diet, with antibiotic use, with illness, with age, and with a range of other factors. When the composition shifts in ways that impair function, reduce diversity, or allow potentially harmful organisms to become disproportionately abundant, the state is called dysbiosis. The evidence linking dysbiosis to a growing catalogue of human diseases has accumulated rapidly since metagenomic sequencing made systematic microbiome characterisation possible.

The most robustly established link between the gut microbiota and disease outside the gut involves the metabolic syndrome: the cluster of obesity, insulin resistance, dyslipidaemia, and hypertension that constitutes the primary driver of type 2 diabetes and cardiovascular disease in developed populations. Germ-free mice are protected against diet-induced obesity. Jeffrey Gordon at Washington University in St Louis performed the definitive experiments in this area, most strikingly in a 2013 paper in Science demonstrating that human microbiota from obese and lean identical twins, when transplanted into germ-free mice, transmitted phenotypic differences in fat deposition: mice receiving the obese-donor microbiota gained more fat than mice receiving the lean-donor microbiota even when caloric intake was identical. The microbiota was a causal factor in the obese phenotype, not merely a correlate.

Inflammatory bowel disease (IBD), comprising Crohn's disease and ulcerative colitis, is characterised by chronic inflammation of the gut wall and is associated with consistent alterations in gut microbiota composition: reduced diversity, reduced abundance of butyrate-producing bacteria, and elevated abundance of potentially inflammatory species. Whether dysbiosis causes IBD or results from the inflammatory environment is a question of active investigation, and the answer is likely to be a bidirectional interaction rather than a simple causal chain.

Faecal microbiota transplantation (FMT), the transfer of the faecal microbiota from a healthy donor into the gut of a recipient, was first described in medical literature in 1958 by Ben Eiseman for the treatment of severe pseudomembranous colitis. Its modern clinical application for recurrent C. difficile infection achieves cure rates of approximately 90 percent in patients who have failed multiple courses of antibiotics, making it one of the most effective interventions for this indication. FMT works by restoring colonisation resistance: reintroducing a diverse commensal community that can outcompete C. difficile and prevent its re-expansion.

The most surprising frontier in microbiome research is the relationship between the gut microbiota and the brain. The two organs are separated by the blood-brain barrier, which excludes most bacteria and their metabolites from direct access to the central nervous system. Yet an accumulating body of evidence from animal models, and a smaller but growing body of evidence from human studies, indicates that the gut microbiota influences brain function, behaviour, mood, stress responses, and potentially the risk of neurological and psychiatric conditions.

The anatomical substrate for gut-brain communication is the enteric nervous system: a network of approximately 500 million neurons embedded in the walls of the gastrointestinal tract, more neurons than are present in the spinal cord, which is why the gut is sometimes called the second brain. The enteric nervous system communicates with the central nervous system via the vagus nerve, which carries signals in both directions between the gut and the brainstem. Approximately 80 to 90 percent of the fibres in the vagus nerve carry information upward, from the gut to the brain, rather than downward. The gut is predominantly sending signals to the brain, not receiving them.

The gut microbiota communicates with the enteric nervous system and with the vagus nerve through multiple chemical channels. Approximately 95 percent of the body's total serotonin is produced not in the brain but in the gut, by enterochromaffin cells lining the gut wall, whose serotonin production is regulated by the microbiota. Certain gut bacteria produce gamma-aminobutyric acid (GABA), the primary inhibitory neurotransmitter in the brain. The SCFAs produced by microbial fermentation of dietary fibre, particularly butyrate, have direct effects on gene expression in the brain through their activity as histone deacetylase inhibitors. The gut microbiota also regulates the hypothalamic-pituitary-adrenal (HPA) axis, the hormonal system governing the stress response.

The evidence from germ-free animal studies is striking. Germ-free mice show exaggerated stress responses, elevated corticosterone levels in response to stress compared to conventionally raised mice, and alterations in the expression of genes involved in neurotransmitter metabolism in the hippocampus and hypothalamus. John Cryan and Ted Dinan at University College Cork, who coined the term psychobiotics for microorganisms with mental health benefits, have reviewed evidence suggesting that specific bacterial strains, particularly certain Lactobacillus and Bifidobacterium species, can reduce anxiety-like behaviour in animal models through vagus-nerve-dependent mechanisms.

The translation of these findings to humans remains an active and sometimes contested research area. The most rigorous studies suggest modest but real effects of the gut microbiota on mood, anxiety, and stress reactivity in healthy humans. What is not in doubt is that the gut and the brain are in continuous bidirectional communication, and that the microbial community in the gut is an active participant in that conversation.

Infants born vaginally are colonised first by the vaginal and faecal microbiota of their mother, acquiring bacteria including Lactobacillus species, Bifidobacterium species, and a diversity of gut-associated bacteria. Infants born by Caesarean section are colonised instead primarily by skin and environmental bacteria, including Staphylococcus species. The difference in initial colonisation between vaginal and Caesarean-section births is detectable in microbiome composition for months to years after birth. Epidemiological data show consistently elevated rates of immune-mediated conditions including asthma, allergies, and type 1 diabetes in Caesarean-born children, associations that may reflect the early microbiome difference but cannot yet be attributed to it causally.

Breastfeeding is the most important determinant of the infant microbiome after birth. Human breast milk contains not only nutrients and antibodies but also human milk oligosaccharides (HMOs), complex carbohydrates that human infants cannot digest but that selectively promote the growth of specific beneficial bacteria, particularly Bifidobacterium infantis, which has the metabolic machinery to consume HMOs that most other bacteria lack. Bruce German and David Mills at the University of California Davis have shown that breast milk is partially a prebiotic: a food designed not for the infant but for the bacteria the infant needs to establish. HMOs constitute the third most abundant solid component of breast milk after lactose and fat. They have no direct nutritional value to the infant. They are nutrients for B. infantis.

The first 1,000 days of life, from conception to approximately age 2, represent the critical window for microbiome establishment and immune system calibration. Antibiotic use during this window has been associated with elevated risk of obesity, allergies, and asthma in a dose-dependent manner across multiple longitudinal studies. The biological plausibility is strong: antibiotics during this period disrupt the establishing microbial community during the developmental window in which the immune system is being calibrated against commensal microorganisms.

The skin microbiome is organised by microenvironment. Sebaceous areas of the skin are dominated by Cutibacterium acnes, which metabolises the sebum secreted by sebaceous glands. Moist areas are dominated by Staphylococcus and Corynebacterium species. Dry areas carry a lower diversity and lower density microbial community. Julia Segre and Elizabeth Grice at the National Human Genome Research Institute produced the first comprehensive characterisation of skin microbiome biogeography in 2009, demonstrating that the ecological zones of the skin are as ecologically distinct from each other as tropical rainforest is from arctic tundra.

The commensal skin microbiota protects against pathogens by occupying ecological niches that potential pathogens would otherwise colonise, by producing antimicrobial molecules, and by training the local immune system. Staphylococcus epidermidis, one of the most abundant commensal skin bacteria, produces molecules that inhibit the formation of biofilms by Staphylococcus aureus, a major pathogen. It also educates skin-resident T cells to be tolerant of the commensal while remaining responsive to the pathogen. The commensal and the pathogen are closely related species sharing a genus. The immune system maintains a distinction between them that is molecularly precise.

The oral microbiome of approximately 700 species forms a community whose pathological disruption is among the most prevalent diseases in the world. Dental caries is caused by acid produced by specific bacteria, particularly Streptococcus mutans, fermenting dietary sugars. The organism is not inherently pathogenic. It performs the same fermentation chemistry it always has. The pathological outcome depends on diet: in the absence of fermentable sugars, S. mutans causes no caries. The disease is the product of an interaction between a specific organism and a specific environmental condition, with the host as the terrain in which both operate.

The development of antibiotics across the 20th century was among the most consequential achievements in the history of medicine. Alexander Fleming's observation in 1928 that the mould Penicillium notatum produced a substance that inhibited bacterial growth, and the subsequent development of penicillin as a clinical agent by Howard Florey and Ernst Chain, inaugurated an era in which infections that had been reliably fatal became reliably treatable. The death toll from bacterial infections in developed countries plummeted.

The cost of that weapon was not understood until the tools to measure it existed. Broad-spectrum antibiotics do not discriminate between pathogens and commensals. A course of ciprofloxacin or ampicillin eliminates a substantial fraction of the gut microbiota alongside whatever pathogen it is targeting, reducing microbial diversity for weeks to months. In most healthy adults, the microbiota recovers to something approximating its pre-treatment composition after several months. In some individuals recovery is incomplete, with lasting reductions in diversity and changes in community composition.

The global antibiotic resistance crisis is a direct consequence of the evolutionary principles discussed in Artifact 3. Antibiotic resistance genes arise by mutation or horizontal gene transfer, and antibiotic use creates precisely the selection pressure that drives resistance genes to high frequency. The World Health Organisation has designated antimicrobial resistance as one of the ten greatest threats to global health. The gut microbiome of an individual who has recently received antibiotics is temporarily enriched in antibiotic-resistant bacteria, a consequence of the selection pressure that clears susceptible organisms while leaving resistant ones to expand.

The microbiome is not only bacteria. Every human gut also contains a virome: a community of viruses, the majority of which are bacteriophages, viruses that infect bacteria rather than human cells. The gut virome is estimated to contain approximately 10¹⁵ virus particles, outnumbering bacterial cells in the gut by up to 10 to 1. These bacteriophages play a crucial role in regulating bacterial population sizes within the microbiota: by infecting and lysing specific bacterial hosts, phages act as a selective pressure that prevents any single bacterial species from dominating the community. The phage community and the bacterial community co-evolve in a continuous arms race, driving microbial diversity through an ongoing process of ecological pressure.

The mycobiome, the fungal component of the microbiome, is far less studied than the bacterial component but is not trivial. Candida species are consistently present in the gut of most humans, normally at low abundance held in check by bacterial commensals and by the immune system. Antibiotic treatment, which reduces bacterial competitors, often allows Candida populations to expand, sometimes to levels that cause clinical disease. The relationship between the bacterial and fungal communities of the gut is an active competitive dynamic, not a stable coexistence.

The full ecological complexity of the human microbiome is a scientific frontier whose depth is just becoming apparent. The majority of microbial species identified in the human gut by metagenomic sequencing have never been cultured in the laboratory. Their metabolic capabilities, their roles in the community, and their specific effects on the host are known, if at all, only from sequence data. A significant fraction of the gut microbiome remains, in the most literal sense, unknown.

The rapid increase in the prevalence of a cluster of conditions in developed countries over the past half century, a period too short for genetic explanations, has focused attention on environmental changes as drivers. The conditions include allergies, asthma, inflammatory bowel disease, type 1 diabetes, multiple sclerosis, and obesity, all of which involve dysregulation of immune function or metabolic regulation and all of which have increased substantially in the past 50 to 100 years.

The microbiome hypothesis for this cluster of conditions holds that modern practices including widespread antibiotic use, Caesarean delivery, formula feeding, and processed diets low in fermentable fibre have disrupted the human microbiome in ways that impair immune calibration and metabolic regulation. The supporting evidence is substantial in animal models and correlational in humans.

Graham Rook's old friends hypothesis specifically proposes that the organisms whose loss is most consequential are not general environmental microbes but specific groups, including helminths (parasitic worms), mycobacteria from soil, and Lactobacillus species from fermented foods and the birth canal, that co-evolved with humans over millions of years and played specific roles in training the regulatory circuits of the immune system. These old friends are not pathogens in any normal sense. They are organisms that the immune system evolved to tolerate and that, in tolerating, helped calibrate itself. Their disappearance from the modern environment may have removed a regulatory input that the immune system evolved to depend on.

The microbiome redraws the boundary of the self in ways that have no precedent in the classical picture of the human organism.

The classical picture is of a body with clear boundaries: inside is self, outside is not. The skin and gut mucosa are the borders. Everything outside is the environment. Everything inside is the organism. The immune system enforces this boundary. The microbiome reveals that the boundary was always more complicated than this picture suggested.

The 38 trillion microbial cells in the gut are inside the boundary in a spatial sense but are not self in any genetic sense. The immune system tolerates them not because it cannot detect them but because it has evolved elaborate mechanisms to identify them as residents rather than invaders. This tolerance is actively maintained and conditional: it can break down, as it does in inflammatory bowel disease.

The holobiont concept, proposed by Lynn Margulis and developed by Scott Gilbert and others, holds that the relevant unit of selection in ecology and evolution is not the host organism alone but the host together with its microbial community: a composite entity with composite genetics, composite metabolism, and composite ecology.

The self, at the biological level, is not a singular organism. It is an ecology. What looks from the outside like a human being is, from the molecular level inward, a negotiated cohabitation of human and microbial cells, each depending on the other, together constituting an entity that neither could be alone. The boundary between the human organism and the microbial world is not where anatomy places it. It is wherever ecology currently draws it. And ecology, unlike anatomy, does not hold still.