In sickness and in health: making the most of our microbial marriage

Executive summary

The microbiome has moved rapidly from academic anonymity to scientific celebrity. The term describes the genomic totality of microbial communities (bacteria, fungi, archaea and viruses) occupying every accessible nook and cranny of our being. People are fascinated by the previously unimagined breadth, intimacy and consequence of their molecular conversation with microbes. Propelled by the accelerating pace of laboratory discovery, the narrative is moving fast. If the research is correct, the microbiome, and particularly the bacterial microbiome, can profoundly influence many aspects of human physiology, from the immune system to the brain, from metabolism to drug response, in ways that can both promote health and engender disease. As the closeness of our microbial relationship and its systemic influence over health is revealed through scientific inquiry, a well-funded and highly competitive race has begun to leverage these discoveries and create microbiome-based consumer and medical products to improve well-being and treat disease.

The rise of the microbiome research insights and challenges

The human body provides ecological opportunity to the most opportunistic and adaptable of lifeforms and is considered a holobiont. The first microbial citizens transfer from mother to infant during birth, with the microbiota of naturally birthed infants differing from those delivered by C- section. These early colonisers, which notably include the bacterial genera Bifidobacterium and Lactobacillus, populate our digestive tract, and in some instances are selectively advantaged over co-located rival interlopers by oligosaccharides and IgA found in breast milk. Thus, a co- evolutionary paradigm is established wherein commensal microbes find habitats within, or on, their human host. They are tolerated immunologically, in exchange for functional favours that include food digestion, vitamin production, detoxification and the provision of a benign living barrier to potentially more threatening microbes.

Given the deep complexity of the human microbiome, which assembles through early childhood to typically several hundred species, its study requires a detailed genomic description of the microbial ecosystem, its functional activity and its connectivity to host physiology. Only when key enabling technologies became widely available, was microbiome science able to flourish.

The transformation came through access to culture-independent ‘omics technologies which proffered a big-data lens through which both the genetic composition and functional activity of complex microbial communities can be studied at resolutions not previously within reach. Initially, next generation sequencing of the 16S ribosomal RNA gene informed the descriptions of microbial taxonomic signatures commonly observed in healthy or diseased subjects. More latterly, and where budgets permit, shotgun metagenomic sequencing has become the preferred approach. From metagenomic data, not only are microbial taxonomies identified more fully and with more precision than 16S sequencing, but microbial genes are also broadly annotated, providing greater insight into the functional potential of the microbial community. This can be critically important given the functional variability seen across different strains of a single species. Metabolomic, transcriptomic and other ‘omics analyses can be additionally performed in parallel to DNA sequencing. An as yet limited number of studies provide early evidence that these complementary data sets may be integrated to form a deeper, more holistic representation of the functional activity of the microbiome and its influence over host biology.

However, despite this rapid progress, much needs to be done to ensure, as far as is possible, that hypotheses drawn from microbiomic data accurately speak to the true underlying biology. For example, while an individual’s microbiome is ordinarily quite taxonomically stable in adulthood, only around 30% of microbial composition is shared by individuals in different households (although functional commonality is considerably higher, estimated in the region of 70%). Therefore, to avoid bias and adequately accommodate inter-individual variability, microbiome studies need to be adequately powered. Furthermore, increasing use of standardised sample collection and extraction methodologies, inclusion of control samples, application of bioinformatic and biostatistical tools that manage the high-dimensionality, sparsity and compositionality of the data, are necessary. Together with more complete genomic and metabolomic reference databases, and more reliable disease models these improvements will help to strengthen the underlying discovery process and reduce variability between studies.

Improvements in methodological rigour notwithstanding, there remain significant challenges in distinguishing whether or not disease-associated microbiome motifs drive a pathologic process, or instead themselves result from a microbiome-independent pathology. In many instances, both aspects may be contained in the data and to varying degrees dependent upon the timing of sample acquisition.

Further, stool analysis is frequently used as a convenient proxy for the entire gut microbiome. However, the digestive tract is composed of many microenvironments along its longitudinal and cross-sectional axes. Each of these niches offer distinct chemical, nutrient, oxygen and other parameters (including host anti-microbials) that influence which bugs are found in any given anatomical site. Thus, the microbiome of the jejunum is much different to that of the descending colon. Equally, the microbiome of the colonic lumen is distinct from that of the adjacent epithelial mucosa. Stool can only poorly reflect these local, but physiologically important, spatial relationships between microbes and host.

Despite the clear challenges, mining ‘omics data has repeatedly identified key taxonomic and metabolite patterns that associate with good health and with various disease states (dysbioses). Increasingly, these insights, combined with improved microbial husbandry, have allowed hypotheses originally built around association to be more tightly aligned to causality, with accompanying mechanistic insights.

Microbiome and host physiology

Cohort and experimental data are provoking a more holistic view of human physiology, wherein the microbiome is recognised as a potent auxiliary metabolic organ with the ability to influence both proximal and distal tissues. Sitting within the digestive tract, the gut microbiome contributes to an already biologically dense ecosystem, one that is both well innervated and vascularized, and that accommodates an estimated 70% of total immunocytes organized within the gut-associated lymphoid tissue. Thus, the microbiome is tightly linked to both bidirectional communication between enteric and central nervous system (the so-called gut-brain axis) and to extensive interplay with host metabolism and immunity. It is this latter role that has gained most attention to date.

From a very early age, the assembling gut microbiota plays an important role in educating and modulating host immunity. Indeed, germ-free mice, which lack a microbiome, display a number of immunological deficiencies. Maintenance of intestinal homeostasis requires careful choreography. Various innate and adaptive immunocytes, mucosa, epithelia, microbes and nutrients collectively contribute to tolerate antigens from food and microbes, permitting the adsorption of nutrients and metabolites, but at the same time imposing a tight barrier to pathogen invasion across a surface area estimated at 200 m2. Gut bacteria are closely tied to the governance of T and B cell biology, and to the production of secretory IgA, of which some 3-5g is delivered to the gut daily. Intensive efforts have already identified a number of immunomodulatory bacteria and as described below, some of these are under evaluation as primary or adjunct live biotherapeutics for serious disease.

Similarly, research efforts have begun to elucidate life-long links between the gut microbiome and the brain. The vagus nerve is the main channel of communication between gut and brain, with nerve terminals (predominantly of sensory afferent neurons) dispersed widely within the intestinal wall. These interact with endocrine, immune, endothelial, dietary, and also microbial components, providing information on gut status to the brain, which can in turn signal in the opposite direction via efferent neurons within the same vagal conduit. Strikingly, microbes themselves are able to communicate directly with neurons by producing neurotransmitters, most notably gamma- aminobutyric acid, noradrenalin, serotonin, acetylcholine and dopamine.

There is also experimental evidence that microbes can communicate with the central nervous system independently of the vagus nerve. Well-defined microbial metabolites like secondary bile acids, short-chain fatty acids and tryptophan derivatives can reach the brain directly via the circulation, as well as influence enteroendocrine cells, which in-turn secrete gut hormones that can both activate the vagus nerve and directly activate the brain.

In animals, microbes are shown to influence behaviours (such as anxiety), development and gene expression within the nervous system, and have been implicated, with varying degrees of evidentiary robustness, in Parkinson’s disease, autism and schizophrenia.

Diet

Studies in twins indicate host genetics plays a relatively limited role in shaping the structure and activity of the gut microbiome. In contrast, the role of diet, alongside other environmental factors such as antibiotic use and more sanitised living conditions, is thought to be more influential. Certainly, microbiomes colonising individuals living in developed economies are far different from those seen in peoples living in closer contact with their natural environment, which generally possess greater taxonomic diversity and can vary compositionally across the year reflecting seasonal availability of food. Investigators posit that the taxonomically eroded microbiota found in more industrialised settings have lost key microbial homeostatic functionalities that would otherwise help prevent increasingly common non-communicable conditions like allergy, asthma, obesity and autoimmunity, many of which are strongly linked epidemiologically to undefined environmental factors.

It is no surprise then, given the highly competitive nature of the microbial ecosystem, that switching to predominantly plant- or animal-based diets can quickly alter gut microbial community composition and functionality. It is significant that raw and cooked forms of the same plant foods are shown to shape the microbiome in different ways, again illustrating the plasticity of the microbiome in adapting to environmental pressures. On a more molecular level, individual dietary compounds like polyphenols and bioactive glycans can both influence microbiome composition and promote the production of microbial metabolites, such as short chain fatty acids, that can directly impact host physiology. How longer-term dietary changes affect the microbiome, or how long changes promoted by acute dietary change persist, is less studied, requiring longer-term human dietary intervention and follow-up. However, the collective evidence clearly identifies diet as an appealingly accessible means through which the microbiome, and by extension host physiology, might be influenced for health benefits.

A major historical challenge to successful dietary interventions is the high variability in individual metabolic response to food intake. Recent work by a number of independent groups indicates that the microbiome can play a key role in determining divergent metabolic profiles (specifically, glucose and triglyceride metabolism) recorded after subjects consume the same dietary input. This observation has potentially important implications for dietary recommendations and suggests baseline microbiome, alongside other data inputs like age and sex, can combine to predict likely dietary response. Such predictive tools would be of significant help in moving beyond ubiquitous one-size-fits-all general dietary guidelines toward the formulation of personalised diets optimally geared around the metabolic response traits of the individual. To reach this point much larger, deeper and longitudinal studies need be undertaken.

Translational opportunities – modulating the microbiome

Microbiome-derived products have the potential to address broad unmet needs of both general consumers and patients.

Consumer products. In the consumer space, microbiome product opportunities include dietary supplements, foods, beverages and skin care products. Consumer probiotics are well established, with a global market of $40 billion in 2017. The market is primarily based on various ‘generally regarded as safe’ (GRAS) such as Bifidobacterium, Lactobacillus and Streptococcus strains. Although Bifidobacteria and Lactobacilli are common gut commensals, the resident strains are not typically those found in probiotic products, and probiotic strains tend only to be transient members of the gut microflora, although can persist for at least 5 months when taken after antibiotic treatment, where they have been reported to delay reassembly of the pre-antibiotic microbiome.

Sales of probiotics have increased markedly in recent years and revenues are forecast to continue to expand, boosted by association with a host of health promoting effects including immune and intestinal functionality and microbial balance. However, the quality of scientific evidence supporting their utility is patchy and observations in the laboratory have often not consistently translated in human studies.

Similarly, prebiotics, most often in the form of dietary fibers, e.g. non-digestible polysaccharides and oligosaccharides that are not directly digested by humans but that are fermented by some bacteria, also occupy a well-established but highly fragmented market niche, but likewise suffer from a lack of consistent evidentiary support for a raft of associated health benefits.

Thus, despite a large literature spanning diverse aspects of health and across different disease states and geographies, the extent and type of benefits accruing from the use of pre- and probiotic products remain somewhat controversial. Historically, the field has been hindered by a number of challenges. In the first instance, as compared to the pharmaceutical sector, consumer products are generally lower-cost, lower-margin commodities with limited barriers to market entry. These aspects combine to constrain research and development budgets in food companies, which can limit the size and depth of investigative studies. In addition, study populations (age, sex, baseline microbiome, diet), bacterial strains, manufacturing and formulation protocols, dosing schedules, endpoints and statistical powering are highly heterogeneous from study to study, making even meta-analyses of the pre-/probiotic scientific landscape problematic. For beneficial effects to be credible, they must be confirmed across homogeneous independent human studies.

It is to be hoped that, as understanding of the microbiome increases, better trials can be designed. For example, gut colonisation can now be measured more accurately, as can the influence of the pre-/probiotic on the resident gut microbiome. More recent studies have clearly shown how the degree of gut colonisation by probiotic strains varies across a population, a metric that is likely to significantly impact how the host responds. Equally, it is becoming more plausible to stratify trial subjects on the basis of baseline microbiome composition, opening the possibility to identify favourably responsive sub-populations and ultimately thereby matching individuals with consumer microbiome products that have the greatest chance of benefiting them.

Importantly, better informed consumers now expect (and indeed deserve) the coming next generation of new pre- and probiotic products, born from improved understanding of our commensal flora, to be supported by convincing efficacy data collected from robust clinical studies that employ the latest analytical tools. It is to be hoped that raising the scientific bar will bring rigor and clarity to the benefit of consumers.

To secure credibility with consumers, achieve market differentiation and attract premium pricing, manufacturers will likely be forced to seek regulated health claims, a path that has previously proven notoriously difficult to navigate in the pre-/probiotic setting. Consumer-focused microbiome products face a complex regulatory challenge, with significant differences encountered territorially. Probiotics making health claims around treatment of disease or disorder are regulated as medical or pharmaceutical products. Otherwise, probiotics are regulated as food supplements and regulation is focused on the acceptability of any claims, rather than efficacy, safety and quality of the product.

In Europe, probiotics are defined as functional foods and form part of the European Food Safety Authority’s Qualified Presumption of Safety list. In the United States (US) GRAS products are approved by the Federal Drug Administration (FDA), but dietary supplements are considered as “food” and regulated by the Dietary Supplement Health and Education Act. Therapeutic probiotics are regulated by the FDA. In some cases, a product can be classified as a medical food intended for the dietary management of a specific disease. This category is regulated much less stringently than pharmaceutical products. Conversely, in Japan, where public awareness of microbiome- related products is much higher, probiotic health claims are awarded on a regular basis.

Notable next generation probiotic consumer products presently in development include single microbial species (Eubacterium hallii and Akkermansia muciniphilia) for the management of pre- diabetes and Cutibacterium acnes strains to resolve acne vulgaris.

Therapeutic products. In the therapeutic disease setting, broadly two types of microbiome products are conceived; ‘bugs as drugs’ and ‘drugs from bugs’. In the former, live microbial preparations, running the gamut from whole fecal transplants, rationally designed consortia, and individual microbes, are delivered orally or via enema to remedy deficits in the preexisting microbiota. In the latter approach, functional metabolic products of the microbiome are identified, possibly derivatized to improve performance, and formulated as drugs, again often to make up for an inherent deficiency in the microbiome.

Fecal microbial transplants (FMT) have a long, unregulated clinical history. Their utility in a randomised well-controlled trial setting was not confirmed until 2013, in the impressive rescue of patients suffering from serious recurrent Clostridium difficile infection. No regulator approved FMT product is currently marketed, but FMTs are frequently used in C. difficile infections that are refractory to standard therapies. Although more than 10,000 FMT procedures annually are estimated to occur in the US alone, the regulatory picture remains unsettled with treatment of recurrent C. difficile infections allowed under an interim FDA policy. FMT for other indications cannot be performed without an investigational new drug submission. A number of credible stool banks now offer more standardised and thoroughly screened fecal material for the clinic. Safety concerns were re-emphasised during 2019 when two patients (one of whom died) in different trials developed antibiotic-resistant Escherichia coli bacteremia after receiving tainted stool from the same donor.

FMT product consistency and safety challenges have persuaded most live biotherapeutic product developers to nominate single species (native or engineered) or consortia of selected bacteria species as candidate drugs. These efforts seek to treat not only C. difficile infection, but also a host of other conditions such as inflammatory bowel disease, acute graft versus host disease, Parkinson’s disease and asthma.

Screening platforms to systematically identify specific functionalities that underpin microbial influence of host biology are also now emerging. These have the advantage of creating more ‘drug- like’ products, typically single chemical entities that may be further derivatized to improve potency and pharmacokinetics. However, those developing live biotherapeutics argue that the microbiome’s influence over host biology is more orchestral, with multiple pathways engaged simultaneously. It appears likely that both schools will ultimately find their niche. For example, live biotherapeutic approaches may be preferred when dramatic reconfiguration of a severely depleted microbiome is required, while single chemical agents may be effective where defined microbial pathways, metabolites or host targets, are clearly implicated in mechanism of action.

Latterly, there has been a growing focus on microbiome therapeutics as adjunct therapies that increase the proportion of patients responding favourably to a primary drug. This paradigm draws upon the microbiome’s influence over systemic ‘tone’ in individual patients, which in turn affects biology around the mechanism of action of the primary drug. A case in point is immune checkpoint inhibitor (ICI) therapy of cancer. Here multiple groups studying anti-PD1/L1 therapy in melanoma, non-small cell lung cancer and renal cancer have uncovered associations (albeit of different bacterial taxa) between patient baseline microbiome composition and response/non- response to therapy. Several trials are now underway with live biotherapeutics to determine if supplementing the microbiota of patients with bacteria that associate with favourable clinical outcomes can increase the overall proportion of patients benefiting from ICIs. Additional cohort studies suggest this pharmacomicrobiomic stratification of patients may reasonably extend to other disease/drug combinations that are driven by and/or resolved through the immune system.

Delivering on the promise

The microbiome has not lacked for media exposure, nor has it escaped hype. The nascent industry that sprung-up around key scientific innovation has attracted significant venture and strategic investment, in aggregate more than $2 billion, with the most advanced programs already in the clinic. This is therefore a critical time. Clear clinical successes will reinforce convictions and accelerate the influx of entrepreneurs and capital, whereas failure will alter the perceived risk profile and see valuations and funding fall. The earliest clinical trial data with commercial products, such as that reported in the treatment of C. difficile infection, have not thus far been fully convincing and this is starkly reflected in the re-pricing of publicly traded microbiome companies. Pivotal clinical data are expected from a number of trials during 2020 and these outcomes will be vital in setting the near-to-mid-term trajectory of the industry.

Microbiome translation is a young field and, not unexpectedly, major challenges persist. The manufacture and use of complex mixtures of live biotherapeutic organisms present unique challenges. Implementing thorough screens for virulence genes and mobile genetic elements is essential to ensure patient safety and product consistency. Designing dosing regimens and promoting engraftment within variable recipient microbiomes to the extent needed for clinical efficacy also present hurdles and may require a period of trial and error to optimise. In addition, there is a paucity of information around how microbial phenotypes, such as prototrophies, auxotrophies, gene expression and anatomical location, impinge upon the success of live bacterial products in the complex, diverse and highly competitive environment of the gastrointestinal tract.

It is probable that a second generation of prebiotics and dietary guidelines will emerge to be used in concert with live biotherapeutics to promote engraftment and help maximise clinical efficacy of live biotherapeutics across a patient population.

Given the current vagaries of live biotherapeutic product development, the reduction of the microbiome to its constituent functionalities holds obvious appeal, in other words, “drugs from bugs”. Broader use of metagenomic and metabolomic data facilitates a push in this direction. Bio-active microbial metabolites can be identified through big data informed hypotheses and through ‘pharma-like’ functional screens. While this approach for microbiome-based drug development may seem relatively more controlled, it obviously lacks the poly pharmacy modalities that are offered by live biotherapeutic consortia, and which may be necessary in some settings for clinical efficacy. Although there are likely to be individual niches to which live biotherapeutic or small molecule solutions are best suited, which approach is likely to become the richest source of products in the near-to-medium term is difficult to predict. However, in the fullness of time, once mechanistic pathways are more completely elucidated, it seems likely that microbiome-derived small molecule agents will come to dominate the marketplace.

From an investor perspective, the microbiome remains for now a young, rapidly evolving and incredibly exciting field that offers high-risk high-reward investment opportunity. Investment decisions must be objectively well-measured and based on solid evidentiary data, including, where appropriate, convincing preclinical/clinical efficacy packages. Companies should be focused, and possess strong commercial rationale for product development, clear competitive advantage, including credible intellectual property filings. Active interaction with regulators, physicians, patients and payers is essential from an early stage to create a navigable path for what are a new class of therapeutics. Valuations need to reflect the attendant risks.

Scientists have variously associated the microbiome with almost every aspect of human physiology, and there is clear and present danger of letting the story get ahead of the facts. However, there is now a realistic possibility that by carefully uncovering and judiciously applying the secrets of the microbiome we can live healthier lives.

Authors: Anthony Williamson, Head of Biotechnology and Health Science, and Irene Corthesy Malnoë, Head of Advanced Nutrition