Microbiome research: overhyped or the great hope?

The rise of the field

Over the past 20 years the microbiome has moved from a fledgling field to a flourishing area of research. In the past 10 years alone approximately US $1.7 billion has been spent on microbiome research, primarily in the USA who have invested around $1billion – much of which funded the Human Microbiome Project [1].

The papers resulting from this investment have confirmed the importance of the microbiome in health and disease and have challenged our conceptions of microorganisms as solely the agents of infectious disease. However, there has been some skepticism about some of the research that’s been carried out, in addition, there are many unanswered questions that remain.

The microbiome has been associated with many conditions including cancer, autism spectrum disorder, obesity, anxiety and depression [2–4]; however, do changes in the microbiome represent the causes rather than the effects of these conditions? Is all research carried out to the highest standard? And is the microbiome more complicated than first thought? In this article, we take a look at some of the promise unlocked by microbiome research and investigate some of the challenges it faces in the future.

The great hope?

Few areas of science have been translated as quickly into therapeutic medicine – for example, microbiome-based therapy, fecal microbial transplant (FMT), is now one of the key treatments for recurrent Clostridium difficile infections, aiming to restore a healthy balance of bacteria in the gut of infected individuals by introducing enteric bacteria from healthy donors [5]. Moreover, the advance of FMT has sparked research and development into microbiome therapeutics, aiming to achieve the same outcome of restoring gut bacteria but with the advantage of a consistent and tested safety profile. There are several microbiome-based candidates currently in the clinical pipeline for C. difficile [6] but microbiome therapeutics or preventatives have been mooted for several conditions, including the use of probiotics for sepsis and ulcerative colitis [7,8].

Despite no microbiome-based biotherapeutics or diagnostic products being approved for use to-date, the current value of human-microbiome-based products and interventions for diagnostic and therapeutic use has been estimated at US $275–400 million worldwide, possibly rising to $2.2 billion by 2025 – demonstrating the opportunities and potential this area holds [1,9].

“Despite no microbiome-based biotherapeutics or diagnostic products being approved for use to-date, the current value of human-microbiome-based products and interventions for diagnostic and therapeutic use has been estimated at US $275–400 million worldwide.”

The microbiome has also been linked to how individuals respond to certain drugs, including chemotherapy [10].  Although this field is not yet impacting on the frontline of care, if microbiome–drug interactions could be better understood perhaps this information could play a central role in the move towards a precision medicine approach. With the microbiome being linked to so many conditions and diseases, a knowledge of microbiome composition (in combination with other factors) could allow therapies and doses to be specifically tailored to the individual, improving treatment.

Finally, it is also thought that an understanding of the microbiome could allow certain microbial patterns or species to be used as biomarkers of disease. The association of the microbiome in many different conditions suggests that either it could be therapeutically manipulated, if it was causative of the disease, or that changes in the microbiome could be use as early signs of the disease, if the disease was causing microbiome alterations. With a better understanding of the mechanisms by which microbiome alterations come about in disease states, this could be a huge field that has as-yet been unresearched and is filled with diagnostic potential.

Or overhyped?

What is a healthy microbiome?

Despite enthusiasm for manipulating the microbiome and pursing a healthy one, one issue facing microbiome research is the lack of definition for a ‘healthy’ microbiome. The Human Microbiome Project, launched in 2007, was aiming to address this question of the healthy microbiome; however, the project concluded that the taxonomic composition of the microbiome was often not a good correlate with host phenotype [12]. This lack of definition: makes it challenging to compare the microbiome in diseased states without a status quo; makes it challenging for developing potential therapeutics, as the endpoint they are aiming for is unknown; and also highlights how little we really understand about this living organ. It is likely that there is no single definition for a healthy microbiome, leading us on to the next challenge, is our community of microbes more complex than first thought?

Catalogues vs communities

Initial research into the microbiome often focused on specific species, these had perhaps previously, or were then subsequently, labelled ‘good’ or ‘bad’ bacteria. However, it quickly emerged that the microbiome was more complex than the impact of a single species. This realization has been compared with the field of genomics [13], which promised an ability to understand the genetic basis of disease; however, initial connections were discovered to be more complicated after further research, only have weak influence on phenotype or require combination of genes to influence phenotype.

William Hanage (Harvard University, MA, USA) commented: “The history of science is replete with examples of exciting new fields that promised a gold rush of medicines and health insights but required skepticism and years of slogging to deliver.” [13]

“With regards to the microbiome, composition is unique to each individual, can change across an individual’s lifetime and can be influenced by a range of factors including diet, location, birth method and potentially other as-yet uncovered determinants.”

With regards to the microbiome, composition is unique to each individual, can change across an individual’s lifetime and can be influenced by a range of factors including diet, location, birth method and potentially other as-yet uncovered determinants. Moreover, the microbiome is home to approximately 300–1000 [14,15] different species of bacteria, all of which will interact with each other, and the host, via metabolites and substances. Therefore, there is a need to understand the complex and mutable ecological and evolutionary relationships between microbes and the host as well as between the microbes themselves.

Before we can harness the potential of the microbiome therapeutically, do we need a deeper understanding of the complex dynamics and interactions that underpin it? What impact do microbe–microbe interactions and predator–prey relations between microbes have? And how does a community of microbes operate as a whole?

Correlation vs cause

A major argument that has been raised in the field of microbiome research, particularly in light of multiple studies showing associations between a disease state and an altered microbiome composition, is whether the microbiome changes are causes, rather than effects of, differences in health? This is an age-old scientific principle – many people will remember science teachers telling them ‘correlation does not imply causation’ or will have been shown graphs demonstrating this principle by linking two spurious trends [11]. The complexity of the microbiome, and the fact that there’s no definition for a ‘healthy’ one, makes it extremely challenging to determine a cause–effect relationship. However, if correlation does imply a causal relationship, what could the mechanism be?

One potential mechanism that’s been hypothesized is communication between the gut and brain via the vagus nerve. This hypothesis suggests that neuromodulatory chemicals released by gut bacteria could signal to the brain by influencing the activity of this nerve, potentially influencing neurological conditions such as anxiety and autism [16]. Gut bacteria also release chemicals that can act on hormone-producing cells, it has been suggested that this could cause these cells in turn to release hormones that will reach and influence the brain, via the bloodstream [17]. Finally, gut bacteria have been shown to produce metabolites that can influence the immune system and could therefore impact the host via this mechanism [18].

“Many links have only been explored in mice and not in humans; however, results from animal models should be interpreted with caution as they are not necessarily translatable.”

While many studies have drawn correlations by comparing groups of people, a big research tool for this field is the germ-free mouse – a mouse raised in a sterile environment that can then be exposed to particular microbes, and the impact assessed. These experiments have demonstrated that transplanting microbiota from individuals with specific conditions, such as anxiety, can induce anxiety-like symptoms in the mice [19], and have also been used to link specific strains of bacteria to particular conditions. Many links have only been explored in mice and not in humans; however, results from animal models should be interpreted with caution as they are not necessarily translatable [20]. Specifically, humans and mice are different with respect to body sizes, diet, gut anatomy and gut functions [21], moreover, several animal models of human digestive tract diseases such as colitis, obesity and diabetes are only superficially related to their human counterparts.

Quality control

Finally, it has been argued that microbiome researchers have not yet broadly embraced quality control practices for their data in a way that would make results comparable between studies but also reproducible by other groups [1]. Coupled with a growing concern about reproducibility more broadly in the biomedical sciences, this has been an area of concern; in order to transition from basic research to the clinic, technologies and computational methods for assessing human-associated microbial communities must be standardized and quality controlled.

Currently, looking across studies, variation at each step in the pipeline is huge – from physical specimen collection and processing, to the computational quantification of microbial communities – and there is little research about how these variations could impact the resulting findings [22]. To advance this field, it is essential to provide a palette of appropriate standards for each step of the assay process, and programs such as the Microbiome Quality Control project are already starting to address this [23].

FMT is another area where quality control could lead to variable results. A number of clinical trials for FMT have been carried out in obesity and irritable bowel syndrome, however, the results have been mixed [24]. Moreover, in C. difficile infection concerns have been raised around the variable safety profile and limited screening of FMT. This was exemplified recently when the US FDA issued a safety communication warning over the risk of serious infections caused by multidrug-resistant organisms in investigational FMT. This warning came after two immunocompromised adults who had received FMT developed invasive infections due to the transmission of multidrug-resistant organisms, sadly leading to the death of one of these cases [25].

Yoav Golan (Tufts University School of Medicine, MA, USA) commented on FMT: “It’s potentially hazardous. Although we screen the feces for some pathogens, other pathogens may still exist in the donated feces as well as other risk factors in the microbiome linked to other diseases that currently we don’t quite understand.” [26]

Conclusions

“This field shows great potential for diagnostics and therapeutics but some of the hype around microbiome research has led to concerning trends.”

The hopes generated when the first Human Microbiome data were published are reminiscent of the excitement when the first draft of the human genome was released in 2001. However, these datasets are only beginning in understanding our microbiomes and the impacts they have on our health. This field shows great potential for diagnostics and therapeutics but some of the hype around microbiome research has led to concerning trends. For example, warnings have been raised regarding individuals making ill-informed decisions, such as trying out ‘DIY’ FMT procedures, in addition, pseudo-science websites have been launched claiming to be ‘experts’ and offering advice or selling supplements, books or videos based on ‘harnessing’ the microbiome [20].

Although some of the links with the microbiome may just be correlations, Pinaki Panigrahi (Georgetown University, DC, USA) believes that some of these associations will pan out to be causative: “Compare microbiome therapeutics with antibiotics and ask: how many years and how many billions of dollars have been spent [in that field] to come to the stage where we can make strong statements? So, when we come to microbiome, we are really just at the beginning so it is not fair to say that we can establish causality – correlation is how you start.”

This research is in its infancy and although the field shows great promise – who knows which correlations might pan out? –  all players should guard against hype. Scientists and researchers need to develop improved experimental model and standardize methods to effectively evaluate conclusions, industry must be honest about the potential benefits of microbiome-based products and therapeutics, and journalists should stop the selective reporting of stories and also ensure that the content of their pieces is not exaggerated – correlation does not imply causation.