Elsevier

Current Opinion in Microbiology

Volume 54, April 2020, Pages 119-126
Current Opinion in Microbiology

Thinking about the microbiome as a causal factor in human health and disease: philosophical and experimental considerations

https://doi.org/10.1016/j.mib.2020.01.018Get rights and content

Relationships between hosts and host-associated microbial communities are complex, intimate, and associated with a wide variety of health and disease states. For these reasons, these relationships have raised many difficult questions and claims about microbiome causation. While philosophers and scientists alike have pondered the challenges of causal inference and offered postulates and rules, there are no simple solutions, especially with poorly characterized, putative causal factors such as microbiomes, ill-defined host effects, and inadequate experimental models. Recommendations are provided here for conceptual and experimental approaches regarding microbiome causal inference, and for a research agenda.

Section snippets

Communities and relationships

As a general rule, microbes live in communities. And as evolving, capable, and adaptive entities, communities shape and are shaped by biotic and abiotic components of their environment. Communities in particular, but any complex system in general, rely on multiple types of interactions. Nearly all microbial communities have a heterogenous composition, and they are structured by interactions among members and between members and their environment. In the case of host-associated communities, the

An illustrative case in complex causes and effects (Gil-Cruz)

In any complex ecosystem, a transient perturbation can lead to sustained effects because of compensatory or reinforcing processes. In these settings, one expects multiple ‘causal’ factors and evolving causal processes over time. This is the case with a wide variety of chronic inflammatory and auto-immune conditions in humans associated with maladaptive responses of the microbiome. Gil-Cruz et al. illustrate some of the challenges in teasing apart the possible causal role of the microbiome in

Philosophical underpinnings of causal inference

The Scottish philosopher, David Hume (1711-1776) is credited with providing the first modern critical re-assessment of the conceptual basis for causal inference. In the first Enquiry, he drew a distinction between causal associations and associations based on either ‘resemblance’ or on ‘contiguity in time and place’ [5]. He pointed to the fundamental importance of causal inference for human understanding because, as he argued, it provides the only path beyond present facts, and allows for

Traditional approaches and requirements for assessing causal relationships

Most discussions about proof of causation in biomedical science refer back to the postulates of Jakob Henle and Robert Koch [9,10]. Drawing from Koch’s work on anthrax and tuberculosis, these postulates were intended as a standard for testing the role of a microbe in disease:

  • (i)

    The parasite occurs in every case of the disease in question and under circumstances which can account for the pathological changes and clinical course of the disease.

  • (ii)

    The parasite occurs in no other disease as a fortuitous

Alternative approaches

The field of epidemiology provides useful guidance for addressing causal relationships in complex ecosystems such as humans, when a putative causal factor such as an environmental factor or a microbiome is not easily physically isolated or defined. Epidemiology is a population-based science concerned with the distribution and determinants of health-related states and events, or literally, the study of ‘what befalls a population’ [14]. Considerations of the environment as a source of important

Recent discussions by others about the microbiome and causality

Needless to say, the importance of examining experimental approaches for microbiome causal inference has been recognized by others. Byrd and Segre discussed the mitigating effects of host-associated communities on would-be pathogens through the long-recognized property of colonization resistance [22]. Recent work on this property has sought to identify individual commensals or microbial products with the means of inhibiting pathogens, directly or indirectly. Yet colonization resistance is

Further consideration of the case study and disease models

In examining the work of Gil-Cruz et al. [2] in light of the proposals and considerations discussed above, a few additional points might be useful. Aside from clarification about the exact nature of the putative agent, and before one might address the specificity, sensitivity and temporality of its association with an effect, there are issues surrounding the suitability of the murine models. The choice of the ‘self-epitope’ for construction of the transgenic mouse would appear relevant and

Path forward and needs

Where does this leave us? While true proofs of causation in biology are impossible, formal treatment of causal inference for complex ‘actors’ and effects, such as host-associated microbiomes and host health phenotypes, is an important and useful goal for which additional theoretical and practical guidance is needed. The traditional postulates that Koch proposed for microbial pathogens can be are helpful in highlighting general qualities of associations like sensitivity and specificity that

Conflicts of interest

DAR is a consultant to ArcBio, Cepheid, ClickDiagnostics, Karius, ProdermIQ, and holds stock or stock options in Evelo Biosciences, NanoBio, Second Genome, and Seres.

Funding

This work was supported by funding from the Chan Zuckerberg Biohub Microbiome Initiative and the Thomas C. and Joan M. Merigan Endowment at Stanford University.

References and recommended reading

Papers of particular interest, published within the period of review, have been highlighted as:

  • • of special interest

  • •• of outstanding interest

Acknowledgements

For Stanley Falkow, the history, importance, and intellectual challenge of causal inference for pathogens and disease was compelling. It created a rationale for much of his work to discern mechanisms of virulence, and offered a potential guide for experimental strategies. But Stanley always thought in much broader terms about the nature of host (animal)-microbe relationships. As much as one might be inclined to attribute a host outcome to an affiliated microbe or a microbial gene product,

References (34)

  • D. Sprockett et al.

    Role of priority effects in the early-life assembly of the gut microbiota

    Nat Rev Gastroenterol Hepatol

    (2018)
  • B. Callahan et al.

    Reproducible research workflow in R for the analysis of personalized human microbiome data

    Pac Symp Biocomput

    (2016)
  • R. Koch

    Ueber bakteriologische Forschung

    Verh. X. Int. Med. Congr. Berlin, 1890

    (1892)
  • A.S. Evans

    Causation and disease: the Henle-Koch postulates revisited

    Yale J Biol Med

    (1976)
  • D.N. Fredricks et al.

    Sequence-based identification of microbial pathogens: a reconsideration of Koch’s postulates

    Clin Microbiol Rev

    (1996)
  • S. Falkow

    Molecular Koch’s postulates applied to microbial pathogenicity

    Rev Infect Dis

    (1988)
  • S. Falkow

    Molecular Koch’s postulates applied to bacterial pathogenicity—a personal recollection 15 years later

    Nat Rev Microbiol

    (2004)
  • Cited by (3)

    • Diet leaves a genetic signature in a keystone member of the gut microbiota

      2022, Cell Host and Microbe
      Citation Excerpt :

      Given the chemical variety of dietary plant polysaccharides, the high genetic repertoire provided by a high microbial diversity in the gut is linked to robust metabolism and improved host health, while a low gut microbial diversity is generally associated with disease (Le Chatelier et al., 2013; Cotillard et al., 2013; Turnbaugh et al., 2009a). Many factors, such as drugs, lifestyle, and inflammatory responses to infections, can perturb microbiota composition and often result in decreased diversity (Relman, 2020). Diet changes the microbiota, and the impacts go beyond the known direct effect on body weight and metabolic imbalances (Kreuzer and Hardt, 2020; Sonnenburg and Sonnenburg, 2014).

    View full text