Hostname: page-component-745bb68f8f-kw2vx Total loading time: 0 Render date: 2025-02-11T10:05:53.813Z Has data issue: false hasContentIssue false

Increasing reproducibility and interpretability of microbiota-gut-brain studies on human neurocognition and intermediary microbial metabolites

Published online by Cambridge University Press:  15 July 2019

Esther Aarts
Affiliation:
Centre for Cognitive Neuroimaging, Donders Institute for Brain, Cognition and Behaviour, Radboud University, 6525 EN Nijmegen, The Netherlands. esther.aarts@donders.ru.nlwww.ru.nl/donders/pac/fac
Sahar El Aidy
Affiliation:
Department of Molecular Immunology and Microbiology, Groningen Biomolecular Sciences and Biotechnology Institute, University of Groningen, 9747 AG Groningen, The Netherlandssahar.elaidy@rug.nlhttps://www.rug.nl/research/microbial-physiology/el-aidy-group/

Abstract

In this commentary, we point to guidelines for performing human neuroimaging studies and their reporting in microbiota-gut-brain (MGB) articles. Moreover, we provide a view on interpretational issues in MGB studies, with a specific focus on gut microbiota–derived metabolites. Thus, extending the target article, we provide recommendations to the field to increase reproducibility and relevance of this type of MGB study.

Type
Open Peer Commentary
Copyright
Copyright © Cambridge University Press 2019 

In a relatively new field, researchers have now started to use human neuroimaging techniques, like functional magnetic resonance imaging (fMRI), to study the microbiota-gut-brain (MGB) axis. A quick search with the terms “microbiome” and “fMRI” yields at least eight studies since 2013 that have linked task-related or resting fMRI to gut microbiota measures or interventions (Aarts et al. Reference Aarts, Ederveen, Naaijen, Zwiers, Boekhorst, Timmerman, Smeekens, Netea, Buitelaar, Franke, van Hijum and Arias Vasquez2017; Ahluwalia et al. Reference Ahluwalia, Wade, Heuman, Hammeke, Sanyal, Sterling, Stravitz, Luketic, Siddiqui, Puri, Fuchs, Lennon, Kraft, Gilles, White, Noble and Bajaj2014; Bagga et al. Reference Bagga, Aigner, Reichert, Cecchetto, Fischmeister, Holzer, Moissl-Eichinger and Schöpf2018a; Reference Bagga, Reichert, Koschutnig, Aigner, Holzer, Koskinen, Moissl-Eichinger and Schopf2018b; Osadchiy et al. Reference Osadchiy, Labus, Gupta, Jacobs, Ashe-McNalley, Hsiao and Mayer2018; Pinto-Sanchez et al. Reference Pinto-Sanchez, Hall, Ghajar, Nardelli, Bolino, Lau, Martin, Cominetti, Welsh, Rieder, Traynor, Gregory, De Palma, Pigrau, Ford, Macri, Berger, Bergonzelli and Bercik2017; Tillisch et al. Reference Tillisch, Labus, Kilpatrick, Jiang, Stains, Ebrat, Guyonnet, Legrain-Raspaud, Trotin, Naliboff and Mayer2013; Reference Tillisch, Mayer, Gupta, Gill, Brazeilles, Le Nevé, van Hylckama Vlieg, Guyonnet, Derrien and Labus2017). However, out of those eight studies, only three based their analyses on groups of more than 20 participants, only two shared their neuroimaging data (in line with journal requirements), and only two (using an intervention) preregistered their design and analyses. As the field is still evolving, we would like to take this opportunity to make a plea for reproducible and interpretable MGB findings, pointing to guidelines for preregistration, results reporting, and data sharing in human neuroimaging studies and making suggestions to increase functional MGB interpretations, thus going beyond the many valid criticisms reported by Hooks et al. and actually providing recommendations.

Many MGB intervention studies register their human trials in a clinical trial register, but this is not common yet for observational MGB studies. However, it is important for reproducibility to preregister the main experimental question, hypotheses, design of the study, justification of the sample size, and the primary and secondary analyses. This limits the researcher's degrees of freedom and, hence, (uncorrected) multiple testing and presentation of only desirable results (p-hacking) or post hoc hypothesis generation that is presented as a priori (HARKing) (Forstmeier et al. Reference Forstmeier, Wagenmakers and Parker2017). Preregistration also allows the presentation of null results, which are crucial for a field to develop. Naturally, journals play an important role in allowing null results to be presented and preventing publication bias. Registered reports (i.e., peer-reviewed preregistrations) are helpful for eventually reporting possible null results, as reviewers have deemed the design and sample size valid for answering the research questions (Chambers et al. Reference Chambers, Feredoes, Muthukumaraswamy and Etchells2014). Of course, findings based on unplanned and unregistered analyses are still relevant to generate new hypotheses, but these should be reported as exploratory and confirmed in future studies designed to answer this specific new research question.

Similar to gut microbiome data, fMRI data are characterized by high dimensionality, with thousands of voxels and, hence, thousands of statistical tests performed. Therefore, when reporting fMRI data, the type of multiple comparison correction should be clearly described (Poldrack et al. Reference Poldrack, Fletcher, Henson, Worsley, Brett and Nichols2008). For cluster-based inference, the right use of the cluster-defining threshold is essential (Eklund et al. Reference Eklund, Nichols and Knutsson2016).

Open science practices – such as sharing of data, analysis scripts, and preprints of publications – have many benefits, including easier replication, increased availability of data for theory building and meta-analyses, and increased possibility of review, before and after publication of an article. For human MGB studies using MRI, we can recommend the advice by the Committee on Best Practice in Data Analysis and Sharing (COBIDAS) (Nichols et al. Reference Nichols, Das, Eickhoff, Evans, Glatard, Hanke, Kriegeskorte, Milham, Poldrack, Poline, Proal, Thirion, Van Essen, White and Yeo2017).

Increasing research on the gut microbiota, including MGB research, has also resulted in massive amounts of data being generated and shared (Editorial 2017). For example, consortia efforts such as the Earth Microbiome Project, Human Microbiome Project, Tara Oceans, and MetaHIT, as well as laboratory-level projects, have generated extensive data pools. Particularly, because of the size, complexity, and diverse formats that come with generating massive data, accessibility and accuracy of these data remain problematic to understand existing data sets. Increasing awareness of the different databases available for the data types and appropriate analyses to use (as listed online: https://www.nature.com/sdata/policies/repositories) could help overcome this obstacle. Equally important is to report how samples were collected, handled, and stored and what methodology was applied to analyze them, as these factors have a dramatic influence on the results.

Many MGB studies propose microbial metabolites as intermediary mechanisms that consolidate the link made between brain, cognition or mood, and microbes (Aarts et al. Reference Aarts, Ederveen, Naaijen, Zwiers, Boekhorst, Timmerman, Smeekens, Netea, Buitelaar, Franke, van Hijum and Arias Vasquez2017; Bagga et al. Reference Bagga, Reichert, Koschutnig, Aigner, Holzer, Koskinen, Moissl-Eichinger and Schopf2018b; Osadchiy et al. Reference Osadchiy, Labus, Gupta, Jacobs, Ashe-McNalley, Hsiao and Mayer2018; Waclawikova & El Aidy Reference Waclawikova and El Aidy2018). However, bioavailability of microbial metabolites remains poorly understood. For example, there is scarce evidence on whether, when, how, and where these metabolites cross the epithelial barrier and blood-brain barrier and how tightly this process in regulated. Despite the remarkable progress in developing high-throughput techniques to identify microbial-derived metabolites, the majority are yet unidentified, and most of the identified ones remain functionally uncharacterized. The latter is related to the challenges of culturing bacterial species (Lagier et al. Reference Lagier, Dubourg, Million, Cadoret, Bilen, Fenollar, Levasseur, Rolain, Fournier and Raoult2018). Although many bacteria in the gut remain uncultured, the current advances in the culturomic approach have enabled the culture of hundreds of new commensal bacteria, thus providing exciting new perspectives on their metabolic activity.

Another challenge confronting MGB studies is the interaction between gut microbes and dietary components – including precursors of neurotransmitters – and the effect on their metabolic products. Currently, all functional MGB studies have been limited to single or limited bacterial species and have not taken diet composition into account because of the extreme complexity of the gut organ system and technical limitations. Microbial-derived compounds, which are mainly products of their breakdown of diet, signal not only to the host cells, but also to other gut bacteria in a beneficial or adverse way (Adair & Douglas Reference Adair and Douglas2017). Considering that more than 1,000 different bacterial species are estimated to reside in the human gastrointestinal tract, this gives an enormous amount of possible variations in inter-microbial communication by produced metabolites (Postler & Ghosh Reference Postler and Ghosh2017). One way to facilitate interpretation of the complex MGB interactions and ultimately allow new therapeutic approaches to treat MGB-related disorders is by developing high-throughput gut and brain organoid systems obtained from urine and blood samples of individuals and cultured with stool samples of the same individual (Dutta et al. Reference Dutta, Heo and Clevers2017). Results could be used to explain inter-individual differences in human neurocognition and its response to – for example, nutritional or pharmacological – interventions, providing a functional and interpretable MGB link.

References

Aarts, E., Ederveen, T. H. A., Naaijen, J., Zwiers, M. P., Boekhorst, J., Timmerman, H. M., Smeekens, S. P., Netea, M. G., Buitelaar, J. K., Franke, B., van Hijum, S. A. F. T. & Arias Vasquez, A. (2017) Gut microbiome in ADHD and its relation to neural reward anticipation. PLoS ONE 12(9):e0183509. Available at: https://doi.org/10.1371/journal.pone.0183509.Google Scholar
Adair, K. L. & Douglas, A. E. (2017) Making a microbiome: The many determinants of host-associated microbial community composition. Current Opinion in Microbiology 35:2329. Available at: https://doi.org/10.1016/j.mib.2016.11.002.Google Scholar
Ahluwalia, V., Wade, J. B., Heuman, D. M., Hammeke, T. A., Sanyal, A. J., Sterling, R. K., Stravitz, R. T., Luketic, V., Siddiqui, M. S., Puri, P., Fuchs, M., Lennon, M. J., Kraft, K. A., Gilles, H., White, M. B., Noble, N. A. & Bajaj, J. S. (2014) Enhancement of functional connectivity, working memory and inhibitory control on multi-modal brain MR imaging with rifaximin in cirrhosis: Implications for the gut-liver-brain axis. Metabolic Brain Disease 29(4):1017–25. Available at: https://doi.org/10.1007/s11011-014-9507-6.Google Scholar
Bagga, D., Aigner, C. S., Reichert, J. L., Cecchetto, C., Fischmeister, F. P. S., Holzer, P., Moissl-Eichinger, C. & Schöpf, V. (2018a) Influence of 4-week multi-strain probiotic administration on resting-state functional connectivity in healthy volunteers. European Journal of Nutrition. Available at: https://doi.org/10.1007/s00394-018-1732-z.Google Scholar
Bagga, D., Reichert, J. L., Koschutnig, K., Aigner, C. S., Holzer, P., Koskinen, K., Moissl-Eichinger, C. & Schopf, V. (2018b) Probiotics drive gut microbiome triggering emotional brain signatures. Gut Microbes 9(6):486–96. Available at: https://doi.org/10.1080/19490976.2018.1460015.Google Scholar
Chambers, C. D., Feredoes, E., Muthukumaraswamy, S. D. & Etchells, P. J. (2014) Instead of “playing the game” it is time to change the rules: Registered Reports at AIMS Neuroscience and beyond. AIMS Neuroscience 1(1):417.Google Scholar
Dutta, D., Heo, I. & Clevers, H. (2017) Disease modeling in stem cell-derived 3D organoid systems. Trends in Molecular Medicine 23(5):393410. Available at: https://doi.org/10.1016/j.molmed.2017.02.007.Google Scholar
Editorial (2017) Overcoming hurdles in sharing microbiome data. Nature Microbiology 2(12):1573. Available at: https://doi.org/10.1038/s41564-017-0077-3.Google Scholar
Eklund, A., Nichols, T. E. & Knutsson, H. (2016) Cluster failure: Why fMRI inferences for spatial extent have inflated false-positive rates. Proceedings of the National Academy of Sciences USA 113(28):7900–905. Available at: https://doi.org/10.1073/pnas.1602413113.Google Scholar
Forstmeier, W., Wagenmakers, E. J. & Parker, T. H. (2017) Detecting and avoiding likely false-positive findings: A practical guide. Biological Reviews: Cambridge Philosophical Society 92(4):1941–68. Available at: https://doi.org/10.1111/brv.12315.Google Scholar
Lagier, J. C., Dubourg, G., Million, M., Cadoret, F., Bilen, M., Fenollar, F., Levasseur, A., Rolain, J.-M., Fournier, P.-E. & Raoult, D. (2018) Culturing the human microbiota and culturomics. Nature Reviews Microbiology 16:540–50. Available at: https://doi.org/10.1038/s41579-018-0041-0.Google Scholar
Nichols, T. E., Das, S., Eickhoff, S. B., Evans, A. C., Glatard, T., Hanke, M., Kriegeskorte, N., Milham, M. P., Poldrack, R. A., Poline, J.-B., Proal, E., Thirion, B., Van Essen, D. C., White, T. & Yeo, B. T. (2017) Best practices in data analysis and sharing in neuroimaging using MRI. Nature Neuroscience 20(3):299303. Available at: https://doi.org/10.1038/nn.4500.Google Scholar
Osadchiy, V., Labus, J. S., Gupta, A., Jacobs, J., Ashe-McNalley, C., Hsiao, E. Y. & Mayer, E. A. (2018) Correlation of tryptophan metabolites with connectivity of extended central reward network in healthy subjects. PLoS ONE 13(8):e0201772. Available at: https://doi.org/10.1371/journal.pone.0201772.Google Scholar
Pinto-Sanchez, M. I., Hall, G. B., Ghajar, K., Nardelli, A., Bolino, C., Lau, J. T., Martin, F.-P., Cominetti, O., Welsh, C., Rieder, A., Traynor, J., Gregory, C., De Palma, G., Pigrau, M., Ford, A. C., Macri, J., Berger, B., Bergonzelli, G. & Bercik, P. (2017) Probiotic Bifidobacterium longum NCC3001 reduces depression scores and alters brain activity: A pilot study in patients with irritable bowel syndrome. Gastroenterology 153(2):448–59.e8. Available at: https://doi.org/10.1053/j.gastro.2017.05.003.Google Scholar
Poldrack, R. A., Fletcher, P. C., Henson, R. N., Worsley, K. J., Brett, M. & Nichols, T. E. (2008) Guidelines for reporting an fMRI study. Neuroimage 40(2):409–14.Google Scholar
Postler, T. S. & Ghosh, S. (2017) Understanding the holobiont: How microbial metabolites affect human health and shape the immune system. Cell Metabolism 26(1):110–30. Available at: https://doi.org/10.1016/j.cmet.2017.05.008.Google Scholar
Tillisch, K., Labus, J., Kilpatrick, L., Jiang, Z., Stains, J., Ebrat, B., Guyonnet, D., Legrain-Raspaud, S., Trotin, B., Naliboff, B. & Mayer, E.A. (2013) Consumption of fermented milk product with probiotic modulates brain activity. Gastroenterology 144(7):1394–401.Google Scholar
Tillisch, K., Mayer, E. A., Gupta, A., Gill, Z., Brazeilles, R., Le Nevé, B., van Hylckama Vlieg, J. E. T., Guyonnet, D., Derrien, M. & Labus, J. S. (2017) Brain structure and response to emotional stimuli as related to gut microbial profiles in healthy women. Psychosomatic Medicine 79(8):905–13. Available at: https://doi.org/10.1097/PSY.0000000000000493.Google Scholar
Waclawikova, B. & El Aidy, S. (2018) Role of microbiota and tryptophan metabolites in the remote effect of intestinal inflammation on brain and depression. Pharmaceuticals (Basel) 11(3):63. Available at: https://doi.org/10.3390/ph11030063.Google Scholar