No single analysis can adequately cover all facets of a complex interdisciplinary field, but we believe there remain a few topics that warrant inclusion in the present discussion of microbiota-gut-brain (MGB) research. Foremost is developmental analysis of the MGB in parents, offspring, and parent-offspring relations. Development is arguably a necessary element in understanding any biological system. None of the “25 most cited” papers (Table 1 in Hooks et al.) comprise developmental analysis, that is, studies designed to elucidate pathways or processes by which the offspring gut is inoculated with bacteria and in which we see the MGB and behavior emerge and differentiate in interaction with their surroundings. Such analyses would include whether and how the stages of microbiota development stimulate neural, neuroendocrine, and behavioral activity at corresponding stages and that, in turn, reciprocally affect the gut bacteria. Another, important aspect of mammalian development is parent-offspring interaction: whether and how a parent-offspring microbiome may form and especially how the microbiota of each individual affects the other, with routes including behavior that is influenced by individual and familial microbiomes.
We emphasize that performing a manipulation at an early age and measuring effects at a later age or ages does not provide “developmental” information. Likewise, manipulations at various time points can produce different results at different life stages, but again, this says little about process or pathways. Such factors and myriad cascading and interactive events could contribute mechanistically to the development of disordered social behavior.
Another non-deferrable topic is the need for more integrative and informed use of behavior as a measure and as a mechanism. Hooks et al. appropriately identify some of the limitations of the handful of behavioral measures currently used to model social-emotional health and disease. Now is a time to discuss how to validate and apply behavioral measures as more precise and informative metrics of function, especially in social-emotional processes. We need to acknowledge the import and the sophistication of careful behavioral analyses. Needed are methods that are sensitive, objective, and sufficiently nuanced to capture social behavior and the nebulous emotional forces that shape and help regulate it. This need is not unique to MGB investigations; it is widespread in the life sciences, where we seek to understand relations among genetic, cellular, system-level, and organismal functions. The MGB as a system is observed and evaluated via behavior, so improved behavioral tools will deepen our ability to see and measure microbiome effects. In developmental contexts, relevant behavioral methods include those designed specifically to titrate dyadic interactions, particularly for mammalian species where internal gestation and extended parental care set the stage for continuous, dynamic interactions. Parent-offspring interactions include behaviors that can modulate and shape social behavior. We must not only measure behavior, but also recognize it as an active agent in development.
Parental behaviors are associated with offspring neural and endocrine development and with epigenetic impacts, and they are themselves affected by the offspring. Sibling interactions are similarly influential. Include microbiota effects and the picture is even more functionally integrated. Such interdependencies prevent individual offspring within a litter from serving as statistically independent observations (e.g., Abbey & Howard Reference Abbey and Howard1973).
Hooks et al. define the goal of MGB research as understanding the characteristics of the microbial community in bidirectional regulation with the gut-brain axis and the mechanisms by which such influence occurs. They question whether MGB research is validly advanced by some common approaches, including the use of germ-free (GF) animals. Their characterization of the purpose and use of GF animal models is limited. Sophisticated, experimental use of GF animals is critical and includes testing the impact of specific colonizing microbiota, for example, the microbiota of healthy versus affected individuals on an animal's phenotype, as well as interventions that rescue the phenotype. Controlled colonization of GF animals can enhance study of the tissue-specific role of characterized microbiota – a line of study that Hooks et al. rightly indicate should be better developed in MGB research. Further, they ask: Is the goal of MGB research to identify the impact on neurobehavior of a specific causal organism or a microbial community? In our view, MGB research rightly includes both approaches. That is, while considering the microbial community in its entirety, it is valid and important to determine whether specific members of the microbial community are missing or over-represented in affected individuals compared to healthy controls, and whether this difference can alone explain the phenotype of interest. This direction of inquiry helps refine understanding of mechanism and targets potential therapeutic approaches.
We wholly concur that advances in MGB require application of state-of-the-art measurement using well-validated methods of DNA profiling of microbial communities. Current methods are nicely described by Knight et al. (Reference Knight, Vrbanac, Taylor, Aksenov, Callewaert, Debelius, Gonzalez, Kosciolek, McCall, McDonald, Melnik, Morton, Navas, Quinn, Sanders, Swafford, Thompson, Tripathi, Xu, Zaneveld, Zhu, Caporaso and Dorrestein2018) and involve either amplicon sequencing using a specific region of the bacterial 16S rDNA gene or shotgun sequencing. The former provides rapid, low-cost characterization of bacterial abundance and diversity. Shotgun sequencing deepens the data obtained from each sample by characterizing all genes, whether bacterial, archaeal, viral, or eukaryotic. Further advances are occurring through use of complementary methods that increase insight into what the microbiota are doing (i.e., their functions). For this reason, pairing metagenomic analysis with other -omics, such as metabolomics, metatranscriptomics, or proteomics, is highly recommended and should be incorporated into MGB research whenever possible.
If we are to address disorders of human behavior, we must devise tests that identify a dysregulated and a regulated healthy mouse, so we can better see the shared mechanisms that make a dysregulated and a regulated, healthy person. Despite their valuable observations and critiques, we were disappointed that the authors called for more justification for using data from nonhuman animals to penetrate some of the mysteries of human disorders. They seem to lack appreciation for the power gained from recognizing evolutionary conservation of core mechanisms. Indeed, we think that expanding the use of diverse species into MGB research will contribute importantly to our understanding of behavioral diversity in the natural world (Ezenwa et al. Reference Ezenwa, Gerardo, Inouye, Medina and Xavier2012), as well as enhance our translational insight into MGB phenomena in humans.
No single analysis can adequately cover all facets of a complex interdisciplinary field, but we believe there remain a few topics that warrant inclusion in the present discussion of microbiota-gut-brain (MGB) research. Foremost is developmental analysis of the MGB in parents, offspring, and parent-offspring relations. Development is arguably a necessary element in understanding any biological system. None of the “25 most cited” papers (Table 1 in Hooks et al.) comprise developmental analysis, that is, studies designed to elucidate pathways or processes by which the offspring gut is inoculated with bacteria and in which we see the MGB and behavior emerge and differentiate in interaction with their surroundings. Such analyses would include whether and how the stages of microbiota development stimulate neural, neuroendocrine, and behavioral activity at corresponding stages and that, in turn, reciprocally affect the gut bacteria. Another, important aspect of mammalian development is parent-offspring interaction: whether and how a parent-offspring microbiome may form and especially how the microbiota of each individual affects the other, with routes including behavior that is influenced by individual and familial microbiomes.
We emphasize that performing a manipulation at an early age and measuring effects at a later age or ages does not provide “developmental” information. Likewise, manipulations at various time points can produce different results at different life stages, but again, this says little about process or pathways. Such factors and myriad cascading and interactive events could contribute mechanistically to the development of disordered social behavior.
Another non-deferrable topic is the need for more integrative and informed use of behavior as a measure and as a mechanism. Hooks et al. appropriately identify some of the limitations of the handful of behavioral measures currently used to model social-emotional health and disease. Now is a time to discuss how to validate and apply behavioral measures as more precise and informative metrics of function, especially in social-emotional processes. We need to acknowledge the import and the sophistication of careful behavioral analyses. Needed are methods that are sensitive, objective, and sufficiently nuanced to capture social behavior and the nebulous emotional forces that shape and help regulate it. This need is not unique to MGB investigations; it is widespread in the life sciences, where we seek to understand relations among genetic, cellular, system-level, and organismal functions. The MGB as a system is observed and evaluated via behavior, so improved behavioral tools will deepen our ability to see and measure microbiome effects. In developmental contexts, relevant behavioral methods include those designed specifically to titrate dyadic interactions, particularly for mammalian species where internal gestation and extended parental care set the stage for continuous, dynamic interactions. Parent-offspring interactions include behaviors that can modulate and shape social behavior. We must not only measure behavior, but also recognize it as an active agent in development.
Parental behaviors are associated with offspring neural and endocrine development and with epigenetic impacts, and they are themselves affected by the offspring. Sibling interactions are similarly influential. Include microbiota effects and the picture is even more functionally integrated. Such interdependencies prevent individual offspring within a litter from serving as statistically independent observations (e.g., Abbey & Howard Reference Abbey and Howard1973).
Hooks et al. define the goal of MGB research as understanding the characteristics of the microbial community in bidirectional regulation with the gut-brain axis and the mechanisms by which such influence occurs. They question whether MGB research is validly advanced by some common approaches, including the use of germ-free (GF) animals. Their characterization of the purpose and use of GF animal models is limited. Sophisticated, experimental use of GF animals is critical and includes testing the impact of specific colonizing microbiota, for example, the microbiota of healthy versus affected individuals on an animal's phenotype, as well as interventions that rescue the phenotype. Controlled colonization of GF animals can enhance study of the tissue-specific role of characterized microbiota – a line of study that Hooks et al. rightly indicate should be better developed in MGB research. Further, they ask: Is the goal of MGB research to identify the impact on neurobehavior of a specific causal organism or a microbial community? In our view, MGB research rightly includes both approaches. That is, while considering the microbial community in its entirety, it is valid and important to determine whether specific members of the microbial community are missing or over-represented in affected individuals compared to healthy controls, and whether this difference can alone explain the phenotype of interest. This direction of inquiry helps refine understanding of mechanism and targets potential therapeutic approaches.
We wholly concur that advances in MGB require application of state-of-the-art measurement using well-validated methods of DNA profiling of microbial communities. Current methods are nicely described by Knight et al. (Reference Knight, Vrbanac, Taylor, Aksenov, Callewaert, Debelius, Gonzalez, Kosciolek, McCall, McDonald, Melnik, Morton, Navas, Quinn, Sanders, Swafford, Thompson, Tripathi, Xu, Zaneveld, Zhu, Caporaso and Dorrestein2018) and involve either amplicon sequencing using a specific region of the bacterial 16S rDNA gene or shotgun sequencing. The former provides rapid, low-cost characterization of bacterial abundance and diversity. Shotgun sequencing deepens the data obtained from each sample by characterizing all genes, whether bacterial, archaeal, viral, or eukaryotic. Further advances are occurring through use of complementary methods that increase insight into what the microbiota are doing (i.e., their functions). For this reason, pairing metagenomic analysis with other -omics, such as metabolomics, metatranscriptomics, or proteomics, is highly recommended and should be incorporated into MGB research whenever possible.
If we are to address disorders of human behavior, we must devise tests that identify a dysregulated and a regulated healthy mouse, so we can better see the shared mechanisms that make a dysregulated and a regulated, healthy person. Despite their valuable observations and critiques, we were disappointed that the authors called for more justification for using data from nonhuman animals to penetrate some of the mysteries of human disorders. They seem to lack appreciation for the power gained from recognizing evolutionary conservation of core mechanisms. Indeed, we think that expanding the use of diverse species into MGB research will contribute importantly to our understanding of behavioral diversity in the natural world (Ezenwa et al. Reference Ezenwa, Gerardo, Inouye, Medina and Xavier2012), as well as enhance our translational insight into MGB phenomena in humans.