Study finds a correlation between early infant gut microbiota composition and temperament traits in toddlers

Scientists in the FinnBrain research project of the University of Turku, Finland, discovered that the gut microbes of a 2.5-month-old infant are associated with the temperament traits manifested at six months of age.

Temperament describes individual differences in expressing and regulating emotions in infants, and the study provides new information on the association between behaviour and microbes.

A corresponding study has never been conducted on infants so young or on the same scale.

Rodent studies have revealed that the composition of gut microbiota and its remodelling is connected to behaviour.

In humans, gut microbes can be associated with different diseases, such as Parkinson’s disease, depression and autism spectrum disorders, but little research has been conducted on infants.

Doctoral Candidate, Doctor Anna Aatsinki from the FinnBrain research project at the University of Turku, Finland, discovered in her study on 303 infants that different temperament traits are connected with individual microbe genera, microbial diversity and different microbe clusters.

“It was interesting that, for example, the Bifidobacterium genus including several lactic acid bacteria was associated with higher positive emotions in infants.

Positive emotionality is the tendency to experience and express happiness and delight, and it can also be a sign of an extrovert personality later in life,” says Aatsinki.

Temperament Can Predict Later Development

One of the findings was that greater diversity in gut bacteria is connected to lesser negative emotionality and fear reactivity.

The study also considered other factors that significantly affect the diversity of the microbiota, such as the delivery method and breastfeeding.

The newly published study is first of its kind and gives evidence of an association between normal variation in behaviour and gut microbiota. The image is in the public domain.

The findings are interesting as a strong fear reaction and negative emotionality can be connected to depression risk later in life.

However, the association with later diseases is not straightforward and they are also dependent on the environment.

“Although we discovered connections between diversity and temperament traits, it is not certain whether early microbial diversity affects disease risk later in life. It is also unclear what are the exact mechanisms behind the association,” adds Aatsinki.

“This is why we need follow-up studies as well as a closer examination of metabolites produced by the microbes.”

The newly published study is first of its kind and gives evidence of an association between normal variation in behaviour and gut microbiota.

This might indicate that there is activity on the gut-brain axis already in infancy.

The gut-brain axis is a functional relationship between gut flora and central nervous system via the nervous, immune and endocrine systems.

The FinnBrain research project of the University of Turku studies the combined influence of environmental and genetic factors on the development of children. Over 4,000 families participate in the research project and they are followed from infancy long into adulthood. The newly published study is part of FinnBrain’s gut-brain axis sub-project led by Adjunct Professor, Child and Adolescent Psychiatrist Linnea Karlsson.

---

Animals harbor a diverse community of microbes, in their gut and in almost any other site, both on and within, their bodies. Host and gut microbiota (GM) interact symbiotically (Sommer and Bäckhed, 2013).

The GM contribute to host health and fitness, playing a major role in diverse host functions including development, fecundity, metabolism and immunity; for the microbes, animal intestines are a favorable niche (Shapira, 2016).

By removing the need for culturing microorganisms for identification, next-generation sequencing methods have massively advanced the characterization of the human GM and are continuously expanding our knowledge about endosymbiotic microbes.

We have just recently grasped that the human GM comprise hundreds of microbial species, with Firmicutes and Bacteroidetes as the dominant phyla (Lozupone et al., 2012), and that microbial life may thus have roles in multiple physiological processes including those related to mental health (Cryan and Dinan, 2012).

This has opened the possibility of applying a “gestalt perspective” allowing us to understand physiological, behavioral and cognitive processes as part of an integrated whole (Koffka, 2013).

Recent technological developments actually allow for the integration of data from various sources such as the genome, transcriptome, proteome, epigenome and microbiome in what has been termed “Gestaltomics”, as a useful approach to the understanding of psychiatric disorders at different levels of organization (Gutierrez Najera et al., 2017).

Aware of the perils of our inference of adaptive significance from proximate control of behavior (Dewsbury, 1999), our review proceeds as follows: in view of the accumulating evidence linking biological processes between micro- and macroorganisms, first, we suggest that the use of the concept of a holobiont as a unit of selection should be applied to the conglomerate of organisms involved in such relationship.

Second, we describe the neurobiology of social behavior highlighting the possible pathways through which microbiota and, particularly, GM may affect social behavior, including macroorganisms’ development. Even under the light of recent and excellent reviews on the topic (Ezenwa et al., 2012; Archie and Tung, 2015), knowledge of the mechanisms by which microbiota in different sites of a host’s body influence behavior is still lacking.

In contrast, there are at least three purported pathways suggesting how the GM interact with an individual’s central nervous system (CNS), modifying behavior in general, and social interactions in particular (Sampson and Mazmanian, 2015).

For this reason, we focus our review on GM and expand this perspective by describing three areas where social behavior can in turn influence individual microbial profiles: (1) due to stressing events in hosts’ social life; (2) because of differences between solitary and social life; and (3) due to social structure.

Figure ​1 introduces to key elements of these interactions.

GM Influence on Development and Social Behavior

In humans, GM establishment during early life occurs primarily through vertical transmission from mother to offspring (Nuriel-Ohayon et al., 2016).

Recent studies defy the long-held dogma that the intrauterine environment is sterile and newborns are germ-free, suggesting that microbiota is transferred from mother to fetus (Reviewed in Perez-Muñoz et al., 2017).

Nonetheless, these discoveries have been strongly contested, and the “sterile womb” paradigm, proposing that the sterile fetus first acquires an early microbiome during and after birth, prevails (Perez-Muñoz et al., 2017).

All the same, the mother’s GM, as well as vaginal microbiota, are known to be dramatically remodeled during pregnancy (Nuriel-Ohayon et al., 2016), supporting the suggestion that exposure to microbial metabolites and compounds originating from the maternal gut play an important role in offspring’s development (Gomez de Agüero et al., 2016; Perez-Muñoz et al., 2017). During vaginal delivery, microbiota from the maternal vagina and gut inoculate the newborn’s GM (Penders et al., 2006; Dominguez-Bello et al., 2010).

In contrast, the GM of newborns delivered by C-section rather resemble maternal skin and oral microbiota (Penders et al., 2006; Dominguez-Bello et al., 2010)

During infancy, the GM increase their complexity via eating and uptake of microbes from the environment (Lozupone et al., 2012).

Thus, the specific GM of early life converge toward adult-like GM around the age of three (Yatsunenko et al., 2012).

Perinatal exposure to the maternal GM apparently plays an important role in the establishment of a “pioneer” infant GM, with major implications for infant brain development (Degroote et al., 2016).

Human cohort studies reported an increased risk of autism spectrum disorder (ASD) in infants from obese mothers (Connolly et al., 2016) and from mothers who received antibiotic treatment during pregnancy (Atladóttir et al., 2012).

Studies in rodent models consistently demonstrated that offspring exposed prenatally to maternal high fat diet or antibiotics showed impaired sociality (Buffington et al., 2016). Buffington et al. (2016) also showed that the social impairment was mediated by mouse-pup GM, which differed between offspring from mothers fed a high fat diet and offspring from mothers fed a regular diet.

Transfer of GM of offspring from mothers fed a regular diet to offspring from mothers fed a high fat diet restored normal social behavior at weaning (4 weeks), but not in adulthood (8 weeks).

Furthermore, offspring from mothers fed a HFD had reduced hypothalamic oxytocin levels; treatment of these offspring with Lactobacillus reuteri, the most drastically reduced strain of their GM, increased hypothalamic oxytocin levels and normalized their social deficit. T

his suggests that L. reuteri improves social behavior by promoting oxytocin-mediated functions (Buffington et al., 2016). L. reuteri treatment also improved wound healing, supposedly by a vagally-mediated increase of oxytocin (Poutahidis et al., 2013).

Recent findings from the same group suggest that L. reuteri viability is not essential for the regulation of oxytocin; thus, a peptide or metabolite produced by these bacteria may be sufficient (Varian et al., 2017).

Degroote et al. (2016) observed that Wistar rat offspring exposed periconceptionally to antibiotics spent 50% less time in social interactions than controls.

Leclercq et al. (2017) found that a low dose of penicillin administrated to mice dams from the last week of pregnancy to weaning of the pups (perinatally), reduced sociality and preference for social novelty in the offspring.

Some males exposed perinatally to antibiotics exhibited a surprising aggressive behavior when physically stressed by an unfamiliar male aggressor (defeat paradigm), different from control mice, who exhibited a submissive posture (Leclercq et al., 2017).

Perinatally antibiotics-treated male and female mice featured a substantially increased expression of vasopressin receptor 1b (Leclercq et al., 2017), known to be involved in social and aggressive behaviors (Wersinger et al., 2002), in the frontal cortex. Interestingly, concurrent supplementation with the probiotic strain Lactobacillus rhamnosus (JB-1), which supposedly regulates emotional behavior in healthy mice via the vagus nerve (Bravo et al., 2011), counteracted some of the antibiotic treatment effects.

JB-1 supplementation concurrently to antibiotic treatment in fact prevented the decrease in sociality and social novelty preference in offspring, i.e., there was no significant difference between offspring of control or antibiotic treated dams in the defeat paradigm.

The GM of offspring perinatally exposed to antibiotic or antibiotic + JB-1 were largely altered and clustered separately from the control group.

Another recent study also demonstrated that JB-1 treatment decreased stress-induced anxiety-like behavior and prevented deficits in social interaction (Bharwani et al., 2017).

These findings warrant further investigation of the influences of Lactobacillus strains on social behavior and their potential to prevent or attenuate features of social impairments.

Antibiotic depletion of mice GM in another neurodevelopmental window, adolescence, was not associated with decreased sociality, but with impaired social memory (Desbonnet et al., 2015).

This correlated with reduced mRNA levels of oxytocin and vasopressin in the hypothalamus (Desbonnet et al., 2015). GM also modulate myelination in the prefrontal cortex. Hoban et al. (2016) found abnormal hypermyelinated axons in male germ-free mice.

Genes involved in myelination and myelin plasticity were upregulated specifically in the prefrontal cortex, which was paralleled by increased protein levels, and thicker myelin sheaths in the prefrontal cortex of germ-free mice.

This coincided with an upregulation of neural activity-induced pathways (Hoban et al., 2016).

In male germ-free mice bacterially colonized post weaning (postnatal day 21), none of these genes were differentially regulated, suggesting a dynamic influence of host GM on myelin-related and activity-induced genes. Increased myelin protein abundance, however, could not be reversed (myelin formation in mice occurs around postnatal day 10; Hoban et al., 2016).

Although host genetics can influence GM (Stewart et al., 2005), adult monozygotic twins do not have more similar GM than adult dizygotic twins, and genetically unrelated cohabiting partners have more similar GM than unrelated individuals (Turnbaugh et al., 2009; Song et al., 2013).

This emphasizes the importance of the environment, including diet, drinking water, sanitation, hygiene and antibiotics, in shaping the GM after early life events (Martínez et al., 2015).

GM differences among societies or communities could reflect particularities in the exposure to such factors (Martínez et al., 2015) while social structure and behavior may determine the flow of microbes among individuals (i.e., horizontal transmission, the transmission of endosymbionts from one individual to another; Song et al., 2013). In this sense, natural selection may have favored GM promoting their transmission via social interactions (Stilling et al., 2014).

Many of the currently available studies supporting the hypothesis of an influence of the GM on social behavior have been conducted in germ-free mice.

The germ-free mouse model presents the major advantage of in proof-of-principle studies, as well as the possibility of introducing certain microbiota or a defined bacterial consortium at various time points of host development (Luczynski et al., 2016).

It is worth noting, however, that the possibility of translating such studies is limited, as no equivalent condition exists in wild mammals or humans. Furthermore, upbringing of germ-free mice may induce irreversible neurodevelopmental deficits, along with a range of other impairments that may limit the suitability of the model for specific scientific queries.

Also, studies in a germ-free model do not allow disentangling cause and effect; in non-germ-free laboratory mice, other model organism or wild animals’ changes in behavior and/or in the brain may influence the type of bacteria present in the gut.

The alternatives used, i.e., antibiotic treatment and probiotic feeding, as in some of the other studies presented, can be regarded as potentially more relevant with respect to translation than the use of germ-free mice.

Further alternatives include fecal transplantation and mouse humanization (Cryan and Dinan, 2012). Nevertheless, all these findings suggest that the GM is critical for the development and modulation of the neurobiological substrate of social behavior, and that specific microbial strains might promote host social behavior.

They further suggest that the prenatal and postnatal periods are the most critical neurodevelopmental windows for GMs’ influence on social behavior. Thus, microbial replenishment until adolescence might, to some extent, rescue social deficits based on GM dysbiosis (Table ​(Table1).1).

These deficits may include transitory, yet incapacitating, mental states associated with dysbiosis. For instance, symptoms of long-term depression in rodents can be facilitated by the blocking of cellular endocannabinoid uptake (Gerdeman et al., 2002), which can take place during antibiotic-induced dysbiosis.

Resembling the lack of motivation found in subjects with bowel disorders, dysbiosis can result in impaired sociality and depression-like symptoms due to neurochemical and functional modifications in the hippocampus; remarkably all could be reversed by administering a probiotic, leading to a normalization of such neurochemical and behavioral modifications (Guida et al., 2018).

Table 1

Perturbation of the gut microbiota (GM) can affect social behavior in rodent models.

Model/intervention	Effects on social behavior	Neural correlates	Reference
Prenatal period
Maternal immune activation	Offsprings exhibited reduced sociability and reduced preference for social novelty	–	Hsiao et al. (2013)
Maternal high fat diet	Offsprings had fewer social interactions, exhibited reduced sociability and reduced prefeference for social novelty	Reduced oxytocin levels in the hypothalamus	Buffington et al. (2016)
Maternal antibiotic treatment	Offsprings had fewer and shorter social interactions	–	Degroote et al. (2016)
Maternal antibiotic treatment (1 week before delivery to 3 weeks after delivery)	Offsprings exhibited reduced sociability and reduced preference for social novelty. Male offsprings exhibited increased aggressive behavior.	Increased mRNA expression of arginine vasopressin receptor 1b in the frontal cortex	Leclercq et al. (2017)
Maternal antibiotic treatment (1 week before delivery to 3 weeks after delivery)	Prevented decrease in sociability and preference for social novelty in offsprings. No effect on male offsprings aggressivity observed.	Non significant (p = 0.1) trend of decreased mRNA expression of arginine vasopressin receptor 1b in the frontal cortex compared to offsprings exposed to antibiotics only.	Leclercq et al. (2017)
Post natal period
Germ-free mice	Increased sociability and increased preference for social novelty	–	Arentsen et al. (2015)
Germ-free mice	Reduced sociability and reduced preference for social novelty	–	Desbonnet et al. (2014)
Childhood and adolescence
Bacterial colonization of socially impaired germ-free mice (at 3 weeks)	Restores sociability but not preference for social novelty, suggesting impaired social memory	–	Desbonnet et al. (2014)
Colonization of socially impaired mice with healthy mice GM (at 4 weeks)	Restores sociability and preference for social novelty	–	Buffington et al. (2016)
Probiotic administration (Lactobacillus reuteri) to socially impaired mice (at 4 weeks)	Restores sociability and preference for social novelty	Enhanced oxytocin levels in the hypothalamus	Buffington et al. (2016)
Antibiotic treatment (from 3 weeks onwards)	Normal sociality, reduced social memory	Reduced oxytocin and vasopressin levels in the hypothalamus	Desbonnet et al. (2015)
Adulthood
Colonization of socially impaired mice with healthy mice GM (at 8 weeks)	Fails to restore sociability and preference for social novelty	–	Buffington et al. (2016)

Other findings in rodent models are consistent with knowledge of human brain development. Neurogenesis and neural migration occur prenatally.

Synaptogenesis and glycogen synthesis start before birth and continue postnatally, with synaptic density reaching its maximum at 2 years of age (Borre et al., 2014).

Furthermore, primates (including humans) show a late and prolonged postnatal neurodevelopment, showing sensitivity to environmental insults (e.g., a chemically-driven developmental impairment) up to the end of adolescence (Borre et al., 2014; Morin et al., 2017), as prefrontal cortex maturation concludes; prefrontal cortex continues its development up until 20 years of age (Marín, 2016).

The ASD is usually characterized by pronounced disturbances of social behavior, and much of the evidence for a connection between GM and social behavior arose from investigations of this malady in human epidemiological studies and from biomedical studies in rodent models.

Individuals with ASD present deficits in social communication and interactions, along with stereotypic behavior. Several studies reported that ASD patients have an altered GM composition (Vuong and Hsiao, 2017) and a higher prevalence of inflammatory bowel disease and other gastrointestinal disorders compared to controls (Doshi-Velez et al., 2015). Hsiao et al. (2013) showed that the gastrointestinal symptoms also co-occurred with symptoms in the CNS in an ASD mouse model.

They further demonstrated that treatment with the gut bacterium Bacteroides fragilis ameliorated GM dysbiosis, corrected gastrointestinal abnormalities, and improved some of the autism-associated behavioral impairments, although deficits in sociability and social preference remained. Desbonnet et al. (2014) demonstrated that germ-free animals exhibited reduced sociability and reduced preference for social novelty. Whereas post-weaning bacterial colonization of germ-free mice reversed social avoidance, it had no effect on social memory impairment (Desbonnet et al., 2014).

Findings in a similarly designed study, however, deviated from those of Desbonnet et al. (2014) as germ-free mice exhibited increased sociability and increased preference for social novelty (Arentsen et al., 2015).

Interestingly, ASD is typically diagnosed before 2 years of age (Marín, 2016). From then until the end of adolescence, the brain undergoes a process of neurodevelopmental reorganization by synaptic pruning.

This makes it vulnerable to environmental deficiencies, including malnutrition (Morin et al., 2017) and GM dysbiosis.

Thus, adolescence is a critical period for the onset of several neuropsychiatric disorders including schizophrenia, depression and obsessive-compulsive disorder (Marín, 2016).

Another exploratory open label study evaluating an investigative microbial transfer in 18 children with ASD (7–16 years) yielded promising results: the treatment produced significant improvements in both gastro-intestinal and autism-related symptoms, and the GM composition of the ASD children approached that of neurotypical children (Kang et al., 2017).

Perhaps, one of the most significant social impairments may be that of social isolation, often worst endured by the elderly (Weldrick and Grenier, 2018).

Evidence suggests, however, that elders who interacted more often with the people in their communities have GM resembling those of younger individuals (Kinross and Nicholson, 2012). As described above, the GM is modified throughout the life cycle.

The GM of children lack the complexity found in adult individuals, whereas advancing age has been associated with lower proportions of bifidobacteria and higher proportions of bacteroides (Hopkins et al., 2002).

While infant’s GM are represented by C. leptum and C. coccoides, those of elderly individuals show higher proportions of E. coliand Bacteroidetes, suggesting an evolving ratio of Firmicutes to Bacteroidetes across the life cycle (Mariat et al., 2009). Within Firmicutes, significant variations in the butyrate producing genera have been found in elders, including Actinobacteria, Feacalibacterium and Proteobacteria, butyrate-producing taxa providing a major energy source to the intestinal epithelium (O’Toole, 2012).

Because of its close relationship with immunity, differences in the composition of GM across different ages may also contribute to the progression of the frailty and poor health observed in old age (Biagi et al., 2010).

Therefore, aging may be characterized by a reduction of core GM, accompanied by increments in subdominant and pro-inflammatory species, but also by an enrichment of other health-related bacteria (i.e., Akkermansia, Bifidobacterium, or Christensenellaceae) that may give extreme elders (>104 years old) some kind of “longevity adaptation” (Biagi et al., 2016).

---

Source:
University of Turku
Media Contacts: 
Anna Aatsinki – Neuroscience News
Image Source:
The image is in the public domain.

Original Research: Open access
“Gut microbiota composition is associated with temperament traits in infants”. Anna Aatsinki et al.
Brain, Behavior, and Immunity. doi:10.1016/j.bbi.2019.05.035