Association between Gut Microbiota and Infant’s Temperament in the First Year of Life in a Chinese Birth Cohort

Infant temperament characteristics play a critical role in children’s developmental pathways and can predict adulthood psychopathology. The diversity and composition of the gut microbiota are associated with human temperament in both adults and young children. However, the relationship between the gut microbiota and temperament in 12-month-old infants is rarely studied; this developmental period is when temperament reaches a relatively stable stage. We used high-throughput sequencing methods to explore whether temperament characteristics were associated with gut microbiota diversity and composition. Infants’ fecal samples were collected at 12 months of age for the gut microbiota analysis. Based on the primary caregivers’ reports, infants’ temperaments were measured using the Infant Behavior Questionnaire-revised (IBQ-R). This study included 51 infants, including 20 boys and 31 girls, with a mean age of 12.25 months. Results showed that soothability was positively correlated with maternal education level (β = 0.29, p = 0.043, adjust p = 0.025) and the abundance of Bifidobacterium genera (β = 0.62, p = 0.004, adjust p = 0.002). Conversely, cuddliness was negatively correlated with the abundance of Hungatella genera. There was no significant difference in temperament based on gender. This study demonstrated that gut microbiota composition was associated with temperament in 12-month-old infants. These results point to the importance of gut microbiota balance. Future studies on the mechanisms behind the gut microbiota affecting temperament are warranted.


Introduction
Studying temperament is an effective approach to study the differences in individual behavioral characteristics [1]. Based on the widely used Infant Behavior Questionnaire (IBQ) [1] and Infant Behavior Questionnaire-revised (IBQ-R) [2], cultural and sociodemographic modulatory factors could shape the development of temperament in early life. These factors include maternal race, maternal education level, caregivers' psychosocial characteristics, family economic status, numbers of siblings, and feeding type [2][3][4][5]. However, infant temperament differences are hypothesized to be biologically based [6]. In studying infants' temperaments, previous studies were restricted by infants' age, sex, and birth order, and there was less focus on the exploration of the natural, biological mechanisms associated with temperament such as gut microbiota, immune factors, and inflammatory cytokines [2,7,8].
The bidirectional microbiota-gut-brain axis (MGB) connects the gut microbiota with the human brain, and this synergy in turn influences human behaviors [9,10]. The MGB axis [11] includes the

Data Collection
Every maternal participant completed three questionnaires face-to-face, including demographic characteristics (e.g., maternal age, education status, child's sex), infants' early life events (feeding type, probiotic consumption at age 6 months), and an assessment of infants' temperament at an age of 12 months.
Our previous study [23] revealed that early life Clostridium butyricum consumption was associated with infants' gut microbiota during later life. We included the probiotic consumption data for analysis in this study. The newborn consumption of probiotics containing Clostridium butyricum (C. butyricum, at least 75 × 10,000,000 CFU) was to facilitate defecation and prevent jaundice in early neonates (0-1 months after birth). In this study, 42 newborns were prophylactically given probiotics.
The Infants Behavior Questionnaire-Revised (IBQ-R) Chinese version was used to assess infants' temperament at 12 months [2]. The IBQ-R was developed for infants aged 3-12 months old. This measure includes 14 dimensions of temperament that are subdivided into three composite scales: extraversion, negative affectivity, and orienting/regulation. This measure has good reliability and validity [2].
Infants' fecal samples were collected at participants' homes by a trained researcher; samples were taken from infants' diapers according to the Human Microbiome Project (HMP) protocol [24]. Temperament questionnaires were completed at the time of sample collection. After collection, fecal samples were transported in a cooler (+4 • C, 1.5 h on average) to the laboratory and were frozen at −80 • C prior to further analysis. In total, 51 fecal samples were collected for high quality Illumina sequencing. Please refer to our previous study [23] for the processing of DNA extraction, polymerase chain reaction (PCR) amplification, sequencing, and data analysis.

Statistical Methods
We compared infant temperaments between boys and girls using the Mann-Whitney U test. The multivariate analysis of the general liner model (GLM) was used to analyze the significant factors of infant temperament. The Spearman correlation (denoted by r) was used to investigate the association between infant temperament and infant microbiota. We subsequently chose the relevant gut microbiota and prior predictors (e.g., infants' sex and maternal education status) to identify potential associations in temperament using multiple linear regressions, controlling for delivery mode, feeding type, and probiotic consumption.
For the gut microbiota data analysis, the distribution of microbial community was visualized by Circos at the genus level. Principal coordinate analysis (PCoA) was conducted using the OTU information from each sample according to the Bray-Curtis distance matrix (gender: male vs. female, delivery mode: cesarean vs. vaginal, probiotic consumption: yes vs. no, and feeding type: breast vs. mixed). A permutational multivariate analysis of variance (PERMANOVA) was used to analyze the explaining degree of different grouping factors to the sample difference based on OTU level using Bray-Curtis distance matrix. A genus relative abundance ≥ 1% was selected for the abundance analysis.
The data for genera relative abundance were the centered log ratio (CLR) transformed before all analyses. We set the significance level at α = 0.05. Multiple hypothesis tests were adjusted using the Benjamini-Hochberg false discovery rate (FDR), and associations were considered significant below an FDR threshold of 0.05. These analyses were performed using SPSS version 21 (IBM, Chicago, IL, USA) and R [25].

Participant Characteristics
This study included 51 infants (20 boys and 31 girls). These infants had a mean age of 12.3 (mean ± SD: 12.3 ± 0.25) months. In this study, all the mothers were of Han nationality; the mean maternal age at the time of delivery was 32.3 (Mean ± SD: 32.3 ± 4.51) years. All mothers were married (Table 1).

Infant Gut Microbiota at 12 Months Old
In total, 2,541,440 usable sequences and 308 OTUs were obtained from these infants using Illumina MiSeq. Pan analysis indicated that only a small number of new shared phylotypes would be expected with additional sequencing. The value of Good's coverage for all children was 99.6%.
The bacterial distribution was characterized in terms of the relative taxonomic abundances. A total of 10 phyla, 18 classes, 25 orders, 46 families, and 128 genera were obtained from 51 infants' samples at 12 months of age. Figure 1 shows the taxonomic compositions of the dominant bacteria at the genus level. In terms of the relative abundance, 15 genera were above 1%, and the five most abundant genera were Veillonella (22.6%), Bifidobacterium (17.04%), Bacteroides (14.75%), Escherichia-Shigella (7.87%), and the Ruminococcus gnavus group (5.14%).

Figure 2.
Principal coordinate analysis (PCoA) of infant gut microbiota using Bray-Curtis distance matrix. A two-dimensional PCoA was used to describe the relative abundance of infant gut microbiota based on sex, delivery mode, feeding type, and probiotic consumption. Each point represents a single sample and is color-coded into different groups.

Figure 2.
Principal coordinate analysis (PCoA) of infant gut microbiota using Bray-Curtis distance matrix. A two-dimensional PCoA was used to describe the relative abundance of infant gut microbiota based on sex, delivery mode, feeding type, and probiotic consumption. Each point represents a single sample and is color-coded into different groups.

Association between Infant Temperament and the Gut Microbiota
Among 14 dimensions and three composite scales of IBQ-R, six genera were associated with infant temperament (Figure 3). Multiple linear regressions were conducted when infant temperament was identified as the outcome, and both relevant genera and other potential confounding factors (e.g., delivery mode, feeding type, and probiotic consumption) were controlled in the model.

Discussion
Boys and girls showed no significant differences in infant temperament at 12 months old. This finding was consistent with previous studies [7,26,27]. Gartstein [7] explored the stability of infants' temperament by gender in a sample of 315 children from South Korea, which was the only non-Western study, and found no differences in temperament between boys and girls. However, research in Western countries found that temperament ratings differed significantly by infant gender. The first study of gender differences in temperament was measured by IBQ-R in 2003 [2], and this study showed that female infants showed higher scores on the fear subscale and that male infants were rated higher on the activity and high-intensity pleasure subscales. Another study from the United States [20] found significant differences in temperament based on gender. Boys had higher ratings in high-intensity pleasure and motor activation, while girls reported higher ratings for inhibitory control and soothability. Specifically, compared to boys, girls were rated lower on the surgency/extraversion subscale, and girls were rated much higher on the effortful control subscale. Thus, the discrepancy in temperament differences between boys and girls in these various studies could be a result of cultural differences.
In this study, infant soothability in 12-month-olds increased with higher maternal education. Soothability refers to children's reduction of fussing, crying, or distress when caregivers use soothing techniques [2]. In practice, caregivers' soothing techniques are key factors that influence the development of greater soothability in older children. Furthermore, more highly educated mothers may have a more positive view of their parent-child interactions, thus leading to a higher level of intimacy and pleasure for both infants and mothers. Additionally, multiple studies have suggested that maternal education level contributes to positive parent-infant interaction by influencing parenting styles and timing of paternal investment. [28,29]. One study observing 139 Chinese infants aged 12-24 months found that highly educated mothers were more likely to utilize guiding behaviors, and they developed higher levels of intimacy with their infants compared to less-educated mothers [30]. When caring for their children, highly educated mothers tended to lead their children in playing or learning in a variety of ways. The ability to love and feel beloved among highly educated mothers was higher, and these mothers were more likely to perceive that they exerted a positive influence on their children through parenting. Although parenting styles and mothers' responsiveness to children's behaviors vary by culture [31], maternal education level is associated with infants' physical health and psychological behaviors across different cultural contexts; this is consistent with the results of the present study. Behaviors and health outcomes associated with the mother's education level include infants' mental health problems [32], watching TV during meals [33], vocabulary acquisition [34], and cognitive and motor outcomes [35]. Overall, our findings indicated that infants' temperaments were closely related to maternal education level.
This study found that high soothability was associated with a predominant relative abundance of Bifidobacterium in infants' gut microbiota at 12 months old. Meanwhile, the temperament of cuddliness (ability to express enjoyment and the molding of one's body to being held by caregivers) was negatively associated with an abundance of Hungatella. In accordance with our study results, previous studies have shown that the human gut microbiota is associated with temperament. In a study of 60 healthy Korean adults, Kim and Park were the first to promote the existence of temperament [19] differences between enterotypes 1 and 2 in adult populations. Compared to enterotype 2, enterotype 1 had significantly higher scores in novelty seeking and reward dependence behaviors. They suggested that differences in temperament and stability of temperament might be shaped by interactions with the gut microbiota during early life stages (e.g., infancy). Furthermore, one study of 77 children aged 18-27 months old [20] found that gut microbiota diversity and composition were associated with temperament as assessed by the Early Childhood Behavior Questionnaire (ECBQ) [36]; among these children, those with a greater phylogenetic diversity in their gut microbiome exhibited greater surgency/extraversion. Some gender differences were also found; in girls, Rikenellaceae were positively correlated with levels of fear, and in boys, sociability was positively correlated with the abundances of Ruminococcaceae phylum and Parabacteroides. Additionally, there were positive relationships between high-intensity pleasure and Dialister, Rikenellaceae phyla, as well as between activity level and Dialister, Rikenellaceae phyla. In agreement with these previous studies, our research further confirmed the possible effects of the gut microbiota on infant temperament at 12 months old.
Bifidobacterium, which accumulates in large amounts in the human gastrointestinal tract, is an integral part of the human gut microbiota. Multiple studies have verified its remarkable function in immune-modulation [37], obesity [38], infection [39], balance of gut microbiota [40], and especially in mental health disorders [41]. In fact, changes in emotional processing modulated by the gut microbiota have been established by a number of observational studies on probiotic consumption. One study in mice [42] suggested that Bifidobacterium longum 1714 and Bifidobacterium breve 1205 supplements were effective treatments for decreasing anxiety-like behaviors, and they were as effective as the anti-anxiety medication escitalopram. Additionally, a study in rats [43], using an electric shock model, found that consumption of Bifidobacterium longum R0175 over a 30-day period led to a significant decrease in anxiety. In animal models of Alzheimer's disease [44], it was observed that Bifidobacterium breve treatment may improve cognitive function and counteract some degree of impairment. In a human study, the consumption of Bifidobacterium was shown to effectively improve depression-anxiety conditions, increase defecation frequency, and decrease body mass index in healthy elderly people [45]. Similarly, supplementation with the prebiotic Bifidobacterium longum 1714 resulted in an improvement of the frontal midline electroencephalographic mobility and hippocampus-dependent memory performance in healthy volunteers [46]. Together, these studies suggest that the human anxiety behaviors, depression behaviors, and cognitive function were altered by Bifidobacterium species. This is further supported by our results finding high soothability in infants to be associated with significantly dominant Bifidobacterium in their gut microbiome.
Hungatella is a functional genus that carries genes related to polysaccharide metabolism and is involved in the production of acetate and propionate. Numerous studies have proved that it is responsible for enhancing efficient energy harvesting and gastrointestinal mobility [47,48]. Additionally, in patients experiencing constipation, Hungatella was significantly more abundant compared to healthy controls [49]. Similarly, work in patients with Parkinson's disease [50] also found that there was a significant positive correlation between the relative abundance of Hungatella and patients' constipation, which was independent of their Parkinson's disease. Meanwhile, there has been far less research on the association between mental behaviors and Hungatella. In an observational study of Chinese navy soldiers, it was found [51] that after a long voyage mission, the abundance of Hungatella in the gut microbiota of soldiers had significantly decreased. These findings opened exciting possibilities for exploring the relationship between an increased abundance of Hungatella and lower cuddliness in infants. It is important to explore the relationships between Hungatella and personality or temperament-related behaviors.
There were some limitations that reduced the generalizability of these study findings. The full study provided a small-scale test of associations between infants' gut microbiota and temperament at the age of 12 months. Additionally, the majority of infants in our study were taking probiotics at the time of sample collection, which may affect the generalizability of the study. Our understanding of these connections could be strengthened by exploring the gut microbiota as it relates to temperament in older children. Future research should replicate this study with a larger sample, using a longitudinal study design to explore the link between infant temperament and gut microbiota and its impact on children's later life. In addition, this study did not address the role of diet quantity or quality at the time of stool sample collection. Although parents fed their children according to standardized guidelines for infants' diets, the potential mechanisms of food quantity may account for some of the differences in the associations.

Conclusions
This study examined the association between infants' temperament and the gut microbiota at the age of 12 months. We found that soothability was positively correlated with the abundance of Bifidobacterium and maternal education level. There was a negative correlation between cuddliness and an abundance of Hungatella. This study contributed to our understanding of microbiota-behavior/physiology interactions and confirmed that this relationship is stable and robust by 12 months of age. It also provides a new approach to studying behavioral characteristics in teens, which should start in infancy. We should emphasize the intestinal health and gut microbiota balance before the transition to adult foods. The deficiencies of the study include infants' probiotic consumption and the small sample size, which somewhat limits the study's conclusions; consequently, further research in this area may be warranted. From the MGB viewpoint, this study was not designed to address detailed causal interaction pathways between the gut microbiota and temperament. Thus, future research should explore the possibility of incorporating metagenomics and metabolomics and analyzing these associations to understand the causes behind the appearance of this new evidence in a large infant cohort.