Gut Microbiota and Fear Processing in Women Affected by Obesity: An Exploratory Pilot Study

Abstract

The microbiota-gut-brain axis extends beyond visceral perception, influencing higher-order brain structures, and ultimately psychological functions, such as fear processing. In this exploratory pilot study, we attempted to provide novel experimental evidence of a relationship between gut microbiota composition and diversity, and fear-processing in obesity, through a behavioral approach. Women affected by obesity were enrolled and profiled for gut microbiota, through 16S rRNA amplicon sequencing. Moreover, we tested their ability to recognize facial fearful expressions through an implicit-facial-emotion-recognition task. Finally, a traditional self-report questionnaire was used to assess their temperamental traits. The participants exhibited an unbalanced gut microbiota profile, along with impaired recognition of fearful expressions. Interestingly, dysbiosis was more severe in those participants with altered behavioral performance, with a decrease in typically health-associated microbes, and an increase in the potential pathobiont, Collinsella. Moreover, Collinsella was related to a lower expression of the persistence temperamental trait, while a higher expression of the harm-avoidance temperament, related to fear-driven anxiety symptoms, was linked to Lactobacillus. Once confirmed, our findings could pave the way for the design of innovative microbiome-based strategies for the treatment of psychological and emotional difficulties by mitigating obesity-related consequences and behaviors.

Keywords: dysbiosis; facial emotion recognition; fear; gut microbiota; obesity; temperament.

Figure 1

Behavioral results of the experimental task. For each experimental condition (bilateral incongruent emotional, bilateral incongruent neutral, bilateral congruent, and unilateral), the reaction time in milliseconds (upper panel), and the level of accuracy in percentage (lower panel), are shown. The minimum, lower quartile, median, upper quartile, maximum, and outliers are displayed.

Figure 2

Phylum- and family-level composition of the gut microbiota of women affected by obesity, compared to normal-weight women. Relative abundance profiles of the gut microbiota of women affected by obesity (PTS) compared to age-matched, normal-weight women from the same geographical area (HC), at phylum (A) and family (B) level. For each panel, left: bar graphs of the individual profiles; right: pie charts showing average values.

Figure 3

Diversity and taxonomic signatures of the gut microbiota of women affected by obesity, compared to normal-weight women. (A) Boxplots showing the distribution of alpha diversity, according to the inverse Simpson and Shannon index, in women affected by obesity (PTS), compared to age-matched, normal-weight women from the same geographical area (HC). A lower value was found in participants (p ≤ 0.04, Wilcoxon test). (B) PCoA plot of beta diversity, based on Bray–Curtis dissimilarity between the genus-level profiles. A separation between groups was found (p = 0.0005, PERMANOVA). Samples were identified with colored dots, as in A. Ellipses include 95% confidence area based on the standard error of the weighted average of the sample coordinates. (C) Boxplots showing the relative abundance distribution of differentially represented taxa between groups. * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001, Wilcoxon test.

Figure 4

Scatter plots of correlation between relative taxon abundances and experimental task parameters, temperamental traits, and anthropometric and biochemical variables in women affected by obesity. Only statistically significant correlations (p ≤ 0.05), based on the Kendall rank correlation test, are shown for experimental task parameters (A), temperamental traits (B), and anthropometric and biochemical variables (C). BMI, body mass index.

Figure 5

Variation of the gut microbiota of women affected by obesity in relation to experimental task parameters. Study participants (PTS) were stratified by high or low value of each behavioral parameter ((A), reaction time (RT); (B), level of accuracy (ACC)), and compared to age-matched, normal-weight women (HC). For each panel, left: boxplots showing the distribution of alpha diversity, according to the inverse Simpson and Shannon index, in the study groups; center: PCoA plot of beta diversity, based on the Bray–Curtis dissimilarity between the genus-level profiles, with ellipses including 95% confidence area based on the standard error of the weighted average of sample coordinates; right: boxplots showing the relative abundance distribution of differentially represented taxa between groups. An asterisk next to the family name indicates unclassified genera. * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001, ^# p ≤ 0.1, Wilcoxon test.

Figure 6

Variation of the gut microbiota of women affected by obesity in relation to temperamental traits. Study participants (PTS) were stratified by T-scores in high, medium or low expression for each of the temperamental traits ((A), reward dependence (RD); (B), persistence (P); (C), harm avoidance (HA); (D), novelty-seeking (NS)), and compared to age-matched, normal-weight women from the same geographical area (HC). For each panel, top left: boxplots showing the distribution of alpha diversity, according to the inverse Simpson and Shannon index, in the study groups; top right: PCoA plot of beta diversity, based on Bray–Curtis dissimilarity between the genus-level profiles, with ellipses including 95% confidence area based on the standard error of the weighted average of sample coordinates; bottom: boxplots showing the relative abundance distribution of differentially represented taxa between groups. * p ≤ 0.05, ** p ≤ 0.01, ^# p ≤ 0.1, Wilcoxon test.

Figure 6

Variation of the gut microbiota of women affected by obesity in relation to temperamental traits. Study participants (PTS) were stratified by T-scores in high, medium or low expression for each of the temperamental traits ((A), reward dependence (RD); (B), persistence (P); (C), harm avoidance (HA); (D), novelty-seeking (NS)), and compared to age-matched, normal-weight women from the same geographical area (HC). For each panel, top left: boxplots showing the distribution of alpha diversity, according to the inverse Simpson and Shannon index, in the study groups; top right: PCoA plot of beta diversity, based on Bray–Curtis dissimilarity between the genus-level profiles, with ellipses including 95% confidence area based on the standard error of the weighted average of sample coordinates; bottom: boxplots showing the relative abundance distribution of differentially represented taxa between groups. * p ≤ 0.05, ** p ≤ 0.01, ^# p ≤ 0.1, Wilcoxon test.

Figure 6

Variation of the gut microbiota of women affected by obesity in relation to temperamental traits. Study participants (PTS) were stratified by T-scores in high, medium or low expression for each of the temperamental traits ((A), reward dependence (RD); (B), persistence (P); (C), harm avoidance (HA); (D), novelty-seeking (NS)), and compared to age-matched, normal-weight women from the same geographical area (HC). For each panel, top left: boxplots showing the distribution of alpha diversity, according to the inverse Simpson and Shannon index, in the study groups; top right: PCoA plot of beta diversity, based on Bray–Curtis dissimilarity between the genus-level profiles, with ellipses including 95% confidence area based on the standard error of the weighted average of sample coordinates; bottom: boxplots showing the relative abundance distribution of differentially represented taxa between groups. * p ≤ 0.05, ** p ≤ 0.01, ^# p ≤ 0.1, Wilcoxon test.