Resilience or susceptibility to traumatic stress: Potential influence of the microbiome

Abstract

Exposure to traumatic stress is a major risk factor for development of neuropsychiatric disorders in a sub-population of individuals, while others remain resilient. The mechanisms and contributing factors differentiating between these phenotypes are still unclear. We hypothesize that inter-individual differences in the microbial composition and function contribute to host resilience or susceptibility to stress-induced psychopathologies. The current study aimed to characterize gut microbial community before and after exposure to traumatic stress in an animal model of PTSD. Sprague-Dawley male rats were randomly divided into unstressed controls and experimental group subjected to Single Prolonged Stress (SPS). After 14 days, behavioral analyses were performed using Open Field, Social Interaction and Elevated Plus Maze tests. Based on the anxiety measures, the SPS group was further subdivided into resilient (SPS-R) and susceptible (SPS-S) cohorts. The animals were sacrificed after the last behavioral test and cecum, colon, hippocampus, and medial prefrontal cortex were dissected. Prior to SPS and immediately after Open Field test, fecal samples were collected from each rat for 16S V3-V4 ribosomal DNA sequencing, whereas urine samples were collected before SPS, 90 min into immobilization and on the day of sacrifice to measure epinephrine and norepinephrine levels. Analyses of the fecal microbiota revealed significant differences in microbial communities and in their predictive functionality among the groups before and after SPS stressors. Before SPS, the SPS-S subgroup harbored microbiota with an overall pro-inflammatory phenotype, whereas SPS-R subgroup had microbiota with an overall anti-inflammatory phenotype, with predictive functional pathways enriched in carbohydrate and lipid metabolism and decreased in amino acid metabolism and neurodegenerative diseases. After SPS, the gut microbial communities and their predictive functionality shifted especially in SPS cohorts, with volatility at the genus level correlating inversely with Anxiety Index. In line with the alterations seen in the gut microbiota, the levels of cecal short chain fatty acids were also altered, with SPS-S subgroup having significantly lower levels of acetate, valerate and caproate. The levels of acetate inversely correlated with Anxiety Index. Interestingly, urinary epinephrine and norepinephrine levels were also higher in the SPS-S subgroup at baseline and during stress, indicative of an altered sympathoadrenal stress axis. Finally, shorter colon (marker of intestinal inflammation) and a lower claudin-5 protein expression (marker for increased blood brain barrier permeability) were observed in the SPS-S subgroup. Taken together, our results suggest microbiota is a potential factor in predisposing subjects either to stress susceptibility or resilience. Moreover, SPS triggered significant shifts in the gut microbiota, their metabolites and brain permeability. These findings could lead to new therapeutic directions for PTSD possibly through the controlled manipulation of gut microbiota. It may enable early identification of individuals more likely to develop prolonged anxiogenic symptoms following traumatic stress.

Keywords: Anxiety; Gut microbiota; Short chain fatty acids; Single prolonged stress; Stress resilience; Urinary catecholamines.

Graphical abstract

Fig. 1

Experimental design. The animals were allowed to accommodate to the animal facility for 14 days upon arrival. On day 15, they were randomly assigned into unstressed control or exposed to SPS group. The SPS group was left undisturbed for 7 days, after which they were kept with normal bedding changes for the remainder of the experiment. On day 31 the controls and the SPS group were exposed to battery of behavioral tests in the following order: Open Field (OF), Social Interaction (SI), and Elevated Plus Maze (EPM). On day 39, the animals were sacrificed by decapitation and different organs were collected. Animals were divided into SPS-R and SPS-S subgroups based on their performance on OF and EPM tests. Stool samples were collected before SPS and immediately after OF test, and urine samples were collected before SPS, 90 min into immobilization and on the day of sacrifice.

Fig. 2

Effect of SPS on anxiety-like behavior measured by Open Filed (OF). Fourteen days after SPS, the animals were tested for: (A) Time spent in the center of arena, and (B) Number of entries into the center of the arena. Unpaired t-test was performed for comparison of the means. After being subdivided into SPS-R and SPS-S subgroups: (C) Time spent in the center of and (D) Number of entries into the center by SPS subgroups. Number of entries into the center by SPS subgroups data didn't pass the normality test and were analyzed using Kruskal-Wallis test followed by Dunn's multiple comparisons test. Time spent in the center by SPS subgroups data passed the normality test and were analyzed using one way-ANOVA, followed by Tukey's multiple comparisons test. Each dot represents value for an individual animal. Blue-Controls, green-SPS-R, red-SPS-S. All data are expressed as means ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 3

Effect of SPS on anxiety-like behavior measured by Elevated Plus Maze (EPM) Animals were tested on the EPM 23 days after SPS. (A) Percent duration in open arm (OA), (B) Percent number of entries into the OA, (C) Anxiety Index. The EPM data passed the normality test and were analyzed using one way-ANOVA, followed by Tukey's multiple comparisons test. Each dot represents value for an individual animal. Blue-Controls, green-SPS-R, red-SPS-S. All data are expressed as means ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 4

Effect of SPS on social impairment measured by Social Interaction (SI) Test. Social interaction test was performed 15 days after SPS. (A) Duration engaged in social interaction and (B) number of approaches/interactions initiated by the test rat. The SI data passed the normality test and were analyzed using one way-ANOVA, followed by Tukey's multiple comparisons test. Each dot represents value for an individual animal. Blue-Controls, green-SPS-R, red-SPS-S. All data are expressed as means ± SEM. *p < 0.05. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 5

Differences in gut microbial alpha and beta diversities among the groups. The fecal 16S sequencing was used to determine the microbial composition of each group before and after SPS. (A) Alpha diversity measured by Inverse Simpson Index, (B) PCA plot of Beta diversity before SPS measured by Aitchison distance, (C) PCA plot of Beta diversity after SPS measured by Aitchison distance, (D) Differences in volatility among the groups, (E) Correlation between volatility and Anxiety Index, (F) Correlation between volatility and % time spent in the closed arms of EPM, (G) Correlation between volatility and % time spent in the open arms of EPM. Inverse Simpson before and after SPS was analyzed by Two-ways ANOVA, mixed effects, followed by Šídák's and/or Tukey's multiple comparisons test. FDR was used to correct between-tests p values. For PCA plots, data points were projected into the space spanned by the first two principal components. Correlations were performed using Pearson's correlation. Blue-Controls, Green-SPS-R, Red-SPS-S. All data are expressed as means ± SEM. *p < 0.05. Values, 2 SD away from the mean were excluded from analysis. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 5

Differences in gut microbial alpha and beta diversities among the groups. The fecal 16S sequencing was used to determine the microbial composition of each group before and after SPS. (A) Alpha diversity measured by Inverse Simpson Index, (B) PCA plot of Beta diversity before SPS measured by Aitchison distance, (C) PCA plot of Beta diversity after SPS measured by Aitchison distance, (D) Differences in volatility among the groups, (E) Correlation between volatility and Anxiety Index, (F) Correlation between volatility and % time spent in the closed arms of EPM, (G) Correlation between volatility and % time spent in the open arms of EPM. Inverse Simpson before and after SPS was analyzed by Two-ways ANOVA, mixed effects, followed by Šídák's and/or Tukey's multiple comparisons test. FDR was used to correct between-tests p values. For PCA plots, data points were projected into the space spanned by the first two principal components. Correlations were performed using Pearson's correlation. Blue-Controls, Green-SPS-R, Red-SPS-S. All data are expressed as means ± SEM. *p < 0.05. Values, 2 SD away from the mean were excluded from analysis. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 6

Differences in gut microbial communities before and after SPS. The 16S sequencing was used to determine the microbial composition of each group at genus levels. (A) Relative abundances of Lactobacillus, (B) Correlation between relative abundance of Lactobacillus and Anxiety Index (AI) before SPS, (C) Relative abundance of Vampirovibrio, (D) Relative abundance of Lachnospiraceae_Incertae_Sedis, (E) Correlation between relative abundance of Lachnospiraceae_Incertae_Sedis and Anxiety Index (AI) before SPS, (F) Relative abundance of Coprobacillus, (G) Relative abundance of Anaeroplasma, (H) Relative abundance of Bacteroides, (I) Correlation between relative abundance of Bacteroides and Anxiety Index (AI) before SPS, (J) Relative abundance of Barnesiella, (K) Correlation between relative abundance of Barnesiella and Anxiety Index (AI) before SPS, (L) Relative abundance of Butyricicoccus, (M) Relative abundance of Asaccharobacter, (N) Relative abundance of Mucispirillum, (O) Relative abundance of Clostridium IV, (P) Relative abundance of Streptococcus. All relative abundances are clr-transformed. Data were analyzed by Two-way ANOVA followed by Šídák's and Tukey's multiple comparisons. FDR was used to correct between-tests p value. Correlation between relative abundances and AI was performed by Pearson's correlation. Blue-Controls, green-SPS-R, red-SPS-S. All data are expressed as means ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 6

Differences in gut microbial communities before and after SPS. The 16S sequencing was used to determine the microbial composition of each group at genus levels. (A) Relative abundances of Lactobacillus, (B) Correlation between relative abundance of Lactobacillus and Anxiety Index (AI) before SPS, (C) Relative abundance of Vampirovibrio, (D) Relative abundance of Lachnospiraceae_Incertae_Sedis, (E) Correlation between relative abundance of Lachnospiraceae_Incertae_Sedis and Anxiety Index (AI) before SPS, (F) Relative abundance of Coprobacillus, (G) Relative abundance of Anaeroplasma, (H) Relative abundance of Bacteroides, (I) Correlation between relative abundance of Bacteroides and Anxiety Index (AI) before SPS, (J) Relative abundance of Barnesiella, (K) Correlation between relative abundance of Barnesiella and Anxiety Index (AI) before SPS, (L) Relative abundance of Butyricicoccus, (M) Relative abundance of Asaccharobacter, (N) Relative abundance of Mucispirillum, (O) Relative abundance of Clostridium IV, (P) Relative abundance of Streptococcus. All relative abundances are clr-transformed. Data were analyzed by Two-way ANOVA followed by Šídák's and Tukey's multiple comparisons. FDR was used to correct between-tests p value. Correlation between relative abundances and AI was performed by Pearson's correlation. Blue-Controls, green-SPS-R, red-SPS-S. All data are expressed as means ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 6

Differences in gut microbial communities before and after SPS. The 16S sequencing was used to determine the microbial composition of each group at genus levels. (A) Relative abundances of Lactobacillus, (B) Correlation between relative abundance of Lactobacillus and Anxiety Index (AI) before SPS, (C) Relative abundance of Vampirovibrio, (D) Relative abundance of Lachnospiraceae_Incertae_Sedis, (E) Correlation between relative abundance of Lachnospiraceae_Incertae_Sedis and Anxiety Index (AI) before SPS, (F) Relative abundance of Coprobacillus, (G) Relative abundance of Anaeroplasma, (H) Relative abundance of Bacteroides, (I) Correlation between relative abundance of Bacteroides and Anxiety Index (AI) before SPS, (J) Relative abundance of Barnesiella, (K) Correlation between relative abundance of Barnesiella and Anxiety Index (AI) before SPS, (L) Relative abundance of Butyricicoccus, (M) Relative abundance of Asaccharobacter, (N) Relative abundance of Mucispirillum, (O) Relative abundance of Clostridium IV, (P) Relative abundance of Streptococcus. All relative abundances are clr-transformed. Data were analyzed by Two-way ANOVA followed by Šídák's and Tukey's multiple comparisons. FDR was used to correct between-tests p value. Correlation between relative abundances and AI was performed by Pearson's correlation. Blue-Controls, green-SPS-R, red-SPS-S. All data are expressed as means ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 7

Predictive functionality of gut microbiota before and after SPS Heat map showing the significantly different KEGG pathways among the cohorts (A) Before SPS. (B) After SPS. Lipid metabolism, non-homologous end-joining, primary immunodeficiency, novobiocin biosynthesis, galactose metabolism, sphingolipids, galactose, LPS biosynthesis, bacterial invasion into epithelial cells, ubiquinone and other terpenoid-quinone biosynthesis were non-parametrically distributed and were analyzed using Kruskal-Wallis test. The remaining data passed the normality test and were analyzed by one-way ANOVA. Both testes were followed by multiple comparison tests. The comparison between the groups is presented as (-log FDR). Values, 2 SD away from the mean were excluded from the analysis. *Cellular processes, amino acid metabolism, carbohydrate metabolism, lipid metabolism, xenobiotics metabolism, glycan metabolism, biosynthesis of secondary metabolites, genetic information processing, environmental information processing, human diseases, organismal systems, co-factors and vitamin metabolites, terpenoids and polyketides metabolism.

Fig. 8

Urinary epinephrine and norepinephrine levels before SPS and 90 min into immobilization were higher in SPS-S subgroup. Urine samples were collected before SPS, 90 min into immobilization, and on the day of dissection to measure urinary epinephrine and norepinephrine levels of individual animals. (A) Relative Epinephrine levels (B) Relative Norepinephrine levels. Data were analyzed by two-way ANOVA repeated measures with post-hoc Šídák's multiple comparisons. Each dot represents value for an individual animal. Blue-Controls, green-SPS-R, red-SPS-S. All data are expressed as means ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 9

Exposure to SPS decreased the expression of brain tight junction protein Claudin-5 in SPS-S subgroup. Ventral hippocampus (vHipp) and medial prefrontal cortex (mPFC) of each animal were dissected and Western blot was performed to analyze the expression of tight junction proteins. Expression of Claudin-5 in (A) ventral hippocampus and (B) medial prefrontal cortex, (C) correlation between volatility and claudin-5 expression in vHipp; Expression of Occludin in (D) ventral hippocampus and (E) medial prefrontal cortex. Representative Western blots are shown. Claudin-5 and Occludin protein expression data in mPFC were non-parametrically distributed and were analyzed using Kruskal-Wallis test followed by Dunn's multiple comparison test. Claudin-5 and Occludin protein expression data in hippocampus passed the normality test and were analyzed using one way-ANOVA, followed by Tukey's multiple comparisons test. Each dot represents values for an individual animal. Blue-Controls, green-SPS-R, red-SPS-S. All data are expressed as means ± SEM. ns = not significant, *p < 0.05, **p < 0.01. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 10

Colon length was altered in SPS-S subgroup. Colonic measurements are expressed as total colon length. The data passed the normality test and were analyzed using one way-ANOVA, followed by Tukey's multiple comparisons test. Each dot represents value for an individual animal. Blue-Controls, Green-SPS-R, Red-SPS-S. All data are expressed as means ± SEM. *p < 0.05. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 11

SPS altered the levels of cecal SCFA Cecal content was used for SCFA analysis. (A) Levels of Acetate, (B) correlation of Anxiety Index with Acetate levels, (C) Levels of Butyrate, (D) Levels of Propionate, (E) Levels of Valerate, (F) Levels of Caproate. Data from SCFA levels passed the normality test and were analyzed using one way-ANOVA, followed by Tukey's multiple comparisons test. Blue-Controls, green-SPS-R, red-SPS-S All data are expressed as means ± SEM. *p < 0.05, **p < 0.01. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)