Abnormal intestinal milieu in posttraumatic stress disorder is not impacted by treatment that improves symptoms

Abstract

Posttraumatic stress disorder (PTSD) is a psychiatric disorder, resulting from exposure to traumatic events. Current recommended first-line interventions for the treatment of PTSD include evidence-based psychotherapies, such as cognitive processing therapy (CPT). Psychotherapies are effective for reducing PTSD symptoms, but approximately two-thirds of veterans continue to meet diagnostic criteria for PTSD after treatment, suggesting there is an incomplete understanding of what factors sustain PTSD. The intestine can influence the brain and this study evaluated intestinal readouts in subjects with PTSD. Serum samples from controls without PTSD (n = 40) from the Duke INTRuST Program were compared with serum samples from veterans with PTSD (n = 40) recruited from the Road Home Program at Rush University Medical Center. Assessments included microbial metabolites, intestinal barrier, and intestinal epithelial cell function. In addition, intestinal readouts were assessed in subjects with PTSD before and after a 3-wk CPT-based intensive treatment program (ITP) to understand if treatment impacts the intestine. Compared with controls, veterans with PTSD had a proinflammatory intestinal environment including lower levels of microbiota-derived metabolites, such as acetic, lactic, and succinic acid, intestinal barrier dysfunction [lipopolysaccharide (LPS) and LPS-binding protein], an increase in HMGB1, and a concurrent increase in the number of intestinal epithelial cell-derived extracellular vesicles. The ITP improved PTSD symptoms but no changes in intestinal outcomes were noted. This study confirms the intestine is abnormal in subjects with PTSD and suggests that effective treatment of PTSD does not alter intestinal readouts. Targeting beneficial changes in the intestine may be an approach to enhance existing PTSD treatments.

NEW & NOTEWORTHY This study confirms an abnormal intestinal environment is present in subjects with PTSD. This study adds to what is already known by examining the intestinal barrier and evaluating the relationship between intestinal readouts and PTSD symptoms and is the first to report the impact of PTSD treatment (which improves symptoms) on intestinal readouts. This study suggests that targeting the intestine as an adjunct approach could improve the treatment of PTSD.

INTRODUCTION

Posttraumatic stress disorder (PTSD) is a debilitating psychiatric disorder, which results from exposure to traumatic events (1). Veterans are more likely to suffer from PTSD than civilians, with those who served in combat roles at particular risk for the development of PTSD (2, 3). PTSD is implicated in long-term mental and physical health problems (4–6).

First-line interventions for the treatment of PTSD are evidence-based cognitive-behavioral psychotherapies, such as cognitive processing therapy (CPT) (7–10). Psychotherapies are effective for reducing PTSD symptoms; however, they are not universally efficacious (11). Approximately two-thirds of veterans who complete these treatments continue to meet diagnostic criteria for PTSD (11), and there are high rates of dropout from front-line psychotherapies for PTSD (12). Consequently, intensive treatment programs (ITP) delivered over 2–3 wk have become increasingly popular as a way to increase treatment completion (13, 14). Understanding what factors may sustain PTSD symptoms and consequently impact responsiveness to psychotherapy may lead to opportunities for adjunct approaches to improve current treatment strategies for PTSD (15).

One area of interest for consideration is the intestinal microbiota and the intestinal milieu. There is well-documented, bidirectional communication between the microbiota and the brain, and changes in the intestine can impact the brain and vice versa. For example, physical and psychological stress can cause proinflammatory changes in the intestinal microbiota (called dysbiosis) (16, 17). Perhaps not surprisingly, studies characterizing the intestinal microbiota in humans with PTSD report that the PTSD-associated microbiota is characterized by high abundance of proinflammatory bacteria and pathobionts (18–21). Proinflammatory changes in the intestinal microbiota can promote systemic and neuroinflammation and can influence behavior including depression and anxiety (22–28). Given the substantial overlap between psychiatric disorders (e.g., depression and anxiety) with PTSD, it is also plausible that proinflammatory changes in the intestinal microbiota may contribute to sustained PTSD symptoms (22–28).

Microbiota-host interactions can impact the brain via several mechanisms. Microbiota components [such as lipopolysaccharide (LPS) found in the outer membrane of Gram-negative bacteria] can profoundly impact the immune system and promote inflammation via mechanisms including LPS-binding protein (LBP), which presents LPS to cell surface pattern recognition receptors. Microbiota-derived metabolites such as short-chain fatty acids (SCFA), branched-chain fatty acids (BCFA), and other bacterial-derived metabolites are also biologically impactful. These metabolites influence the intestinal barrier (which is responsible for containing the proinflammatory contents within the intestine) and directly and indirectly influence the immune system with resulting consequences on inflammation. Another mechanism of gut-brain communication is extracellular vesicles (EVs), which are increasingly recognized as important mediators of cell-cell communication (29). Mammalian cells, including intestinal epithelial cells, produce EV and the number and contents of EV are influenced by the microenvironment in which they are produced making them credible candidates for mediating communication between the intestine and the brain. In addition, numerous cell types (intestinal epithelial cells, hepatocytes, neurons, and glia) produce high mobility group box protein 1 (HMGB1) that is robustly associated with inflammation, the promotion of LPS-induced inflammation, and tissue damage (30–33).

In this study, we evaluated markers of the intestinal microenvironment in control and PTSD subjects by 1) comparing control versus PTSD subjects, 2) examining the relationship between the intestinal microenvironment and symptoms of PTSD, depression, anxiety, and insomnia, and 3) assessing pre- to posttreatment changes in the intestinal milieu in subjects with PTSD enrolled in a 3-wk CPT-based ITP.

METHODS

Subjects and Assessments

The study was approved by the Institutional Review Board of Rush University Medical Center. Samples in this study were obtained from existing biorepositories, and written informed consent was obtained from subjects before providing samples.

Subjects.

Samples from control (n = 40) and PTSD (n = 40) subjects were obtained from two sources. Samples from veterans with PTSD were obtained from the Sample Biorepository of the Road Home Program at Rush University Medical Center. Subjects with PTSD completed a 3-wk intensive treatment program (ITP) offered as part of the Road Home Program between January and September 2017 (34, 35). Samples from control, nonveteran subjects without PTSD were obtained from the Injury and Traumatic Stress (INTRuST) Consortium Biorepository. Control subjects were matched to subjects with PTSD based on sex, age, and race/ethnicity (Table 1). There were no significant differences in demographic characteristics between control and PTSD subjects (Table 1). Questionnaires were completed by veterans to self-report symptoms of PTSD, depression, anxiety, insomnia, as well as alcohol use, and drug use. Medication use data was also collected before the ITP. Eight veterans reported use of no medications, with 32 others taking one or more medications (psychiatric and other types) (which were held consistent through the 3-wk ITP), antibiotic use was not noted in any subjects with PTSD (Table 1).

Table 1.

Demographic data

	Control	PTSD
Number, n	40	40
Sex
Female	18	18
Male	22	22
Age
Average	40.8 ± 1.6	41.8 ± 1.6
18–24	1 (2.5%)	1 (2.5%)
25–34	11 (27.5%)	11 (27.5%)
35–44	14 (35%)	13 (32.5%)
45–54	8 (20%)	10 (25%)
55–64	6 (15%)	5 (12.5%)
Ethnicity
Not Hispanic or Latino	34 (85%)	34 (85%)
Hispanic or Latino	4 (10%)	6 (15%)
Unknown	2 (5%)	0 (0%)
Race
American Indian/Alaska Native	0 (0%)	2 (5%)
Asian	1 (2.5%)	0 (0%)
Black or African American	8 (20%)	7 (17.5%)
White	29 (72.5%)	29 (72.5%)
Other/Unknown	2 (5%)	2 (5%)
PTSD characteristics (pretreatment): median (range)
PCL-5 (PTSD symptoms)	n/a	55.5 (30–75)
PHQ-9 (Depression symptoms)	n/a	16 (8–26)
GAD-7 (Anxiety symptoms)	n/a	14 (6–21)
ISI (Insomnia symptoms)	n/a	20 (6–28)
AUDIT-C (Alcohol use)	n/a	1 (0–11)
DAST-10 (Drug use)	n/a	0 (0–6)
Medication use
None	NA	8 (20%)
Antidepressant	NA	19 (47.5%)
Benzodiazepine	NA	3 (7.5%)
Anticonvulsant	NA	7 (17.5%)
Antipsychotic	NA	4 (10%)
Stimulant	NA	1 (2.5%)
Hypnotic	NA	6 (15%)
Prazosin	NA	9 (22.5%)
Steroid	NA	3 (7.5%)
Antihistamine	NA	6 (15%)
Anxiolytic	NA	2 (5%)
Narcotic	NA	3 (7.5%)
Antimanic	NA	1 (2.5%)
Antibiotics	NA	0 (0%)
Other	NA	24 (60%)

NA = not available. PCL (0–80): >31–33 = significant PTSD and >50 severe PTSD. PHQ (0–27): 5–9 = mild, 10–14 = moderate, 15–19 = moderate severe, >20 = severe. GAD (0–21): 5–9 = mild, 10–19 = moderate, 20 = severe. ISI (0–28): 8–14 = subthreshold, 15–21 moderate severe, 22–28= severe. AUDIT (0–12): >3 female or >4 male = hazardous drinking. DAST (drug use, 0–10): 0 = no problem, 1–2 = low level, 3–5 = moderate level, 6–8 substantial level, 9–10 = severe level. Some individuals were taking more than one medication; thus, the totals exceed 100%.

Intensive treatment program.

Subjects with PTSD attended a 3-wk ITP. The ITP consisted of 14 sessions of individual cognitive processing therapy (CPT), 13 sessions of group CPT, 13 sessions of mindfulness, and 12 sessions of yoga. Patients were also offered secondary interventions, including psychoeducation, art therapy, nutrition/fitness, case management, acupuncture, and pharmacotherapy. Prior studies report large reductions in PTSD and depression symptoms from pre- to post-ITP and that the symptom improvements are maintained up to 12 mo after treatment (34–37).

PTSD diagnosis and assessments.

PTSD diagnosis was confirmed by psychological assessments conducted as part of standard clinical procedures (38, 39). Previously validated self-report measures used in this study assessed for PTSD symptoms (PCL-5) (38, 40), depression (PHQ-9) (41), anxiety (GAD-7) (42), and insomnia (ISI) (43), as well as alcohol (AUDIT-C) (44) and drug use (DAST-10) (45). Baseline questionnaires (pretreatment) were completed within 2 wk before starting the ITP and on the final day of the ITP (day 18, posttreatment). For the PCL-5, subjects were asked to rate their PTSD symptoms for the past month (pretreatment) or during the past week (posttreatment). For PHQ-9, GAD-7, and ISI, subjects were asked to rate symptoms during the past 2 wk at both pre- and posttreatment.

Serum Collection and Assessments

Blood collection.

Blood samples were collected on the morning of the second day of the ITP (day 2) and on the last day of ITP (day 18). Blood samples were drawn, centrifuged, and serum collected/aliquoted and stored at −80°C until use.

Bacterial metabolites.

Quantitation was performed using isotope dilution GC-MS/MS by using MRM mode. The absolute quantity of each SCFA was determined using calibration curves measured for each analyte using methods we previously described (46, 47). Samples were analyzed by using the Thermo TSQ-Evo triple quadrupole in tandem with the Trace 1310 gas chromatograph (Thermo Fisher Scientific). Chromatographic separation was achieved by using an HP-5MS fused-silica capillary column (30 m × 0.250 mm × 0.25 µm; Agilent Technologies, Santa Clara, CA) coated with 5% phenylmethyl siloxane. Each extract (1 µL) was injected in split mode (10:1). Helium as carrier gas flow was 1 mL/min. The GC oven temperature program was as follows. The initial temperature of 40°C was held for 2 min after injection before it was increased up to 50°C at 3°C/min, followed by increase to 110°C at 5°C/min, then 250°C at 30°C/min and 310°C at 70°C/min, and then held at 310°C for 3 min. Argon was used as collision gas. The injector, transfer line, and ion source temperature were set at 260, 290, and 230°C, respectively. The mass spectrometer was tuned to an electron impact ionization energy of 70 eV in the MRM mode with the following parent-to-daughter ion transitions: m/z 61.0 → 43.0 for acetic acid, m/z 63.0 → 45.0 for [¹³C₂]-acetic acid, m/z 71.0 → 41.0 for butyric acid, m/z 78.1 → 46.1 for D₇-butyric acid, m/z 85.1 → 57.1 for isovaleric acid, m/z 87.1 → 59.1 for D₂-isovaleric acid, m/z 135.1 → 45.1 for lactic acid, m/z 138.1 → 48.0 for D₃-lactic acid, m/z 75.1 → 57.0 for propionic acid, m/z 77.1 → 59.0 for D₂-propionic acid, m/z 101.1 → 55.0 for succinic acid, and m/z 105.1 → 57.0 for D₆-succinic acid. Quantification of acetic acid, butyric acid, propionic acid, isovaleric acid, lactic acid, and succinic acid was performed using isotope dilution GC-MS/MS. The absolute quantity of each SCFA was determined using calibration curves measured for each analyte. Samples were analyzed using the Thermo TSQ-Evo triple quadrupole in tandem with the Trace 1310 gas chromatograph (Thermo Fisher Scientific).

Intestinal barrier.

LPS is a component of the outer membrane of Gram-negative bacteria and is used as a marker of intestinal barrier integrity. Serum LPS levels were assessed using the Pyrogen Recombinant Factor C kit according to manufacturer instructions (Lonza, Walkersville, MD). LBP is an acute-phase protein produced in the liver that binds to LPS in conjunction with other cell-surface pattern recognition receptors to initiate inflammation. LBP is reliably used as a long-term indicator of LPS exposure. Serum LBP was assessed using ELISA (HK315; Hycult Biotech, detection range: 4.4–50 ng/mL) according to manufacturer instructions.

Intestinal epithelial cells/neuroinflammation.

High mobility group box protein 1 (HMGB1) is a nuclear protein that is released by numerous cell types including intestinal epithelial cells, neurons, and glial cells in response to stress. HMGB1 activates receptors for advanced glycation end products (RAGE) and Toll-like receptor 4 (TLR4) that can initiate and sustain inflammation. HMGB1 was analyzed by Elisa (SEA399Hu; Cloud-Clone Corp, Detection Range: 62.5–4,000 pg/mL) according to manufacturer instructions.

Extracellular vesicles.

As an exploratory analysis, EVs were examined in a subset of subjects. EVs are produced by numerous cell types and found in the blood and are one mechanism of organ-organ communication. A random subset of blood samples [n = 5 control, n = 5 PTSD (before ITP)] was used for this analysis. In brief, EVs were isolated by the ExoQuick Ultra kit (EQULTRA-20A-1, System Biosciences, San Francisco, CA) followed by isolation of A33+ exosomes. A33+ and Tgs101 EV were subsequently quantified using Western blot [anti-GPA33 antibody (EPR4240, ab108938, Abcam), anti-TGS101 antibody (C-2, sc-7964, Santa Cruz Biotechnology] (48, 49).

Analysis

Data were analyzed via Student’s t test, paired t test, Wilcoxon matched-pairs rank test, or Spearman correlation, as appropriate. All analyses were conducted using GraphPad Prism (version 9.0.1 for Windows, GraphPad Software, San Diego, CA, www.graphpad.com). Outliers (i.e., ± 2 standard deviations from the mean) were removed before analysis. Statistics were corrected for multiple testing using two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli to control for the false discovery rate (FDR). Adjusted P values (q values) were considered significant when q < 0.05.

RESULTS

Intestinal Microbiome Metabolic Function of Subjects with PTSD Differed from Controls

Subjects with PTSD (pretreatment) were compared with controls to understand the intestinal microenvironment in PTSD including microbiota function, intestinal barrier function, and intestinal epithelial cell function.

Subjects with PTSD had evidence of altered microbiome function. The SCFA acetic acid was significantly lower in subjects with PTSD than in controls (P = 0.001, q < 0.001) but no changes were observed in levels of the SCFA butyric acid (P = 0.467, q = 0.262) or propionic acid (P = 0.844, q = 0.426) (Fig. 1, A–C). The level of the branched-chain fatty acid isovaleric acid was not different between controls and subjects with PTSD (P = 0.141, q = 0.089); however, both lactic acid (P = 0.010, q = 0.008) and succinic acid (P < 0.001, q < 0.001) were significantly lower in subjects with PTSD than in control subjects (Fig. 1, D–F). Taken together, these data suggest that compared with controls, PTSD is associated with altered microbial metabolic function.

Figure 1.

Subjects with PTSD have an altered intestinal milieu compared with controls. Serum samples were obtained from controls and veterans with PTSD and analyzed for markers of bacterial metabolites (A–F) including short-chain and branched-chain fatty acids, markers of intestinal barrier integrity (G–H), a marker of neuroinflammation high mobility group box protein 1 (HMGB1; I), and intestinal epithelial cell-derived extracellular vesicles (EVs; J) (an exploratory analysis in a subset of subjects). Outliers were omitted from the analysis with statistical outliers defined as ± 2 standard deviations from the mean. The actual number of samples included in each analysis after omitting outliers are indicated for each group (n = number of samples in each group). Student’s t test was used for all analyses. P value and the FDR corrected P value (i.e., q value) are indicated on each graph (ns = not significant). LBP, LPS binding protein; PTSD, posttraumatic stress disorder.

Subjects with PTSD also had evidence of intestinal barrier dysfunction. LPS was significantly higher in subjects with PTSD than in controls (P < 0.001, q < 0.001) (Fig. 1G). The higher levels of LPS in subjects with PTSD may reflect an increase in Gram-negative bacteria or intestinal barrier dysfunction. In addition, LBP (an acute early phase protein that is critical for the immune response to LPS and is thought to reflect long-term intestinal barrier function) was significantly higher in subjects with PTSD than in controls (P < 0.001, q < 0.001) (Fig. 1H). The intestinal epithelial barrier is (in part) maintained by intestinal epithelial cells; thus, intestinal epithelial cells were examined. Specifically, analysis focused on two mechanisms by which intestinal epithelial cells can impact the brain: EV and HMGB1.

Subjects with PTSD had evidence of perturbation of intestinal epithelial cells. Examination of HMGB1 revealed that subjects with PTSD had higher levels of HMGB1 than controls (Fig. 1I, P < 0.001, q < 0.001). As an exploratory assessment, the number of intestinal epithelial cell-derived EV (A33) was evaluated in a subset of subjects (n = 5/group). This revealed that subjects with PTSD had higher intestinal epithelial cell-derived EVs compared with total EVs than controls (p = 0.036, q = 0.026) (Fig. 1J).

Taken together, these data indicate that subjects with PTSD have changes in microbiota metabolic function, intestinal barrier dysfunction, and an increase in HMGB1 with a concurrent increase in epithelial cell production of EV. To investigate if there was a relationship between severity of symptoms and intestinal readouts, correlation analysis was conducted. No significant correlations were identified between any assessment and intestinal readouts (data not shown, all P > 0.05, q > 0.05) indicating that the severity of the symptoms was not associated with intestinal milieu.

Impact of ITP on the Intestinal Microenvironment

We next evaluated if the ITP (which significantly reduced symptoms of PTSD) was associated with changes in the intestinal milieu. Symptoms assessed via PCL-5, PHQ-9, GAD-7, and ISI reduced significantly from pre- to posttreatment (Table 2, all P < 0.05, q < 0.05). However, intestinal readouts were unaltered by ITP including bacterial metabolites (i.e., acetic acid, butyric acid, propionic acid), intestinal barrier integrity (i.e., LPS, LBP), nor HMGB1 (all P > 0.05, q > 0.05) (Table 2). Thus, the beneficial effects of ITP on symptoms were not associated with concurrent changes in the intestine.

Table 2.

Pretreatment to posttreatment assessments

	PTSD Pretreatment	PTSD Posttreatment	P Value	q Value
Questionnaire Assessments (Wilcoxon Matched-Pairs Signed-Rank Test) (Median) (Range) (n)
PCL-5	55.50 (30–75) (40)	31.50 (1–74) (40)	<0.001	0.004
PHQ-9	16.00 (8–26) (40)	10.00 (2–27) (40)	<0.001	0.004
GAD-7	14.00 (6–21) (40)	8.5 (0–21) (40)	0.001	0.002
ISI	20.00 (6–28) (40)	17.00 (4–27) (40)	0.011	0.020
Intestinal Microenvironment (paired t test) (mean ± SE) (n)
Acetic acid, %	100.000 ± 12.249 (35)	294.559 ± 117.349 (35)	0.104	0.148
Butyric acid, %	100.000 ± 13.157 (37)	80.065 ± 9.595 (37)	0.135	0.161
Propionic acid, %	100.000 ± 9.534 (38)	99.318 ± 8.892 (38)	0.941	0.665
LPS, %	100.000 ± 14.063 (38)	106.311 ± 12.542 (38)	0.533	0.471
LBP, %	100.000 ± 3.675 (36)	101.010 ± 3.549 (36)	0.742	0.583
HMGB1, %	100.000 ± 4.513 (38)	104.744 ± 4.513 (38)	0.381	0.385

HMGB1, high mobility group box protein; LBP, LPS-binding protein; LPS, lipopolysaccharide; PTSD, posttraumatic stress disorder. Bold type indicates significant between-group differences.

DISCUSSION

PTSD Is Associated with Changes in the Intestine

Results in this study indicate that the intestine is perturbed in individuals with PTSD including altered levels of bacterial metabolites, intestinal barrier dysfunction, and higher levels of intestine-derived EV with concurrent higher levels of HMGB1. This adds to an increasing body of literature highlighting the gut-brain axis in psychiatric disorders like PTSD, depression, and anxiety (23–27, 50–52).

In this study, subjects with PTSD had intestinal barrier dysfunction including elevated LPS and LBP. This finding is in agreement with a publication by Bajaj et al. (20), which reports that subjects with PTSD and cirrhosis combined have increased abundance of pathobionts including Escherichia/Shigella compared with controls. The increased abundance of these Gram-negative, LPS-containing bacteria (Escherichia/Shigella) could, at least partially, account for the increased levels of serum LPS that were observed in this study but high levels of LPS (and consequently high levels of LBP) are likely also the consequence of changes in the intestinal barrier.

HMGB1 was elevated in veterans with PTSD compared with controls, which is consistent with previous reports demonstrating high levels of HMGB1 in patients with PTSD (53–55). HMGB1 is produced by numerous cell types including intestinal epithelial cells, neurons, and glia (among others) under conditions of stress. The origin of the HMGB1 in this study cannot be identified; however, HMGB1 (regardless of cellular source) may play an important role in gut-brain axis communication leading to sustained inflammation in PTSD. HMGB1 originating from neurons and glia in the brain may promote systemic and neuroinflammation, exacerbate LPS-induced inflammation, and disrupt intestinal barrier integrity (30, 31). HMGB1 produced by intestinal epithelial cells may further perturb intestinal barrier integrity and promote inflammation and tissue injury (as has been shown in intestine-derived HMGB1-induced lung injury) (32, 33). No matter the origin of the HMGB1, the high levels observed in subjects with PTSD are likely to have proinflammatory consequences. Additional investigations into HMGB1 [and the specific origin(s) of HMGB1] and how it relates to the gut-brain axis are warranted (55, 56).

The intestinal barrier is maintained through several mechanisms but critically important are intestinal epithelial cells. Outcomes in this study demonstrate that subjects with PTSD had higher levels of intestinal epithelial cell-derived EVs. Numerous cell types (including intestinal epithelial cells) produce EVs and the number and contents (e.g., RNA, DNA, proteins, lipids) of the EVs are dictated by the environment of the cell producing the EVs (48, 57). Higher levels of intestine-derived EVs in this study are consistent with a perturbed intestinal microenvironment (e.g., microbiota dysbiosis, barrier dysfunction). A recent study found that veterans with PTSD have distinct profiles of microRNAs contained within plasma-derived EVs (58), but to date, no previously published reports have examined intestine-derived EVs in subjects with PTSD. Evaluations of the content of the intestine-derived EVs will be required to fully understand how they may contribute to gut-brain axis communication. Emerging data indicate that microbiota-derived EVs contribute to gut-brain communication (59, 60) but far less is known about EVs produced by intestinal epithelial cells. The current study and others (61) demonstrate that intestine-derived EVs are found in the systemic circulation, and it is plausible that EV contents from intestinal epithelial cells (e.g., proteins, RNA, DNA, lipids) may influence neuroinflammation and/or glia and neurons in the brain. Importantly, EVs (not intestinal-epithelial cell-specific) administered systemically to mice can be found in the brain and influence neurons and microglia in a PD mouse model (62) and emerging data suggest that EVs may contribute to neurodegeneration (63, 64). However, if intestine-derived EVs reach the brain to elicit a direct biological impact or if other indirect mechanisms may be important (e.g., modulation of the immune system) remains to be determined.

Studies of the microbiota-gut-brain axis reveal alterations in the intestinal microbiota in patients with PTSD (18–21). The intestinal microbiota was not evaluated in the current study but the differences in bacterial metabolites suggest changes in intestinal microbiota function. Previous reports have investigated the metabolic profile in individuals with PTSD and demonstrate differences in phospholipids, fatty acid metabolites, and bile acids (65–70). This study identified three metabolites as significantly lower in veterans with PTSD than in controls: lactic acid, succinic acid, and acetic acid.

Each metabolite will be discussed in turn. Lactic acid was significantly lower in veterans with PTSD than in controls, which is in discordance with a previously published report indicating that lactic acid levels are higher in male combat veterans with PTSD (65). A subanalysis of only males (to mirror the previous study) again showed that lactate levels were lower in male subjects with PTSD than in controls (control: 141.6 ± 9.4 vs. PTSD: 114.1 ± 4.0). Therefore, we cannot reconcile the difference between the current and the previous study based on sex nor are there any other factors (e.g., depression) that can readily account for the differences between these studies. Nonetheless, lactic acid can induce panic attacks and flashbacks in individuals with PTSD (71–74), therefore, it is of value to understand if microbiota-derived lactic acid may influence PTSD. Succinic acid was lower in veterans with PTSD than in controls. To the best of our knowledge, no reports have identified altered succinic acid levels in humans with PTSD. Succinic acid has anxiolytic-like effects in rodents and it is possible that the lower levels of succinic acid in individuals with PTSD may promote anxiety (75). Another metabolite that was significantly different between veterans with PTSD and controls was acetic acid. In 2017, Hemmings et al. (18) reported reduced abundance of bacteria in the phyla Actinobacteria and Verrumicrobia in the PTSD-associated microbiome compared with controls. Many acetate-producing bacteria are in the phyla Actinobacteria (e.g., Bifodobacterium bifidum, Bifidobacterium longum, Bifidobacterium breve, etc.) and Verrucomicrobia (e.g., Akkermansia mucinophilia) (76) and the lower levels of acetic acid in subjects with PTSD in this study are consistent with these previously reported changes in the microbiome. Acetic acid attenuates LPS-induced neuroinflammation (77), interleukin-1β production (78), and is protective against neurodegeneration in a mouse model of Alzheimer’s disease (79). Therefore, the consequences of low levels of acetate could contribute to sustained neuroinflammation in PTSD (especially in the context of elevated levels of LPS and barrier dysfunction) (80).

Changes in the intestine (microbiota/metabolites, barrier) are often linked to systemic inflammation. Studies widely report that individuals with PTSD have elevated levels of proinflammatory markers such as interleukin-1β, interleukin-6, C-reactive protein, and tumor necrosis factor-α in the blood (81–84). It is plausible that changes in the intestine (e.g., LPS, EV, HMGB1) contribute to systemic inflammation in PTSD.

Reducing PTSD Symptoms Did Not Alter Intestinal Readouts

Psychological stress impacts the intestine including the microbiota and the barrier (17, 85, 86). Surprisingly, ITP (which significantly reduced PTSD symptoms) was not associated with changes in intestinal outcomes. The lack of changes in the intestinal microenvironment during the ITP suggests the following possibilities: 1) additional time is needed to observe changes in the intestine (i.e., beneficial effects of the ITP on symptoms take longer than 3 wk to emerge in the intestine) or 2) it is possible that ITP does not impact the intestine.

The lack of treatment-induced changes in the intestinal outcomes could explain previously reported paradoxical findings. For example, it is reported that patients with PTSD have increased cytokine levels during treatment despite reduced psychological stress (87). Therefore, it is intriguing to consider that targeting the intestinal microenvironment may be an adjunct approach to complement existing treatment strategies.

Beneficially altering the intestinal microbiota has proven successful for symptom relief in other disorders (e.g., autism) (88) and the same could be true for PTSD. Probiotics reduce depression and anxiety, improve mood, and reduce stress responsivity (although this effect is not observed in all studies, indicating results are population-specific) (89–97). Importantly, a recent study demonstrated that administration of the probiotic Lactobacillus reuteri to veterans with traumatic brain injury and PTSD decreased inflammation and influenced heart rate in response to a stressful task with a relatively small sample size (n = 15 placebo, n = 16 probiotic) (89). These reports underscore that additional studies are needed to fully understand the clinical utility of microbiota-directed treatments. Based on results from this study, it would be interesting to determine if 1) strategies to increase acetic or succinic acid alleviate symptoms of PTSD or 2) strategies that reduce LPS concurrently reduce intestinal epithelial cell-derived EVs and inflammation.

Limitations

There are limitations to note. The relatively small sample size is one important limitation (n = 40/group), particularly for the exploratory analysis of EVs (n = 5/group). Another limitation is that the subjects with PTSD (Road Home Program) and the control (INTRuST Program) subjects are from different studies. Veterans with PTSD who attend the ITP within the Road Home Program reside throughout the United States and are not local to the Chicago area. However, we cannot discount that regional effects may account for some of the observed between-group differences. Also, we do not have a detailed trauma history or medication use for control subjects. However, if controls experienced traumatic events this would minimize the between-group differences and omitting controls with trauma would be expected to increase between-group differences that were observed. Another limitation is that no dietary data were collected in this study. The abundance of different bacteria and production of bacterial metabolites are dependent on the intake of dietary substrates; therefore, future studies must include collection of dietary information to determine if diet contributes to these differences. In addition, veterans with PTSD were taking a variety of medications that may have contributed to within-group variability and future evaluations of how medications impact the intestine (and response to the ITP) would be informative.

Summary

This report demonstrates that compared with controls, veterans with PTSD have an abnormal intestine milieu (altered production of bacterial metabolites and intestinal barrier dysfunction) with concurrent evidence of changes in intestinal epithelial cell function that could directly impact the brain (HMGB1 and EVs). These disturbances could be the consequence of the original trauma or a result of the physiological or psychological response to the trauma. The data are in agreement with other reports indicating changes in the intestinal microbiome in patients with PTSD. However, a 3-wk ITP that significantly improved symptoms in patients with PTSD did not impact the intestinal microenvironment. Going forward, additional investigations to evaluate the microbiota and approaches to modify the intestinal microenvironment may be a useful adjunct approach to complement existing treatments for PTSD.

GRANTS

This study was in part funded by NIH R24AA026801 (to A.K.), NIH R01AG056653 (to R.M.V.), NIH R01DK130227 (to J.M.B.), and philanthropic funding from Mr. and Mrs. Larry Field, Mr. and Mrs. Philip Glass, the Sylar family, Chuck and Joan Johnson, and Harlan Berk.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

R.M.V., A.K.Z., C.B.F., and A.K. conceived and designed research; S.R., L.Z., J.M.B., and R.A.B. performed experiments; R.M.V., A.K.Z., and S.R. analyzed data; R.M.V., A.K.Z., P.H., M.H.P., and A.K. interpreted results of experiments; R.M.V. prepared figures; R.M.V. drafted manuscript; R.M.V., A.K.Z., S.R., L.Z., J.M.B., C.B.F., P.H., M.H.P., and A.K. edited and revised manuscript; R.M.V., A.K.Z., S.R., L.Z., J.M.B., C.B.F., R.A.B., P.H., M.H.P., and A.K. approved final version of manuscript.