Post-traumatic stress disorder (PTSD) is highly prevalent in military populations and frequently associated with deployment-related mild traumatic brain injury (mTBI) (Hoge et al., 2008; Schneiderman et al., 2008). The risk for the development of PTSD after exposure to traumatic experiences varies among individuals and populations (Holmes and Singewald, 2013), and biological mechanisms underlying susceptibility to PTSD development remain largely unknown. Finding biomarkers associated with persistent PTSD symptoms following mTBIs may shed light on the underlying pathobiology and may lead to novel molecular targets for the development of personalized therapies.

Extracellular vesicles (EVs) such as exosomes and microvesicles are released by cells throughout the body, including neurons and glia (Fauré et al., 2006; Krämer-Albers et al., 2007). EVs have a lipid bilayer and contain a cargo that includes proteins (e.g., cytokines and growth factors) and microRNAs (miRNAs) (Taylor and Gercel-Taylor 2013; Snijders et al., 2018), which are small non-coding RNAs that are thought to play a major role in the regulation of gene expression and epigenetic alterations (Taylor and Gercel-Taylor, 2013). EVs have been linked to the pathology of age-related neurodegenerative conditions, play a role in angiogenesis and have immunodulatory properties (Perez-Gonzalez et al., 2012; Dinkins et al., 2017; DeLeo and Ikezu, 2018; Oggero et al., 2019). Moreover, neuroinflammatory activity can be initiated by EVs, which contain biologically active cytokines in their cargo and surface (Gupta and Pulliam, 2014; Fitzgerald et al., 2018). Importantly, EVs readily cross the blood brain barrier (BBB) to the peripheral circulation, where they can be accessed and isolated, allowing for the quantification of their protein and miRNA content (Gupta and Pulliam, 2014; Fitzgerald et al., 2018).

TBI of all severities has been associated with remote neurodegeneration, persistent inflammation, and vascular changes (Johnson et al., 2013; Woodcock and Morganti-Kossmann, 2013; Mckee and Daneshvar, 2015; Jassam et al., 2017). PTSD has been linked to inflammation and immune dysregulation, with reports of elevated peripheral blood levels of cytokines in association with PTSD symptoms in participants with TBI (Gill et al., 2010; Passos et al., 2015; O’Donovan et al., 2017; Rodney et al., 2020). Our group has previously reported correlations between PTSD symptom severity and EV levels of neurofilament light chain (NfL), a protein found in large myelinated axons and a marker of axonal injury and degeneration in military populations with a positive history of mTBI (Menke et al., 2015; Zetterberg 2016; Guedes et al., 2020a). Differential expression of circulating miRNAs has been observed in mTBI as well as PTSD (Bhomia et al., 2016; Gheysarzadeh et al., 2018; Lee et al., 2019; Snijders et al., 2019). Nevertheless, the potential of EV proteins and miRNAs as biomarkers of PTSD and/or PTSD symptoms in chronic TBI populations remains largely unexplored.

In this study, we aimed to analyze the molecular signature associated with severity of persistent PTSD symptoms in a cohort of combat-exposed service members and Veterans (SMs/Vs) with and without remote mTBI(s). The levels of 798 miRNAs in peripherally circulating EVs were evaluated. In addition, we measured the EV and plasma levels of eight proteins: NfL, Tau, Amyloid beta (Aβ) 42, Aβ40, interleukin-6 (IL-6), IL-10, tumor-necrosis factor-alpha (TNFα), as well as vascular endothelial growth factor (VEGF). NfL, tau, Aβ42, and Aβ40 are candidate TBI biomarkers that have been linked to neurodegenerative processes (Brody et al., 2015; Mustapic et al., 2017; Khalil et al., 2018); whereas IL-10, IL-6, TNFα are cytokines implicated in inflammatory responses (Morganti-Kossmann et al., 2002; Taylor and Gercel-Taylor 2014); and VEGF is an angiogenesis and vascular injury marker (Arany et al., 2008; Vempati et al., 2014).

This cross-sectional study evaluated peripheral blood EV levels of miRNAs and a panel of proteins among SMs/Vs with prior combat deployment, and their associations with mTBI history and current PTSD symptoms measured by the PCL-5 questionnaire. Participants who sustained one or more mTBIs were divided into two groups according to their PCL-5 scores. The control group consisted of those negative for any lifetime TBI and having lower PTSD symptom severity scores (PCL-5 below 38). We observed elevated EV and plasma levels of NfL in SMs/Vs with mTBI and more severe PTSD symptoms (PCL-5 score of 38 or higher). Moreover, changes in expression profiles of EV miRNAs were linked to PTSD symptom burden. These findings contribute to the efforts to develop prognostic biomarkers of TBI-related behavioral health disorders, and shed light on possible molecular mechanisms associated with the development of persistent PTSD symptoms in TBI populations.

EV biological functions in health and disease are yet to be completely understood (Margolis and Sadovsky, 2019; Guedeset al., 2020b). Short and long distance cell-to-cell communication mediated by EVs has been established as one of their important biological functions (Valadi et al., 2007; Frühbeis et al., 2013; Raposo and Stoorvogel, 2013; Cocucci and Meldolesi, 2015; Fitzgerald et al., 2018). EVs might also contribute to central nervous system (CNS) protective mechanisms as they assist in clearance processes such as the removal of unwanted proteins (Bulloj et al., 2010; Dinkins et al., 2017). They may also play a role in the development of neurodegenerative conditions as EVs have been linked to packaging and spreading of misfolded proteins, a mechanism shared by a number of neurodegenerative diseases (Perez-Gonzalez et al., 2012; Budnik et al., 2016; Thompson et al., 2016; Dinkins et al., 2017). In Alzheimer’s disease (AD), EVs have been implicated in the lateral and long-distance propagation of tau and might contribute to the biogenesis of Aβ fragments (Sharples et al., 2008; Bulloj et al., 2010; Perez-Gonzalez et al., 2012). Additionally, EVs play roles in inflammatory responses and angiogenesis (Gupta and Pulliam, 2014; Todorova et al., 2017; Fitzgerald et al., 2018; Oggero et al., 2019). Our group has explored the potential of EVs as biomarkers in TBI in previous studies. We have shown higher levels of EV NfL (Guedes et al., 2020a) and EV tau (Kenney et al., 2018) in individuals with history of 3 or more mTBIs when compared to those with 1 or 2 mTBIs. In addition, we observed links between higher levels of EV NfL and EV tau and more severe neurobehavioral and PTSD symptoms (Kenney et al., 2018; Guedes et al., 2020a). Interestingly, we have also shown associations between EV levels of specific miRNAS, including hsa-miR-139–5p, and severity of neurobehavioral symptoms (Devoto et al., 2020).

In this study, levels of EV NfL were higher among participants with TBI history and more severe PTSD symptoms when compared to those with less PTSD symptoms and controls. EV NfL was also associated with symptom severity in correlation analysis and regression models controlling for demographics, number of TBIs and time since last injury. In addition, a marginally significant correlation between EV Aβ40 and PCL-5 scores was also found. A broad literature supports the relevance of NfL as a biomarker of nervous system pathologies involving axonal injury or degeneration, including mTBI (Neselius et al., 2012; Bacioglu et al., 2016; Shahim et al., 2016; Yuan et al., 2017). Higher levels of NfL in blood have been observed in several neuroinflammatory and neurodegenerative conditions such as AD (Menke et al., 2015; Meeter et al., 2016; Preische et al., 2019). Elevated NfL in blood has also been shown to correlate with severity of chronic depression, postconcussive and PTSD symptoms in mTBI patients (Guedes et al., 2020a; Shahim et al., 2020). Here, we also observed links between the severity of PTSD symptoms and the levels of miRNAs that have been previously associated with some pathologies of neurodegenerative diseases. Our findings suggest a connection between the severity of persistent chronic PTSD symptoms years after mTBI and neurodegenerative changes. PTSD has been associated with structural alterations in areas of the frontal cortex, hippocampus and amygdala in animal models and human subjects, and functional changes within neural networks (Pitman et al., 2012; Fan et al., 2019; Zandvakili et al., 2020). We hypothesized that changes in brain connectivity with chronic axonal pathology might be, at least in part, underlying links between TBI and persistent PTSD symptoms. Further studies are necessary to evaluate possible roles of neurodegenerative changes and cell signaling mechanisms involving EVs in the development of persistent PTSD symptoms following TBI.

In the miRNA analysis, hsa-miR-3196, hsa-miR-372–3p, hsa-miR-615–5p, hsa-miR-1185-1-3p, hsa-miR-3190–3p, and hsa-miR-139–5p were differentially regulated in participants with more severe TBI symptoms. We observed a downregulation of hsa-miR-374a-3p in participants with mTBI and more severe PTSD symptoms in comparison to those with mTBI and less severe symptoms. miR-374a-3p has been linked to downregulation of pro-inflammatory markers associated with insulin resistance (Doumatey et al., 2018), and targets molecules that are part of oxidative phosphorylation, mitochondrial dysfunction, neuroinflammation, and amyotrophic lateral sclerosis canonical signaling pathways, among others. In addition, we found that hsa-miR-139–5p was downregulated in the mTBI groups when compared to controls and linked to severity of PTSD symptoms in regression models. Moreover, regression models including EV or plasma NfL were improved with the inclusion of hsa-miR-139–5p. Importantly, hsa-miR-139–5p is also suppressed in the blood of rats resilient to chronic stress, suggesting a link between this miRNA and vulnerability to stress (Chen et al., 2015), but possible associations between hsa-miR-139–5p and TBI-related PTSD symptoms had not been previously reported.

Our finding of downregulated hsa-miR-139–5p in the mTBI groups also confirms previous studies in animal models, showing a downregulation of hsa-miR-139–5p in the dentate gyrus in association with chronic mTBI (Puhakka et al., 2017). Decrease in peripheral blood EV hsa-miR-139–5p in Alzheimer’s disease (AD) patients when compared to controls has also been reported (Lugli et al., 2015). hsa-miR-139–5p has been implicated in neurodegenerative processes in an AD model via the targeting of metabolism- and circadian rhythm-related genes, and is considered a candidate biomarker for prion diseases (Noh et al., 2014). hsa-miR-139–5p targets CASP3, which encodes the caspase 3 protein. CASP3 is part of canonical pathways functionally associated with mitochondrial dysfunction, oxidative phosphorylation, amyotrophic lateral sclerosis, Huntington’s disease Signaling, and Parkinson’s disease signaling, among others. Caspase 3 is important in the activation cascade of caspases in apoptosis, and a major mediator of apoptosis in neurons (Brentnall et al., 2013; D’Amelio et al., 2009). Moreover, caspase 3 is involved in the cleavage of β-amyloid precursor protein (APP), which is targeted by hsa-miR-372–3p and linked to cell death in AD (Nishimura et al., 2002). APP cleavage by caspase 3 also occurs in rodent models of traumatic axonal injury (Stone et al., 2002). Other components of the amyloid processing pathway are the calpain small subunit 1 gene (CAPNS1), a hsa-miR-3196 target, and protein kinase cAMP-activated catalytic subunit beta (PRKACB), which is also part of axonal guidance and synaptogenesis signaling pathways and targeted by hsa-miR-372–3p.

The cAMP Response Element-Binding Protein (CREB) signaling in neurons canonical pathway is modulated by hsa-miR-3196, hsa-miR-372–3p, hsa-miR-615–5p, hsa-miR-1185-1-3p, hsa-miR-3190–3p, hsa-miR-374a-3p, and hsa-miR-139–5p. CREB acts as a transcription factor binding to the cAMP response element (CRE) of the promoters of its target genes (Wang et al., 2018). In neurons, CREB signaling is associated with processes such as proliferation, differentiation, neurogenesis, and plasticity (Sakamoto et al., 2011; Wang et al., 2018). CREB is a major regulator of neurotrophins such as brain derived growth factor (BDNF), and BDNF promotes the activation of CREB through tropomyosin receptor kinase (Trk) B receptors (Yossifoff et al., 2008; Wang et al., 2018). SHC1 (SHC adaptor protein 1) is part of the CREB signaling in neurons canonical pathway and a target of hsa-miR-139–5p. SHC1 is also involved in several canonical pathways that include synaptogenesis, actin/cytoskeleton, VEGF, IL-6, estrogen receptor, and glucorticoid receptor signaling. The Shc family of adaptor proteins has multiple domains that allow the recruitment of multiple signaling molecules, and plays a major role in cell signaling mediated by integrins and growth factors (Ravichandran, 2001; Ahmed and Prigent, 2017).

In this study, all miRNAS that were differentially regulated in participants with more severe PTSD symptoms have target molecules that are part of the Glucocorticoid receptor (GR) signaling canonical pathway. GR, in addition to mineralocorticoid receptors, mediate the effects of Glucocorticoids (GCs), steroid hormones that are synthesized in the adrenal cortex and have potent anti-inflammatory effects (Yehuda 2009; Sundahl et al., 2015; Scheschowitsch et al., 2017). The hypothalamic–pituitary–adrenal (HPA) axis coordinates hormonal and inflammatory stress responses to stress. The activation of the HPA in response to acute stress culminates in GC release, which plays a role in the coordination of hormonal and behavioral stress responses (Leistner and Menke, 2020). Changes in the HPA axis are a characteristic of PTSD pathophysiology (Yehuda, 2009), and associations between PTSD and changes in GR sensitivity involving the FK506 binding protein 5 (FKBP5) gene recently emerged (Li et al., 2020). PTSD has also been associated with an increased risk of cardiovascular disease (CVD) (Edmondson et al., 2013; Scherrer et al., 2019; Ebrahimi et al., 2021), which might be due to comorbid conditions (Scherrer et al., 2019), and insulin resistance (Michopoulos et al., 2016; Blessing et al., 2017). Accordingly, in this study, miRNAs that were differentially regulated in participants with more severe PTSD symptoms were also associated with insulin secretion signaling, cardiac hypertrophy and cardiac beta-adrenergic signaling canonical pathways.

PTSD has been previously linked to low grade systemic inflammation, characterized by elevated blood levels of inflammatory markers (Maes et al., 1999; Passos et al., 2015; Speer et al., 2018). Chronic PTSD symptoms have been associated with peripheral blood elevations of IL-6 in civilians and military populations (Gill et al., 2010; Lindqvist et al., 2014; Rodney et al., 2020). Cytokines mediate cell-to-cell communication as soluble factors and in association with EVs, where they might be surface-bound or encapsulated (Fitzgerald et al., 2018). In this study, we observed higher levels of EV IL-6 in those with mTBI and more severe PTSD symptoms, but group comparisons were only marginally significant. miR-1185-1-3p modulates the mRNA levels of GSK3B, which has been linked to peripheral and central inflammatory diseases, and promotes the expression of cytokines such as TNF-α and IL-6 (Beurel 2011; Garcia-Lacarte et al., 2019). In addition, hsa-miR-3196, hsa-miR-372–3p, hsa-miR-615–5p, and hsa-miR-139–5p were associated with the IL-6 canonical signaling pathway as seen in the miRNA target analysis. Further investigation of links between EV levels of cytokines and the development of PTSD symptoms in military populations are warranted.

Limitations of this study include the relatively small sample size and number of participants with severe PTSD symptoms, though the largest to date exploring this novel biomaker, and the variability in number of mTBIs and years since last TBI, and results must be considered exploratory and hypothesis generating in this Biomarker Discovery dataset. Additionally, we evaluated PTSD symptoms using a questionnaire that allows for quantification of symptom severity, but is not a stand-alone diagnostic tool. Our cohort was predominantly white and male, which limits our ability to generalize our findings. Further, all were combat-deployed which limits our findings to combat-related TBI and PTSD symptom severity and is not generalizable to civilian populations. Levels of miRNA in peripheral blood may be influenced by other factors, such as medication use, that were not evaluated in this study. Furthermore, proteins and miRNAs analyzed here are found at low concentrations in the peripheral circulation. Even though we selected methods with high sensitivity to measure their concentrations, EV proteins and miRNAs were not detectable in some samples. The methods used in this study do not allow us to identify the tissue of origin, or to distinguish miRNA and proteins that are located inside the vesicles from those that are bound to the EV membrane. Future technical advancements could further improve EV analysis as well as detection of molecules found at low levels in the peripheral circulation that might reflect pathological processes underlying the development of PTSD symptoms following mTBIs.

Despite these limitations, our findings could inform subsequent larger prospective studies examining longitudinal changes in EV biomarkers, which are warranted to validate our results and to understand the role of EVs in the pathophysiology of chronic mTBI and PTSD. Longitudinal studies could identify relationships between levels of biomarkers and PTSD symptoms over time, and biomarkers that could predict those most at risk of developing severe PTSD symptoms at later timepoints. This line of research may lead to novel avenues for the treatment of PTSD, and may facilitate clinical interventions prior to the onset of symptoms and underlying pathological processes.

Here, we found links between levels of NfL and the severity of PTSD symptoms in persons with military combat deployment and varying mTBI histories. We also observed an association between persistent PTSD symptoms and the EV levels of miRNAs, especially hsa-miR-139–5p. Our results shed light on possible mechanisms underlying individual susceptibility to the development of persistent or later-in-life PTSD symptoms. Specifically, NfL and hsa-miR-139–5p were linked to the severity of PTSD symptoms in group comparisons, correlation analysis and regression models controlling for potential confounders. NfL is a marker of axonal damage that is elevated in neurodegenerative conditions. Similarly, hsa-miR-139–5p has been previously associated to neurodegenerative processes. Our results suggest a possible role for axonal degeneration and neurodegenerative changes in the development of persistent chronic PTSD symptoms years after mTBI.