Thirty years ago, the field of stress research was surprised by the first publication of an unexpected finding by Yehuda and colleagues (
1) suggesting that posttraumatic stress disorder (PTSD), a disorder that by definition manifests in response to severe stress exposure, is characterized by markedly decreased, rather than increased, secretion of the stress hormone cortisol. This seminal finding was published at a time when the general consensus was that high-stress experience was equivalent to increased cortisol release and that stress-related disorders occur as a consequence of elevated stress mediators. The counterintuitive observation of lower-than-normal cortisol concentrations in patients with PTSD sparked intense discussion and controversy; these results were in stark contrast to the state of knowledge regarding stress biology, and they forced the field to rethink prevailing models of stress and disease. With replication, the findings propelled a paradigmatic shift in how we conceptualize stress-related disorders. This paradigm shift involved the acknowledgment that stress is not necessarily equal to high cortisol secretion, which fueled novel theories that PTSD is not merely a normative response to extreme stress, but indeed reflects a nonnormative and inadequate response to severe stressors, with its pathophysiology involving maladaptation or dysfunction in stress-regulatory systems (
2).
Soon thereafter, intensive efforts focused on 1) the mechanisms that contribute to the development of hypocortisolism, for example, increased feedback inhibition versus depletion of the adrenals, and 2) the impacts that low cortisol availability would have on neural mechanisms that underlie the symptoms of PTSD (
3,
4). A prevailing model proposes that a relative deficiency or lack of the regulatory and adaptive effects of cortisol at the neural level results in disinhibition of noradrenergic and corticotropin-releasing factor circuitry connecting brainstem nuclei with the amygdala and the hypothalamus, and that such disinhibition in turn amplifies 1) vigilance, hyperarousal, and stress sensitization, 2) fear acquisition and/or lack of fear extinction, and 3) glutamate-mediated neurotoxicity, leading to structural brain changes (hippocampal volume loss) culminating in the clinical phenotype of PTSD (
5,
6).
These insights into the pathophysiology of PTSD paved the way for novel strategies for early risk detection. Defining measurable markers for early risk detection is essential, not only for identifying individuals in need of preventive treatment but also for providing targets to develop mechanistically driven interventions that directly mitigate this risk. Prospective studies provided further evidence that elevated heart rate and deficient cortisol secretion at the time of a traumatic exposure increase the risk for developing subsequent PTSD (
7). Consequently, Pitman and colleagues (
8) showed that the administration of the centrally acting beta-blocker propranolol shortly after exposure to trauma reduced PTSD symptom severity and physiological reactivity to traumatic reminders, although it did not prevent PTSD. Several studies confirmed attenuation of physiological reactivity after trauma by propranolol (
9). The administration of hydrocortisone in the immediate aftermath of a traumatic exposure was shown in a number of studies (
9,
10) to be an effective pharmacological strategy to prevent the development of PTSD. Hydrocortisone administration also has therapeutic efficiency in patients who suffer from PTSD, in particular as an augmentation strategy to exposure therapy, likely by facilitating extinction learning (
11). The history of hypocortisolism in PTSD exemplifies how an initially counterintuitive observation sparked a novel conceptualization of stress-related disease that was translated into using measurable markers for early risk detection and mechanistically driven strategies for preventing this risk or reversing the manifested symptoms.
The study by Michopoulos et al. in this issue of the
Journal (
12) is remarkable in that it translates the approach of identifying biological markers for early risk detection concerning the development of chronic PTSD immediately after trauma exposure to the immune system, thus expanding the scope of this line of research beyond stress response systems. A growing body of research has evaluated potential alterations of the immune system in PTSD (
13,
14). Acute stress induces a profound adaptive immune response in order to fight injuries and pathogens, and PTSD is often comorbid with medical diseases that are associated with inflammation or autoimmune processes. Although findings regarding immune changes in PTSD are somewhat conflicting, a recent meta-analysis confirmed that PTSD is associated with increased blood-based concentrations of interleukin 1β, interleukin 6, tumor necrosis factor α (TNFα), and interferon-γ (IFNγ), suggesting a state of systemic inflammation in PTSD (
13). Several theories posit that peripheral proinflammatory mediators contribute to the pathophysiology of PTSD by way of passage to the CNS, where they interact with microglia, resident immune cells in the brain, that in turn release neuroinflammatory mediators. Via complex interactions with neurons, neuroinflammatory states can modulate neurotransmitter systems, augment symptoms of stress and depression, modulate fear learning, and potentiate neurotoxicity and apoptosis (
14). Against the backdrop of this literature, one would expect that elevated levels of systemic inflammation in the immediate aftermath of a traumatic exposure would predict the development of PTSD, with the prospect of using anti-inflammatory drugs as a potential early intervention strategy to mitigate this risk.
In their prospective study, Michopoulos et al. evaluated a broad panel of blood-based immune mediators, including proinflammatory and anti-inflammatory cytokines as well as chemokines, within 3 hours after a traumatic exposure in 274 adults recruited in an emergency department setting, and assessed predictors of falling into one of three groups after up to 12 months of follow-up: 1) those who were resilient and did not develop PTSD, 2) those who developed symptoms but recovered, and 3) those who developed chronic PTSD symptoms. The authors found that blood-based levels of proinflammatory cytokines as a whole, but in particular TNFα and IFNγ, immediately after the traumatic exposure distinguished those individuals who went on to develop chronic PTSD symptoms from those who were resilient or had recovered from their symptoms at the time of follow-up. Remarkably, however, these proinflammatory cytokines were not increased, but instead were decreased, in the group of individuals who developed chronic PTSD when compared with the resilience and recovery groups. In other words, the resilience and recovery groups mounted a solid inflammatory response to the acute extreme stressor, whereas the risk group exhibited a deficient inflammatory response. This suggests that the individuals who went on to develop PTSD may have an inability or failure to activate an innate immune response to challenge. The effects were maintained when the authors compared subgroups of the resilience and recovery groups with the risk group, selecting cases carefully matched in relation to potential confounders. The observed effects were controlled for type of trauma (interpersonal, noninterpersonal) and prior exposures. It can be concluded that the degree of proinflammatory response to acute trauma has predictive value for identifying individuals at risk for chronic PTSD; however, the direction of the effect is opposite from what would be expected.
The results of the study are striking, as they stand in stark contrast to current knowledge and prevailing models concerning the role of inflammation in PTSD. The findings are reminiscent of the initial reports of hypocortisolism and the ensuing confusion in the field of psychoneuroendocrinology (
15). The results force us to broaden our current concepts and consider a potential role of “hypoinflammatory” signaling during acute stress as a neuroimmune risk mechanism that drives the development of PTSD and potentially other stress-related disorders. But how can we make sense of these findings? What are the determinants of a deficient inflammatory response to severe stress? Do peripheral measures of inflammation reflect neuroinflammation? Why is a robust inflammatory response to acute trauma protective, whereas a compromised response increases risk for PTSD? And what are the implications for preventive interventions? The results by Michopoulos et al. challenge us to address these questions and generate new answers.
One potential determinant of a deficient inflammatory response to acute trauma could involve heightened glucocorticoid responses to acute trauma in the risk group compared with the resilience and recovery groups, as discussed by the authors. However, this would contradict the findings that deficient glucocorticoid availability at the time of trauma is a predictor of subsequent PTSD. Another possibility is that supersensitive leukocyte glucocorticoid receptors dampen proinflammatory responses, even in a state of decreased cortisol secretion (
3–
6). Studies are needed to scrutinize the origins of the blunted immune response to acute trauma, but potential preceding or disposing factors should also be considered, such as early developmental stressors, including immune challenge, or genetic factors. Early-life stress, such as childhood adversity, is a potent risk factor for developing PTSD in response to later trauma, and elevated peripheral markers of inflammation are one of the best-replicated findings in children and adults with early-life stress (
15). However, little is known regarding the responsiveness of the immune system to subsequent challenge after early-life stress. A noteworthy recent study by Marrocco et al. (
17) showed that early-life stress leads to global restriction of translational reactivity in hippocampal tissue of rodents in response to additional acute stress, encompassing genes relevant for glucocorticoid receptor binding and inflammatory response, among others. Such findings should guide future studies to scrutinize potential “cross-system” failure of adaptive response and its relation to the prospective development of PTSD.
Before evaluating potential impacts of deficient inflammatory responsiveness on neural mechanisms that might promote PTSD risk, the extent to which the findings reported by Michopoulos et al., which are derived from peripheral blood cells, can be transferred to the brain must be discussed. Under stressful conditions, peripheral monocytes can infiltrate the brain and produce macrophages that secrete cytokines and promote neuroinflammation. Peripherally produced inflammatory mediators as well as glucocorticoids can stimulate microglia. Activated microglia in turn produce cytokines and other substances that are important for adaptive regulation (
18,
19). It may be deduced that both deficient peripheral inflammatory and glucocorticoid responses, as observed in those trauma-exposed individuals who are at risk for PTSD, might favor compromised microglia activation and contribute to deficient neuroinflammatory response. However, this notion is speculative and remains to be systematically studied. In humans, microglial activation can be directly measured only in postmortem brain tissue. Neuroimaging methods, such as positron emission tomography, are available to estimate neuroinflammation in vivo; however, such studies would be difficult to implement in an emergency department setting after a traumatic exposure.
If a deficient neuroimmune response to acute trauma can be confirmed, it must be considered in what ways such a deficiency could be a risk mechanism for developing PTSD, whereas an adequate neuroimmune response would hypothetically be protective. While a comprehensive discussion of the complex functions of microglia is outside the scope of this commentary, it should be noted that microglia, as the brain’s primary defense system, are involved in various forms of pathology or challenge to the CNS and are critically important for recovery and regeneration. In the damaged brain, resting microglia undergo a transformation and become active phagocytes that eliminate damaged cells and remove synapses from injured neurons. Activated microglia produce a variety of substances, including cytokines and trophic factors, that influence pathological processes during acute and chronic phases of the disease process and during subsequent regeneration (
20,
21). Of note, several studies have shown that TNFα produced by microglia contributes to structural and functional recovery after traumatic brain injury (
22). Blockade of TNFα signaling leads to enhancement of pathology in a mouse model of neurodegeneration (
23). Moreover, microglia can switch to an anti-inflammatory phenotype over the course of the disease progression (
20,
21). In recent years, it has been increasingly recognized that microglia activation and the substances produced by microglia, including TNFα, are critical elements for neuroimmune cross-talk and that proper function of microglia is necessary for adaptive plasticity, learning, and maintaining homeostasis in the CNS (
20). Taken together, although peripheral inflammatory responses, activation of microglia, and subsequent neuroinflammation are generally considered “pathogenic” in stress research, these findings provide a foundation to posit a new hypothesis suggesting that “lower than normal” microglial activation in states of acute stress could be detrimental, impairing adaptive and neuroprotective capacities that are critical for recovery from traumatic stress. Interesting in this regard is the observation that preexisting experiences, such as early developmental insults, can elicit microglial “priming” that leads to a dormant memory for potentiated but also decreased neuroimmune response if the challenge reoccurs (
24).
Finally, the findings by Michopoulos et al., in the context of the above considerations, challenge the idea that anti-inflammatory drugs might be beneficial for the treatment of PTSD or may be useful for preventing PTSD risk in the direct aftermath of a traumatic exposure. It is noteworthy that in a small clinical trial in soldiers with PTSD, symptom improvement during psychotherapy was accompanied by increases in peripheral TNFα (
25). Perhaps it might be speculated that a certain degree of inflammation is necessary for adaptive plasticity, and blocking such processes could be detrimental in the direct aftermath of traumatic exposure.
In conclusion, generating predictive models of PTSD that allow for the detection of individuals at risk in the immediate aftermath of a traumatic exposure is indispensable for devising preventive strategies and improving public health. With ongoing conflicts and war, refuge-related trauma, and high rates of civilian traumas, PTSD is one of the major epidemics of our time. Current and future studies need to capture the complexity of the determinants and mechanisms contributing to PTSD risk by using big-data approaches and machine-learning methods to build predictive models (
26). While it is clear that such predictive models would include markers of immune response, they should also include potentially important moderators of immune response, such as mild traumatic brain injury, physical injury, and changes in cell distributions with trauma, as well as predisposing factors, such as early developmental insults, including immune challenges. While big-data approaches certainly have unforeseeable potential in helping us understand the complexity in modeling disease risk, we still need to make sense of identified predictive markers in terms of their biological function and in the context of specific disease mechanisms.
Thirty years after its initial conception, the role of hypocortisolism in the development of stress-related disorders is widely accepted and has paved the way for new clinical strategies. The study by Michopoulos et al. is important, as it has the potential to induce a novel paradigmatic shift of similar magnitude in the field of psychoneuroimmunology. It places in the spotlight the notion of a deficient inflammatory response to acute trauma as a neuroimmune mechanism that drives the development of PTSD and possibly other stress-related disorders, at a time when elevated inflammation, by and large, is considered in the field to be a pathogenic factor. The findings remind us, once more, that biological responses to stress are not universally damaging but also have important adaptive functions.
Acknowledgments
The author receives funding from the German Ministry of Research and Education (BMBF 01GL174301 and 01GL1903), the Deutsche Forschungsgemeinschaft (DFG)–funded NeuroCure Cluster of Excellence (EXC 2049), and NIH (P50-HD089922).