Stress response is defined as an emotional experience involving physiological, cognitive, and behavioral changes in adaptation to intrinsic and extrinsic stimuli. Stress response involves interworks of the autonomic nervous system, the hypothalamic–pituitary axis, and the immune system, activating both central and peripheral immune cells to release cytokines. Above all, the brain is a key organ that determines behavioral and physiological changes, resulting in atrophic changes in the hippocampus, prefrontal cortex, and amygdala under sustained chronic stressor (1). Appropriate responses of stress systems to a stressor are necessary to maintain homeostasis of the body and sense of wellbeing. However, inappropriate responses to a stressor may disturb the normal growth and development and even cause psychiatric disorders (2). Substantial evidence supports that stress response alteration plays an important role in the development of major depressive disorder (MDD), panic disorder (PD), and post-traumatic stress disorder (3). MDD is a disease with high prevalence (5.1% in females and 3.6% in males), chronic course, and high treatment cost associated with poor social function. It has a very high social disease burden due to decreased productivity (4, 5). MDD is a multifactorial disorder in which adaptation to stress plays an important role in its manifestation. In almost 80% of major depressive episodes, major stressful life events precede within 3 months before the onset of episode. Before major depressive onset, stressors are more than two times more frequent in those with MDD than in healthy controls during the same period (6). As the number of negative events and the severity increase, the risk of developing depression also increases and the onset occurs more rapidly (7). Chronic stressors are associated with poor prognosis, frequent relapse, and poor symptom relief (8, 9). The prevalence of PD is estimated to be 2.7% in 12 months and 4.7% during lifetime according to the National Comorbidity Survey-Replication reports. Its etiology is complex, having interactions with biological factors and environmental factors (10, 11). Panic disorder occurs frequently with other psychopathological comorbidities such as major depression, bipolar illness, and other anxiety disorders. In addition, it is associated with an increased risk of suicide (12). Compared to health controls, PD patients report more life events in the period before onset of PD, and 80% of patients with PD report at least one stressful life event within 6 months before the first panic attack (13, 14).

Since adaptation to stress seems to play an important role in the onset, prognosis, recovery, and even relapse of MDD and PD, the reduction of stress response during the course of the disease can be a good predictor of better prognosis or treatment outcome. There is evidence that an intense acute stressor can stimulate the pro-inflammatory cytokine network, including increased levels of cortisol, interleukin (IL)-6, and tumor necrosis factor (TNF)-α (15). Subjects with psychological stress-induced distress and anxiety show significantly greater increases of interferon (IFN)-γ and decreases in IL-10 than those without such distress or anxiety (16). Leptin and adiponectin are known to have interactions with other stress mechanisms of stress-related disorders (17). Identifying peripheral biomarkers that can predict stress response reduction has useful values in predicting the prognosis and treatment outcome of patients with MDD and PD. However, studies determining the relationship between reduction of stress response and predictive biomarkers are sparse, especially in psychiatric disorders.

Thus, the aim of this study was to compare the Stress Response Inventory (SRI) score at every visit to see any differences by groups. Reduction of stress response was defined as the difference in SRI score between baseline and week 12 (visit 5) divided by the baseline score. We hypothesized that reduction of stress response could be predicted using baseline peripheral biomarkers in patients with MDD and PD.

To our best knowledge, this is the first study to identify adiponectin as a predictive peripheral biomarker for stress response reduction in PD patients. The inverse relationship of baseline adiponectin and stress response change was still robust even after adjusting for confounding factors. Biological functions of adiponectin appear to play a more critical role in regulating stress response in PD than other related factors.

Adiponectin is one of the adipocyte-derived peptide hormones synthesized mainly by white adipose tissues. It has anti-inflammatory, insulin-sensitizing, and anti-atherogenic effects (22). Decreased levels of adiponectin are associated with chronic metabolic conditions such as obesity, diabetes, atherosclerosis, cardiovascular disorder, and hypertension (23, 24). A limited but growing body of evidence has suggested the role of adiponectin in various psychiatric disorders including mood disorder, attention deficit hyperactive disorder, post-traumatic stress disorder, obsessive-compulsive disorder, eating disorder, and even neuropsychiatric disorders (25–29). According to a systematic review, adiponectin levels are lower in patients with anxiety, mood, trauma, and stress-related disorders than in healthy controls (30). Studies investigating the role of adiponectin in PD are sparse. One cross-sectional study has compared levels of metabolic parameters and adiponectin in participants with PD and paired healthy controls and found that mean adiponectin levels are significantly lower in the PD group than in healthy controls even after adjusting for BMI and sex (31).

The amygdala is the center for fear processing. It is closely associated with the pathogenesis of panic attack and stress response system. Functional neuroanatomical studies have suggested that the amygdala functions as a center of integrating incoming signals from the peripheral immune system (32). Peripheral cytokines are known to change amygdaloid activity and increase anxiety-like behaviors. TNF-α, IL-6, and CRP are regarded as the main pro-inflammatory cytokines (33). Adiponectin has an inverse relationship with TNF-α and IL-6. It has been demonstrated that adiponectin can exert a variety of anti-inflammatory activities, including inhibition of pro-inflammatory cytokine production and induction of anti-inflammatory factors (34, 35). In addition, adiponectin can cross brain–blood barriers. It appears to be able to directly regulate not only energy and glucose homeostasis but also neurogenesis and synaptic plasticity in various grain regions (36). Recent evidence has suggested that adiponectin can regulate contextual fear extinction in the hippocampus via receptor AdipoR2 and modulate dopamine neuronal activity in ventral tegmental area and anxiety-related behaviors through receptor AdipoR1 (37, 38). Taken together, it seems reasonable to speculate that an increase of adiponectin in PD patients might alleviate stress response by reversing the effects of some pro-inflammatory cytokines, activating anti-inflammatory effects, and directly functioning in related brain regions.

Interestingly, only somatization among subscales of SRI had a clinical correlation with adiponectin levels. Some animal studies have suggested an adiponectin-mediated analgesic effect through regulating neuronal activities at the spinal cord level (39, 40). Studies on migraine and osteoarthritis have revealed an association between adiponectin and decreased severity of pain or better treatment response (41, 42). Adiponectin has also been found to regulate norepinephrine and autonomic function through acting on specific nucleus in the hypothalamus. Higher adiponectin levels are associated with better cardiovascular autonomic functions (43, 44). Dysfunction of the autonomic nervous system has been hypothesized to be associated with symptom manifestation and panic attacks in PD. Adiponectin might reinforce the capability of regulating autonomic function and raise the threshold for somatization. As another explanation, adiponectin might go through several stages of action to modulate stress response. Thus, a relatively short period of 3 months might not be long enough for assessing reduction of stress response in domains such as emotion and cognition.

In our study, more than 50% of subjects had values of IL-10, TNF-α, and IFN-γ at zero, which might have limited us to explore their possible correlations with stress response change. When we additionally evaluated non-zero values of these three cytokines in the total study population, baseline IFN-γ levels were negatively associated with 12-week changes in stress response with a correlation coefficient of −0.4266. IFN- γ has been used as a therapeutic modality in malignant tumors for its anti-proliferative and anti-tumor functions. It is also involved in regulating amygdala reactivity to emotional stimuli (45, 46). Although the number of non-zero data was small, findings of a moderate negative correlation (r = −0.4266) in this study underlie the possible important role of IFN- γ in stress regulation system.

In patients with MDD, baseline peripheral biomarkers showed no significant associations with stress response changes. For plausible explanations, the high heterogeneity of MDD, sharing inflammation theory as a key mechanism with stress response, and more chronic or add-on impacts on brain structures from both MDD and stress response can be suggested (47, 48). Responses to stressors in MDD characterized by episodic mood changes and those in PD characterized by anxiety reactions might have different brain regions activated in their mechanisms of action and associated biomarkers. Behavioral- and biological-level changes derived from stressors have been suggested to be more obvious when stressors are uncontrollable (49). Characteristics of unexpected panic attacks with anticipatory anxiety in patients with PD reflect such uncontrollability. In MDD, the response to stressors is not simply an anxiety response due to uncontrollability, but rather a wide range of emotional, cognitive, physical, and behavioral changes according to mood changes. Interactions with the hypothalamic–pituitary axis, autonomic nervous system, immune system, and activation of limbic areas seem to be involved in a complex way. Since response patterns to stress are different between MDD and PD, approaches can be differentiated for assessing and dealing with stress responses. In patients with PD, the use of anxiolytics to reduce stress response resulting from uncontrollability might be a good strategy. However, in patients with MDD, the use of antidepressants seems to be a more proper way to deal with overall stress response.

This study is novel in that it reveals baseline adiponectin as a predictor of reduction of stress response in patients with PD. However, it has several limitations. First, this study had a relatively small sample size with a low power to find differences between groups. Applying the results of this study to clinical setting requires validation and replication through larger population-based studies. Second, SRI is focused on stress responses based on the most recent month without providing a start point or duration of stressor. Thus, we could not know whether the stress response was acute or chronic. Also, reduction of stress response was calculated based on a duration of 3 months, which might be too short to reflect changes of stress response. Third, information on medication regimens such as start date and maintenance period of medication was not fully provided. Thus, we could not exclude possible modulating effects of antidepressants or antipsychotics on activities of cytokines or stress responses. Finally, inflammatory diseases may affect cytokine levels. About one-fifth of the MDD group and one-quarter of the PD group were reported to have inflammatory diseases such as hyperlipidemia or diabetes mellitus, but we could not confirm the possible modulating effects because of the relatively small sample size of this study.

In conclusion, greater baseline adiponectin levels predicted more reduction of stress response in patients with PD, but not in those with MDD. It implies that these two highly comorbid disorders, PD and MDD, might have different etiologies in stress response. Thus, differentiated approaches for regulating stress response in PD and MDD might be useful. Further large population-based and longitudinal studies with evaluation of medication and comorbidity are necessary to confirm the role and mechanisms of action of adiponectin in reduction of stress response in PD.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by the Institutional Review Board (IRB) of Samsung Medical Center (IBR No. 2015-07-151). The patients/participants provided their written informed consent to participate in this study.

MP and HJ: conceptualization. EJ, AK, and SB: data curation. HK and HSK: investigation. MP: formal analysis and writing—original draft. HY and HJ: supervision. All authors: writing—review and editing. All authors contributed to the article and approved the submitted version.

This work was partly supported by a Grant (No. 2015-0-00062, Development of skin adhesive patches for monitoring and prediction of mental disorders) of the Institute for Information and Communications Technology Promotion (IITP) funded by the Korean government (MSIT) and an Original Technology Research Program for Brain Science through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (No. NRF-2016M3C7A1947307; HJ). It was also supported by a Grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI) funded by the Ministry of Health and Welfare, Republic of Korea (No. HR21C0885).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.