In addition to the HPA, the SNS is also activated during a stress response. In the brain, this system induces the release of its effectors, epinephrine (E) and norepinephrine (NE), from the locus coeruleus via stimulation from the brainstem [4]–[6]. The SNS can be assessed in various ways, for example, measuring NE and E levels, galvanic skin response, heart rate, blood pressure, or salivary alpha-amylase (sAA) levels [7], [8].

Numerous behavioral paradigms have been used to induce psychosocial stress. To date, the most frequently used laboratory stressor is the Trier Social Stress Test (TSST: Kirschbaum, et al., 1993), a combination of an impromptu speech and mental arithmetic in front of an audience. The TSST elicits a peak cortisol response 20 to 30 minutes after onset, which returns to baseline approximately 60 minutes post activation [3], [9]. In contrast, SNS markers such as sAA, show peak values in close temporal proximity to the stressor, and a rapid return to baseline levels after the cessation of the stressful stimulus [10]–[12].

Recently, increased attention has been paid to the interaction between the two systems. One review has highlighted the anatomical link between the two systems in the central nervous system, with fibers linking CRH releasing neurons in the hypothalamus with NE releasing neurons in the brainstem [13]. Furthermore, NE levels have been shown to affect HPA regulation by stimulating CRH release from the hypothalamus [5], [6]. Although this evidence suggests a significant interplay between the SNS and HPA, the exact mechanism of this relationship under stress remains to be elucidated.

In addition to the physiological and endocrine reaction to stress, there is also the subjective emotional experience of stress [14]. This characteristic of stress is of great importance as the perception of the situation as stressful is the initial trigger of the physiological and endocrinological response to stress. Hence, there should be a strong correlation between the psychological components of stress (the perception of a situation as stressful) and the physiological and endocrine reactivity to stress. However, this relationship is not consistently observed - a recent meta analysis has found this link to be rather weak and inconsistent [14]. The authors observed an association between the emotional experience of stress and cortisol in only 27% of the studies (N=30), and in only 25% of studies with cardiovascular measures (N=12) [14]. The authors note the low number of studies and advise a dynamic approach of assessing the psychological aspects of stress, to go beyond a simple pre and post measure methodology.

This discordance between subjective and objective measures is not only observed in stress research, but has also been highlighted in various other disciplines (for example, subjective pain and sleep experiences [15], [16]). Understanding the mismatch between subjective and objective ratings may help better understanding these various states, and eventually lead to improved interventions.

Our group has previously contributed to the literature on the interaction of the physiological and psychological responses to stress. Using a newly developed stress paradigm by combining the dexamethasone suppression test and the TSST, resulting in “The combined Dexamethasone/TSST paradigm,” we could investigate the impact of suppressing the peripheral HPA axis response on the subjective, and the physiological reactivity to stress [17]. In that study, the groups pretreated with dexamethasone the night before testing showed a suppressed cortisol stress response, together with an elevated heart rate throughout the stress protocol compared to the placebo group, suggesting an inverse relationship between the two stress systems [17]. A trend for higher subjective stress responses to stress could also be observed. These results supported the hypothesis of an inverse association between the HPA and SNS, and we thus wanted to follow-up the original study by conducting the reverse manipulation, and suppress the SNS.

Thus, the goal of the current study was to further investigate the interaction between the SNS, the HPA and the subjective stress response by suppressing the SNS using propranolol, tentatively called: “The combined Propranolol/TSST paradigm.” Propranolol is a non-selective β-antagonist (β₁ & β₂) commonly used to treat hypertension, anxiety and tremors [18], [19]. β-blockers have been shown to reduce overall cardiac reactivity and salivary alpha-amylase (sAA) [20]–[23]. Previous human studies have shown an increase in cortisol following propranolol administration [20], [21], [24]–[26], however have not included systematic assessments of stress at the endocrine, physiological and behavioral levels in combination with propranolol administration.

The significant decrease in heart rate and sAA levels in the PROP group confirms the effectiveness of the pharmacological procedure. However, no significant differences between the groups for the diastolic blood pressure measurements were observed, together with a strong trend for systolic blood pressure. This lack of a clear effect on the blood pressure may be due to the indirect effect of propranolol on the α₁-adrenoceptor, responsible for vasoconstriction [39]. As propranolol blocks the β₁ & β₂ –adrenoceptors, it results in increased synaptic NE and subsequent α₁-adrenoceptors activation. Accordingly, propranolol is discussed as a direct β-blocker and indirect α₁ agonist [39]–[43]. Therefore, the combination of these two effects may have mutually antagonized any net effect on blood pressure, explaining the lack of clear group differences on this set of variables.

In line with our finding, other studies similarly reported inconsistent effects on SNS biomarkers changes after propranolol. For instance, Maheu et. al. [24] did not observe significant changes in heart rate and blood pressure after propranolol administration in combination with a TSST paradigm. A potential additional explanation might be that the here used SNS measures were all indirect, thus increasing variability in the target variables, resulting in smaller effect sizes and thus lower chances for significance [17].

The typical cortisol patterns observed in both groups corroborate the effectiveness of the TSST in eliciting a stress response [3]. More importantly, in-line with previous literature, increased cortisol levels were observed in the PROP group [20], [21], [24]–[26]. Different mechanisms have been proposed to explain this effect. First, the SNS could have a bi-regulatory role (excitatory and inhibitory) on the adrenal cortex, where in case of an abnormally elevated cortisol response, the SNS would inhibit the HPA activity [44]. In this case, since propranolol inhibited the SNS, it would have prevented this inhibitory action, explaining the increased cortisol levels in the PROP group. In addition, by preventing NE's binding to the β-adrenoceptors with propranolol, an increased compensatory E production could result, stimulating increased CRH release and subsequent HPA axis activation [44]. The synergistic effect of both mechanisms could explain the rise in cortisol post propranolol administration; nonetheless, further investigation is required to fully comprehend these underlying processes.

The increased cortisol in the propranolol group may have significant implications for the use of propranolol as a treatment for cardiovascular disease and/or social anxiety in the long-term. The chronic effects of excessive cortisol exposure on learning, memory, and cognition are typically considered detrimental, together with adverse effects on neuronal plasticity and dendritic branching. Thus, the short-term benefit of buffered heart-rate response might be offset by the long-term detrimental effects of elevated cortisol exposure. We are not aware of any studies that have investigated this issue, but we think that this is an important topic to consider for future research.The observed subjective stress response findings are is in line with previous studies, where peak subjective stress levels were observed during the TSST, and declined quickly thereafter [45]. Additionally, in the current study subjective stress levels did not appear to be closely related to either HPA, or SNS activity. This was not in line with our expectations, and we thus have to reject the hypothesis that SNS suppression would be associated with lower levels of subjective stress. While the PROP group failed to show an increase in heart rate in response to the stress, this group did show a normal increase in subjective stress levels. Further, in line with previous studies, the magnitude of the cortisol response was found unrelated to subjective stress levels, suggesting a relative independence of subjective stress levels from the activity of the HPA axis. It is of course conceivable that other markers of HPA axis activity, like CRH, would show a closer correlation with the subjective stress levels. The fact that we looked at the most downstream marker of HPA axis activity (cortisol), which also is in large temporal distance to the onset of stress, might have further masked this relationship. Of note, methodological approaches like cross-correlations have been suggested to bridge the temporal gap, and make the cortisol/subjective stress relationship more visible [46].

The relationship between perception of stress and its associated physiological response has been dealt with on an empirical as well as theoretical level. The two-factor theory of emotion, for example, proposes that the perception of the bodily arousal contributes to the emotional. In the context of stress, a decreased arousal (through partial SNS suppression) should thus have led to a decreased subjective stress experience. Our current findings can not support this theory. First, independent of SNS suppression, the appraisal of the stressor as threatening led to a subjective stress response. Second, the levels of subjective stress were even nominally higher in the propranolol group at the onset of the experiment (although this difference was not significant), suggesting if anything an effect in the opposite direction. It could be speculated that this might have been due to a central effect of CRH, since the HPA axis activity at the time of the stressor onset is limited to CRH, which was likely increased at this time already. However, CRH was not measured in the current study, thus this has to remain speculative at this point. Since CRH measures can only be obtained through lumbar puncture, which is rarely feasible within the limitations of a laboratory stress task, an experimental design where both systems are suppressed in the presence of a psychological stress would allow to test this hypothesis instead.

One of the limitations of the current study is that we focused solely on men. Due to the known regulatory effects of hormones involved in the menstrual cycle on the HPA and SNS in women, we decided to first establish the main effects in a male population sample. As a consequence, our conclusions are limited to males only. Future studies should include women in various phases of the menstrual cycle, and women using birth control. Another limitation is the testing time - cortisol stress reactivity is best tested in the afternoon, when levels are lower and allow for a greater cortisol increase over baseline. In the current protocol however, the morning testing time was chosen to be comparable to other ongoing studies in the laboratory. One of these studies employed the dexamethasone suppression test, where dexamethasone was administered at night and the subsequent stress testing had to be done in the morning while the system is still suppressed. Therefore, to allow for later comparison between the different drug regimens (dexamethasone, propranolol and placebo), time of testing was kept identical across all groups.