The notion of ghrelin as an important junction of metabolism, stress, mood and reward (1) has raised interest concerning its role in psychiatric conditions like addiction, depression, anxiety disorders and post-traumatic stress disorder (PTSD) (2, 3). These disorders are often associated with metabolic dysregulation and show a high level of comorbidity with obesity and eating disorders as well as with each other (4, 5). This suggests a common neurobiological pathology, which represents an interesting possibility to study potential diagnostic biomarkers and novel therapeutic targets (6). Dysregulation of the ghrelin system in psychiatric disorders, and anxiety- and trauma-related disorders in particular, has been reported in several studies. For example, polymorphisms in the pre-proghrelin gene have been linked to an increased risk of panic disorder (7) or symptom severity in PTSD (8). Some studies reported altered plasma ghrelin levels in different anxiety disorders (9–11), while a recent population-based trial found that total serum ghrelin levels were positively associated with physiological anxiety, but negatively associated with pathological anxiety (12). Also, one study reported that therapy resistance in panic disorder and depression was correlated with elevated plasma acyl ghrelin (11). However, to date, it remains unclear whether a modulation of the ghrelin system could be beneficial in the treatment of anxiety- and trauma-related disorders (6). Several rodent studies point to a positive effect of ghrelin receptor (GHSR) agonism on fear- and anxiety-like behaviors (13–17). Moreover, accumulating evidence from preclinical studies suggests that caloric restriction or short-term fasting can reduce anxiety-like behaviors and attenuate fear, probably in a ghrelin-dependent manner (13, 18–21). Interestingly, also in healthy human participants overnight fasting before a fear extinction session reduced the return of fear, which was correlated with plasma ghrelin levels (22). Along these lines, it appears as a particularly interesting therapeutic avenue to target the ghrelin system as an adjunct to psychotherapy to improve treatment efficacy: While cognitive behavioral therapy, which relies on fear extinction, is a gold-standard treatment for anxiety disorders and PTSD, many patients experience only limited treatment success and relapse after time (23, 24), which is likely due to individual deficits in the ability to acquire, consolidate or retrieve extinction memories (25, 26). However, so far, it has not been studied whether GHSR agonism or short-term fasting can rescue deficient fear extinction. One relevant rodent model to investigate this, is the 129S1/SvImJ mouse strain, which exhibits severe deficits in Pavlovian fear extinction paradigms (25, 27, 28). In the current study, we tested the functionality of the ghrelin system in S1 mice compared to the well-characterized responses of C57BL/6J (BL6) mice to GHSR agonist administration and short-term fasting and investigated GHSR agonist MK0677 and overnight fasting as potential means to attenuate fear in the S1 mouse strain. Because we previously found no effect of MK0677 administration or fasting on fear extinction in BL6 mice (29), we did not include additional BL6 groups in fear conditioning and extinction experiments in this study.

Many studies in rodents suggest that ghrelin and the GHSR play a role in the regulation of fear and anxiety-like behaviors. Which role, however, is still a matter of debate. We previously reported that a loss of GHSR function, induced either via exposure to a high-fat diet (31) or a knock-out of GHSR, had no significant effect on fear extinction or anxiety-like behaviors in mice on a BL6 background (32). This led us to conclude that intact GHSR signaling is no pre-requisite to extinguish fear and that GHSR dysfunction per se has no influence on fear and anxiety-related behaviors (32). Nevertheless, this did not exclude the possibility that a manipulation of the ghrelin system could have anxiolytic or fear-reducing effects, as suggested by several rodent studies [for summary see (6)]. In the current study, we investigated for the first time whether GHSR agonism and overnight fasting would be sufficient interventions to rescue the extinction-impaired phenotype of the S1 mouse strain.

Before testing whether an activation of the ghrelin system can ameliorate the extinction-impaired phenotype of the S1 strain, we wanted to confirm its functionality in comparison to the responses of (typically extinction-competent) BL6 mice to GHSR agonism and overnight fasting. In our experiments, the GHSR agonist MK0677 induced food intake in S1 similar to BL6 mice (but had no effect on locomotion that could potentially confound behavioral readouts), while overnight fasting stimulated ghrelin release in both mouse strains, suggesting a responsive ghrelin system in the S1 line. Interestingly, while MK0677 increased food intake in both, S1 and BL6 mice, baseline food intake shortly before injection of saline or MK0677 was lower in S1 mice. Overnight food intake prior to the experiment, on the other hand, was similar in both mouse strains. The circadian distribution of wakefulness and sleep is strongly altered in S1 compared to BL6 animals (33), thus, it is likely that also the peak times of food intake may differ between the strains. This is furthermore corroborated by the finding that the baseline distance traveled was lower in S1 than BL6 mice, which indicates decreased home cage activity at the beginning of the light period. In agreement with this finding, reduced spontaneous locomotion and home cage activity in the S1 strain has been reported before (34–36). We did not observe spontaneous freezing in the S1 animals (data not shown, but also see no pre-CS habituation during fear conditioning) as food intake measurements were conducted in the home cage to minimize anxiogenic effects during the test. Importantly, it should be considered that alterations in circadian rhythms may not only influence food intake but also ghrelin secretion. Overnight fasting stimulated ghrelin release in both mouse strains, however, we found elevated acyl ghrelin levels in S1 compared to BL6 mice regardless of their feeding status. Ghrelin secretion in rodents follows a diurnal rhythm in rodents (37, 38), therefore we cannot exclude that these differences in strain are merely a byproduct of different circadian sleep/wake behaviors in S1 and BL6 lines.

Interestingly, elevated plasma acyl ghrelin has been suggested as a persistent biomarker for chronic stress exposure that mediates vulnerability to stress enhanced fear in a translational study (39, 40) and another study in patients found that therapy resistance in panic disorder and depression was correlated with increased acyl ghrelin levels (11). While stress (and chronic stress in particular) increases the acyl ghrelin concentration in the blood (13, 41–43), it should be taken into consideration that chronically elevated plasma acyl ghrelin at baseline may also represent a feature of individuals that are particularly susceptible for the development of stress-related behaviors or psychiatric disorders. Indeed, some studies found elevated (acyl) plasma ghrelin levels in different anxiety disorders (9–11). A recent population-based cohort study reported that total ghrelin levels were negatively associated with pathological anxiety, but (non-significantly) higher in subjects with clinically relevant anxiety symptoms than in those without (12). Generally, these findings suggest that ghrelin plays a role in the regulation of fear and anxiety-like behaviors and/or is altered by either previous stress exposure or generally elevated in stress-susceptible individuals. Along these lines, it seems plausible that impaired fear extinction, which was proposed as a risk factor for treatment resistance in anxiety and PTSD patients (44, 45), could also be associated with elevated acyl ghrelin levels. In our study, we did not directly address this hypothesis by correlating behavioral parameters of extinction-impaired S1 mice with plasma acyl ghrelin levels, so it remains to be tested in (other) rodent models and, most importantly, psychiatric patients.

Although systemic MK0677 administration induced an eating response in S1 as well as BL6 mice strains, it did not attenuate the extinction-impaired phenotype of S1 mice. This finding is in line with our results from experiments in extinction-competent BL6 mice, in which neither systemic nor central (ventral tegmental area) injection of MK0677 had an effect on fear extinction (29). Other authors, however, previously reported that MK0677 reduced fear memory strength in unstressed rats and enhanced fear extinction C57BL/6 mice after systemic or intra-amygdalar application (14, 19). MK0677 doses in our experiments (2–10 mg/kg) were in the range other studies reporting central drug effects used for i.p. administration in rodents (14, 43, 46). Moreover, these doses were already associated with a strong instatement of hunger, as shown here and in our previous study (29), which could potentially confound behavioral readouts. Thus, we do not believe that testing higher doses would have experimental or translational value. While only a few studies are published which tested behavioral effects of small molecule GHSR agonists like MK0677, there are abundant reports of anxiolytic, fear-attenuating and antidepressant effects of ghrelin peptide administration (13, 15–17, 47). Other studies contradict these findings and some even suggest anxiogenic or fear-enhancing effects after peripheral or central administration of acyl ghrelin (43, 48–51). Interestingly, most studies that reported a beneficial effect of GHSR agonism on fear-, anxiety-, and depressive-like behaviors were conducted in previously stress-exposed rodents, while findings in unstressed rodents were more mixed [for summary see (6)]. Thus, it seems likely that the effect of GHSR agonists may depend on various (experimental) factors like previous stress exposure or also handing routines and different behavioral paradigms, which per se could be perceived as more or less stressful by the tested animals. Other reasons for the inconsistent findings could involve differences in age, sex and strain of the used rodent models, various administration routes (e.g., central vs. peripheral), the time of day of testing and also the feeding status of the animals (6).

As an alternative to GHSR agonism, we studied whether overnight fasting (as a method to enhance endogenous ghrelin signaling) would rescue the extinction-impaired phenotype of S1 mice. Several studies have demonstrated that caloric restriction or short-term fasting strongly facilitate within-session fear extinction and extinction memory formation in extinction-competent mice (18–20) or reduce anxiety-like behaviors (13, 21) via different mechanisms, but probably also in a GHSR-dependent manner (13, 19, 21). In contrast, we previously found no effect of overnight fasting on fear extinction or extinction retrieval in BL6 or GHSR-KO mice (29). Also, in S1 mice, overnight fasting did not influence within-session extinction or extinction memory consolidation. In the fasted state, S1 mice even showed higher freezing levels than non-fasted controls, while in extinction retrieval sessions, when all animals were fed, there was no significant difference in the freezing behavior of both treatment groups. This suggests that that the stressful nature of an overnight fast may have had adverse effects in this anxiety-prone mouse strain (52), which resolved after re-feeding. The fact that we could not confirm the reports of a beneficial effect of fasting in unstressed rodents raises the question to which experimental parameters this effect may be susceptible and whether it is robust enough to translate into clinical studies. A recent meta-analysis found that fasting or caloric restriction decreased anxiety, stress and depression levels in a few small population-based studies (53) and one study in healthy human participants reported that short-term fasting improved extinction retention and reduced return of fear phenomena, which was correlated with plasma ghrelin levels (22). However, while these preliminary results may seem encouraging, they do not form a solid foundation to recommend fasting as a suitable treatment option at this point. To our best knowledge, no successful clinical trial of a fasting intervention in anxiety patients has been conducted to date. Also, with regard to the translational value of fasting experiments, it should be considered that forced fasting of experimental animals, who do not comprehend the context of their distress, and controlled food restriction within the framework of a clinical study may not deliver comparable results.

In summary, our data suggest that an activation of the ghrelin system does not facilitate fear extinction (in unstressed individuals), but underline the possibility that elevated acyl ghrelin could be a biomarker for fear-susceptible and extinction-impaired individuals. With regard to the divergent reports of ghrelin’s role in stress and mood regulation (6), our results highlight once more the need to investigate which factors may influence the effects of fasting and GHSR agonism on fear- and anxiety-like behaviors in order to better understand the consequences of (therapeutically) manipulating the ghrelin system and guarantee the greatest benefit for patients from future clinical trials.

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

This animal study was reviewed and approved by Austrian Federal Ministry of Science and Research.

NS and DD conceptualized and overviewed the project. EF performed all experiments and data analysis with methodological input from AP. EF wrote the manuscript. DD and NS reviewed and revised all sections. All authors read and approved the final manuscript.