Introduction

Excessive alcohol use is a major economic and public health problem that is aggravated in individuals with post-traumatic stress disorder (PTSD) [1,2,3]. Individuals with PTSD display the greatest prevalence rates for alcohol use disorder (AUD) [1] and show poorer health outcomes [4] and higher rates of suicidality [5]. These negative consequences are magnified among military Veterans with comorbid PTSD/AUD including higher rates of AUD [6], 3-fold higher suicidal ideations [7], and higher aggression indices [8]. Thus, there is a need to understand the underlying mechanisms to assist in developing therapeutic strategies to alleviate trauma-related alcohol drinking.

Recent advances have suggested that FK506-binding protein 51 (FKBP5), a glucocorticoid receptor (GR) co-chaperone, as a promising novel therapeutic target [9, 10] to counteract over-active shared stress systems associated with alcohol use [9,10,11,12]. Epidemiological studies have also suggested that FKBP5 gene variants predict greater alcohol use [13]. A function of FKBP5 is to decrease the ligand binding affinity of GR, in turn delaying nuclear translocation and decreasing GR transcriptional activity [14, 15]. Enhanced FKBP5 activity is associated with increased corticosterone and blunted negative feedback inhibition of the stress axis [16], providing a basis to study FKBP5 inhibition as a potential avenue to increase negative feedback of stress systems and reduce comorbid stress disorders and alcohol use. In rodent models, stress exposure also increases FKBP5 expression in numerous brain regions associated with drug use [17]. Mice with a history of maternal separation stress and alcohol exposure also showed higher levels of Fkbp5 gene expression in cortical regions [13]. Thus, here, we examined the role of stress and alcohol history on the sensitivity to compounds that inhibit FKBP5 in our adapted model of comorbid PTSD/AUD [18].

Inhibitory avoidance-based shock stress models have been widely used to study stress-mediated drinking and anxiety across species of rodents [19,20,21,22]. We recently developed a familiar avoidance-based shock stress model of PTSD that elicits heightened alcohol drinking, fear overgeneralization, and irritability after rats experience “2-hit” traumatic shocks in a familiar environment [18, 23]. This translationally relevant model utilizes Pavlovian and instrumental conditioning of inhibitory avoidance that encompasses Pavlovian non-associative sensitization to fear, and instrumental aspects related to negative reinforcement [24]. The changes in alcohol drinking and anxiety-like phenotypes also translationally mimic psychogenic symptoms of comorbid PTSD/AUD. In addition, this model encompasses sex differences whereby stressed females display increased hyperarousal and decreased sleep maintenance during abstinence from alcohol when compared to stressed males [18].

We utilized this model of comorbid PTSD/AUD in a familiar contextual environment in male and female rats [18] and tested two FKBP5 inhibitors in two separate studies: benztropine (Study 1), an FDA-approved drug (Cogentin®) with repurposing potential that disrupts the association of FKBP5 with the glucocorticoid receptor (GR)-heat shock protein 90 complexes [25]; and SAFit2 (Study 2), a target-selective FKBP5 inhibitor [26]. Changes in alcohol drinking, hyperarousal, diurnal sleep patterns, fear overgeneralization, and irritability-like behavior were measured following administration of benztropine or SAFit2 during abstinence from alcohol. To better understand effects of treatment on peripheral stress indices, we analyzed plasma corticosterone. Overall, we hypothesized that the FKBP5 inhibitors would reduce comorbid phenotypes of PTSD/AUD, and that these effects would demonstrate important sex differences in trauma-related behavioral responses and potentially in corticosterone.

Methods and materials

The Supplementary material includes detailed descriptions of all experimental procedures.

Animals

A total of 176 adult male (~300 g) and female (~200 g) Wistar rats were purchased from Charles River Laboratories. All rats were pair-housed separated by a perforated clear Plexiglas® divider and had access to food and water ad libitum. All experimental procedures were approved by The Scripps Research Institute Institutional Animal Care and Use Committee (Protocol #: 19-00024 and ACURO #:5-312-216591). All procedures followed the National Institutes of Health Guide for the Care and Use of Laboratory Animals (8th edition).

Experimental design overview

This report consists of two studies conducted in separate cohorts of unstressed and stressed rats (n = 8–10 per group). Study 1 examined the effects of benztropine (Cogentin®), a broad-acting and FDA-approved drug with FKBP5 inhibiting properties, and Study 2 examined a selective FKBP5 inhibitor, SAFit2, on alcohol drinking, diurnal sleep cycles, hyperarousal, fear overgeneralization, and irritability following familiar inhibitory avoidance shock stress in male and female rats. Unstressed control animals underwent similar procedures without the presence of the shock. Administration of FKBP5 inhibitors occurred on the last day of alcohol drinking and again prior to each PTSD-like behavioral test that was assessed 1 week apart. Blood plasma was collected to assess the acute effects of FKBP5 inhibitors on circulating corticosterone levels on the last day of alcohol drinking. See Fig. 1 for experimental design overview.

Fig. 1: Experimental design overview.
figure 1

To evaluate the role of stress and alcohol history on sensitivity to FKBP5 inhibitors, we administered benztropine (Study 1), a broad-acting FKBP5 inhibitor, or SAFit2 (Study 2), a target-selective FKBP5 inhibitor. Rats first received an inhibitory avoidance-based “2-hit” familiar shock stress. Two weeks later, all male (n = 8–10 per group) and female (n = 8–10 per group) rats then received 2-bottle choice (2BC) alcohol drinking for 4 weeks (20%; 2 h; Monday, Wednesday, Friday). On the last day of alcohol access, rats received an acute administration of vehicle or benztropine (0, 5, 10 mg/kg; IP) or SAFit2 (0, 10, 20 mg/kg; IP). Plasma corticosterone was also assessed at the end of the alcohol test day. The rats then received a 1-week break period and then were tested on trauma-related behaviors which include sleep bouts cycles using the comprehensive lab animal monitoring system (CLAMS), acoustic startle, fear context overgeneralization, and bottle-brush irritability. Benztropine or SAFit2 were re-administered on each trauma-related behavioral test day. Each behavioral test was separated 1-week apart. Unstressed controls underwent similar procedures without the presence of the familiar shock stress and were conducted as separate cohorts.

Full size image

Drugs

Benztropine was administered 2 h prior to each test [27], and the doses selected were 5 and 10 mg/kg based on prior work [27]. SAFit2 was administered 16 h prior to each test [28], and the doses selected were 10 and 20 mg/kg based on prior literature [29, 30]. The different pretreatment intervals (2 vs 16 h) were based on previous studies with these compounds [26,27,28, 31,32,33].

Familiar shock stress

The “2-hit” foot shock stress procedures were conducted using a familiar context as previously reported [18]. The foot shock stress (3.0 mA for 2 s) occurred on two single shock instances in the same contextual environment (familiar) 48 h apart. Latency to cross (s) to the dark compartment was recorded.

Voluntary alcohol regimen

Two weeks after the first foot shock, rats received an initial 48 h acclimation to the alcohol (20% v/v) in addition to the water bottle followed by intermittent (Monday, Wednesday, Friday), 2-bottle choice (2BC) limited access (2 h) for 4 weeks in their home cage as previously described [18]. Blood alcohol levels were measured on the last 2BC alcohol session before rats entered their abstinence period. To validate measures, intake of rats that still showed detectable alcohol levels by the end of the 2 h session were compared to those that no longer showed detectable levels (see Supplementary Fig. S6).

Comprehensive Lab Animal Monitoring System (CLAMS)

Rats were placed in a OXYmax-CLAMS units (Columbus Instruments) to assess their sleep maintenance during their inactive phase (lights on). Number of sleep bouts, average bout duration (min), max bout duration (min), total sleep (min), and total sleep (%) were recorded as previously reported [34].

Acoustic startle

Acoustic startle responses were recorded on a SR-Lab system (SR-LAB) as previously reported [18, 34]. Rats underwent a 5-min acclimation period in the startle chamber before the beginning of 75 pseudorandomized trials across a 30-min testing window.

Fear overgeneralization

To measure fear-related context memory, all rats were tested in a fear overgeneralization paradigm in a modified inhibitory avoidance box and were given a maximum of 10 min to enter the dark compartment as previously described [18, 23]. Latency to cross (s) into the dark compartment was recorded, and animals that never crossed were censored at 10 min (600 s).

Bottle brush irritability

Behaviors assessed were aggressive behaviors, which included biting, boxing, following, and mounting, as well as defensive behaviors, including startling, digging, freezing, climbing cage walls, vocalizing and attempting to escape and general explorative behaviors such as grooming, rearing, and exploring as described in ref. [18].

Corticosterone analysis

Plasma samples were analyzed using commercially available MilliPlex® kits (Millipore Sigma) specific for corticosterone as previously reported [18]. Blood plasma collection was conducted for all rats immediately after drinking (4 h into their dark cycle) on the same day of the treatment with FKBP5 inhibitors. The samples were analyzed on a MAGPIX® system using xPONENT® software. The intra-assay (%) CVs were <10%.

Statistical analysis

Separate generalized linear models (ANOVA or rank regression) were used to analyze outcome data from each experiment and sex in Study 1 (Benztropine) and Study 2 (SAFit2). Model 1 analysis only involved stressed groups and included drug (Benztropine: 0, 5 or 10 mg/kg and SAFit2: 0, 10 or 20 mg/kg) as between-subject factors. Model 2 included stress condition (unstressed versus stressed), drug (Benztropine: 0 or 10 mg/kg and SAFit2: 0 or 20 mg/kg), and the interaction of stress and drug as between-subject factors. Corticosterone data were analyzed in a similar manner. A sub-analysis derived from Model 2 was used to compare vehicle-treated unstressed versus stress groups in both Study 1 and Study 2. Behavioral outcome data were performed using SAS version 9.4 (SAS Institute Inc.). Blood plasma data were analyzed on GraphPad Prism Version 8. Adjustments for multiple comparisons were conducted by using the false discovery rate (FDR) correction method. Full linear model results tables are illustrated in Supplementary Tables S1–S7. We report significance values throughout the results and figure legends. All graphs were generated using GraphPad Prism version 8.

Results

“2-hit” shock stress in a familiar environment increased alcohol intake and some trauma-related behaviors

We compared vehicle-treated unstressed versus shock-stressed rats (from Model 2 analysis) and found stress phenotypes generally consistent in direction and magnitude with prior work using the familiar “2-hit” model described in Steinman et al. (2021). To allow comparison to prior work and across cohorts, we report in Supplementary Tables S1 and S2 the mean differences for the observed effect of stress in vehicle-treated controls.

FKBP5 inhibitors reduce alcohol drinking after familiar shock stress history

We measured whether benztropine (Study 1) or SAFit2 (Study 2) decreased 2BC alcohol drinking following familiar shock stress history. In Study 1, our Model 1 analysis revealed that benztropine significantly reduced alcohol preference in shock-stressed males at the 5 mg/kg dose versus vehicle-treated controls (p < 0.05; Fig. 2b and Supplementary Table S4) but not alcohol intake (p > 0.05; Fig. 2a and Supplementary Table S4). There were no differences observed in our Model 2 analysis across male groups for alcohol intake or preference (p’s > 0.05; Fig. 2c, d and Supplementary Table S5). Our Model 1 analysis in shock-stressed females revealed that benztropine also reduced alcohol preference at the 5 and 10 mg/kg dose versus vehicle-treated controls (p < 0.05; Fig. 2f and Supplementary Table S4) but not alcohol intake (p > 0.05; Fig. 2e and Supplementary Table S4). Our Model 2 analysis revealed that benztropine generally reduced alcohol preference in females (p < 0.05; Fig. 2h and Supplementary Table S5), but not alcohol intake (p > 0.05; Fig. 2g and Supplementary Table S5).

Fig. 2: FKBP5 inhibitors reduce alcohol drinking after familiar shock stress history.
figure 2

The top panels (a–h) reflect 2BC alcohol drinking following benztropine administration (Study 1: n = 8–10 per group; 0, 5, or 10 mg/kg) and the bottom panels (i–p) reflect alcohol drinking following SAFit2 administration (Study 2: n = 8–10 per group; 0, 10, or 20 mg/kg). Data are expressed as mean (±SEM) for alcohol intake (g/kg/2 h) or preference (%). The asterisks (*) denote a significant difference from vehicle-treated rats and the number sign (#) reflects a difference from unstressed control, *p < 0.05, **p < 0.01, ***p < 0.001 or #p < 0.05.

Full size image

In Study 2, our Model 1 analysis in shock-stressed males revealed that SAFit2 reduced alcohol intake and preference at the 10 and 20 mg/kg dose versus vehicle-treated controls (p’s < 0.05; Fig. 2i, j and Supplementary Table S6). Similarly, our Model 2 analysis also revealed that SAFit2 reduced alcohol intake in shock-stressed males at the 20 mg/kg dose versus vehicle-treated controls (p < 0.05; Fig. 2k and Supplementary Table S7). Across female groups, Model 1 and Model 2 identified no differences produced by SAFit2 for alcohol intake or preference (p’s > 0.05; Fig. 2m–p and Supplementary Tables S6 and S7).

FKBP5 inhibitors reduce some trauma-related behaviors after familiar shock stress and alcohol drinking history

FKBP5 inhibitors do not alter sleep bout cycles

We measured the effects of FKBP5 inhibitors on sleep bout cycles after familiar shock stress and alcohol drinking using CLAMS. In Study 1, our Model 1 and Model 2 analysis revealed that benztropine did not produce significant differences in the longest sleep bout (min) across male and female groups (p’s > 0.05; Supplementary Fig. S1a–d and Supplementary Tables S4 and S5).

In Study 2, Model 1 and Model 2 analysis also revealed that SAFit2 did not produce significant differences in the longest sleep bout (min) across male and female groups (p’s > 0.05; Supplementary Fig. S1e–h and Supplementary Tables S6 and S7).

FKBP5 inhibitors reduce startle response outcomes after familiar shock stress and alcohol drinking history

One week after sleep bout cycles assessments, we measured whether the FKBP5 inhibitors reduced hyperarousal states after familiar shock stress and alcohol drinking. In Study 1, our Model 1 analysis in shock-stressed males revealed that benztropine reduced 120 dB startle responses (mV) during trial 1 and final block at the 10 mg/kg dose versus vehicle-treated controls (p’s < 0.05; Supplementary Fig. S2a, c and Supplementary Table S4) but not across trial 2–6 (p > 0.05; Supplementary Fig. S2b and Supplementary Table S4) and 80–105 dB series (p > 0.05; Supplementary Fig. S3a and Supplementary Table S4). Our Model 2 analysis also revealed that benztropine produced reductions in 120 dB startle responses (mV) in males during trial 1 and final block and across 80–120 dB startle responses versus vehicle-treated groups (p’s < 0.05; Supplementary Figs. S4a, c and S5a and Supplementary Table S5) but not across trial 2–6 (p > 0.05; Supplementary Fig. S4b and Supplementary Table S5). In shock-stressed females, our Model 1 analysis revealed that benztropine reduced 120 dB startle response during the final block phase and across 80–105 dB stimuli (p’s < 0.05; Supplementary Figs. S2g and S4b and Supplementary Table S5) at the 5 and 10 mg/kg dose versus vehicle-treated controls, but not across trial 1 or trials 2–6 (p’s > 0.05; Supplementary Fig. S2e, f and Supplementary Table S6). Model 2 analysis also revealed that females display general reductions in 120 dB startle responses during the final block phase and across 80–120 dB startle responses (p < 0.05; Supplementary Figs. S4g and S5b and Supplementary Table S5). There were no differences produced by benztropine for prepulse inhibition across male and female groups (p’s > 0.05; Supplementary Figs. S2d, h and S4d, h and Supplementary Table S5).

In Study 2, our Model 1 analysis revealed that SAFit2 reduced startle responses across 80–105 dB startle responses in shock-stressed males at the 20 mg/kg dose versus vehicle-treated controls (p < 0.05; Supplementary Fig. S3c and Supplementary Table S6). There were no differences in our Model 1 analysis produced by SAFit2 across trial 1, trial 2–6, and final block (p’s > 0.05; Supplementary Fig. S2i–l and Supplementary Table S6). Our Model 2 analysis revealed that SAFit2 produced general reductions during 120 dB trials 2–6 and across 80–120 dB startle responses at the 20 mg/kg dose versus vehicle-treated groups (p’s < 0.05; Supplementary Figs. S4j and S5c and Supplementary Table S7) but not across trial 1 and final block (p’s > 0.05; Supplementary Fig. S3i, k and Supplementary Table S7). Lastly, we also observed no significant differences in Model 1 and Model 2 analysis produced by SAFit2 in males and females during prepulse inhibition (p’s > 0.05; Supplementary Figs. S2p and S3p and Supplementary Tables S6 and S7).

The selective FKBP5 inhibitor, SAFit2, reduces fear overgeneralization after familiar shock stress and alcohol drinking history

These data are illustrated as violin plots and were analyzed using rank regression modeling to account for non-normal distribution of the data where many rats were censored at 600 s during testing. In Study 1, our Model 1 analysis revealed no significant differences in shock-stressed males following benztropine administration versus vehicle-treated controls (p > 0.05; Fig. 3a). However, Model 2 analysis revealed that benztropine administration generally increased the latency to cross (min) into the dark compartment of the apparatus in males at the 10 mg/kg dose versus vehicle-treated controls (p < 0.05; Fig. 3b). Furthermore, our Model 1 and Model 2 analysis in females revealed no differences produced by benztropine for latency to cross into the dark compartment (p > 0.05; Fig. 3c, d).

Fig. 3: Selective inhibition of FKBP5 reduces fear overgeneralization after familiar shock stress and alcohol drinking history.
figure 3

The top panels (a–d) reflect violin plots for fear overgeneralization in males and females following benztropine administration (Study 1: n = 8–10 per group; 0, 5 or 10 mg/kg) and the bottom panels (e–h) reflect fear overgeneralization in males and females following SAFit2 administration (Study 2: n = 8–10 per group; 0, 10, or 20 mg/kg). Data are expressed as violin plots that display median and quartiles for latency to cross (s) into the dark compartment of the fear overgeneralization apparatus. The asterisks (*) denote a significant difference from vehicle-treated rats and the number sign (#) reflects a difference from unstressed controls, *p < 0.05 or ##p < 0.05.

Full size image

In Study 2, our Model 1 analysis in shock-stressed males revealed that SAFit2 reduced the latency to cross (min) into the dark compartment at the 20 mg/kg dose versus vehicle-treated controls (p < 0.05; Fig. 3e and Supplementary Table S6). There were no differences observed in our Model 2 analysis regarding SAFit2 effects on fear overgeneralization across male groups (p > 0.05; Fig. 3f and Supplementary Table S7). Furthermore, our Model 1 analysis in shock-stressed females revealed that SAFit2 reduced the latency to cross (min) into the dark compartment at the 10 mg/kg versus vehicle-treated controls (p < 0.05; Fig. 3e and Supplementary Table S6). Lastly, our Model 2 analysis revealed no differences produced by SAFit2 in females (p > 0.05; Fig. 3h).

Benztropine reduces aggressive bouts in females but not males

In Study 1, our Model 1 analysis in shocked stress males revealed no differences produced by benztropine across doses (p > 0.05; Fig. 4a and Supplementary Table S4). However, our Model 2 analysis revealed that benztropine generally reduced aggressive bouts in males at the 10 mg/kg dose versus vehicle-treated controls (p < 0.05; Fig. 4b and Supplementary Table S5). Furthermore, our Model 1 analysis in shock-stressed females revealed that benztropine reduced aggressive bouts at the 5 and 10 mg/kg dose versus vehicle-treated controls (p’s < 0.05; Fig. 4c and Supplementary Table S4). Our Model 2 analysis revealed similar marginal reductions in shock-stressed females in aggressive bouts at the 10 mg/kg dose versus vehicle-treated controls (p = 0.058; Fig. 4d and Supplementary Table S5).

Fig. 4: Benztropine reduces aggressive bouts after familiar shock stress and alcohol drinking in females.
figure 4

The top panels (a–d) reflect aggressive bouts in the bottle brush test following benztropine administration (Study 1: n = 8–10 per group; 0, 5, or 10 mg/kg) and the bottom panels (e–h) reflect aggressive bouts following SAFit2 administration (Study 2: n = 8–10 per group; 0, 10, or 20 mg/kg). Data are expressed as mean (±SEM) for aggressive bouts. The asterisks (*) denote a significant difference from vehicle-treated control rats and the number sign (#) reflects a difference from unstressed controls, *p < 0.05, **p < 0.01 or #p < 0.05.

Full size image

In Study 2, our Model 1 and Model 2 analysis revealed no significant differences by SAFit2 in aggressive bouts across male and female groups (ps > 0.05; Fig. 4e–h and Supplementary Tables S6 and S7).

FKBP5 inhibitors decrease corticosterone levels after familiar shock stress and alcohol drinking history

We assessed circulating levels of corticosterone from blood plasma collected at the end of the last 2BC session after administration of benztropine (Study 1) or SAFit2 (Study 2). In Study 1, our Model 1 and Model 2 analysis revealed that benztropine administration decreased corticosterone in stressed males at the 10 mg/kg dose (p’s < 0.05; Fig. 5a, b and Supplementary Table S3). In females, our Model 2 analysis revealed that benztropine generally increased corticosterone at the 10 mg/kg dose versus vehicle-treated groups (p < 0.05; Fig. 5d and Supplementary Table S3). Model 1 revealed that benztropine did not change corticosterone levels in stressed females (p > 0.05; Fig. 5c and Supplementary Table S3).

Fig. 5: FKBP5 inhibitors suppress plasma corticosterone after familiar shock stress and alcohol drinking.
figure 5

The top panels (a–d) reflect plasma corticosterone levels following benztropine administration (Study 1: n = 8–10 per group; 0, 5, or 10 mg/kg) and the bottom panels (e–h) corticosterone levels following SAFit2 administration (Study 2: n = 8–10 per group; 0, 10, or 20 mg/kg). Data are expressed as pg/mL (±SEM) for corticosterone. The asterisks (*) denote a significant difference from vehicle-treated controls and the number signs (#) denote a difference from unstress controls, *p < 0.05, **p < 0.01, *** p < 0.001 or, #p < 0.05, or ###p < 0.001.

Full size image

In Study 2, our Model 1 and Model 2 analysis revealed that SAFit2 reduced corticosterone in stressed drinking males at the 10 and 20 mg/kg dose versus vehicle-treated groups (p’s < 0.05; Fig. 5e, f and Supplementary Table S3). Similarly, our Model 1 and Model 2 analysis revealed that SAFit2 also reduced corticosterone levels in stressed females at the 20 mg/kg dose versus vehicle-treated groups (p’s < 0.05; Fig. 5g, h and Supplementary Table S3). Lastly, we acknowledge differences in our corticosterone results between unstressed controls across Study 1 and Study 2. This discrepancy may relate to differences in our pretreatment injection intervals, cohort variability, and prior benztropine or SAFit2 treatment effects.

Discussion

This report studied the effects of two FKBP5 inhibitors in an adapted model of comorbid PTSD/AUD [18]. Our model elicited similar magnitude of changes in alcohol drinking, fear overgeneralization, and irritability-like behavior as seen in our prior work [18]. Table 1 summarizes behavioral effects produced by FKBP5 inhibitors in stressed males and females. In Study 1, benztropine decreased alcohol preference in stressed males and females, while decreasing aggressive-like behavior in stressed females. Benztropine also reduced startle responses to acoustic stimuli in stressed males and females. Interestingly, benztropine only reduced circulating corticosterone in stressed males but not females. We also observed opposing effects of stress response following benztropine administration in unstressed drinking males via increased anxiety-like behavior on the fear overgeneralization test, suggesting off-target effects or anxiogenic-like effects of benztropine under previously unstressed conditions. In Study 2, the selective FKBP5 inhibitor, SAFit2, reduced alcohol drinking in stressed males but not females. Importantly, SAFit2 reduced fear overgeneralization in both stressed males and females as well as corticosterone levels. Neither FKBP5 inhibitor altered sleep bout structure across males or females. Taken together, these findings demonstrate that FKBP5 inhibitors, with some sexually dimorphic action, show promising therapeutic efficacy to reduce some aspects of trauma-related behaviors and alcohol drinking in a model of comorbid PTSD/AUD.

Table 1 Summary of FKBP5 inhibitor effects after familiar shock stress history.
Full size table

FKBP5 has gained significant interest for its putative involvement in modulating alcohol drinking in humans [11, 13] and rodents [13]. Thus, we hypothesized that administration of FKBP5 inhibitors, benztropine or SAFit2, would normalize alcohol drinking in stressed animals. Here, we demonstrated that the FKBP5 inhibitors decrease trauma-related alcohol drinking in stressed males and females. The reductions in alcohol drinking appear to be selective to animals exposed to traumatic stress history since we observed no drug effects on drinking in unstressed rats following FKBP5 inhibitors. Similar lines of evidence with FKBP5 inhibitors have shown that SAFit2 administration reduces alcohol consumption using 2BC procedures in male mice [31]. The same report also found that selective inhibition of FKBP5 decreased relapse-induced drinking [31]. Other reports suggest opposing effects of FKBP5 in knock-out mice, which exhibited increases in alcohol consumption across a range of alcohol concentrations [35] or a lack of effects produced by FKBP5 inhibition in selectively-bred high drinking mice [36]. The latter finding might suggest an inverse relationship in the role of FKBP5 in stressed versus unstressed conditions to promote alcohol drinking, increased sensitivity to FKBP5 inhibitors after stress history, or, alternatively, opposing effects resulting from constitutive loss versus acute pharmacological reduction of FKBP5 function.

Since FKBP5 is thought to have a strong pathophysiological role in PTSD [10, 14, 37,38,39], we administered the FKBP5 inhibitors, benztropine or SAFit2 and also assessed PTSD-like behaviors during abstinence from alcohol. We hypothesized that benztropine and SAFit2 would also reduce various trauma-related behaviors involved in PTSD, including sleep bout cycles, hyperarousal, fear overgeneralization, and irritability-like behavior [40,41,42]. Furthermore, we found that FKBP5 inhibitors reduced some aspects of PTSD-related behaviors such as hyperarousal and irritability-like behaviors (females only), with stronger effects exerted by benztropine. We also found that SAFit2 reduced fear overgeneralization in stressed males and females. Thus, targeting FKBP5 might help improve some trauma-related symptoms of PTSD which has received increasing interest due to its role in PTSD. A preclinical study using FKBP5 KO mice showed greater resiliency to the effects of social defeat stress and improved stress coping behaviors using forced swim test paradigms [43]. Moreover, co-administration of SAFit2 and escitalopram (i.e., selective serotonin reuptake inhibitor) increased mobility time on the forced swim test apparatus, an indication of improvement of stress coping behavior produced by these drugs [29].

We also demonstrated that FKBP5 inhibitors have some sex-specific effects on trauma-related behaviors and alcohol drinking. The reduction in alcohol drinking following SAFit2 administration were only observed in stressed males but not females. The reductions in aggressive bouts following benztropine were only observed in females. We also consider that trauma-induced increase in drinking were predominantly observed in the males but not females. This pattern of results may be explained by higher basal vulnerability rates of drinking in females in the model where females generally display enhanced drinking patterns compared to males [18]. Our results might suggest that some aspects of trauma and alcohol drinking may be differentially driven by FKBP5 mechanisms in males and females. There is also evidence that FKBP5 interacts not only with glucocorticoid receptors, but also with other steroid receptors, including progesterone, estrogen, and androgen. FKBP5 may inhibit progesterone receptor activity [44], while promoting estrogen and androgen activity [45, 46]. Indeed, gonadal hormones may relate to these differences, as a rodent study demonstrated that FKBP5 inhibition decreased stress-induced drug reinstatement at a timepoint when estrogen was low in females (during metestrus/diestrus) [47]. Future studies are needed to assess the interaction between estrous cyclicity/gonadal hormones on the sensitivity to FKBP5 inhibitors that may influence their behavioral effects on comorbid PTSD/AUD.

The rationale to consider potential sex differences in FKBP5 inhibition for normalizing PTSD responses and reducing alcohol drinking is based on a strong human literature on sex differences regarding disrupted physiological activity and in pharmacotherapy responses for PTSD/AUD. For example, alcohol-induced disruptions in biophysiological markers of adaptive stress response are more common in women than men [48]. The nature and extent of many alterations are also sex-specific (e.g., blunted physiological responses to stress cues, alcohol cues, and alcohol exposure; sensitized emotional response to stress; alterations in hormonal fluctuations) [49]. Furthermore, pharmacotherapy with naltrexone for AUD found that women exhibit poorer treatment response and experience a higher rate and mean number of adverse effects than men, yet while experiencing a significantly greater reduction in craving scores than men [50, 51]. Similarly, PTSD pharmacotherapies with serotonin reuptake inhibitors and several other medication types appear more effective in men than women, a sex difference that remains controversial despite clear differences in medication metabolism and pharmacodynamics [52, 53].

While broadly similar in reducing drinking and PTSD-like behaviors, benztropine and SAFit2 produced non-overlapping reductions on some behavioral assays (i.e., 2BC alcohol drinking, hyperarousal, and irritability-like behavior). This discrepancy might be explained by pharmacokinetic differences and pharmacological actions of these compounds among males and females. For instance, it is known that benztropine has non-FKBP5 actions and it is classically considered a blocker of cholinergic and histaminergic signaling as well as dopamine transport [54]. However, the functional relevance of its inhibition of FKBP51 remains important since benztropine has shown to increase GR activity in numerous human cell types, increase GR nuclear translocation in mouse neuronal culture through FKBP5 inhibition, and contains significant selectivity for FKBP51 over its functional opposing homolog, FKBP52 [25, 55]. Furthermore, SAFit2 modulates other non-GR actions of FKBP5 suggesting modulation of PTSD/AUD via non-direct GR mechanisms [56, 57]. Thus, this novel mechanism of inhibiting FKBP5 suggests that FKBP5 inhibitors could potentially alleviate PTSD and AUD comorbidity. In support of this, benztropine analogs can reduce cocaine self-administration in rats [58], and during cocaine withdrawal Fkbp5 mRNA is upregulated in brain regions associated with drug use [59]. Therefore, these studies support the notion that inhibiting FKBP5 may be effective in reducing PTSD and AUD comorbidity, distinctly across males and females, and some trauma-related symptoms.

There are some potential mechanistic links to consider based on prior work showing increased CeA GABAergic transmission after stress and alcohol drinking history [18, 23]. Here, we speculate that familiar shock stress and alcohol history synergistically may affect/alter CeA GABAergic signaling, possibly through FKBP5/GR mechanisms (e.g., facilitating ligand activation of GR and its subsequent translocation into the nucleus) [14]. This notion is also supported by other reports indicating that GR antagonists prevent and reverse increased alcohol self-administration associated with alcohol dependence [60,61,62,63,64] and genetically-selected, alcohol preferring models [36], reduced CeA GABA transmission in alcohol dependent rats [65], as well as Fkbp5 gene expression [66]. Future work is needed to identify the specific brain regions involved, and assess the interaction between GR translocation, feedback and expression following treatment with FKBP5 inhibitors in PTSD/AUD comorbidity. We also cannot exclude that other non-GR mediated functions of FKBP5 may contribute to observed effects, including interactions with cholinergic and histaminergic systems [54] as well as serine/threonine protein kinase AKT sites [56, 57].

Collectively, our results indicate that FKBP5 inhibitors reduced some trauma-related behaviors and alcohol drinking in a model of comorbid PTSD/AUD. Our work has strong translational value that suggests the repurposing of benztropine (Cogentin®) or investigation of SAFit2 may be beneficial in suppressing comorbid PTSD/AUD. Since there are no current FDA-approved treatments to alleviate comorbid PTSD/AUD [67], the applicability of this work might suggest that approaches to target shared molecular systems such as FKBP5 is one strategy to ameliorate these co-occurring disorders. Our work also suggests that males and females demonstrate unique biological differences that imply distinct responding to treatment with FKBP5 inhibitors, suggesting that treatment for comorbid PTSD/AUD might require a tailored approach in males and females. Lastly, future studies are required to provide a deeper understanding of the molecular and synaptic mechanisms of FKBP5 and its inhibitors.