Pregnenolone effects on provoked alcohol craving, anxiety, HPA axis, and autonomic arousal in individuals with alcohol use disorder

Abstract

Rationale

Chronic alcohol intake down-regulates GABAergic transmission and reduces levels of neuroactive steroids. These changes are associated with greater stress dysregulation and high alcohol craving which in turn increases relapse risk.

Objectives

This study tested whether potentiation of the neurosteroid system with pregnenolone (PREG), a precursor to neuroactive steroids and known to increase GABAergic transmission, will normalize chronic alcohol-related stress adaptations in the hypothalamic–pituitary–adrenal (HPA) axis and autonomic responses and reduce alcohol craving to significantly impact relapse risk.

Methods

Forty-three treatment-seeking individuals with alcohol use disorder (AUD) were randomized to placebo (PBO) or supraphysiologic pregnenolone doses of 300 mg or 500 mg treatment using a parallel-between subject design as part of a larger 8-week pilot clinical trial. In week 2, they participated in a 3-day laboratory experiment where on each day they self-administered the assigned study drug in the laboratory and were then exposed to 5-min personalized guided imagery provocation of stress, alcohol, or neutral/relaxing cues, one condition per day on separate days, in a random, counterbalanced order. Repeated assessments of alcohol craving, anxiety, HPA axis, heart rate (HR), systolic (SBP), and diastolic blood pressure (DBP) and serum pregnenolone levels were made on each day.

Results

Pregnenolone levels were significantly increased in the PREG groups versus PBO. PREG treatment decreased stress- and alcohol cue-induced craving and dose-specifically reduced stress-induced anxiety in the 300 mg/day group. Both PREG doses compared to PBO also normalized CORT/ACTH and increased stress-induced HR, stress- and cue-induced SBP, and in the 300 mg PREG group cue-induced DBP responses relative to neutral condition.

Conclusions

Findings indicate that pregnenolone decreases stress- and alcohol cue-provoked craving and normalizes HPA axis and autonomic arousal in individuals with AUD, thereby supporting the need for further assessment of pregnenolone in the treatment of AUD.

Introduction

Approximately 9% of the US population suffers from current alcohol use disorders (AUD) (Grant et al. 2017), with $249 billion/year in alcohol-related health costs (Substance Abuse and Mental Health Services Administration (US) and Office of the Surgeon General (US) 2016). While acute low to moderate alcohol in social drinkers stimulates autonomic and neuroendocrine arousal (see (Blaine and Sinha 2017) for review), chronic alcohol use results in significant adaptations in stress arousal with increased basal autonomic tone and blunted autonomic and HPA axis responsivity (Adinoff et al. 1995; Ingjaldsson et al. 2003; Lee and Rivier 1997; Richardson et al. 2008; Shively et al. 2007; Sinha 2008; Sinha et al. 2009; Thayer et al. 2006; Wand and Dobs 1991). Furthermore, this multilevel stress dysfunction occurs along with increases in provoked and basal alcohol craving (Breese et al. 2005; Fox et al. 2007, 2005), thereby suggesting a link between stress pathophysiology and reward dysfunction that is associated with increased compulsive alcohol motivation.

Previous work in our laboratory has shown that exposure to stress and alcohol cues consistently increases alcohol craving along with blunted autonomic and HPA axis responses in early abstinent individuals with AUD relative to social drinkers (Sinha 2008; Sinha et al. 2009; Sinha et al. 2011). This multilevel stress dysregulation also parallels blunted corticostriatal stress and reward circuitry to stress and cues, and both peripheral and neural stress pathophysiology along with high alcohol craving have been found to predict great relapse risk and poor treatment outcomes (Blaine et al. 2020; Seo et al. 2013; Sinha et al. 2011). Importantly, recent findings also show high alcohol craving at treatment entry and daily increased levels of stress-related alcohol craving prospectively predict poor outpatient alcohol use outcomes (Martins et al. 2022; Wemm et al. 2019). These findings suggest that normalizing the multilevel alcohol-related stress dysregulation alongside reductions in alcohol craving may improve alcohol relapse risk and drinking outcomes (Blaine and Sinha 2017; Milivojevic and Sinha 2017).

Stress responses are centrally regulated by GABA (Herman et al. 2004), with differential effects in acute versus chronic stress and alcohol exposure states. Normal GABAergic transmission is a potent modulator of autonomic responses and neuroactive steroids that enhance GABAergic transmission and are increased in response to acute stress and acute alcohol exposure are involved in modulating HPA axis responsivity and return to homeostasis (Biggio et al. 2007; Girdler et al. 2001; Morrow et al. 1999, 2006, 2001). Chronic alcohol and chronic stress states, on the other hand, result in neuroadaptations in GABAergic transmission and in neuroactive steroids with opposing effects via downregulation of GABAergic transmission and blunted/suppressed neuroactive steroid levels (Milivojevic et al. 2019; Morrow et al. 2006; Purdy et al. 1991). Neuroactive steroids are synthesized from the precursor pregnenolone, which is rapidly converted to downstream GABAergic neuroactive steroids such as allopregnanolone and pregnanolone (Sripada et al., 2013) but also additional downstream steroid metabolites that have active neural actions and may be excitatory or opposite to GABAergic action (Mellon 1994). This suggests that enzymatic activity within the complex neuroactive steroid pathway may be involved in modulating differential neuroactive steroid functions (Compagnone and Mellon 2000; Rupprecht 2003). Preclinical studies have shown that repeated alcohol exposure and withdrawal decrease GABAergic neuroactive steroids and downregulate the enzymes involved in conversion of pregnenolone to its downstream metabolites, such as allopregnanolone (Cagetti et al., 2003). These data suggest that neuroactive steroids may play an important role in regulating stress and that this system may be perturbed by chronic alcohol exposure.

Previous research from basic science studies indicates that pregnenolone and its GABAergic metabolites normalize the physiologic stress response and reduce anxiety and also regulate alcohol motivation by reducing alcohol-seeking and intake (Besheer et al. 2010; Janak et al. 1998; Morrow et al. 2006). These preclinical findings suggest that pregnenolone may be acting by normalizing stress system upregulation and potentially decreasing related compulsive alcohol-seeking, but whether pregnenolone would have beneficial effects in individuals with AUD has not been tested thus far. Therefore, we conducted a pilot experimental study in treatment engaged individuals with AUD who were part of an ongoing preliminary 8-week double-blind placebo-controlled dose finding trial of two supraphysiologic doses of pregnenolone (PREG; 300 mg/day, 500 mg/day) versus placebo (PBO). In week 2 of the trial, patients participated in a 3-day laboratory experiment where on each day they brought in their morning study medication dose and self-administered the assigned study drug in the laboratory prior to the experimental exposure to 5-min personalized guided imagery provocation of stress, alcohol, or active control neutral/relaxing cues, one condition per day on separate days, in a random, counterbalanced order. Stress- and alcohol cue-provoked alcohol craving, subjective anxiety, HPA axis, and autonomic responses were assessed. On the basis of the previous research cited above, we hypothesized that acute pregnenolone dose administration compared to placebo in the context of chronic dosing will reduce stress- and alcohol cue-provoked craving and anxiety and normalize HPA axis and autonomic responses in treatment engaged individuals with AUD.

Methods

Participants

Forty-three treatment-seeking individuals (31 M/12F) with alcohol use disorder (AUD) who responded to local advertisements around the New Haven area participated in the study (Fig. 1 CONSORT diagram). Inclusion criteria consisted of men and women 18–68 years old, with current AUD and the ability to read English. Current AUD criterion was determined using the Structured Clinical Interview for the Diagnostic and Statistical Manual of Mental Disorders 5 (SCID-5; (First et al. 2015) and confirmed with positive urine toxicology screens for alcohol metabolite (ethyl glucuronide (EtG)) collected during the initial eligibility assessment period. Exclusion criteria included: DSM-5 substance use disorder for any psychoactive substance, other than alcohol or nicotine, including opiate use disorder and including heroin (assessed and confirmed via urine toxicology screen in addition to SCID-5); any psychotic disorder or current Axis I psychiatric symptoms requiring specific attention; and significant underlying medical conditions such as cerebral, renal, or cardiac pathology which in the opinion of study physician would preclude patient from fully cooperating or be of potential harm during the course of the study. Stable use of antidepressants and medications for medical conditions without known study drug interaction effects that did not preclude the participant from participating in the study as determined by the study physician were allowed to ensure generalizability of the study findings (see Table 1). All individuals underwent stringent medical assessments including electrocardiography and laboratory tests of renal, hepatic, pancreatic, hematopoietic and thyroid function, and a physical exam conducted by the study physician to determine study eligibility. Written and verbal consent was obtained from all participants, and the Human Investigation Committee of the Yale University School of Medicine approved the study.

Fig. 1.

CONSORT diagram shows individuals screened, consented and enrolled for this placebo-controlled experimental study of the effects of two doses of pregnenolone on stress- and cue-provoked craving, anxiety, HPA axis, and autonomic response

Table 1.

Demographic characteristics of the AUD sample (N = 43)

	Placebo (N = 14)	300 mg PREG (N = 14)	500 mg PREG (N = 15)
Gender (male [%])	10 [71%]	11 [79%]	10 [67%]
Race
Caucasian [%]	6 [43%]	7 [50%]	9 [60%]
African American [%]	6 [43%]	5 [36%]	5 [33%]
Hispanic [%]	2 [14%]	1 [7%]	1 [7%]
Others [%]	0	1 [7%]	0
Age (± SD)	44.6 ± 13.5	46.7 ± 11.8	48.1 ± 12.3
Years of education (± SD)	14.6 ± 2.7	13.9 ± 2.7	14.1 ± 2.3
No. of regular smokers [%]	6 [43%]	8 [57%]	8 [53%]
Years of alcohol use (± SD)	20.1 ± 13.5	22.3 ± 13.7	17.1 ± 11.8
CIWA score (± SD)	3.43 ± 3.65	3.36 ± 3.13	3.20 ± 2.98
Lifetime mood disorder [%]	5 [36%]	4 [29%]	6 [40%]
Lifetime anxiety disorder (incl. PTSD) [%]	5 [36%]	5 [36%]	6 [40%]
Number on concurrent medications [%]#	9 [64%]	6 [43%]	9 [60%]
0 Meds. (Nper group, [%])	5 [36%]	8 [57%]	6 [40%]
1 Med. (Nper group, [%])	5 [36%]	2 [14%]	4 [27%]
2 + Meds. (Nper group, [%])	4 [29%]	4 [29%]	5 [33%]

All variables: p > 0.05

^#Specific medications (Med.[s]) patients were on (identifies number of patients taking the specific medication): Adderall (1), fluticasone propionate/salmeterol (1), loratadine (1), allopurinol (1), amlodipine (6), umeclidinium bromide/vilanterol (1), thyroid (1), atenolol (1), atorvastatin (1), cannabidiol (1), celexicob (1), chlorthalidone (1), clonidine (1), cyclobenzaprine (1), dextroamphetamine (1), diclofenac sodium (1), gabapentin (1), hydrochlorothiazide (1), Sitagliptin (1), empagliflozin (1), escitalopram (2), lidocaine (1), lisinopril (2), losartan (1), metformin (3), mirtazapine (1), nabumetone (1), omeprazol (1), ramelteon (1), ropinirole (1), rosuvastatin (1), montelukast (1), sumatriptan (1), levothyroxine (2), trazodone (2), doxy-lamine (1), bupropion (1). Note: There were no significant differences by medication group in frequency of participants taking medications (p > .74) or in mean number of medications being taken by patients in each study medication group (p > .84)

Study procedures

Upon determination of eligibility, participants were randomized to receive daily placebo (PBO) or one of two doses of pregnenolone, 300 mg/day or 500 mg/day, in a double-blind manner for a period of 8 weeks as part of the larger clinical trial (NCT03872128), and those who volunteered for this experimental study were enrolled for the current project. For the current experimental study, scripts for the individualized guided imagery induction were developed (see below) using the well-established standardized procedures, as described in previous studies (Sinha 2009; Sinha et al. 2009) during week 1 of their treatment. In week 2, a 3-day laboratory experiment involving participation in 3 separate experimental testing sessions on 3 separate days was conducted. Research staff were blind to order of imagery condition presented per day and subjects also remained blind until imagery presentation. Order of imagery condition was randomized and counterbalanced across subjects.

Study medication dosing and compliance/adherence

Identical pregnenolone (150 mg and 250 mg strength) and placebo capsules that also contained 25 mg of Riboflavin were formulated by the Yale University research pharmacist (Investigational Drug Services, IDS) and prepared for dispensing in 1-week bottles. Participants took the study medication orally, twice per day. Study participants and investigators were blind to the medication condition. Medication randomization was conducted by the Yale Stress Center Biostatistician and randomization was balanced for age, sex, smoking status, AUD severity, and education. The Biostatistician also provided de-identified dummy subject IDs for the Medication group so that the current analyses could be conducted in a blinded manner. This experimental study was embedded within the larger study period of 8 weeks, where the study medication was provided in 1-week bottles for self-administration at home in the morning and evening at 8AM and 8PM during the 8-week trial. To ensure administration of study medication on laboratory experiment days, research participants were asked to bring in their medication bottle and take the morning dose on the 3 experimental days at 1PM in the presence of research staff before the experiment. This also allowed for the assessment of acute medication effects in the current experiment within the context of chronic dosing. Medication compliance with daily chronic dosing during the trial was achieved in 3 ways: (a) riboflavin detection in weekly urine provided by participants in the clinic; (b) using the smartphone-based video monitoring tool eMocha Mobile Health, Inc. (Baltimore, MD); and (c) blood levels of pregnenolone collected on each of the three laboratory experimental days.

Imagery script development procedures

Imagery script development was conducted in week 1 in a session prior to the laboratory experiment. Procedures are based on methods developed by Lang and his colleagues (Lang et al. 1980, 1983) and further adapted and validated in our previous studies (Fox et al. 2005; Sinha et al. 2009, 2000, 1992, 2003). Briefly, the stress imagery script was based on subjects’ descriptions of a recent “most stressful” adverse personal event that made them “sad, mad, or upset” that they were not able to control in the moment. “Most stressful” was determined by having the subjects rate their perceived stress on a 10-point Likert scale where 0 = not at all stressful and 10 = the most stress they felt recently in their life. Only situations rated as 8 or above were accepted as appropriate for script development (e.g. being fired from their job, marital conflict situation). The alcohol-related cue scripts were developed by having subjects identify a recent situation that included alcohol-related stimuli and resulted in subsequent alcohol use (e.g., walking by their favorite bar; watching others drink alcohol). Alcohol-related situations that were associated with negative affect or psychological distress were not allowed. A relaxing, non-physiologically arousing and non-alcohol related script was developed from the subjects’ description of a personal, relaxing situation (e.g., being at the beach; fall afternoon reading at the park). In addition to the script development, on the day of the first laboratory session, subjects were brought into the testing room in order to acclimatize them to specific aspects of the study procedures including the subjective rating forms and training in relaxation and imagery procedures, as previously described in Sinha et al. (2009).

Laboratory sessions (conducted across 3 separate days)

Participants were instructed to abstain from drinking alcohol after midnight prior to the laboratory sessions, which was confirmed with a negative breathalyzer test on the day of each laboratory session. Participants were brought into the testing room at 1:30PM, after a standard lunch provided at 1PM after self-administration of study drug. Patients who were smokers were allowed a smoke break immediately prior to 1:30PM in order to reduce potential nicotine withdrawal during the session. After settling in a sitting position on a reclining chair in an experimental testing room, a heparin-treated intravenous (IV) catheter was inserted at 2PM by the research nurse in the antecubital region of the subject’s non-preferred arm in order to periodically obtain blood samples. A Critikon Dinamap 120 Patient Monitor was also placed on the subject’s preferred arm, including a pulse sensor which was placed on the subject’s forefinger. This was followed by a 40-min adaptation period during which the subjects were provided relaxation instructions to ensure stable psychophysiological state prior to each guided imagery provocation per day. Immediately following the adaptation period, subjects were provided headphones and given the following instructions for the 5-min imagery procedure: “Close your eyes and imagine the situation being described, ‘as if’ it were happening right now. Let your body and mind get completely involved in the situation, doing what you would do in the real situation.” Alcohol craving and subjective emotion ratings, heart rate (HR), systolic (SBP) and diastolic blood pressure (DBP), and blood samples were collected at baseline (− 20 and − 5 min prior to imagery), immediately following imagery presentation (0) and every 15 min after the imagery period, up to 75 min (+ 15, + 30, + 45, + 60, + 75). See Fig. 2 for a schematic illustration of the laboratory sessions timeline.

Fig. 2.

Schematic illustration of the timeline for each of the 3-day experimental sessions conducted on separate days, with order of imagery condition randomized and counter balanced. Medication group was a between-subjects factor where subjects were randomized in parallel to either PREG 300 mg/day, PREG 500 mg/day, or PBO. Imagery condition was a within-subjects factor, where each subject was exposed to stress, alcohol, or neutral cue imagery on separate days. Subjects self-administered their morning study medication at 1:00PM on each day prior to the experimental sessions. Alcohol craving and subjective emotion ratings, heart rate (HR), systolic (SBP) and diastolic blood pressure (DBP), and blood samples were collected at baseline (− 20 and − 5 min prior to imagery), immediately following imagery presentation (0) and every 15 min after the imagery period, up to 75 min (+ 15, + 30, + 45, + 60, + 75)

Laboratory assessments

Alcohol craving

The desire for using alcohol was assessed using the Alcohol Urge Questionnaire (AUQ)-Brief, a brief well validated 8-item self-report craving scale (Bohn et al., 1995).

Anxiety

Study participants were asked to rate how tense, anxious, and/or jittery they feel using a 10-point visual analog scale (VAS) in which 0 = “not at all” and 10 = “more than ever.”

Heart rate and blood pressure

A Critikon Dinamap 120 Patient Monitor (GE Medical Systems, Tampa, FL) was used to assess heart rate, systolic blood pressure (SBP), and diastolic blood pressure (DBP) at the specific timepoints outlined above in the “Laboratory sessions” section.

HPA axis markers

Blood samples for measurement of adrenocorticotropic hormone (ACTH) and cortisol were obtained in heparinized tubes. All tubes were placed on ice immediately after drawing, and then aliquoted after being centrifuged at 4 °C within 30 min of collection. Blood samples for HPA axis measures were stored at − 80 °C and processed at the Yale Center for Clinical Investigation Core Laboratories using commercially available cortisol and ACTH radio-immuno-assay (RIA) kits from MP Biomedicals, LLC (Solon, OH, USA). For cortisol, the Corti-Cote Cortisol Solid Phase RIA kit was used, which has a sensitivity of 0.07 μg/dL and a 100.0% antiserum specificity for cortisol. For ACTH, the ImmuChem Double Antibody hACTH RIA kit was used, with a sensitivity of 5.7 pg/mL and a 100.0% specificity for ACTH.

Pregnenolone levels

Baseline (− 20 timepoint before imagery) and recovery time point (+ 75) blood samples were collected on each laboratory day to measure serum pregnenolone levels. All serum blood collection tubes were kept at room temperature and then aliquoted after being centrifuged after 30 min of collection. Samples were then stored at − 80 °C until processing at the Yale Center for Clinical Investigation Molecular Core Laboratories using a commercially available pregnenolone enzyme-linked immunosorbent assay (ELISA) kit (Eagle Biosciences, Inc., Nashua, NH). The ELISA kit has a sensitivity of 0.05 ng/mL and a 100% specificity for pregnenolone. The coefficients of variation (CV) of intra-assay and inter-assay were < 10.6% and < 14.5%, respectively. For this ELISA, the cross reactivity between pregnenolone and other steroids was 6% for progesterone, 4.7% for 5 alpha-androstanediol, 0.4% for pregnenolone sulfate, 0.3% for androstanedione, 0.2% for DHEAS, and less than 0.1% for several other steroids (e.g. androsterone, aldosterone, androstenedione, cholesterol, corticosterone, 5alpha-DHT, 17beta-estradiol, testosterone).

Pregnenolone levels at the − 20 and + 75 timepoints were averaged across all three days to assess pregnenolone doses effects across all participants. There was a high degree of consistency between initial pregnenolone levels across days (Cronbach’s alpha = 0.98) indicating excellent stability and association of pregnenolone levels.

Data and statistical analysis

All statistical analyses were performed using SPSS software (SPSS Inc., Version 26, Chicago, IL, USA) and using linear mixed effects (LME) models. Within-subjects factors of imagery condition (stress, alcohol, neutral cues), timepoint (− 20, − 5, 0, + 15, + 30, + 45, + 60, + 75) and the between-subjects factors of medication group (300 mg PREG, 500 mg PREG, placebo) were the fixed effects. Subjects represented the random effect. The Bonferroni test for multiple comparisons was used to analyze simple effects. Analysis of variance (ANOVA) was used to compare pregnenolone levels between medication groups and across timepoints. ANOVA and chisquare analyses were used to compare the medication groups on demographic variables. Figures were created with Graph-Pad Prism 9 (GraphPad Software Inc., San Diego, CA). Any non-significant findings were not explicitly reported in the results. Moreover, as significant effects of condition and timepoint (without medication group) were expected given our previous reports on validation of the experimental paradigm (Sinha et al., 2009, 1992, 2000, 2003, 2011), these effects are not specifically reported in the manuscript but are listed in Supplemental Table 1.

Results

Participants

The medication groups did not significantly differ on any of the demographic characteristics (Table 1). About one-third of the sample was female, and the sample was racially diverse with 51% of participants being Caucasian, 37% African American, and 9% Hispanic. Importantly, race distributions were not significantly different across the groups. The average age was 46.5 years, about half of the sample consisted of regular nicotine smokers (21 non-smokers, 22 smokers), and the prevalence of regular smoking was balanced across the groups and did not differ significantly. The sample reported a long history of regular alcohol use and also included about one-third of individuals with lifetime rates of mood, anxiety, and post-traumatic stress disorders and those who were stable on medications for depression or other medical conditions (see Table 1). Importantly, the study medication groups were balanced on these factors and not different from one another.

Pregnenolone levels

A main effect of medication group (F(2,37) = 8.55; p < 0.0001) showed that circulating levels of pregnenolone were significantly higher 105 min after acute study drug in the 300 mg PREG group (p < 0.05) and the 500 mg PREG group (p < 0.002) compared to the placebo group and did not differ significantly between the PREG groups. They remained significantly higher 210 min after acute study drug in the 300 mg PREG group (p < 0.05) and the 500 mg PREG group (p < 0.0001) compared to the placebo group (Fig. 3). For the − 20 baseline timepoint (105 min after acute study drug administration), mean (± standard error) pregnenolone levels were 2.08 ng/mL (± 0.491) in the placebo group, 3.55 ng/mL (± 0.552) in the 300 mg PREG group, and 4.36 ng/mL (± 0.508) in the 500 mg PREG group. For the last, + 75 recovery timepoint (210 min after acute study drug administration), mean (± standard error) pregnenolone levels were 1.99 ng/mL (± 0.491) in the placebo group, 3.83 ng/mL (± 0.554) in the 300 mg PREG group, and 5.38 ng/mL (± 0.508) in the 500 mg PREG group.

Fig. 3.

Circulating pregnenolone levels as a function of placebo vs. PREG treatment groups at 105 min (A) and 210 min (B) after acute study drug administration. Pregnenolone levels were significantly higher 105 min after acute study drug in the 300 mg PREG group (p = 0.05) and the 500 mg PREG group (p < 0.002) compared to placebo and remained significantly higher at 210 min after study drug in the 300 mg PREG group (p = 0.05) and the 500 mg PREG group (p < 0.0001) compared to placebo. All data are displayed as mean ± S.E.M

Baseline differences between medication groups on outcomes

There were no significant medication group differences in baseline responses for any of the outcomes below.

Alcohol craving

There was a significant medication group × condition interaction (F(4,990) = 2.5; p = 0.04). Further analysis of the interaction showed that the PBO group had significantly higher alcohol craving in the stress condition (p < 0.001) and the alcohol cue condition (p = 0.002) compared to neutral, while the 300 mg and 500 mg PREG groups did not see such increases in stress and cue-provoked craving compared to the neutral condition (p = n.s.); see Fig. 4A. Importantly, there were no significant medication group differences in alcohol craving in the neutral condition alone.

Fig. 4.

A Provoked alcohol craving as a function of imagery condition and placebo vs. PREG treatment groups. Alcohol craving was significantly higher in the stress condition (p < 0.001) and the alcohol cue condition (p = 0.002) compared to the neutral condition in the placebo group. In the 300 mg PREG and 500 mg PREG groups, stress and alcohol cue-provoked craving was not significantly different than neutral (p = n.s.). B Anxiety as a function of imagery condition and placebo vs. PREG treatment groups. Stress-induced anxiety was significantly higher in the placebo (p < 0.001) and the 500 mg PREG groups compared to the neutral condition, but not in the 300 mg PREG group (p = n.s.). All data are displayed as mean ± S.E.M

Anxiety

There was a significant medication group × condition interaction (F(4,991) = 2.7; p = 0.03). Further analysis of the interaction showed that stress-induced anxiety was higher in the PBO group (p < 0.001) and the 500 mg PREG group (p < 0.001) compared to the neutral condition, whereas the 300 mg group did not experience such stress-induced increases in anxiety (p = n.s.). Moreover, anxiety was higher in the stress compared to alcohol cue conditions in the 300 mg (p < 0.001) and the 500 mg PREG (p = 0.003) groups but not in PBO; see Fig. 4B. Importantly, there were no significant group differences in anxiety in the neutral condition alone.

HPA axis response

There was a significant medication group × condition interaction (F(4,742) = 5.2; p < 0.001). Further analysis of the interaction showed that there were no significant differences between conditions in the PBO group (p = n.s.) in CORT/ACTH ratio, indicative of blunted overall HPA axis response. However, the CORT/ACTH ratio was significantly higher in the stress relative to neutral condition in the 300 mg PREG (p = 0.005) and 500 mg PREG (p = 0.0002) groups and in the stress relative to the alcohol cue condition in the 300 mg PREG (p = 0.02) and the 500 mg PREG (p < 0.0001) groups. Across dose groups, CORT/ACTH response to stress was significantly higher in the 300 mg PREG group relative to PBO (p = 0.01); see Fig. 5. Additionally, there were no significant group differences in CORT/ACTH ratio in the neutral condition alone.

Fig. 5.

CORT/ACTH ratio as a marker of adrenal sensitivity. The placebo group (left) had blunted CORT/ACTH response across all three conditions with no significant differences across imagery conditions. The CORT/ACTH ratio was significantly higher in the stress relative to neutral condition (p = 0.005) and the stress relative to the alcohol cue condition (p = 0.02) in the 300 mg PREG group (middle) and the stress relative to neutral (p = 0.0002) and the stress relative to the alcohol cue condition (p < 0.0001) in the 500 mg PREG group (right). CORT/ACTH response to stress was significantly higher in the 300 mg PREG group relative to placebo (p = 0.01). All data are displayed as mean ± S.E.M

Heart rate

There was a significant medication group × condition interaction (F(4,849) = 7.7; p < 0.001). Further analysis of the interaction showed significant increases in HR in response to stress compared to the neutral condition in the 300 mg PREG group (p < 0.001) and in the 500 mg PREG group (p < 0.001), whereas in the PBO group, the HR response to the stress condition was not significantly different than neutral (p = n.s.). In response to the alcohol cue condition, higher increases in HR were observed compared to the neutral condition in both the PBO (p = 0.04) and the 300 mg PREG group (p < 0.001), whereas the 500 mg PREG group did not have such increase in response to the alcohol cue condition compared to neutral (p = n.s.); see Fig. 6A. In addition, there were no significant group differences in heart rate in the neutral condition alone.

Fig. 6.

A Heart rate (HR) response as a function of imagery condition and placebo vs. PREG treatment groups group. For the stress relative to neutral condition response (left), HR increased significantly in the 300 mg PREG group (p < 0.001) and the 500 mg PREG group (p < 0.001), but not in the placebo group (p = n.s.). For the alcohol cue relative to neutral condition response (right), HR increased significantly in both the placebo (p = 0.002) and the 300 mg PREG group (p < 0.001), but not in 500 mg PREG group (p = n.s.). B Systolic blood pressure (SBP) response as a function of imagery condition and placebo vs. PREG treatment groups. In response to the stress condition (left), SBP increased significantly compared to the neutral condition in the 300 mg PREG group (p < 0.001) and the 500 mg PREG group (p = 0.007), but not in the placebo group (p = n.s.). In response to the alcohol cue condition (right), SBP increased significantly compared to the neutral condition in the 300 mg PREG group (p < 0.001) and the 500 mg PREG group (p = 0.013), but not in the placebo group (p = n.s.). C Diastolic blood pressure (DBP) response as a function of imagery condition and medication group. The 300 mg PREG (p < 0.001) and the 500 mg PREG (p < 0.001) groups as well as placebo (p < 0.001) had significant increases in DBP in the stress condition compared to neutral. In response to alcohol cue compared to neutral, the 300 mg PREG group had significantly higher DBP response (p < 0.001), whereas the 500 mg PREG group and the placebo group did not (p = n.s.). All data are displayed as mean ± S.E.M

Systolic blood pressure

There was a significant medication group × condition interaction (F(4,850) = 6.4; p < 0.001). Further analysis of the interaction showed that the 300 mg PREG group had higher increases in mean systolic blood pressure (SBP) in stress (p < 0.001) and alcohol cue (p < 0.001) conditions compared to the neutral condition. Similarly, the 500 mg PREG group had higher mean SBP increases in stress (p = 0.007) and alcohol cue (p = 0.013) conditions compared to the neutral condition, while the PBO group had no such increases in SBP (p = n.s.); see Fig. 6B. Importantly, there were no significant group differences in SBP in the neutral condition alone.

Diastolic blood pressure

There was a significant medication group × condition interaction (F(4,849) = 5.0; p < 0.001). Further analysis of the interaction showed that in response to stress, both the 300 mg PREG (p < 0.001) and the 500 mg PREG (p < 0.001) groups as well as PBO (p < 0.002) had significant increases in DBP compared to neutral. In response to alcohol cue compared to neutral, the 300 mg PREG group had significantly higher DBP response (p < 0.001), whereas the 500 mg PREG and the PBO groups did not (p = n.s.); see Fig. 6C. Lastly, there were no significant group differences in DBP in the neutral condition alone.

Discussion

This is the first study to assess the physiologic, neuroendocrine, and subjective effects of acute oral pregnenolone (150 mg and 250 mg) doses versus placebo in the context of chronic twice daily dosing of pregnenolone doses in treatment seeking adults engaged in an 8-week pilot study of pregnenolone treatment for AUD. Effects of the acutely administered pregnenolone versus placebo on laboratory stress- and alcohol cue-induced alcohol craving and anxiety, as well as stress-induced endocrine and physiologic arousal, were evaluated. In addition, because we looked at a verified acute dose on the laboratory days in the context of chronic dosing, the effect of an acute pregnenolone administration is not clear in this study as we cannot rule out the findings resulting from the chronic dosing regimen. It is certainly possible that chronic dosing regimen led to more sustained system-level changes such as adaptations in the expression of enzymes that metabolize pregnenolone to downstream GABAergic and non-GABAergic neuroactive steroids or other system-level changes. Future studies may assess acute versus chronic dosing of pregnenolone to tease apart any potential differences in stress and cue reactivity related to dosing regimens. Nonetheless, current findings indicate that an acute oral administration of exogenous supraphysiologic pregnenolone in the context of chronic dosing results in significantly increased and stable peripheral circulating pregnenolone levels in a dose-dependent manner. Furthermore, using our validated experimental protocol of stress- and alcohol-cue provocation, we found that stress- and cue-induced alcohol craving increased in the placebo group as expected, while no such increases were observed in the pregnenolone groups. Moreover, both pregnenolone doses also had strong effects on the HPA axis response to stress by increasing the stress-induced CORT/ACTH ratio, a proxy for adrenal sensitivity, and altered the autonomic response to stress and alcohol cue, such that pregnenolone increased stress-induced heart rate responses and stress- and alcohol cue-induced systolic and diastolic blood pressure responses. Subjective anxiety responses showed a dose-specific effect with lower anxiety associated with 300 mg dose of pregnenolone relative to placebo. These experimental findings of provoked alcohol craving and anxiety and biological stress measures suggest that specific pregnenolone doses may have clinically relevant benefits in reducing craving and anxiety in the real world and in normalizing chronic alcohol-related peripheral adaptations in the HPA axis and autonomic stress response systems that manifest in AUD.

The neuroactive steroid system plays a critical role in many vital functions of the human body, including direct modulation of GABAergic neurotransmission (Majewska et al. 1986) and regulation of the physiologic arousal response to stress (Besheer et al. 2010; Janak et al. 1998; Morrow et al. 2006; Regier et al. 2014). Neuroactive steroids are responsive to the acute effects of alcohol and also show alcohol tolerance as dependence develops (Milivojevic et al. 2014, 2011; Morrow et al. 2020, 2001). In animal models, alcohol at high doses increases neuroactive steroids in the periphery and in the brain (Barbaccia et al. 1999; Morrow et al. 1999; Sanna et al. 2004; VanDoren et al. 2000), but not at low doses (Porcu et al. 2010). In humans, plasma neuroactive steroids are increased following severe intoxication (Torres and Ortega 2003; 2004), but not moderate intoxication (Holdstock et al. 2006; Nyberg et al. 2005; Pierucci-Lagha et al. 2006), and their activation plays an important role in the subjective response to alcohol (Milivojevic et al. 2014; Morrow et al. 2001). In contrast, chronic exposure to alcohol and other drugs leads to dynamic adaptations in the neuroactive steroid system, such as decreases in neuroactive steroid levels in the brain and periphery (Milivojevic et al. 2019; Morrow et al. 2006, 2001; Purdy et al. 1991), including significant decreases in pregnenolone levels. As such, potentiation of the neuroactive system in individuals with AUD may have merit by offering a promising mechanism to normalize some of the physiologic, endocrine, and subjective adaptations that occur with chronic alcohol use.

We observed markedly lower provoked alcohol craving in both pregnenolone groups, which was not seen in the placebo group. In animal studies, neuroactive steroids such as allopregnanolone have been reported to influence alcohol-seeking and intake (Ford et al. 2008a, 2008b; Martin-Garcia et al. 2007), and additional preclinical evidence showed that administration of pregnenolone to alcohol conditioned rats reduced alcohol intake (Besheer et al. 2010), thereby suggesting that increasing levels of neuroactive steroids and their precursor pregnenolone in alcohol-altered states may reduce alcohol-seeking and consumption. While clinical research in AUD that directly targets neuroactive steroid potentiation is lacking, we previously reported that high dose progesterone versus placebo increased plasma levels of allopregnanolone in men and women with co-morbid alcohol and cocaine use disorder and that increased levels of allopregnanolone were associated with normalized HPA axis responses to stress, improved cognitive performance in response to stress and cue, and overall decreased craving in the laboratory (Milivojevic et al. 2016). However, the current study is the first that assessed and found beneficial effects of pregnenolone on stress- and alcohol cue-induced craving in individuals with AUD. Alcohol craving is a well-established clinical symptom of AUD, and both daily and provoked alcohol craving are predictive of alcohol relapse and treatment failure (Martins et al. 2022; Monterosso et al. 2001; Sinha et al. 2011; Wemm et al. 2019), thereby supporting its utility as an intermediate phenotype for assessment of alcohol relapse and treatment failure risk in human laboratory studies. Thus, current findings support the significance of assessing provoked alcohol craving as a key medications development target in efforts to improve alcohol use outcomes.

Pregnenolone dose-specifically reduced stress-induced anxiety at the 300 mg dose, but not at the 500 mg dose. The reduction in anxiety is clinically relevant, as alcohol use disorder and excessive drinking are associated with increased anxiety-related comorbidity (Grant et al. 2004), and a greater sensitivity to stress and anxiety and stress-induced and alcohol cue-induced anxiety has been observed in patients with alcohol use disorder compared with social drinkers (Glautier et al. 1992; Sinha et al. 2009). Furthermore, anxiety and negative mood in the laboratory has been shown to predict alcohol relapse risk in prospective outcome studies (Breese et al. 2005; Cooney et al. 1997; Fox et al. 2007; Fox and Sinha 2009; Sinha et al. 2011). It is important to note, however, that while there were significant effects on provoked anxiety in the current study, the magnitude of anxiety change from neutral was quite small, and these findings should be replicated in future studies and their relevance fully evaluated in the clinical context.

Neuroactive steroids have potent anxiolytic properties similar to those induced by other G ABA_A receptor potentiating drugs (Lambert et al. 2001; Majewska et al. 1986). Consistent with this previous research, current findings show stress-induced increases in anxiety in the placebo group, and the ability of 300 mg/day pregnenolone to reduce this sensitized anxiety response to stress may have tangible clinical relevance in the overall beneficial effects on treatment outcome and relapse risk. Interestingly, the 500 mg pregnenolone group did not demonstrate reduction in stress-induced anxiety. Since the medication groups did not differ on demographic characteristics, such as prevalence of anxiety or mood disorders or in concurrent medications, one possibility for this lack of effect in the 500 mg dose group may be that individuals in the 300 mg pregnenolone group could have had a higher conversion of pregnenolone to its anxiolytic metabolites, such as the highly anxiolytic allopregnanolone (Girdler and Klatzkin 2007), compared to the 500 mg pregnenolone group. The metabolic conversion of pregnenolone to downstream neuroactive steroids is regulated by enzymes such as 5α-reductase and 3α-hydroxysteroid dehydrogenase (HSD), which have consequential single nucleotide polymorphisms (SNPs) in their genetic code: these genetic variants have been shown to affect the protein expression of the enzyme and/or their enzymatic activity and therefore the production of anxiolytic neuroactive steroids (Milivojevic et al., 2014). While we did not assess the genetic variation in neurosteroid enzyme genes in this study, future studies in AUD individuals should also examine gene expression and assess whether pregnenolone-related anxiety effects in AUD individuals may be modulated by the genetic variation in neuroactive steroid synthesis genes. It is also possible that there is an optimal threshold beyond which other regulatory enzymes may modulate downstream metabolites that in turn may have opposing excitatory effects on anxiety and possibly alcohol intake. For example, there is evidence that the downstream neuroactive steroid dehydroepiandrosterone (DHEA) increases cocaine use in cocaine abusing individuals (Shoptaw et al., 2004). Importantly, unlike GABAergic neurosteroids, DHEA is a negative modulator of GABA and therefore an excitatory neuroactive steroid (Yadid et al., 2010) and as such may have opposite effects to those of GABAergic neurosteroids and possibly pregnenolone levels. While the current findings indicate that the 300 mg dose had optimal effects on anxiety in response to provocation in a laboratory setting, future studies may benefit from assessing downstream metabolites of pregnenolone and also the effects of pregnenolone on anxiety in real-world outcomes in individuals with AUD,

We found that pregnenolone at both doses significantly increased stress-induced CORT/ACTH ratio, a marker of adrenal sensitivity and HPA axis responsivity. Extensive evidence has demonstrated that chronic exposure to alcohol leads to significant HPA axis adaptations, marked by blunted stress-induced release of cortisol and ACTH (Adinoff et al. 1995; Errico et al. 1993; Lee and Rivier 1997; Richardson et al. 2008; Sinha et al. 2009; Sinha et al. 2011; Thayer et al. 2006; Wand and Dobs 1991) in individuals with AUD compared to healthy controls. Thus, treatment targets that normalize this stress dysfunction may contribute to improved control over cravings along with positively affecting abstinence and relapse rates (Milivojevic and Sinha 2018). The findings in the current study show that increases in pregnenolone may indeed normalize these chronic alcohol-related HPA axis adaptations and enhance overall cortisol response to stress-cue and alcohol-cue, thereby potentially restoring select components of the HPA dysregulation observed with chronic alcohol use.

Pregnenolone also had marked effects on autonomic responses to stress and alcohol cue provocation. This is an important finding, as individuals with AUD consistently show adaptations in the autonomic nervous system (Milivojevic and Sinha 2018; Shively et al. 2007; Sinha et al. 2009; Thayer et al. 2006). The autonomic nervous system is an important pathway that mediates the biological response to stress and to cue reactivity, in which the sympathetic component mobilizes arousal with increases in heart rate and blood pressure (Sinha 2008). Chronic alcohol use dysregulates the autonomic nervous system and is known to blunt the autonomic response to alcohol/cue and stress-cue exposure (for review see (Milivojevic and Sinha 2018)). For example, previous work has found that inpatient treatmentengaged, recovering individuals with alcohol use disorder displayed reduced heart rate responses to stress provocation relative to healthy controls (Fox et al. 2009; Sinha et al. 2009). Current findings suggest that pregnenolone may be reversing some of these autonomic adaptations and restoring the blunted heart rate and systolic and diastolic blood pressure responses to stress and cue provocation observed in the placebo group. For example, both pregnenolone doses increased stress-induced heart rate responses and systolic blood pressure stress and alcohol cue responses compared to placebo. Conversely, heart rate response to alcohol cue was blunted in the 500 mg pregnenolone group compared to both placebo and the 300 mg group, and diastolic blood pressure response to alcohol cue was increased only in the 300 mg pregnenolone group. As there is very little previous experimental assessment of autonomic arousal effects of pregnenolone in humans, we are unable to speculate about the source of this dose-related divergent responding. However, it does appear that autonomic responses to stress and alcohol cue have distinct differences that need further study to understand differential effects of stress and alcohol cues and also to assess their clinical relevance in AUD.

The present study has a number of strengths. The study utilized a well-controlled and validated human laboratory experimental design to assess effects of stress and alcohol cue reactivity in a double-blinded, placebo controlled, dose-finding study of pregnenolone. This study represents a real-world sample that included participants with AUD and included those with certain co-occurring psychiatric or medical condition who were stabilized on medication treatments. While close to half of individuals were on additional medications, the proportion on other medications across study medication groups were similar and thus cannot explain the current findings. In fact, despite close to half the participants being on a disparate range of other medications, the findings show a robust and consistent multi-level effect of pregnenolone doses relative to placebo on provoked laboratory responses, thereby supporting generalizability of the current findings to the broader population of individuals with AUD. We also tested these effects of pregnenolone on a racially diverse population. Limitations of the study include recruitment of low numbers of women that prevented us from assessing sex differences. Furthermore, while we carefully assessed pregnenolone levels during the experiment, we cannot speak to differences in pregnenolone levels with acute versus chronic dosing. Future studies should also assess baseline levels of pregnenolone prior to study drug administration. Despite these limitations, this is the first controlled experimental study to show the efficacy of two doses of pregnenolone in reducing alcohol craving and normalizing physiological and neuroendocrine reactivity to stress and to alcohol cues in individuals with alcohol use disorder. Findings support the further development and investigation of pregnenolone doses in evaluating potential therapeutic benefit in alcohol use treatment outcomes.