Mifepristone as a pharmacological intervention for stress-Induced alcohol craving: a translational crossover randomized trial

Carolina L. Haass-Koffler, Molly Magill, Nazzareno Cannella, Joshua C. Brown, Elie G. Aoun, Patricia A. Cioe, Rajita Sinha, Robert M. Swift, Roberto Ciccocioppo, Lorenzo Leggio
doi: https://doi.org/10.1101/2023.01.02.23284122

Abstract

Preclinical and clinical work suggests that mifepristone (glucocorticoid receptor antagonist), may be a viable treatment for alcohol use disorder (AUD). The aim of this work was to translate our preclinical mifepristone study using yohimbine (α2 receptor antagonist) stress-induced reinstatement of alcohol-seeking to a clinical setting. This was a Phase 1/2, outpatient, cross-over, randomized, double-blind, placebo-controlled trial with non-treatment-seeking individuals with AUD (N=32). We investigated the safety, alcohol craving and consumption after oral administration of mifepristone (600mg daily for a week) in a human laboratory study comprised of administration of yohimbine in a cue-reactivity procedure and alcohol self-administration. Outcomes were assessed using Generalized Estimating Equations and mediation and moderation analyses assessed mechanisms of action and precision medicine targets. We did not observe serious adverse events related to the study drugs or study procedure and mild to moderate non-serious adverse events were reported by both study conditions. Also, there was no statistically-significant difference between the mifepristone and placebo in the hemodynamic response, alcohol subjective effects and pharmacokinetics parameters. Mifepristone significantly reduced alcohol craving and increased cortisol levels. Mifepristone-induced cortisol increase was not a mediator of alcohol craving. Moderation analysis with family history density of AUD (FHDA) and mifepristone, suggested that reduced craving was present in individuals with low, but not high FHDA. Mifepristone, compared to placebo, did not reduce alcohol consumption in the laboratory or in a naturalistic setting. This study successfully translated a preclinical paradigm to a human laboratory study confirming safety, tolerability and efficacy of mifepristone in an alcohol paradigm. Mediation analysis showed that the effect of mifepristone on craving was not related to mifepristone-induced increases in cortisol and moderation of FHDA suggested the importance of evaluating AUD endophenotypes for pharmacotherapies.

Clinical trial registration Clinicaltrials.gov; NCT02243709

IND/FDA 121984, mifepristone and yohimbine (Holder: Haass-Koffler)

Introduction

Stress on biological systems has been linked to depression, anxiety1 and alcohol use disorder (AUD)2,3. Stress combined with re-exposure to priming or to environment and cues previously associated with alcohol exacerbates reoccurring drinking episodes both in rodents4-6 and humans7,8.

Mifepristone, a glucocorticoid receptor antagonist, is a medication approved by the FDA for the treatment of hyperglycemia, secondary to endogenous Cushing syndrome, in adults who have failed surgery or are not candidates for surgery, as well as for the termination of early pregnancy. Mifepristone has been under investigation as a potential treatment for many neuropsychiatric disorders such as psychotic depression (data pooled from three studies: mifepristone n=833; placebo n=627)9, including AUD (mifepristone n=28; placebo n=28)10.

In our preclinical work, we demonstrated that systemic administration of mifepristone, as well as its infusion in the central nucleus of the amygdala, reduced yohimbine-induced reinstatement of alcohol-seeking in alcohol-dependent Long Evans rats11. The effect of mifepristone on reducing alcohol-seeking was also supported by studies in alcohol-dependent Wistar rats10.

One of the most challenging aspects in designing a human laboratory study is the inclusion of an acute stress condition intended to represent a comprehensive naturalistic environment in individuals with AUD. In this study, we utilize yohimbine, rather than other stressors (psychological, physical), to maximize the translational efforts from our preclinical model11. Yohimbine is a well-validated pharmacological tool12 that has been widely employed in preclinical alcohol research studies to evaluate the effect of noradrenergic activation as a proxy for physiologic stress11,13,14. As a pharmacological challenge, yohimbine was shown to activate the hypothalamic-pituitary-adrenal (HPA) axis, in addition to increasing sympathetic nervous system activity15 and increasing alcohol craving16. This model builds on the evidence that the glucocorticoid receptor plays a role in the reinstatement behaviors elicited by yohimbine11, and that noradrenergic activation has been linked to alcohol intake and reoccurrence during abstinence17.

Translational studies should evaluate discrepancies between preclinical and clinical studies to reduce the knowledge gap between preclinical and clinical setting. In a study of a rhesus macaque AUD model18, cortisol was reported as potential mediator of mifepristone effect on alcohol self-administration whereas in a trial with pooled clinical studies, mifepristone plasma concentration was a more powerful mediator of mifepristone effects on psychiatric symptoms than cortisol9. It is also important to evaluate possible discrepancies between different preclinical models of AUD to better evaluate precision medicine approaches based on AUD endophenotype. Mifepristone reduced alcohol self-administration in Wistar rats trained to consume alcohol, however, marchigian sardinian alcohol preferring (msP) rats, a genetically selected AUD model, were less responsive to mifepristone’s ability to reduce alcohol self-administration19. This information when translated in human laboratory study may inform pharmacotherapy approached based on family contribution to the development of AUD20.

The aim of this work was to translate our preclinical study on the effect of mifepristone on yohimbine stress-induced reinstatement of alcohol-seeking11 in a human laboratory study. The primary outcome of this study was to test in individuals diagnosed with AUD, the safety of oral administration of mifepristone (600mg/day/7-day), compared to placebo. We used a human laboratory paradigm comprised of oral administration of yohimbine (32.4mg/1-day) paired to a cue-reactivity procedure, a priming alcohol dose and alcohol self-administration in an open bar laboratory. Secondary outcomes included: assessment of alcohol craving and consumption during the laboratory procedures. Finally, mediation and moderation analyses were conducted to assess potential mechanisms of action and precision medicine targets.

Materials and methods

Study design, setting and approval

The study was a Phase 1/2, outpatient, cross-over, randomized, double-blind, placebo-controlled, human laboratory study (Figure 1) and reported following the Consolidated Standards of Reporting Trials (CONSORT) extension21. A crossover design was chosen for this study because the within-subject variation is less than the between subject variation and allows for recruitment of less participants. The study was conducted at the Center for Alcohol and Addiction Studies, Brown University, Providence, RI, USA from 2014-2021. The trial was approved by the Brown University Institutional Review Board, conducted under an FDA Investigational New Drug application (IND121984) and registered at clinicaltrials.gov (NCT02243709). Important amendment to the original clinical study included the inclusion of females within childbearing age, using non-hormonal, barrier contraceptive, to improve female representation in the study (February-1, 2019).

Figure 1
Figure 1 Schematic outline of the laboratory paradigm procedures.

Visit 1 (screening), Visit 2 (randomization, mifepristone or placebo), Visit 3: laboratory 1 (mifepristone or placebo), washout period (21 days), Visit 4 (follow-up and second condition: placebo or mifepristone) and Visit 5: laboratory 2 (opposite condition, counter balanced), Visit 6 (follow up).

Legend: ADS, Alcohol Dependence Scale; ACQ, Alcohol Craving Questionnaire; AUQ, Alcohol Urge Questionnaire; BAES, Biphasic Alcohol Effects Scale; BrAC, breath alcohol concentration; BTQ: Brief Trauma Questionnaire; CIWA-Ar, Clinical Institute Withdrawal Assessment for Alcohol-revised; ECG, electrocardiogram; FHDA, Family History Density of Alcoholism; HAMA/HAMD, Hamilton Anxiety and Depression Rating Scale; LEC, Life Event Checklist; PLC, PTSD Checklist; PSS, Perceived Stress Scale; SAFE-T, Suicide Assessment Five-Step Evaluation and Triage; SCID, Structured Clinical Interview for DSM-IV, STAI, Spielberger State Trait Anxiety; TLFB, Timeline Followback; UA, urine analysis

Participants

After signing a written informed consent document and performing a screening to assess inclusion/exclusion criteria (Methods S1), eligible, non-abstinent, individuals with AUD who were no seeking treatment, were randomly assigned, by computer allocation, to 7-day treatment with either daily 600mg mifepristone or placebo. After a 3-week washout period to allow cortisol levels to return to baseline after mifepristone administration22 and avoiding carryover effect, participants returned to the laboratory and received the crossover condition.

Study procedures, drugs, dose justification and compliance

Details of visit 1 (screening), visit 2 (randomization), visit 3 and 5 (laboratory sessions) and visit 4 and 6 (follow up) are published23 and reported in Methods S1. Mifepristone and matching placebo were provided by Corcept Therapeutics (Menlo Park, CA). The oral dose of yohimbine was based on studies that examined neuroendocrine activation in humans16,24,25 and compounded by a local pharmacy.

Statistical analysis

For all outcomes, we utilized an intention-to-treat (ITT) approach, where participants were examined based on their a priori randomized protocol and received at least one dose of the study medication (mifepristone or placebo)26. All analyses were conducted using Generalized Estimating Equations (GEE)27 with robust standard errors, and an unstructured correlation matrix, unless noted.

Primary outcomes

safety and tolerability assessed by the number of adverse events (AEs) of oral administration of mifepristone was assessed after 7 days in an outpatient setting, and when it was administered with yohimbine and alcohol during the laboratory paradigms. Additional safety and tolerability analysis included hemodynamic response, alcohol pharmacokinetics and subjective effects of alcohol between the mifepristone and placebo condition (χ2 test).

Secondary outcomes

Craving measures included alcohol craving questionnaire short form-revised (ACQ-SF-R)28. Additional craving analysis included alcohol urge questionnaire (AUQ)29, and cue-induced saliva output. Values of the water trials for each dependent variable were inserted as covariate in the model (allowed for the dependent variable to be specific for alcohol), time coded: t1=alcohol trial 1 and t2=alcohol trial 2. In the bar laboratory, alcohol consumption was measured by number of drinks consumed (t-test). In the outpatient setting, alcohol consumption was measured by self-report using the timeline follow back method, reported as heavy drinking days and drinks per week.

Mediation

analyses for cortisol level, on ACQ, AUQ and cue-induced salivary output, were conducted using a regression-based, Macro Estimating Model30 that estimated the indirect effect of a within-participant manipulation on outcomes. Mediation was tested using standard procedures (product of the a and b path coefficients), but difference scores were created for the mediator and outcome under mifepristone/placebo conditions. The dependent variables were ACQ, AUQ and cue-elicited salivary output and the mediator was the cortisol level after 7-day mifepristone (M1) or placebo (M2) administration. The indirect effect was tested with Monte Carlo CI (interval estimate confidence 95%).

Moderator

analyses for family history density of alcoholism (FHDA) on ACQ, AUQ and salivary output were conducted using dichotomous predictor variables31. Consistent with previous work32, we used the Family Tree Questionnaire33 to calculate the FHDA and dichotomized (median split into low/high)34. FHDA was set as the moderator (m0=FHDA low; m1=FHDA high) and drug (mifepristone/placebo) as within-subject factors. For the outcomes in the laboratory (ACQ, AUQ, cue-elicited salivary output and neuroendocrine variations), the model was specified to evaluate: main effect of FHDA, drug, and the moderator by drug interaction. For additional statistical details, see Method S1.

Results

Participants’ characteristics

The CONSORT diagram is reported in Figure 2 and sociodemographic and baseline clinical characteristics of the participants in Table S1. One hundred-fifty-five participants were screened on the telephone, 46 were screened in person and 32 were randomized. They received at least one dose of the study medication and included in ITT analysis. Additional information on participant retention can be found in the Results S1.

Figure 2
Figure 2 Consolidated Standards of Reporting Trials (CONSORT) extension for cross-over trials.

Primary outcome

Safety and tolerability of mifepristone, yohimbine and alcohol

There were no serious AEs when the study medication was co-administered with yohimbine and alcohol in the laboratory. We observed three non-serious AEs (mifepristone: n=0, 0%; placebo: n=3, 10%; p>.05). Two individuals had an emesis episode after yohimbine and alcohol administration and one individual experienced increased blood pressure after yohimbine administration but before alcohol administration, however, blood pressure normalized after the priming alcohol administration.

The safety and the tolerability of the laboratory procedures were also assessed by monitoring the hemodynamic function (SBP, DBP and HR), alcohol pharmacokinetic parameters (Cmax, Tmax and AUC0-40) and subjective response to alcohol (stimulation/sedation) (Figure 3). For SBP, we found no significant main effect for drug, a significant time effect such that SBP increased from baseline after cue-reactivity (p<.013), and a time by drug interaction, where these increases were observed only in the placebo after the cue-reactivity (t60min, p=.020) (Figure 3A). For DBP, we found a significant main effect for drug, such that DBP was lower in the mifepristone condition compared to placebo (p=.005), a main effect for time such that DBP increased from baseline to before (t30min, p=.001) and after (t60min, p=.002) the cue-reactivity, and a time by drug interaction, where these increases were observed only in the placebo after the cue-reactivity (t60min, p<.001) (Figure 3B). Finally, for HR, there was no significant main effect for drug, time or drug by time interaction (p’s>.05) (Figure 3C).

Figure 3
Figure 3 Hemodynamic function, subjective response and pharmacokinetic parameters after administration of yohimbine paired to a cue reactivity and alcohol self-administration paradigm.

A) SBP: no significant main effect for medication, a significant time effect such that SBP increased from baseline to after cue-reactivity (t60min, b=5.96; CI=1.27; 10.65; p<.013), and a time by medication interaction, where these increases were observed only in the placebo condition after the cue-reactivity (t60min, b=8.149; CI=1.27, 15.03, p=.020). B) DBP, a significant main effect for medication, such that DBP was lower in the mifepristone condition compared to placebo (b=-4.01; CI=-6.80, -1.215; p=.005), a main effect for time such that DBP increased from baseline to before (t30min, b=3.639; CI=1.41, 10.21; p=.001) and after (t60min, b=4.79; CI=1.83, 7.75, p=.002) the cue-reactivity, and a time by medication interaction, where these increases were observed only in the placebo condition after the cue-reactivity (t60min, b=6.221; CI=2.57, 9.88, p<.001). C) HR: no significant main effect for medication, time or medication by time interaction (p’s>.05). All data presented as mean±SEM. *p<.05 main effect; $p<.05 interaction.

After the yohimbine administration and cue-reactivity procedure, to assess the safety and tolerability of mifepristone when co-administered with alcohol, we measured alcohol pharmacokinetics via BrAC after the administration of a prime alcohol drink designed to rise BrAC to 0.03-0.05mg/l. There was no difference in mifepristone condition compared to placebo in the BrAC pharmacokinetic curve parameters (AUC, Tmax, Cmax) (p’s>05) (Figure 3D). Also, there was no significant difference in the mifepristone compared to placebo (main effect and interaction), on the alcohol subjective effect both in stimulation and sedation (p’s>.05) scales (Figure 3E-F). During the 7-day administration of mifepristone or placebo in an outpatient setting, with mifepristone, we did not observe serious AEs related to the study drugs or study procedure (Table S2, Results S1). Mild to moderate non-serious AEs were reported by both study conditions throughout the trial, with no difference (p’s>.05).

Secondary outcomes

Alcohol craving, urge and cue-elicited salivary output

After the administration of yohimbine, the effect of mifepristone, compared to placebo, on alcohol craving was assessed during a cue-reactivity procedure using ACQ. Additional analysis included AUQ and cue-elicited salivary response. Analysis of ACQ showed no main effect for drug (p>.05), but a significant main effect for time (p<.001) where increases of craving were observed in alcohol trial 2. A time by drug interaction suggested decrease of craving for the mifepristone compared to placebo (p=.007) at the alcohol trial 1 (Figure 4A). Analysis of AUQ showed a main effect for drug (p=.010) where a decrease in urge was observed in mifepristone compared to placebo, but not a main effect for time (p>.05). We found, however, a drug by time interaction in the alcohol trial 1 (p=.040), suggesting a lower urge in mifepristone compared to placebo. (Figure 4B). Analysis of cue-elicited salivary output revealed a significant main effect for drug (p<.001), where lower saliva output was observed in the mifepristone compared to placebo, and a significant main effect for time showed salivary decreases at the alcohol trial 2 (p<.001). A time by drug interaction was observed both at the alcohol trial 1 (p<.001) and alcohol trial 2 (p<.001), with decrease of salivary output in the mifepristone condition compared to placebo (Figure 4C).

Figure 4
Figure 4 The effect of mifepristone compared to placebo on alcohol craving, urge and cue-elicited salivary output and cortisol and mediation modle.

A) ACQ: no main effect for medication (p >.05), but a main effect for time (p<.001) and a time by medication interaction (p=.007) at the alcohol trial 1. B) AUQ: a main effect for medication (p=.010), no main effect for time (p>.05), but a medication by time interaction in the alcohol trial 1 (p=.040). C) cue-elicited salivary output: a main effect for medication (p<.001), a main effect for time (p<.001) at the alcohol trial 2, and a time by medication interaction at the alcohol trial 1 (p<.001), and alcohol trial 2 (p<.001). D) cortisol: a significant main effect for medication (b=.54; CI=.22, .85; p<.001), no main effect for time (p>.05), but there was a medication by time interaction both before (t30min: b=.634; CI=.31, .99: p<.001) and after (t60min: b=.49; CI=.29, -.91 p<.001) the cue reactivity. Mediation Model. Increase of cortisol level as mediator of alcohol craving, urge and salivary output after 7-day mifepristone administration before initiating any laboratory procedure. E) Total effect (c path); F) Direct effect (c’ path); Indirect effect (a x b path). ↓↑ represent effect of mifepristone on outcomes. All data presented as mean±SEM. *p<.05 main effect; $p<.05 interaction.

Cortisol

Analysis of saliva cortisol during the laboratory procedures, revealed a significant main effect for drug (p<.001), such that higher cortisol was observed in the mifepristone, compared to placebo, no main effect for time (p>.05), but there was a drug by time interaction, indicating higher cortisol levels both before (t30min p<.001) and after (t60min p<.001) the cue-reactivity (Figure 4D). This result further supports individuals’ medication compliance as cortisol increases with mifepristone administration22.

Finally, to test if participants responded to the laboratory procedures, the increase of the HPA was confirmed when we compared the value of cortisol levels collected at the screening visit (basal) to the values collected during the laboratory visits (stress) in the placebo (Results S1, Figure S1).

Cortisol as mediator of alcohol craving, urge and cue-elicited salivary output

The above results support the total effect of mifepristone on ACQ, AUQ, and cue-elicited salivary output outcomes at alcohol trial 1 (c path) (Figure 4E), as well as an effect of mifepristone on cortisol level (a path, described above Figure 4D). However, b path results did not show a relationship between cortisol and ACQ (p>.05), AUQ (p>.05) and cue-elicited salivary output (p>.05) at the alcohol trial 1. We also tested the effect combining alcohol 1 and 2: ACQ (p>.05), AUQ (p>.05) and cue-elicited salivary output (p>.05). As a result, the Monte Carlo CI around the product of the a and b path coefficients were non-significant (p’s>.05) (Figure 4F).

FHDA as moderator of alcohol craving, urge, cue-elicited salivary output and cortisol

For ACQ, there was a main effect for FHDA (p=.023) such that craving was higher for those with high FHDA compared to low FHDA. There was also a FHDA by drug interaction, suggesting lower craving in the mifepristone, compared to placebo, in those with low FHDA (p=.007) (Figure 5A). Similarly, for AUQ, there was a main effect for FHDA (p=.018), with increased urge in the high FHDA compared to the low FHDA group, and a FHDA by drug interaction (p=.028), suggesting lower urge in the low FHDA group for mifepristone, compared to placebo condition (Figure 5B). For cue-elicited salivary output, we found a main effect for FHDA (p=.035) with increased salivation in the high FHDA compared to the low FHDA group. In the FHDA by drug interaction analysis, we found that mifepristone, compared to placebo, reduced salivary output both in the low FHDA (p=.025) and in the high FHDA (p=.002) groups with greater reductions for high FHDA participants (Figure 5C). When we tested if FHDA was a moderator of HPA-axis response, we found a main effect of FHDA (p=.049), such that lower cortisol was observed in the FHDA high compared to the FHDA low participants. There was also a FHDA by drug interaction, such that there was higher cortisol level both in the FHDA low (p=.042) and in the FHDA high (p=.036) groups with greater increases for high FHDA participants (Figure 5D).

Figure 5
Figure 5 FHDA moderation analysis: the effect of mifepristone compared to placebo on alcohol craving, urge and cue-elicited salivary output.

A) ACQ: a main effect for FHDA (b=17.36; CI=2.43, 32.29; p=.023), a FHDA by medication interaction in those with low FHDA (b=-18.65; CI=-32.10, -5.20; p=.007). B) AUQ: a main effect for FHDA (b=-11.096; CI=1.89, 20.30; p=.018) in the high FHDA compared to the low FHDA, a FHDA by medication interaction (b=-9.781; CI=-18.53, -1.03; p=.028) in the low FHDA in the mifepristone group, compared to placebo. C) Salivary output: a main effect for FHDA (b=1.623; CI=.12, 3.13; p=.035) in the high FHDA compared to the low FHDA, a FHDA by medication interaction both in the low FHDA (b=-.588; CI=-1.17, -.20; p=.025) and in the high FHDA (b=-2.122; CI=-.43, -3.22, p=.002). D) Cortisol: a main effect of FHDA (b=-0.328; CI=-32.34, -5.22; p=.049), a FHDA by medication interaction, both in the FHDA low (b=0.413; CI=2.10, 5.20; p=.042) and in the FHDA high (b=.708; CI=2.15, 5.43, p=.036). All data presented as mean±SEM. *p<.05 main effect; $p <.05 interaction.

Participants responded to stress cue-reactivity paradigm (Results S1, Figure S2), FHDA was not related to differences in age, age of onset of AUD, alcohol dependence, baseline of craving (ACQ), urge (AUQ), stress (PSS, STAI trait and state), anxiety (HAMA) (all VIF<5)35 or trauma (assessed by the BTQ and LEC, criterion A for PTSD) (Spearman, p>.05), therefore indicating that FHDA was likely an independent moderator.

Alcohol consumption

In the open bar laboratory session, participants consumed a small number of standard alcohol drinks both in the mifepristone (0.8±0.3) and placebo (0.5±0.2) conditions, with no significant difference between conditions (p>.05). In an outpatient setting, neither mifepristone or placebo reduced alcohol consumption (p’s>.05) (Results S1, Figure S3).

DISCUSSION

The major finding of this study is that mifepristone, administered with yohimbine and alcohol, was safe for individuals diagnosed with AUD. We also demonstrated that mifepristone, compared to placebo, attenuates combined yohimbine/cue-induced alcohol craving in the laboratory. Mifepristone effect in reducing yohimbine-induced alcohol craving is independent of the mifepristone-induced increase of cortisol level and significantly reduced craving only in individuals with low family history of AUD (FHDA).

At the research level, our results cast important relevance as this was the first time that a preclinical paradigm that included the use of a noradrenergic pharmacological stressor (yohimbine)12, paired to a cue-reactivity (cue-reinstatement) and alcohol self-administration, was successfully and safely translated in a human laboratory study in individuals with diagnosis of AUD. We effectively translated an extensively used preclinical paradigm to a human laboratory study that investigates stress-induced alcohol craving, including a comprehensive model capable of integrating contributions of each alcohol-proximal variable (e.g., neuroendocrine variations, family history, stress and trauma factors) that may influence craving36. The robustness of the laboratory procedures was supported by the increase of blood pressure from baseline post cue-reactivity. The analysis of the laboratory session from baseline (screening) demonstrated that the participants responded to the procedures, as it is known that both cue-reactivity37 and yohimbine24 increase blood pressure.

At the clinical level, it was critical to assess the safety and tolerability of study drug when co-administered with alcohol, given the potentially serious consequences of drug-alcohol interactions. This assessment is important for novel38 and repurposed39 medication for treating AUD and consistent with the FDA40 and the European Community (EMA/CHMP/EWP/20097/2008) guidelines on the development of medications to treat AUD. Clinical screening included laboratory analysis to evaluate possible overlapping enzymatic pathways and limited renal activity for excretion of multiple drugs. Under this paradigm, we did not encounter serious AEs in either the mifepristone or in the placebo, and non-serious AEs were encountered in both conditions. After the alcohol prime, mifepristone compared to placebo did not did not alter the alcohol pharmacokinetics and did not affect the stimulation/sedation effects of alcohol, supporting the safety of mifepristone when co-administered with alcohol.

Mifepristone reduced the self-reported alcohol craving (ACQ and AUQ) at the first alcohol cue exposure. It is known that the second alcohol trial boosts alcohol craving41-43, in this study it is possible that the effect of mifepristone was washed out under the strong stress procedures. The cue-elicited salivary output, however provided an objective biomarker for the effect of mifepristone in craving reduction, in both trials. This observation is in line with original cue-reactivity studies showing that salivation, rather than be associated to conscious attention to alcohol (vision, smell and tactile experience), it is more pronounced in individuals with serious AUD and is a strong predictor of alcohol consumption in the first period after detoxification44. Furthermore, to support the validity of salivation as biomarker, it should be noted, that, higher mifepristone-induced cortisol level may induce acetylcholine production and, therefore, increased salivation (statistically reduced in both trials in the mifepristone, compared to the placebo, condition)45.

While this trial was developed based on our original preclinical11 and other translational10 literature related to mifepristone in AUD, subsequent preclinical studies in AUD model of rhesus monkeys showed that mifepristone decreased daily alcohol self-administration and the effect was mediated by mifepristone-induced increase in cortisol18. When we tested this hypothesis in our human study, first, we confirmed that cortisol increased after mifepristone administration, which is consistent with results in clinical setting9,22, human alcohol research10 and preclinical AUD models (alcohol dependent rats10 and monkeys18). However, we did not find that the effect of mifepristone on alcohol craving outcomes was mediated by increasing cortisol. Our results align with other clinical data showing that mifepristone’s effects on reducing psychotic symptoms were independent of the increased plasma cortisol and adrenocorticotropin hormone9. As this is a translational trial, we should evaluate further the discrepancy from the monkey study. First, the yohimbine administration was not done in monkeys. Also, it is possible that salivary cortisol (humans) is a more appropriate measure for adrenocortical function than blood (monkeys), as the salivary and serum total cortisol concentration have a non-linear relationship due to the rapid increase in salivary concentration once the serum cortisol-binding globulin is saturated46. Our clinical results, however, align with our preclinical data, which demonstrated that infusion of mifepristone directly into the amygdala suppressed yohimbine-induced reinstatement of alcohol-seeking, with corticosterone levels unaffected11. The involvement of the amygdala, rather than the effect on the negative feedback effect on the hypothalamus, was also later further supported by other translational AUD10 and psychiatric9 studies. Here, the mediation analysis suggests that the mifepristone effect in reducing yohimbine-induced alcohol craving is independent of the mifepristone-induced increase of cortisol level.

The second translational analysis was based on the work conducted in genetically selected msP rats, which demonstrated that mifepristone was less efficacious in reducing alcohol consumption compared to Wistar rats (rats trained to consume alcohol)19. This effect of mifepristone in msP rats is consistent with our clinical findings where mifepristone significantly reduced craving only in patients with low FHDA. In fact, the msP rat line is a well-characterized model of excessive voluntary alcohol drinking, and these lines are based on repeated generations of selective breeding for alcohol preference47. Therefore, mifepristone is less likely to be effective in patients who have supposedly a greater family contribution to the development of AUD20,48. It is also possible that mifepristone should be administered at a higher dose10 to block craving after repeated challenges. The dose response of mifepristone was also reported in a clinical study in patients with psychotic depression, where psychotic symptoms were reduced by mifepristone (1,200mg/day) and the effect was dependent on the blood level of mifepristone attained9.

This study provides exploratory and preliminary, yet promising, results on the potential role of GR antagonists in the treatment of AUD in a subtype population. There are possible confounding effects in this study by the placebo, as lower alcohol use was reported during placebo treatment for subjects with low FHDA. It is possible that the low FHDA group exhibited a differentially potent placebo effect, due to motivating conditions of the clinical trial49.

One of the major strengths of this study is the within-subject, cross-over design which provided the same set of participants acting as their own controls, increasing power and reducing variability23. The cortisol level also provided robust results for medication compliance, as it increases with mifepristone administration22. This is also the first time that yohimbine was paired to a cue-reactivity paradigm in a AUD population12, highlighting the translational efforts of this RCT between animal and human models50. The premise and the logistics of this work was paved by our preclinical study11 using the same medications paired to cue-induced reinstatement, stress-induced reinstatement and alcohol self-administration paradigms. The robustness of the study paradigm was also determined by our participants’ retention, regardless of the long washout period (3-week), required to allow the cortisol to reset at the basal level after the mifepristone administration.

A major limitation of this study was not having control placebo-conditions for yohimbine. Therefore, we could not determine if the effect of stress induction was due to the interaction between yohimbine combined to a cue-reactivity. Independently, however, both yohimbine and cue-reactivity are well-established and validated procedures for stress-induced alcohol craving and consumption12. Also, we did not collect mifepristone blood level in this work to ensure that clinical effects may be dependent on the blood level of mifepristone attained9. Finally, the low number of females enrolled in this study did not permit us to evaluate sex as a biological variable. In summary, this study provides important information on the safety of mifepristone as a medication to treat AUD. Our findings support the safety of mifepristone-alcohol combination in a human laboratory setting. The lack of serious or severe AEs, together with the low number of dropouts, indicate that mifepristone may be an acceptable medication in the treatment of AUD. Similarly, the safety data of this trial support the use of translational integration of yohimbine, combined to a cue-reactivity and alcohol self-administration protocol. In terms of efficacy, consistent with the earlier clinical work10, we found an effect in reducing alcohol craving, an important behavioral marker and now a diagnostic criterion in the DSM-551. This translational trial fits with previous clinical studies that have utilized mifepristone in different psychiatry disorders, as mifepristone’s effects on reducing craving were independent of the increased plasma cortisol9. Future studies, with higher dose, are warranted to identify possible biomarkers of mifepristone’s response in patients with AUD and to best identify potential patients who will be responders or non-responders.

Data Availability

All data produced in the present study are available upon reasonable request to the authors

Authors Contributions

CLH-K: Lead for conceptualization, funding acquisition, investigation, methodology, project administration, analysis, writing-original draft, writing-review and editing. CLH-K, PC, JCB, EGA and RMS carried out the experiments and collected the experimental data. CLH-K and MM analyzed the data. CLH-K, EGA, NC, RC, RS, JCB, and LL provided substantial intellectual input for the interpretation of the results. All authors reviewed content and approved the final version of the manuscript.

Funding

This study was fully funded by the National Institute on Alcohol Abuse and Alcoholism (K01 AA023867 to CLH-K) and the Research Excellence Award from the Center of Alcohol and Addiction Studies, Brown University (CLH-K). Dr. Haass-Koffler is also supported by the National Institute on Alcohol Abuse and Alcoholism (K01 AA023867; R01 AA026589; R01 AA027760; R21 AA027614) and by the National Institute of General Medical Sciences (NIGMS), Center of Biomedical Research Excellence (COBRE, P20 GM130414). Drs. Magill and Cioe are supported by the National Institute on Alcohol Abuse and Alcoholism (R01 AA027760; R21 AA027614). Dr. Leggio is supported by the National Institute on Drug Abuse Intramural Research Program and the National Institute on Alcohol Abuse and Alcoholism Division of Intramural Clinical and Biological Research (ZIA DA000635 and ZIA AA000218).

Competing interests

The medication (mifepristone and matching placebo) was kindly provided by Corcept Therapeutics. Corcept Therapeutics did not have any role in the study design, execution or interpretation of the results, and this publication does not necessarily represent the official views of Corcept Therapeutics. CLH-K traveled to CA to attend and present the data to the Corcept Therapeutic Conference (September 2022). The views expressed herein are those of the authors and do not reflect the official policy or position of the funding agencies. The other authors declare no competing interests.

Acknowledgements

The authors thank Michael Brickley, Danielle Giovenco, Victoria Long, Zoe Brown (Brown University) and Karen Hafez (Corcept Therapeutics) for providing logistics support during the course of the trial. Also, we would like to thank Prabhjot Singh and Talia Vasaturo-Kolondier (Brown University) for providing help with the database.

References

  1. 1.↵
    Vinkers CH, Kuzminskaite E, Lamers F, Giltay EJ, Penninx B. An integrated approach to understand biological stress system dysregulation across depressive and anxiety disorders. J Affect Disord. Mar 15 2021;283:139–146. DOI:10.1016/j.jad.2021.01.051
    OpenUrlCrossRef
  2. 2.↵
    Haass-Koffler CL, Bartlett SE. Stress and addiction: contribution of the corticotropin releasing factor (CRF) system in neuroplasticity. Front Mol Neurosci. 2012;5:91. DOI:10.3389/fnmol.2012.00091
    OpenUrlCrossRefPubMed
  3. 3.↵
    Koob GF, Schulkin J. Addiction and stress: An allostatic view. Neurosci Biobehav Rev. Nov 2019;106:245–262. DOI:10.1016/j.neubiorev.2018.09.008
    OpenUrlCrossRef
  4. 4.↵
    Liu X, Weiss F. Stimulus conditioned to foot-shock stress reinstates alcohol-seeking behavior in an animal model of relapse. Psychopharmacology (Berl). Jul 2003;168(1-2):184–191. DOI:10.1007/s00213-002-1267-z
    OpenUrlCrossRefPubMed
  5. 5.
    Zironi I, Burattini C, Aicardi G, Janak PH. Context is a trigger for relapse to alcohol. Behav Brain Res. Feb 15 2006;167(1):150–5. DOI:10.1016/j.bbr.2005.09.007
    OpenUrlCrossRefPubMedWeb of Science
  6. 6.↵
    Janak PH, Chaudhri N. The Potent Effect of Environmental Context on Relapse to Alcohol-Seeking After Extinction. Open Addict J. Jan 1 2010;3:76–87. DOI:10.2174/1874941001003010076
    OpenUrlCrossRefPubMed
  7. 7.↵
    Laberg JC. What is presented, and what prevented, in cue exposure and response prevention with alcohol dependent subjects? Addict Behav. 1990;15(4):367–86. DOI:10.1016/0306-4603(90)90046-z
    OpenUrlCrossRefPubMedWeb of Science
  8. 8.↵
    Rohsenow DJ, Niaura RS, Childress AR, Abrams DB, Monti PM. Cue reactivity in addictive behaviors: theoretical and treatment implications. Int J Addict. 1990;25(7A-8A):957–93. DOI:10.3109/10826089109071030
    OpenUrlCrossRefPubMed
  9. 9.↵
    Block TS, Kushner H, Kalin N, Nelson C, Belanoff J, Schatzberg A. Combined analysis of mifepristone for psychotic depression: plasma levels associated with clinical response. Biological psychiatry. 2018;84(1):46–54.
    OpenUrl
  10. 10.↵
    Vendruscolo LF, Estey D, Goodell V, et al. Glucocorticoid receptor antagonism decreases alcohol seeking in alcohol-dependent individuals. J Clin Invest. Aug 3 2015;125(8):3193–7. DOI:10.1172/JCI79828
    OpenUrlCrossRefPubMed
  11. 11.↵
    Simms JA, Haass-Koffler CL, Bito-Onon J, Li R, Bartlett SE. Mifepristone in the central nucleus of the amygdala reduces yohimbine stress-induced reinstatement of ethanol-seeking. Neuropsychopharmacology. Mar 2012;37(4):906–18. DOI:10.1038/npp.2011.268
    OpenUrlCrossRefPubMedWeb of Science
  12. 12.↵
    Curley DE, Vasaturo-Kolodner TR, Cannella N, Ciccocioppo R, Haass-Koffler CL. Yohimbine as a pharmacological probe for alcohol research: a systematic review of rodent and human studies. Neuropsychopharmacology. Jun 27 2022;DOI:10.1038/s41386-022-01363-9
    OpenUrlCrossRef
  13. 13.↵
    Bertholomey ML, Verplaetse TL, Czachowski CL. Alterations in ethanol seeking and self-administration following yohimbine in selectively bred alcohol-preferring (P) and high alcohol drinking (HAD-2) rats. Behav Brain Res. Feb 1 2013;238:252–8. DOI:10.1016/j.bbr.2012.10.030
    OpenUrlCrossRefPubMed
  14. 14.↵
    Richards JK, Simms JA, Bartlett SE. Conditioned cues and yohimbine induce reinstatement of beer and near-beer seeking in Long-Evans rats. Addict Biol. Apr 2009;14(2):144–51. DOI:10.1111/j.1369-1600.2008.00139.x
    OpenUrlCrossRefPubMed
  15. 15.↵
    Stanford S. Central noradrenergic neurones and stress. Pharmacology & therapeutics. 1995;68(2):297–342.
    OpenUrl
  16. 16.↵
    Umhau JC, Schwandt ML, Usala J, et al. Pharmacologically induced alcohol craving in treatment seeking alcoholics correlates with alcoholism severity, but is insensitive to acamprosate. Neuropsychopharmacology. May 2011;36(6):1178–86. DOI:10.1038/npp.2010.253
    OpenUrlCrossRefPubMed
  17. 17.↵
    Haass-Koffler CL, Swift RM, Leggio L. Noradrenergic targets for the treatment of alcohol use disorder. Psychopharmacology (Berl). Jun 2018;235(6):1625–1634. DOI:10.1007/s00213-018-4843-6
    OpenUrlCrossRef
  18. 18.↵
    Jimenez VA, Walter NAR, Shnitko TA, et al. Mifepristone Decreases Chronic Voluntary Ethanol Consumption in Rhesus Macaques. J Pharmacol Exp Ther. Nov 2020;375(2):258–267. DOI:10.1124/jpet.120.000169
    OpenUrlAbstract/FREE Full Text
  19. 19.↵
    Benvenuti F, Cannella N, Stopponi S, et al. Effect of Glucocorticoid Receptor Antagonism on Alcohol Self-Administration in Genetically-Selected Marchigian Sardinian Alcohol-Preferring and Non-Preferring Wistar Rats. Int J Mol Sci. Apr 17 2021;22(8)DOI:10.3390/ijms22084184
    OpenUrlCrossRef
  20. 20.↵
    Kenna GA, Haass-Koffler CL, Zywiak WH, et al. Role of the alpha1 blocker doxazosin in alcoholism: a proof-of-concept randomized controlled trial. Addict Biol. Jul 2016;21(4):904–14. DOI:10.1111/adb.12275
    OpenUrlCrossRef
  21. 21.↵
    Dwan K, Li T, Altman DG, Elbourne D. CONSORT 2010 statement: extension to randomised crossover trials. BMJ. Jul 31 2019;366:n4378. DOI:10.1136/bmj.l4378
    OpenUrlCrossRef
  22. 22.↵
    Gallagher P, Watson S, Elizabeth Dye C, Young AH, Nicol Ferrier I. Persistent effects of mifepristone (RU-486) on cortisol levels in bipolar disorder and schizophrenia. J Psychiatr Res. Oct 2008;42(12):1037–41. DOI:10.1016/j.jpsychires.2007.12.005
    OpenUrlCrossRefPubMedWeb of Science
  23. 23.↵
    Haass-Koffler CL, Perciballi R, Magill M, et al. An inpatient human laboratory study assessing the safety and tolerability, pharmacokinetics, and biobehavioral effect of GET 73 when co-administered with alcohol in individuals with alcohol use disorder. Psychopharmacology (Berl). Nov 3 2021;DOI:10.1007/s00213-021-06008-1
    OpenUrlCrossRef
  24. 24.↵
    Greenwald MK, Lundahl LH, Steinmiller CL. Yohimbine increases opioid-seeking behavior in heroin-dependent, buprenorphine-maintained individuals. Psychopharmacology (Berl). Feb 2013;225(4):811–24. DOI:10.1007/s00213-012-2868-9
    OpenUrlCrossRef
  25. 25.↵
    Wangelin BC, Powers MB, Smits JA, Tuerk PW. Enhancing exposure therapy for PTSD with yohimbine HCL: protocol for a double-blind, randomized controlled study implementing subjective and objective measures of treatment outcome. Contemp Clin Trials. Nov 2013;36(2):319–26. DOI:10.1016/j.cct.2013.08.003
    OpenUrlCrossRef
  26. 26.↵
    Schulz K, Altman D, Moher D. Consolidated Standards of Reporting Trials Group CONSORT 2010 Statement: Updated Guidelines for Reporting Parallel Group Randomised Trials. Bmj. 2010;340:c332.
    OpenUrlFREE Full Text
  27. 27.↵
    Zeger SL, Liang K-Y. Longitudinal data analysis for discrete and continuous outcomes. Biometrics. 1986:121–130.
  28. 28.↵
    Singleton E. Alcohol craving questionnaire, short-form (revised; ACQ-SF-R): background, scoring, and administration. Baltimore, MD, USA. 1995;
  29. 29.↵
    Bohn MJ, Krahn DD, Staehler BA. Development and initial validation of a measure of drinking urges in abstinent alcoholics. Alcohol Clin Exp Res. Jun 1995;19(3):600–6. DOI:10.1111/j.1530-0277.1995.tb01554.x
    OpenUrlCrossRefPubMedWeb of Science
  30. 30.↵
    Montoya AK, Hayes AF. Two-condition within-participant statistical mediation analysis: A path-analytic framework. Psychological Methods. 2017;22(1):6.
    OpenUrl
  31. 31.↵
    Rhemtulla M, Brosseau-Liard PE, Savalei V. When can categorical variables be treated as continuous? A comparison of robust continuous and categorical SEM estimation methods under suboptimal conditions. Psychol Methods. Sep 2012;17(3):354–73. DOI:10.1037/a0029315
    OpenUrlCrossRefPubMedWeb of Science
  32. 32.↵
    Capone C, Kahler CW, Swift RM, O’Malley SS. Does family history of alcoholism moderate naltrexone’s effects on alcohol use? J Stud Alcohol Drugs. Jan 2011;72(1):135–40. DOI:10.15288/jsad.2011.72.135
    OpenUrlCrossRefPubMed
  33. 33.↵
    Mann RE, Sobell LC, Sobell MB, Pavan D. Reliability of a family tree questionnaire for assessing family history of alcohol problems. Drug Alcohol Depend. May 1985;15(1-2):61–7. DOI:10.1016/0376-8716(85)90030-4
    OpenUrlCrossRefPubMedWeb of Science
  34. 34.↵
    Haass-Koffler CL, Giovenco DE, Zywiak WH, et al. ROLE OF COMBINED TOPIRAMATE-ARIPIPRAZOLE ADMINISTRATION ON FREE-CHOICE ALCOHOL USE AND EXPLORATORY PHARMACOGENETICS. Alcoholism-Clinical and Experimental Research. JUN 2016 2016;40:95A.
    OpenUrl
  35. 35.↵
    Miles J. Tolerance and variance inflation factor. Wiley StatsRef: Statistics Reference Online. 2014;
  36. 36.↵
    Haass-Koffler CL, Leggio L, Kenna GA. Pharmacological approaches to reducing craving in patients with alcohol use disorders. CNS Drugs. Apr 2014;28(4):343–60. DOI:10.1007/s40263-014-0149-3
    OpenUrlCrossRef
  37. 37.↵
    Fox HC, Hong KI, Siedlarz K, Sinha R. Enhanced sensitivity to stress and drug/alcohol craving in abstinent cocaine-dependent individuals compared to social drinkers. Neuropsychopharmacology. Mar 2008;33(4):796–805. DOI:10.1038/sj.npp.1301470
    OpenUrlCrossRefPubMedWeb of Science
  38. 38.↵
    Haass-Koffler CL, Goodyear K, Long VM, et al. A Phase I randomized clinical trial testing the safety, tolerability and preliminary pharmacokinetics of the mGluR5 negative allosteric modulator GET 73 following single and repeated doses in healthy volunteers. European Journal of Pharmaceutical Sciences. 2017;109:78–85.
    OpenUrl
  39. 39.↵
    Haass-Koffler CL, Goodyear K, Zywiak WH, Leggio L, Kenna GA, Swift RM. Comparing and Combining Topiramate and Aripiprazole on Alcohol-Related Outcomes in a Human Laboratory Study. Alcohol Alcohol. May 1 2018;53(3):268–276. DOI:10.1093/alcalc/agx108
    OpenUrlCrossRef
  40. 40.↵
    Guidance D. Alcoholism: Developing Drugs for Treatment. Center for Drug Evaluation and Research (CDER). 2015;
  41. 41.↵
    Haass-Koffler CL, Long VM, Farokhnia M, et al. Intravenous administration of ghrelin increases serum cortisol and aldosterone concentrations in heavy-drinking alcohol-dependent individuals: Results from a double-blind, placebo-controlled human laboratory study. Neuropharmacology. Nov 1 2019;158:107711. DOI:10.1016/j.neuropharm.2019.107711
    OpenUrlCrossRef
  42. 42.
    Leggio L, Zywiak WH, Fricchione SR, et al. Intravenous ghrelin administration increases alcohol craving in alcohol-dependent heavy drinkers: a preliminary investigation. Biol Psychiatry. Nov 1 2014;76(9):734–41. DOI:10.1016/j.biopsych.2014.03.019
    OpenUrlCrossRefPubMed
  43. 43.↵
    Leggio L, Zywiak WH, McGeary JE, et al. A human laboratory pilot study with baclofen in alcoholic individuals. Pharmacol Biochem Behav. Feb 2013;103(4):784–91. DOI:10.1016/j.pbb.2012.11.013
    OpenUrlCrossRefPubMed
  44. 44.↵
    Rohsenow DJ, Monti PM, Abrams DB, et al. Cue elicited urge to drink and salivation in alcoholics: Relationship to individual differences. Advances in behaviour research and therapy. 1992;14(3):195–210.
    OpenUrlCrossRef
  45. 45.↵
    Walker SW, Lightly ER, Clyne C, Williams BC, Bird IM. Adrenergic and cholinergic regulation of cortisol secretion from the zona fasciculata/reticularis of bovine adrenal cortex. Endocr Res. 1991;17(1-2):237–65. DOI:10.1080/07435809109027200
    OpenUrlCrossRefPubMedWeb of Science
  46. 46.↵
    Vining RF, McGinley RA, Maksvytis JJ, Ho KY. Salivary cortisol: a better measure of adrenal cortical function than serum cortisol. Ann Clin Biochem. Nov 1983;20 (Pt 6):329–35. DOI:10.1177/000456328302000601
    OpenUrlCrossRefPubMed
  47. 47.↵
    Borruto AM, Stopponi S, Li H, Weiss F, Roberto M, Ciccocioppo R. Genetically selected alcohol-preferring msP rats to study alcohol use disorder: Anything lost in translation? Neuropharmacology. Mar 15 2021;186:108446. DOI:10.1016/j.neuropharm.2020.108446
    OpenUrlCrossRefPubMed
  48. 48.↵
    Garbutt JC, Greenblatt AM, West SL, et al. Clinical and biological moderators of response to naltrexone in alcohol dependence: a systematic review of the evidence. Addiction. Aug 2014;109(8):1274–84. DOI:10.1111/add.12557
    OpenUrlCrossRefPubMed
  49. 49.↵
    Goodyear K, Vasaturo-Kolodner TR, Kenna GA, Swift RM, Leggio L, Haass-Koffler CL. Alcohol-related changes in behaviors and characteristics from the baseline to the randomization session for treatment and non-treatment seeking participants with alcohol use disorder. Am J Drug Alcohol Abuse. Nov 2 2021;47(6):760–768. DOI:10.1080/00952990.2021.1961799
    OpenUrlCrossRef
  50. 50.↵
    Nieto SJ, Grodin EN, Aguirre CG, Izquierdo A, Ray LA. Translational opportunities in animal and human models to study alcohol use disorder. Transl Psychiatry. Sep 29 2021;11(1):496. DOI:10.1038/s41398-021-01615-0
    OpenUrlCrossRef
  51. 51.↵
    Haass-Koffler CL, Kenna GA. Bacchus by Caravaggio as the Visual Diagnosis of Alcohol Use Disorder from the Fifth Edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5). Front Psychiatry. 2013;4:86. DOI:10.3389/fpsyt.2013.00086
    OpenUrlCrossRef
Back to top
Posted January 04, 2023.
Download PDF

Supplementary Material

Data/Code
Email
Mifepristone as a pharmacological intervention for stress-Induced alcohol craving: a translational crossover randomized trial
Carolina L. Haass-Koffler, Molly Magill, Nazzareno Cannella, Joshua C. Brown, Elie G. Aoun, Patricia A. Cioe, Rajita Sinha, Robert M. Swift, Roberto Ciccocioppo, Lorenzo Leggio
medRxiv 2023.01.02.23284122; doi: https://doi.org/10.1101/2023.01.02.23284122
Twitter logo Facebook logo LinkedIn logo Mendeley logo
Citation Tools

Subject Area