Preclinical evidence to support repurposing everolimus for craving reduction during protracted drug withdrawal

Abstract

Cue-elicited drug-craving is a cardinal feature of addiction that intensifies (incubates) during protracted withdrawal. In a rat model, these addiction-related behavioral pathologies are mediated, respectively, by time-dependent increases in PI3K/Akt1 signaling and reduced Group 1 metabotropic glutamate receptor (mGlu) expression, within the ventromedial prefrontal cortex (vmPFC). Herein, we examined the capacity of single oral dosing with everolimus, an FDA-approved inhibitor of the PI3K/Akt effector mTOR, to reduce incubated cocaine-craving and reverse incubation-associated changes in vmPFC kinase activity and mGlu expression. Rats were trained to lever-press for intravenous infusions of cocaine or delivery of sucrose pellets and then subjected to tests for cue-reinforced responding during early (3 days) or late (30–46 days) withdrawal. Rats were gavage-infused with everolimus (0–1.0 mg/kg), either prior to testing to examine for effects upon reinforcer-seeking behavior, or immediately following testing to probe effects upon the consolidation of extinction learning. Single oral dosing with everolimus dose-dependently blocked cocaine-seeking during late withdrawal and the effect lasted at least 24 h. No everolimus effects were observed for cue-elicited sucrose-seeking or cocaine-seeking in early withdrawal. In addition, everolimus treatment, following initial cue-testing, reduced subsequent cue hyper-responsivity exhibited observed during late withdrawal, arguing a facilitation of extinction memory consolidation. everolimus’ “anti-incubation” effect was associated with a reversal of withdrawal-induced changes in indices of PI3K/Akt1/mTOR activity, as well as Homer protein and mGlu1/5 expression, within the prelimbic (PL) subregion of the prefrontal cortex. Our results indicate mTOR inhibition as a viable strategy for interrupting heightened cocaine-craving and facilitating addiction recovery during protracted withdrawal.

Profiling prefrontal cortex protein expression in rats exhibiting an incubation of cocaine craving following short-access self-administration procedures

“…Also consistent with Chiu et al (47), no cocaine-or time-related differences in the expression of the monomer form of mGlu5 were detected within the PL (Figure 2B; Group X Withdrawal ANOVA: all p's > 0.120), although mGlu5 dimer expression was significantly lower overall in the rats tested on WD30 vs. WD3 (Figure 2C) [Withdrawal effect: F (1,62) = 7.922, p = 0.007; Group effect and interaction, p's > 0.258]. However, in contrast to both Miller et al (29) and Chiu et al (47), no significant correlations were noted between cue-elicited responding and the PL levels of either mGlu receptor monomer (for mGlu1 monomer, r = −0.005, p = 0.968; for mGlu5 monomer, r = −0.098, p = 0.454) or the mGlu5 dimer (r = −0.203, p = 0.111). Taken together with the results of Chiu et al (47), the incubation of cocaine craving under short-access IV-SA procedures is insufficient to elicit changes in the PL expression of Group 1 mGlu receptors.…”

“…In contrast to our prior study employing the 6-h model ( 27), but consistent with our recent report using the mixedaccess model (47), we detected no significant group differences in the PL expression of the mGlu1 monomer (Group effect and interaction, p's>0.060), although mGlu1 monomer expression tended to be lower overall on WD30 vs. WD3 (Figure 2A; Withdrawal effect: p = 0.055). Also consistent with Chiu et al (47), no cocaine-or time-related differences in the expression of the monomer form of mGlu5 were detected within the PL (Figure 2B; Group X Withdrawal ANOVA: all p's > 0.120), although mGlu5 dimer expression was significantly lower overall in the rats tested on WD30 vs. WD3 (Figure 2C) [Withdrawal effect: F (1,62) = 7.922, p = 0.007; Group effect and interaction, p's > 0.258]. However, in contrast to both Miller et al (29) and Chiu et al (47), no significant correlations were noted between cue-elicited responding and the PL levels of either mGlu receptor monomer (for mGlu1 monomer, r = −0.005, p = 0.968; for mGlu5 monomer, r = −0.098, p = 0.454) or the mGlu5 dimer (r = −0.203, p = 0.111).…”

“…In the IL, we detected a time-dependent reduction in the levels of the monomer forms of both Group 1 mGlu receptors, as well as the dimer form of mGlu5, irrespective of the cocaine history of the rats (Figures 5A-C) [for mGlu1, Withdrawal effect: F (1,57) = 7.369, p = 0.009; Group effect and interaction, p's > 0.633; for mGlu5 monomer, Withdrawal effect: F (1,61) = 7.203, p = 0.01; Group effect and interaction, p's > 0.078; for mGlu5 dimer, Withdrawal effect: F (1,60) = 6.995, p = 0.011; Group effect and interaction, p's > 0.165]. However, consistent with Chiu et al (47), no correlations were detected between cue-elicited responding on test day and the IL expression of either mGlu receptor monomer [for mGlu1, r = 0.058, p = 0.664; for mGlu5, r = −0.147, p = 0.253] or the mGlu5 dimer (r = −0.161, p = 0.215). GluN1 expression was also lower overall at the 30-day withdrawal time-point [Withdrawal effect: F (1,61) = 4.596, p = 0.036], but was higher overall in both cocaine IV-SA groups vs. Sham controls (Figure 5D In contrast, we detected no significant cocaine-or withdrawal-related changes in the expression of any of the other glutamate-related proteins examined herein (see Table 1) [Group X Withdrawal ANOVAs: for Homer1b/c, all p's > 0.120; for Homer2a/b, all p's > 0.150; for GluN2a, all p's > 0.450; for GluN2b, all p's > 0.510, for GluA1, all p's > 0.400; for GluA2, all p's > 0.500) nor did we detect any significant correlations between our glutamate-related proteins and cue-elicited responding on test day (for Homer1b/c, r = −0.223 p = 0.077; for Homer2a/b, r = −0.122, p = 0.339; for GluA1, r = −0.150, p = 0.257; for GluA2, p = 0.009, p = 0.947; for GluN2a, r = −0.004, p = 0.973; for GluN2b, r = −0.002, p = 0.987].…”

“…Total AKT levels were unchanged in the PL (Figure 4A; Group X Withdrawal ANOVA, all p's > 0.14) and there was no correlation between behavior during testing and total Akt expression in this PFC subregion [r = −0.131, p = 0.301]. However, akin to our prior results for both the 6-h (16) and mixed-access model (47), we observed group differences in a time-dependent increase in p(Ser473)-Akt1 expression within the PL as indexed either by total phospho-protein expression (Figure 4B) [Group X Withdrawal: F (2,64) = 8.12, p = 0.001] or the relative expression of the phospho-protein (Figure 4C) [Group X Withdrawal: F (2,63) = 5.354, p = 0.007]. Indeed, deconstruction of these interactions along the Group factor Examination of changes in CaMKII expression within the PL detected no group or time-related differences for total CaMKII expression within the PL (Figure 4D; Group X Withdrawal ANOVA, all p's > 0.616).…”

“…The other IV-SA procedure (Mixed) tested in this study was identical to that described in Chiu et al (47). Following IV catheterization, rats underwent a 6-h self-administration session on Day 1 of training and for the next 9 days, rats underwent the same 2-h self-administration training procedures as those described for the 2-h model (see Section 2-h IV-SA procedure) and were randomly assigned to either the WD3 or WD30 test groups in such a way as to ensure comparable cocaine intake over the last 3 days of self-administration.…”