Ranking the contribution of behavioral measures comprising oxycodone self-administration to reinstatement of drug-seeking in male and female rats

Abstract

Introduction: Rates of relapse to drug use during abstinence are among the highest for opioid use disorder (OUD). In preclinical studies, reinstatement to drug-seeking has been extensively studied as a model of relapse-but the work has been primarily in males. We asked whether biological sex contributes to behaviors comprising self-administration of the prescription opioid oxycodone in rats, and we calculated the relative contribution of these behavioral measures to reinstatement in male and female rats.

Materials and methods: Rats were trained to self-administer oxycodone (8 days, training phase), after which we examined oxycodone self-administration behaviors for an additional 14 days under three conditions in male and female rats: short access (ShA, 1 h/d), long access (LgA, 6 h/d), and saline self-administration. All rats were then tested for cue-induced reinstatement of drug-seeking after a 14-d forced abstinence period. We quantified the # of infusions, front-loading of drug intake, non-reinforced lever pressing, inter-infusion intervals, escalation of intake, and reinstatement responding on the active lever.

Results: Both male and female rats in LgA and ShA conditions escalated oxycodone intake to a similar extent. However, males had higher levels of non-reinforced responding than females under LgA conditions, and females had greater levels of reinstatement responding than males. We then correlated each addiction-related measure listed above with reinstatement responding in males and females and ranked their respective relative contributions. Although the majority of behavioral measures associated with oxycodone self-administration did not show sex differences on their own, when analyzed together using partial least squares regression, their relative contributions to reinstatement were sex-dependent. Front-loading behavior was calculated to have the highest relative contribution to reinstatement in both sexes, with long and short inter-infusion intervals having the second greatest contribution in females and males, respectively.

Discussion: Our results demonstrate sex differences in some oxycodone self-administration measures. More importantly, we demonstrate that a sex- dependent constellation of self-administration behaviors can predict the magnitude of reinstatement, which holds great promise for relapse prevention in people.

Keywords: escalation; female; front-loading; opioid; relapse; sex difference; vulnerability.

FIGURE 1

Experimental Schematic. The IVSA paradigm is depicted with the orange boxes, forced abstinence with the red arrow, and the reinstatement test with the green box. A blow up of one example day is shown below the overall schematic, with an example IVSA session represented as long access (LgA), with 6 bars representing 6 h [First blue bar indicates a short access (ShA) session)]. Below that, 1 h of one session is represented by 4 smaller bars representing 15-min bins. To the right of the blow up of a representative session is an example of infusion patterns, with blue ticks representing infusions. Text boxes represent the behavioral measures analyzed for sex differences and used in regression analyses to determine contribution to reinstatement responding and ability to predict magnitude of reinstatement.

FIGURE 2

Female and male rats escalate oxycodone self-administration under both long access (LgA, 6 h) and short access (ShA, 1 h) conditions. Oxycodone self-administration behavior is shown as # of infusions/day (A–C) and # of active lever presses/day (D–F) for male and female rats across the 22 days of the self-administration paradigm. Groups included rats that self-administered oxycodone under LgA conditions (A,D) or under ShA conditions (B,E) during the escalation phase (days 9–22) and rats that self-administered saline (C,F). Under both LgA and ShA, but not saline, conditions, there was a main effect of Day. There were no sex differences in total infusions (G) or active lever presses (H) during the escalation phase (d9–22). There were no sex differences in the change in number of infusions (I) or active lever presses (J) from escalation day 9 to escalation day 22, although the net change for LgA and ShA, but not saline, rats was >0, indicating escalation of intake (see Supplementary Figure 2). ^#p < 0.05, ^###p < 0.001, main effect of Day. N: LgA (22 females, 26 males); ShA (15 females, 13 males); Sal (10 females, 9 males).

FIGURE 3

Reinstatement responding after 14 days of forced abstinence: effects of sex and prior self-administration condition. Data are shown as the average number (±SEM) of active lever presses during the first hour of the Reinstatement Test in male and female rats that had self-administered oxycodone in the escalation phase under LgA or ShA conditions or self-administered saline (Sal). Across treatment groups, females had higher levels of reinstatement responding than males (main effect of Sex). Further, treatment group dictated the level of reinstatement responding, with LgA > ShA > Sal (main effect of Treatment). ^#p < 0.05, ^##p < 0.01, ^###p < 0.001, main effects of sex and treatment, as indicated by brackets. N: LgA (19 females, 20 males); ShA (15 females, 13 males); Sal (8 females, 6 males).

FIGURE 4

Female and male rats front-load oxycodone intake. Data are shown as the average number (±SEM) of infusions (A–C) and active lever presses (D–F) in 15-min bins from the first hour of escalation day 9 (ESC d9) and day 22 (ESC d22) in male and female rats. For each treatment condition: LgA (A,D), ShA (B,E), and Sal (C,F), 3-way ANOVAs (Sex × ESC day × Time) were calculated. For oxycodone self-administration under LgA and ShA conditions, rats self-administered more oxycodone in the first 15-min bin compared to the last 15-min bin (i.e., front-loading), and overall oxycodone intake was greater on ESC d22 compared to ESC d9. This was determined through main effects or interactions of ESC day and Time. ^#p < 0.05, ^##p < 0.01, main effect of Time; **p < 0.01, Dunnett’s multiple comparisons to the 0–15 min bin; ^δp < 0.05, ^δδp < 0.01 Bonferroni’s multiple comparisons of ESC d9 to ESC d22 per time bin. N: LgA (19 females, 20 males); ShA (15 females, 13 males); Sal (8 females, 6 males).

FIGURE 5

Female and male rats self-administer oxycodone throughout the 6-h LgA sessions during the escalation phase. The average number (±SEM) of oxycodone infusions (A) and active lever presses (B) for each of the 6 h of escalation days 9 (ESC d9) and 22 (ESC d22) are plotted for males and females. Infusions and active lever presses are higher for both sexes on ESC d22 compared to ESC d9 (main effect of Day), although there is an overall, yet modest, decrease in responding across the 6-h (main effect of Hour). No sex differences were observed. ^#p < 0.05, ^##p < 0.01, ^###p < 0.001, main effects of day and hour, as indicated. N: LgA (19 females, 20 males); ShA (15 females, 13 males); Sal (8 females, 6 males).

FIGURE 6

Males exhibit higher levels of non-reinforced active lever pressing than females under LgA self-administration conditions. Non-reinforced active lever pressing is expressed as: [(non-reinforced active lever presses—infusions)/infusions]. Data are shown as non-reinforced active lever pressing per self-administration day (d1–22) for LgA (A), ShA (B), and Sal (C) escalation phase treatment groups and as the total of non-reinforced active lever pressing during the escalation phase (d9–22) for LgA (D), ShA (E), and Sal (F) escalation phase treatment groups. Under LgA conditions, non-reinforced active lever pressing depended on main effects of Sex and Day (A), with males showing significantly higher responses than females (D); unpaired t-test). ^#p < 0.05, ^##p < 0.01, main effects of sex and day, as indicated; *p < 0.05 unpaired t-test. N: LgA (22 females, 26 males); ShA (15 females, 13 males); Sal (10 females, 9 males).

FIGURE 7

Latency between bursts of oxycodone infusions decreases in females and males under LgA, but not ShA, self-administration conditions. Schematics illustrating methods to calculate interval latencies between oxycodone infusions are generated from representative data and shown in (A–C). In panel (A), raster plots show oxycodone infusions (blue ticks) from the escalation days 9 and 22 from a LgA rat. All inter-infusion intervals (s) are quantified example intervals demarcated with red lines (A). Frequency distributions of log-transformed inter-infusion intervals across the 6-h self-administration days reveal bimodal distributions that consist of a population of long (long mode), and a population of short (short mode), intervals (B,C). Mo₁ represents the inter-infusion interval in the short mode with the highest frequency of occurrence, and Mo₂ represents the inter-infusion interval in the long mode with the highest frequency of occurrence. Data in panels (D–G) are dot and line plots representing average inter-infusion intervals for the long mode (D,E) and short mode (F,G) on escalation days 9 (ESC d9) and 22 (ESC d22) for each rat under LgA (D,F) and ShA (E,G) conditions. 2-way ANOVAs (Sex × ESC day) show a main effect of ESC day and a significant decrease in Long Mode inter-infusion intervals in both males and females under LgA conditions on ESC d22 compared to ESC d9 (D). *p < 0.05, **p < 0.01, Bonferroni’s multiple comparisons of ESC d9 to ESC d22 per sex. LgA (22 females, 26 males); ShA (15 females, 13 males).

FIGURE 8

Relative contribution of behavioral measures associated with oxycodone self-administration to post-abstinence reinstatement responding. The 6 behavioral measures quantified in the previous figures [Infusions, front-loading, non-reinforced active lever pressing, Inter-infusion intervals (short and long modes), Escalation] were segregated by sex, z-scored, and correlated with reinstatement responding (Pearson’s). For both females (A) and males (B), histograms across the diagonal represent z-scored distributions of each measure. The inset (C) provides enlarged visualization of each component of the matrix. In the histograms, the X-axis represents the z-scored measures, while the Y-axis represents the number of rats. The correlation plot in panel (C) shows the relationship between z-scored active lever presses in the reinstatement test (RT: active lever) and z-scored oxycodone infusions (infusion). The correlation matrix (below the diagonal of histograms) represents the cross-correlation between all the measures, while the matrix on top shows the correlation coefficients (R). Significant correlations are indicated with asterisks. Gray highlighted squares show comparisons of reinstatement test responding with each behavioral measure: correlation plots are in the gray-shaded column and correlation coefficients (R) are in the gray-shaded row. (D) Schematic showing the partial least square regression model used to determine the role of each measure to reinstatement responding. The dot plots in panel (D) (orange circles, females and blue circles, males) were generated with the partial least square regression model and demonstrate a tight fit between observed reinstatement responses vs predicted ones. Bar plots were generated indicating the relative contribution of each behavioral measure to reinstatement test responding in female and male rats (E). N: 30 females, 29 males.