Unmasking Convergent Oxycodone Seeking and Consumption Driving Augmented Intake during Extended Access to Oral Operant Self-Administration

Abstract

Individual vulnerability to opioid intake escalation is a critical but poorly understood aspect of addiction. Using genetically diverse inbred rat strains, we investigated operant oral oxycodone self-administration, identifying 'Augmenter' phenotypes that dramatically increased consumption during extended (16h vs. 4h) access-a vulnerability not predicted by standard motivation tests. A key innovation was applying lick microstructure analysis (LMA) to operationally distinguish 'consumption' from 'seeking' lick clusters within the inter-reward interval. During extended access, Augmenters of both sexes exhibited a striking surge in the frequency of both consumption and seeking clusters (p<0.0001), driving their escalated intake. Notably, female Augmenters also showed larger seeking cluster sizes (p=0.006), suggesting enhanced reward value specifically linked to seeking behavior. In contrast, interlick interval (a palatability measure) did not differentiate phenotypes. This LMA-based approach reveals that an increased drive to seek out additional oxycodone, rather than altered hedonic impact alone, underlies the augmentation of opioid intake, offering a nuanced rodent model of heightened vulnerability and a powerful tool to dissect reward dynamics.

Figure 1.. Augmenter and Non-Augmenter rat strains that greatly increased their oxycodone intake during 16h self-administration (SA) sessions were identified from a large panel of inbred strains.

Panel A shows the progressive increase in mean oxycodone intake across stages (i.e., increasing concentration and duration of access to drug) SA by all female vs male rat strains in the full panel. Across the panel, oxycodone intake (mg/kg) increased as the dose and session length increased (p < 2.0e-16). Females showed higher overall intake than males (p < 2.0e-16). Panels B and C compare oxycodone intake by the two intake phenotypes, Augmenter strains (n = 9) and Non-Augmenter strains (n = 9), in 4h vs 16h SA sessions. In both sexes, intake was higher in Augmenters: females [panel B; main effects of intake phenotype (p = 0.011) and Schedule (4h vs 16h; p < 2.0e-06)]; males [panel C; main effects of intake phenotype (p = 0.026) and Schedule (p < 3.4e-06)]. In both sexes, there was no difference in 4h oxycodone intake between Augmenters and Non-Augmenters (females, p = 0.873 ; males, p = 0.795).

Figure 2.. Total Licks on the active and inactive spouts by Augmenters vs Non-Augmenters in each sex by stage of oxycodone SA (panels A and B) and the dynamic ranges of lick cluster size (CS) and interlick interval (ILI) by stage of SA (panels C and D, respectively).

In both sexes, a main effect of stage and an interaction of stage x intake phenotype was identified (panel A female, p<0.001 and p = 0.001, respectively; panel B, male, p<0.001 for both). In both sexes at 16h, the total number of licks was greater in Augmenters than Non-Augmenters (female and male, p<0.001). Licks on the inactive spout (panels A and B) did not differ between the two intake phenotypes in either sex. The dynamic range from 4h 0.025 mg/ml to 16h 0.1 mg/ml was significantly greater for CS (panel C) than ILI (panel D), p < 0.001.

Figure 3.. Classifying consummatory vs. seeking behavior based on the post-reward microstructure of active spout lick clusters.

A. Latency is the time (sec) from a reward-associated cluster generated on the active spout to all subsequent non-reward clusters occurring before the very next reward cluster. Active spout non-reward clusters (orange) were mostly emitted during the 20 sec time-out interval, whereas a vastly smaller number occurred after the time-out. In order to display data for the entire 16h session (inset), the latencies of each cluster were log transformed. B. Analysis of lick cluster size (licks/cluster; CS) in clusters emitted consecutively after a reward-associated cluster (position 0) and the latency (sec) of each of these clusters emitted within the 20-sec timeout period revealed a distinct pattern. The reward-triggering cluster (i.e. 0) and the first two subsequent clusters (dark blue) were significantly larger than later clusters (light blue, positions 3–11), with the most substantial decrease in CS occurring after cluster number 2. This temporal decay in CS (shown by circle size) informed the operational classification, designating the initial, larger clusters (0–2) as “Consumption” clusters and the later, smaller clusters (3+) as “Seeking.” In each panel, the total number of clusters (n = 5,013) emitted on the active spout by Augmenter and Non-Augmenter strains in both sexes are included. Data points are mean ± SEM.

Figure 4.. During 16h sessions of oxycodone SA, the frequency of lick clusters during both consumption and seeking on the active spout differentiates Augmenters from Non-Augmenters.

Inclusive analysis of the four panels showed the number of lick clusters during Consumption and Seeking varied across stages (p < 2.2e-16) and differed between sexes (female > male; p = 0.031) and intake phenotypes (Augmenter: orange circle; Non-Augmenter: green circle; p = 6.4e-05). While cluster counts were comparable between phenotypes during standard 4h access at 0.1mg/ml, Augmenters in both sexes exhibited a pronounced escalation in the frequency of both consumption and seeking clusters relative to Non-Augmenters, specifically during the extended 16h access condition (consumption and seeking, p < 0.0001). Data are mean ± SEM.

Figure 5.. Lick Cluster Size (CS) shows a limited association with intake phenotype.

Overall, consumption clusters (defined in Fig. 3) were consistently much larger (i.e.,licks/cluster) than seeking clusters (p < 2.2e-16). Combined analysis of all 4 panels found main effects of stage (p = 2.8e-05), with CS generally increasing at the 16h stage; sex (p = 0.003), with male CS greater than female; and intake phenotype (p = 0.0003), with Augmenters having slightly larger CS than Non-Augmenters. CS generally increased during extended (16h) access compared to prior sessions and differed slightly between sexes and phenotypes. At 16h, an increase in seeking CS was observed in female Augmenters vs Non-Augmenters (p = 0.0062). Data are mean ± SEM.

Figure 6.. interlick interval remains stable between Augmenters vs Non-Augmenters despite divergent intake levels.

Based on a four panel analysis of the average interlick interval (ILI, sec) within consumption and seeking clusters, a potential indicator of reward valuation, we found main effects of sex (p = 5.3e-06), with longer ILIs (i.e. slower licking) in females (0.190 ± 0.002 s) than males (0.175 ± 0.001 s), and a small effect of operant action (p = 0.033), with seeking cluster ILIs (0.184 ± 0.002 s) slightly longer than consumption clusters (0.180 ± 0.002 s). However, there were no main effects of stage (p = 0.13) or intake phenotype (p = 0.08). Although males generally displayed faster licking (shorter ILI) than females, no significant differences in ILI were found between Augmenter and Non-Augmenter phenotypes for either cluster type. Data are mean ± SEM.