Introduction: Women are more vulnerable than men in many aspects of opioid use disorder (OUD); a major theory of sex differences in substance use disorders is that these differences are due to ovarian hormones with estradiol enhancing vulnerability in females. However, most of this evidence is for psychostimulants and alcohol; evidence with opioids is sparse. Thus, the goal of this study was to determine the impact of estradiol on vulnerability in females in a rat model of OUD. Method: Following self-administration training, ovariectomized (OVX) females with (E) or without (V) estradiol replacement were given extended (24 h/day), intermittent access (2, 5-min trials/h) to fentanyl for 10 days. Then, the development of three key features of OUD were assessed, including physical dependence, defined by the magnitude and time course of weight loss during withdrawal; an enhanced motivation for fentanyl, assessed using a progressive-ratio schedule; and relapse vulnerability, assessed using an extinction/cue-induced reinstatement procedure. These later two characteristics were examined following 14 days of withdrawal when the phenotypes are known to be highly expressed. Results: OVX+E females self-administered markedly higher levels of fentanyl under extended, intermittent-access conditions and showed a longer time course of physical dependence, a greater increase in motivation for fentanyl, and an enhanced sensitivity to the reinstating effects of fentanyl-associated cues compared to OVX+V rats. Severe health complications were also observed in OVX+E, but not OVX+V females, during withdrawal. Conclusion: These results indicate that, as with findings with psychostimulants and alcohol, estradiol enhances vulnerability in females to developing opioid addiction-like features and serious opioid-related health complications.

© 2023 S. Karger AG, Basel

Introduction

Opioid use disorder (OUD) is a major epidemic in the USA with opioid-involved overdose deaths reaching the highest number ever recorded in the 12-month period ending in March 2022, which was primarily driven by fentanyl, a synthetic opioid [1]. Although men have historically had higher rates of OUD than women, these differences have been narrowing, and in 2020, a greater number of women (18 years and older) misused opioids than men [2]. Additionally, women appear to be more vulnerable than men on many aspects of the disease process including meeting the criteria and/or seeking treatment for OUD more rapidly after initial use (i.e., the telescoping effect; [3–9]), exhibiting more negative affective symptoms during attempts to stop using opioids [9–11], experiencing higher levels of cue- and stress-induced opioid craving [12, 13], and suffering more serious drug-related medical and psychological consequences than men [14–18].

Despite the severity of this problem in women, the vast majority of preclinical studies on OUD have only included males. However, recent studies have started to consider sex as a biological variable in OUD and, as with clinical findings, have reported notable sex differences in the development and expression of an addiction-like phenotype. For example, females self-administer higher levels of opioids, including heroin, fentanyl, and oxycodone, under extended-access conditions compared to males [19–23]. There are also sex differences in the pattern of opioid use under extended-access conditions with females showing greater escalation of heroin intake under the long-access procedure (i.e., fixed-ratio 1 access to the drug for 6 or more hours/day; [19, 22]) and having greater fentanyl intake within active trials (or binge intake) under the intermittent-access (IntA) procedure (fixed-ratio 1 access during 2, 5-min trials/hr, 24 h/day; [23]) than males. Sex differences in levels and patterns of opioid intake are most apparent under conditions that maximize individual differences, such as under low doses and using procedures that do not limit the number of infusions/hour or day. Additionally, following extended-access self-administration, females show a prolonged course of physical dependence during withdrawal [23, 24] and a greater vulnerability to cue-induced relapse, especially when tested in the estrus phase of their estrous cycle [23, 24].

A major theory of sex differences in substance use disorders is that differences are due to ovarian hormones with estradiol enhancing vulnerability in females. This hypothesis is supported by numerous studies showing that in females, ovariectomy (OVX), which depletes ovarian hormones, decreases extended-access drug self-administration for multiple drugs including cocaine, nicotine, and alcohol and that estradiol replacement restores levels of drug self-administration [25–32]. Additionally, estradiol is critical for the development of key features of an addiction-like phenotype such as an enhanced motivation for the drug and a preference for the drug over other reward alternatives. For example, our previous studies with cocaine show that both OVX and pharmacological blockade of estradiol via treatment with the selective estrogen receptor modular tamoxifen in ovary-intact females effectively block the development of an enhanced motivation for cocaine following extended-access self-administration and 14 days of withdrawal, which are optimal conditions for inducing this addiction-like phenotype [26, 27, 33]. Additionally, as with effects on drug intake, this phenotype can be rescued in OVX females with estradiol replacement [26, 27]. OVX has also been reported to prevent the development of a preference for cocaine over food which can be enhanced by estradiol replacement [34]. Thus, estradiol enhances both drug use and vulnerability to developing an addiction-like phenotype.

One major caveat, however, is that most of the evidence implicating estradiol in substance use disorders is for psychostimulants and alcohol. Evidence with opioids is sparse and restricted to effects on drug use under short-access self-administration conditions (<2 h/day access), which likely reflects vulnerability to drug use/reinforcement but not necessarily vulnerability to developing OUD. To our knowledge, no studies have examined the impact of estradiol in a rat model of OUD that has been optimized/validated for inducing an addiction-like phenotype similar to that observed in humans with an OUD (e.g., compulsive use, an enhanced motivation to use the drug, vulnerability to relapse). Thus, the goals of this study were to determine the impact of estradiol on fentanyl self-administration under extended-access conditions and the subsequent development of three key addiction-like features including physical dependence, as defined by the magnitude and time course of weight loss during withdrawal, an enhanced motivation for fentanyl, as assessed using a progressive-ratio schedule, and vulnerability to relapse, as assessed using an extinction/cue-induced reinstatement procedure. These later two characteristics were examined following 14 days of withdrawal, when levels of motivation for the drug and drug-seeking are high. We also used an extended (24 h/day), IntA fentanyl self-administration procedure (2, 5-min trials/h, 10 days) that mimics patterns of drug use observed in humans (i.e., binge-abstinent patterns of use and repeated spiking drug levels; [35]) and has been buprenorphine-validated to induce an addiction-like phenotype in both females and males [i.e., vulnerability to relapse is attenuated following buprenorphine treatment during a 14-day withdrawal period; [36]). Based on previous reports of estradiol increasing extended-access drug self-administration and the subsequent development of an addiction-like phenotype with psychostimulants and alcohol [25–34], we predicted that estradiol would increase fentanyl intake and subsequent vulnerability to developing addiction-like features with fentanyl.

MethodsSubjects

Sexually mature OVX female Sprague-Dawley rats (N = 53; Charles River) were purchased from Charles River and arrived at the vivarium within 1 week of surgery. Rats were individually housed in operant test chambers (Med Associates, St. Albans, VT, USA) for the duration of the experiment and maintained on a 12-h light/dark cycle with lights on at 7:00 a.m. and ad libitum access to water and food (Teklad LM-485 7912). To accelerate the acquisition of fentanyl self-administration, rats were pretrained to lever-press for sucrose pellets (45 mg) under a fixed-ratio 1 schedule using methods previously described [37]. The health of the rats was monitored daily throughout the study, and the rats were weighed at least three times a week. Body weight was used as an indicator of overall health throughout the study and as a measure of physical dependence to fentanyl during opioid withdrawal as previously described [23]. All procedures were conducted within animal care guidelines set by the National Institute of Health and were approved by the University of Virginia Animal Care and Use Committee.

ProceduresDrugs

Fentanyl hydrochloride was obtained from the National Institute on Drug Abuse (Research Triangle Park, NC) and dissolved in sterile saline at a concentration of 6.25 µg/mL. Fentanyl solutions were sterile filtered (0.22 μm; Millipore, Billerica, MA, USA) and stored at 4°C. Rats were weighed Monday, Wednesday, and Friday, and the duration of infusion was adjusted for changes in body weight to ensure that the mg/kg dose was consistent throughout the study. 17B-estradiol was purchased from Sigma-Aldrich (St Louis, MO) and dissolved in corn oil (administered as 5 µg/0.1 mL).

Estradiol Replacement and Vaginal Cytology

Upon arrival, rats were randomly assigned to the estradiol or vehicle group and treatment was started the next day. Rats received either a subcutaneous injection of estradiol (5 µg/day; OVX+E; N = 28) or an equal volume of corn oil (0.1 mL; OVX+V; N = 25) at 11:00 a.m. 5 days a week (Monday–Friday). We have previously shown that this dose of estradiol enhances extended-access cocaine self-administration and restores the development of an enhanced motivation for cocaine in OVX females [26, 27]. In order to confirm successful OVX and estradiol replacement, at the start of the study, daily vaginal swabs were performed at noon for a minimum of 5 days as described previously [26, 27].

Surgery and Catheter Maintenance

Rats underwent catheterization surgery after lever pretraining using methods previously described [37]. The catheters were flushed with heparinized saline 3 days a week, which helped to maintain and verify patency throughout the study. Methohexital (1.5 mg/kg) was used to confirm patency when necessary. Any right jugular catheter that was no longer patent was replaced with a new catheter implanted into the left, external jugular vein, and behavioral testing re-started following recovery.

Fentanyl Self-Administration Training

Following recovery from surgery, rats were trained to self-administer fentanyl (0.25 µg/kg/infusion) as previously described [36]. Sessions were conducted daily until the acquisition requirement was met, which was defined as 5 consecutive days wherein all 40 infusions were obtained. Moderate food restriction (85% of free-feeding body weight) was used when necessary to encourage acquisition. Two OVX+V females failed to meet the acquisition requirements within 28 days, and one OVX+V female developed an issue with patency, resulting in their removal from the study and a final group size of 28 OVX+E and 22 OVX+V females.

Motivation for Fentanyl

Following acquisition, in a subset of OVX rats (Fig. 1a; OVX+E, N = 17 and OVX+V, N = 14), a baseline level of motivation for fentanyl was established prior to IntA fentanyl self-administration (pre-IntA PR) using a progressive-ratio schedule as previously described [38]. Briefly, the response requirement to obtain a fentanyl infusion increased throughout the session in the following steps: 1, 2, 4, 6, 9, 12, 15, 20, 25, 32, 40, 50, 62, 77, 95, 118, 145, 178, 219, 268, 328, 402, etc. These sessions continued daily until a stable baseline was achieved, which was defined as no increasing or decreasing trend in the number of infusions obtained over three consecutive sessions (typically 3 to 4 sessions).

Fig. 1.

Summary of experimental events. Ovariectomize female rats with and without estradiol replacement were trained to self-administer fentanyl (FR train; 0.25 μg/kg/infusion). Following acquisition, a baseline level of motivation for fentanyl was established using a progressive-ratio schedule (pre-IntA PR). Then, rats were given extended, intermittent access to fentanyl (IntA SA; 0.25 μg/kg/infusion; 24 h/day, 2, 5-min trials/h) for 10 days. Following the last fentanyl self-administration session, physical dependence was evaluated over a 14-day withdrawal period and on withdrawal day 15 motivation for fentanyl was reassessed using the same progressive-ratio schedule (post-IntA PR test; a). A different subset of the rats underwent the same procedure except vulnerability to relapse was only assessed on withdrawal day 15 using an extinction/reinstatement procedure (b).

Extended, IntA Fentanyl Self-Administration

Once the acquisition criteria were met or a stable, baseline level of motivation for fentanyl was established, rats were given extended, 24 h/day access to fentanyl (0.25 μg/kg/infusion) for ten consecutive days under an IntA procedure that mimics patterns of drug use observed in humans by inducing a binge-abstinent pattern of drug use with repeated spiking drug levels as previously described [36]. One OVX+E female developed an issue with patency, resulting in their removal from the study and a final group size of 27 OVX+E and 22 OVX+V females.

Assessing the Development of Addiction-Like Features

Physical Dependence to Fentanyl. Following IntA fentanyl self-administration, physical dependence to fentanyl was evaluated over the 14-day withdrawal period by measuring the magnitude and time course of weight loss during early (W1, 24 h after IntA self-administration), intermediate (W7, days 6–8), and protracted (W14, days 14 or 15) withdrawal relative to the end of IntA fentanyl self-administration (IntA10, at the start or end of the last day of IntA self-administration) as previously described [23]. Four OVX+E females developed severe health complications in response to spontaneous fentanyl withdrawal, resulting in their removal from the study and a final group size of 23 OVX+E and 22 OVX+V females.

An Enhanced Motivation for Fentanyl. Following IntA fentanyl self-administration and 14 days of withdrawal, motivation for fentanyl was reassessed in a subset of rats screened prior to IntA self-administration (Fig. 1a; 13 OVX+E and 12 OVX+V) using the same progressive-ratio schedule and conditions as those used for the pre-IntA PR test, except that, due to health and patency complications, only one progressive-ratio session was conducted.

Vulnerability to Relapse. Following IntA fentanyl self-administration and 14 days of withdrawal, fentanyl-seeking was assessed in a different subset of rats (Fig. 1b; 10 OVX+E and 10 OVX+V females) using a within-session extinction/cue-induced reinstatement procedure as previously described [36]. If responding did not extinguish within 9 sessions (two OVX+E rats), the session terminated and extinction responding was reassessed the next day. For the two OVX+E females that did not extinguish within 9 extinction sessions, data from the first day of extinction testing were used in the analyses of hourly extinction responses, whereas the second day was used for the last extinction session and reinstatement. Technical issues during relapse testing prevented the inclusion of the reinstatement data from one OVX+E rat, and it should also be noted that one of the OVX+E and two of the OVX+V females underwent progressive-ratio testing prior to IntA fentanyl self-administration; these animals underwent relapse testing following withdrawal rather than progressive-ratio testing because patency issues during withdrawal prevented post-IntA PR testing. These rats did not differ from the other rats within their groups in levels of fentanyl intake during the extended-access period or in responding during the extinction/reinstatement test.

Analysis

We first determined whether there was an effect of estradiol on the acquisition of fentanyl self-administration using the non-parametric Mann-Whitney U t test to compare the average number of days to acquire self-administration between OVX+E and OVX+V females that met the acquisition criteria. We also compared percent group acquisition between OVX+E and OVX+V females within the first 5 training sessions and by the end of the 28-day training period using a χ2 test. Then, repeated measures ANOVA was used to analyze group differences in levels and patterns of fentanyl intake, including the number of infusions per session and per active trial or binge as defined by trials with one or more infusions, and the number of active trials or binges per day, using session as the repeated measure and treatment group as the between-subject factor. Post-IntA tests, including the post-IntA PR test or extinction/reinstatement test, were included as an additional factor in the acquisition and IntA analyses; however, since no significant overall effects of test were observed, the data were collapsed across test and presented as OVX+E versus OVX+V.

Next, repeated measures ANOVAs were used to determine the impact of estradiol on the development of an opioid addiction-like phenotype, with separate analyses used for each feature (physical dependence to fentanyl, an enhanced motivation for fentanyl, and vulnerability to relapse). Specifically, group differences in the magnitude and time course of physical dependence were determined by comparing percent change in body weight during early, intermediate, and protracted withdrawal relative to the end of IntA fentanyl self-administration. Group differences in the development of an enhanced motivation for fentanyl were also determined using repeated measures ANOVA by comparing the number of infusions obtained during the pre-versus post-IntA PR testing phase. We also used a univariate ANOVA to examine the percent change in the number of infusions obtained during the post-IntA PR test relative to group averages for the OVX+V and OVX+E rats during the pre-IntA PR test. Both analyses focused on the first day of PR testing within each phase. Additionally, group differences in vulnerability to relapse were determined using repeated measures ANOVA by comparing the number of responses during each of the 6 extinction sessions as well as during the last extinction session versus the reinstatement session. We also examined the effects of estradiol on total extinction responding across all extinction sessions run (1–9) using a univariate ANOVA, and effects on the likelihood of reinstatement in response to cues were determined by comparing the percentage of rats within each group that had higher responses during the reinstatement session as compared to the last extinction session using a χ2 test.

Lastly, the effects of estradiol on body weights were also determined at each of the major phases of the study including arrival, training (days 3–5), IntA fentanyl self-administration (early, days 1–3, IntA1; intermediate, days 4–6, IntA5; late, days 14–15, IntA10), and withdrawal (early, 24 h after IntA self-administration, W1; intermediate, days 6–8, W7; protracted, days 14–15, W14) using repeated measures ANOVA. In order to determine whether the incidence of severe health complications, defined as labored breathing or death, in OVX+E was significantly higher than that observed in OVX+V, we analyzed percent group survival as a function of phase of study using a Kaplan-Meier survival analysis and the Log-rank (Mantel-Cox) statistic. Pearson correlations were also conducted to determine the association between the frequency of drug use (number of infusions or active trials) and the expression of addiction-like features (development of enhanced motivation, relapse vulnerability, and physical dependence). The analysis was performed collapsed across groups since the univariant ANOVA determined there was not a significant difference in the correlation coefficients for OVX+E and OVX+V females. A one-tailed test was used for a priori predicted differences (i.e., OVX+E females would have a higher level of motivation for fentanyl than OVX+V females following protracted withdrawal from IntA fentanyl self-administering and there would be a significant correlation between frequency of drug use and the expression of addiction-like features); all other tests were two-tailed. All post hoc comparisons were corrected for multiple comparisons using Tukey’s method. Statistical analyses were performed using SPSS (V26) with alpha set at 0.05.

ResultsEffect of Estradiol on Acquisition

Both OVX+E and OVX+V females acquired fentanyl self-administration quickly under these training conditions, with the number of days to acquire not differing between the groups for the rats that met the acquisition criteria (Fig. 2a). However, a greater percentage of OVX+E females met the acquisition criteria within the first five training sessions than OVX+V females (U = 4.0, p < 0.05; Fig. 2b). While a greater percentage of OVX+E versus OVX+V females also acquired self-administration by the end of the acquisition testing period (within a maximum of 28 sessions), this difference was not statistically significant. Thus, although both OVX+E and OVX+V rats acquired fentanyl self-administration relatively quickly under these conditions, OVX+E females were more likely to acquire fentanyl self-administration rapidly than OVX+V females.

Fig. 2.

Effect of estradiol on fentanyl acquisition. Mean (±SEM) number of days to meet the acquisition criteria for OVX+E (n = 28) and OVX+V (n = 22) females that acquired fentanyl self-administration (a) and percent of OVX females that reached the fentanyl acquisition criteria within the first five training sessions and the end of the acquisition training period for OVX+E (n = 28) and OVX+V (n = 24) females (b). *p < 0.05, versus OVX+V (b).

Effect of Estradiol on Extended, IntA Fentanyl Self-Administration

OVX+E females self-administered markedly more fentanyl than OVX+V females during IntA self-administration (1.6 times more fentanyl intake than OVX+V females; overall effect of group, F(1, 47) = 9.8, p < 0.01; Fig. 3a). This difference is most pronounced during later IntA sessions (interaction of session and group, F(9, 423) = 2.3, p < 0.05) and appears to be driven by a greater increase, or escalation, of intake from the initial sessions to later ones in OVX+E females (65% increase from session 1 vs. 10, p < 0.05), but not in OVX+V females, which maintained a similar level of fentanyl intake across the IntA period. Further analysis confirmed that fentanyl intake was similar between OVX+E and OVX+V females in session one, but significantly greater in OVX+E compared to OVX+V females by session 10 (p < 0.01). Thus, OVX+E females escalated their fentanyl intake over the IntA period, resulting in significantly greater fentanyl intake compared to OVX+V females.

Fig. 3.

Effect of estradiol on extended, intermittent-access (IntA) fentanyl self-administration. Mean (±SEM) number of infusions (a), active trials (b), and number of infusions per active trials (c) for each of the 10 IntA sessions for OVX+E (n = 27) and OVX+V (n = 22) females. *p < 0.01, versus OVX+V (a, b). ^p < 0.05, versus session 10 (a–c).

To further explore differences in patterns of fentanyl self-administration between OVX+E and OVX+V females, we analyzed the number of active trials (or binges) per session across the IntA period. OVX+E females had more binges per session than OVX+V females during the IntA period (Fig. 3b; overall effect of group, F(1, 47) = 18.438, p < 0.001). Similar to the effects observed with infusions, this difference is more pronounced during later IntA sessions (interaction of session and group, F(9, 423) = 2.8, p < 0.05) and appears to be driven by a greater increase in binges per session from initial sessions to later ones in OVX+E versus OVX+V females. While both groups showed an increase in binges/session from days 1 to 10 of the IntA period (p’s < 0.001), analysis within days 1 and 10 confirmed that on day 1 the number of binges per session was similar between OVX+E and OVX+V females, but on session 10, it was significantly greater in OVX+E compared to OVX+V females (p < 0.001). Fentanyl intake within each binge/active trial also tended to be higher in OVX+E versus OVX+V rats (Fig. 3c; F(1, 47) = 3.0, p = 0.09). In contrast to the effects observed for the number of binges/session (or active trials), intake within each binge, or active trial, decreased from initial to later sessions (overall effect of session, F(1, 47) = 2.9, p < 0.01; session 2 vs. 10, p < 0.05). Thus, compared to OVX+V females, OVX+E females had a higher frequency of binges, particularly during later sessions, and tended to take more fentanyl per active binge.

Effect of Estradiol on the Development of Features of an Addiction-Like Phenotype

Physical Dependence to Fentanyl. OVX+E females lost a greater percentage of their body weight during withdrawal compared to OVX+V females (Fig. 4a; overall effect of group, F(2, 43) = 16.8, p < 0.001); this effect appears to be due to OVX+E females having a prolonged course of weight loss (overall effect of withdrawal time point, F(2, 86) = 30.3, p < 0.001; interaction of withdrawal time point and group, F(2, 86) = 9.2, p < 0.001). Post hoc analysis during early withdrawal confirmed that both OVX+E and OVX+V females lost a significant percentage of their body weight relative to the end of the IntA period (or 0; p < 0.05). During intermediate withdrawal, the body weight of OVX+E females remained significantly below their pre-withdrawal baseline (relative to the end of the IntA period, or 0; p < 0.05); whereas, OVX+V females had not only regained their lost weight, they also gained additional body weight relative to the end of the IntA period (or 0; p < 0.05), resulting in a significant difference in the percent change in body weight between the OVX+E and OVX+V females during intermediate withdrawal (p < 0.05). By protracted withdrawal, the body weight of OVX+E females returned to baseline relative to the end of the IntA period (or 0; p < 0.05) and the body weight of OVX+V females remained increased from baseline relative to the end of the IntA period (or 0; p < 0.05), resulting in a significant difference in the percent change of body weight between the OVX+E and OVX+V females. Thus, OVX+E and OVX+V females lost a similar percentage of body weight during early withdrawal, indicating that physical dependence was expressed similarly following fentanyl self-administration; however, this weight loss persisted longer in OVX+E females compared to OVX+V females, indicating that estradiol may prolong the course of physical dependence with fentanyl.

Fig. 4.

Effect of estradiol on the development of addiction-like features. Physical Dependence: Mean (±SEM) percent change in body weight (g) during early (W1), intermediate (W7), and protracted (W14) withdrawal relative to end of IntA fentanyl self-administration (IntA10) for OVX+E (n = 23) and OVX+V (n = 22) females (a). An Enhanced Motivation: Mean (±SEM) number of infusions obtained under a progressive-ratio schedule prior to (pre-IntA) and following IntA fentanyl self-administration and protracted withdrawal (post-IntA; b) and percent change in number of infusions obtained under a progressive-ratio schedule post-IntA relative to pre-IntA (c) for OVX+E (n = 13) and OVX+V (n = 12) females. Vulnerability to Relapse: Mean (±SEM) number of responses made on the lever formerly associated with fentanyl during the first six 1-h extinction sessions (d) and the number of responses made on the lever formerly associated with fentanyl during the last extinction session versus the reinstatement session (e) for OVX+E (n = 10) and OVX+V (n = 10) females. *p < 0.05, versus OVX+V (a–c, e). +p < 0.05, versus baseline (or 0; A), pre-IntA (b), hours 2–6 (d), and the last extinction session (e). The dashed lines in c represents the threshold for the development of an enhanced motivation for fentanyl (15% increase) and dashed line and 100% and 60% values in (e) represent the percentage of animals within each group to show an increase in responses from the last extinction session to the reinstatement session.

An Enhanced Motivation for Fentanyl. As predicted, OVX+E and OVX+V females obtained a similar number of infusions under the PR schedule prior to IntA fentanyl self-administration, but after IntA self-administration and withdrawal, OVX+E females obtained more infusions than OVX+V females (Fig. 4b). Results from the repeated measures ANOVA revealed significant effects of phase (F(1, 23) = 67.3, p < 0.001) and phase by group (F(1, 23) = 4.1, p < 0.05). Post hoc comparisons within each phase confirmed a group difference in the number of infusions obtained during post-IntA PR testing (p < 0.05) but not pre-IntA PR testing. However, within-group comparisons of infusions obtained during pre-IntA versus post-IntA PR testing revealed significant increases in both OVX+E and OVX-V females (p’s < 0.01), indicating that both groups developed an enhanced motivation for fentanyl following extended-access self-administration and protracted withdrawal. The analysis of the percent change in the number of infusions obtained during post-IntA PR testing (relative to the average number of infusions obtained by the OVX+E and OVX+V females during pre-IntA PR testing) also confirmed an overall effect of group with OVX+E females having a larger increase in motivation for fentanyl following IntA fentanyl self-administration and protracted withdrawal than the OVX+V females (F(1, 23) = 5.7, p < 0.05; Fig. 4c). Thus, both OVX+E and OVX+V females developed an enhanced motivation for fentanyl following IntA fentanyl self-administration and protracted withdrawal, but its expression was more robust in the OVX+E females than the OVX+V females.

Vulnerability to Relapse. OVX+E and OVX+V females had similar levels of responding during the first six extinction sessions (Fig. 4d) with results from the repeated measures ANOVA comparing responding during the first six extinction sessions revealing an overall effect of session (F(5, 90) = 18.0, p < 0.001) but no overall or interactive effect of group. Post hoc analysis revealed that responding was highest in both groups during the first extinction session compared to later sessions (session 1 vs. sessions 2–6, p’s < 0.05). Analysis of total extinction responding confirmed no significant overall and interactive effect of group (data not shown).

Both OVX+E and OVX+V females reinstated fentanyl-seeking upon the presentation of fentanyl-associated cues with results from the repeated measures ANOVA comparing responding during the last extinction session to the reinstatement session revealing a significant overall effect of session (Fig. 4e; F(1,17) = 26.2, p < 0.001). There was no overall or interactive effect of group; however, analysis of percent group reinstatement using χ2 revealed that a greater percentage of OVX+E females (100%) were reinstated compared to OVX+V females (60%; x2 = 48.4, p < 0.001). Thus, both the OVX+E and OVX+V females developed an enhanced vulnerability to relapse following IntA fentanyl self-administration and protracted withdrawal, but the OVX+E females showed an enhanced sensitivity to the reinstating effects of fentanyl-associated cues compared to OVX+V females.

Effect of Estradiol on Physical Health (Body Weight) and Probability of Survival

There were also significant differences in general health effects with the analysis on body weight revealing that OVX+E females weighed less than OVX+V females (Fig. 5a; overall effect of group, F(1, 43) = 72.5, p < 0.001). This difference appears to be driven by less weight gain in the OVX+E females during later phases of the study, including during IntA fentanyl self-administration and withdrawal (overall effect of phase, F(7, 301) = 100.7, p < 0.001; interaction of phase and group, F(7, 301) = 47.8, p < 0.001). Post hoc analysis within the different phases of the study confirmed there was no difference in body weight upon arrival, but that OVX+E females weighed significantly less than OVX+V females during each subsequent phase (p’s < 0.05). Notably, a greater proportion of OVX+E females developed severe adverse health effects, including labored breathing (3 of 27) and death (4 of 27), during withdrawal than OVX+V females, where neither issue was observed (Fig. 5b; x2 = 6.2, p < 0.05). Thus, estradiol appears to blunt weight gain and exacerbate opioid withdrawal-related adverse health consequences. These life-threating health complications also appear to be most pronounced during later withdrawal time points when weight differences are the largest.

Fig. 5.

Effect of estradiol on physical health and probability of survival. Mean (±SEM) body weight (g) at arrival, training (train), early (IntA1), intermediate (IntA5) and late (IntA10) phases of intermittent-access (IntA) fentanyl self-administration, and early (W1), intermediate (W7), and protracted phases of withdrawal (W14; a) for OVX+E (n = 23) and OVX+V (n = 22) females and percent survival prior to withdrawal at the end of IntA self-administration (IntA10), early withdrawal (days 1–7), and protracted withdrawal (days 8–15) for OVX+E (n = 27) and OVX+V (n = 22) females (b). *p < 0.05 versus OVX+V (a, b).

Association between Frequency of Fentanyl Intake and the Expression of Features of an Addiction-Like Phenotype

As with our previous study [23], the frequency of fentanyl use, as defined by the average number of fentanyl infusions obtained during the IntA period, was predictive of the subsequent development of physical dependence (Fig. 6a; r = −0.41, p < 0.001), as defined by the percent body weight loss during early withdrawal (or W1) relative to the end of the IntA period (or IntA10), development of an enhanced motivation for fentanyl (Fig. 6b; r = 0.40, p < 0.05), as defined by an increase in the number of infusions obtained under the PR schedule post-IntA and withdrawal relative to pre-IntA, and development of enhanced relapse vulnerability (Fig. 6c; r = 0.45, p < 0.05), as defined by reinstatement responses; these relationships were similar between OVX+E and OVX+V females (non-significant interaction of group). Notably, the frequency of binge intake, as defined by the average number of active trials during the IntA period, was also predictive of the development of physical dependence to fentanyl (Fig. 6d; r = −0.34, p < 0.05), the development of enhanced motivation for fentanyl (Fig. 6e; r = 0.50, p < 0.05), and relapse vulnerability (Fig. 6f; r = 0.40, p < 0.05); these relationships were also similar between OVX+E and OVX+V females (non-significant interaction of sex). Thus, the frequency of opioid use appears to be a relatively strong predictor for the severity of OUD in females.

Fig. 6.

Association between frequency of fentanyl use and the development of addiction-like features. Number of infusions self-administered during the intermittent-access (IntA) period was predictive of the development of physical dependence, defined as the percent change in body weight during early withdrawal relative to the end of IntA (a), an enhanced motivation, defined as percent change in number of infusions obtained under a progressive-ratio schedule following IntA fentanyl self-administration and protractive withdrawal (post-IntA) relative to prior to IntA fentanyl self-administration (pre-IntA; b), and vulnerability to relapse, defined as the number of response on the formerly active lever upon the presentation of fentanyl-associated cues (reinstatement responses; c), and for OVX+E (n = 13, 10, and 23, respectively) and OVX+V (n = 12, 10, and 22, respectively) females. Notably, the number of active trials (or binges) during the IntA period was also predictive of the development of physical dependence (d), enhanced motivation (e), and vulnerability to relapse (f), and for OVX+E (n = 13, 10, and 23, respectively) and OVX+V (n = 12, 10, and 21, respectively) females. *p < 0.05, significant association (a–f).

Discussion

The purpose of this study was to determine the impact of estradiol on fentanyl intake and the development of an opioid addiction-like phenotype in females. Similar to findings with psychostimulants and alcohol (for review, see [39]), we found that estradiol replacement in OVX females increased fentanyl use and the expression of several key features of an addiction-like phenotype. More specifically, compared to OVX+V, OVX+E females were more likely to acquire fentanyl self-administration rapidly and to escalate their fentanyl use over time under extended, IntA conditions, indicating that estradiol may accelerate transitions from initial to regular opioid use and from regular opioid use to escalated/dysregulated opioid use. Following IntA fentanyl self-administration, physical dependence (weight loss) was similarly expressed in OVX+E and OVX+V females; however, the weight loss persisted significantly longer in OVX+E females, indicating that estradiol prolongs the course of physical dependence in response to fentanyl. Additionally, following protracted withdrawal, OVX+E females expressed a greater increase in motivation for fentanyl and were more sensitive to the reinstating effects of fentanyl-associated cues than OVX+V females. A significant percentage of OVX+E females also developed severe health complications over withdrawal (27%), and 4 of the 7 sick rats succumbed to illness which is striking considering that signs of severe illness were not observed in OVX+V females. Together, these results indicate that estradiol enhances vulnerability in females to developing opioid addiction-like features and serious opioid-related health complications.

The most striking difference between OVX females with and without estradiol replacement was the markedly higher levels of fentanyl intake in OVX+E females during the extended, IntA period. This is significant because these extended-access conditions are thought to model the “binge-abstinence” pattern of opioid use observed in humans with OUD [35]. We also saw evidence of loss of control of fentanyl intake in OVX+E but not OVX+V females, which is another key feature of OUD in humans that has been defined in animal models as an increase, or escalation, of drug intake over time. More specifically, OVX+E females increased their fentanyl intake by an average of 65% from the first to the last session, whereas OVX+V females maintained relatively constant levels of intake across the IntA period. The escalation effect observed in OVX+E females appears to be attributable to an increase in the frequency of fentanyl use (or the number of infusions and active trials) which is significant since we know from our previous work [23] and from findings in this study that frequency of use is predictive of an enhanced vulnerability for the development of addiction-like features. Notably, these effects of estradiol on drug intake are consistent with findings with other addictive drugs, including cocaine [25–28], nicotine [29], and alcohol [29–32], showing that while OVX robustly decreases drug intake under extended-access condition, estradiol replacement increases drug intake to levels similar to intact females. That similar findings were observed here with opioids provides support for the hypothesis that estradiol’s effect on vulnerability in females is similar for opioids as compared to psychostimulants and alcohol. These findings also indicate that estradiol may accelerate the transition from regular use to escalated/dysregulated opioid use; therefore, this effect of estradiol may also explain the clinical phenomenon termed the telescoping effect where women progress faster from initial opioid use to meeting the criteria for OUD and/or seeking treatment for OUD than men [3–9].

Since the main goal of this study was to determine the effects of estradiol on the development of addiction-like features following extended-access drug self-administration, the training conditions used were designed to induce rapid and maximal rates of acquisition rather than maximize individual differences in rates of acquisition. None-the-less, group differences were apparent with OVX+E females showing a greater likelihood of acquisition during the first five training sessions as compared to OVX+V females (86 vs. 67%, respectively). Our low fentanyl dose may have contributed to this finding, considering that sex and other individual differences during the acquisition of drug self-administration are more likely to be observed under low dose conditions. Indeed, a previous study also found that estradiol replacement in OVX females facilitated the acquisition of heroin self-administration at a relatively low dose (7.5 µg/kg/infusion; [40]). This is also consistent with a large body of work with psychostimulants and alcohol indicating that estradiol enhances initial vulnerability to drug use as defined by faster rates of acquisition and greater percent group acquisition (for review, see [39]). However, it is possible that the effects of estradiol differ for rates of acquisition versus initial levels of use, considering that several studies have reported no differences between OVX+E and OVX+V females in initial levels of heroin self-administration under short-access conditions (2–3 h/day, fixed-ratio 1 schedule; [41, 42]), even when assessed across a broad range of doses (0.3–50 µg/kg/infusion). Notably, this provides additional evidence for estradiol accelerating transitions between the stages of addiction in females with OVX+E females transitioning faster from initial to regular (or stable) opioid use than OVX+V females, similar to the transition from regular to escalated/uncontrolled opioid use.

As predicted, estradiol also increased vulnerability to developing addiction-like features with fentanyl; however, in contrast to effects with cocaine [26], where OVX+E, but not OVX+V, females developed an addiction-like phenotype, defined as an enhanced motivation for the drug, both OVX+V and OVX+E females developed features of an addiction-like phenotype with fentanyl. More specifically, despite OVX+E and OVX+V females expressing a similar degree of physical dependence (percentage of body weight lost) following fentanyl self-administration, OVX+V females recovered their body weight quickly over the first week of withdrawal. In contrast, weight loss persisted well into the second week of withdrawal for OVX+E females, indicating that estradiol may prolong the course of physical dependence with fentanyl, similar to ours and others previous findings in intact females compared to males [23, 24]. Additionally, both OVX+E and OVX+V females developed an enhanced motivation for fentanyl following protracted withdrawal from IntA fentanyl self-administration, but this phenotype was more robustly expressed (1.7 times higher) and more likely to occur in OVX+E versus OVX+V females given that 100% of OVX+E females developed this phenotype (≥15% increase relative to baseline; [26, 27]) versus 83% of OVX+V females. Similarly, both OVX+E and OVX+V females expressed high levels of fentanyl-seeking during extinction and reinstatement testing following IntA fentanyl self-administration and protracted withdrawal; however, responding was reinstated by fentanyl-associated cues in a higher percentage of OVX+E females (100%) compared to OVX+V females (60%). These findings indicate that OVX+E females also have enhanced sensitivity to the reinstating effects of fentanyl-associated cues compared to OVX+V females. Taken together, these findings suggest that while estradiol enhances the expression of opioid addiction-like features in females, it is not necessary for their development. The development of a loss of control over drug use may be one exception, considering that only females with estradiol showed an escalation of fentanyl intake over the 10-day extended, IntA self-administration phase.

One of the most important findings of this study from a translational perspective is that risk of life-threating health complications was markedly higher in OVX+E versus OVX+V females. It is notable that this risk was observed exclusively during withdrawal, and not during extended, IntA fentanyl self-administration. The most obvious explanation for this effect of estradiol is that the lower body weights of OVX+E females may have made them more susceptible to adverse health effects associated with fentanyl withdrawal. This effect of estradiol on body weight is well known [43], and it was apparent in this study prior to fentanyl self-administration and withdrawal. As such, body weight loss during early withdrawal, even though at a similar percentage, would be expected to have a bigger impact in OVX+E females. Indeed, unlike OVX+V females, which quickly recovered their body weight over the first week of withdrawal, OVX+E females showed a prolonged course of weight loss that lingered into the second week of withdrawal, as discussed above. Another possibility, however, is that the greater intake of fentanyl in OVX+E females during the extended-access period may have contributed to both the adverse health effects and the prolonged course of physical dependence in OVX+E versus OVX+V females. It is also possible that estradiol enhances the risk of adverse health effects during fentanyl withdrawal by exacerbating physical dependence/opioid withdrawal severity, considering findings in humans with OUD or alcohol use disorder showing that women experience greater adverse withdrawal effects and require more support to manage withdrawal effects in response to extended-release naltrexone, a long-acting opioid antagonist, presumably due to effects of ovarian hormones in women [44, 45]. Additionally, high levels of estradiol may enhance immunosuppression induced by opioid withdrawal (see [46] for review) and result in both a prolonged course of physical dependence (as observed in OVX+E females) and an increased vulnerability to infections and/or sepsis [47]. Finally, it is possible that the enhanced risk of adverse health effects in OVX+E females was mediated independent of withdrawal, considering that both intravenous drug use and estrogen replacement therapy are risk factors for health complications such as deep vein thrombus which can lead to life-threatening sequela including pulmonary embolisms [48–51]. While future research is necessary to determine the mechanism for the enhanced vulnerability in the OVX+E females to the adverse health effects that developed over fentanyl withdrawal, it is notable that women with substance use disorders have also been reported to experience greater adverse health effects and/or an accelerated course of drug-related medical consequences than men with substance use disorders, including opioids [52–55].

In summary, as with the effects with psychostimulants and alcohol, our findings indicate that estradiol enhances vulnerability in females to opioid use and the development of an opioid addiction-like phenotype. Our findings also indicate that estradiol prolongs the course of physical dependence and increases the likelihood of developing life-threatening health complications during withdrawal. These effects are very concerning, and future studies are needed to understand the mechanisms for these adverse health effects. Further research is also needed to understand the neurobiological mechanism underlying the behavioral differences found here.

Acknowledgments

The authors would like to thank Katriel E. Cho for creating the graphical abstract.

Statement of Ethics

This study protocol was reviewed and approved by the Animal Care and Use Committee at the University of Virginia, approval number 3492-07-20.

Conflict of Interest Statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Funding Sources

This work was supported by NIDA R01 grants R01DA24716 and DA050094 (Lynch), NIDA F30 grant 1F30DA057808-01, and a Medical Sciences Training Grant T32GM007267 (Towers).

Author Contributions

E.B.T. and W.J.L. designed the study and wrote the manuscript. E.B.T., B.S., and W.J.L. performed the statistical analysis. E.B.T., B.S., and W.J.L. collected the data. All authors contributed to manuscript revision and read and approved the submitted version.

Data Availability Statement

The data that support the findings of this study are openly available in zenodo at https://doi.org/10.5281/zenodo.7664951. Further inquiries can be directed to Dr. Wendy Lynch.

Copyright: All rights reserved. No part of this publication may be translated into other languages, reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording, microcopying, or by any information storage and retrieval system, without permission in writing from the publisher.
Drug Dosage: The authors and the publisher have exerted every effort to ensure that drug selection and dosage set forth in this text are in accord with current recommendations and practice at the time of publication. However, in view of ongoing research, changes in government regulations, and the constant flow of information relating to drug therapy and drug reactions, the reader is urged to check the package insert for each drug for any changes in indications and dosage and for added warnings and precautions. This is particularly important when the recommended agent is a new and/or infrequently employed drug.
Disclaimer: The statements, opinions and data contained in this publication are solely those of the individual authors and contributors and not of the publishers and the editor(s). The appearance of advertisements or/and product references in the publication is not a warranty, endorsement, or approval of the products or services advertised or of their effectiveness, quality or safety. The publisher and the editor(s) disclaim responsibility for any injury to persons or property resulting from any ideas, methods, instructions or products referred to in the content or advertisements.