A total of 39 Long-Evans rats were used in the current study. Four pregnant dams were ordered at gestation day 10 from Charles River Labs (Wilmington, MA, United States). All dams were kept in a temperature-controlled vivarium (22–24°C) at Arizona State University on a standard reverse light-dark cycle (12 hr light/12 hr dark). Rats were single housed in conventional polycarbonate cages until pups were weaned. All animals were given standard rodent chow and water ad libitum for the duration of the study. All experiments conducted were approved by the Institutional Animal Care and Use Committee at Arizona State University.

A total of 4 litters were used in the current study. At PND2, the sex of each pup was determined via anogenital distance. To minimize potential genetic influences, pups from all litters were randomly assigned to one of two conditions (LBN or control), and sex was kept at a 1:1 ratio. To minimize potential litter influences, dams assigned to the LBN group were placed with their assigned pups in a standard cage containing a wire mesh floor designed to maintained animals ∼2.5 cm above the cage floor, which contained a thin layer of standard bedding. However, due to some pups becoming entrapped in the mesh during the 1st day of LBN procedures (PND 2), wire meshes were removed from the cages and dams assigned to the LBN condition were provided only with ∼100 ml of standard bedding and a single folded paper towel to use as nesting material (Walker et al., 2017) for the remainder of LBN procedures (PND3-9). Control dams and pups were housed in cages containing standard amounts of bedding and nesting material. Dams and pups were returned to normal housing during PND10–21. On PND21, all pups were weaned and same sex pair housed until PND70, where animals either underwent surgical implantation of jugular vein catheters for intravenous heroin self-administration (Experiment 1), or stereotaxic injection of retrograde AAV expressing enhanced green fluorescent protein (EGFP; Experiment 2). Following surgical procedures, all animals were single housed for the duration of the experiment. All dams were only used once to avoid influences of LBN on maternal behaviors in subsequent litters.

All animals used in Experiment 1 (n = 24; 12 males, and 12 females) were anesthetized with isoflurane (5% induction, 2–3% maintenance) vaporized in oxygen at a flow rate of 2 L/min and implanted with polyurethane catheters (BTPU-040; Instech Laboratories, Plymouth Meeting, PA) 2–2.5 cm into the right jugular vein and were tunneled subcutaneously to the posterior side of the animal and connected to an indwelling back port (Instech). Dental cement was used to adhere the catheter to the indwelling port, and the port was sutured ∼2 cm caudal from the shoulder blades using 4–0 vicryl braided sutures. Catheters were flushed daily with heparin (70 U/ml) containing timentin (66.6 mg/ml) during recovery to maintain patency and minimize postoperative infections. Meloxicam (1 mg/kg, s.c.) was administered immediately following surgery and once daily for the first 3 days of recovery to minimize discomfort. Throughout the remainder of the self-administration phase of the experiment, heparin (10 U/ml) was administered daily. Animals used in Experiment 2 (n = 15; 8 males, and 7 females) were anesthetized as described above underwent stereotaxic injections of the retrograde serotype AAV expressing EGFP (AAVrg-CAG-EGFP; titer >10¹² vg/ml; Addgene #37825) into the nucleus accumbens core (NAc) at a volume of 300—400 nl and at the following coordinates in mm from bregma and skull surface: A/P: + 2.5; M/L ± 1.6; DV: −6.1. The AAV was delivered over a 5 min period and the injector was left in place for an additional 10 min to allow for diffusion. The wound was then sutured closed and triple antibiotic ointment was applied to the suture site. Meloxicam (1 mg/kg s.c.) was administered immediately following surgery and for 6 consecutive days during recovery.

All self-administration procedures were conducted in Med Associates operant conditioning chambers as described elsewhere (Nagy et al., 2020). Animals were allowed to self-administer heroin (0.06 mg/kg/inf) via nosepoke on an FR1 reinforcement schedule in 3 hr daily sessions. A stimulus light and audible cues (2900 Hz tone and the sound of the syringe pump motor) were presented during each heroin infusion. After 15 days of self-administration, rats were subjected to progressive ratio (PR) testing where an exponential increase in the number of active nosepokes (i.e., 1, 2, 4, 6, 9, 12, 15, 20, 25, 32, etc.) was required to obtain each successive infusion. PR sessions were conducted in 3 hr daily sessions across for three consecutive days, and resulting data were averaged across days for analysis. Animals were then allowed to re-establish baseline IVSA patterns on an FR1 schedule for 2 days. Next, animals underwent extinction procedures where active nosepokes no longer produced heroin infusions or cue presentations. After reaching extinction criteria (>70% reduction in the number of active nosepoke responses as compared to the average value on the final 2 days of self-administration) animals were placed in their home cages and left undisturbed for 2 days prior to cue-induced reinstatement of heroin-seeking. For reinstatement sessions, previously paired heroin-associated cues (stimulus light, tone, and sound of the syringe pump motor) were presented upon active nosepokes. On the day following cue reinstatement testing procedures, active nosepoke responding was again extinguished using the same procedures and criteria above. A timeline of these procedures is outlined in Figure 1.

Prior to whole-cell electrophysiology, animals used for Experiment 2 were randomly assigned to receive once daily intraperitoneal (i.p.) injections of either saline or heroin (2 mg/kg) for five consecutive days. This dosage of heroin was selected given its ability to produce conditioned place preference (CPP) in rodents (Schlussman et al., 2008). One hour following the last injection, rats were asphyxiated with CO₂, rapidly decapitated and brains were extracted. Brains were placed into ice-cold carbogen (95% O₂/CO₂) saturated cutting solution containing the following: 120 mmol/l NaCl, 25 mmol/l NaHCO₃, 10 mmol/l dextrose, 3.3 mmol/l KCl, 1.23 mmol/l NaH₂PO₄, 1.8 mmol/l CaCl₂, and 2.4 mmol/l MgCl₂, osmolarity adjusted to 295 ± 5 mOsm and pH adjusted to 7.40 ± 0.03. Coronal slices (300 μm) containing the insular cortex were then prepared in ice-cold cutting buffer using a vibrating tissue slicer (Leica, VT 1000S). Slices were then transferred to a holding chamber filled with recording artificial cerebrospinal fluid (aCSF) containing in the following: 120 mmol/l NaCl, 25 mmol/l NaHCO₃, 3.3 mmol/l KCl, 1.23 mmol/l NaH₂PO₄, 0.9 mmol/l CaCl₂, 2.0 mmol/l MgCl₂, and 10 mmol/l dextrose, osmolarity adjusted to 295 ± 5 mOsm and pH adjusted to 7.40 ± 0.03. In the holding chamber, the recording solution was continuously saturated with carbogen and incubated at 34°C for 45 min, then allowed to cool at room temperature before being transferred to the recording chamber. Once transferred to the recording chamber, slices were perfused continuously with carbogen saturated recording solution. Pyramidal cells within the insular cortex expressing EGFP were visualized using an Olympus BX51WI microscope with infrared DIC microscopy and equipped with a blue collimated LED (ThorLabs). Whole-cell recordings were made from the soma of pyramidal neurons expressing EGFP using recording pipettes (7–15 mΩ) made from thin-walled capillary tubes containing an intracellular solution containing the following: 135 mmol/l K-gluconate, 12 mmol/l NaCl, 1 mmol/l K-EGTA, 10 mmol/l HEPES, 2 mmol/l Mg-ATP, and 0.38 mmol/l tris-GTP. Osmolarity was adjusted to 285 ± 5 mOsm and pH adjusted to 7.30 ± 0.01. Upon establishing a giga-seal, the cell membrane was ruptured and cells were allowed to equilibrate for 5 min until a steady state resting membrane potential was reached. In voltage clamp mode, a 10 mV square pulse was delivered via the recording pipette and 100 capacitive transients were recorded and averaged to determine membrane capacitance/resistance and series resistance. Recordings were then switched to current clamp mode, where rheobase and input-output (I-O) curves were assessed to determine the amount of injection current required to elicit action potential spiking. A total of 15 depolarizing current pulses (150 ms duration, stepped from −35 pA to 175 pA in 15 pA increments) were used. Action potential (AP) spike threshold, amplitude, rise time, and decay time were also measured, as were after-hyperpolarization (AHP) amplitude and duration. Each of these protocols have been described in detail in our prior studies (Leyrer-Jackson et al., 2020, 2021). Successful recordings from a total of 70 labeled cells (n = 17–18 cells/group).

Self-administration data were analyzed utilizing a two-way ANOVA with session and rearing condition (LBN and control) as considered factors. Extinction and re-extinction data were analyzed using a survival analysis, where LBN and control reared animals were compared. Cue-induced reinstatement data were analyzed utilizing a two-way ANOVA with session (extinction and cue-induced reinstatement) and rearing condition (LBN and control) as considered factors. Electrophysiological data were analyzed using a two-way ANOVA, where group (LBN and control) and treatment (saline and heroin) were considered factors. Bonferroni-corrected post hoc t-tests were conducted where appropriate and p-values <0.05 were considered statistically significant. The number of post hoc comparisons for all two-way ANOVAs was six. Additionally, unpaired t-tests were used to assess changes in rheobase, action potential amplitude, action potential rise time and action potential decay time between treatment groups (saline and heroin). Lastly, a two-way ANOVA with treatment (heroin and saline) and input current amplitude as factors was used to assess changes in excitability. Bonferroni’s post hoc comparisons were utilized to examine changes in the number of action potentials elicited at each current input. Statistical tests were performed in GraphPad Prism 8.0, and p < 0.05 was considered statistically significant. Values are represented as mean ± standard error of the mean (SEM).

All recordings were conducted from layer V pyramidal cells within the insula that expressed EGFP from retrograde viral labeling to the core, indicating that all recorded cells had axonal projections targeting the NAc. Representative labeling of AIC-NAc projecting cells can be seen in Figure 4A. Overall, AIC-NAc projection neurons had an average cellular capacitance of 25.6 ± 0.5 pA, resting membrane potential of −66.4 ± 0.4 mV and spiking threshold of −41.5 ± 1.1 mV. A two-way ANOVA, where rearing condition and drug were considered factors, revealed a significant effect of drug (F_1,61 = 4.3; p < 0.05) but not treatment (F_1,61 = 2.2; p < 0.05) or interaction (F_1,61 = 0.04; p < 0.05) between drug and treatment on cellular capacitance. A post hoc Bonferroni comparison revealed no significant differences between groups (p > 0.05; Figure 4B). A two-way ANOVA, where rearing condition and drug were considered factors, revealed a significant effect of rearing condition (F_1,61 = 9.8; p < 0.01; Figure 4C). A Bonferroni’s post hoc revealed a significant difference between control reared saline treated animals and LBN reared heroin-treated animals (t₆₁ = 3.3; p < 0.01; Figure 4C). Interestingly, for all other measurements described below, we found no significant differences between LBN and control reared Heroin or Saline treated animals. Thus, for all measurements described below, we have grouped Heroin animals (LBN + Control) as well as Saline animals (LBN + Control) for all analyses.

Heroin users have been found to experience higher levels of early life stress including childhood trauma such as physical punishment, physical neglect and penetrative sexual abuse (Grice et al., 1995; Moustafa et al., 2021). Exposure to ELS in rodents have validated such findings, demonstrating that ELS increases drug self-administration and relapse-related behaviors (Duffing et al., 2014; Koob and Schulkin, 2019). Furthermore, others have shown that maternal deprivation in rat pups increases the rewarding effects of morphine and promotes morphine seeking propensities and dependence in adulthood (Vazquez et al., 2005, 2006). For a more in depth review regarding ELS and the risks for opioid misuse, we direct the readers to the recent review by Oswald and colleagues (Oswald et al., 2021). Most similar to the study conducted here, the Mahler laboratory has also shown that LBN had no effect on overall heroin intake, yet promotes cue-induced reinstatement of heroin-seeking as well as impaired extinction learning in rats (Levis et al., 2019), which is consistent with the results we have reported here. In addition, we also demonstrated that LBN reared animals show higher breakpoints for heroin (increased number of heroin infusions earned) compared to control reared animals, demonstrating that these animals show elevated motivation for heroin. Taken together, our results suggest that LBN does not alter acquisition of heroin self-administration, yet increases motivation for the drug under conditions of increasing behavioral demand, reduces inhibitory learning with regards to heroin seeking (i.e., extinction) and increases the propensity for relapse (i.e., reinstatement) induced by heroin-associated cues.

Previous studies have shown that insular structure and function are adversely affected by both opiates and ELS. In fact, a meta-analysis of 12 neuroimaging studies of opiate-dependent individuals has recently shown evidence for gray matter atrophy within the insular cortex, accompanied by abnormal functional connectivity, resting state activity as well as white matter composition (Wollman et al., 2017). Similarly, individuals exposed to ELS showed reduced cortical thickness and volume in the insula (Baker et al., 2013; McLaughlin et al., 2014; Mackey et al., 2017; Lim et al., 2018), and rodents subjected to maternal stress or early life social isolation show neuronal loss in the AIC (Poeggel et al., 2005). ELS and repeated opiate exposure also affects the morphology and physiology of neurons in efferent targets of the AIC, including the NAc (Aleksic et al., 2016; Thompson et al., 2021). In the current study, we show that ELS alone did not have an effect on basic cellular physiology or excitability of AIC-NAc projecting pyramidal neurons. However, continuous exposure to heroin for five consecutive days resulted in decreased excitability of AIC-NAc projection neurons. We and others have found similar effects within the prefrontal cortex, where heroin decreases the AMPA/NMDA ratios in layer V pyramidal neurons through reductions in AMPA receptor expression and activity (Van den Oever et al., 2008) and decreases layer V pyramidal cell excitability of commissural projecting pyramidal neurons (Leyrer-Jackson et al., 2021). Further, non-contingent morphine exposure has also been shown to reduce cellular excitability within the mPFC as reflected by reductions in activity-related immediate early gene c-fos expression (Piao et al., 2017). Furthermore, heroin has been shown to increase GABAergic inhibition in the mPFC, therefore reducing pyramidal cell excitability (Van den Oever et al., 2008). Taken together, these studies highlight that heroin exposure has the ability to decrease pyramidal cell excitability within the prefrontal cortex through mechanisms that may also be observed within the anterior insula.

In the current study, we did not examine electrophysiological changes in AIC-NAc pyramidal neurons following heroin self-administration, extinction or reinstatement. Given that experimenter delivered and IVSA can have different effects on cellular excitability, it is possible that AIC-NAc pyramidal neurons display differences following each of these phases of IVSA as compared to experimenter delivered used in the current study. While there are limited studies addressing the differences in AIC-NAc pathway following contingent versus non-contingent heroin delivery, studies assessing effects of heroin on other pathways have reported some physiological differences. For example, non-contingent heroin delivery was found to reduce GABAergic neuronal firing within the ventral tegmental area (VTA), whereas heroin self-administration (contingent) initially results in a burst of GABAergic activity during nose-poke/approach activity, which was followed by a dramatic reduction in firing, below pre-approach levels (Steffensen et al., 2006). As well, other studies utilizing contingent and non-contingent cocaine and 3,4-methyledoxymethamphetamine (MDMA) delivery have shown enhanced dopamine transporter binding within the accumbens and VTA (Miguéns et al., 2008), and reduced dopamine levels within the NAc (Orejarena et al., 2009) following contingent administration relative to non-contingent administration of each drug, respectively. Given that these studies demonstrate physiological changes that vary between contingent and non-contingent drug administration paradigms, assessment of self-administered heroin effects on AI-NAc projections is an important future avenue to pursue. However, such a paradigm would need to incorporate the degree of heroin intake and parse the effects of low vs. high drug intake to make definitive claims in regards to the results presented in the current study. Nonetheless, the electrophysiological experiments conducted here are an important first step in understanding how heroin exposure can alter this specific reward pathway.

In the current study, we did not observe differences between AIC-NAc projection neurons recorded from heroin-treated LBN and control reared animals. This finding suggests that heroin significantly reduces the excitability of this pathway, regardless of LBN exposure. However, these lack of effects are possibly due to the short duration (i.e., 5 days) of heroin exposure. Given that LBN vs. control animals showed no differences in heroin IVSA in Experiment 1, it is not surprising that experimenter delivered heroin did not produce any differences in AIC-NAc excitability between ELS and control reared animals. It is likely that longer exposure (15 + days as used in the IVSA paradigm) to heroin as well as withdrawal would reveal differences in AIC-NAc projection neuron excitability between ELS and control reared animals. To fully explore this hypothesis, future studies exploring the effects long-term heroin exposure as well as heroin withdrawal on this pathway are necessary. Additionally, these studies would lend insight into potential neuronal underpinnings that lead to hindered extinction to heroin-paired cues as well as