Enhanced expression of parvalbumin and perineuronal nets in the medial prefrontal cortex after extended‐access cocaine self‐administration in rats

The medial prefrontal cortex (mPFC) drives cocaine‐seeking behaviour in rodent models of cocaine use disorder. Parvalbumin (PV)‐containing GABAergic interneurons powerfully control the output of the mPFC, yet few studies have focused on how these neurons modulate cocaine‐seeking behaviour. Most PV neurons are surrounded by perineuronal nets (PNNs), which regulate the firing of PV neurons. We examined staining intensity and number of PV and PNNs after long‐access (6 h/day) cocaine self‐administration in rats followed by either 8–10 days extinction ± cue‐induced reinstatement or short‐term (1–2 days) or long‐term (30–31 days) abstinence ± cue‐induced reinstatement. The intensity of PNNs was increased in the prelimbic and infralimbic PFC after long‐term abstinence in the absence of cue reinstatement and after cue reinstatement following both daily extinction sessions and after a 30‐day abstinence period. PV intensity was increased after 30 days of abstinence in the prelimbic but not infralimbic PFC. Enzymatic removal of PNNs with chondroitinase ABC (ABC) in the prelimbic PFC did not prevent incubation of cue‐induced reinstatement but decreased cocaine‐seeking behaviour at both 2 and 31 days of abstinence, and this decrease at 31 days was accompanied by reduced c‐Fos levels in the prelimbic PFC. Increases in PNN intensity have generally been associated with the loss of plasticity, suggesting that the persistent and chronic nature of cocaine use disorder may in part be attributed to long‐lasting increases in PNN intensity that reduce the ability of stimuli to alter synaptic input to underlying PV neurons.

cocaine-seeking behaviour in rodent models of cocaine use disorder, [1][2][3] and the prelimbic PFC contributes to the maintenance of cocaine-associated memories. 4The output of mPFC pyramidal neurons is powerfully modulated by GABAergic interneurons, 5,6 but much less focus has been on how these inhibitory interneurons modify circuit output in substance use disorder studies.GABAeric neurons receive dopaminergic input 7 and are involved in cocaine-induced plasticity. 8,9In particular, parvalbumin (PV)-containing GABAergic interneurons generate fast-spiking activity needed for precise firing and for the coordination of network function. 10e majority of PV neurons in the cortex are encased in a specialized extracellular matrix (ECM) called the perineuronal net (PNN) that allows for their fast spiking capabilities and precise firing. 11PNNs form in the central nervous system during the critical period of development to stabilize neural networks in adulthood, 12 but they also play a central role in plasticity during learning and memory. 130][21][22][23] Decreases in PV intensity have also been associated with increased plasticity, including retrieval of long-term memory. 23,24Only a few studies have examined the impact of voluntary intake using the cocaine self-administration paradigm on PNNs. 25,26These studies incorporated the extended access protocol over a 6-h period, which promotes escalation of cocaine consumption over days and increases motivation to take cocaine. 27In addition, 6-h cocaine self-administration can induce an 'incubation' of cocaine-seeking behaviour after several weeks of abstinence. 280][31] Despite these studies, none have compared abstinence versus extinction sessions after extended cocaine access to determine the extent to which these manipulations alter the intensity of PNNs or their underlying PV interneurons in the mPFC, nor have they examined subsequent cue-induced reinstatement.Further, no studies have tested whether PNN removal after extended access prevents the incubation of cocaine seeking during cue-induced reinstatement.We conducted three experiments to (1) test the impact of extinction and cue-induced reinstatement on PNNs and PV neurons in the mPFC; (2) test the impact of 1-and 30-day abstinence on PNNs and PV interneurons in the mPFC; and (3) determine whether removing PNNs in the mPFC after training but prior to abstinence would reduce incubation of cue-induced reinstatement.

| Animals
A total of 85 adult male Sprague-Dawley rats were used across 12 treatment groups.Rats (Envigo, Livermore, CA) weighing 275-300 g at the start of the experiment were housed under a reverse light-dark cycle (lights off at 08:00 h) and given food and water ad libitum except when placed in the self-administration apparatus.Animals were doubly housed until surgery, after which they were singly housed to prevent damage to the intravenous (iv) catheters.Experiments were approved by the Washington State University Vancouver Institutional Animal Care and Use Committee according to the NIH Guide for the Care and Use of Laboratory Animals.All efforts were made to reduce animal suffering and to reduce the number of animals used.

| Drugs
Cocaine hydrochloride (NIDA) was a gift from the National Institute on Drug Abuse and dissolved as the weight of the salt in saline and given by the iv route (0.5 mg/kg/infusion in 0.10 mL over 6 s).Saline (iv) was used as a control.Protease-free chondroitinase-ABC (ABC; Sigma-Aldrich) was dissolved in sterile saline (vehicle) to a final concentration of 0.09 U/μL. 4 ABC degrades the glycosaminoglycan side chains of chondroitin sulphate proteoglycans of PNNs and the ECM. 32

| Surgery
Rats were anaesthetised with ketamine/xylazine (ketamine 87 mg/kg/ xylazine 13 mg/kg) and implanted with a chronic indwelling iv catheter into the right jugular vein. 33In rats that received microinjection, bilateral cannula were also implanted 1 mm above the prelimbic PFC using the following coordinates: A/P + 3.0 mm from bregma; M/L ± 0.7 mm from midline; D/V À2.5 mm from the skull surface.Rats were given carprofen gel cups (MediGel) 1 day prior to surgery and for 2 days after surgery.They were allowed to recover for 7 days prior to beginning experiments.Microinjection needles 1 mm longer than cannula were used to deliver 0.4 μL of saline vehicle or ABC bilaterally over 2 min followed by a 1 min post-infusion period.

| Self-administration
Self-administration chambers (Med Associates) were equipped with active and inactive levers, cue lights over each lever and a house light.
All studies were conducted during the same time of day for each group of animals, during the dark phase of the light cycle between 08:00 and 16:00 h.Correct active lever press resulted in reward delivery and activation of the cue light and a 20-s time-out period.Active lever presses during the time-out period and inactive lever presses had no consequences.Daily self-administration sessions were 6 h, 5 days a week, with a fixed-ratio 1 schedule (FR1) for 14-15 days.Extinction sessions were 2 h/day for 8-10 days.During extinction, reward and the cue light were omitted.

| Image acquisition and processing
Images of the prelimbic and infralimbic PFC were taken using a Leica  34 or PipsqueakTM was run in 'semi-automatic mode' to select ROIs to identify individual c-Fos, PV cells and PNNs, which were then verified by a trained experimenter who was blinded to the experimental conditions.The plug-in compiles this analysis to identify single-, double-and triple-labelled neurons (https://ai.RewireNeuro.com).

| Statistics
Data were statistically analysed using GraphPad Prism 9.For behavioural analysis, a two-way repeated measures analysis of variance (ANOVA) was used, with treatment as the between-subjects factor (saline vs. cocaine in Experiments 1 and 2 and vehicle vs.ABC in Experiment 3) and day as the within-subjects factor followed by a Šídák's post-hoc test.For WFA and PV intensity levels, distributions of log-normalized intensities (due to non-normal distributions) were compared between saline and cocaine groups using Student's twotailed t-test when comparing only two groups (Figures 2 and 3  We next tested whether short-or long-term abstinence from cocaine self-administration would alter PV or PNN expression.Rats were given 14-15 days of 6-h saline or cocaine access as described above. We tested the impact of 1 day (saline n = 6, cocaine n = 5) or 30 days of abstinence (saline n = 6, cocaine n = 5) on PV and PNN expression.
During the 30-day abstinence, rats were briefly handled 3 days/week.

| RESULTS
Table 1 shows that, during the training period, there were no differences among the four cocaine treatment groups of rats not given surgery for prelimbic PFC cannulae implantation (Experiments 1 and 2; total number of active (p = 0.3603) and inactive (p = 0.6849) lever presses and number of infusions (p = 0.1556)).There were also no differences among the four saline treatment groups for total number of active (p = 0.5608) or inactive (p = 0.8177) lever presses or for infusions (p = 0.6462).Within the four groups given prelimbic PFC cannulae for vehicle or ABC (Experiment 3), there were no differences during the training period for total number of active (p = 0.495) or inactive (p = 0.3127) lever presses or for infusions (p = 0.7324).When all six groups were compared, there was a difference in active lever presses (p = 0.044) and number of infusions (p < 0.0001) but not inactive lever presses (p = 0.8291) during training.We attributed these differences to decreased lever pressing behaviour in rats with chronic guide cannulae, which may have reduced activity or motivation to obtain cocaine.
Overall, these results indicate that, within each of the three experiments conducted, there were no differences in training between experimental and control groups.Table 2 shows a summary of all significant changes found in PV and PNNs after cocaine self-administration or ABC treatments, which are discussed in detail below.
3.1 | Experiment 1: Extinction and cue reinstatement of cocaine seeking alters PV and WFA intensity in the prelimbic and infralimbic prefrontal cortex

| Behaviour after extinction and cue reinstatement
Figure 1A shows the timeline for Experiment 1.The number of active lever presses (Figure 1B) and cocaine infusions (Figure 1C) were increased over 14 days of training for cocaine versus saline.Inactive lever presses during training (Figure 1D) were slightly but significantly reduced in rats responding to cocaine versus saline on days 6, 8 and 10.
Figure 1E,F shows extinction behaviour over 8 days for active and inactive lever presses, respectively.Responding for both levers was extinguished in cocaine self-administering rats over the first 1-2 days.
Figure 1G shows differences in active lever presses between the last extinction session and the cue reinstatement session.Both groups showed significant cue-induced reinstatement compared with their last extinction day; however, the cocaine group reinstated higher than the saline group.The number of cue 'rewards' (cue light only) was also higher in cocaine versus saline rats (Figure 1H).Inactive lever pressing was increased across days in saline but not in cocaine self-administering rats.Thus, although both cocaine and saline animals were reinstated following extinction, the number of lever presses and cue 'rewards' for cocaine animals was more than twice that of saline controls.

| PV and PNNs in prelimbic and infralimbic PFC after extinction and cue reinstatement
We measured the intensity and number of PV and WFA doublelabelled neurons to assess cocaine-induced changes in the level of PV in neurons surrounded by PNNs and in the level of PNNs. Figure 2A,B shows that PV intensity in WFA-surrounded cells in the prelimbic PFC decreased in cocaine versus saline animals after extinction, whereas Figure 2C,D shows that there were no differences between groups after cue reinstatement.Figure 2E-H shows that WFA intensity surrounding PV cells in the prelimbic PFC increased relative to saline animals after both extinction and cue reinstatement, with an apparent greater increase after cue reinstatement.
In the infralimbic PFC, Figure 2I,J shows that PV intensity in WFA-surrounded cells decreased in cocaine versus saline groups after extinction, whereas Figure 2K,L shows that there were no differences in PV intensity between treatment groups after cue reinstatement.Figure 2M-P shows that WFA intensity surrounding PV cells in both extinction and cue reinstatement groups increased in cocaine versus saline animals that underwent extinction, and as in the prelimbic PFC, the increase in cocaine animals after cue reinstatement appeared to be greater than after extinction alone.We next determined the extent to which PNNs and PV were altered following acute (1 day) and extended (30 days) abstinence from cocaine in rats that were trained to self-administer saline or cocaine during daily 6-h sessions.Figure 3A shows the timeline for Experiment 2. The number of active lever presses (Figure 3B) and cocaine infusions (Figure 3C) were increased over 14 days during training for cocaine relative to saline.Inactive lever presses during training (Figure 3D) were slightly but significantly reduced in rats responding to cocaine versus saline on one of the training days (day 8).
After training, animals were randomly assigned to 1 day or 30 days of forced abstinence.The prelimbic and infralimbic PFC were examined for PV and WFA staining intensity and cell number after either 1 day or 30 days of abstinence.Figure 3E,F shows that PV intensity in WFA-surrounded cells in the prelimbic PFC was not altered after 1 day of abstinence.However, Figure 3G,H shows that PV intensity increased after 30 days of abstinence.Figure 3I,J shows that WFA intensity surrounding PV cells was decreased after 1 day of abstinence but increased after 30 days of abstinence (Figure 3K,L).
In the infralimbic PFC, Figure 3M-P shows no changes in PV intensity in WFA-surrounded cells after 1 day or 30 days of abstinence.
However, although WFA intensity was not altered after 1 day of abstinence (Figure 3Q,R), WFA intensity was increased at 30 days (Figure 3S,T).In summary, we found an incubation-like effect (increased intensity) after 30 days of abstinence for PV in the prelimbic cortex and for WFA intensity in both prelimbic and infralimbic PFC.Although there were no changes in the number of WFA-surrounded PV neurons after cocaine treatment for most conditions, the increased intensity of WFA may have conferred changes in the properties of the underlying PV cells.
T A B L E 2 Summary of changes in PV and PNNs.We then tested whether PNN removal after training for cocaine selfadministration but prior to the 30-day abstinence period would prevent incubation of cue reinstatement.Figure 5A shows the timeline for Experiment 3. The number of active lever presses (Figure 5B) and the number of cocaine infusions (Figure 5C) were increased over 14 days during training for cocaine relative to saline.Inactive lever presses during training (Figure 5D) were not different across days.
During cue reinstatement, we again observed an incubation effect from 2 to 31 days of abstinence, which occurred in both vehicle and ABC groups.Post-hoc analysis showed that there was a reduction in the ABC group at 31 days but not at 1 day of abstinence (Figure 5E).
However, it was apparent that there was a weak trend toward reduced active lever presses in the ABC group at 1 day of abstinence (p = 0.1157).A similar picture emerged for the number of cue 'rewards' (Figure 5F), but in this case, there was only an incubation effect in the ABC group, even though the ABC group obtained a

| c-Fos activated cells after PNN depletion
We also measured the number of cells activated in the prelimbic PFC after cue reinstatement, as detected by c-Fos staining.were not able to assess c-Fos after 2 days of abstinence due to aberrant c-Fos staining near the site of microinjection, observed for both vehicle and ABC-treated rats, suggesting that the damage induced by microinjection led to aberrant c-Fos staining 24 h later.The number of total PV cells was not altered across the incubation period or after ABC treatment (Figure 6G).The number of total WFA-stained cells was reduced with ABC treatment after 2 days (as expected) but not after 31 days of abstinence (Figure 6H).The number of total WFA-stained cells was increased from 2 to 31 days of abstinence in both vehicle-and ABC-treated rats.
For double-labelled PV/WFA cells, there was an expected decrease after ABC at 2 days of abstinence and an expected increase in ABC-treated rats across time as PNNs slowly re-form (Figure 6I).

| Long access to cocaine produces long-term changes in PNN intensity
The extent to which changes in PNNs occur after cocaine exposure appears to depend on the dose and route of administration (for review see previous studies 35,36 ).In our previous work using investigatoradministered cocaine, a single intraperitoneal cocaine injection decreased PNN intensity 2 h later, and five daily cocaine injections increased PNN intensity 2 h later, 14 but this effect disappeared by 24 h.In the current study, the increased PNN intensity found 30 days after the last self-administration day under basal conditions is consistent with this earlier study but suggests that the effects after repeated 6-h cocaine exposures are more robust and much longer lasting.
Two other studies have tested whether long-term abstinence from cocaine self-administration altered PNNs.One study examined the intensity and number of PNNs in the mPFC after 1 and 30 days of 6-h cocaine self-administration in rats; these findings are somewhat in contrast to what we report here.In that study, the number of PNNs was increased in the prelimbic and infralimbic PFC after 1 day and was no longer increased in the prelimbic PFC after 30 days of abstinence from cocaine, although the increase in the number of PNNs was dependent on the intensity and hemisphere. 26In the cerebellar cortex, PNN intensity was decreased in Golgi interneurons after 1 day following short access to cocaine (1 h/day) but was increased over 28 days of abstinence after long access to cocaine (6 h/day). 25few studies have also tested whether long-term changes in PNNs occur after investigator-administered cocaine.In the cerebellum, repeated cocaine exposure increased the intensity of PNNs around glutamatergic neurons in the medial nucleus when examined 1 day after the last cocaine, 16 but the same cocaine exposure decreased the intensity around the same type of neurons when a cocaine challenge was given after 1-month abstinence, 15 suggesting that abstinence induces further plasticity in cerebellar output neurons.One other study examined PNN/ECM components after heroin self-administration 37 after a 3-week abstinence period or an equivalent period of extinction sessions; they reported decreases in the mPFC of the ECM proteins brevican and/or tenascin R, proteins found within PNNs.Interestingly, a 30-min cue-induced reinstatement reversed the decreases in brevican. 37Together with our findings, this latter study suggests that drug cues alone during reinstatement can induce rapid increases in ECM proteins and PNNs, but the mechanism for these rapid changes remains unknown.One intriguing possibility is that ECM/PNN components are recycled via endocytosis and resurfacing in neurons, as was recently shown for tenascin R via its binding to ß1-integrin. 382 | Long access to cocaine alters PV intensity after 30 days of abstinence PV neurons and PNNs underwent similar changes after extinction, while PV intensity was increased in the prelimbic but not infralimbic PFC after 30 days of abstinence.Therefore, abstinence may alter PV neuronmediated function in the prelimbic PFC.In addition, extinction, but not abstinence, reduced PV levels in both brain regions.The functional consequences of these alterations in PV levels remain unknown.On the one hand, PV levels are highly correlated with GAD-67 levels, and low PV levels have been associated with early stages of new learning, whereas high-intensity PV levels have been associated with completed or 'definite' learning that interferes with the acquisition of new memories.39 The reduced PV levels after extinction but not after reinstatement points to the intriguing idea that extinction may produce a persistent, high-plasticity state that is maintained by lower PV expression.
Cocaine-induced increases in dopamine may set into motion the changes in PV neurons and PNNs.Dopamine increases the excitability of PV fast-spiking neurons in the PFC 40 and stimulation of dopamine D1 receptors induces proteolysis of the PNN components aggrecan and brevican. 41Changes in PV neuron activity may alter PNN intensity levels through different mechanisms, including activity-dependent recycling of PNN components 38 or alterations in the orthodenticle homeobox protein 2 (OTX2) 42 or neural pentraxin-2. 43

| Prelimbic versus infralimbic PFC changes in PV and PNNs
We observed nearly identical responses in PV and WFA staining intensity between the prelimbic and infralimbic PFC when animals were given extinction followed by a cue reinstatement session.In contrast, when rats were not given extinction sessions but instead remained in the home cage for 30 days during forced abstinence, PV intensity was increased in the prelimbic PFC with no changes in the infralimbic PFC.
Roura-Martinez et al 26 found in the prelimbic PFC increases in specific intensities of PNNs at 1 day of abstinence, but they found decreases in specific intensities of PNNs at 30 days of abstinence using a similar 6-h cocaine self-administration protocol as used in the current study, and these changes were dependent on the hemisphere.In the infralimbic PFC, they found that abstinence (combined 1 day + 30 days) increased specific intensities of PNNs in cocaine versus saline self-administering rats.One difference in our study versus the Roura-Martinez study is that their study was conducted during the light phase, whereas ours was conducted during the dark phase, and we and others have reported diurnal variation in PNNs. 44,45In addition, although there appeared to be comparable active lever-pressing during the 6-h period between the two studies and the dose of cocaine was only slightly higher in the Roura-Martinez study (0.75 vs. 0.5 mg/kg in our study), it is unclear how much behavioural incubation occurred in the Roura-Martinez study, as it was not specifically reported.
Several studies have implicated differential roles for the prelimbic versus infralimbic PFC in cocaine-seeking behaviour, 46 and their roles of prelimbic and infralimbic PFC can be altered after a forced abstinence period. 47In particular, after 30 days of abstinence, inhibition of the dorsal mPFC did not alter cue reinstatement. 48However, other studies support a role for the dorsal mPFC after abstinence. 3,49A role for the infralimbic PFC in cocaine cue-induced reinstatement after abstinence is mixed.Although pharmacological inhibition of the infralimbic PFC blocked cue-induced reinstatement after 30-day abstinence, 48 maturation of silent synapses from infralimbic PFC to NAc shell projections appeared to provide an 'anti-relapse' function after prolonged abstinence. 3This latter study is consistent with the finding that, after abstinence, optogenetic stimulation of projections from the infralimbic PFC to NAc shell inhibited cue-induced reinstatement. 50However, there are some exceptions in which optogenetic or chemogenetic activation of the infralimbic PFC did not alter cue-induced reinstatement after abstinence. 51,52The results from these studies need to be considered in light of several others that collectively suggest that the infralimbic PFC contributes more nuanced effects that can either promote or inhibit behaviours, depending on whether the behaviour is adaptive. 47

| Removal of PNNs reduces cue-induced reinstatement
Removal of PNNs with ABC reduced cue-induced reinstatement at both early and late abstinence times, but ABC treatment did not prevent the incubation of cue-induced reinstatement.Over several weeks, PNNs slowly return in number but not in intensity, 4 and therefore are not likely to resume normal functioning even 31 days later. 20 typical memory studies that use relatively weak stimuli versus strong stimuli, such as drugs of abuse or footshock, PNN removal enhances plasticity and new learning 13,20,[53][54][55] (but see Paylor et al. 56 ).Paradoxically, however, most studies involving administration of drugs of abuse show that PNN removal reduces responses to drugs or drug-associated cues. 35,36For example, PNN removal with ABC in the prelimbic but not infralimbic PFC reduced the acquisition and reconsolidation of cocaine-induced CPP. 4 PNN degradation with ABC in the lateral hypothalamus blocked the acquisition and cue-induced reinstatement of cocaine self-administration. 33,57 PNN removal with ABC infusion into the amygdala decreased cocaine-primed reinstatement in both CPP and self-administration models and also reduced morphine-primed reinstatement of CPP and heroin-primed reinstatement in a self-administration model. 58In the insular cortex, PNN removal with ABC increased sensitivity to quinine-induced aversion to ethanol. 59Thus, most studies, including the current study, support a role for PNN removal to reduce drug-induced behaviours and likely the plasticity states induced by drugs of abuse.
We found positive correlations between cue-induced behavioural responding and WFA intensity around PV neurons only in the ABCtreated groups but not in the vehicle-treated groups.This was apparent at both 2 and 31 days of abstinence, overall suggesting that reduced PNN intensity may drive lower cue-induced responding.0][31] Although speculative, the mechanism by which PNN removal diminished cue-induced reinstatement may instead rest on an inability of PV neurons to synchronize pyramidal neuron output necessary to form discrete activated ensembles expressed during memory.A reduced ability to fire at high rates would disrupt their precise firing patterns.This disruption may in turn produce less sparse and consequently overlapping neural ensembles that lead to less specificity of ensembles representing a particular stimulus or memory. 60,61In the current study, this may manifest as reduced cue-induced reinstatement.Additionally, PNN removal may induce changes in ion channel function.Although degradation of PNNs with hyaluronidase treatment did not alter the lateral mobility of alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) receptors on aspiny interneurons on the short time scale of minutes, 62 it may do so over a longer period of time.Reducing levels of the PNN component brevican reduces potassium ion channels and AMPA receptor subunit expression in PV neurons, 63 which would be expected to alter the output of PV neurons.

| PNN removal reduces c-Fos levels after forced abstinence followed by cue reinstatement
The levels of c-Fos are increased after cocaine cue-induced reinstatement. 64In our study, ABC cue reinstatement activated fewer prelimbic PFC neurons after 31 days of abstinence, as assessed by the number of c-Fos-labelled cells.Although we did not assess c-Fos levels in the absence of reinstatement for comparison, the decreases in total c-Fos and c-Fos/PV/WFA triple-labelled cells after ABC suggest a global reduction in the activity of both pyramidal and PV neurons.

| Limitations
A limitation to interpreting our findings is that the ABC enzyme degrades chondroitin sulphate glycosaminoglycans on chondroitin sulphate proteoglycans 65 also degrades the loose ECM. 66However, studies using knockout mice for PNN components that are more specific for PNNs still show enhanced plasticity states, 23,67 and additional ABC treatment in knockout mice lacking the gene for the PNNspecific protein Crtl1 showed no further changes in object recognition memory. 20These studies strongly suggest that the enhanced plasticity is due to changes in PNNs.A second limitation of these studies is that we tested the impact of self-administration on PNNs and PV neurons only in male rats.1][72][73][74] Thus, we might expect a greater impact of cocaine self-administration in female versus male rats, but future studies would also need to take into account the estrous phase. 75,76

| CONCLUSIONS
In conclusion, our studies demonstrated a role for PNNs in cueinduced reinstatement after 6-h cocaine self-administration.Increases in PNN intensity have generally been associated with the loss of plasticity, suggesting that the persistent and chronic nature of cocaine use disorder may in part be attributed to long-lasting increases in PNN intensity that reduce the ability of stimuli to alter synaptic input to underlying PV neurons.Future studies should focus on an understanding of how PNNs alter network dynamics across the brain to reduce the influence of drug-related cues on relapse.
) or two-way ANOVA test within a given stain type when there were two levels of analysis (Figure5L-O).For WFA and PV cell numbers, Student's two-tailed t-test was performed when comparing only two groups (Figures4 and 6C-F) or a two-way ANOVA test within a given stain type when there were two levels of analysis (Figure6G-I).Data for Figure 5P-S were analysed using a Pearson correlation.Differences were considered significant at p < 0.05.Statistical results are presented in the figure legends, with the exception of those presented in

Figure
Figure 2Q-X shows confocal images of prelimbic and infralimbic PFC PV and WFA staining after extinction or cue reinstatement in saline and cocaine self-administering rats.In summary, both prelimbic and infralimbic PFC demonstrated nearly identical changes, with a decrease in PV intensity in PV/PNN cells following extinction but a return to saline control levels following cue reinstatement, whereas WFA intensity around PV cells increased following both extinction and cue reinstatement.

3. 2 |
Experiment 2: Extended abstinence from cocaine increases WFA and PV intensity in the prelimbic and infralimbic PFC

Figure
Figure4A-D shows that the number of PV/WFA double-labelled cells was not different between saline and cocaine groups after extinction or cue reinstatement or at 1 or 30 days of abstinence in the prelimbic cortex.In the infralimbic PFC, we found an increase in the number of PV/WFA cells in the cocaine versus saline group after 30 days of abstinence (Figure4H) but not in any other treatment groups (Figure4E-G).

ac- 3 . 3 |
Fos staining decreased in: total c-Fos cells, c-Fos/PV cells, c-Fos/WFA cells and c-Fos/PV/WFA cells.F I G U R E 1 Experiment 1 extinction and cue reinstatement behaviour.(A) Timeline of experiment.(B) Training: Active lever presses for cocaine versus saline increased across the 6-h cocaine self-administration training sessions on days 4-14.There was a main effect of treatment (F 1, 34 = 117.2,p < 0.0001), day (F 2.878, 97.86 = 8.004, p < 0.0001), and a treatment Â day interaction (F 13, 442 = 6.065, p < 0.0001).(C) Training: Number of cocaine versus saline infusions increased across the 6-h self-administration training sessions, with a main effect of treatment (F 1, 34 = 234.6,p < 0.0001), day (F 3.878, 131.8 = 36.38,p < 0.0001), and a treatment Â day interaction (F 13, 442 = 24.30,p < 0.0001).(D) Training:Inactive lever presses during 6-h training sessions were slightly but significantly lower in cocaine versus saline animals on days 6, 8 and 10, with a main effect of day (F 4.469, 147.5 = 2.694, p < 0.0280) and a treatment Â day interaction (F 13, 429 = 3.529, p < 0.0001).(E) Extinction: Active lever presses during 2 h extinction sessions were increased in cocaine versus saline animals on days 1 and 2 of extinction and were similar to saline animals across remaining extinction days, with an effect of treatment (F 1,34 = 22.00, p < 0.0001), day (F 3.133,106.5= 13.29,p < 0.0001) and a treatment Â day interaction (F 7,238 = 13.94,p < 0.0001).(F) Extinction: Inactive lever presses across extinction days was increased in cocaine versus saline animals on day 1 of extinction and similar to saline animals across remaining extinction days, with an effect of treatment (F 1,34 = 7.258, p = 0.0109), day (F 2.383,81.01= 5.028, p = 0.0058), and a treatment Â day interaction (F 7,238 = 8.218, p < 0.0001).(G) Reinstatement: active lever presses during last extinction session and cue-induced reinstatement.There a main effect of treatment (F 1, 20 = 13.23,p < 0.0016), day (F 1, 20 = 119, p < 0.0001), and a treatment Â day interaction (F 1, 20 = 38.02,p < 0.0001).Active lever pressing after cue-induced reinstatement was higher than extinction in both saline (p = 0.0063) and cocaine groups (p < 0.0001), with higher active lever responding in the cocaine versus saline group.(p < 0.0001).(H) Reinstatement: Number of rewards was higher for the cocaine versus saline group (p = 0.0005).(I) Reinstatement: inactive lever presses in the saline, but not cocaine, group showed a slight but significant (p = 0.0005) increase from last extinction day to cue reinstatement day, with a main effect of day (F 1, 20 = 9.102, p < 0.0068).N = 11/group; *p < 0.05F I G U R E 2Experiment 1: Extinction decreases PV intensity but increases PNN intensity, whereas cue reinstatement increases PNN intensity in prelimbic and infralimbic PFC.Data show median and upper and lower quartiles; for additional clarity comparing percent change from saline controls, bar graphs show mean ± standard error of the mean (SEM).(A, B) extinction decreased PV intensity in PV/WFA double-labelled cells in the prelimbic PFC (p < 0.0001).(C, D) Cue reinstatement did not alter PV intensity in PV/WFA doublelabelled cells in the prelimbic PFC.(E, F) Extinction increased WFA intensity around PV/WFA doublelabelled cells in the prelimbic PFC (p < 0.0001).(G, H) Cue reinstatement increased WFA intensity around PV/WFA doublelabelled cells in the prelimbic PFC (p < 0.0001).(I, J) Extinction decreased PV intensity in PV/WFA double-labelled cells in the infralimbic PFC (p = 0.0054).(K, L) Cue reinstatement did not alter PV intensity in PV/WFA doublelabelled cells in the infralimbic PFC.(M, N) Extinction slightly increased WFA intensity around PV/WFA double-labelled cells in the infralimbic PFC (p = 0.0211).(O, P) Cue reinstatement increased WFA intensity around PV/WFA doublelabelled cells in the infralimbic PFC (p < 0.0001).Intensity data were normalized to saline groups at the same time point within the same brain region.(Q-X) Confocal images showing PV-(red) and WFA-(green)labelled cells in the prelimbic and infralimbic PFC after extinction or cue reinstatement.N = 11/group; *p < 0.05 Experiment 3: Removal of PNNs attenuates incubation of cue reinstatement 3.3.1 | Behaviour after PNN depletion in the prelimbic PFC prior to incubation

F I G U R E 3
Experiment 2: Long-term abstinence increases PV intensity in prelimbic PFC and WFA intensity in prelimbic and infralimbic PFC.(A) Timeline of experiment.(B) Active lever presses for cocaine versus saline increased across the 6-h cocaine self-administration training sessions on days 4-10, 13 and 14.There was a main effect of treatment (F 1, 22 = 35.49,p < 0.0001) and a treatment Â day interaction (F 13, 286 = 1.95, p < 0.0247).(C) Number of cocaine versus saline infusions increased across the 6-h self-administration training sessions on days 3-14, with a main effect of treatment (F 1, 22 = 106.8,p < 0.0001), day (F 3.467, 76.28 = 15.72,p < 0.0001), and a treatment Â day interaction (F 13, 286 = 10.10,p < 0.0001).(D) Inactive lever presses for cocaine versus saline during training were lower on day 10.There was a main effect of day (F 1.441, 31.71= 7.662, p = 0.0044) and a treatment Â day interaction (F 13, 286 = 2.430, p < 0.0039).(E-T) PV and WFA intensity in double-labelled cells.Data show median and upper and lower quartiles; for additional clarity comparing percent change from saline controls, bar graphs show mean ± SEM. (E, F) 1 day of abstinence did not alter PV intensity in PV/WFA double-labelled cells in the prelimbic PFC.(G, H) 30 days of abstinence increased PV intensity in PV/WFA double-labelled cells in the prelimbic PFC (p < 0.0001).(I, J) 1 day of abstinence decreased WFA intensity in WFA/PV double-labelled cells in the prelimbic PFC (p = 0.0038).(K, L) 30 days of abstinence increased WFA intensity in WFA/PV double-labelled cells in the prelimbic PFC (p < 0.0001).(M, N) 1 day of abstinence did not alter PV intensity in PV/WFA double-labelled cells in the infralimbic PFC.(O, P) 30 days of abstinence did not alter PV intensity in PV/WFA double-labelled cells in the infralimbic PFC.(Q, R) 1 day of abstinence did not alter WFA intensity in WFA/PV double-labelled cells in the infralimbic PFC.(S, T) 30 days of abstinence increased WFA intensity in WFA/PV double-labelled cells in the infralimbic PFC p < 0.0001).Intensity data were normalized to saline groups at the same time point within the same brain region. 1 day saline N = 7; 1 day cocaine N = 6; 30 day saline N = 6; 30 day cocaine N = 5. *p < 0.05 reduced number of cue 'rewards' at 31 days, with a trend toward reduced cue 'rewards' at 1 day ( p = 0.0765).For inactive lever presses, only the vehicle group showed an incubation effect.Overall, although ABC treatment did not prevent incubation across time, the ABC group responded less during cue-induced reinstatement at 31 days of abstinence.

Figure 5H -
Figure 5H-K shows staining for PV and WFA at 2 and 31 days of abstinence (1 and 30 days after vehicle or ABC treatment, respectively).Apparent are bright WFA-stained neurons in vehicle-treated rats, with much weaker WFA staining and fewer WFA/PV neurons stained after ABC treatment.Figure 5L,M shows that PV intensity was increased in the vehicle group between 2 and 31 days of abstinence.ABC treatment slightly increased PV intensity at 2 days but slightly decreased PV intensity at 31 days of abstinence.Figure 5N,O shows that WFA intensity around PV cells was increased in the Figure 6A,B shows confocal microscope images with PV, WFA and c-Fos staining F I G U R E 4 Experiments 1 and 2: Long-term abstinence increases the number of PV/WFA double-labelled cells in infralimbic PFC.(A, B) Neither extinction (A) nor cue reinstatement (B) altered the number of PV/WFA double-labelled cells in the prelimbic PFC.(C, D) Neither day (C) nor days (D) abstinence altered the number of PV/WFA double-labelled cells in the prelimbic PFC.(E, F) Neither extinction (E) nor cue reinstatement (F) altered the number of PV/WFA double-labelled cells in the infralimbic PFC.(G) 1 day of abstinence did not alter the number of PV/WFA double-labelled cells in the infralimbic cortex, whereas (H) 30 days of abstinence increased the number of PV/WFA double-labelled cells in the infralimbic cortex (p = 0.0445).All data are expressed as mean ± SEM.N sizes are as in Figures 1 and 3. *p < 0.05 after vehicle or ABC treatment in the prelimbic PFC after vehicle or ABC 31 days later.At 31 days of abstinence, ABC decreased the of c-Fos-labelled cells (total c-Fos single-labelled, c-Fos/PV or c-Fos/WFA double-labelled and c-Fos/PV/WFA triple-labelled) (Figure 6C-F).We

F I G U R E 5
Experiment 3: PNN removal decreases cocaine-seeking behaviour and alters PV and PNN intensity but does not block incubation.(A) Timeline of experiment.(B) Active lever presses during 6-h cocaine self-administration training sessions increased across time in all groups, with a main effect of day (F 3.651,76.66= 3.963, p = 0.0071).(C) Number of cocaine infusions during 6-h self-administration training sessions increased across time in all groups, with a main effect of day (F 4.912,103.1 = 36.14,p < 0.0001).(D) Inactive lever presses during 6-h cocaine self-administration training sessions did not change across time.(E) Active lever presses during cue reinstatement show that both treatment groups demonstrated incubation of cocaine-seeking behaviour, with a main effect of treatment (F 1, 21 = 10.9, p = 0.0034) and day (F 1,21 = 20.45,p = 0.0002); (across days, vehicle p = 0.0089; ABC p = 0.0078).Responding was reduced in the ABC versus vehicle group at 31 days (p = 0.0259).(F) Number of cocaine infusions during cue reinstatement showed that the ABC-treated rats showed incubation with a main effect of treatment (F 1, 21 = 11.25, p = 0.003) and day (F 1, 21 = 11.25, p = 0.003); (across days, ABC p = 0.0386).Number of infusions was reduced in the ABC versus vehicle group at 31 days (p = 0.0361).(G) Inactive lever presses during cue reinstatement revealed an effect of day (F 1, 21 = 11.39,p = 0.0029), with an incubation in the vehicle group (p = 0.0119).(H-K) confocal microscope images showing PV and WFA staining after vehicle or ABC treatment in the prelimbic PFC 2 days or 31 days later.White arrows show double-labelled neurons; note faint staining of PNNs after ABC.White line = 100 μm.(L-P) PV and WFA intensity in PV/WFA and WFA/PV double-labelled cells.Data show median and upper and lower quartiles; for additional clarity comparing percent change from saline controls, and bar graphs show mean ± SEM. (L, M) 31 days of abstinence increased PV intensity in PV/WFA double-labelled cells in the vehicle but not ABC group, with a treatment Â day interaction (F 1,1,539 = 11.31,p = 0.0008).PV intensity was increased from day 2 to day 31 in the vehicle-treated group (p < 0.0001).Post-hoc analysis also revealed at 2 days of abstinence a slight increase in PV intensity in the ABC group (p = 0.0233) but at 31 days of abstinence a slight decrease in PV intensity after ABC (p = 0.0191).(N, O) ABC treatment reduced WFA intensity in remaining WFA/PV double-labelled cells at both 2 and 31 days of abstinence, with a main effect of treatment (F 1,1,539 = 82.45,p < 0.0001) and day (F 1,1,539 = 32.37,p < 0.0001).(across days, vehicle p < 0.0001, ABC p = 0.0144).31 days of abstinence increased WFA intensity in WFA/PV double-labelled cells relative to 2 days of abstinence in the vehicle (p < 0.0001) and ABC groups (p < 0.0001).(P-S) Correlations between WFA intensity in WFA/PV double-labelled cells and active lever presses or rewards for ABC groups at 2 and 31 days.All data are expressed as mean ± SEM.Intensity data were normalized to PV intensity at 2 days of abstinence in vehicle-treated rats.2-day vehicle N = 5; 2-day ABC N = 6; 31-day vehicle N = 5; 31-day ABC N = 9. *p < 0.05 Here, we report several novel findings from rats given 14 days of 6-h cocaine self-administration: (1) Repeated extinction sessions after cocaine self-administration increased PNN intensity in the prelimbic and infralimbic PFC, (2) 30 days of abstinence from cocaine selfadministration increased PV intensity in the prelimbic but not infralimbic PFC, (3) Incubation of lever-pressing behaviour during cue reinstatement after a 30 day abstinence period was accompanied by increases in PNN intensity in the prelimbic and infralimbic PFC, (4) enzymatic removal of PNNs in the prelimbic PFC did not block incubation of cocaine-seeking behaviour but did attenuate cueinduced reinstatement at 31 days of abstinence, (5) PNN intensity after ABC was correlated with cue-induced reinstatement after both short-and long-term abstinence and 6) enzymatic removal of PNNs reduced c-Fos levels in the prelimbic mPFC in response to cueinduced reinstatement after 30 days of abstinence.

Experiment 3 :
PNN removal reduces the number of c-Fos cells in the prelimbic PFC after cue-induced reinstatement at 31 days of abstinence.(A, B) confocal microscope images showing PV, WFA, and c-Fos staining after vehicle or ABC treatment in the prelimbic PFC after vehicle or ABC 31 days later.White arrows indicate triple-labelled cells.Note faint staining of PNNs after ABC treatment.(C-F) ABC reduced the number of c-Fos (p = 0.0081), c-Fos/PV double-labelled cells (p = 0.0317), c-Fos/WFA double-labelled cells (p = 0.0154) and c-Fos/PV/WFA triple-labelled cells (p = 0.0168), respectively, after cue-induced reinstatement at 31 days of abstinence in the prelimbic PFC.(G) Number of total PV single-labelled cells was not altered across abstinence days or after vehicle or ABC treatment.(H) Number of total WFA single-labelled cells was altered across abstinence days and ABC treatment, with a main effect of treatment (F 1,21 = 35.79,p < 0.0001), day (F 1,21 = 48.63,p < 0.0001) and a treatment Â day interaction (F 1,21 = 9.135, p = 0.0065).Number of WFA cells was reduced at 2 days (p < 0.0001) but not at 31 days of abstinence.Number of WFA cells was increased across abstinence after cue-induced reinstatement in both vehicle (p = 0.0352) and ABC (p < 0.0001) groups.(I) Number of WFA/PV double-labelled cells was altered across abstinence days and ABC treatment, with a main effect of treatment (F 1,21 = 18.47, p = 0.0003), day (F 1,21 = 31.42,p < 0.0001), and a treatment Â day interaction (F 1,21 = 4.593, p = 0.0440).Number of PV/WFA cells was reduced after ABC at 2 days (p = 0.0005) with an increase in these cells across abstinence (p < 0.0001).All data are expressed as mean ± SEM. 2-day vehicle N = 5; 2-day ABC N = 6; 31-day vehicle N = 5; 31-day ABC N = 9. *p < 0.05

Table 1 ,
presented in the Results section.Total lever presses and infusions during cocaine self-administration training.