Prosapip1-Dependent Synaptic Adaptations in the Nucleus Accumbens Drive Alcohol Intake, Seeking, and Reward

Abstract

The mammalian target of rapamycin complex 1 (mTORC1), a transducer of local dendritic translation, participates in learning and memory processes as well as in mechanisms underlying alcohol-drinking behaviors. Using an unbiased RNA-seq approach, we identified Prosapip1 as a novel downstream target of mTORC1 whose translation and consequent synaptic protein expression are increased in the nucleus accumbens (NAc) of mice excessively consuming alcohol. We demonstrate that alcohol-dependent increases in Prosapip1 levels promote the formation of actin filaments, leading to changes in dendritic spine morphology of NAc medium spiny neurons (MSNs). We further demonstrate that Prosapip1 is required for alcohol-dependent synaptic localization of GluA2 lacking AMPA receptors in NAc shell MSNs. Finally, we present data implicating Prosapip1 in mechanisms underlying alcohol self-administration and reward. Together, these data suggest that Prosapip1 in the NAc is a molecular transducer of structural and synaptic alterations that drive and/or maintain excessive alcohol use.

Keywords: Actin; Addiction; Alcohol; Dendritic spines; Plasticity; Prosapip1; mTORC1.

Figure 1. RNA-Seq Identified mRNAs Whose Translation Is Increased by Alcohol and depends on mTORC1 Activity in the NAc

(A) Mice experienced 7–8 weeks of IA20%-2BC. Control animals had access to water only. Three hours before the end of the last 24 hr of alcohol withdrawal, mice were systemically administered with 20 mg/kg of rapamycin or vehicle. The NAc from the four groups of mice (water/vehicle [blue], water/rapamycin [hatched blue], alcohol/vehicle [red], alcohol/rapamycin [hatched red]) were dissected 3 hr after the administration of rapamycin and were subjected to polysomal purification and RNA-seq analysis. (B) RNA-seq data depicting fragments per kilobase of transcript per million mapped reads (FPKM) value of Prosapip1, TAFA-3, Gucy1a3, and Tsnax. Data are presented as the average FPKM value ± SEM and expressed as the percentage of water/vehicle. Two-way ANOVA showed a significant interaction between alcohol and rapamycin (respectively, Prosapip1 F_(1,8) = 8.874, p = 0.018; TAFA-3 F_(1,8) = 10.02, p = 0.013, Gucy1a3 F_(1,8) = 7.98, p = 0.022; Tsnax F_(1,8) = 11.83, p = 0.009) and post hoc Student–Newman–Keuls test detected a significant difference between water and alcohol within the vehicle group (respectively, Prosapip1 p = 0.003; TAFA-3 p = 0.002; Gucy1a3 p = 0.002; Tsnax p = 0.007) and a significant difference between vehicle and rapamycin within the alcohol group (respectively, Prosapip1 p = 0.018; TAFA-3 p = 0.003; Gucy1a3 p = 0.005; Tsnax p = 0.04). n = 3. (C) A new cohort of animals that underwent the same paradigm described in (A) was used to measure polysomal mRNA level of Prosapip1, TAFA-3, Gucy1a3, and Tsnax by qRT-PCR. Data are presented as the average ratio of each transcript to GAPDH ± SEM, and expressed as the percentage of water/vehicle. Two-way ANOVA showed a significant interaction between alcohol and rapamycin (respectively, Prosapip1 F_(1,16) = 5.943, p = 0.027; TAFA-3 F_(1,16) = 4.654, p = 0.047, Gucy1a3 F_(1,28) = 7.6, p = 0.01; Tsnax F_(1,28) = 6.97, p = 0.013) and post hoc Student–Newman–Keuls test detected a significant difference between water and alcohol within the vehicle group (respectively, Prosapip1 p = 0.013; TAFA-3 p = 0.017; Gucy1a3 p = 0.006; Tsnax p = 0.005) and a significant difference between vehicle and rapamycin within the alcohol group (respectively, Prosapip1 p = 0.004; TAFA-3 p = 0.007; Gucy1a3 p = 0.007; Tsnax p = 0.007). n = 5 for Prosapip1 and TAFA-3; n = 8 for Gucy1a3 and Tsnax. *p < 0.05, **p < 0.01.

Figure 2. Alcohol-Dependent Translation of Prosapip1 Corresponds to Increases in Protein Levels that Are Maintained during Withdrawal

After 7–8 weeks of IA20%-2BC, NAc of mice were dissected 4 hr after the beginning (Binge, B) (A–D) or 24 hr after the end of the last drinking session (Withdrawal, WD) (E and F). Control animals had access to water only. (A and B) Prosapip1 mRNA levels in response to a binge session were determined by qRT-PCR in the polysomal fraction (A) and total fraction (B). Data are presented as the average ratio of Prosapip1 to GAPDH ± SEM and expressed as the percentage of water control. Significance was determined using two-tailed unpaired t test. (A) polysomal Prosapip1 mRNA t₍₆₎ = 3.207, p = 0.0184. n = 4. (B) total Prosapip1 mRNA t₍₆₎ = 0.07, p = 0.9464. n = 4. (C–F) Prosapip1 protein levels after binge (C and D) and withdrawal (E and F) were determined by western blot analysis. ImageJ was used for optical density quantification. Data are presented as the average ratio of Prosapip1 to GAPDH ± SEM (C and E) or Prosapip1 to actin ± SEM (D and F) and are expressed as the percentage of water control. Significance was determined using two-tailed unpaired t test. Prosapip1 in the total homogenate (C) t₍₁₄₎ = 4.359, p = 0.0007, n = 8 per group; and in the synaptic fraction (D) t₍₇₎ = 3.7, p = 0.0077, n = 4 water, 5 binge. Prosapip1 in the total homogenate (E) t₍₁₅₎ = 3.656, p = 0.0023, n = 8 water, 9 withdrawal; and in the synaptic fraction (F) t₍₇₎ = 3.515, p = 0.0098. n = 4 water, 5 withdrawal. *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 3. Prosapip1 Affects Actin Dynamics and Alcohol Drinking Promotes F-actin Assembly via Prosapip1

(A and B) F- and G-actin content in N2A cells after Prosapip1 transfection without (A, hatched green) and with (B, hatched purple) co-transfection with a shRNA sequence targeting Prosapip1 (shProsapip1), compared to cell expressing the empty plasmid (CTL) (A, green) or a scrambled sequence (SCR, purple) (B). F- and G-actin contents were determined by western blot analysis, and quantification was conducted as in Figure 2. Data are presented as the average ratio of F-actin or G-actin to total actin (F+G) ± SEM and expressed as the percentage of the corresponding control. Significance was determined using two-tailed unpaired t test. (A) F-actin t₍₁₀₎ = 4.255, p = 0.0017; G-actin t₍₁₀₎ = 40.255, p = 0.0017. n = 6. (B) F-actin t₍₆₎ = 7.714, p = 0.0002; G-actin t₍₆₎ = 4.22, p = 0.0056. n = 4. (C and D) F- and G-actin content after bilateral infection of NAc neurons with a lentivirus expressing Prosapip1 (Prosapip1, hatched green) or an empty GFP plasmid (CTL, green) (1 × 10⁸ pg/mL) (C), or after infection with Ltv-shProsapip1 (shProsapip1, hatched purple) or Ltv-SCR (SCR, purple) (1 × 10⁸ pg/mL). (E) F- and G-actin contents were determined in the NAc of mice after binge drinking of alcohol and in the water control group. Data are presented as the average ratio of F-actin or G-actin to total actin (F + G) ± SEM and expressed as the percentage of the corresponding control. Significance was determined using two-tailed unpaired t test. (C) F-actin t₍₇₎ = 3.52, p = 0.0097; G-actin t₍₇₎ = 3.52, p = 0.0097, n = 4 Ltv-Prosapip1, 5 Ltv-CTL. (D) F-actin t₍₈₎ = 3.389, p = 0.0095; G-actin t₍₈₎ = 3.389, p = 0.0095, n = 5 per group. (E) F-actin t₍₁₀₎ = 4.078, p = 0.0022; G-actin t₍₁₀₎ = 5.098, p = 0.0005, n = 6 per group. (F) The NAc of mice were infused with Ltv-SCR or Ltv-shProsapip1 (1 × 10⁸ pg/mL). Three weeks after surgery, mice underwent 2 weeks of CIE exposure or air exposure as control (CTL) (Figure S5A) and F- and G-actin content were determined as described above. Two-way ANOVA showed a significant main effect of CIE (F-actin: F_(1,12) = 16.83, p < 0.01) and shProsapip1 (F-actin: F_(1,12) = 18.18, p < 0.01;) but no significant interaction (F-actin: F_(1,12) = 0.04, p = 0.85). Post hoc Student-Newman-Keuls test detected a significant difference between air and CIE within the SCR group (CTL/SCR versus CIE/SCR, p < 0.05), a significant difference between SCR and shProsapip1 within the air group (CTL/SCR versus CTL/shProsapip1, p < 0.05) and within the CIE group (CIE/SCR versus CIE/shProsapip1, p < 0.05). *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 4. Prosapip1 in the NAc Contributes to Dendritic Spine Morphology and Alterations of Spine Structure by Alcohol Are Mediated by Prosapip1

(A) Mice underwent 4 weeks of IA20%-2BC. Mice consuming water only were used as control. Ltv-shProsapip1 or Ltv-SCR at low titer (1 × 10⁵ pg/mL) was infused into the NAc and after 1 week of recovery, mice were subjected to 3 more weeks of IA20%-2BC. Infection of the NAc with a low titer of Ltv-shProsapip1 did not affect alcohol intake (Ltv-SCR: 15.1 ± 1.57 g/kg/24 hr, Ltv-shProsapip1: 14.5 ± 0.97 g/kg/24 hr). Four hours after the beginning of the last drinking session, mice were transcardially perfused, and MSN morphology was analyzed in NAc shell. (B) Representative ×100 confocal z stack images of dendritic segments bearing spines for all four experimental conditions. Scale bar, 5 μm. (C) Average spine area of NAc shell MSNs of mice (Water/SCR [blue], Water/shProsapip1 [hatched blue], Alcohol/SCR [red], Alcohol/shProsapip1 [hatched red]). Data are presented as average spine area ± SEM and are expressed in μm². Two-way ANOVA showed a significant main effect of alcohol (F_(1,23) = 33.16, p < 0.001) and shProsapip1 (F_(1,23) = 31.19, p < 0.001) but no interaction (F_(1,23) = 1.052, p = 0.32). Post hoc Student-Newman-Keuls test detected a significant difference between water and alcohol in the SCR group (Water/SCR versus Alcohol/SCR, p < 0.001), a significant difference between SCR and shProsapip1 within the water group (Water/SCR versus Water/shProsapip1, p < 0.01), and within the alcohol group (Alcohol/SCR versus Alcohol/shProsapip1, p < 0.001), but no difference between Water/SCR and Alcohol/shProsapip1 (p = 0.903). (D) Percentage of filopodia, thin, stubby, and mushroom-type spines of NAc shell MSNs. Data are presented as the average percentage of each type of spines across the four conditions ± SEM. Filopodia-type spines, two-way ANOVA showed a significant main effect of shProsapip1 (F_(1,23) = 30.14, p < 0.001) but no main effect of alcohol (F_(1,23) = 0.67, p = 0.42) and no interaction (F_(1,23) = 0.09, p = 0.76); thin spines, two-way ANOVA showed a significant main effect of alcohol (F_(1,23) = 47.66, p < 0.001) and shProsapip1 (F_(1,23) = 104.6, p < 0.001) but no interaction (F_(1,23) = 0.32, p = 0.58). Post hoc Student-Newman-Keuls test detected a significant difference between water and alcohol within the SCR group (Water/SCR versus Alcohol/SCR, p < 0.001), between SCR and shProsapip1 within the water group (Water/SCR versus Water/shProsapip1, p < 0.001) and within the alcohol group (Alcohol/SCR versus Alcohol/shProsapip1, p < 0.001), and a significant difference between Water/SCR and Alcohol/shProsapip1 (p < 0.05); stubby spines, two-way ANOVA showed no significant main effect of alcohol (F_(1,23) = 0.38, p = 0.55) or shProsapip1 (F_(1,23) = 0.63, p = 0.43) and no interaction (F_(1,23) = 0.30, p = 0.58); mushroom spines, two-way ANOVA showed a significant main effect of alcohol (F_(1,23) = 36.71, p < 0.001) and shProsapip1 (F_(1,23) = 103, p < 0.001) but no interaction (F_(1,23) = 2.15, p = 0.15). Post hoc Student-Newman-Keuls test detected a significant difference between water and alcohol within the SCR group (Water/SCR versus Alcohol/SCR, p < 0.001), between SCR and shProsapip1 within the water group (Water/SCR versus Water/shProsapip1, p < 0.001) and within the alcohol group (Alcohol/SCR versus Alcohol/shProsapip1, p < 0.001), and a significant difference between Water/SCR and Alcohol/shProsapip1 (p < 0.01). n = 6–9 neurons, n = 4 mice per group. *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 5. Prosapip1 Is Required for Alcohol-Dependent Change in AMPAR Subunit Composition at Glutamate Synapses

(A and B) The NAc of mice were infected bilaterally with Ltv-shProsapip1 or Ltv-SCR (1 × 10⁸ pg/mL). Twenty-one days after virus administration, NAc slices were prepared and infected MSNs were patched. Electrically evoked AMPAR-mediated EPSCs were recorded from MSNs in the NAc shell at +40 mV and 70 mV from each group and rectification index was calculated. (A) The averages ± SEM of the rectification index are plotted for each group with symbols showing the values from individual cells. shProsapip1 (purple) 0.93 ± 0.05, SCR (black) 1.08 ± 0.05; significance difference was determined using two-tailed unpaired t test. t₍₂₉₎ = 2.194, p = 0.036, n = 16 cells/4 mice (shProsapip1) and 15 cells/4 mice (shSCR). *p < 0.05. (B) Representative traces of AMPAR-mediated EPSC recorded at +40 mV and at 70 mV. The dotted lines show the expected amplitude for non-rectifying AMPAR-mediated currents. (C and D) Mice underwent at least 8 weeks of IA20%-2BC. Mice consuming water only were used as control. Electrically evoked AMPAR-mediated EPSCs were recorded from MSNs in the NAc shell at +40 mV and 70 mV. (C) The averages ± SEM of the rectification index were plotted for each group with symbols showing the values from individual cells. Water (blue) 1.12 ± 0.04, alcohol (red) 1.28 ± 0.04; significance was determined using two-tailed un paired t test. t₍₅₉₎ = 2.693, p = 0.009, n = 28 cells/6 mice (water) and n = 33 cells/6 mice (alcohol). (D) Representative traces of AMPAR-mediated EPSCs recorded at +40 mV and 70 mV of MSNs from water (blue) or alcohol (red) group. Dotted lines depict the expected amplitude for non-rectifying AMPAR-mediated currents. (E and F) The NAc of mice were infected bilaterally with Ltv-SCR. 21 days after virus administration, mice underwent 2 weeks of CIE exposure to alcohol vapor or air. Electrically evoked AMPAR-mediated EPSCs were recorded from GFP-positive MSNs expressing Ltv-SCR. (E) The averages ± SEM of the rectification index are plotted for each group with symbols showing the values from individual cells. Air (blue) 1.17 ± 0.06, CIE (red) 1.43 ± 0.09; Significance difference was determined using two-tailed unpaired t test. t₍₃₅₎ = 2.349, p = 0.025, n = 19 cells/4 mice (air) and 18 cells/4 mice (CIE). (F) Representative traces of AMPAR-mediated EPSC recorded at +40 mV and at 70 mV. Dotted lines depict the expected amplitude for non-rectifying AMPAR-mediated currents. (G and H) The NAc of mice were infected bilaterally with Ltv-SCR or Ltv-shProsapip1. Twenty-one days after virus administration, mice underwent 2 weeks of CIE exposure to alcohol. Electrically evoked AMPAR-mediated EPSCs were recorded from GFP-negative (uninfected) or -positive (infected) MSNs from each group. (G) The averages ± SEM of the rectification index are plotted for each group with symbols showing the values from individual cells. Ltv-SCR uninfected (red, empty) 1.327 ± 0.058, Ltv-SCR infected (red, filled) 1.432 ± 0.093, Ltv-shProsapip1 uninfected (gray, empty) 1.449 ± 0.086, Ltv-shProsapip1 (black, filled) 1.124 ± 0.058; one-way ANOVA indicates significant difference among means. F_{(3, 55)} = 3.991, p = 0.012. Post hoc multiple comparison indicates significant difference between Ltv-SCR infected versus Ltv-shProsapip1 infected, p = 0.019; Ltv-shProsapip1 uninfected versus Ltv-shProsapip1 infected, p = 0.028. n = 12 cells (GFP), 18 cells (GFP+)/4 mice (SCR) and 12 cells (GFP), 16 cells (GFP+)/4 mice (shProsapip1). (H) Representative traces of AMPAR-mediated EPSC recorded at +40 mV and at 70 mV. Dotted lines depict the expected amplitude for non-rectifying AMPAR-mediated currents. *p < 0.05, **p < 0.01.

Figure 6. Prosapip1 in the NAc Drives Alcohol Self-Administration and Seeking

(A) Mice that underwent IA20%-2BC for 7–8 weeks were trained to self-administer 20% alcohol in operant chambers. After reaching a stable baseline of responding (see Figure S7), mice received an infusion of Ltv-SCR or Ltv-shProsapip1 in the NAc. Following 2 weeks of recovery, self-administration of alcohol was resumed for an additional 2 weeks. (B) Number of active and inactive lever presses on the last self-administration session. shProsapip1 reduces the number of active (t₍₁₈₎ = 2.43, p = 0.026) but not inactive (t₍₁₈₎ = 0.73, p = 0.47) lever presses. (C) Cumulative number of active lever presses. Two-way ANOVA indicates a significant main effect of shProsapip1 (F_(1,18) = 5.67, p < 0.05) on cumulative number of active lever presses. (D) Frequency of active lever presses. shProsapip1 reduces the frequency of active lever presses (t₍₁₈₎ = 2.25, p = 0.038). (E) Number of port entries. shProsapip1 reduces the number of port entries (t₍₁₈₎ = 2.32, p = 0.033). (F) Cumulative number of port entries. Two-way ANOVA indicates a significant main effect of shProsapip1 on cumulative number of port entries (F_(1,18) = 4.04, p = 0.05). (G) Frequency of port entries. shProsapip1 reduces the frequency of port entries (t₍₁₈₎ = 2.10, p = 0.05). (H) Alcohol intake (g/kg/2 hr). shProsapip1 reduces the amount of alcohol consumed (t₍₁₈₎ = 2.45, p = 0.025). (I) Cumulative number of alcohol deliveries. Two-way ANOVA indicates a significant main effect of shProsapip1 on cumulative number of alcohol deliveries (F_(1,18) = 8.79, p < 0.05). (J) Frequency of alcohol delivery. shProsapip1 reduces the frequency of alcohol delivery (t₍₁₈₎ = 2.67, p = 0.016). (K) Latency of first active lever press. shProsapip1 increases the latency of first active lever press (t₍₁₈₎ = 2.24, p = 0.038). (L) Latency of the first port entry. shProsapip1 does not affect the latency of first port entry (t₍₁₈₎ = 0.25, p = 0.81). (M) Latency of the first alcohol delivery (min). shProsapip1 increases the latency of first alcohol delivery (t₍₁₈₎ = 2.02, p = 0.05). (N–P) An independent cohort of mice was trained to self-administer 1% sucrose (see Figure S7) and received an infusion of Ltv-SCR or Ltv-shProsapip1 in the NAc. (N) Number of active and inactive lever presses on the last self-administration session. shProsapip1 did not affect the number of active (t₍₂₆₎ = 0.09, p = 0.92) or inactive (t₍₂₆₎ = 0.04, p = 0.97) lever presses. (O) Number of port entries. shProsapip1 did not affect the number of port entries (t₍₂₆₎ = 0.03, p = 0.97). (P) Sucrose intake (mL/kg/2 hrs). shProsapip1 did not affect the amount of sucrose self-administered (t₍₂₆₎ = 0.24, p = 0.81). Data are represented as the average ± SEM (B–M) n = 10 per group and (N–P) n = 14 per group. *p < 0.05 versus SCR group.

Figure 7. Prosapip1 in the NAc Contributes to Alcohol Reward

(A) Four weeks after intra-NAc administration of Ltv-shProsapip1 or SCR, mice underwent a CPP experiment in which they were daily administered (intra-peritoneum, i.p.) with alcohol (1.8 g/kg) or saline prior to the confinement in the drug- or non-drug-paired compartment for 5 min. Following conditioning (8 days), a 15 min post-conditioning test was conducted. (B) Distance traveled (cm) in the CPP apparatus prior to the conditioning phase. shProsapip1 does not alter ambulatory activity (t₍₃₁₎ = 0.01, p = 0.99). n = 15–18 per group. (C) CPP score is expressed as the time (s) spent in the drug-paired compartment during the post-conditioning minus the time spent in the same compartment during the pre-conditioning period. Two-way ANOVA showed a main effect of conditioning (saline or alcohol; F_(1,31) = 16.32, p < 0.001) and no interaction between conditioning and virus treatment (F_(1,31) = 1.51, p = 0.23). Post hoc Student-Newman Keuls testing detected a significant difference between saline and alcohol in the scramble group (Saline/SCR versus Alcohol/SCR, p < 0.01) and a significant difference between animals who received shProsapip1 or scramble in the alcohol group (Alcohol/SCR versus Alcohol/shProsapip1, p < 0.01). n=7–10 per group. (D) Four weeks after intra-NAc administration of Ltv-shProsapip1 or Ltv-SCR, mice underwent a CPA experiment in which they were administered daily (s.c.) with lithium chloride (LiCl, 130 mg/kg) or saline prior to confinement in the drug- or non-drug paired compartment for 45 min. Following conditioning (3 days, two sessions per day), a 20 min post-conditioning test was conducted. (E) CPA score is expressed as the time (s) spent in the drug-paired compartment during post-conditioning minus the time spent in the same compartment during pre-conditioning. Two-way ANOVA showed a main effect of condi tioning (saline or LiCl; (F_{(1, 26)} = 15.17, p = 0.0006), no effect of virus treatment (F_(1,26) = 0.002, p = 0.963), and no interaction between virus and LiCl (F_(1,26) = 0.024, p = 0.877). Further analysis by the method of contrast reveals a significant difference for the LiCl/SCR group and LiCl/shProsapip1 versus Saline/SCR group (respectively, p < 0.05) but no significant difference between LiCl/SCR and LiCl/shProsapip1 group (p = 0.99). Data are represented as the average ± SEM (B and C) n = 7–10 per group and (E) n = 7–8 per group. *p < 0.05 and ***p < 0.001.