Cortical GluN2B deletion attenuates punished suppression of food reward-seeking

Abstract

Rationale: Compulsive behavior, which is a hallmark of psychiatric disorders such as addiction and obsessive-compulsive disorder, engages corticostriatal circuits. Previous studies indicate a role for corticostriatal N-methyl-D-aspartate receptors (NMDARs) in mediating compulsive-like responding for drugs of abuse, but the specific receptor subunits controlling reward-seeking in the face of punishment remain unclear.

Objectives: The current study assessed the involvement of corticostriatal GluN2B-containing NMDARs in measures of persistent and punished food reward-seeking.

Methods: Mice with genetic deletion of GluN2B in one of three distinct neuronal populations, cortical principal neurons, forebrain interneurons, or striatal medium spiny neurons, were tested for (1) sustained food reward-seeking when reward was absent, (2) reward-seeking under a progressive ratio schedule of reinforcement, and (3) persistent reward-seeking after a footshock punishment.

Results: Mutant mice with genetic deletion of GluN2B in cortical principal neurons demonstrated attenuated suppression of reward-seeking during punishment. These mice performed normally on other behavioral measures, including an assay for pain sensitivity. Mutants with interneuronal or striatal GluN2B deletions were normal on all behavioral assays.

Conclusions: Current findings offer novel evidence that loss of GluN2B-containing NMDARs expressed on principal neurons in the cortex results in reduced punished food reward-seeking. These data support the involvement of GluN2B subunit in cortical circuits regulating cognitive flexibility in a variety of settings, with implications for understanding the basis of inflexible behavior in neuropsychiatric disorders including obsessive-compulsive disorders (OCD) and addictions.

Fig. 1

Schematic and cartoon of behavioral procedures. Mice were trained to lever-press for a 14-mg food pellet on a FR3 schedule of reinforcement then a tested for rewarded lever-pressing during “reward-available” and “reward-unavailable” periods during the final three sessions of training. b The next day, motivation for reward was assayed from lever-pressing breakpoints on a progressive ratio (PR) shedule. Responding on an FR3 schedule was reestablished on the next session. c The following day, punished suppression of reward seeking was tested with a 0.4-mA footshock

Fig. 2

Deletion of GluN2B on cortical principal neurons attenuated punished suppression of reward-seeking. a GluN2B was deleted in cortical principal neurons by crossing GluN2B-floxed mice with CaMKIIα-Cre mice. b Genotypes reached the training criterion in a similar number of sessions. c Both genotypes pressed the active lever more often during signaled reward-available than reward-unavailable phases. d Breakpoints attained on a progressive ratio (PR) schedule were similar between genotypes. e Punishment suppressed responding in both genotypes, but did so to a lesser extent in GluN2B^CxNULL than GluN2B^FLOX controls. **P<.01 versus available phases or unpunished baseline, #P<.05 GluN2B^CxNULL versus GluN2B^FLOX. n = 13–15 per genotype. Data are means ±SEM

Fig. 3

Deletion of GluN2B on forebrain interneurons did not affect measures of compulsive-like behavior a GluN2B was deleted in cortical interneurons by crossing GluN2B-floxed mice with Ppp1r2-Cre mice. b Genotypes reached the training criterion in a similar number of sessions. c Both genotypes pressed the active lever more often during signaled reward-available than reward-unavailable phases. d Breakpoints attained on a progressive ratio (PR) schedule were similar between genotypes. e Punishment suppressed responding in both genotypes. *P<.05, **P<.01 versus available phases or unpunished baseline. n=10–12 per genotype. Data are means ±SEM

Fig. 4

Deletion of GluN2B on striatal neurons did not affect measures of compulsive-like behavior. a GluN2B was deleted in striatal neurons by crossing GluN2B-floxed mice with Rgs9- Cre mice. b Genotypes reached the training criterion in a similar number of sessions. c Both genotypes pressed the active lever more often during signaled reward-available than reward-unavailable phases. d Breakpoints attained on a progressive ratio (PR) schedule were similar jbetween genotypes. e Punishment suppressed responding in both genotypes. **P<.01 versus available phases or unpunished baseline. n=9–11 per genotype. Data are means±SEM