Neurofibromin regulates corticostriatal inhibitory networks during working memory performance

Abstract

Neurofibromatosis type I (NF1) is one of the most common single-gene causes of learning disabilities. Here, we use behavioral working memory probes and electrophysiological studies in a mouse model of NF1 (Nf1 heterozygous null mutants; Nf1(+/-)) to demonstrate that (i) Neurofibromin regulates prefrontal and striatal inhibitory networks, specifically activity-dependent GABA release and (ii) is required for working memory performance, with inhibition-dependent working memory deficits seen in Nf1(+/-) mice. We find that increased inhibition in medial prefrontal cortex (mPFC) is sufficient to alter persistent activity in a biophysical model of an mPFC microcircuit, suggesting a possible mechanism for Nf1(+/-) working memory deficits. Accordingly, working memory assays applied during functional MRI (fMRI) studies in human subjects with NF1 reveal hypoactivation of corticostriatal networks, which is associated with impaired working memory performance. Collectively, these integrative mouse and human studies reveal molecular and cellular mechanisms contributing to working memory deficits in NF1.

Fig. 1.

Whole-cell recordings in mPFC of Nf1^+/− mice demonstrate specific increases in inhibition. (A) Nf1^+/− cells (n = 13) show increased frequency of sIPSCs compared with WT (n = 16) across all amplitude bins (RM ANOVA P < 0.001; *P < 0.05; **P < 0.01, Tukey’s post hoc) and in overall sIPSC frequency (Inset *P < 0.05, t test). Representative traces of sIPSC recordings shown above. (B) No difference in frequency of sEPSCs is seen between Nf1^+/− cells (n = 12) and WT (n = 13) across amplitude bins or overall (Inset). Representative traces of sEPSC recordings shown above. (C) Interneuronal Nf1 deletion in Dlx-Cre+;Nf1^flox/+ cells (n = 8) leads to increased frequency of sIPSCs compared with control groups: Dlx-Cre+;WT (n = 9), Dlx-Cre-;Nf1^flox/+ (n = 14), or Dlx-Cre-;WT (n = 11) (P = 0.006, RM ANOVA; **P < 0.01, Tukey’s post hoc). (D) In the presence of glutamate receptor antagonists CNQX/APV, Nf1^+/− cells (n = 10) show increased frequency of sIPSCs compared with WT (n = 8) across all amplitude bins (RM ANOVA P = 0.009; *P < 0.05, Tukey’s post hoc). Error bars represent SEM.

Fig. 2.

Neurofibromin modulation of the Ras signaling pathway selectively regulates sIPSC frequency in mPFC. (A Left) U0126, an inhibitor of Ras signaling, decreases overall sIPSC frequency in both Nf1^+/− (n = 18) and WT (n = 13) cells (RM ANOVA drug × sIPSC frequency interaction P = 0.003; *P < 0.05, Tukey’s post hoc). (Right) U0126 causes a larger within-cell decrease in sIPSC frequency in Nf1^+/− compared with WT cells (P = 0.045, one-tailed t test) (B Left) U0126 decreases overall sEPSC frequency across Nf1^+/− (n = 13) and WT (n = 8) cells (RM ANOVA drug × sEPSC frequency interaction P = 0.012; *P < 0.05, Tukey’s post hoc). (Right) Within-cell decrease in sEPSC frequency is not different between genotypes (P = 0.357, one-tailed t test). Error bars represent SEM.

Fig. 3.

Whole cell recordings of sIPSCs and sEPSCs onto cells in the striatum of Nf1^+/− mice demonstrate specific increases in inhibition. (A) Nf1^+/− MSNs (n = 11) show increased frequency of sIPSCs compared with WT (n = 10) across all amplitude bins (RM ANOVA P < 0.001; *P < 0.05; **P < 0.01, Tukey’s post hoc) and in overall sIPSC frequency (Inset *P < 0.05, t test). Representative traces of sIPSC recordings shown above. (B) No difference in frequency of sEPSCs is seen between Nf1^+/− MSNs (n = 10) and WT (n = 9) across amplitude bins or overall (Inset). Representative traces of sEPSC recordings shown above. Error bars represent SEM.

Fig. 4.

Working memory impairments caused by increased inhibition in the Nf1^+/− mice. (A–C) Delayed win-shift radial arm maze errors, summed across 10 d of task performance, are compared between Nf1^+/− (n = 12) and WT (n = 12) mice for training phase errors (P = 0.883, t test), testing phase within-phase errors (P = 0.047, t test), and testing phase across-phase errors (P = 0.403, t test). (D) Nf1^+/− mice make more testing phase within-phase errors during the first 6 d of the task (RM ANOVA genotype × day × error; P = 0.024, *P < 0.05, Tukey’s post hoc). (E) Picrotoxin improves within-phase errors (P = 0.009) in Nf1^+/− mice (RM ANOVA drug × genotype × day × error). Comparisons made across four groups: saline (n = 20) and picrotoxin (n = 18) treated Nf1^+/− mice and saline (n = 20) and picrotoxin treated WT mice (n = 19). (F) In the operant delayed nonmatch to sample task, Nf1^+/− mice (n = 12) show impaired accuracy (RM ANOVA genotype × delay × %correct P = 0.043) compared with WT (n = 10). (G) Picrotoxin improves accuracy in Nf1^+/− mice (n = 11, RM ANOVA drug × genotype × delay × %correct P = 0.037) compared to WT (n = 9). (Left) Accuracy of Nf1^+/− and WT mice administered saline or picrotoxin (0.025 mg/kg) in a counterbalanced within-subject design. (Right) Within-subject percent change in accuracy at the 3–6 s delay (*P < 0.05; one-tailed t test). Error bars represent SEM.

Fig. 5.

Hypoactivation of corticostriatal structures and working memory deficits in NF1 individuals. (A) Individuals with NF1 (n = 14) show decreased testing phase accuracy in both working memory maintenance and manipulation conditions compared with age matched controls (n = 12; RM ANOVA group × %correct P = 0.019). (B) Individuals with NF1 show decreased testing phase accuracy in the spatial capacity task across loads (RM ANOVA group × %correct P = 0.03; *P < 0.05; **P < 0.01, Tukey’s post hoc). (C) Center: Regions of interest in which activation was measured during performance of both the SCAP and StMNM. Individual graphs: BOLD signal changes during each task (*P < 0.05; **P < 0.01, RM ANOVA group × region). Error bars represent SEM.

Fig. 6.

Degree of DLPFC hypoactivation predicts degree of performance impairment in working memory maintenance tasks in individuals with NF1. Significant association between right DLPFC activation and task performance during (A) the maintenance condition of the MNM task (F_1,11 = 8.69, P = 0.013; r² = 0.32) and (B) the SCAP task (F_1,11 = 50.86, P < 0.001; r² = 0.54).