Schizophrenia risk gene CAV1 is both pro-psychotic and required for atypical antipsychotic drug actions in vivo

Abstract

Caveolin-1 (Cav-1) is a scaffolding protein important for regulating receptor signaling cascades by partitioning signaling molecules into membrane microdomains. Disruption of the CAV1 gene has recently been identified as a rare structural variant associated with schizophrenia. Although Cav-1 knockout (KO) mice displayed no baseline behavioral disruptions, Cav-1 KO mice, similar to schizophrenic individuals, exhibited increased sensitivity to the psychotomimetic N-methyl-D-aspartate receptor antagonist phencyclidine (PCP). Thus, PCP disruption of prepulse inhibition (PPI) and PCP-induced mouse locomotor activity were both enhanced by genetic deletion of Cav-1. Interestingly, genetic deletion of Cav-1 rendered the atypical antipsychotics clozapine and olanzapine and the 5-HT(2A)-selective antagonist M100907 ineffective at normalizing PCP-induced disruption of PPI. We also discovered that genetic deletion of Cav-1 attenuated 5-HT(2A)-induced c-Fos and egr-1 expression in mouse frontal cortex and also reduced 5-HT(2A)-mediated Ca(2+) mobilization in primary cortical neuronal cultures. The behavioral effects of the 5-HT(2A) agonist (2,5-dimethoxy-4-iodoamphetamine) including head twitch responses and disruption of PPI were also attenuated by genetic deletion of Cav-1, indicating that Cav-1 is required for both inverse agonist (that is, atypical antipsychotic drug) and agonist actions at 5-HT(2A) receptors. This study demonstrates that disruption of the CAV1 gene--a rare structural variant associated with schizophrenia--is not only pro-psychotic but also attenuates atypical antipsychotic drug actions.

Figure 1

Genetic deletion of caveolin-1 (Cav-1) sensitizes mice to the behavioral effects of the psychotomimetic phencyclidine (PCP). (a) Cav-1 knockout (KO) mice exhibit increased sensitivity to PCP disruption of sensory motor gaiting. Littermate wild-type (WT) and Cav-1 KO mice were injected with saline, 4 or 6 mg kg⁻¹ PCP and immediately placed into acoustic startle response chambers. In all, 6 mg kg⁻¹ PCP significantly decreased (disrupted) prepulse inhibition (PPI) at all prepulse levels in WT and Cav-1 KO mice. However, a lower dose of 4 mg kg⁻¹ PCP significantly disrupted PPI in only Cav-1 KO but not WT mice (mean±s.e.m., n=8 to 12 littermate pairs; ^*P<0.05 versus WT saline or Cav-1 KO saline at each prepulse level). (b, c) PCP-induced hyperlocomotion is increased in Cav-1 KO mice. Littermate WT and Cav-1 KO mice were placed into open field locomotion chambers and allowed to acclimate; 30 min later mice were injected with PCP and the total distance traveled (cm) was determined. Cav-1 KO mice exhibited a greater sensitivity to 4 mg kg⁻¹ PCP than WT mice (panel b, mean±s.e.m., n=9 littermate pairs; ^*P<0.05 versus WT mice at time indicated). Cav-1 KO mice also displayed increased total distance traveled in response to 4 and 6 mg kg⁻¹ PCP compared with WT mice (panel c, mean±s.e.m., n=8 to 12 littermate pairs; ^*P<0.05 versus WT mice in each group). (d, e) Stereotypy movement is increased in Cav-1 KO mice in response to 4 mg kg⁻¹ PCP (panel d, mean±s.e.m., n=9 littermate pairs). Cav-1 KO mice displayed increased total stereotypy counts in response to 4 and 6 mg kg⁻¹ PCP compared with WT mice (panel e, mean±s.e.m., n=8 to 12 littermate pairs; ^*P<0.05 versus WT mice in each group).

Figure 2

Caveolin-1 (Cav-1) knockout (KO) attenuates atypical antipsychotic and 5-HT_2A antagonist drug actions in the phencyclidine (PCP) mouse model of psychosis. (a, b) Cav-1 KO renders clozapine and M100907 less effective at normalizing disrupted sensory motor gaiting. Littermate wild-type (WT) and Cav-1 KO mice received injections of vehicle, vehicle +6 mg kg⁻¹ PCP, clozapine (0.5 or 1.0 mg kg⁻¹) +PCP or M100907 (0.5 or 1.0 mg kg⁻¹) + PCP followed by measurement of acoustic startle responses. 6 mg kg⁻¹ PCP significantly decreased (disrupted) prepulse inhibition (PPI) in both genotypes at all prepulse levels. Clozapine normalized the disrupted PPI to near vehicle levels in WT mice; however, clozapine was ineffective at normalizing PCP disrupted PPI in Cav-1 KO mice at 4 and 8 dB prepulse levels (panel a, mean±s.e.m., n=12 littermate pairs; ^*P<0.05 versus vehicle-treated WT or vehicle-treated Cav-1 KO mice). Similarly, M100907 normalized the disrupted PPI to near vehicle levels in WT mice but this was prevented in Cav-1 KO mice at 4 and 8 dB prepulse levels (panel b, mean±s.e.m., n=10 littermate pairs, ^*P<0.05 versus vehicle-treated WT or vehicle-treated Cav-1 KO mice). (c–f) Clozapine and M100907, but not haloperidol, are less effective at normalizing PCP-induced hyperlocomotion in Cav-1 KO mice. Clozapine treatments significantly decreased the total distance traveled after PCP in WT but not Cav-1 KO mice; however, haloperidol was equally effective in both genotypes (panels c and d, mean±s.e.m., n=8 littermate pairs; ^*P<0.05 versus vehicle WT mice or vehicle Cav-1 KO mice). M100907 was also less effective at normalizing PCP-induced hyperlocomotion in Cav-1 KO mice compared with the drug activity in WT animals (panels e and f; mean±s.e.m., n=8 littermate pairs; ^*P<0.05 versus WT 0.1 M100 + PCP; ^#P<0.05 versus WT M100 + PCP).

Figure 3

Mouse behavioral responses to the 5-HT_2A agonist 2,5-dimethoxy-4-iodoamphetamine (DOI) are attenuated by genetic deletion of caveolin-1 (Cav-1). (a) Mouse head twitch responses to the 5-HT_2A agonist DOI are decreased by Cav-1 knockout (KO). Littermate wild-type (WT) and Cav-1 KO mice were injected with 1.25 mg kg⁻¹ DOI and head twitches scored over a 35-min period; DOI-induced head twitches were attenuated in Cav-1 KO mice at all-time points (panel a, mean±s.e.m., n=12 littermate pairs; ^*P<0.05 versus WT mice). (b) Dose responses of DOI on total mouse head twitches over a 35-min period are shown; head twitches were attenuated in Cav-1 KO mice only at higher concentrations (panel b, mean±s.e.m., n=6 to 8 littermate pairs; ^*P<0.05 versus WT mice at each concentration). (c) DOI disruption of sensory motor gaiting in mice is prevented by Cav-1 KO. Littermate WT and Cav-1 KO mice were injected with saline, 1 or 5 mg kg⁻¹ DOI and 10 min later placed into acoustic startle response chambers to assess prepulse inhibition (PPI). DOI significantly decreased (disrupted) PPI at 4 and 8 dB prepulse levels in WT but not Cav-1 KO mice (panel c, mean±s.e.m., n=8 littermate pairs; ^*P<0.05 versus WT saline-treated mice at each prepulse level). NS, nonsignificant.

Figure 4

5-HT_2A signaling responses to the agonist 2,5-dimethoxy-4-iodoamphetamine (DOI) are attenuated in vivo and in vitro by genetic deletion of caveolin-1. (a, b) Littermate wild-type (WT) and caveolin-1 (Cav-1) knockout (KO) mice were injected with 1.0 mg kg⁻¹ DOI, 1 h later frontal cortex was isolated, mRNA extracted and transcript levels of c-Fos and egr-1 determined by real-time PCR. DOI-induced c-Fos and egr-1 gene expression in frontal cortex was decreased in mice heterozygous for Cav-1; DOI-induced gene expression was prevented in KO mice (panels a and b, mean±s.e.m., n=5 littermate mice for each group; ^*P<0.05 versus saline-treated mice in each group). (c) Neuronal 5-HT_2A receptor calcium signaling is decreased by genetic deletion of Cav-1. Primary cortical neurons were isolated from WT and Cav-1 KO mice and cultured in 96-well plates. Neurons were infected with lentiviruses encoding 5-HT_2A receptors and intracellular calcium release was determined by fluorescence calcium dye imaging using high-content imaging microscopy. Images of cells are shown before (left panels) and after (middle panels) an automated addition of the 5-HT_2A agonist DOI. WT neurons expressing 5-HT_2A receptors (middle panels) exhibited a robust increase in intracellular calcium dye fluorescence after DOI. In comparison, KO neurons expressing 5-HT_2A receptors (lower panels) displayed a reduced intracellular calcium dye fluorescence after agonist. The calcium dye fluorescence in individual responding cells in the wells (segmenting) was determined. (d) Quantification of calcium dye fluorescence in individual cells is shown and expressed as fold over initial fluorescence; KO neurons exhibited a reduced peak and total calcium response (Figure 4d, mean fluorescence intensity±s.e.m. from >90 cells per group from n=3 cell preparations).