mGlu1 tonically regulates levels of calcium-permeable AMPA receptors in cultured nucleus accumbens neurons through retinoic acid signaling and protein translation

Abstract

In several brain regions, ongoing metabotropic glutamate receptor 1 (mGlu1) transmission has been shown to tonically suppress synaptic levels of Ca²⁺ -permeable AMPA receptors (CP-AMPARs) while pharmacological activation of mGlu1 removes CP-AMPARs from these synapses. Consistent with this, we previously showed in nucleus accumbens (NAc) medium spiny neurons (MSNs) that reduced mGlu1 tone enables and mGlu1 positive allosteric modulation reverses the elevation of CP-AMPAR levels in the NAc that underlies enhanced cocaine craving in the "incubation of craving" rat model of addiction. To better understand mGlu1/CP-AMPAR interactions, we used a NAc/prefrontal cortex co-culture system in which NAc MSNs express high CP-AMPAR levels, providing an in vitro model for NAc MSNs after the incubation of cocaine craving. The non-specific group I orthosteric agonist dihydroxyphenylglycine (10 min) decreased cell surface GluA1 but not GluA2, indicating CP-AMPAR internalization. This was prevented by mGlu1 (LY367385) or mGlu5 (MTEP) blockade. However, a selective role for mGlu1 emerged in studies of long-term antagonist treatment. Thus, LY367385 (24 hr) increased surface GluA1 without affecting GluA2, whereas MTEP (24 hr) had no effect. In hippocampal neurons, scaling up of CP-AMPARs can occur through a mechanism requiring retinoic acid (RA) signaling and new GluA1 synthesis. Consistent with this, the LY367385-induced increase in surface GluA1 was blocked by anisomycin (translation inhibitor) or 4-(diethylamino)-benzaldehyde (RA synthesis inhibitor). Thus, mGlu1 transmission tonically suppresses cell surface CP-AMPAR levels, and decreasing mGlu1 tone increases surface CP-AMPARs via RA signaling and protein translation. These results identify a novel mechanism for homeostatic plasticity in NAc MSNs.

Keywords: GluA1; group I metabotropic glutamate receptors; homeostatic plasticity; primary culture; receptor trafficking.

Figure 1.

Acute DHPG application reduces surface GluA1 levels. (A) Representative phase contrast image of NAc/PFC co-cultures. The blue rectangle shows the fixed length of process (15 μm) on a medium spiny neuron (MSN) used to analyze surface GluA1 and GluA2 levels. Scale bar = 20 μm. (B) Enlarged representative fluorescent images of surface GluA1 (red) and GluA2 (green) on MSNs in NAc/PFC co-cultures under control conditions (CON; media) or after treatment with DHPG (50 μM, 10 or 20 min). Scale bar = 5 μm. (C) Left: Quantification of surface GluA1 (Top) and GluA2 (Bottom) levels following treatments described in (A). Acute application of DHPG for 10 minutes significantly decreased surface GluA1 but not GluA2, indicating internalization of CP-AMPARs. After 20 minutes of DHPG treatment, the decrease in GluA1 surface expression was not as pronounced and there was a trend towards increased GluA2 surface expression relative to the 10 min time-point, suggesting insertion of CI-AMPARs containing GluA1 and GluA2. Data are shown as % control (mean ± SEM). *p<0.05 versus control. Right: Scatter plots for GluA1 (top) and GluA2 (bottom) showing individual data points for all cells in each treatment group. Data are shown as % control and lines represent each group mean. n=21–24 cells/group.

Figure 2.

The DHPG-induced reduction in surface GluA1 is mediated by mGlu1 and mGlu5 and mimicked by mGlu1 or mGlu5 PAMs. (A) Left: Quantification of surface GluA1 (Top left) and GluA2 (Bottom left) levels following a 10 minute treatment with media (control, CON), the group I mGluR agonist DHPG (50 μM), DHPG plus the mGlu1 antagonist LY367385 (LY, 100 μM), or DHPG plus the mGlu5 antagonist MTEP (1 μM) (antagonists added 5 min prior to DHPG). Blocking either mGlu5 or mGlu1 prevented the acute DHPG-induced decrease in GluA1 surface expression while having no effect on GluA2 surface expression. Data are shown as % control (mean ± SEM). *p<0.05 versus control. Right: Scatter plots for GluA1 (top) and GluA2 (bottom) showing individual data points for all cells in each treatment group. Data are shown as % control and lines represent each group mean. n=20–25 cells/group. (B) Left: Quantification of surface GluA1 (Top left) and GluA2 (Bottom left) levels following a 10 minute treatment with media (control), the group I mGluR agonist DHPG (50 μM), the mGlu1 positive allosteric modulator (PAM) Ro67–7476 (Ro67, 3 μM), or the mGlu5 PAM CDPPB (10 μM). Consistent with the antagonist study, both mGlu1 and mGlu5 PAMs produced a decrease in GluA1 surface expression similar to that observed following DHPG application, with no change in GluA2 levels, indicating internalization of CP-AMPARs. Data are shown as % control (mean ± SEM). *p<0.05 versus control. Right: Scatter plots for GluA1 (top) and GluA2 (bottom) showing individual data points for all cells in each treatment group. Data are shown as % control and lines represent each group mean. n=20–22 cells/group.

Figure 3.

Long-term reduction in mGlu1 transmission, but not mGlu5 transmission, leads to upregulation of surface GluA1. Left: Quantification of surface GluA1 (Top) and GluA2 (Bottom) levels following long-term treatment (24–48 h) with media (control, CON), the mGlu1 antagonist LY367385 (LY, 100 μM) or the mGlu5 antagonist MTEP (1 μM). Both 24 and 48 h incubation with the mGlu1 antagonist LY367385 (but not the mGlu5 antagonist MTEP) increased GluA1 (left) but not GluA2 (right) surface levels in cultured NAc neurons, suggesting that long-term inhibition of mGlu1 transmission leads to scaling up of CP-AMPARs. Data are expressed as % control (mean ± SEM). *p<0.05 versus control. Right: Scatter plots for GluA1 (top) and GluA2 (bottom) showing individual data points for all cells in each treatment group. Data are shown as % control and lines represent each group mean. n=17–24 cells/group.

Figure 4.

Upregulation of surface GluA1 after long-term blockade of mGlu1 requires protein translation and retinoic acid signaling. Cultures were treated for 24 h with the mGlu1 antagonist LY367385 (LY, 100 μM) in the presence of the protein synthesis inhibitor anisomycin (Aniso, 40μM) (A) or the inhibitor of retinoic acid synthesis DEAB (10 μM) (B). Controls (CON) were treated with media. Inhibiting either protein synthesis or the production of retinoic acid completely blocked the LY-induced increase in surface GluA1 expression. Left: Data are expressed as % control (mean ± SEM). *p<0.05 versus control. Right: Scatter plots for GluA1 and GluA2 showing individual data points for all cells in each treatment group. Data are shown as % control and lines represent each group mean. n=17–20 cells/group (A) and n=17–25 cells/group (B).