Calcineurin mediates homeostatic synaptic plasticity by regulating retinoic acid synthesis

Homeostatic synaptic plasticity is a form of non-Hebbian plasticity that maintains stability of the network and fidelity for information processing in response to prolonged perturbation of network and synaptic activity. Prolonged blockade of synaptic activity decreases resting Ca²⁺ levels in neurons, thereby inducing retinoic acid (RA) synthesis and RA-dependent homeostatic synaptic plasticity; however, the signal transduction pathway that links reduced Ca²⁺-levels to RA synthesis remains unknown. Here we identify the Ca²⁺-dependent protein phosphatase calcineurin (CaN) as a key regulator for RA synthesis and homeostatic synaptic plasticity. Prolonged inhibition of CaN activity promotes RA synthesis in neurons, and leads to increased excitatory and decreased inhibitory synaptic transmission. These effects of CaN inhibitors on synaptic transmission are blocked by pharmacological inhibitors of RA synthesis or acute genetic deletion of the RA receptor RARα. Thus, CaN, acting upstream of RA, plays a critical role in gating RA signaling pathway in response to synaptic activity. Moreover, activity blockade-induced homeostatic synaptic plasticity is absent in CaN knockout neurons, demonstrating the essential role of CaN in RA-dependent homeostatic synaptic plasticity. Interestingly, in GluA1 S831A and S845A knockin mice, CaN inhibitor- and RA-induced regulation of synaptic transmission is intact, suggesting that phosphorylation of GluA1 C-terminal serine residues S831 and S845 is not required for CaN inhibitor- or RA-induced homeostatic synaptic plasticity. Thus, our study uncovers an unforeseen role of CaN in postsynaptic signaling, and defines CaN as the Ca²⁺-sensing signaling molecule that mediates RA-dependent homeostatic synaptic plasticity.Synaptic plasticity, a fundamental feature of the nervous system, is defined as modification of synaptic strength based on experience and activity history. Homeostatic synaptic plasticity is a type of compensatory mechanism activated during chronic elevation or reduction of network activity to modulate synaptic strength in the opposite direction (for example, reduced network activity leads to increased synaptic strength) (1, 2). Retinoic acid (RA) is a key signaling molecule in a form of homeostatic synaptic plasticity induced by reduced excitatory synaptic activity (3–5). Prolonged inhibition of excitatory synaptic transmission leads to a compensatory increase in synaptic excitation and a decrease in synaptic inhibition (4, 5). Both of these processes require RA synthesis. It also has been shown that dendritic Ca²⁺ levels directly govern the synthesis of RA; basal Ca²⁺ levels maintained by normal synaptic transmission are sufficient to suppress RA synthesis. Upon synaptic activity inhibition, reduced Ca²⁺ levels de-repress RA synthesis and activate RA-dependent homeostatic synaptic mechanisms (6). Thus, a Ca²⁺-dependent signaling molecule that is sensitive to changes in basal Ca²⁺ levels must be involved in synaptic RA signaling.Ca²⁺-dependent protein kinases and phosphatases are critical components of signaling pathways involved in synaptic plasticity (7, 8). For example, regulation of the phosphorylation of key serine residues in the C-terminal sequences of the α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) type glutamate receptor (AMPAR) subunit GluA1 by kinases (i.e., PKA, PKC, and CaMKII) and phosphatases [i.e., calcineurin (CaN) and PP2A] is thought to play major roles in governing AMPAR trafficking in and out of the synaptic membrane and to mediate the expression of well-established forms of synaptic plasticity, such as long-term potentiation (LTP) and long-term depression (LTD) (9–12). A recent study examining the stoichiometry of AMPAR phosphorylation revealed surprisingly low levels of phosphorylated GluA1 in neurons (13), however, suggesting that the phosphorylation of AMPARs might not be involved in homeostatic plasticity, and that other mechanisms may be in play.In the present study, we identified CaN as the Ca²⁺-dependent signaling molecule that regulates RA synthesis in neurons in an activity-dependent manner. Inhibition of CaN activity triggers RA synthesis, suggesting that basal CaN activity, supported by normal synaptic transmission, is sufficient to suppress RA synthesis in an active neural network. In CaN-deficient neurons, synaptic activity blockade-induced RA-dependent forms of homeostatic synaptic plasticity are absent. Similar to direct RA application, CaN inhibition enhances excitatory synaptic transmission and reduces inhibitory synaptic transmission. Blocking RA synthesis or genetic deletion of the RA receptor RARα prevents CaN inhibitor-induced regulation of synaptic strength, indicating that CaN acts upstream of RA. Importantly, neurons bearing GluA1 S831A or S845A knockin (KI) mutations, which eliminate phosphorylation at these two serine residues, respond normally to CaN inhibitors and RA treatment. Taken together, our results indicate that CaN participates in RA-dependent homeostatic synaptic plasticity through regulation of RA synthesis independent of the modulation of GluA1 phosphorylation.