Cdk5 is highly expressed in the hippocampus. Image reprinted from Nature Neuroscience.
Sometimes you have to lose to win. Most knockout mice show functional impairment, but some genetic losses cause functional gains. Hawasli et al. report improved spatial memory in mice lacking cyclin-dependent kinase 5 (Cdk5) in a recent article in Nature Neuroscience.
Cdk5 is a serine/threonine protein kinase that phosphorylates many neuronal proteins, including cytoskeletal, cell adhesion and membrane trafficking proteins. Mice lacking Cdk5 during development have structural deficits in the cortex, cerebellum and hippocampus and die perinatally.
The authors knocked down Cdk5 expression in adult mice. They crossed mice with Cdk5 flanked by loxP recombination sites with mice expressing hydroxytamoxifen-inducible Cre recombinase (Cre-ERT) under the control of the prion promoter. Control mice expressed floxed Cdk5 but not Cre-ERT, and both Cdk5 knockdown and control mice were treated with hydroxytamoxifen. Cdk5-deficient mice had less Cdk5 mRNA and protein in cortex, cerebellum and hippocampus. More than 90% of hippocampal pyramidal neurons lacked Cdk5 in Cdk5-deficient mice.
Cdk5-deficient mice showed improved hippocampal-dependent memory. Mice freeze with fear in response to a context previously associated with a shock. Cdk5-deficient mice showed more context-dependent freezing than did control mice. After repeated exposure to the context without shock, mice extinguish the shock memory and show reduced freezing behavior. Cdk5-deficient mice extinguished context-dependent fear conditioning faster than did control mice. Cdk5-deficient and control mice performed similarly in the Morris water maze, in which mice learn the location of a submerged platform in a pool of opaque water. When the platform location was changed, Cdk5-deficient mice adapted faster than did control mice. Because the water maze is an NMDA-receptor-dependent task, these data suggest changes in glutamatergic neurotransmission in Cdk5-deficient mice.
CA1 hippocampal neurons from Cdk5-deficient mice showed enhanced synaptic plasticity and NMDA neurotransmission. Tetanus induced long-term potentiation (LTP) in hippocampal slices from Cdk5-deficient and control mice. However, theta-burst stimulation that was too weak to induce LTP in CA1 neurons from control mice could induce LTP in CA1 neurons from Cdk5-deficient mice, suggesting that Cdk5-deficiency reduced the threshold for LTP in CA1 neurons. CA1 hippocampal neurons showed similar AMPA receptor-mediated excitatory postsynaptic currents (EPSCs) but enhanced NMDA receptor-mediated EPSCs in Cdk5-deficient relative to control mice.
NMDA receptors contain NR1 as well as NR2A and NR2B subunits. The NR2B-specific inhibitor ifenprodel blocked the increase in NMDA EPSCs and the enhanced LTP in Cdk5-deficient mice, suggesting that NR2B is responsible for enhanced hippocampal transmission in Cdk5-deficient mice. Cdk5-deficient mice showed increased hippocampal NR2B protein but not NR2B mRNA relative to control mice, suggesting that Cdk5 is important in the post-transcriptional processing of NR2B. NMDA activates the calcium-dependent protease calpain to degrade NR2B. Cdk5-deficient hippocampal slices showed diminished NMDA induction of NR2B degradation. Cdk5 activated and immunoprecipitated but did not phosphorylate calpain in vitro.
Together, these data suggest that Cdk5 normally activates calpain-mediated degradation of NR2B, which is important in moderating hippocampal synaptic transmission involved in learning. Although targeted inhibition of Cdk5 might be beneficial for people with learning disorders, universal inhibition of Cdk5 should be avoided because Cdk5 has many neuronal substrates.