3 December 2006. Neurons cannot afford to be spineless. Not only are those dendritic protrusions essential for transmitting incoming signals from other neurons, but their continual growth and retraction endows neural networks with the plasticity essential for learning and memory. In fact, reduction in dendritic spine density is one of the earliest pathological markers in mouse models of Alzheimer disease (see ARF related news story) and possibly in human disease, too, where learning and memory problems surface long before amyloid plaques and neurofibrillary tangles. So whatever factors drive loss of dendritic spines could be key pieces to the AD puzzle.
One of those factors is none other than the serine/threonine kinase Cdk5. In today’s Nature Neuroscience online, Nancy Ip and colleagues at the Hong Kong University of Science and Technology report that the kinase is crucial for dendritic spine retraction in hippocampal neurons. The finding places Cdk5, already linked to Aβ and tau pathology, at the scene of perhaps one of the earliest AD pathologies.
Ip and colleagues, including collaborators Michael Greenberg and coworkers at Children’s Hospital, Boston, discovered the spineless side to Cdk5 when studying the Eph family of receptor tyrosine kinases, which play key roles in synaptic remodeling. When activated by ephrins, EphBs help drive spine morphogenesis, while EphAs, in particular EphA4, drive spine retraction in hippocampal neurons (see Murai et al., 2003). Though what happens downstream of EphA4 activation has remained a bit of a mystery, the researchers wondered if Cdk5 might be involved, given that the serine/threonine kinase localizes to postsynaptic regions. To test this idea, joint first authors Wing-Yu Fu and Yu Chen added the Cdk5 inhibitor roscovitine to hippocampal slices, then challenged them with ephrin-A1. They found that roscovitine completely blocked spine retraction. They were also able to prevent spine loss in cultured hippocampal neurons by knocking down Cdk5 with RNAi.
The researchers went on to find that Cdk5 and its activator p35 interact with EphA4, and that activation of the latter by ephrin causes an increase in Cdk5 activity, most likely by phosphorylating the kinase at tyrosine 15—in cells expressing a Y15F Cdk5, mutant ephrin-A1 failed to have any effect on dendritic spine number. The researchers found that downstream of this interaction, Cdk5 was necessary for activation of the GTPase RhoA and for clustering of EphA4 on the cell membrane, an important step in recruitment and activation of other downstream signaling molecules including the guanidine exchange factor ephexin-1. In fact, Fu, Chen, and colleagues determined that Cdk5 phosphorylates ephexin-1, thereby inducing activation of RhoA, which is intimately involved in regulating cytoskeletal actin in dendritic spines—RhoA has also been implicated in ectodomain shedding and activation of γ-secretase (see ARF related news story). All told, the data point to Cdk5 as a major conduit for ephrin signals to the cytoskeleton.
Cdk5 is perhaps best known for regulating neurodevelopment, and its involvement in spine retraction would fit that role, especially as the authors found that ephrin-A1 causes spine loss in hippocampal neurons as early as postnatal day 7. The effects also seem physiologically significant because Ip and colleagues report that ephrin-driven spine retraction causes a decrease in synapses and a concomitant reduction in mini-excitatory postsynaptic currents (these are widely accepted as reflecting synapse number). In this respect, it is worth noting that p35-/- mice exhibit impaired long-term potentiation and long-term depression, which might be explained by poor spine regulation, though it might also be a direct result of neurodevelopmental problems.
Whether Cdk5-mediated spine loss has any pathological significance is not yet clear. But there has been much interest in the role of Cdk5 in neurodegenerative diseases of late, especially because of the potential involvement of the kinase in tau and Aβ pathology in AD. That dark side of Cdk5 has been linked to p25, the highly stable fragment of p35. Constitutive activation of Cdk5 by p25 can drive phosphorylation of tau (see ARF related news story) and production of intraneuronal Aβ (see Cruz et al., 2006 and ARF related news story). But interestingly, Ip and colleagues report that though Cdk5-p35 interacts with the EphA4 receptor, Cdk5-p25 does not, suggesting that spines may be spared in situations where p25 is upregulated, as in AD brain (see ARF related news story).—Tom Fagan.
Fu W-Y, Chen Y, Sahin M, Zhao X-S, Shi L, Bikoff JB, Lai K-O, Yung W-H, Fu AKY, Greenberg ME, Ip NY. Cdk5 regulates EphA4-mediated dendritic spine retraction through an ephexin1-dependent mechanism. Nature Neurosci. Advanced online publication. December 3, 2006. Abstract