One poster that stood out at last month’s 35th Annual Conference of the Society for Neuroscience was that of Andre Fischer, Farahnaz Sananbenesi, and colleagues from the labs of Li-Huei Tsai at Harvard Medical School and Bai Lu at NIMH in Bethesda, Maryland. Presenting the latest chapter in the evolving story of what the kinase Cdk5 might do in neurodegeneration and Alzheimer disease, the poster rose above the din of the conference hall by offering a surprising twist that revives an old idea. P25, the small protein that the field has come to view as a pathological activator of the kinase Cdk5, is actually good for learning and memory when it acts briefly and then disappears, the scientists found. Indeed, it might be part of the brain’s compensatory effort at sharpening its cognition in the face of stressors, and turn into a destructive force only when active in runaway fashion. The study appears today in the journal Neuron (Fischer et al., 2005).
In the past 6 years, a series of studies by several labs has placed the serine/threonine kinase Cdk5 amid the intersecting web of hypotheses of what causes AD. Cdk5 is known to be important for brain development and has been implicated in various brain functions in adult life. Late in life, it is suspected of participating in the pathogenesis of AD when its normal, tightly regulated activation by the short-lived protein p35 gets effectively turned into constitutive activation by the long-lived p35 truncation fragment p25. Links to neurodegeneration, and to both major AD pathologies, abound. For example, neurotoxic stimuli induce the protease calpain, which in turn cleaves p35 into p25. The protein tau is among Cdk5’s many substrates, and chronic p25/Cdk5 activation in the forebrain causes neurofibrillary pathology, gliosis, and severe neurodegeneration in mouse models. The connection to APP or the Aβ peptide is less clearly established to date, though some research has placed Aβ upstream of p25. Likewise, high levels of p25 in postmortem AD brain have been described by Tsai’s and one other group but are not widely reproduced as yet.
To sort out what role p25/Cdk5 might play in AD pathogenesis, it would surely help to understand its function. In other words, what is all that phosphorylating good for? The present study addresses this issue by focusing squarely on synaptic plasticity. Prior studies have already implicated Cdk5 in long-term potentiation (LTP) and other aspects of synaptic function. For example, Cdk5 phosphorylates the postsynaptic protein PSD-95 as part of its function in clustering NMDA receptors; PSD-95 also appears to be reduced in the context of accumulating intraneuronal Aβ (see ARF related SfN story; Almeida et al., 2005). Similarly, Cdk5 phosphorylates hippocampal NR2A, an NMDA receptor subunit, as well as other synaptic proteins. To find out how p25/Cdk5 might affect the dynamics of synapses, Fischer et al. used transgenic mice in whose forebrain they could turn p25 production on or off at will by feeding the antibiotic tetracycline or its analog doxycycline. The scientists toggled this switch such that they were able to compare young adult mice that expressed p25 for 6 weeks to others that only made p25 for 2 weeks and then lived another month without it. The researchers assessed cognitive performance, LTP, dendritic spine density, and synapse number.
As expected, Fischer et al. found that prolonged expression of p25 severely impaired LTP in the hippocampus. The mice performed poorly in a fear-conditioning test and the Morris water maze. They also suffered astrogliosis, lost synapses, and their neurons died in droves. The surprise came with the other mice, in whom transient exposure to p25 turned out to be an altogether good thing. They aced the associative and the spatial learning and memory tests, and cultured slices of their hippocampus produced enhanced LTP. They had more dendritic spines on their neuronal dendrites, and more synapses, than control mice. Their neurons stayed alive and well. Intriguingly, their improved performance lasted for a month until after p25 was shut off again, as if its temporary presence had indelibly changed the synapses. The cellular mechanisms for this require further investigation. Early hints and prior work, however, suggest that NR2A phosphorylation and a subsequent facilitation of NMDA receptor signaling may have much to do with the improved LTP. Formation of new spines and synapses, possibly mediated by the Cdk5 substrate PAK-1, could further improve cognitive performance, Fischer and colleagues write.
In summary, the data show that regulated Cdk5 activity is important not only for development but also for learning and memory, writes Frank LaFerla of the University of California, Irvine, in an accompanying news and views article. A critical next step for unraveling the role of Cdk5 in AD pathogenesis will be to analyze how aging mice respond to p25 induction, LaFerla writes. But even now, he concludes, the available body of work on Cdk5 points to this kinase as a central player in memory, plasticity, and the pathologies of AD.
Not all findings in this study fit neatly into a clear-cut picture. The authors point to the paradox that mice with prolonged p25 expression retain high numbers of dendritic spines even as they perform poorly, suffer neurodegeneration, and show electrophysiological deficits. Perhaps this could mean that the additional spines formed in response to p25 induction are not functional. Their increase might reflect the brain’s attempt to compensate for the overall decline in synaptic activity while p25 levels stay high, the authors speculate. Whatever it will turn out to mean, the present data, together with cell culture experiments in this study and prior papers, suggest that p35/Cdk5 normally regulates the formation of dendritic spines, and that p25 stokes this process, the authors note. This would be part of the cellular mechanisms of synaptic plasticity regulated by Cdk5, which can be beneficial with short-term use of p25 but detrimental with chronic exposure.
Finally, the authors leave the reader with some speculation on the broader implications of their data. Could p25 have a physiological role in synaptic plasticity? Other investigators have described evidence for altered neuroplasticity in AD and in reaction to brain injury. Perhaps p25 starts out as a compensatory response to risk factors, but then goes bad? Some evidence exists suggesting that insults such as Aβ, excitotoxicity, and ischemia trigger p25 expression, the authors note. In some people, p25 levels might increase in response to flagging Cdk5 activity with advancing age. A genetic component appeared in the literature recently, when Rosa Rademakers and colleagues at University of Antwerp in Belgium identified an SNP in the Cdk5 gene in some Dutch and Swedish people with familial AD. To test the idea of compensatory p25 production, it will be important to measure p25 levels in patients with MCI but prior to neuronal loss.
P25, then, could be an example for how the very processes of neural plasticity represent an Achilles heel for the thinking, aging brain. The authors point out that this idea goes back to Marek-Marsel Mesulam and Thomas Arendt. These investigators postulated years ago that mechanisms that enhance neural plasticity can become maladaptive when used chronically. In essence, the hypothesis holds that those areas of the brain that show the most plasticity eventually cave in under its burden.—Gabrielle Strobel
- Almeida CG, Tampellini D, Takahashi RH, Greengard P, Lin MT, Snyder EM, Gouras GK. Beta-amyloid accumulation in APP mutant neurons reduces PSD-95 and GluR1 in synapses. Neurobiol Dis. 2005 Nov;20(2):187-98. PubMed.
- Fischer A, Sananbenesi F, Pang PT, Lu B, Tsai LH. Opposing roles of transient and prolonged expression of p25 in synaptic plasticity and hippocampus-dependent memory. Neuron. 2005 Dec 8;48(5):825-38. PubMed.