Despite its long and distinguished history in Alzheimer’s research, the microtubule-stabilizing protein tau still poses many a riddle to scientists. They do know that excessive phosphorylation of tau somehow figures in neurodegeneration, yet which kinase enzymes start this process, which ones egg it on, and where along the way the neuron sustains damage are among the questions that still confound them. A paper in today’s Neuron moves the story forward. It introduces a new animal model of tauopathy and proposes that the kinase CDK5, while not the sole instigator, is key to tangle formation and greatly worsens the progression of neurofibrillary pathology, at least in these mice.

Wendy Noble, working with Karen Duff at the Nathan S. Kline Institute of New York University in Orangeburg and colleagues there and elsewhere, crossed mice transgenic for increased CDK5 activity (Ahlijanian et al., 2000) with other transgenics overexpressing the mutant human P301L form of tau (Lewis et al., 2000) that are predisposed to tau pathology. The researchers did that because prior reports have implicated CDK5 activity driven by its cofactor p25 to be elevated in AD, and have shown that CDK5 can phosphorylate tau at sites relevant to AD (see, for example, ARF related news story). Tantalizingly, the CDK5 inhibitor roscovitine reduces tau pathology and neurodegeneration in a mouse model (see ARF related news story). And yet, CDK5 overexpression in mice is insufficient to produce tangles, indicating that other factors are at play (Bian et al., 2002; ARF related news story).

The combined model presented here shows a more pronounced phenotype, Noble and colleagues report. The mice do not have elevated CDK5 expression, but the enzyme’s activity is up roughly twofold. That enhancement came with a marked increase in tau hyperphosphorylation and tau aggregation in the brainstem and in the cortex. The double-transgenics also had more tangles in the brainstem-though no tangles in the cortex-than did the single-transgenic P301L mice, Noble et al. report. (The P301L single-transgenic have no tau pathology in the cortex, the authors point out.) Noble and colleagues also report data on GSK3, another leading suspect among tau-phosphorylating kinases and a drug target (see related news story). Like CDK5, GSK3 also showed increased activity in the double transgenics, and confocal microscopy suggests that both kinases co-localize with tau and even with each other in the cell body and some processes of cultured neurons, Duff’s team writes.

In spite of the intensified neurofibrillary pathology in the brainstem of the double-transgenics, these animals did not suffer an accelerated version of the dystonia seen in the P301T mice, at least by one year of age, the researchers note.

Putting their data in context, the authors propose a sequence of events whereby, once aberrant phosphorylation by CDK5 and/or other kinases has begun, tau proteins no longer bind microtubules, but instead become cytoplasmic and then redistributed away from the axon to the cell body, where they polymerize, fibrillize, and aggregate. During this entire time, tau phosphorylation by GSK3, CDK5, and other kinases may well continue even while tangles are already forming, the authors suggest.

Numerous animal models of tau exist by now, ranging from fly to mouse, from CDK5 to GSK3, from transgenes of mouse tau to human tau, from expression in motor and corticohippocampal neurons to sensory neurons. While this variation makes extrapolation to human tau pathogenesis difficult, the scientists consider compelling their evidence that increased CDK5 activity can promote tangle formation in mice that are predisposed to tauophathy.

How does all this tie in with AD and the amyloid hypothesis? Many researchers think that amyloid can drive tau pathology, perhaps via CDK5 activity. For example, some have proposed that Aβ induces tau hyperphosphorylation, and that the protease calpain-which leads to increased CDK5 activity-is elevated in AD brain (see ARF related news story; Town et al., 2002). P301T mice develop more severe neurofibrillary pathology when crossed with mice producing excess Aβ (see ARF related news story). None of these pathways are proven to occur in human neurodegenerative diseases involving tau, but the cumulative evidence is suggestive enough to explore the use of kinase inhibitors to counteract the progression of neurofibrillary pathology, Duff and colleagues write.—Gabrielle Strobel


  1. The combination of mutant tau-P301L with the CDK5-activating cofactor p25 in brain of double-transgenic mice is proven here to increase phosphorylation of tau and its aggregation into filaments. This outcome is not totally unexpected, and confirms the fact that CDK5 was proposed and identified as tau-kinase II—and GSK3β as tau-kinase I (Ishiguro et al., 1992). This was confirmed by many studies since then, at least in cell culture. Nevertheless, in brain in vivo, the situation was and is more complex, since even overexpression of CDK5 and p35 with human tau-4R in brain of triple-transgenic mice was not sufficient to increase tau-phosphorylation appreciably (Van den Haute et al., 2001) as opposed to GSK3β (Spittaels et al., 2000).

    Given the phenotype of the parental transgenic mouse strains, and of other strains as published, comparing them to the current presented double-transgenic strain brings up some interesting questions.

    First, the criteria used here to define increased CDK5 activity as pathological are biochemical and histological in nature. No mention is made of neurotoxicity or neuronal loss, defects in LTP or cognition, deviations in behavior or in any other aspect that would indicate that more aggregated tau is bad for neuronal health (although certainly refraining from claiming that it is in any way good or healthy).

    Second, both parental mouse strains suffer motoric and health problems at an early age: Whole-body exertion tremors in p25 mice at age four to nine weeks (Ahlijanian et al., 2000) and absence of escape extension at age 6.5 months in hemizygous tau-P301L mice (Lewis et al., 2000). These phenotypic aspects appear to be missing (completely?) in the double-transgenic mice analyzed here, despite their advanced age of 12 months. Could that mean that moderately increased CDK5 activity has the same effect as moderately increased GSK3β activity, namely rescue of the motoric problems (Spittaels et al., 2000)?

    Third, the relative levels of kinases and their site of action, both in terms of subcellular site and in brain regional site, will be determining their actual activity and specificity. For CDK5 and GSK3β to induce or promote "tau pathology" as defined by intraneuronal tau aggregates might indeed require "complexes" of some sort, as tentatively identified and proposed by the authors. And I also agree with them when they state that we need to implement in the models the actual—and still elusive—link of increased tau phosphorylation to amyloid (or vice versa, depending on your school of thought …) (Terwel et al., 2002).


    . Tau protein kinase I converts normal tau protein into A68-like component of paired helical filaments. J Biol Chem. 1992 May 25;267(15):10897-901. PubMed.

    . Coexpression of human cdk5 and its activator p35 with human protein tau in neurons in brain of triple transgenic mice. Neurobiol Dis. 2001 Feb;8(1):32-44. PubMed.

    . Glycogen synthase kinase-3beta phosphorylates protein tau and rescues the axonopathy in the central nervous system of human four-repeat tau transgenic mice. J Biol Chem. 2000 Dec 29;275(52):41340-9. PubMed.

    . Axonal transport, tau protein, and neurodegeneration in Alzheimer's disease. Neuromolecular Med. 2002;2(2):151-65. PubMed.

  2. This paper is very relevant to our work and that of others. It begins to bring together the roles of tau dysregulation and aberrant phosphorylation in a mammalian model. It further demonstrates the role for tau hyperphosphorylation in accelerating NFT pathology.

    Although in the discussion it is stated that the other mouse models and fly models are at odds, or inconclusive, they actually are quite in line with these current findings. The one fly model that explores the role of phosphorylation (Jackson et al., 2002) shows that both tau dysregulation and dysregulation of kinase activity (in this case GSK3β, but I suspect CDK5 would be similar) are needed to form NFTs, and that altered kinase activity alone, or wild-type tau overexpression alone, are insufficient. This current paper shows that hyperactive kinase accelerates NFT formation, just as it does in the fly.

    Quite importantly, this demonstrates nicely a role for CDK5 in addition to the previously demonstrated role for GSK3β in NFT formation. It also confirms in vivo that CDK5 activation leads to GSK3β activation, by a mechanism that has yet to be uncovered. From a therapeutic standpoint, this work supports continued development of treatments focused on kinase inhibition. However, given the multifaceted roles of these kinases, the potential toxicity of GSK3β or CDK5 inhibition remains an issue.


    . Human wild-type tau interacts with wingless pathway components and produces neurofibrillary pathology in Drosophila. Neuron. 2002 May 16;34(4):509-19. PubMed.

  3. Tau aggregation is a central issue for understanding tauopathies, including AD. Crossbreeding CDK5-activator p25 transgenics with P301L transgenics resulted in hyperphosphorylation of human tau and induced tau aggregation in neurons. The correlation between hyperphosphorylation and aggregation of tau is not simple. In utero, tau is highly phosphorylated, but not aggregated. The different phosphorylation sites between fetal tau and PHF tau have been reported and may provide answers regarding which kinases are essential for formation of tau aggregates. Regarding this point, this paper did not satisfy the criteria for phosphorylation sites of PHF-tau, because Serine 202 and 404 of tau are phosphorylated in fetal and PHF tau, although they are phosphorylation sites of CDK5. The phosphorylation of Ser422, Ser262, and AT100 epitopes are specific to PHF tau, but the activation of CDK5 alone cannot explain the phosphorylation of these sites, even in synergistic activation with GSK3. For this reason, it is thought that tau phosphorylation by CDK5 in these mice might not itself induce tau aggregation; this may occur through other cascades. Nevertheless, CDK5 inhibitor administration into tau Tg mice may clarify the involvement of CDK5 on tau aggregation.

  4. An interesting observation emanating from the studies of Noble et al. is that the double-transgenic mice overexpressing p25 and mutant tau show an increase in the active form of GSK3β while total GSK3β levels remain unaltered. Although the GSK3β kinase activity was not directly determined in these studies, this observation raises a question: What is the link between elevated p25 levels (which activate CDK5) and the active form of GSK3β? To date, there is no evidence that GSK3β can be directly phosphorylated and activated by CDK5. Nevertheless, these findings suggest that hyperphosphorylation of tau observed in the p25/T double-transgenic animals is likely caused by both kinases. Earlier studies had shown that phosphorylation of tau by CDK5 makes tau a "better" substrate for GSK3β. Thus, these observations suggest two ways by which sustained activation of CDK5 causes hyperphosphorylation of tau: by direct phosphorylation and by activation of GSK3β, which will further phosphorylate tau. Hyperphosphorylated tau is released from microtubules and forms aggregates and tangles leading to neuronal pathology.

    There is, however, a wrinkle to this conclusion. An earlier study by Spittaels et al. showed that constitutive overexpression of human GSK3β in tau-transgenic mice actually reduced the axonal pathology. On the other hand, Lucas et al. observed that conditional expression of Xenopus GSK3β in mice promoted tau hyperphosphorylation and neurodegeneration. Although these differences could be explained by differences in the experimental designs (constitutive vs. conditional expression of transgene, species and strain differences, etc.), these observations exhibit the complexities of interpreting data from transgene studies.

    Finally, the "AβPpists" might wonder what connection AβPP or Aβ might have to these observations. One possibility suggested by the observations of Town et al. is that Aβ stimulates p25 production in a calpain-dependent fashion. Increased levels of p25 will activate CDK5 (and indirectly GSK3β). Considering the "cross-talk" between various kinases, such a simple scenario probably represents oversimplification. Also, direct evidence supporting this chain of "cause-and-effect" remains lacking. Future studies focused on understanding the molecular mechanism by which p25 (and, therefore, activated CDK5) stimulates GSK3β will likely bring us a step closer to understanding tangle formation.


    . Glycogen synthase kinase-3beta phosphorylates protein tau and rescues the axonopathy in the central nervous system of human four-repeat tau transgenic mice. J Biol Chem. 2000 Dec 29;275(52):41340-9. PubMed.

    . Decreased nuclear beta-catenin, tau hyperphosphorylation and neurodegeneration in GSK-3beta conditional transgenic mice. EMBO J. 2001 Jan 15;20(1-2):27-39. PubMed.

    . p35/Cdk5 pathway mediates soluble amyloid-beta peptide-induced tau phosphorylation in vitro. J Neurosci Res. 2002 Aug 1;69(3):362-72. PubMed.

  5. Neurofibrillary tangles consist of tau protein that has forsaken its role as a stabilizer of microtubules to polymerize into abnormal fibrils within neurons. Preclinical and clinical evidence leaves little doubt that abnormal tau polymerization is injurious to neurons. A prominent characteristic of tangles is a high degree of site-specific phosphorylation that is thought to contribute to the dysfunction and polymerization of tau, as well as to the stability of tau filaments. Several kinases are implicated in tau hyperphosphorylation in brain, and one that has garnered attention in Alzheimer's pathogenesis is cyclin-dependent kinase-5 (CDK5). There is evidence in AD that CDK5 is overactivated by an excess of a protein fragment called p25; the resulting increase in phosphorylation is hypothesized to facilitate tau polymerization and tangle formation.

    Transgenic mice are ideal for testing such hypotheses in vivo. Mice overexpressing p25 develop axonopathy and movement dysfunction, but the neurons show no evidence of neurofibrillary tangles (Ahlijanian et al., 2000; Bian et al., 2002), indicating that normal mouse tau is refractory to p25/CDK5-induced tangle formation. Noble and colleagues took the interesting step of crossing p25-transgenics (Ahlijanian et al., 2000) with mice expressing mutant human tau (4R0N, P301L mutation; Lewis et al., 2000). Tau in these dual-transgenic mice is hyperphosphorylated at sites linked to CDK5, and accumulates intraneuronally in several brain areas. The finding that phosphorylated sites are detectable on soluble tau (in addition to the insoluble aggregates) suggests that phosphorylation precedes, and thus may facilitate, atypical polymerization. These findings support the notion that activation of CDK5 (and, as the authors note, probably other kinases) contributes to the hyperphosphorylation and consequent autophilicity of tau in vivo. Surprisingly, the age of onset of dystonia seen in P301L mice was not noticeably advanced by the presence of the p25 transgene (nor is there yet evidence for frank neuronal loss). Perhaps older mice will show motoric (and, with luck, mnemonic) impairments.

    The CDKs regulate a number of important cellular functions such as cell division, differentiation, and apoptosis. In fact, CDK inhibitors are being considered as therapies for cancer, atherosclerosis, and viral infections, among other disorders (Knockaert et al., 2002). Hence, for the treatment of chronic neurodegenerative disorders, highly selective and demonstrably safe inhibitors of CDK5 will be needed. When such agents are found, they will find fertile ground for preclinical efficacy testing in p25/P301L-dual transgenic mice. Ultimately, as with all new, disease-modifying approaches, only clinical trials will prove the benefit of CDK5 inhibition for the treatment of AD and other tauopathies.


    . Hyperphosphorylated tau and neurofilament and cytoskeletal disruptions in mice overexpressing human p25, an activator of cdk5. Proc Natl Acad Sci U S A. 2000 Mar 14;97(6):2910-5. PubMed.

    . Axonopathy, tau abnormalities, and dyskinesia, but no neurofibrillary tangles in p25-transgenic mice. J Comp Neurol. 2002 May 6;446(3):257-66. PubMed.

    . Pharmacological inhibitors of cyclin-dependent kinases. Trends Pharmacol Sci. 2002 Sep;23(9):417-25. PubMed.

    . Neurofibrillary tangles, amyotrophy and progressive motor disturbance in mice expressing mutant (P301L) tau protein. Nat Genet. 2000 Aug;25(4):402-5. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Enzyme Essential to Brain Development Found to Hyperphosphorylate Tau, Kill Neurons
  2. Orlando: It’s Getting Hot around CDK5, Has the Field Noticed Yet?
  3. Lithium Hinders Aβ Generation, Buffing Up GSK as Drug Target
  4. The Calpain Connection
  5. Finally United? Aβ Found to Influence Tangle Formation

Paper Citations

  1. . Hyperphosphorylated tau and neurofilament and cytoskeletal disruptions in mice overexpressing human p25, an activator of cdk5. Proc Natl Acad Sci U S A. 2000 Mar 14;97(6):2910-5. PubMed.
  2. . Neurofibrillary tangles, amyotrophy and progressive motor disturbance in mice expressing mutant (P301L) tau protein. Nat Genet. 2000 Aug;25(4):402-5. PubMed.
  3. . Axonopathy, tau abnormalities, and dyskinesia, but no neurofibrillary tangles in p25-transgenic mice. J Comp Neurol. 2002 May 6;446(3):257-66. PubMed.
  4. . p35/Cdk5 pathway mediates soluble amyloid-beta peptide-induced tau phosphorylation in vitro. J Neurosci Res. 2002 Aug 1;69(3):362-72. PubMed.

Further Reading


  1. . Partial rescue of the p35-/- brain phenotype by low expression of a neuronal-specific enolase p25 transgene. J Neurosci. 2003 Apr 1;23(7):2769-78. PubMed.
  2. . Cdk5: one of the links between senile plaques and neurofibrillary tangles?. J Alzheimers Dis. 2003 Apr;5(2):127-37. PubMed.
  3. . Cyclin-dependent kinase 5 (CDK5) and neuronal cell death. Cell Tissue Res. 2003 Apr;312(1):1-8. PubMed.
  4. . Cyclin-dependent kinase inhibitors: cancer killers to neuronal guardians. Curr Med Chem. 2003 Mar;10(5):367-79. PubMed.
  5. . Characterization of tau phosphorylation in glycogen synthase kinase-3beta and cyclin dependent kinase-5 activator (p23) transfected cells. Biochim Biophys Acta. 1998 Apr 10;1380(2):177-82. PubMed.
  6. . Deregulation of cdk5, hyperphosphorylation, and cytoskeletal pathology in the Niemann-Pick type C murine model. J Neurosci. 2002 Aug 1;22(15):6515-25. PubMed.
  7. . GSK3 takes centre stage more than 20 years after its discovery. Biochem J. 2001 Oct 1;359(Pt 1):1-16. PubMed.

Primary Papers

  1. . Cdk5 is a key factor in tau aggregation and tangle formation in vivo. Neuron. 2003 May 22;38(4):555-65. PubMed.