Yong Shen does an excellent job of summarizing some of the important data supporting an important role for cell death in a number of neurodegenerative diseases. Clearly, there are a myriad of possible scenarios that may lead to neuronal cell death, and some of these mechanisms may overlap with one another. Shen illustrates exceptionally well the vast number of pathways that may lead to the stimulation of cell death pathways, including apoptosis.

With regard to Alzheimer's disease, I would like to address one topic mentioned in the text, namely: Is apoptosis irrelevant because it is a late-stage event in AD?

Most scientists would agree that the earliest known step in the cascade of events leading to neuronal cell death in AD is the formation of toxic Ab fibrils. Several studies have supported the hypothesis that Ab may induce neuronal cell death by activation of multiple pathways of apoptosis; this includes the receptor-mediated pathway and the mitochondrial pathway involving reactive oxygen species [2, 3, 5]. However, it is likely that, due to the arsenal of...  Read more

Yong Shen does an excellent job of summarizing some of the important data supporting an important role for cell death in a number of neurodegenerative diseases. Clearly, there are a myriad of possible scenarios that may lead to neuronal cell death, and some of these mechanisms may overlap with one another. Shen illustrates exceptionally well the vast number of pathways that may lead to the stimulation of cell death pathways, including apoptosis.

With regard to Alzheimer's disease, I would like to address one topic mentioned in the text, namely: Is apoptosis irrelevant because it is a late-stage event in AD?

Most scientists would agree that the earliest known step in the cascade of events leading to neuronal cell death in AD is the formation of toxic Ab fibrils. Several studies have supported the hypothesis that Ab may induce neuronal cell death by activation of multiple pathways of apoptosis; this includes the receptor-mediated pathway and the mitochondrial pathway involving reactive oxygen species [2, 3, 5]. However, it is likely that, due to the arsenal of anti-apoptotic mechanisms neurons possess, the activation of apoptosis and eventual neuronal cell death may be separated in time. We have proposed that during this protracted battle, caspases may continue to be active at low levels and cleave critical cellular proteins that may actually drive the pathology observed in AD [7].

One potential target for caspase-mediated cleavage may be the microtubule-associated protein, tau. Well known in AD circles as being the protein whose hyperphosphorylation leads to neurofibrillary tangle formation, recent evidence suggests that tau is a target for caspase-3-mediated cleavage. Three recent studies now suggest that caspase-3 cleavage of tau occurs in the AD brain, is a relatively early event (preceding PHF formation), and may enhance filament formation in vitro [1, 4, 6]. Furthermore, in the study by Gamblin et al., the authors demonstrate that the link between caspase activation and cleavage of tau is Ab. They demonstrate in primary cultured cortical neurons that application of Ab induces caspase activation and subsequent cleavage of tau. The cleavage of tau by caspase-3 leads to enhanced filament formation in vitro [1].

Supporting these findings is a study currently in press by Rissman et al. (This paper will be released in the July 1st JCI. Due to embargo, I cannot reveal details at this point.) Taken together, these studies now show is that caspase activation may in fact be a proximal event that leads to neurofibrillary tangle formation in AD.

Apoptosis in AD has long been criticized for being a late-stage event, but these studies support the idea that inhibition of caspases may be a viable target for drug development in treating AD. Yes, the morphological features of apoptosis probably do occur late in the disease, but the activation of caspases and cleavage of critical proteins may occur earlier and contribute to the downstream pathology observed in AD. It is noteworthy that numerous other neurodegenerative diseases, including Pick's disease, dementia with Lewy bodies, Progressive Supranuclear Palsy (PSP), Parkinson's and corticobasal degeneration (CBD), are all associated with the abnormal intracellular accumulations either of the tau protein or of alpha-synuclein. We have examined whether caspase-cleavage of tau is a common denominator in these diseases and found the answer to be yes (Unpublished observations). Thus, caspase activation and accumulation of cleaved tau may represent a common pathway leading to neurofibrillary lesions in these neurodegenerative diseases.

1. Gamblin, T.C., Chen, F., Zambrano, A., Abraha, A., Lagalwar, S., Guillozet, A.L., Lu, M., Fu, Y., Garcia-Sierra, F., LaPointe, N., Miller, R., Berry, R.W., Binder, L.I. and Cryns, V.L., Caspase cleavage of tau: linking amyloid and neurofibrillary tangles in Alzheimer's disease, Proc Natl Acad Sci U S A, 100 (2003) 10032-7. Abstract

2. Ivins, K.J., Thornton, P.L., Rohn, T.T. and Cotman, C.W., Neuronal apoptosis induced by beta-amyloid is mediated by caspase-8, Neurobiol Dis, 6 (1999) 440-9. Abstract

3. Miranda, S., Opazo, C., Larrondo, L.F., Munoz, F.J., Ruiz, F., Leighton, F. and Inestrosa, N.C., The role of oxidative stress in the toxicity induced by amyloid beta- peptide in Alzheimer's disease, Prog Neurobiol, 62 (2000) 633-48. Abstract

4. Rissman, R.A., Poon, W.W., Blurton-Jones, M., Oddo, S., Reidun, T., Vitek, M.P., LaFerla, F.M., Rohn, T.T. and Cotman, C.W., Caspase-cleavage of tau is an early event in Alzheimer's disease pathology, J. Clin. Invest., In Press (2004).

5. Rohn, T.T., Head, E., Nesse, W.H., Cotman, C.W. and Cribbs, D.H., Activation of caspase-8 in the Alzheimer's disease brain, Neurobiol Dis, 8 (2001) 1006-16. Abstract

6. Rohn, T.T., Rissman, R.A., Davis, M.C., Kim, Y.-E., Cotman, C. and Head, E., Caspase-9 Activation and caspase cleavage of tau in the Alzheimer's disease brain, Neurobiol Dis, 11 (2002) 341-354. Abstract

7. Rohn, T.T., Rissman, R.A., Head, E. and Cotman, C.W., Caspase activation in the Alzheimer's disease brain: tortuous and torturous, Drug News Perspect, 15 (2002) 549-557. Abstract

8. Roth, K.A., Caspases, apoptosis, and Alzheimer disease: causation, correlation, and confusion, J Neuropathol Exp Neurol, 60 (2001) 829-38. Abstract

View all comments by Troy Rohn

Besides the possibility of continuous low-grade caspase expression, I think the roles of cyclin-dependent kinase 5 (Cdk5) in the brain are critical, particularly under pathological conditions. Li-Huei Tsai's group has found that p25, a truncated form of p35, accumulates in neurons in the brains of patients with Alzheimer's disease (Patrick et al.,1999; Lee et al., 2000, see also recent ARF news update). This finding might indicate a relatively slow neurodegenerative process as the p25/Cdk5 complex hyperphosphorylates tau, which reduces tau's ability to associate with microtubules and also...  Read more

Besides the possibility of continuous low-grade caspase expression, I think the roles of cyclin-dependent kinase 5 (Cdk5) in the brain are critical, particularly under pathological conditions. Li-Huei Tsai's group has found that p25, a truncated form of p35, accumulates in neurons in the brains of patients with Alzheimer's disease (Patrick et al.,1999; Lee et al., 2000, see also recent ARF news update). This finding might indicate a relatively slow neurodegenerative process as the p25/Cdk5 complex hyperphosphorylates tau, which reduces tau's ability to associate with microtubules and also induces cytoskeletal disruption, morphological degeneration and apoptosis. This degenerative process seems to be mediated by amyloid.

Regarding the use of caspase inhibitors in AD animal models, a more straightforward way to prove the concept might be to use gene targeting to knock out caspases in AD mice. For example, one could ask whether cross-breeding mutant tau-transgenic mice with caspase knockout mice would prevent neurofibrillary tangle damage. The Tau P301L-JNPL3 mouse may be a good candidate, since that strain was reported to induce apoptosis (in oligodendrocytes) and activate caspase-3 ( Zehr et al., 2004). To study further whether the longest human tau isoform (T40) is involved in neurodegeneration, and whether a caspase mediates its role, Akihiko Takashima's R406W transgenic mice might be suitable. Recently, Virginia Lee's group developed a R406W mutant strain (Zhang et al., 2004 ). These mice have axonal degeneration, which I think could be mediated by caspases. Therefore, it would be worthwhile crossing either of these strains with a caspase 3 or 8 k.o. strain to see if that reduces axonal degeneration in vivo.

View all comments by Yong Shen

Yong Shen asks whether I consider caspase-mediated cleavage of MAP or tau a normal physiological process or a pathological process. This is an excellent question. We recently published data indicating a role for caspase-mediated cleavage of tau in neuronal cell dispersal and migration (Rohn et al., 2004). In this paper we report an action of caspase-3 involving cell dispersion that is independent of cell death. Upon plating PC12 cells, we found a transient activation of caspase-3 (within a 24-hour window) was required for these cells to disperse properly. Additional experiments showed that tau was in fact a target for caspase cleavage. A restructuring of the cytoskeleton would be expected for cells to be able to migrate and disperse. However, this activity of caspase-3 was not required for PC12 cells to differentiate into a neuronal phenotype. Our work supports other recent studies demonstrating that caspases may have additional functions beyond those described for apoptosis (Zermati et...  Read more

Yong Shen asks whether I consider caspase-mediated cleavage of MAP or tau a normal physiological process or a pathological process. This is an excellent question. We recently published data indicating a role for caspase-mediated cleavage of tau in neuronal cell dispersal and migration (Rohn et al., 2004). In this paper we report an action of caspase-3 involving cell dispersion that is independent of cell death. Upon plating PC12 cells, we found a transient activation of caspase-3 (within a 24-hour window) was required for these cells to disperse properly. Additional experiments showed that tau was in fact a target for caspase cleavage. A restructuring of the cytoskeleton would be expected for cells to be able to migrate and disperse. However, this activity of caspase-3 was not required for PC12 cells to differentiate into a neuronal phenotype. Our work supports other recent studies demonstrating that caspases may have additional functions beyond those described for apoptosis (Zermati et al., 2001); (Fernando et al., 2002). So, my answer to Yong's question would be that the cleavage of tau may serve as a physiological role or process under normal conditions, but if caspase activation is not turned off, then it may convert to a pathological process as observed in AD.

A second excellent question Yong raises regards kinases, such as Cdk5, and hyper-phosphorylation of tau. How does this fit in with the role of caspase-mediated cleavage of tau in AD? I would like to comment on this based on our report due out in JCI, but have to defer this issue until its embargo has lifted. I will just say that caspase cleavage and hyper-phosphorylation of tau may not be two independent events, but serial events.

Finally, Yong responded to my comments on how can we demonstrate in animal models that caspase activation is important for driving AD pathology? He suggests crossing caspase-3 or caspase-8 knockout mice with a mouse AD model. My understanding is that most of the caspase knockouts are lethal, which would prohibit doing such a cross. I have submitted a grant that circumvents this problem using the new AD mouse developed by Frank Laferla's group (Oddo et al., 2003). My proposal involves crossing this mouse with a Tg mouse that specifically over-expresses Bcl-2 in neurons of the CNS. If funded, I think this would provide an excellent opportunity to test this hypothesis. At the same time, I would encourage researchers to test some of the other ideas presented by Dr. Shen. In my mind, it is clear we have reached a "the proof is in the pudding" stage regarding caspase activation and apoptosis in AD. What we really need to do now is prove or disprove these ideas directly using animal models as described above before drug companies will become interested in picking up the ball.

View all comments by Troy Rohn

Yong Shen should be congratulated on this comprehensive and interesting review; it contains several novel connections and ideas. However, I was somewhat disappointed by one passage. Despite a great deal of publication and discussion, misunderstandings about NF-κB are still common. So the author can be forgiven for this misleading statement: "The activation of NF-κB translocation causes apoptosis in vitro." None of the references cited, including the Yang et al. (2002) paper, provides anything other than a correlation between conditions that cause cell death and the activation of NF-κB. This is likely to be because NF-κB is induced in a compensatory fashion. The McCarthy reference that was cited even includes the following statement arguing for a disconnection between NF-κB and cell death: "Mutational analysis revealed the pro-apoptotic function of RIP2 to be restricted to its C-terminal CARD domain, whereas the intact molecule was necessary for NF-κB activation."

There are now scores of reports...  Read more

Yong Shen should be congratulated on this comprehensive and interesting review; it contains several novel connections and ideas. However, I was somewhat disappointed by one passage. Despite a great deal of publication and discussion, misunderstandings about NF-κB are still common. So the author can be forgiven for this misleading statement: "The activation of NF-κB translocation causes apoptosis in vitro." None of the references cited, including the Yang et al. (2002) paper, provides anything other than a correlation between conditions that cause cell death and the activation of NF-κB. This is likely to be because NF-κB is induced in a compensatory fashion. The McCarthy reference that was cited even includes the following statement arguing for a disconnection between NF-κB and cell death: "Mutational analysis revealed the pro-apoptotic function of RIP2 to be restricted to its C-terminal CARD domain, whereas the intact molecule was necessary for NF-κB activation."

There are now scores of reports showing that NF-κB activation actually promotes cell survival, as we first reported (Barger et al., 1995, see also Wang et al., 1998 , Wu et al., 1998 ). Much of the confusion about NF-κB comes from analysis of culture lysates or tissue homogenates that contain non-neuronal cell types. The transcription factor is associated with inflammatory reactions in microglia and astrocytes, which can participate in a "sign change" that results in neurotoxicity, primarily through the release of excitotoxins; (see further below). But even in the activated microglia themselves, NF-κB probably promotes survival (e.g., Lu et al., 2002). There are new reports suggesting subtleties that arise from differences in the components of the "NF-κB" being activated, namely, a distinction between RelA- and c-Rel-containing complexes (Pizzi et al., 2002), but this paper appears to suffer the old problem of glial contamination, as well. Taken together, the more rigorous and well-controlled studies generally show anti-apoptotic, pro-survival roles for NF-κB.

On a separate note, I was gratified to see in Yong Shen's Alzforum statement a section on excitotoxicity. I would like to draw a link between that topic and inflammatory activation of microglia. A significant number of papers have reported the release of glutamate receptor ligands by activated microglia. Glutamate itself is released, apparently through the exchange for cystine through the Xc transport system (Barger & Basile, 2001). Quinolinic acid is another candidate for such ligands. At least in microglia-neuron cocultures, these ligands for glutamate receptors activate the NMDA receptor to a degree that causes substantial excitotoxicity. In vivo, the high capacity for glutamate uptake by astrocytes may help to attenuate this effect, unless they are themselves compromised by ischemia or other pathological conditions. More recently, we determined that D-serine, an agonist at the "glycine" site of the NMDA receptor, is also released by microglia treated with Aβ (Wu et al., 2004, J. Neuroinflammation 1:2). Because the uptake mechanisms for D-serine are poorly characterized, it is unclear that this agonist could be cleared as effectively as glutamate. Furthermore, it may potentiate glutamate effects to the extent that even small elevations of the latter could become excitotoxic.

View all comments by Steve Barger

I am glad Steve Barger brought up the controversy on whether ND-κB causes or counteracts apoptosis. This question in itself is interesting enough to attract more rigorous approaches for definitive study. Most of the studies that found NF-κB to be anti-apoptotic are conducted not only in non-neuronal cells, but also in immune or tumor-related cells. For example, the study by Wu et al., 1998 , was conducted in Jurkat cells. The experiments performed by Wang et al., 1998 were in H710801, a fibrosarcoma cell line. The paper from Pizzi et al., 2002, particularly the key evidence (Figs. 1-3), was done by using cerebellar granule cells. On the other hand, most of the evidence suggesting that NF-κB is involved in apoptosis is from cortical neurons (Schneider et al., Nat Med. 1999; Lipton, Nat Med., 1997), hippocampal neurons (  Read more

I am glad Steve Barger brought up the controversy on whether ND-κB causes or counteracts apoptosis. This question in itself is interesting enough to attract more rigorous approaches for definitive study. Most of the studies that found NF-κB to be anti-apoptotic are conducted not only in non-neuronal cells, but also in immune or tumor-related cells. For example, the study by Wu et al., 1998 , was conducted in Jurkat cells. The experiments performed by Wang et al., 1998 were in H710801, a fibrosarcoma cell line. The paper from Pizzi et al., 2002, particularly the key evidence (Figs. 1-3), was done by using cerebellar granule cells. On the other hand, most of the evidence suggesting that NF-κB is involved in apoptosis is from cortical neurons (Schneider et al., Nat Med. 1999; Lipton, Nat Med., 1997), hippocampal neurons (Yang et al. (2002); Li et al., 2004), or midbrain neurons (de Erausquin et al., 2003).

I believe that the experimental evidence from both sides is correct. It seems that functions of NF-κB are cell type-dependent. Even within the brain, consistent with recent publications (Wu et al., 2004, J. Neuroinflammation 1:2; Pizzi et al., 2002), the responses of NF-κB to neurons and microglia are completely different. Which molecules mediate this discrepancy, and what mechanisms are behind this remains unclear.

What would be good ways to evaluate NF-kappa B? I think we need to look at both NF-κB binding activity as well as NF-κB translocation, because this is what (NF-κB) naturally does. Many studies demonstrate that IKKalpha can cause degradation of IkappaBalpha by phosphorylation, and result in the release of cytoplasmic NF-κB, which is then able to translocate into the nucleus (Miyamoto et al., 1994; Verma et al., 1995; Baeuerle and Baltimore, 1996). Probably p65 binds certain types of nuclear protein and-depending on the cell type or/and pathological conditions-turns on or off apoptotic genes or anti-apoptotic genes.

Lastly, Steve draws a link between that topic and inflammatory activation of microglia. This is a great idea. Clearance and uptake of glutamate by microglia is critical. Appropriate glial activation may help maintain neuronal functions, including synaptic transmission. For example, Roger Nicoll's group (Luscher et al, 1998) found that glial cell responses (in this case glial glutamate transporter currents measured with whole-cell recording) are very sensitive to release of glutamate yet remained constant during LTP by directly measuring changes in the synaptic release of glutamate. Moreover, Robert Malenka's group (Malenka et al., 2003) found that glial TNFa enhances neuron synaptic plasticity in young neurons.

View all comments by Yong Shen

Yong pointed out in his reply posted 6/24/2004 that the role of cyclin-dependent kinase 5 (Cdk5) in the pathology and death of neurons is important. Identified over a decade ago as tau-phosphorylating entities purified from brain lysates, Cdk5 and GSK3 have emerged as the two major kinases that contribute to tau hyperphosphorylation and other tau-associated pathologies in animal models of Alzheimer's disease.

Intriguingly, calpain-mediated cleavage of the 'physiological' cdk5 activator p35 into the 'pathological' cdk5 activator p25 results in upregulation of cdk5 activity and an increased tendency for phosphorylating substrates implicated in neurodegeneration, such as tau. Accordingly, in AD and other neurodegenerative diseases, p25/cdk5 is thought to be an important link between excitotoxicity/calcium (calpain is Ca-dependent) and the downstream pathological processes that result in cell death.

In vivo studies strongly support this view. Karen Duff's group has recently shown that p25 enhances the progression of neurofibrillary pathology in transgenic mice...  Read more

Yong pointed out in his reply posted 6/24/2004 that the role of cyclin-dependent kinase 5 (Cdk5) in the pathology and death of neurons is important. Identified over a decade ago as tau-phosphorylating entities purified from brain lysates, Cdk5 and GSK3 have emerged as the two major kinases that contribute to tau hyperphosphorylation and other tau-associated pathologies in animal models of Alzheimer's disease.

Intriguingly, calpain-mediated cleavage of the 'physiological' cdk5 activator p35 into the 'pathological' cdk5 activator p25 results in upregulation of cdk5 activity and an increased tendency for phosphorylating substrates implicated in neurodegeneration, such as tau. Accordingly, in AD and other neurodegenerative diseases, p25/cdk5 is thought to be an important link between excitotoxicity/calcium (calpain is Ca-dependent) and the downstream pathological processes that result in cell death.

In vivo studies strongly support this view. Karen Duff's group has recently shown that p25 enhances the progression of neurofibrillary pathology in transgenic mice expressing the FTDP-mutant tau (Noble et al., 2003). In addition to tau hyperphosphorylation, p25 can cause cytoskeleton disruption, morphological degeneration, and apoptosis in cultured neurons. An inducible p25 transgenic mouse model that expresses p25 in the postnatal forebrain exhibits severe brain atrophy, up to 40 percent loss of total neurons in the brain, and tau-associated pathology (Cruz et al., 2003). While phosphorylation of tau (and other microtubule-associated proteins) by Cdk5 may contribute to cytoskeleton disruption and neuronal loss, recent studies suggest additional mechanisms by which the p25/Cdk5 kinase induces cell death.

One potential downstream cdk5 target mediating apoptotic events may be MEF2, a pro-survival transcription factor (Gong et al., 2003). Phosphorylation of Ser444 in the trans-activation domain of MEF2 leads to inhibition of MEF2 function in neurons. Ser444 phosphorylation is induced by oxidative stress and excitotoxicity. MEF2 mutations resistant to Cdk5 phosphorylation protect neurons from Cdk5-induced and neurotoxin-induced apoptosis. Another potential target of p25/Cdk5 is the NR2A subunit of NMDA receptors (Li et al., 2001; Wang et al., 2003). Phosphorylation of Ser1232 of NR2A by Cdk5 is induced by ischemic injury of the brain (Wang et al., 2003). Remarkably, expression of a NR2A mutant resistant to Cdk5 phosphorylation can prevent the death of CA1 neurons from ischemic injury.

As discussed by Yong, there are numerous ways for a neuron to die; the same idea applies in p25/cdk5-mediated cell death. Perhaps different Cdk5 substrates contribute to cell death depending on the type of insult or physiological context. Alternatively, these substrates could all become phosphorylated to orchestrate the process of neuronal death. It is likely that additional cdk5 substrates will be identified in the future that contribute to p25/cdk5- mediated cell death.

In the inducible p25 transgenic mice, induction of caspase 3 is observed that coincides with neuronal loss. It was reported recently that inhibition of Cdk5 is protective in necrotic and apoptotic paradigms of neuronal cell death and prevents mitochondrial dysfunction (Weishaupt et al., 2003). Thus, the neuronal death induced by p25/Cdk5 is likely to operate via the mitochondria-dependent mechanism. As neurodegeneration progresses rapidly in the inducible p25 transgenic mice, a rescue of neuronal loss should be relatively easy to detect. Crossing these mice with various caspase mutant mouse strains should help reveal the identity of the caspase family member(s) that mediate(s) p25/Cdk5- induced neuronal death. This information has clear implications in the therapeutic intervention of AD and ischemic brain injury. We have initiated a collaboration with Dr. Junying Yuan at Harvard Medical School to perform this experiment.

References:

Cruz, J.C. et al. (2003) Aberrant Cdk5 activation by p25 triggers pathological events leading to neurodegeneration and neurofibrillary tangles. Neuron 40 (3), 471-483. Abstract

Gong, X. et al. (2003) Cdk5-mediated inhibition of the protective effects of transcription factor MEF2 in neurotoxicity-induced apoptosis. Neuron 38 (1), 33-46. Abstract

Li, B.S. et al. (2001) Regulation of NMDA receptors by cyclin-dependent kinase-5. Proc Natl Acad Sci U S A 98 (22), 12742-12747. Abstract

Noble, W. et al. (2003) Cdk5 is a key factor in tau aggregation and tangle formation in vivo. Neuron 38 (4), 555-565. Abstract

Wang, J. et al. (2003) Cdk5 activation induces hippocampal CA1 cell death by directly phosphorylating NMDA receptors. Nat Neurosci 6 (10), 1039-1047. Abstract

Weishaupt, J.H. et al. (2003) Inhibition of CDK5 is protective in necrotic and apoptotic paradigms of neuronal cell death and prevents mitochondrial dysfunction. Mol Cell Neurosci 24 (2), 489-502. Abstract

View all comments by Li-Huei Tsai

Don't Lock the Barn Door After the Horse is Gone
Yong Shen presents a comprehensive review of molecular phenomena related to neuron death in AD. Yet I suggest that his major contention deals with an event that is not one that leads to the major clinical symptoms but is, rather, the final straw. "We argue that despite all else, it is still neuronal loss in the brain that leads to the major clinical symptoms of these disorders, and that the study of death mechanisms in AD deserves renewed attention," Shen writes. However, just as the death of the individual with AD does not constitute the major clinical symptom of the disease, so is the death of AD affected neurons not a major contributor to the clinical symptomatology of AD. Neuron death is the latecomer in the cellular cascade of AD - the culmination of a series of events that consume decades prior to the death of that cell (e.g. Morsch et al., 1999). It is the series of events that precede neuron death that are central to the cognitive and mnemonic losses that define clinical AD. And loss and dysfunction of...  Read more

Don't Lock the Barn Door After the Horse is Gone
Yong Shen presents a comprehensive review of molecular phenomena related to neuron death in AD. Yet I suggest that his major contention deals with an event that is not one that leads to the major clinical symptoms but is, rather, the final straw. "We argue that despite all else, it is still neuronal loss in the brain that leads to the major clinical symptoms of these disorders, and that the study of death mechanisms in AD deserves renewed attention," Shen writes. However, just as the death of the individual with AD does not constitute the major clinical symptom of the disease, so is the death of AD affected neurons not a major contributor to the clinical symptomatology of AD. Neuron death is the latecomer in the cellular cascade of AD - the culmination of a series of events that consume decades prior to the death of that cell (e.g. Morsch et al., 1999). It is the series of events that precede neuron death that are central to the cognitive and mnemonic losses that define clinical AD. And loss and dysfunction of synapses are vital consequences and players in this cascade that precedes death of a neuron.

Synapses are the means by which the nervous system transmits, processes, and stores information - all functions that are disrupted in AD. Although a synapse cannot exist once its parent neuron has died, it is clear that a synapse can cease to exist or function normally while its parent neuron remains alive, though sick. The argument here, then, is that synapse loss and dysfunction are the critical events in the deficits of AD, and that neuron death is a late event that administers the coup de grace by depleting the few synapses remaining associated with that neuron, whose postsynaptic dendrites have long since regressed (e.g. Buell and Coleman, 1979; Braak and Braak, 1997).

What is the evidence in support of this contention? For over a decade, it has been known that synapse density is a significant correlate of cognitive and mnemonic deficits in AD while neuron density is not (e.g. Terry et al., 1991). The fact that synapses may be lost independently of neuron death has been demonstrated by the reduction by about half of the synapse/neuron ratio in AD (e.g. Bertoni-Freddari et al., 1996) as well as by the demonstration of reduced synaptophysin transcript in living, tangle-bearing neurons (Callahan et al., 2002). In addition to frank loss of synapses in AD, there is an additional decrement in functional capacity of existing synapses, as has been demonstrated by decreased expression of transcripts related to trafficking of synaptic vesicles even when transcripts related to other aspects of synaptic structure, such as PSD-95, remain unaffected (e.g. Yao et al., 2003). It seems probable that altered capacities of existing synapses may be a significant factor, since the correlation of about 0.75 between synapse density and cognitive status yields an r squared that accounts for only about 50 percent of the variance in cognition and memory (DeKosky and Scheff, 1990; Terry et al., 1991; Scheff and Price, 2003 for review).

It remains to be formally determined whether synapse loss and dysfunction in AD is a way station in the cellular progression to neuron death, or whether synapse loss and dysfunction can, and does, take place through molecular cascades independent of the processes leading to neuron death, or indeed whether neuronal death processes are a consequence of local events at the synapse which may then feed back to the soma (e.g. Eberwine et al., 2001; Job and Eberwine, 2001).

It follows that studies that focus on neuron death as the end point may or may not address synapse loss and dysfunction. Such studies may address synaptic deficits if, by chance, they addresses a critical step in the very earliest stages of the cascade to death of the neuron. A focus on neuron death will, unfortunately, fail to address the AD clinical phenotype if it addresses molecules participating in the late stages of the obvious end point of death of a neuron. Likewise, if type 1 and type 2 synaptic deficits (above) are related to cascades distinct from those leading to neuron death. Studies of neuron death are appropriate in their own right and have yielded many valuable insights. At the same time it is vital to place neuron death in proper perspective when dealing with the clinical phenotype of AD and other neurodegenerative diseases. That perspective is as an end game that is the consequence of a long series of events that have had major consequences for the clinical phenotype of AD well in advance of neuron death (Coleman et al., 2004).

References:

Bertoni-Freddari C. Fattoretti P. Casoli T. Caselli U. Meier-Ruge W. Deterioration threshold of synaptic morphology in aging and senile dementia of Alzheimer's type. Analytical & Quantitative Cytology & Histology. 18(3):209-13, 1996. Abstract

Braak E. Braak H. Alzheimer's disease: transiently developing dendritic changes in pyramidal cells of sector CA1 of the Ammon's horn. Acta Neuropathologica. 93(4):323-5, 1997. Abstract

Buell SJ. Coleman PD. Dendritic growth in the aged human brain and failure of growth in senile dementia. Science. 206(4420):854-6, 1979. Abstract

Callahan LM. Vaules WA. Coleman PD. Progressive reduction of synaptophysin message in single neurons in Alzheimer disease. Journal of Neuropathology & Experimental Neurology. 61(5):384-95, 2002. Abstract

Coleman, P.D. Federoff HJ and Kurlan R Neurology, 2004, In Press.

DeKosky ST. Scheff SW. Synapse loss in frontal cortex biopsies in Alzheimer's disease: correlation with cognitive severity. Annals of Neurology. 27(5):457-64, 1990. Abstract

Eberwine J. Miyashiro K. Kacharmina JE. Job C. Local translation of classes of mRNAs that are targeted to neuronal dendrites. Proceedings of the National Academy of Sciences of the United States of America. 98(13):7080-5, 2001. Abstract

Job C. Eberwine J. Identification of sites for exponential translation in living dendrites. Proceedings of the National Academy of Sciences of the United States of America. 98(23):13037-42, 2001. Abstract

Morsch R. Simon W. Coleman PD. Neurons may live for decades with neurofibrillary tangles. Journal of Neuropathology & Experimental Neurology. 58(2):188-97, 1999. Abstract

Scheff SW. Price DA. Synaptic pathology in Alzheimer's disease: a review of ultrastructural studies. Neurobiology of Aging. 24(8):1029-46, 2003. Abstract

Terry RD. Masliah E. Salmon DP. Butters N. DeTeresa R. Hill R. Hansen LA. Katzman R. Physical basis of cognitive alterations in Alzheimer's disease: synapse loss is the major correlate of cognitive impairment. Annals of Neurology. 30(4):572-80, 1991. Abstract

Yao PJ. Zhu M. Pyun EI. Brooks AI. Therianos S. Meyers VE. Coleman PD. Defects in expression of genes related to synaptic vesicle trafficking in frontal cortex of Alzheimer's disease. Neurobiology of Disease. 12(2):97-109, 2003. Abstract

View all comments by Paul Coleman

Yong Shen's review of cell death in Alzheimer's disease (AD) poses a number of interesting questions about the role of cell death, particularly apoptotic cell death, in human neurodegenerative disease. Shen states "despite all else, it is still neuronal loss in the brain that leads to the major clinical symptoms of these disorders." This assessment is challenged in Paul Coleman's comments, who argues, and I concur, that the critical event in AD pathogenesis is synaptic dysfunction, not neuron loss. Several investigators had originally reported that caspase-dependent neuron apoptosis was prominent in the AD brain and that unregulated apoptosis caused many neurodegenerative diseases (reviewed in Roth, Caspases, apoptosis and Alzheimer's disease: causation, correlation and confusion. JNEN 2001; 60:829-838). This hypothesis has been challenged on several fronts and has largely been revised to the proposal that apoptosis-associated molecules, particularly caspases, cause AD through mechanisms that do not acutely lead to apoptosis. This is the view presented by Troy Rohn, who...  Read more

Yong Shen's review of cell death in Alzheimer's disease (AD) poses a number of interesting questions about the role of cell death, particularly apoptotic cell death, in human neurodegenerative disease. Shen states "despite all else, it is still neuronal loss in the brain that leads to the major clinical symptoms of these disorders." This assessment is challenged in Paul Coleman's comments, who argues, and I concur, that the critical event in AD pathogenesis is synaptic dysfunction, not neuron loss. Several investigators had originally reported that caspase-dependent neuron apoptosis was prominent in the AD brain and that unregulated apoptosis caused many neurodegenerative diseases (reviewed in Roth, Caspases, apoptosis and Alzheimer's disease: causation, correlation and confusion. JNEN 2001; 60:829-838). This hypothesis has been challenged on several fronts and has largely been revised to the proposal that apoptosis-associated molecules, particularly caspases, cause AD through mechanisms that do not acutely lead to apoptosis. This is the view presented by Troy Rohn, who proposes that caspases, through chronic low-level activation, cleave important cellular proteins that drive AD pathogenesis. If this hypothesis is correct, caspase inhibition should be a primary therapeutic objective in the pharmacological treatment of AD patients.

There are at least three major issues that require additional investigation before caspase inhibitors can be considered for the treatment of AD: i) a role for caspases in selective axonal or synaptic degeneration, independent of apoptotic death, has not been proven, ii) the specific caspase or caspases involved in AD are unknown, and iii) most importantly, concern that caspase inhibition will trigger other, non-apoptotic forms of neuron death.

It is clear from a variety of studies that caspases play physiological roles in processes other than cell death. Cytokine processing and cell cycle regulation (Woo et al., 2003) are two additional functions of some caspases, and the work of Rohn and others suggest that caspase-3 may be involved in tau and/or Aβ cleavage in the AD brain (Rissman et al., 2004). However, caspases have not yet been convincingly shown to be directly involved in AD-associated synaptic pathology. This will certainly be a prime objective of many laboratories in the future.

It will be critical to establish a link between specific caspases and AD pathogenesis. The difficulties involved with doing this, however, are substantial, as a brief review of caspase-12 and AD will illustrate. In rat and mouse cells, caspase-12 was found to be localized to endoplasmic reticulum (ER) and to regulate cell death in response to ER stress (Nakagawa et al., 2000). Using both in vivo and in vitro models, several investigators implicated caspase-12 in Aβ responsiveness and increased sensitivity to DNA damage in cells containing presenilin-1 mutations (Nakagawa et al., 2000, Chan et al., 2002). This led to the speculation that caspase-12 may be important in the pathogenesis of AD and that the human orthologue of caspase-12 may be a therapeutic target for the treatment of AD. Unfortunately, this speculation proved incorrect as the human caspase-12 was recently found to lack caspase catalytic activity, be uninvolved in ER stress responsiveness in human cells, and to be unassociated with AD (Saleh et al., 2004). Similarly, species and strain-specific effects of other caspases, including caspase-three are well-known (Leonard et al., 2002). Thus, extreme caution is required before extrapolating studies of rodent AD models to humans (Roth, 2001).

Finally, there is significant potential for caspase inhibitors to do more harm than good to AD patients. Under many experimental conditions, inhibition of caspases does not prevent death of cells exposed to apoptotic stimuli. Rather, caspase inhibition prevents the morphological changes that characterize apoptosis but not death per se (Roth et al., 2000). Non-apoptotic, autophagic cell death is observed in several neurodegenerative diseases and this type of cell death is unaffected by caspase inhibition (Zaidi et al., 2001). The complexities of cell death regulation are illustrated by studies of caspase-8. Caspase-8 is important for triggering apoptotic death through the extrinsic apoptotic death pathway and it has been implicated in AD (Roth, 2001). Thus, caspase-8 inhibition would at face value be therapeutically useful in the treatment of AD. Yet, a recent study indicates that caspase-8 inhibition leads directly to autophagic cell death (Yu et al., 2004). This observation led the authors to caution that "clinical therapies involving caspase inhibitors may arrest apoptosis but also have the unanticipated effect of promoting autophagic cell death" (page 1500, Yu et al., 2004). I feel that this warning is particularly relevant to AD, since anti-caspase therapy would necessarily be chronic and would not alleviate the primary stimulus or stimuli driving caspase activation. Under such circumstances, caspase inhibition therapy could convert dysfunctional neurons into dead neurons, eliminating all hope for clinical improvement. In total, a great deal has been learned over the last several years about the molecular regulation of caspase-dependent and -independent neuronal cell death and about Alzheimer's disease pathogenesis. However, my assessment of the field is similar to what it was three years ago when I wrote "an etiologic role for caspases in AD is far from proven" (page 829 Roth, 2001). Until such proof exists, treatment of AD patients with caspase inhibitors should proceed cautiously.

References:

Chan SL, Culmsee C, Haughey N, Klapper W, Mattson MP. Presenilin-1 mutations sensitize neurons to DNA damage-induced death by a mechanism involving perturbed calcium homeostasis and activation of calpains and caspase-12. Neurobiol.Dis. 11, 2-19. 2002. Abstract

Leonard JR, Klocke BJ, D'Sa C, Flavell RA, Roth KA. Strain-dependent neurodevelopmental abnormalities in caspase-3-deficient mice. J.Neuropathol.Exp.Neurol 61[8], 673-677. 2002. Abstract

Nakagawa T, Zhu H, Morishima N, Li E, Xu J, Yankner B A, Yuan J. Caspase-12 mediates endoplasmic-reticulum-specific apoptosis and cytotoxicity by amyloid-b. Nature 2000; (403): 98-103. Abstract

MP, LaFerla FM, Rohn TT, Cotman CW. Caspase-cleavage of tau is an early event in Alzheimer disease tangle pathology. J.Clin.Invest. 114[1], 121-130. 2004. Abstract

Roth K A. Caspases, apoptosis and Alzheimer's disease: causation, correlation and confusion. J Neuropathol Exp Neurol 2001; (60): 829-838. Abstract

Roth K A, Kuan C, Haydar T F, D'Sa-Eipper C, Shindler K S, Zheng T S, Kuida K, Flavell R A, Rakic P. Epistatic and independent functions of caspase-3 and Bcl-XL in developmental programmed cell death. Proc Natl Acad Sci USA 2000; (97): 466-471. Abstract

Saleh M, Vaillancourt JP, Graham RK, Huyck M, Srinivasula SM, Alnemri ES, Steinberg MH, Nolan V, Baldwin CT, Hotchkiss RS, Buchman TG, Zehnbauer BA, Hayden MR, Farrer LA, Roy S, Nicholson DW. Differential modulation of endotoxin responsiveness by human caspase-12 polymorphisms. Nature 429, 75-79. 2004. Abstract

Woo M, Hakem R, Furlonger C, Hakem A, Duncan GS, Sasaki T, Bouchard D, Lu L, Wu GE, Paige CJ, Mak TW. Caspase-3 regulates cell cycle in B cells: a consequence of substrate specificity. Nature Immunol 4[10], 1016-1022. 2003. Abstract

Yu L, Alva A, Su H, Dutt P, Freundt E, Welsh S, Baehrecke EH, Lenardo MJ. Regulation of an ATG7-beclin 1 program of autophagic cell death by caspase-8. Science 304, 1500-1502. 2004. Abstract

Zaidi AU, McDonough JS, Klocke BJ, Latham CB, Korsmeyer SJ, Flavell RA, Schmidt RE, Roth KA. Chloroquine-induced neuronal cell death is p53 and Bcl-2 family-dependent but caspase-independent. J.Neuropathol.Exp.Neurol. 60[10], 937-945. 2001. Abstract

View all comments by Kevin Roth

How does tau induce toxicity in neurons? Tau is highly hydrophilic, and normally binds to and stabilizes microtubules. In the disease state, tau is highly...  Read more

How does tau induce toxicity in neurons? Tau is highly hydrophilic, and normally binds to and stabilizes microtubules. In the disease state, tau is highly phosphorylated and turns into insoluble protein aggregates, resulting in some neurons dying. The activation of the tau kinases (GSK-3, JNK, and CDK5) generates hyperphosphorylated tau, which possibly triggers the cell death signal. This activation is triggered even when Aβ induces neuronal death, because cell death did not occur in Aβ-treated, tau-deficient neurons. Eliminating all other possibilities, the accumulation of tau in cytoplasm, formation of insoluble tau aggregate and hyperphosphorylated tau may either individually or in combination be required for Aβ to induce neuronal death. For example, the accumulation of tau in cytoplasm was reported to affect anterograde vesicle trafficking (Ebneth et al., 1998; Stamer et al., 2002; Trinczek et al., 1999). This may lead to a loss of synaptic function, and consequently induce neuronal death at a later stage.

Our recent study may be able to add yet another possibility for tau toxicity. We prepared N2a cells stably over-expressing P301L tau, and N2a cells over-expressing both P301L tau and CHIP, a ubiquitin ligase that induces ubiquitination of tau. When cells are treated with MG132, a protease inhibitor, they are marked for death. More tau over-expressing cells died than mock or CHIP-overexpressing cells. Thus, tau overexpression increases cell vulnerability to stress while CHIP overexpression protected cells from this vulnerability. Then we examined the SDS-insoluble tau in each of the cells. Tau overexpressing cells showed larger amounts of SDS-insoluble tau, and MG132 treatment enhanced the formation of SDS-insoluble tau, while CHIP over-expression dramatically reduced the formation of SDS-insoluble tau. SDS-insoluble tau increases cell vulnerability under stress, and tau accumulation in cytoplasm induces detergent-insoluble tau formation and vulnerability to insult.

References:
Ebneth, A., Godemann, R., Stamer, K., Illenberger, S., Trinczek, B., and Mandelkow, E. (1998). Overexpression of tau protein inhibits kinesin-dependent trafficking of vesicles, mitochondria, and endoplasmic reticulum: implications for Alzheimer's disease. J Cell Biol 143, 777-794. Abstract

Rapoport, M., Dawson, H. N., Binder, L. I., Vitek, M. P., and Ferreira, A. (2002). Tau is essential to beta -amyloid-induced neurotoxicity. Proc Natl Acad Sci U S A 99, 6364-6369. Abstract Hatakeyama, S., Matsumoto, M., Kamura, T., Murayama, M., Chui, D-H, Planel, E, Takahashi, R., Nakayama, K.I., and Takashima, A. (2004). U-box protein CHIP mediates poly-ubiquitilation on preferentially 4-repeat Tau and is involved in neurodegeneration of tauopathy. J. Neurochem (in press).

Stamer, K., Vogel, R., Thies, E., Mandelkow, E., and Mandelkow, E. M. (2002). Tau blocks traffic of organelles, neurofilaments, and APP vesicles in neurons and enhances oxidative stress. J Cell Biol 156, 1051-1063. Abstract

Trinczek, B., Ebneth, A., Mandelkow, E. M., and Mandelkow, E. (1999). Tau regulates the attachment/detachment but not the speed of motors in microtubule-dependent transport of single vesicles and organelles. J Cell Sci 112 ( Pt 14), 2355-2367. Abstract

View all comments by Akihiko Takashima

Then we...  Read more

Then we discuss who is the "murderer" of cognitive decline, synapse or neuron, or both? While we definitely appreciate the studies that show synapse loss and dysfunction are the critical events in the deficits in AD, there is also much solid evidence to prove that neuronal death or loss contribute to the cognitive deficits of AD (there are 557 references found by searching "neuron death and Alzheimer brains"). Recently, one study reported that early Alzheimer patients still can learn and memorize (reference). The interesting to me is that although synapse deficits occur early in AD brains, these patients still can learn and memorize. This can be interpreted that synapses have strong flexibility (or plasticity) even after injury. Alternatively, it also could be explained by the brain still functioning even after synapses are reduced. However, when neurons are lost (actually neurons are gradually lost, perhaps parallel to synapse deficits, not all lost at one time in the end stages of Alzheimer's disease), brain cognition starts to decline. It is not the other way around-that clinical symptoms start then neurons start to be lost. Thus, it is valuable and important to study mechanisms of neuron death and identify therapeutic target molecules that cause cell death. We may then have an opportunity to slow down the neurodegenerative process of Alzheimer's disease.

View all comments by Yong Shen