Evidence exists that cell death seen in AD arises from neurons attempting to divide, and this is why researchers working in the field believe that AD therapy might benefit from drugs able to arrest cell division in neurons. Until this paper by Varvel and colleagues, however, that hypothesis had not been truly tested. Now, this elegant paper shows that non-steroidal anti-inflammatory drugs (NSAIDs) prevent the emergence of cycling neurons in a transgenic (Tg) mouse model that mimics the alterations of the neuronal cell cycle found in the human AD brain.

This paper is informative in many ways. First, it hints at activated microglia cells as being essential players in the initiation of the neuronal cycle. Second, it indicates that microglia actvation is relevant to the generation of cycling neurons only in the presence of β amyloid. Third, it shows that NSAIDs prevent cell cycle initiation through the blockade of inflammation, but the drugs are not able to reverse the neuronal cycle once it is initiated. There is some good and bad in this demonstration. The bad is that NSAIDs...  Read more

Evidence exists that cell death seen in AD arises from neurons attempting to divide, and this is why researchers working in the field believe that AD therapy might benefit from drugs able to arrest cell division in neurons. Until this paper by Varvel and colleagues, however, that hypothesis had not been truly tested. Now, this elegant paper shows that non-steroidal anti-inflammatory drugs (NSAIDs) prevent the emergence of cycling neurons in a transgenic (Tg) mouse model that mimics the alterations of the neuronal cell cycle found in the human AD brain.

This paper is informative in many ways. First, it hints at activated microglia cells as being essential players in the initiation of the neuronal cycle. Second, it indicates that microglia actvation is relevant to the generation of cycling neurons only in the presence of β amyloid. Third, it shows that NSAIDs prevent cell cycle initiation through the blockade of inflammation, but the drugs are not able to reverse the neuronal cycle once it is initiated. There is some good and bad in this demonstration. The bad is that NSAIDs do not seem valuable in treating subjects diagnosed with AD. The drugs might be potentially useful in preventing/delaying the onset of the disease, but we do not have definitive answers: 1) the Alzheimer's Disease Anti-inflammatory Prevention Trial (ADAPT) had to be stopped ahead of time because of increased risk of cardiovascular and cerebrovascular events (PLoS Clin Trials, 2006); 2) evidence that NSAIDs reduce behavioral impairments in AD Tg mice is very thin (Lim et al., 2001) and, in same cases, it is unrelated to the anti-inflammatory properties of the drugs (Kukar et al., 2007). On the good side of the equation, we have a possible explanation for the disappointing results of clinical trials with NSAIDs in subjects with AD (Aisen et al., 2003). More important, from a research perspective, we might have a tool to monitor the effect of emerging drugs on the initiation and progression of AD-like neuropathology.

Karl Herrup’s lab has provided very convincing evidence that cell cycle events represent a biomarker for the risk of neurodegeneration in AD (Yang et al., 2003). However, because AD Tg mice do not develop frank neuronal loss, at the end we will need to assess how the neuronal cycle is related to the behavioral phenotype of Tg mice. The neuronal cycle is unique in the sense that it may last years. Currently, we are not able to predict to which extent we have to halt the cycle (do neurons need to re-enter quiescence or is it enough to block DNA replication?) to maintain/restore neuronal function. I believe that the upcoming goal is to see whether we can improve the cognitive impairment of AD Tg mice by using drugs that halt the neuronal cycle. The paper by Varvel and colleagues is a significant step toward this goal.

View all comments by Agata Copani

Cell cycle changes can be detected in neurons that are vulnerable for developing neurodegenerative changes that are associated with Alzheimer disease (AD) pathology. The paper by Varvel and colleagues is very interesting, showing the potential of non-steroidal anti-inflammatory drugs (NSAIDs) to prevent neuronal cell cycle changes which can be found in the human AD brain.

Most NSAIDs inhibit the activity of cyclooxygenase (COX). Both isoforms, COX-1 and COX-2, are expressed in the brain with differences in cellular localization. Expression studies of COX-1 and COX-2 in AD brain have shown changes compared to non-demented control brains, suggesting a role for COX-1 and COX-2 in AD pathology. In their paper, Varvel and colleagues primarily focus on the effect of NSAID treatment on microglia activation, while no attention is paid to neuronal COX-2, whose expression is increased in the early stages of AD pathology (Ho et al., 2001; Hoozemans et al., 2001; Yermakova and O'Banion, 2001).

The expression of COX-2 in numerous types of cancer and the effect of selective COX-2...  Read more

Cell cycle changes can be detected in neurons that are vulnerable for developing neurodegenerative changes that are associated with Alzheimer disease (AD) pathology. The paper by Varvel and colleagues is very interesting, showing the potential of non-steroidal anti-inflammatory drugs (NSAIDs) to prevent neuronal cell cycle changes which can be found in the human AD brain.

Most NSAIDs inhibit the activity of cyclooxygenase (COX). Both isoforms, COX-1 and COX-2, are expressed in the brain with differences in cellular localization. Expression studies of COX-1 and COX-2 in AD brain have shown changes compared to non-demented control brains, suggesting a role for COX-1 and COX-2 in AD pathology. In their paper, Varvel and colleagues primarily focus on the effect of NSAID treatment on microglia activation, while no attention is paid to neuronal COX-2, whose expression is increased in the early stages of AD pathology (Ho et al., 2001; Hoozemans et al., 2001; Yermakova and O'Banion, 2001).

The expression of COX-2 in numerous types of cancer and the effect of selective COX-2 inhibitors on tumor growth suggest a role for COX-2 in cell cycle regulation. Interestingly, COX-2 co-localizes with cell cycle proteins in AD neurons (Hoozemans et al., 2002; Mirjany et al., 2002; Hoozemans et al., 2004). Functional studies show that increased neuronal COX-2 expression leads to increased expression of cell cycle mediators in post-mitotic neurons, as shown in a transgenic mouse model with increased neuronal COX-2 expression (Xiang et al., 2002). In addition, COX-2 is required for cyclin D1 expression in neurons after ischemia in vivo and anoxia in vitro (Wu Chen et al., 2004). So while COX-2 and cell cycle proteins co-localize in AD neurons, COX-2 is also actively involved in the regulation of cell cycle control in these cells.

In their study, Varvel and colleagues nicely show that NSAID treatment prevents neuronal cell cycle protein expression by reducing microglia activation. Although it does not exclude that microglia or inflammation are involved in inducing increased expression of cell cycle proteins in neurons, clinico-pathological data indicate that neuronal cell cycle protein expression precedes the widespread activation of microglia in AD brain (Hoozemans et al., 2005). Alternatively, in the human brain, NSAIDs could directly affect cell cycle protein expression in neurons by selective inhibition of neuronal COX-2. Either way, there is increasing evidence that regular use of NSAIDs can lower risk of developing AD by preventing aberrant cell cycle protein expression in neurons.

References:
Ho, L., Purohit, D., Haroutunian, V., Luterman, J. D., Willis, F., Naslund, J., Buxbaum, J. D., Mohs, R. C., Aisen, P. S., and Pasinetti, G. M., 2001. Neuronal cyclooxygenase 2 expression in the hippocampal formation as a function of the clinical progression of Alzheimer disease. Arch Neurol 58, 487-92. Abstract

Hoozemans, J. J., Bruckner, M. K., Rozemuller, A. J., Veerhuis, R., Eikelenboom, P., and Arendt, T., 2002. Cyclin D1 and cyclin E are co-localized with cyclo-oxygenase 2 (COX-2) in pyramidal neurons in Alzheimer disease temporal cortex. J.Neuropathol.Exp.Neurol. 61, 678-688. Abstract

Hoozemans, J. J., Rozemuller, A. J., Janssen, I., de Groot, C. J., Veerhuis, R., and Eikelenboom, P., 2001. Cyclooxygenase expression in microglia and neurons in Alzheimer's disease and control brain. Acta Neuropathol (Berl) 101, 2-8. Abstract

Hoozemans, J. J., Van Haastert, E. S., Veerhuis, R., Arendt, T., Scheper, W., Eikelenboom, P., and Rozemuller, A. J., 2005. Maximal COX-2 and ppRb expression in neurons occurs during early Braak stages prior to the maximal activation of astrocytes and microglia in Alzheimer's disease. J Neuroinflammation 2, 27. Abstract

Hoozemans, J. J., Veerhuis, R., Rozemuller, A. J., Arendt, T., and Eikelenboom, P., 2004. Neuronal COX-2 expression and phosphorylation of pRb precede p38 MAPK activation and neurofibrillary changes in AD temporal cortex. Neurobiol.Dis. 15, 492-499. Abstract

Mirjany, M., Ho, L., and Pasinetti, G. M., 2002. Role of cyclooxygenase-2 in neuronal cell cycle activity and glutamate-mediated excitotoxicity. J Pharmacol Exp Ther 301, 494-500. Abstract

Wu Chen, R., Zhang, Y., Rose, M. E., and Graham, S. H., 2004. Cyclooxygenase-2 activity contributes to neuronal expression of cyclin D1 after anoxia/ischemia in vitro and in vivo. Brain Res Mol Brain Res 132, 31-7. Abstract

Xiang, Z., Ho, L., Valdellon, J., Borchelt, D., Kelley, K., Spielman, L., Aisen, P. S., and Pasinetti, G. M., 2002. Cyclooxygenase (COX)-2 and cell cycle activity in a transgenic mouse model of Alzheimer's disease neuropathology. Neurobiol Aging 23, 327-34. Abstract

Yermakova, A. V. and O'Banion, M. K., 2001. Downregulation of neuronal cyclooxygenase-2 expression in end stage Alzheimer's disease. Neurobiol Aging 22, 823-36. Abstract

View all comments by Jeroen Hoozemans

I would like to take the opportunity to clarify a few points in the review pertaining to our work published in 2008. Over the past few years our goal has been to determine the insult(s) responsible for the induction of neuronal cell cycle events (CCEs) in mouse models of AD. Specifically, we explored the involvement of Aβ in the induction of CCEs. To accomplish this we analyzed CCEs in two different mouse models of AD and compared our findings to those obtained from R1.40 mice maintained on the C57BL/6J genetic background (B6-R1.40). First, R1.40 animals maintained on the DBA/2J (D2-R1.40) genetic background first exhibit neuronal CCEs at 12 months of age, six months after CCEs are first encountered in B6-R1.40 animals. While both B6-R1.40 and D2-R1.40 mice exhibit similar levels of both holo-APP and APP CTFs, the steady-state levels of Aβ are substantially reduced in the D2-R1.40 animals. These data indicate that reductions in Aβ levels delay the induction of CCEs. Second, B6-R1.40 mice deficient for Bace1 (B6-R1.40;Bace1-/-) exhibit no evidence of CTFβ and fail to develop...  Read more

I would like to take the opportunity to clarify a few points in the review pertaining to our work published in 2008. Over the past few years our goal has been to determine the insult(s) responsible for the induction of neuronal cell cycle events (CCEs) in mouse models of AD. Specifically, we explored the involvement of Aβ in the induction of CCEs. To accomplish this we analyzed CCEs in two different mouse models of AD and compared our findings to those obtained from R1.40 mice maintained on the C57BL/6J genetic background (B6-R1.40). First, R1.40 animals maintained on the DBA/2J (D2-R1.40) genetic background first exhibit neuronal CCEs at 12 months of age, six months after CCEs are first encountered in B6-R1.40 animals. While both B6-R1.40 and D2-R1.40 mice exhibit similar levels of both holo-APP and APP CTFs, the steady-state levels of Aβ are substantially reduced in the D2-R1.40 animals. These data indicate that reductions in Aβ levels delay the induction of CCEs. Second, B6-R1.40 mice deficient for Bace1 (B6-R1.40;Bace1-/-) exhibit no evidence of CTFβ and fail to develop neuronal CCEs. These data indicate that CCEs are dependent on β-secretase activity and not γ-secretase activity as stated in the posted article. Genetic removal of Bace1 results in the inability to produce Aβ as well as CTFβ. Therefore, we cannot rule out CTFβ as the Bace1-dependent APP processing product that induces CCEs. However, D2-R1.40 exhibit similar steady-states levels of CTFβ when compared to B6-R1.40 animals. Thus, we concluded that Aβ must be involved in the induction of neuronal CCEs.

References:
Varvel et. al. (2008) Abeta oligomers induce neuronal cell cycle events in Alzheimer's disease. Journal of Neuroscience. 28(43):10786-10793. Abstract

View all comments by Nicholas H. Varvel