The effect of ATM knockdown on ionizing radiation-induced neuronal cell cycle reentry in Drosophila

Neuronal death in many neurodegenerative diseases is preceded by re-expression of cell cycle regulatory proteins and DNA synthesis.1 To illustrate, in post-mortem cerebella of individuals with Ataxiatelangiectasia (A-T), Purkinje neurons at risk for degeneration express proliferating cell nuclear antigen and Cyclin B, and in a mouse model of A-T, DNA replication accompanies expression of these proteins in at-risk neurons.2 A-T is a recessive genetic disease caused by mutation of the A-T mutated (ATM) gene.3 A link between cell cycle regulation and neurodegeneration in A-T is not unexpected because ATM regulates the cell cycle progression of cells that incur DNA damage.4 For example, in response to ionizing radiation (IR)-induced DNA damage, ATM promotes G1 phase cell cycle arrest through phosphorylation and stabilization of p53. Thus, ATM may function to maintain neurons in a quiescent or G0 state, and progressive neurodegeneration that occurs in the absence of ATM may result from cell cycle reentry triggered by the accumulation of DNA damage over time.

Our prior studies of a Drosophila model of A-T, in which ATM expression is knocked down by RNA interference (RNAi), indicate that cell cycle reentry not only correlates with neurodegeneration but also causes neurodegeneration.5 ATM knockdown results in reentry of post-mitotic neurons into S phase of the cell cycle prior to caspase-dependent apoptosis. Blocking apoptosis by expression of the baculovirus p35 caspase inhibitory protein results in accumulation of neurons in G₂/M phase of the cell cycle. Moreover, blocking cell cycle reentry by mutation of the cell cycle activator Cdc25/String inhibits neurodegeneration. These data predict the existence of ATM-regulated events that continuously function to block signals that promote cell cycle reentry of post-mitotic neurons.

A key feature of the Drosophila studies was the use of a method that enabled cell cycle analysis of post-mitotic neurons in vivo. Other studies of the role ATM plays in neurodegeneration have utilized fixed tissue or neurons cultured in vitro, which likely lack cell-cell signaling events that influence the process of cell cycle reentrymediated neurodegeneration in a whole animal. Indeed, in contrast to our finding that ATM knockdown induces cell cycle reentry of neurons in vivo, others have found that pharmacological inhibition of ATM activity does not induce cell cycle reentry of mammalian cortical neurons in vitro.5,6

The Drosophila method involves marking all post-mitotic neurons in flies by expression of green fluorescent protein (GFP) under control of the embryonic lethal, abnormal vision (elav) gene transcriptional regulatory sequences.5,7 In the third instar larval eye imaginal disc, the developmental precursor to the adult eye, elav expression marks cells that have exited the cell cycle and differentiated into photoreceptor neurons.8,9 Fluorescence activated cell sorting (FACS) is then used to determine the DNA content and, thus, the cell cycle status of individual GFP-positive cells produced by dissociation of eye imaginal discs with trypsin. Here we have used this method to test the hypothesis that DNA damage causes cell cycle reentry of post-mitotic neurons and to determine the extent to which ATM knockdown affects DNA damage-induced cell cycle reentry.

Fly larvae were either untreated or irradiated with 50 Gy of gamma rays and allowed to recover for either 0.5–2.5 h or 21 h prior to eye imaginal disc dissection and FACS analysis. Two fly genotypes were analyzed. Flies designated elav-GFP expressed GFP in post-mitotic neurons by means of the GAL4-UAS system; expression of the GAL4 transcription factor from an elav-GAL4 transgene drove expression of a UAS-GFP transgene in post-mitotic neurons. Flies designated elav-GFP-ATMi expressed GFP in the same manner and had reduced ATM expression in post-mitotic neurons due to a UAS-ATMi transgene that expressed a short hairpin RNA that targeted ATM mRNA for degradation by RNAi. For each sample, ∼20 pairs of eye imaginal discs were dissected from wandering third instar larvae, and the DNA content of 5,000 live, GFP-positive, single cells was determined using a Becton Dickinson LSRII flow cytometer.

Consistent with our prior study, a substantial percentage (Ȭ28%) of neurons in untreated elav-GFP eye imaginal discs were cycling (i.e., they were either in S or G₂/M phase) (Table 1). Treatment with IR followed by 0.5–2.5 h of recovery resulted in a statistically significant 11.9% increase in cycling neurons (p < 0.001), most of which were resident in S phase. Recovery for 21 h did not result in a further increase in cycling neurons (relative to 0.5–2.5 h of recovery); however, it did decrease the percentage of S phase neurons to that of untreated larvae and increase the percentage of G₂/M phase neurons, suggesting that neurons continued to progress through the cell cycle during this recovery period. Thus, in the context of eye imaginal discs, IR-induced DNA damage can cause post-mitotic neurons to reenter the cell cycle and undergo DNA replication.

A parallel study of elav-GFP-ATMi eye imaginal discs revealed that treatment with IR and recovery for 0.5–2.5 h did not significantly increase the percentage of cycling neurons over the Ȭ14.8% increase caused by ATM knockdown (p > 0.05) (Table 1). However, recovery for 21 h resulted in a statistically significant Ȭ17.6% increase in cycling neurons (p < 0.001), which had progressed into G₂/M phase. Thus, ATM knockdown attenuates IR-induced neuron cell cycle reentry shortly (0.5–2.5 h) after IR treatment but not after a longer period of time (21 h). Attenuation of DNA damage-induced cell cycle reentry by ATM inhibition was also observed in a study of cortical neurons in vitro; however, in this case, subsequent cell cycle reentry was not observed.6 This difference may be due to environmental context (in vivo versus in vitro) or one of many differences between the experimental protocols.

In summary, IR-induced DNA damage is sufficient to cause post-mitotic neurons in vivo to reenter the cell cycle, and loss of ATM delays but does not block IR-induced neuronal cell cycle reentry. In other words, IR-induced DNA damage can cause neuronal cell cycle reentry through ATM-dependent and ATM-independent mechanisms. The demonstration of an ATM-independent mechanism provides support for the hypothesis that progressive neurodegeneration in A-T results from cell cycle reentry triggered by the accumulation of DNA damage over time.

Abbreviations

A-T	=	Ataxia-telangiectasia
ATM	=	A-T mutated
elav	=	embryonic lethal, abnormal vision
FACS	=	fluorescence activated cell sorting
GFP	=	green fluorescent protein
RNAi	=	RNA interference

Figures and Tables

Table 1 Cell cycle profiles of wild type and ATM knockdown neurons in response to IR

* Data were analyzed using CellQuest (Becton Dickinson) and ModFit (Verify Software House) software. Indicated are the average percentage of photoreceptor neurons in each phase of the cell cycle and the standard error of the mean. p-values were calculated using one-way ANOVA and Tukey's mutiple comparison post-test.

Acknowledgements

We thank K. Schell and the UW-Madison Comprehensive Cancer Center FACS facility for assistance with flow cytometry and R. Tibbetts for insights that greatly improved this work. Funding support was provided by a grant from the NIH (R01 NS059001 to D.A.W.) and a pre-doctoral fellowship from NIH training grant T32 GM08688 (to A.J.P.).