|Home | About | Journals | Submit | Contact Us | Français|
Developmental and pathological death of neurons requires activation of a defined pathway of cell cycle proteins. However, it is unclear how this pathway is regulated and whether it is relevant in vivo. A screen for transcripts robustly induced in cultured neurons by DNA damage identified Sertad1, a cyclin dependent kinase 4 (Cdk4) activator. Sertad1 is also induced in neurons by nerve growth factor (NGF) deprivation and β-amyloid (Aβ). RNAi-mediated downregulation of Sertad1 protects neurons in all three death models. Studies of NGF withdrawal indicate that Sertad1 is required to initiate the apoptotic cell cycle pathway since its knockdown blocks subsequent pathway events. Finally, we find that Sertad1 expression is required for developmental neuronal death in the cerebral cortex. Sertad1 thus appears to be essential for neuron death in trophic support deprivation in vitro and in vivo and in models of DNA damage and Alzheimer’s disease. It may therefore be a suitable target for therapeutic intervention.
Neuronal loss by apoptosis is a physiological process during development (Oppenheim, 1991) and a pathological hallmark of many neurodegenerative disorders such as Alzheimer’s disease (AD) and of additional insults to the nervous system such as DNA damage (Park et al., 1997a). There are striking and mutually informative similarities between the molecular mechanisms that govern neuron death under these various conditions (Greene et al., 2004; Greene et al., 2007). However, the molecular events, particularly those that initiate death of neurons during development and disease/injury are incompletely understood.
One major focus regarding the mechanisms of developmental and disease-associated neuron death has been the aberrant activation of cell cycle-related proteins (Becker and Bonni, 2004; Greene et al., 2004; Herrup et al., 2004; Greene et al., 2007). Past studies have indicated a sequential and multi-step pathway that is activated by various apoptotic insults including NGF deprivation, DNA damage, and Aβ exposure and that is required for neuron death. The first described step is rapid activation of the G1/S kinase Cdk4. This in turn hyperphosphorylates members of the Rb family, leading to dissociation of complexes comprised of Rb family members and E2F transcription factors. Ultimately, these events lead to induction of pro-apoptotic genes such as Bim and to activation of the core apoptotic machinery (Greene et al., 2007).
An important and currently unresolved issue about the apoptotic cell cycle pathway is how Cdk4 is activated in neurons by apoptotic stimuli. Understanding this will not only further elaborate how apoptotic stimuli lead to neuron death, but may also identify additional molecular targets for therapeutic intervention. In addition, it must be recognized that the majority of evidence that causally links the steps in the apoptotic cell cycle pathway has been generated by in vitro studies. Thus, it remains important to demonstrate that the elements that make up this pathway are relevant in vivo.
The protein Sertad1, also known as p34(SEI-1) or Trip-Br1, has been implicated as a regulator of Cdk4 activity. Sertad1 was first identified as an antagonist of p16 INK4a that facilitates the formation and activation of cyclin D-Cdk4 complexes (Sugimoto et al., 1999). Further studies revealed that it directly binds and activates Cdk4 in a concentration dependent manner (Li et al., 2004). Functions in addition to regulation of Cdk4 have been described for Sertad1 including stimulation of the transcriptional activities of E2F1 (Hsu et al., 2001) and p53 (Watanabe-Fukunaga et al., 2005). Sertad1 was also additionally reported to exhibit anti-apoptotic activity by stabilizing XIAP in cancer cells (Hong et al., 2009).
In a screen for genes regulated in neurons following DNA damage, we identified Sertad1 transcripts as being robustly induced. Accordingly, we have examined the potential role of Sertad1 in neuron death induced by apoptotic stimuli relevant both to normal development and to neurodegeneration. We find that Sertad1 is required for neuronal apoptosis both in vitro and in vivo. Furthermore, our findings indicate that ertad1 is essential for initiating the Cdk4-dependent cascade of cell cycle events in neuronal cells that follow from trophic factor deprivation.
Platinum Taq DNA polymerase, V5 antibody, GFP antibody and Lipofectamine 2000 were from Invitrogen (Carlsbad, CA), anti-human NGF antiserum and anti-β-actin antibody were from Sigma (St. Louis, MO), anti-Sertad1 antibody and Bim antibody were from Abcam (Cambridge, MA), anti-ERK1, phospho-Rb and C-Myb antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA), Zsgreen antibody was from Clontech (Mountain View, CA) and the In Situ Cell Death Detection Kit, TMR red was from Roche Applied Science (Indianapolis, IN). pSIREN vector was from BD Biosciences (San Jose, CA). Human recombinant NGF was a kind gift from Genentech (South San Francisco, CA). Camptothecin was obtained from Sigma (St. Louis, MO). CEP11004 was obtained from Cephalon (Frazer, PA). E2F-1 and control siRNAs and E2F-1 antibody were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). p53 mice were genotyped according to published protocols (Aleyasin, et al, 2004).
PC12 cells were cultured and neuronally differentiated as previously described (Greene and Tischler, 1976). For NGF deprivation, after a week of NGF treatment the cultures were washed with NGF-free medium twice and anti-NGF antibody (1:100) was added. Control cells were washed with serum-free medium and maintained in RPMI 1640 medium supplied with NGF without serum. Neonatal rat superior cervical ganglion sympathetic (SCG) neurons were cultured as previously described (Park et al., 1998). HEK293 cells were cultured in DMEM with 10% fetal bovine serum. Embryonic rat and mouse cortical neurons were cultured as previously described (Park et al., 1998).
Total RNA was extracted from cortical neuron cultures using Trizol reagent according to the manufacturer’s instructions (Invitrogen, Carlsbad, CA). RNA was sent to the Ottawa Genomics Innovation Centre Microarray Facility for processing and expression analysis using the Affymetrix Mouse 430 array (Affymetrix, Santa Clara, CA). Probe signals were scaled and normalized according to standard facility procedures.
Total RNA was extracted using TriPure isolation reagent (Roche Applied Science, Indianapolis, IN). 50 ng of total RNA were used for cDNA synthesis and gene amplification reactions using SuperScript One-Step RT-PCR kit (Invitrogen, Carlsbad, CA). cDNA synthesis was performed at 48°C for 45 min, followed by a 2 min initial denaturation step at 94°C. This was followed by 30 cycles (Sertad1) or 25 cycles (S12) at 94°C for 30 s, melting temperature (Tm) 60°C for 30s, and 72°C for 1 min. Targeting primers were as follows: 5′-CGCAAGCGGGAGGAGGAGAC-3′ and 5′-AGGGGCTGGGGGCTGGATGG-3′ for Sertad1, 5′-GGAAGGCATAGCTGCTGG-3′ and 5′-CCTCGATGACATCCTTGG-3′ for S12. Transcript levels were normalized against S12 signals and results were reported as times fold increase in reference to untreated control values. Data are presented as mean ± SEM of three independent experiments.
Each sample of total RNA was isolated from cultured neurons by using TRI reagent (Molecular Research Center, Cincinnati, OH). cDNA was transcribed from total RNA with Superscript RT II (Invitrogen, Carlsbad, CA). The primers used for PCR amplification of rat Sertad1 were 5′-GCCTCCTGGAAGATCTCAGTC-3′ and 5′-CATTCTCAGGGACAGGTTTGA-3′. The primers for α-tubulin were: 5′-ATGAGGCCATCTATGACATC-3′ and 5′-TCCACAAACTGGATGGTAC-3′. Equal amounts of cDNA template were used for each PCR analysis of Sertad1 or α-tubulin. Quantitative PCR was performed using a Cepheid (Sunnyvale, CA) SmartCycler following the manufacturer’s specifications. α-tubulin was used for Sertad1 transcript normalization. cDNA was added to a 25 μl volume reaction mix containing OmniMix HS master mix (Cepheid, Sunnyvale, CA) and SYBR Green I (Invitrogen, Carlsbad, CA) together with appropriate primers at 0.2 μM each. Analyses of growth curves of real-time fluorescence and of melting curves were performed as described previously (Troy et al., 2000).
Neuronal PC12 cells were lysed and protein was analyzed by Western immunoblotting as described previously (Biswas and Greene, 2002). For mouse cortical neurons, Sertad1 was detected using a chicken IgY antibody against Sertad1 (1:1000; Genway, San Diego, CA). Goat-anti-chicken HRP (1:3000) was used as secondary antibody.
Rat Sertad1 was generated by RT-PCR of PC12 cDNA. The primers for the amplification were 5′-AGGATGCTGAGCAAAGGTCT-3′ and 5′-GCGCCCAGGTCCTGGTGGCC-3′. The PCR product was gel purified and cloned into pCDNA3.1 vector (Invitrogen, Carlsbad, CA), then verified by sequencing. Sertad1 was also subcloned into pCMS-EGFP vector(Clontech, Mountain View, CA) by using primers 5′-GATCTCGAGACCATGCTGAGCAAAGGTCTG-3′ and 5′-CTAGTCGACCTAGCGCCCAGGTCCTGGTGG-3′.
Sertad1 shRNAs were prepared in the pSIREN vector by using BD™ Knockout RNAi systems according to the manufacturer’s instructions (BD Biosciences, San Jose, CA). based on the sequences: 5′-CCGTGGCTTCTAGCTCTCT-3′ (#2), 5′-GCTCCACCACAGCCTTCGG-3′ (#3), 5′-CCAGACCTCCGACACCTGG-3′ (#4), 5′-GATCTCAGTCATATTGAGG-3′ (#5). pSIREN-shRNA-RAND-Zsgreen was as described previously (Sproul, 2009). For in utero electroporation (see below), GFP constructs of Sertad1 shRNA and control shRNA were prepared by subcloning the shRNA expression cassette from pSIREN vector into pCMS-EGFP backbone sequence. The (CMV promoter)-MCS sequence in pCMS-EGFP was substituted with the (U6 promoter-shRNA) sequence from pSIREN-RetroQ-zsGreen by subcloning with BglII and EcoRI restriction enzymes. The control shRNA is an inactive mutant of the primary siRNA knockdown construct for GATA2: 5′-GCACCTGATGTCTTCTTCAACC-3′.
DNA was prepared with Plasmid Maxi kits (Qiagen, Valencia, CA). Neuronal PC12 cells were co-transfected with 0.5 μg of plasmid pCDNA-V5, pCDNA-Sertad1-V5, pCMS-EGFP, pCMS-Sertad1-EGFP, pSIREN-shRNA-Sertad1-Zsgreen (#2, #3, #4, or #5), pSIREN-shRNA-Rand-Zsgreen, or pSIREN-shRNA-Luc-Zsgreen in 500 μl serum free medium per well in 24-well dishes using Lipofectamine 2000. 6 hrs later, medium with Lipofectamine 2000 was replaced with fresh complete medium. HEK293 cells were transfected as previously described (Xu et al., 2001). E2F-1 siRNA were transfected as previously described (Zhang et al., 2009).
Mouse Sertad1 specific shRNA oligos (Applied Biosystems Inc, Foster City, CA) were cloned into pSilencer3.0-H1 vector (Applied Biosystems Inc, Foster City, CA). The shRNA fragments containing the H1 promoter were subcloned into the pAdTrack vector. shRNA adenoviruses and DN-c-Jun adenoviruses were constructed as previously described (He et al., 1998).
Lyophilized, HPLC-purified Aβ1-42 was purchased from American Peptide (Sunnyvale, CA) and dodecamer Aβ1-42 was prepared as described previously (Barghorn et al., 2005). Briefly, Aβ1-42 was reconstituted in 100% 1,1,1,3,3,3 hexafluoro-2-propanol (HFIP) to 1 mM, HFIP was removed by evaporation in a Speed Vac, then resuspended to 5 mM in anhydrous DMSO. This stock was then diluted with PBS to a final concentration of 400 μM. SDS was added to a final concentration of 0.2% and the resulting solution was incubated at 37°C for 18–24 hrs. The preparation was diluted again with PBS to a final concentration of 100 μM and incubated at 37°C for 18–24 hrs.
Sprague Dawley rats (Charles River Laboratories, Wilmington, MA) were housed, cared for and electroporated under the guidelines established by Columbia University Medical Center Institutional Animal Care and Use Committee. Timed pregnant rats [embryonic day 16 (E16)] wereanesthetized with ketamine/xylazine (100/10 mixture, 0.1 mg/g body weight, i.p.). The uterine horns were exposed, and the left lateral ventricles of embryos were injected with DNA constructs of a control shRNA or Sertad1 shRNA (1–5 μg/μl) and FastGreen (2 mg/ml; Sigma, St. Louis, MO) by using pulled glass capillaries (Sutter Instrument Co, Novato, CA). Electroporation was accomplishedwith a BTX electro square electroporator, model ECM830 (BTX, Holliston, MA). Each embryo’s head was held between tweezer-type circular electrodes (Harvard Apparatus, Holliston, MA)across the uterus wall and five electrical pulses (amplitude,50 V; duration, 50 ms; intervals, 100 ms) were delivered. Brains from postnatal pups of 5 day old were fixed in4% paraformaldehyde (PFA) by cardiac perfusion.
Rat pups were anesthetized and perfused transcardially with 4% PFA in PBS, pH 7.4. After perfusion, brains were dissected out from the skull and postfixed overnight in fresh fixative. Then brains were washed with PBS, pH 7.4 and cryoprotected in 30% sucrose phosphate buffer. They were frozen with O.C.T. and dry ice and sectioned (20-μm-thick) in the coronal plane with a cryostat. To visualize nuclei with DNA cleavage, residues of fluorescein-labeled nucleotides were catalytically added to DNA fragments by terminal deoxy-nucleotidyl-transferase (TdT). Briefly, sections were fixed in fresh 4% PFA/PBS at room temperature for 20 min, washed in PBS three times for 5 min, permeabilized with proteinase K for 5 min on ice and incubated with nucleotide mix and TdT (In Situ Cell Death Detection Kit, TMR red) at 37°C for 1 hr. Apoptotic cells exhibit strong, nuclear red fluorescence. To visualize GFP positive cells, sections were then immunostained with rabbit anti GFP antibody (1:1000, Invitrogen, Carlsbad, CA) in 3% nonimmune goat serum overnight at 4°C, followed by secondary labeling with goat anti-rabbit antibody (1:4000, Alexa Fluor 488, Molecular Probes, CA) for 1 hr.
Neuronal PC12 cells, sympathetic neurons or cortical neurons were transfected with either pCMS-Sertad1-EGFP, pCMS-EGFP, pSIREN-Sertad1-shRNA, pSIREN-Luc-shRNA, or a Random pSIREN-ZsGreen, and then 48 hrs later deprived of NGF (in case of neuronal PC12 cells and sympathetic neurons) or treated with 1.25 μM dodecamer Aβ (in case of cortical neurons). The numbers of surviving transfected (green) cells per well were assessed just after treatment and at 24 and 48 hrs after NGF deprivation or Aβ exposure as described previously (Biswas et al., 2007). Data represent means ± SEM of three experiments performed in triplicate.
Neuronal PC12 cells were transfected as described above with appropriate constructs of shRNA. 48 hrs later, cells were subjected to NGF withdrawal for 18 hrs and then immunostained as described in Angelastro et al. (2003). Briefly, PC12 cells were fixed with 4% paraformaldehyde for 10 min. After three washes with PBS, cells were blocked in 3% non-immune goat serum for 2 hrs. The cultures were immunolabeled with rabbit anti-Bim (1:1000; Abcam) antibody, Rabbit C-Myb antibody (1:500, Santa Cruz Biotechnology, CA) or Rabbit p-Rb antibody (1:100, Santa Cruz Biotechnology, CA) in 3% nonimmune goat serum overnight at 4°C, followed by secondary labeling with goat anti-rabbit antibody (1:1000, Alexa Fluor 568, Molecular Probes, CA) for 1 hr. For Sertad1 knockdown experiment, the cultures were immunostained with anti-Sertad1 antibody (1:200; Genway) followed by secondary labeling with goat anti-chicken antibody (1:500, Molecular Probes, CA) for 1 hr.
In an initial microarray screen for genes induced in cortical neurons by treatment with the DNA damaging agent camptothecin (8 hrs), we identified Sertad1 as a candidate regulated gene. Induction was 46-fold over controls (Table 1). In the same screen, we detected increases in the Puma and Noxa genes as expected and previously reported (Aleyasin et al., 2004; Cregan et al., 2004). In contrast, β-tubulin did not change (Table 1). To verify the array data, we performed a semiquantitative reverse transcriptase PCR assay. Consistent with the array data, the RT-PCR results demonstrated a robust (~27-fold) increase of Sertad1 message as early as 2 hrs after camptothecin exposure. This increase persisted even at 8 hrs after camptothecin treatment (Figure 1A). A similar magnitude of increase was obtained by quantitative PCR (Figure S1). Next, we asked whether the DNA damage-induced elevation in Sertad1 message is associated with an increase in protein level. As revealed by western blotting (Figure 1B), Sertad1 protein expression was significantly induced by 6 hrs (~3.2 fold) of camptothecin exposure and gradually increased until 18 hrs (~7.2 fold). Under the conditions of our experiments, neuronal death first becomes apparent by 8–10 hrs of camptothecin exposure and 50% of neurons die within 16–20 hrs. Thus, Sertad1 induction is observed early and prior to overt signs of death.
To identify upstream signals that mediate upregulation of Sertad1 mRNA in response to DNA damage, we tested three potential candidates: p53, E2F1 and JNKs. As shown in Figure 2A, germ line deficiency of p53 does not affect camptothecin-induced upregulation of Sertad1 message in cultured cortical neurons. In contrast, siRNA-mediated knockdown of E2F-1 blocks Sertad1 mRNA upregulation by about 50% (Figure 2B). Finally, neither JNK inhibitor CEP11004 nor dominant-negative c-Jun expression affect camptothecin-induced Sertad1 upregulation (Figure S2). Taken together, these findings indicate that E2F1, but not p53 or the JNK pathway at least partially regulates Sertad mRNA expression following DNA damage
Because neuronal death caused by DNA damage shares with death evoked by trophic factor deprivation a reliance on cell cycle proteins (Park et al., 1996; Park et al., 1997b; 1997a; Liu and Greene, 2001; Zhang et al., 2006; Gonzalez et al., 2008), we also investigated whether Sertad1 expression is regulated in response to NGF withdrawal. For this purpose we used neuronally differentiated PC12 cells and primary cultures of superior sympathetic ganglion (SCG) neurons. Both cell types undergo apoptosis (evident starting at about 18 hrs) in response to NGF deprivation (Rukenstein et al., 1991; Xu et al., 2001). Similar to DNA damage, a time course performed by quantitative PCR revealed that Sertad1 mRNA levels were elevated in neuronal PC12 cells as early as half an hour following NGF deprivation and were consistently and significantly increased in SCG neurons after 2 hrs of such treatment (Figures 1C and 1D). Sertad1 protein expression was also elevated by 2–4 hrs in response to NGF deprivation and this change was similar in magnitude (by about 2-fold) to that of mRNA (Figure 1E). Thus, both Sertad1 mRNA and protein expression are induced in response to DNA damage and trophic factor withdrawal, although this response is more modest in the case of NGF deprivation.
Next, we examined whether elevated expression of Sertad1 is sufficient in itself to trigger neuron death in the presence of trophic support. Expression of Sertad1 alone did not induce death of neuronal PC12 cells (Figure 3A), cortical neurons (Figure S3) or cerebellar granule neurons (data not shown). However, over-expression of Sertad1 significantly enhanced the level of neuronal cell death that occurs in response to NGF deprivation (Figure 3B).
Because of the observed up-regulation of Sertad1 in multiple neuronal death paradigms and the sensitization to death with Sertad1 over-expression, we next examined whether neuron death in our systems requires Sertad1. We first studied the role of Sertad1 in NGF deprivation. To achieve this, we prepared several shRNAs specifically targeted to rat Sertad1 and identified three that substantially reduced expression of the over-expressed protein (Figure 4A). Each of the effective shRNAs significantly protected neuronal PC12 cells from death induced by NGF withdrawal (Figure 4B,C). Moreover, most of the Sertad1 shRNA expressing cells preserved overall neuron morphology even after 2 days of NGF deprivation (Figure 4B). This is reminiscent of the preservation of neuronal morphology after NGF deprivation that is achieved with small molecule Cdk inhibitors (Park et al., 1996; Park et al., 1997b) or Cdk4 shRNAs (Biswas et al., 2005). Similar experiments with SCG neurons also showed that down-regulation of Sertad1 by two independent shRNAs significantly blocks death and preserves overall neuron morphology after NGF deprivation (Figure 4D and data not shown).
Similarly, we used a shRNA knockdown strategy to assess whether Sertad1 also plays a required role in neuronal death after DNA damage. In this case, we used adenoviruses to deliver shRNAs to cultured mouse cortical neurons. We screened two potential shRNA sequences directed against mouse Sertad1 to knockdown endogenous Sertad1. As shown in Figure 5A and B, infection of both Sertad1 shRNA viruses, but not a control virus, significantly reduced the Sertad1 immunostaining signal (54% and 33%, respectively) following camptothecin treatment. Survival assays in response to camptothecin treatment were performed using these two Sertad1 shRNAs along with a control shRNA virus. As shown in Figure 5C, expression of the two different Sertad1 shRNAs resulted in significant protection from camptothecin at times up to 48 hrs. This finding is consistent with our observations in the NGF deprivation model and indicates that Sertad1 plays a required role in the neuronal death process.
We additionally examined the importance of Sertad1 in a culture model directly relevant to neurodegeneration. A substantial body of evidence has implicated cell cycle molecules in neuron death associated with Alzheimer’s disease (Herrup et al., 2004; Greene et al., 2007; Zhu et al., 2007; Copani et al., 2008). Moreover, there are striking similarities between the molecular mechanisms of neuron death induced by NGF deprivation, DNA damage and Aβ exposure (Greene et al., 2007) and recent findings support a mechanism in which neuron death triggered by NGF deprivation is mediated by elevated production and release of Aβ (Matrone et al., 2008a; 2008b) . We therefore examined whether Sertad1 is induced and is necessary for death of cultured cortical neurons induced by exposure to aggregated Aβ. Sertad1 transcripts were increased by nearly 4-fold within 3 hrs of Aβ treatment (Figure 6A). Down-regulation of Sertad1 by shRNA significantly protected the cortical neurons from Aβ induced death (Figure 6B). Moreover, Sertad1 shRNA-expressing cells showed preservation of overall neuron morphology in presence of aggregated Aβ (Figure 6C). Taken together, these findings indicate that Sertad1 is induced and plays a required role in in vitro paradigms of both development and pathological neuron death.
The above evidence indicates the importance of Sertad1 in in vitro models of neuronal death. However, its involvement in vivo is unknown. Critically, this uncertainty extends generally to whether the cell cycle pathways may be of importance in developmental death in vivo. Because limiting levels of target derived trophic support appear to regulate neuron death in vivo, our in vitro data led us to test whether Sertad1 is essential for developmental neuron death in the early postnatal cerebral cortex. We used in utero electroporation to deliver DNA encoding Sertad1 shRNA and GFP or a control DNA expressing random shRNA and GFP into the left lateral ventricles of E16 rat embryonic brains. DNA delivered in this manner is taken up by ventricular zone progenitor cells that subsequently differentiate and migrate towards the pial surface. Because maximum cortical neuron death occurs during the first week of development (Spreafico et al., 1995), we sacrificed the DNA-electroporated rat pups on postnatal day 5 (p5) and then analyzed their fixed cortices by TUNEL assays coupled with immunohistochemistry for GFP to assess the proportions of transfected neurons undergoing death. Past studies have established that TUNEL staining faithfully reports apoptotic neurons in developing brain as judged by nuclear morphology and caspase 3 activation (Sophou et al., 2006). Transfected, morphologically identified neurons were mostly present in cortical layers II, III, and IV of the electroporated side of the brain (Figure 7A and Figure S4). Counts of TUNEL positive cells indicated that approximately 14 per 1000 transfected neurons (1.4%±0.13; N=3) were apoptotic in these layers of the cerebral cortex for those animals electroporated with DNA expressing control shRNA (Figures 7A and 7B). This is consistent with a past study that reported about 15 apoptotic (TUNEL positive) nuclei per 1000 neurons in the same cortical layers of normal p5 brains (Spreafico et al., 1995). In contrast, we found an average of only 6 apoptotic nuclei per 1000 Sertad1 shRNA-expressing neurons (0.6%±0.03; N=3) in the same three layers of the cerebral cortex (Fig 7B and 7C). There were no evident effects of the Sertad1 shRNA on the migration or morphology of the transfected neurons. Taken together, our findings indicate that Sertad1 plays an essential role in neuron death during cortical development in vivo.
Our past studies (Park et al., 1997b; Liu and Greene, 2001; Liu et al., 2004; Biswas et al., 2005; Liu et al., 2005; Biswas et al., 2007; Greene et al., 2007) have established a sequential pathway for neuron death caused by NGF deprivation and Aβ exposure in which: 1) activated Cdk4 phosphorylates members of Rb family of transcription-regulating proteins; 2) such phosphorylation leads to dissociation of repressor complexes containing Rb family members, E2F transcription factor proteins and chromatin modifiers; 3) there is consequent derepression and elevated expression of E2F-responsive genes including the transcription factors B- and C-Myb; and 4) the induced mybs activate transcription of the gene encoding the pro-apoptotic BH3-only protein Bim. If Sertad1 acts as anticipated at the proximal end of this pathway by binding and activating Cdk4, we would predict that the downstream responses such as Rb phosphorylation, Myb induction and Bim induction should be blocked by Sertad1 down-regulation. We first assessed the effects of shRNA-mediated Sertad1 knockdown on Rb phosphorylation after NGF deprivation. As shown in Figure 8A and consistent with our past findings, NGF deprivation (18 hrs) raised the proportion of cells with a high level of phospho-Rb immunostaining from about 8% to 65%. Transfection with Sertad1 shRNA reduced this to an average of 12.5%. Similar findings were achieved for expression of C-Myb and its downstream target Bim (Figures 8B and 8C). In each case, shRNAs targeted to Sertad1 strongly suppressed the large increases in the proportions of transfected cells that showed high staining for C-Myb and Bim after NGF deprivation. Taken together, these results support the conclusion that Sertad1 is required for the activation of a neuronal apoptotic pathway in response to NGF deprivation that depends on Cdk-mediated Rb family phosphorylation and consequent induction of E2F-responsive gene Myb and its pro-apoptotic target Bim.
Cell cycle proteins have been implicated as required elements in the mechanisms of postmitotic neuron death, both during normal development and in response to injury, stroke and a range of neurodegenerative diseases (Smith et al., 2004; Neve and McPhie, 2006; Greene et al., 2007; Nunomura et al., 2007; Rashidian et al., 2007). However, the proximal events that initiate the involvement of cell cycle proteins, particularly the activation of Cdk4, have been unclear. Here, we carried out experiments that now implicate the cell cycle regulatory protein Sertad1 in three different paradigms of neuron death: NGF deprivation, DNA damage and Aβ exposure. We find that Sertad1 expression is rapidly elevated in cultured neurons in all three apoptotic paradigms and that down-regulation of Sertad1 by shRNA is protective in each case. We further show that down-regulation of Sertad1 inhibits the increase in Rb phosphorylation and consequent induction of Myb and Bim that mediate neuron death due to NGF deprivation. These findings thus place Sertad1 proximal to the other defined events in the Cdk4-dependent apoptotic cell cycle pathway that is triggered by loss of trophic support.
The steps that comprise the apoptotic cell cycle pathway described here have been determined largely by in vitro experimentation. We used in vivo electroporation to extend our studies to the developing cortex and found that in this instance also, Sertad1 plays an essential role in normally occurring neuronal death. Such findings thus implicate not only Sertad1 but also events downstream of it in developmental neuron death.
Sertad1 has been found to directly bind and activate Cdk4 as well as to render active Cdk4-cyclinD1 complexes resistant to inhibition by p16(INK4a) (Sugimoto et al., 1999). Such actions are entirely consistent with our findings that knockdown of Sertad1 protects neurons in our three death paradigms, each of which requires Cdk4 activation (Greene et al., 2004; Greene et al., 2007). Our studies of the effects of Sertad1 knockdown on cell cycle events triggered by NGF deprivation also support the idea that the capacity of Sertad1 to activate Cdk4 is crucial to its death-promoting actions in neurons. For example, this mechanism would explain why Sertad1 knockdown blocks hyperphosphorylation of Rb, a major cellular target of Cdk4 as well as its inhibition of the subsequent induction of Myb and of Bim. In contrast, other known functions of Sertad1 seem less likely to account for its role in neuron death caused by NGF withdrawal. For instance, Sertad1 binds the E2F1 partner DP-1, thereby enhancing the transactivation of DP-1/E2F1 complexes (Hsu et al., 2001). However, our past findings showed that gene derepression rather than transactivation is required for neuron death triggered by NGF deprivation (Liu and Greene, 2001). Such a mechanism would also not account for our observations that Sertad1 knockdown blocks Rb hyperphosphorylation and induction of Myb (which is repressed by E2F complexes) and Bim. Another potentially relevant action of Sertad1 is its capacity to stimulate p53 transcriptional activity (Watanabe-Fukunaga et al., 2005). However, the major proapoptotic targets of p53, Noxa and Puma, do not appear to be either induced or required for death in response to NGF deprivation (Biswas S.C. and Greene, L.A., unpublished observations) and it is unclear that p53 plays a major role in death associated with this paradigm (Sadoul et al., 1996). In addition to activating Cdk4, Sertad1 was recently reported to stabilize XIAP and thus disinhibit caspases in cancer cells (Hong et al., 2009). Such an action of Sertad1, if occurred in neurons, would have an anti-apoptotic effect and does not seem to be consistent with the role found here of being pro-apoptotic in response to NGF deprivation, DNA damage and Aβ. However, in our death paradigms, either camptothecin or Aβ evokes a series of death pathways which are both caspase-dependent and caspase-independent (Stefanis et al., 1999); (Cregan et al., 2002); (Selznick et al., 2000). Even if Sertad1 confers anti-apoptotic properties, delayed neuronal death still occurs in a caspase-independent fashion in these paradigms.
While Cdk4 activity is essential for neuron death induced by DNA damage, this paradigm of death differs from that caused by trophic factor withdrawal in that it requires both E2F1 and p53 as well as induction of Puma (Park et al., 2000; Wyttenbach and Tolkovsky, 2006). This raises the possibility that in addition to its role in activating Cdk4, Sertad’s capacities to activate E2F1 and p53 may be relevant to neuronal death triggered by DNA damage. Additionally, we found that E2F1 at least partially mediates death induced by camptothecin.
Although Sertad1 over-expression enhanced neuronal death caused by NGF deprivation, it was not sufficient to promote neuron death when over-expressed in healthy neurons. This observation indicates that Sertad1 is necessary, but not sufficient for neuron death and that additional events are required to initiate the death pathway. This could include for instance, modification of Sertad1 triggered by apoptotic stimuli, induction of additional proteins essential for Cdk4 activation, or promotion of Sertad1-independent events such as phosphorylation/dephosphorylation of Cdk4 that are needed for its full activation. A past study has shown that activation of Cdc25A is required for Cdk4 activation and neuron death in response to DNA damage (Zhang et al., 2006), raising the possibility for its involvement at least in the case of this death paradigm.
Past studies have described strong parallels between the mechanisms by which neurons die in response to NGF deprivation and Aβ treatment (Greene et al., 2007). This also appears to be the case for the vulnerable populations of neurons in AD that co-express elevated levels of Cdk4, hyperphosphorylated Rb and Bim protein (Biswas et al., 2007). A unifying explanation has recently been provided by the observations that neurotrophic factor deprivation causes the enhanced formation and release of Aβ which then interacts with cells to trigger an apoptotic pathway (Matrone et al., 2008a; 2008b). In this light, it is significant that Sertad1, as in the case of trophic deprivation, is also required for neuron death triggered by exposure to aggregated Aβ. Such finding further support a common pathway for the two apoptotic stimuli and raise the possibility that Sertad1 may be a therapeutic target for amelioration of neuron death and degeneration in AD. In addition to AD, activation of cell cycle proteins has also been implicated in a number of neurodegenerative diseases and nervous system insults associated with neuron death (Becker and Bonni, 2004; Greene et al., 2004; Herrup et al., 2004; Greene et al., 2007). For example, Sertad1 was among those genes that were induced in a cellular model of Parkinson’s disease (Ryu et al., 2005), raising the possibility of its involvement at least in this disorder and perhaps in others.
Microarray analysis of RNA extracted from cortical neurons following camptothecin (10μM) treatment for 8 hrs or with vehicle (0.1% DMSO). Fold change represents the ratio of signal in camptothecin treated neurons relative to vehicle-treated neurons for each probe set.
This work was supported in part by grants from the NINDS (L.A.G.), ADRC/Taub pilot grant (Columbia University; S.C.B.), Indian Institute of Chemical Biology (S.C.B), Canadian Institute of Health Research (R.S.S., D.S.P.), and Heart and Stroke Foundation Ontario (R.S.S., D.S.P.). We thank Janet Peterson for technical assistance. We also thank Dr. Julio Pozueta for providing us dodecamer β-amyloid.