Peg3/Pw1 is a mediator between p53 and Bax in DNA damage-induced neuronal death.

Neuronal cell death after DNA damage requires p53 and Bax, but the mechanism by which p53 activation leads to Bax translocation and cell death in neurons is not known. We report here that Peg3/Pw1 is up-regulated after DNA damage in cortical neurons in a p53-dependent manner. Overexpression of Peg3/Pw1 leads to decreased neuronal viability. The deleterious effect of Peg3/Pw1 on neuronal survival is abrogated by deletion of either p53 or Bax, indicating an essential role for both in Peg3/Pw1-mediated neuronal death. Moreover, overexpression of a Peg3/Pw1 dominant negative protein inhibits Bax translocation and neuronal cell death after DNA damage. These findings implicate Peg3/Pw1 as a mediator between p53 and Bax in a neuronal cell death pathway activated by DNA damage.

Previous observations in our laboratory have demonstrated that p53 and Bax are essential components of the cell death pathway activated in neurons after DNA damage due to exposure to ionizing radiation or the topoisomerase I inhibitor, camptothecin (1)(2)(3). We and others have also obtained evidence for a downstream role for caspase activation in this cell death pathway under some, but not all circumstances (4 -6). Numerous other agents that are released from mitochondria as a result of Bax translocation have been identified as potential mediators of caspase-independent cell death pathways (7), although their role in neuronal cell death is poorly understood. Moreover, the mechanism by which p53 activation leads to Bax translocation is not known. Although Bax expression can be up-regulated at the transcriptional level by p53, we and others have shown that DNA damaging agents result in Bax translocation without causing a significant increase in the levels of Bax protein in cultured postnatal cortical neurons (2,3,8). Thus, additional mechanisms must exist that link p53 activation to Bax translocation. Identification of molecules that serve as intermediaries between p53 and Bax would increase our understanding of neuronal death mechanisms and possibly provide new therapeutic targets.
One such candidate molecule is Peg3/Pw1, a protein that was originally identified as being involved in the development of the myogenic and neuronal lineages (9). Peg3/Pw1 is thought to be a multifunctional protein with primarily nuclear localization (9,10). The protein contains 12 putative Kruppel-type zinc finger DNA-binding domains, an RER protein interaction domain, and two proline-rich repeat domains, suggesting that Peg3/Pw1 may act as a transcription factor (9,11). The Peg3/ Pw1 gene is an imprinted gene that is expressed ubiquitously from the paternal allele in embryos, and mRNA hybridization studies in mouse and human tissues suggest that it is expressed at highest levels in the ovary, placenta, testis, and brain (9,(11)(12)(13). Multiple isoforms with differential localization have been identified in humans, although only a single isoform has been reported in mice (14). Peg3/Pw1 is highly conserved between mouse and human, suggesting an important and conserved function throughout evolution (12). Disruption of the paternal allele of mouse Peg3/Pw1 results in behavioral abnormalities, growth retardation, and alterations in the number of oxytocin neurons in the hypothalamus (15).
More recently, Peg3/Pw1 was identified as a gene product that is specifically expressed in fibroblasts undergoing p53mediated apoptosis when compared with cells undergoing p53mediated growth arrest (16,17). Evidence suggests that it may cooperate with another p53-regulated gene product, Siah1, to mediate Bax translocation and cell death in a fibroblast cell line (16). In neurons, Peg3/Pw1 expression is up-regulated in vivo and in vitro by hypoxia (18). This observation is of interest because of evidence suggesting that neuronal death occurring after hypoxic insults is p53-dependent (19,20).
Given the putative role of Peg3/Pw1 in p53-mediated apoptosis in fibroblasts and neurons, we have chosen to examine its role in p53-mediated neuronal cell death occurring after DNA damage. We report here that Peg3/Pw1 is up-regulated after DNA damage in neurons in a p53-dependent manner. Enforced overexpression of Peg3/Pw1 protein leads to decreased neuronal viability, an effect that is abrogated by deletion of either the p53 or Bax genes. Moreover, overexpression of a Peg3/Pw1 dominant negative protein (Peg3-DN) containing a truncated C terminus inhibits both Bax translocation and neuronal death occurring after DNA damage. These findings suggest that Peg3/Pw1 acts as a mediator between p53 and Bax in a neuronal death pathway that is activated by DNA damage.

EXPERIMENTAL PROCEDURES
Animals and Cell Culture-Primary neuronal cell cultures derived from wild-type, p53-deficient, or Bax-deficient animals were used for this study. p53 wild-type and knockout mice on a C57BL/6 X 129 SV background were generated as previously described (21). Bax knockout mice were generated from a 129/Sv x C3H background as previously described (22). The genotypes of the mating pairs and all offspring were determined by PCR using DNA extracted from the tail as previously described (23,24). Primary neuronal cultures were established from newborn mouse cortex as described previously (1,2,4). Briefly, cortical tissue was excised, trypsinized, and dissociated by trituration to obtain single cells. Cells were then plated onto poly-D-lysine-coated culture-ware and maintained in neurobasal medium with B27 supplements (Invitrogen). Under these conditions, cultures were shown to contain greater than 95% neurons using neurofilament immunostaining techniques (1). Cultures were maintained for 4 days prior to experimental manipulations.
Adenovirus Preparation-Non-replicative recombinant adenovirus deleted in the E1 region and carrying the human p53 gene under the control of the cytomegalovirus promoter was generated as previously described (25) and was the generous gift of Dr. Toshiyoshi Fujiwara (Okayama University Medical School, Japan). Non-replicative recombinant adenovirus deleted in the E1 region and carrying the AxCALacZ ␤-galactosidase gene was generated as described previously (26) and was kindly provided by Drs. Saito and Kanegae (University of Tokyo, Japan). Adenoviruses were propagated in E1 complementing human embryonic kidney 293 cells, purified, and used as described in our previous publications (1,2). An multiplicity of infection of 125 (a titer that did not produce significant toxicity alone) was used for these studies. The transduction efficiency using this method exceeded 90%, as described previously (1,2).
Immunocytochemistry-Neurons in dissociated cell culture were washed in PBS 1 and fixed in 4% paraformaldehyde for 30 min. The cells were then permeabilized with 0.2% Triton X-100 for 5 min, washed 3 times in PBS (pH 7.4), and blocked in blocking buffer containing the appropriate blocking serum for 1 h. The cultures were incubated in primary antibody diluted in blocking buffer overnight at 4°C. The cultures were then washed in PBS 3 times and incubated for 2 h in secondary antibody diluted in blocking buffer. Peg3/Pw1 immunocytochemistry was performed using a mouse monoclonal anti-Peg3/Pw1 antibody at a 1:50 dilution, 2 followed by a goat anti-mouse Alexa 594labeled secondary antibody at a 1:1000 dilution (Molecular Probes). Fluorescent microscopic images were obtained on a Zeiss Axiovert deconvolution microscope equipped with emission and excitation filter wheels and a cooled CCD camera (Cooke Corp., Auburn Hills, MI). Fluorescent intensity comparisons were made using images obtained with identical exposure time and normalized on the same pixel intensity distribution. All images were deconvolved and/or analyzed using Slidebook image analysis software (Intelligent Imaging Innovations, Denver, CO).
Transfection/Transduction Experiments-An expression construct encoding Bax fused to the enhanced green fluorescent protein (GFP) was created using full-length human Bax cDNA inserted into the vector pEGFP-C3 (CLONTECH, Palo Alto, CA) (27) and generously supplied by Dr. R. Youle (Bethesda, MD). The EGFP-Peg3/Pw1 vector was constructed by fusing the full-length Peg3/Pw1 coding sequence in-frame to the carboxyl terminus of EGFP in pEGFP-C1 (CLONTECH) as de-scribed previously (17). The dominant negative mutant form of Peg3/ Pw1 (Peg3-DN, amino acids 1-592) containing a COOH-terminal truncation has been described previously (28). The vectors were amplified in TOP 10FЈ bacteria using a standard protocol.
Transient transfections were performed using LipofectAMINE 2000 reagent (Invitrogen) in Opti-MEM according to the manufacturer's instructions. Briefly, 2 g of vector DNA diluted in 20 l of Opti-MEM was incubated for 20 min with 2 l of LipofectAMINE 2000 reagent diluted in 80 l of Opti-MEM. This mixture was then diluted to 450 l with neurobasal medium and added to neuronal cultures for a 4 -6-h incubation period. Cells were then washed and again placed into neurobasal growth medium. Transfected cells were viewed using fluorescein isothiocyanate epifluorescence. The transfection efficiency using this method was ϳ2-5%. The Peg3-DN (amino acids 1-592) vector was used in transient co-transfection experiments. Peg3-DN DNA was cotransfected with either EGFP or EGFP-Bax DNA in a 7:3 ratio using the method described above, such that the total amount of transfected DNA was 2 g.
Western Blot-Cultured cells were collected by scraping in ice-cold PBS and spun down as previously described (2). Western blots for Peg3/Pw1 were performed as described previously (17,28).

DNA Damage Up-regulates Peg3/Pw1 Protein Levels in Neu-
rons-To investigate the role of Peg3/Pw1 in p53-mediated neuronal cell death, we examined Peg3/Pw1 expression in cultured primary mouse cortical neurons under control conditions and after genotoxic or excitotoxic injury. Primary cortical neurons were obtained from newborn mice and maintained in dissociated cell culture for 4 days in a chemically defined medium. The cells were then exposed to camptothecin (2.5 M), glutamate (50 M), or vehicle control. After 18 h, the cells were washed in PBS, fixed in 4% paraformaldehyde, and stained for Peg3/Pw1 immunoreactivity using a mouse monoclonal anti-Peg3/Pw1 antibody. As shown in Fig. 1A, there was a low but significant level of Peg3/Pw1 immunoreactivity present in cultured mouse cortical neurons under control conditions. This finding is in agreement with observations on the expression of Peg3/Pw1 mRNA and protein in mouse brain in vivo (10,15). After DNA damage, a significant increase in Peg3/Pw1 immunoreactivity was observed. Interestingly, glutamate exposure, which is thought to activate both p53-dependent and p53independent cell death pathways (2, 4), failed to increase Peg3/ Pw1 immunoreactivity significantly. The differential induction of Peg3/Pw1 expression that was noted between glutamate and FIG. 1. Peg3/Pw1 protein is increased by DNA damage. A, phase-contrast (A1, B1, and C1) and fluorescence (A2, B2, and C2) micrographs of mouse cortical neurons stained for Peg3/Pw1 immunoreactivity under control conditions or after exposure to a topoisomerase I inhibitor (camptothecin, 2.5 M) or glutamate (50 M). Primary cortical neurons were obtained from newborn wild-type mice and maintained in dissociated cell culture for 4 days in a neurobasal medium with B27 supplements. The cells were then exposed to camptothecin (2.5 M), glutamate (50 M), or vehicle control (Me 2 SO). After 18 h, the cells were washed in phosphate-buffered saline, fixed in 4% paraformaldehyde for 30 min, and stained for Peg3/Pw1 immunoreactivity using a mouse monoclonal anti-Peg3/Pw1 antibody as described under "Experimental Procedures." A significant increase in Peg3/Pw1 immunoreactivity was noted after exposure to camptothecin, but not glutamate. B, wild-type primary mouse cortical neurons were dissociated and maintained in a serum-free medium as described. After 4 days, the cultures were treated for 6 h with camptothecin (CPT, 2.5 M) or with vehicle control (Me 2 SO). The cells were then harvested by scraping and the protein analyzed by Western blot as described under "Experimental Procedures." Western blot analysis confirmed a 2-fold increase in Peg3/Pw1 protein in wild-type neurons after a 6-h exposure to camptothecin. Western blot analyses at later time points revealed a persistent and progressive accumulation of Peg3/Pw1 protein with continued exposure to camptothecin (see Fig. 4A).
camptothecin parallels a similar differential induction of Bax protein and caspase activity caused by these agents in cultured postnatal cortical neurons (2,4).
The induction of Peg3/Pw1 protein was confirmed by Western blot analysis (Fig. 1B). Primary cortical neurons were treated with camptothecin (2.5 M) for 6 h, and protein extracts were then collected and analyzed by Western blot as described. A 2-fold increase in Peg3/Pw1 protein was observed 6 h after camptothecin exposure. This increase was persistent, and was also observed after 18 h of camptothecin exposure (see Fig. 3A). Thus, both immunocytochemical and Western blot analysis indicated that Peg3/Pw1 protein levels were increased in primary cortical neurons after DNA damage.
Increased Expression of Peg3/Pw1 Decreases Neuronal Viability-To determine the effect of increased Peg3/Pw1 expression on neuronal viability, primary mouse cortical neurons were transfected with a vector containing Peg3/Pw1 fused to green fluorescent protein (pEGFP-Peg3/Pw1) in a transient transfection assay. Control cells were transfected with a vector containing only pEGFP. After 72 h, the cultures were examined for the presence of fluorescent cells expressing either the Peg3/ Pw1-GFP fusion protein or the GFP control protein. Approximately 5% of the neurons were noted to express GFP (as indicated by the presence of green fluorescence) in control cultures. The EGFP-transfected cells showed the extensive dendritic and axonal outgrowth typical of mouse cortical neurons cultured under these conditions ( Fig. 2A). In contrast, most neurons overexpressing the Peg3/Pw1-GFP fusion protein displayed a small, rounded morphology with short, abortive neuritic processes. Co-staining of fixed neurons with Hoechst dye to identify the nucleus confirmed that the overexpressed Peg3/Pw1-GFP protein was primarily nuclear in its location (data not shown), although some expression in the cytoplasm was observed. When measured at either 24 or 72 h, the mean number of neurons per high power field expressing the Peg3/Pw1-GFP fusion protein was significantly less than the mean number of neurons expressing either GFP or a COOH terminus truncated Peg3/Pw1 dominant negative protein (0.6 versus 7.0 versus 6.9, p Ͻ 0.003, after 72 h, paired t test). More importantly, the percent decline in the mean number of transfected neurons observed between 24 and 72 h was greater in the EGFP-Peg3/ Pw1-transfected cultures (75%) than in the control EGFPtransfected cultures (11%) or the Peg3/Pw1-DN-transfected cultures (13%). Thus, these findings indicated that overexpression of Peg3/Pw1 in primary mouse cortical neurons resulted in a progressive decline in neuronal viability.
Overexpression of a Peg3/Pw1 Dominant Negative Protein Inhibits Neuronal Death-To determine whether Peg3/Pw1 is required for execution of the neuronal death pathway activated by DNA damage, wild-type neurons were transfected with a Peg3/Pw1-DN vector (amino acids 1-592) which encodes a COOH terminus truncated Peg3/Pw1 protein (28). This protein lacks 8 of the 12 zinc finger domains and both of the His-Pro repeat domains, but retains the RER protein-interaction domain. This Peg3/Pw1-DN protein has been previously shown to inhibit Peg3/Pw1-mediated NF-B activation, Peg3/Pw1-mediated apoptosis, and Peg3/Pw1-mediated Bax translocation in fibroblasts (16,17,28). The EGFP vector was co-transfected with the Peg3/Pw1-DN vector to facilitate the identification of transfected cells. Control cultures were transfected with the EGFP vector alone. Approximately 24 h after transfection, the cultures were exposed to camptothecin (2.5 M) for an additional 18 h. Transfected cells were then examined under fluorescence and counted to assess overall survival. In some cases, transfected neurons were fixed and counterstained with Hoechst dye to identify apoptotic nuclei with condensed chroma-tin. As shown in Fig. 3A, transfected neurons expressing the GFP control vector showed marked somatic disruption and neurite degeneration after camptothecin exposure, characteristic of neurons undergoing DNA damage-induced cell death. In contrast, most neurons expressing the Peg3/Pw1-DN protein and exposed to camptothecin retained an intact soma and ex- Control neurons expressing only EGFP showed a neuronal morphology with intact somata and extensive axonal and dendritic outgrowth. In contrast, neurons overexpressing Peg3/Pw1-GFP showed abnormal, rounded, condensed cell bodies with absent dendrites and short, abortive axonal outgrowth. Peg3/Pw1-GFP was localized primarily to the nucleus, although cytoplasmic distribution of the protein was observed in some cells. B, dissociated mouse cortical neurons were transfected with a pEGFP vector, a pEGFP-Peg3/Pw1 vector, or a vector encoding a truncated form of Peg3/Pw1 (Peg3-DN), and were subsequently viewed under fluorescence after 24 or 72 h. The mean number of fluorescent neurons per high-power field (ϫ20 objective) was assessed for each condition. The results represent the mean Ϯ S.D. (8 fields per culture; n ϭ 3 cultures/condition) and are representative of three separate experiments. After 72 h, the mean number of neurons expressing the Peg3/Pw1-GFP fusion protein was significantly less than the mean number of neurons expressing either EGFP or a COOH terminus truncated Peg3/Pw1 dominant negative protein (0.6 versus 7.0 versus 6.9, p Ͻ 0.003, paired t test). The percent decline in the mean number of transfected neurons observed between 24 and 72 h was greater in the EGFP-Peg3/Pw1-transfected cultures (75%) than in the control EGFPtransfected cultures (11%) or the Peg3/Pw1-DN-transfected cultures (13%). tensive neuritic outgrowth similar to untreated neurons. Quantitative analysis (Fig. 3B) indicated that a greater percentage of Peg3/Pw1-DN-expressing neurons displayed healthy appearing nuclei and neurites after camptothecin-induced DNA damage than was observed among control GFP-expressing neurons exposed to camptothecin (78% versus 44%, p Ͻ 0.05, paired t test). Moreover, the number of viable neurons expressing GFP alone declined by 87% after camptothecin treatment (n ϭ 103), but the number of viable cells expressing the Peg3/Pw1-DN protein was not significantly different after camptothecin exposure than that observed under control conditions (n ϭ 98).
Thus, inhibition of Peg3/Pw1 activity by overexpression of the Peg3/Pw1-DN protein significantly inhibited neuronal death after DNA damage.
Peg3/Pw1-mediated Neuronal Death Requires a Functional p53 Protein-We have previously shown that neuronal death after DNA damage requires the presence of a functional p53 protein (1)(2)(3). To determine whether the observed increase in Peg3/Pw1 expression after DNA damage was p53-dependent, primary cortical neurons derived from wild-type or p53-deficient mice were exposed to camptothecin for 18 h. The protein was then collected and analyzed by Western blot for Peg3/Pw1 expression. Protein extracts from control (vehicle treated) neurons were also analyzed. As shown in Fig. 4A, Peg3/Pw1 expression was up-regulated in wild-type, but not in p53-deficient neurons exposed to camptothecin. Thus, a functional p53 protein is required for Peg3/Pw1 induction after DNA damage in neurons.
Studies in fibroblasts containing a temperature-sensitive mutant form of p53 indicated that p53 expression alone was insufficient to induce Peg3/Pw1 expression (16,17). To determine whether this was also true in neurons, we transduced As shown in A, viable neurons with intact somata and healthy neurites were seen after camptothecin exposure (2.5 M) in cultures transfected with the Peg3/Pw1-DN vector when compared with cultures transfected with the EGFP vector. B, primary wild-type cortical neurons were transfected with either a Peg3/Pw1 dominant negative mutant vector (Peg3-DN) or with an EGFP control vector. After 24 h, the cells were exposed to either camptothecin (2.5 M) or vehicle control for 18 h. The cultures were counterstained with Hoechst 33342 dye and viewed under epifluorescence to identify apoptotic cells exhibiting condensed and fragmented nuclei. The total number of apoptotic and non-apoptotic transfected neurons per high-power field was determined and expressed as a percentage. The experiment was performed in quadruplicate. As discussed in the text, neurons expressing GFP and exposed to camptothecin showed a significant decrease in survival when compared with untreated EGFP-expressing neurons (p Ͻ 0.01, paired t test). In contrast, neurons expressing the dominant negative form of Peg3/Pw1 failed to show a significant decline in neuronal viability after 18 h of camptothecin exposure when compared with untreated, Peg3/Pw1-DNexpressing neurons (p Ͻ 0.84, paired t test). The percentage of viable neurons expressing Peg3/Pw1-DN protein was significantly greater than the percentage of viable GFP-expressing neurons after camptothecin exposure (p ϭ 0.05, paired t test).
FIG. 4. p53 is necessary but not sufficient for Peg3/Pw1-mediated neuronal death. A, primary cortical neurons derived from wildtype (ϩ/ϩ) or p53-deficient (Ϫ/Ϫ) mice were dissociated and maintained in a serum-free medium as described. After 4 days, the cultures were treated for 18 h with camptothecin (CAMPTO, 2.5 M) or with vehicle control (CONTROL). The cells were then harvested by scraping and the protein analyzed by Western blot as described under "Experimental Procedures." Western blot analysis revealed a significant increase in Peg3/Pw1 protein in wild-type neurons, but not p53-deficient neurons exposed to camptothecin. B, primary cortical neurons were transduced with an adenovirus containing either the human p53 gene (Ad-p53) or the LacZ ␤-galactosidase gene (Ad-␤gal) at an multiplicity of infection of 125. After ϳ60 h, the cultures were fixed in 4% paraformaldehyde and stained for Peg3/Pw1 immunocytochemistry. Phase-contrast (top row) and fluorescence (bottom row) images are shown. No significant increase in Peg3/Pw1 immunoreactivity was observed. C, dissociated primary cortical neurons derived from wild-type (WT) or p53-deficient (p53Ϫ/Ϫ) postnatal mice were transfected with a pEGFP-Peg3/Pw1 vector, and subsequently viewed under epifluorescence after 24 or 48 h. The mean number of fluorescent neurons per high power field (ϫ20 objective) was assessed for each condition. Data shown are derived from two separate experiments, and are represented as the mean Ϯ S.E. Each condition was performed in triplicate. A significant decrease in the number of wild-type neurons expressing Peg3/Pw1-GFP was noted between 24 and 48 h (p Ͻ 0.05, paired t test), in contrast to p53-deficient neurons expressing Peg3/Pw1-GFP, which failed to show any decrease in neuronal survival over the same period (p Ͻ 0.12, paired t test). cultured mouse cortical neurons with an adenovirus containing either the p53 gene or the LacZ ␤-galactosidase gene as a control. After ϳ60 h, the cultures were fixed and stained for Peg3/Pw1 immunoreactivity as described. Up-regulation of p53 protein expression after transduction with the p53 adenovirus was confirmed by Western blot (data not shown). No increase in Peg3/Pw1 immunoreactivity was observed in p53-transduced cultures when compared with the ␤-galactosidase-transduced cultures (Fig. 4B). These data are consistent with observations in fibroblasts indicating that up-regulation of p53 alone is insufficient to induce Peg3/Pw1 expression (16). Taken together, these findings indicate that p53 is necessary, but not sufficient for Peg3/Pw1 induction after DNA damage in neurons.
Although p53 was required for Peg3/Pw1 induction in neurons after DNA damage, we wondered whether enforced overexpression of Peg3/Pw1 could nevertheless mediate neuronal death in the absence of p53. To investigate this possibility, primary cortical neurons derived from wild-type or p53-deficient mice were transfected with the EGFP-Peg3/Pw1 vector, and the cultures were then examined after either 24 or 48 h to determine the number of transfected neurons that survived. As shown in Fig. 4C, Peg3/Pw1 overexpression significantly decreased neuronal survival in wild-type, but not in p53-deficient neurons at both time points. A statistically significant decline in the number of transfected cells per high powered field was noted in the wild-type, but not in the p53-deficient cultures over the 48-h observation period. Thus, a functional p53 protein is required for Peg3/Pw1-mediated neuronal death, even when Peg3/Pw1 expression is increased by other means.
Peg3/Pw1 Mediates Neuronal Death via Bax Translocation-We have previously shown that neuronal death after DNA damage requires the presence of a functional Bax protein (2,3). Bax-deficient neurons are markedly resistant to cell death occurring after DNA damage. Several reports indicate that Bax translocation is an important step in the genesis of mitochondrial dysfunction and programmed cell death under numerous circumstances (29,30). Indeed, camptothecin-induced DNA damage leads to Bax translocation in primary cultured cortical neurons (8). This finding is of interest, given the reported role of Peg3/Pw1 in mediating Bax translocation in cultured fibroblasts undergoing p53-dependent apoptosis. To examine the role of Bax in Peg3/Pw1-dependent neuronal cell death in neurons, primary cortical neurons derived from wildtype or Bax-deficient mice were transfected with the Peg3/Pw1 vector and subsequently examined to determine the number of transfected neurons that survived. As shown in Fig. 5A, Peg3/ Pw1 overexpression significantly decreased neuronal survival in wild-type, but not in Bax-deficient neurons (p Ͻ 0.05, paired t test). Thus, a functional Bax protein is required for Peg3/Pw1mediated neuronal death.
To determine whether Peg3/Pw1 activity mediates Bax activation and translocation after DNA damage in neurons, we co-transfected wild-type neurons with a vector containing a truncated (amino acids 1-592) Peg3/Pw1-DN protein and with a Bax-GFP vector. The neurons were then exposed to camptothecin for 18 h, and the cells were examined to determine the cytoplasmic distribution of the Bax-GFP protein. As shown in Fig. 5B, The Bax-GFP fusion protein was diffusely distributed throughout the neuronal cytoplasm under control conditions. After camptothecin exposure, a redistribution of Bax-GFP into a punctate distribution was observed, indicating translocation of Bax from the cytoplasm to mitochondria, in agreement with the findings of Morris et al. (8). Co-transfection of Bax-GFP with the Peg3/Pw1-DN construct, however, significantly inhibited Bax translocation. As shown in Fig. 5B, the Bax-GFP protein retained a diffuse cytoplasmic pattern. These data are consistent with observations in fibroblasts indicating that increased Peg3/Pw1 expression induces Bax translocation (17). Taken together, the findings indicate that Peg3/Pw1 mediates cell death by causing Bax translocation to mitochondria in neurons. previous studies (1-3, 8), we and others have obtained experimental evidence that p53 and Bax are essential components of a neuronal cell death pathway that is activated by DNA damage. Among the downstream effectors of this pathway are caspases, although the importance of these cysteine aspartases to neuronal death after DNA damage may vary depending on the stage of neuronal maturation (4, 5). Other downstream effectors apparently contribute to the p53-dependent apoptotic process (7), but these have yet to be clearly identified in neurons. Moreover, the exact relationship between p53 and Bax in this pathway is unclear. Bax is a p53-regulated gene under many (but not all) conditions (31). Bax protein levels are up-regulated in a p53-dependent manner in primary cortical neurons by glutamate, but not by DNA damage due to camptothecin or ionizing radiation (2,3,8). Thus, mechanisms other than direct transcription of the Bax gene by p53 must exist for p53-dependent regulation of Bax activity in primary neurons.

A Role For Peg3/Pw1 in p53-dependent Neuronal Death after Injury-In
In this study, we report the first evidence that Peg3/Pw1 acts as an intermediary between p53 and Bax in a cell death pathway activated by DNA damage in primary mouse cortical neurons. Peg3/Pw1 protein expression is increased by DNA damage in cortical neurons in a manner that requires the presence of p53. cDNA array analysis indicates that mRNA levels for Peg3/Pw1 are also moderately increased in cultured primary cortical neurons after DNA damage, 3 suggesting that this upregulation may occur at the transcriptional level. Support for a direct role for Peg3/Pw1 in mediating neuronal death derives from the observation that enforced overexpression of Peg3/Pw1 decreases neuronal survival, and inhibition of Peg3/Pw1 activity by overexpression of a dominant negative form of Peg3/Pw1 inhibits Bax translocation and neuronal death after DNA damage. Moreover, neuronal death occurring as a result of Peg3/ Pw1 overexpression requires the presence of both p53 and Bax.
Our finding that Peg3/Pw1 plays a role in determining neuronal viability after some forms of injury is supported by the recent finding that Peg3/Pw1 mRNA is up-regulated by hypoxia in primary cortical neurons in vitro and in vivo (18). Prior studies have implicated p53 in a cell death pathway activated in neurons by hypoxia (19,20,32). Yamaguchi et al. (18) also reported that overexpression of a Peg3/Pw1-DN mutant protein inhibited hypoxia-induced cell death in SK-N-SH neuroblastoma cells. Taken together with the finding presented in the current study, that the neuronal death pathway activated by DNA damage involves Peg3/Pw1, there is growing evidence for a role for Peg3/Pw1 in mediating neuronal death after selected forms of injury.
However, Peg3/Pw1 may not be involved in all forms of injury-induced neuronal death. The circumstances under which Peg3/Pw1 is induced in injured neurons may vary, depending upon the nature of the stimulus. In contrast to the marked up-regulation in Peg3/Pw1 immunoreactivity observed in primary cortical neurons after DNA damage, we did not observe a significant increase in Peg3/Pw1 immunoreactivity after glutamate exposure. In support of this finding, overexpression of a Peg3/Pw1 dominant negative mutant protein failed to prevent glutamate-induced neuronal death in cultured primary cortical neurons (data not shown). Similarly, Yamaguchi et al. (18) reported significant increases in Peg3/ Pw1 mRNA in astroglial cells after hypoxia and heat stress, but not after exposure to tunicamycin or the calcium ionophore A23187. We have previously reported that although neuronal death occurring after low dose glutamate exposure and after camptothecin treatment both depend upon the presence of p53, there are differences in the regulation of downstream p53regulated gene products, such as Bax and caspases (2,4). These findings suggest that the specific cell death pathway activated by p53 may depend critically upon the nature of the inciting stimulus, and they point to the likely existence of co-regulating factors that determine the specific nature of p53-dependent cell death pathways in neurons.
The Relationship between p53 and Peg3/Pw1-Although DNA damage-induced up-regulation of Peg3/Pw1 protein was critically dependent upon the presence of p53, overexpression of p53 alone using adenovirus-mediated transduction was insufficient to up-regulate Peg3/Pw1 immunoreactivity in cultured primary cortical neurons. The finding that increased p53 expression is necessary, but not sufficient to induce Peg3/Pw1 protein levels is similar to that observed in fibroblasts containing a temperature-sensitive mutant form of p53 (16), and suggests that cofactors, in addition to p53, are necessary for Peg3/ Pw1 up-regulation. This result differs from that obtained by Yamaguchi et al. (18), however, who found a significant upregulation of Peg3/Pw1 mRNA in SK-N-SH neuroblastoma cells transduced with a p53 adenovirus. The reasons for this discrepancy are unclear, although one possible explanation is that Yamaguchi et al. (18) assayed Peg3/Pw1 mRNA levels, in contrast to the immunocytochemical analysis of Peg3/Pw1 protein expression performed in the current study. Several reports indicate that changes in mRNA levels do not always correspond to changes in protein expression (33). Another explanation for this discrepancy may lie in the fact that Yamaguchi et al. (18) used a neuroblastoma cell line for the adenovirus experiments, in contrast to the primary cortical neurons used in the present study. Using a different neuroblastoma cell line (SH-SY5Y), we have observed differences between neuroblastoma cells and primary neurons in the sensitivity to camptothecin and in the ability of caspase inhibitors to promote survival after DNA damage (4), indicating subtle differences in the p53-dependent cell death pathways activated by DNA damage in neuroblastoma cell lines when compared with primary neurons. In support of this hypothesis, it is notable that Yamaguchi et al. (18) also reported a concomitant increase in Bax mRNA levels in SK-N-SH cells after p53 overexpression. This contrasts with findings from our laboratory and others indicating no change in Bax protein levels in primary neurons after adenovirus-mediated overexpression of p53 or after DNA damage (2,3,8). Further studies are clearly needed to determine the factors governing the p53-dependent up-regulation of Peg3/Pw1 in neurons after injury.
The finding that p53-deficient neurons are resistant to cell death caused by enforced overexpression of Peg3/Pw1 suggests that Peg3/Pw1 cooperates with other p53-regulated factors to effect cell death. The pro-apoptotic protein, Bax, is clearly one such factor. The Bax gene contains a p53 consensus binding site in its promoter, and p53 can directly increase Bax transcription (31). Bax protein levels are unchanged in primary neurons after DNA damage, although the p53-dependent cell death pathway has an obligate requirement for Bax in order for it to be executed (2,3,8). Because Bax protein is present at significant levels in p53-deficient cortical neurons (2), there must be yet additional p53-regulated factors that are required for Peg3/Pw1 to mediate cell death.
One possible such cofactor is Siah1a (seven in absentia homolog 1), another p53-induced gene product. Siah1 is differentially induced by p53 in a leukemic cell line undergoing growth arrest or apoptosis (34), or by genotoxic stress (35). Interestingly, several members of the Siah1 family have been shown to have ubiquitin targeting activity (36). Other studies suggest that p53 and Siah1 share a common mechanism of tumor suppression and induction of apoptosis that involves protein folding, unfolding, and trafficking (37). Siah1a has previously been shown to bind directly to Peg3/Pw1 by yeast two-hybrid assays and by direct co-immunoprecipitation experiments in a fibroblast cell line (16). Whereas overexpression of Peg3/Pw1 in the absence of p53 does not induce apoptosis in fibroblasts (a finding that parallels our observations in neurons), co-trans-fection of both Peg3/Pw1 and Siah1 induces apoptosis in fibroblasts in the absence of a functional p53 protein (16). It will be of interest to determine whether Siah1 also plays a role in neuronal death occurring after DNA damage.
Peg3/Pw1 has also been shown to bind to TRAF2, another RER domain protein that is believed to regulate NF-B activation through interactions with the tumor necrosis factor superfamily of death receptors. However, the significance of the Peg3/Pw1-NF-B interaction to cell survival is unclear, since cells derived from mice lacking Peg3/Pw1 show normal NF-B activation and apoptosis in response to tumor necrosis factor (38).
Fortin et al. (6) have reported that DNA damage in embryonic cortical neurons leads to the p53-dependent up-regulation of APAF-1 expression. APAF-1 is a component of the apoptosome, and cooperates with cytochrome c and ATP to activate caspases (39). It is unlikely that APAF-1 is the p53-regulated factor required for Peg3/Pw1-mediated Bax activation in the paradigm used in the current study, however, since APAF-1dependent caspase activation generally occurs downstream of Bax translocation (39). Moreover, we and others have shown that postnatal primary cortical neurons such as those used in this study die by a p53-dependent, Bax-dependent, but caspase-independent mechanism (4,5).
The Relationship between Peg3/Pw1 and Bax-Previous studies in fibroblasts indicated that Peg3/Pw1 overexpression induced Bax translocation, and that the use of antisense mRNA to block Peg3/Pw1 protein expression inhibited such translocation during p53-mediated apoptosis (17). However, it was unclear from those studies whether there was an obligate requirement for Bax in Peg3/Pw1-induced cell death. In the present study, we report that Bax is required for Peg3/Pw1-dependent cell death. Neurons derived from Bax-deficient mice failed to show a decline in viability after enforced overexpression of Peg3/Pw1. We have also found that inhibition of Peg3/Pw1 activity using a truncated form of the Peg3/Pw1 protein inhibits both Bax translocation and neuronal death after DNA damage. This latter finding parallels the results using antisense mRNA obtained previously in fibroblasts (17), and using a dominant negative Peg3/Pw1 protein in SK-N-SH cells exposed to hypoxia (18). Taken together, these data indicate that Peg3/ Pw1 and Bax cooperate to bring about neuronal death after DNA damage.
The mechanism by which Peg3/Pw1 causes Bax translocation is unclear. The Peg3/Pw1 protein contains 12 zinc finger domains and an RER protein interaction domain (9). This domain structure, combined with the fact that localization of Peg3/Pw1 is primarily nuclear, suggests that this protein may affect gene transcription. However, it seems unlikely that a direct effect on transcription of the Bax gene is involved in Peg3/Pw1-dependent cell death, given that there is no change in Bax protein levels in cultured postnatal cortical neurons after camptothecin exposure. However, the possibility remains that Peg3/Pw1 may regulate the transcription of other genes that are important for neuronal viability. In this context, it is of interest that another set of nuclear proteins, the cyclin dependent kinases (cdk's), have been shown to play an important role in neuronal death occurring after DNA damage. Inhibition of cdk's using cdk inhibitors delays DNA damage-induced neuronal cell death (8,40). During development, the maximal expression of Peg3/Pw1 in developing myoblasts is cell cyclespecific, occurring primarily in late M phase (9). The cdk's, which are nuclear proteins critical for cell cycle control, would thus be potential candidates for cooperating with Peg3/Pw1 in mediating neuronal cell death.
It is also possible that direct protein/protein interactions in nuclear or extranuclear compartments could be involved in the mechanism of Peg3/Pw1-induced cell death. As mentioned above, potential interactions of Peg3/Pw1 with Siah1, TRAF2, or other as yet unidentified proteins may be involved. Additional studies are needed to elucidate the mechanism by which increased Peg3/Pw1 expression leads to Bax translocation. A Model for Neuronal Death after DNA Damage-Taken together, these findings detail a striking interdependence of Peg3/Pw1, Bax, and p53 in neuronal death after DNA damage. We propose a model (Fig. 6) in which DNA damage leads to p53 activation, followed by increased expression of Peg3/Pw1 and a yet to be identified cofactor (e.g. Siah1, TRAF2, cdk's, etc). Peg3/Pw1 would then cooperate with this factor(s) to cause Bax translocation to mitochondria, resulting in mitochondrial dysfunction, a decline in intracellular ATP levels, cytochrome c release, and caspase activation. Based on previous studies, we believe that Bax translocation also activates a caspase-independent cell death pathway that involves other factors (e.g. p53AIP, endonuclease G, etc.) that remain to be specifically identified as having a role in p53-dependent neuronal cell death. Studies are currently underway to elucidate additional details of p53-dependent neuronal cell death pathways after injury.
FIG. 6. The relationship between Peg3/Pw1 and cell death mediators induced in cortical neurons by DNA damage. The present results demonstrate that Peg3/Pw1 is induced in a p53-dependent manner in response to DNA damage. Previous work from our laboratory demonstrated that p53-mediated cell death was dependent on the presence of the pro-apoptotic Bcl-2 family member Bax (2). An important component of Bax-dependent cell death is its translocation from the cytosol to the mitochondria where it is promotes changes in the permeability of the outer mitochondrial membrane. This leads to the release of apoptogenic factors from the mitochondria such as cytochrome c. Cytochrome c and Apaf1, which is also induced by DNA damage in a p53-dependent manner (6), contribute to the formation of the apoptosome, a necessary step in the activation of caspases. Our data also demonstrates that Peg3/Pw1 is involved in the Bax translocation step. However, Peg3/Pw1 activity appears to require additional uncharacterized factors, because Peg3/Pw1 expression by itself cannot induce neuronal cell death in the absence of p53 or Bax. This suggests that p53 induces other genes in neurons that are essential for cell death following DNA damage. Nevertheless, changes in the expression or distribution of Peg3/Pw1, Bax, and Apaf1, suggest that mitochondrial involvement plays a critical role in neuronal cell death induced in response to DNA damage.