Aneuploid neurons populate the normal adult brain, but the cause and the consequence of chromosome abnormalities in the CNS are poorly defined. In the adult cerebral cortex of three genetic mutants, one of which is a mouse model of the human neurodegenerative disease ataxia-telangiectasia (A-T), we observed divergent levels of sex chromosome (XY) aneuploidy. Although both A-T mutated (Atm)- and transformation related protein 53 (Trp53)-dependent mechanisms are thought to clear newly postmitotic neurons with chromosome abnormalities, we found a 38% increase in the prevalence of XY aneuploidy in the adult Atm-/- cerebral cortex and a dramatic 78% decrease in Trp53-/- mutant mice. A similar 43% decrease in adult XY aneuploidy was observed in DNA repair-deficient Xrcc5-/- mutants. Additional investigation found an elevated incidence of aneuploid embryonic neural progenitor cells (NPCs) in all three mutants, but elevated apoptosis, a likely fate of embryonic NPCs with severe chromosome abnormalities, was observed only in Xrcc5-/- mutants. These data lend increasing support to the hypothesis that hereditary mutations such as ATM-deficiency, which render abnormal cells resistant to developmental clearance, can lead to late-manifesting human neurological disorders.
The overproduction and clearance of embryonic neural progenitor cells (NPCs) and newly postmitotic neurons are essential aspects of cerebral cortical neurogenesis (Blaschke et al., 1996; Kuida et al., 1996; Thomaidou et al., 1997; Pompeiano et al., 2000; Li et al., 2003). Recent findings suggest that some aneuploid NPCs are cleared whereas others survive and give rise to genetic mosaicism in the mature cerebral cortex (Rehen et al., 2001; Osada et al., 2002; Kaushal et al., 2003; Yang et al., 2003). Other studies have implicated altered genetic mosaicism, in the form of chromosome abnormalities and chromosome instability, in the pathogenesis of human neurodegenerative (Potter, 1991; Li et al., 1997; Geller and Potter, 1999; Rolig and McKinnon, 2000) and neuropsychiatric disease (Burd et al., 1985; DeLisi et al., 1994; Lewis et al., 1995; Konstantareas and Homatidis, 1999; Yurov et al., 2001; Oliveira et al., 2003). Methodological considerations, such as the difficulty of identifying aneuploid human neurons in vivo and of obtaining high-quality human brain tissue, limit the direct study of neural aneuploidy and human disease; however, studies in ataxia-telangiectasia (A-T) mutated (Atm)-/- mice suggest that late-manifesting neurodegeneration in A-T patients may reflect the generation and failed clearance of NPCs and newly postmitotic neurons with chromosome abnormalities (Herzog et al., 1998; Rolig and McKinnon, 2000; Allen et al., 2001). To explore a link between neural aneuploidy and A-T, we determined the incidence and character of neural aneuploidy in Atm-/- mutant mice relative to wild-type (WT) cohorts.
A complex array of cerebral cortical phenotypes have been reported for mice lacking Atm or transformation related protein 53 (Trp53), which encodes the tumor suppressor protein p53, and genes that encode nonhomologous end-joining (NHEJ) DNA repair proteins such as Xrcc5, known also as Ku80/86 (Chun and Schatz, 1999a,b). Three NHEJ mutants, Xrcc5-/-, Ku70-/-, or Prkdc-/-, have comparatively mild neurodevelopmental phenotypes (Gu et al., 2000; Chechlacz et al., 2001) that are worsened when Atm is also absent, leading to early embryonic lethality (Sekiguchi et al., 2001). Yet, two other NHEJ mutants, DNA ligase 4 (Lig4-/-) or Xrcc4-/-, present with massive neurodevelopmental apoptosis that is partially rescued by the loss of either Atm or Trp53 (Frank et al., 2000; Gao et al., 2000; Lee et al., 2000; Sekiguchi et al., 2001; Lee and McKinnon, 2002). One explanation for these divergent findings is that the various DNA repair deficits differentially sensitize embryonic NPCs and newly postmitotic neurons to oxidative DNA damage and resultant apoptosis (Karanjawala et al., 2002a,b). It follows then that loss of cell cycle control and related apoptotic signaling, attributable to Atm or Trp53 deficiency, complements lethal DNA repair deficits and permits the survival of Lig4-/- and Xrcc4-/- neural cells with chromosome abnormalities. This reasoning led us to hypothesize that neural aneuploidy would be similarly affected in both Atm-/- and Trp53-/- mutants. To better elucidate the individual aspects of Atm signaling [i.e., combined DNA repair, cell cycle control, and apoptotic signaling (Kastan et al., 2000; Shiloh, 2003)] during neurodevelopment, we compared neural aneuploidy in Atm-/- mutants with neural aneuploidy in Xrcc5-/- [i.e., predominantly DNA repair deficits (Lieber et al., 2003)] and Trp53-/- [i.e., predominantly cell cycle control and apoptotic deficits (Vogelstein et al., 2000)] mutants.
Materials and Methods
Animals. All animal procedures were performed according to protocols approved by the Salk Institute for Biological Studies or the University of California, San Diego animal care and use committee. A. Nussenzweig (National Cancer Institute, Bethesda, MD) kindly provided Xrcc5-/- mice (C57BL/6J background) (Nussenzweig et al., 1996). Atm-/- mice (129S6/SvEvTac background) were generated as described previously (Barlow et al., 1996). Age-matched, male 129S3/SvImj and 129S3/SvImj-Trp53 mice were purchased from Jackson Laboratories (Bar Harbor, ME). Trp53-/- embryos (129S6/SvEvTac background) were a generous gift from J. Wang (University of California, San Diego, La Jolla, CA). All animals were derived from heterozygote breedings within each mutant background with controls derived from littermates of the same background. Females were monitored daily for the presence of a copulation plug. This day was deemed embryonic day 0.5 (E0.5) and embryos were collected at the time of gestation indicated. Only tumor-free animals were analyzed. All analyses were performed in a blinded manner.
Spectral karyotyping. Cerebral cortical hemispheres were dissected from individual embryos between E12.5 and E14.5 and maintained at 37°C in 0.5 ml of OptiMEM media containing 0.02 m glucose, 0.055 m 2-mercaptoethanol, 1× penicillin-streptomycin-glutamine solution, 100 ng/ml Colcemid, and 50 ng/ml basic fibroblast growth factor-2 (all from Invitrogen, Carlsbad, CA) for 3 hr. Hemispheres were dissociated by gentle trituration, and cells were collected by centrifugation at 350 × g for 5 min. The cell pellet was resuspended by drop-wise addition of 0.75 ml of hypotonic solution (0.075 m KCl) while vortexing slowly. After 15 min at room temperature (RT), three drops of fixative (three parts 100% methanol:one part glacial acetic acid) were added and cells were collected by centrifugation. The supernatant was removed to a residual volume of 0.05 ml, and 0.75 ml of RT fixative was added drop-wise while vortexing. After overnight incubation at 4°C, the cells were washed twice in RT fixative, as described above. Cells were stored at 4°C until spread preparation. Adult NPCs were prepared as described by Allen et al. (2001).
Prometaphase-metaphase chromosome spreads were prepared as described previously (Henegariu et al., 2001; Rehen et al., 2001). Fixed cells from appropriate genotypes were selected and blinded until analysis was complete. Results were pooled from at least three individual embryos. Spectral karyotyping (SKY) (Liyanage et al., 1996) was performed on slides that were aged for 3-7 d at RT and then hybridized with mouse SKY paint [Applied Spectral Imaging (ASI), Carlsbad, CA] at 37°C for 24 hr. Excess probe was removed and hapten-labeled nucleotides were detected using the concentrated antibody detection kit (ASI) and following manufacturer's instructions. Karyotypes were determined from 100× micrographs captured using a Zeiss Axioplan 2 (Zeiss, Thornwood, NY), Spectracube interferometer and charge-coupled device camera (ASI), and Spectral Imaging and SKYview software (ASI).
XY fluorescence in situ hybridization. Cerebral cortices were dissected from at least three age-matched WT and mutant adult (8-14 weeks of age) male mice. Nuclei were isolated by detergent lysis, affixed to glass slides, and hybridized with whole chromosome paints (ASI), as described previously (Rehen et al., 2001; Kaushal et al., 2003; Yang et al., 2003). Samples (200-400 nuclei) from each animal were scored for the presence or absence of sex chromosomes and pooled for statistical analysis. No significant difference was detected in the proportions of aneuploid nuclei in the C57BL/6J and 129S6/SvEvTac strains, so we pooled data from these WT nuclei and performed statistical analysis on Atm-/- and Xrcc5-/- mutant strains versus pooled WT cohorts.
Histochemistry. Embryos or adult brains were embedded in Tissue-Tek OCT (Sakura Finetek, Torrance, CA), frozen using dry ice, and stored at -80°C. Before fixation, 10 μm sections were sliced using a cryostat. Matched sections from each genotype were mounted on each slide. “In situ end-labeling plus” (ISEL+) was performed as described previously (Blaschke et al., 1996; Chun and Blaschke, 1997). Sections were counter-stained with 4′,6-diamidino-2-phenylindole (DAPI). At least three animals from each genotype were analyzed, and representative micrographs are presented. Images were acquired using a Zeiss Axiocam and Zeiss imaging software and prepared using Photoshop 6.0 (Adobe Photo Systems, San Jose, CA).
Data analysis. Fisher's exact test was performed using R software (Free Software Foundation's GNU project, Boston, MA). Other statistical analyses were performed, and graphs were prepared using Excel 2000 (Microsoft, Redmond, WA) and Photoshop 6.0. χ2 analysis used 2 × 2 contingency tables (euploid vs aneuploid) to calculate predicted values from observed data. Student's t test was performed on parametric data assuming a two-tailed distribution and homoscedastic variance.
The prevalence of adult XY aneuploidy is increased in Atm-/- mutants and decreased in Xrcc5-/- and Trp53-/- mutants
To study the consequences of impaired DNA repair and/or cell cycle and apoptotic signaling on the prevalence of aneuploidy in the adult cerebral cortex, XY fluorescence in situ hybridization (XY FISH) (Rehen et al., 2001; Kaushal et al., 2003) was performed on nuclei isolated from the cerebral cortex of adult male mice (subsequently referred to as nuclei) and mutant cohorts with germ-line null mutations in Atm, Xrcc5, and Trp53. XY FISH uses X chromosome and Y chromosome whole chromosome paints, conjugated with distinct fluorophores, to determine the rates of gain and loss for each member of the chromosome pair and estimate the overall prevalence of aneuploidy in interphase nuclei. This method allowed us to determine the following: (1) the overall prevalence of adult XY aneuploidy, measured by the percentage of nuclei that had gained (XXY or XYY) or lost (OY or XO) a sex chromosome and (2) the character of adult XY aneuploidy, measured by the relative proportions of XY versus OY versus XXY versus XYY versus XO nuclei. We found an increased prevalence of aneuploid nuclei in Atm-/- mutants (WT, 51 of 902 nuclei, 5.7% aneuploidy vs mutant, 70 of 910 nuclei, 7.7% aneuploidy) and a decreased prevalence in both Xrcc5-/- (WT, 49 of 814 nuclei, 6% aneuploidy vs mutant, 21 of 624 nuclei, 3.4% aneuploidy) and Trp53-/- (WT, 64 of 910 nuclei, 7% aneuploidy vs mutant, 14 of 918 nuclei, 1.5% aneuploidy) mutants (Fig. 1, Table 1). In addition, statistical analysis confirmed a significant difference in the character of XY aneuploidy that was observed in each mutant relative to WT cohorts (Fig. 1, Table 1).
The relationship between adult XY aneuploidy and adult XX aneuploidy or autosomal aneuploidy is not yet known. Although sex chromosome aneuploidy (DeLisi et al., 1994; Samango-Sprouse, 2001) and X chromosome mosaicism (Gleeson et al., 2000; Samango-Sprouse, 2001) have been specifically linked to neurological disease, the mechanism through which XY mosaicism affects normal or pathological brain function is not clear. Altered levels of XY aneuploidy as a reflection of overall aneuploidy can, however, direct further study of the neurodevelopmental parameters that shape overall aneuploidy. It is likely that many of the aneuploid adult nuclei we observed were isolated from the progeny of aneuploid embryonic NPCs (Davis and Temple, 1994; Price and Willshaw, 2000; Qian et al., 2000). Therefore, we next asked whether an altered prevalence of aneuploid embryonic NPCs could account for altered adult XY aneuploidy in these mutants.
Embryonic NPC aneuploidy is elevated when the DNA damage response is compromised
If the prevalence of adult XY aneuploidy is a direct consequence of the prevalence of embryonic NPC aneuploidy, then diminished embryonic NPC aneuploidy should correlate with diminished adult XY aneuploidy and vice versa. Because embryonic NPCs are mitotic, we used SKY (Liyanage et al., 1996; Rehen et al., 2001) to determine the prevalence of embryonic NPC aneuploidy in Atm-/-, Xrcc5-/-, and Trp53-/- embryos relative to WT cohorts. We observed a statistically significant increase in the percentage of aneuploid embryonic NPCs in Atm-/- (+89%Δ) (Fig. 2A vs B) and Xrcc5-/- (+110%Δ) (Fig. 2C vs D) embryos. A similar trend toward increased embryonic NPC aneuploidy was observed in Trp53-/- embryos (+39%Δ) (Fig. 2E vs F). In all cases, most aneuploid NPCs were hypoploid. These data show that in the absence of Atm there is a clear correlation between elevated embryonic NPC aneuploidy and elevated XY aneuploidy in the adult cerebral cortex. Paradoxically, in Xrcc5-/- and Trp53-/- mutants, embryonic and adult aneuploidy were inversely correlated.
Chromosome translocations in vivo are detected in the absence of Atm
One possible explanation for these divergent findings is that the character of aberrant embryonic NPCs (chromosome gain and loss vs chromosome translocations) in some way influences their ability to survive and populate the adult CNS. Previous studies have reported that Atm or Xrcc5 deficiency results in dividing cells that accumulate gains or losses of whole chromosomes as well as chromosomal translocations (Difilippantonio et al., 2000; Allen et al., 2001; Hande et al., 2001; Sekiguchi et al., 2001); however, in Trp53-/- dividing cell populations, chromosome translocations are rarely detected (Ferguson et al., 2000a,b; Sekiguchi et al., 2001). In addition to chromosome gains and losses, the use of SKY allowed us to analyze acutely dissociated embryonic NPCs for the presence or absence of chromosome translocations. Among acutely dissociated Atm-/- embryonic NPCs, chromosome translocations were observed in 2 of 23 aneuploid Atm-/- embryonic NPCs (Fig. 2B, inset), but none were observed in 14 euploid embryonic NPCs. Although chromosome translocations have been observed in Xrcc5-/- and other NHEJ-deficient mouse embryonic fibroblasts (MEFs) (Difilippantonio et al., 2000; Ferguson et al., 2000a,b), we were unable to identify any chromosome translocations in either euploid or aneuploid (n = 31 total studied) Xrcc5-/- embryonic NPCs. Importantly, consistent with analysis of Trp53-/- MEFs and lymphocytes (Ferguson et al., 2000a,b; Sekiguchi et al., 2001), no chromosome translocations were observed among 36 Trp53-/- embryonic NPCs. In a previous study, we observed chromosome translocations and aneuploidy in cultured Atm-/- adult NPCs (Allen et al., 2001); therefore, we prepared and studied similarly cultured Xrcc5-/- adult NPCs in which we also observed chromosome translocations (Fig. 2D, inset) and aneuploidy (mutant, 91% vs WT, 46%; data not shown). The observation of chromosome translocations in cultured, aneuploid Xrcc5-/- adult NPCs, but not in acutely isolated, aneuploid Xrcc5-/- embryonic NPCs, suggests that although either Atm-/- or Xrcc5-/- cells with translocations can survive and proliferate in vitro, aneuploid Xrcc5-/-, but not Atm-/-, embryonic NPCs harboring chromosome translocations are cleared rapidly in vivo.
Despite similar levels of embryonic NPC aneuploidy, only Xrcc5 deficiency leads to elevated embryonic NPC apoptosis
Recently, studies using terminal deoxynucleotidyl transferase-mediated nick end labeling (TUNEL) to detect apoptotic cells demonstrated that mice deficient in Xrcc5-/- have higher levels of apoptosis among newly postmitotic neurons (Gu et al., 2000; Karanjawala et al., 2002a). In contrast, in Trp53-/- mutants, no evidence of elevated neurodevelopmental apoptosis was reported using TUNEL (Frenkel et al., 1999; Klocke et al., 2002) or using a more sensitive assay (M. Pompeiano, S. Rehen, and J. Chun, unpublished observations), termed in situ end-labeling plus (Blaschke et al., 1996; Chun and Blaschke, 1997). We used ISEL+ to definitively establish whether there were elevated or reduced levels of apoptosis in Xrcc5-/- and Atm-/- embryonic NPCs. An 81% increase in the number of ISEL+-labeled cells was detected in the ventricular zone (VZ) of E15.5 Xrcc5-/- (Fig. 3) embryos. In contrast, Atm-/- embryos showed staining similar to WT controls (Fig. 3A). Thus, although embryonic NPC aneuploidy is elevated in all genotypes, only Xrcc5-/- mutants show clear evidence of resultant apoptosis during cortical development. Taken together, these data suggest that the decreased adult XY aneuploidy in Xrcc5-/- mutants reflects a culling of embryonic NPCs with deleterious chromosome abnormalities. In contrast, increased adult XY aneuploidy in Atm-/- mutants is likely the consequence of both an elevated occurrence and a failure to eliminate karyotypically aberrant embryonic NPCs.
The hypothesis that karyotypically aberrant neural cells are involved in the pathology of A-T is based on SKY analysis of cultured Atm-/- adult NPCs (Allen et al., 2001) and inferred from the failed clearance of presumably abnormal neural cells after irradiation (Herzog et al., 1998). The main goal of this study was to directly examine the karyotypes of acutely isolated cells from the developing and mature Atm-/- CNS. In Atm-/- mutant mice, we found an elevated incidence of aneuploidy and chromosome translocations among embryonic NPCs (Fig. 2), no evidence for the neurodevelopmental clearance of additional aneuploid embryonic NPCs (Fig. 3), and elevated XY aneuploidy in the adult cerebral cortex (Fig. 1, Table 1). These observations demonstrate that Atm deficiency promotes both the ontogenesis and survival of aneuploid embryonic NPCs and their progeny. Comparison of Atm-/- and Xrcc5-/- neurodevelopment suggests that DNA repair functions to prevent translocations and limit aneuploidy in a subset of embryonic NPCs, and that Atm-dependent signaling pathways play a significant role in the clearance of severely aberrant embryonic NPCs.
What determines the mosaic composition of the mature cerebral cortex? Although our data (Blaschke et al., 1996; Pompeiano et al., 2000; Rehen et al., 2001) and that of others (Kuida et al., 1996; Thomaidou et al., 1997; Li et al., 2003) demonstrate that apoptotic mechanisms clear many aneuploid embryonic NPCs during development, the basis for this decision is unknown. XY FISH measures aneuploidy for only 1 of the 20 mouse chromosome pairs; therefore, it is only an estimate of the overall prevalence and character (i.e., the mosaic composition) of neural aneuploidy. However, three observations from this analysis suggest that selection based on genetic factors shapes the mosaic composition of the adult cerebral cortex. First, there is a propensity for Y chromosome aneuploidy (Fig. 1, OY + XXY < XYY + XO in all cases) as opposed to X chromosome aneuploidy in the male cerebral cortex. This observation is consistent with a minimal genetic impact of Y chromosome aneuploidy on selection during neurodevelopment. Other mechanisms, such as loss of heterozygousity (Kaushal et al., 2003) or ploidy-dependent gene expression (Galitski et al., 1999) among autosomes, may dictate the prevalence of X chromosome aneuploidy. Second, a conserved bias for chromosome loss is observed in both the developing (Fig. 2) (Rehen et al., 2001) and mature cerebral cortex (Fig. 1, XO + OY > XXY + XYY in all cases). This may indicate that selection during neurodevelopment favors chromosome loss rather than chromosome gain. A third observation that extends this notion is that many additional hyperploid nuclei were observed only in Atm-/- mutants (Table 1, Atm-/- XXY, 2.7-fold change and Atm-/- XYY, 2.2-fold change). This observation is consistent with the hypothesis that normal selection-clearance mechanisms select against hyperploid cells and that these normal mechanisms are compromised during Atm-/- neurodevelopment.
Why is adult XY aneuploidy diminished in Trp53-/- mice when, unlike Xrcc5-/- mutants, there is no evidence for elevated clearance of embryonic NPCs during Trp53-/- development? Because p53 and p53 family proteins affect neuronal survival (Miller et al., 2000), one possibility is that Trp53 is more important for the survival of aneuploid than of euploid NPCs and/or neurons. In this scenario, net developmental cell death would be constant, whereas adult XY aneuploidy is diminished (Table 1, Fig. 1). Additional possibilities include late manifesting apoptosis of aneuploid Trp53-/- neurons (Amson et al., 2000) or an altered contribution of the VZ relative to other germinal zones [e.g., the ganglionic eminence (Anderson et al., 2002)] in the Trp53-/- cerebral cortex. Regardless of the underlying events leading to diminished Trp53-/- adult XY aneuploidy, the additional aneuploidy, chromosome translocations, and resultant cell death observed in Xrcc5-/-, but not Trp53-/-, mutants make it clear that impaired DNA repair and impaired cell cycle control diminish adult XY aneuploidy via distinct mechanisms. The divergent levels of adult XY aneuploidy that we observed in the Atm-/- and Trp53-/- cerebral cortex demonstrates that these proteins have distinct roles in cell cycle control and apoptotic signaling during cerebral cortical neurogenesis and suggests that novel neurodevelopmental signaling may distinguish Atm-dependent rescue of Lig4 and Xrcc4-deficient neurogenesis from Trp53-dependent rescue of these mutants.
This study has significant clinical implications for the treatment of hereditary neurological diseases. In the CNS of A-T patients, altered genetic mosaicism may manifest as neurodegeneration later in life. Moreover, related consequences of altered genetic mosaicism, not necessarily limited to specific genes, could underlie the purported links between chromosome abnormalities and Alzheimer's disease (Potter, 1991; Geller and Potter, 1999), schizophrenia (DeLisi et al., 1994; Yurov et al., 2001), and/or autism (Burd et al., 1985; Konstantareas and Homatidis, 1999; Oliveira et al., 2003). If the mosaic composition of the cerebral cortex is altered during human development, as our study demonstrates for DNA damage-signaling mutant mice, then it may be challenging to use only pharmacological means to treat such patients. In this setting, stem cell-based replacement or augmentation may be a more efficacious therapy for afflicted individuals.
This work was supported by the National Institute of Neurological Disorders and Stroke, V Foundation, and Searle Scholars Program (C.B.); the National Institute of Mental Health, Helen L. Dorris Child and Adolescent Neural and Psychiatric Disorder Institute, and an unrestricted gift from Merck Research Laboratories (J.C.); predoctoral support from a National Institute of General Medical Sciences Pharmacology training grant (M.J.M. and A.H.Y.) and the National Science Foundation (D.K.); postdoctoral support from the PEW Latin American Fellows (S.K.R.); and a Neuroplasticity of Aging Training grant (M.A.K.). We are grateful to Dr. H. L. Borges for critical reading of this manuscript and to F. He and the University of California, San Diego Biostatistics and Informatics Consulting Laboratory for assistance with statistical tests. We thank Dr. F. H. Gage for helpful advice and Dr. H. L. Borges, M. Hemming, G. Kennedy, M. Fontanoz, and D. Mitchell for technical assistance.
Correspondence should be addressed to Dr. Carrolee Barlow, The Salk Institute for Biological Studies, Laboratory of Genetics, La Jolla, CA 92037. E-mail:.
D. Kaushal's present address: Howard Hughes Medical Institute, The Children's Hospital, Boston, MA 02446.
M. J. McConnell, A. H. Yang, M. A. Kingbury, S. K. Rehen, and J. Chun's current address: Department of Molecular Biology and Helen L. Dorris Institute, The Scripps Research Institute, La Jolla, CA 92037.
Copyright © 2004 Society for Neuroscience 0270-6474/04/248090-07$15.00/0