Skip secondary menu

The Role of DNA Damage Response Defects in Neurogenetic Diseases


The Role of DNA Damage Response Defects in Neurogenetic Diseases
Co-sponsored by:
A-T Children's Project
National Institute of Neurological Disorders and Stroke
Tarrytown Conference Center, Tarrytown, New York
July 29 - August 1, 2001

DNA Meeting Photo




Sunday, July 29

4:30-6:30 pm: Registration: Registration Lobby

7:00 pm: Reception: West Terrace, Biddle House

8:00 pm: Dinner: Sun Porch, Biddle House

10:00 pm: Sleepy Hollow Pub


Monday, July 30

7:30 pm: Full Breakfast: Winter Palace, Biddle House

8:30 pm: Introduction: Brad Margus, Dave Frohnmayer

8:40 pm: The link between DNA damage, disease and neurodegeneration: Remarks from Session Chairs: Yossi Shiloh, Alan Lehman

8:45 pm: A-T: Yossi Shiloh
9:00 pm: A-TLD and NBS-1: Pat Conconnan
9:15 pm: Excision repair (XP/CS/TTD): Alan Lehman
9:30 pm: Helicase deficiency: Nathan Ellis
9:45 pm: Bloom syndrome: James German
10:00 pm: Fanconi anemia: Arleen Auerbach
10:15 pm: Werner syndrome: Ray Monnat

10:30 pm: Break

10:45 pm: Discussion

12:00 noon: Lunch: Biddle House

1:30 pm: Responses to DNA damage: Remarks from Session Chair: Mike Kastan
1:35 pm: General response pathways: Al Fornace
1:50 pm: DS break response pathways: Peggy Jeggo
2:05 pm: Substrates of ATM: Mike Kastan
2:20 pm: ATM function in the IR pathway: Martin Lavin
2:35 pm: ATM and ATR: Responding to damage: Kathy Brumbaugh
2:50 pm: Fanconi anemia and genomic stability: Alan D'Andrea
3:05 pm: Macromolecular response machinery: Jun Qin

3:20 pm: Break

3:35 pm: Homologous recombination repair: Larry Thompson
3:50 pm: DNA damage response in yeast: Ted Weinert
4:05 pm: BRCA1 and DNA damage: Junjie Chen
4:10 pm: Telomeres: Tej Pandita
4:15 pm: DNA damage response in Aspergillus: Steve Harris
4:20 pm: UV damage responses: Kathleen Dixon

4:30-5:30 pm: Discussion

7:00 pm: Reception: Statue Garden

8:00 pm: Dinner: Winter Palace, Biddle House

10:00 pm: Sleepy Hollow Pub


Tuesday, July 31

7:30 pm: Full Breakfast: Winter Palace, Biddle House

8:30 pm: Initiation of DNA damage in the nervous system: Remarks from Session Chair: Fred Alt

8:35 pm: Links between the immune and nervous system: Fred Alt
8:50 pm: DNA damage during CNS development: Peter McKinnon
9:05 pm: Endogenous DNA breaks/ V(D)J-like damage: Jerold Chun:
9:20 pm: DNA repair in neural cells: PJ Brooks

9:35 pm: Discussion

10:30 pm: Break

10:45 pm: DNA damage selectivity in the nervous system: Karl Herrup
11:00 pm: DNA damage selectivity in the nervous system: Nat Heintz
11:15 pm: Discussion

12:00 noon: Lunch: Winter Palace, Biddle House

1:30 pm: DNA damage and neural apoptosis: Remarks from Session Chair: Gene Johnson

1:35 pm: Control of neural apoptosis: Eugene Johnson
1:50 pm: Survival pathways and neural development: Michael Greenberg
2:05 pm: Mechanisms of DNA damage induced neural death: David Park

2:20 pm: Discussion

3:30 pm: Break

3:45 pm: Therapeutic approaches to CNS treatment: Xandra Breakefield
4:00 pm: Therapeutic approaches to CNS treatment: Allen Mandir, Tom Crawford
4:15 pm: Possible roles of ATM in homeostasis:: Yossi Shiloh
: Implications on treating A-T patients

4:30-5:30 pm: Discussion

7:00 pm: Reception and Barbecue: West Terrace, Biddle House

10:00 pm: Sleepy Hollow Pub


Wednesday, August 1

7:30 pm: Full Breakfast: Winter Palace, Biddle House

8:30 pm: Oxidative stress and neurodegeneration: Remarks from Session Chair: Mark Mattson

8:35 pm: Oxidative stress: Rod Levine
8:50 pm: Oxidative stress & neurodegeneration: Mark Mattson
9:05 pm: ATM & oxidative stress: Carrolee Barlow
9:20 pm: ATM & oxidative stress: Yossi Shiloh

9:35 pm: Discussion

10:30 pm: Break

10:45 pm: Giovanna Spinella

11:00 pm: Future Directions
Discussion Leaders: Gene Johnson, Mike Kastan

12:00 noon: Lunch: Winter Palace, Biddle House




The purpose of this meeting was to highlight past and current studies on A-T and other genetic diseases in which proteins involved in DNA damage responses are compromised, as well as research in the field of neurobiology. The meeting brought together a seemingly disparate group of researchers with the goal of delineating hypotheses and model systems that might uncover links between DNA damage response defects and neurogenetic diseases. Of course, a prime example of such a disease is ataxia telangiectasia (A-T).

A-T is a chromosomal instability syndrome that affects multiple organ systems. Patients diagnosed with A-T exhibit immunodeficiency, increased cancer susceptibility, premature aging, and neurodegeneration. It is the neurodegeneration that usually leads to the early demise of these patients, most often in the second or third decade of life. The protein product of the gene mutated in A-T, A-T-mutated (ATM), is a large protein kinase involved in cell cycle checkpoint and genotoxic stress responses. Although research has unveiled some of the signaling pathways governed by ATM in checkpoint and stress responses, little is known about the role ATM plays in surveillance of the nervous system. In part, this is due to the lack of a suitable model system. Atm -/- mice display signs that mimic the symptoms of A-T patients, but do not recapitulate all aspects of the disease, particularly the neurodegenerative syndrome.

The conference began with opening remarks from Brad Margus, President of the A-T Children's Project, and Dave Frohnmayer, co-founder of the Fanconi Anemia Research Fund, Inc. Brad Margus explained that the A-T Children' Project was organized based on the model provided by the Fanconi Anemia Research Fund. Due to the cancer-predisposition of both A-T and Fanconi Anemia patients, a meeting invitation was extended to researchers and members of the Fanconi Anemia group.


Session I: Chairs Yossi Shiloh and Alan Lehmann

The link between DNA damage, disease, and neurodegeneration

The opening session centered on a number of syndromes in which mutation of a protein involved in DNA damage responses leads to a disease phenotype. Where applicable, a link to neurodegeneration was highlighted. There are a number of syndromes whose phenotype resembles A-T. Patients with Nijmegen breakage syndrome (NBS) are immunodeficient and exhibit increased cancer incidence, but are not ataxic. The nervous system is affected in NBS, however, as these patients present with microcephaly and mental retardation. Recent studies demonstrated that the gene mutated in NBS, Nbs-1, is a downstream substrate of the ATM protein kinase, and is one member of a protein complex that binds to DNA double-strand breaks (dsb). The other two members of this Nbs-1 complex are the Rad50 and Mre11 gene products. Interestingly, another disease that bears significant similarity to A-T, the A-T-like disorder (ATLD), is characterized by Mre11 deficiency. Patients with ATLD are more A-T-like than NBS-like in that they are ataxic, but they do not present with telangiectasies. There are yet other patients who present with an NBS-like disorder (NBSLD) that is described by radiosensitivity, cancer predisposition, growth and mental retardation, with a lack of immunodeficiency and chromosomal translocations. These diseases, which highlight the importance of the ATM signaling pathway and its functional linkages to the Nbs-1/Mre11/Rad50 complex, predict that similar syndromes may arise when other genes in this pathway are mutated.

In addition to the A-T and A-T-related diseases, other cancer-prone genetic disorders were discussed. In two disorders in which patients display DNA repair-deficiencies, Xeroderma pigmentosa (XP) and Cockayne's syndrome (CS), patients are extremely sensitive to UV light and exhibit an increased incidence of skin cancer. XP patients have a defect in nucleotide excision repair (NER) but display no neurologic sequelae. CS patients show defects in transcription-coupled repair (TCR) and the repair of oxidative base damage, and intriguingly, bear neurologic pathologies, including an ataxic gait and demyelination. In the two helicase deficiencies discussed, Bloom's and Werner's syndromes, patients have an increased cancer predisposition but no neurologic disease is evident.

Fanconi anemia is grouped with these other syndromes due to the heightened cancer incidence in Fanconi patients. Although no motor defects are described for these patients, they do exhibit growth retardation, microcephaly, and learning disabilities.

Overall, the common factor in the above disorders is that the disease endows patients with a predisposition to cancer. In only some of these syndromes do patients actually exhibit neurodegenerative features, which suggests that any defect in the response to genotoxic stress does not lead to neurodegeneration. Rather, certain tissues may be more dependent on a specific type of damage response. For instance, terminally-differentiated cells are likely more dependent on genotoxic stress responses than tissues with proliferative potential. Cells that are actively proliferating, however, are usually extremely sensitive to genotoxic insults, unless the cell can repair damage prior to cell division. It is also possible that mutagenesis is an intrinsic problem in these cells or tissues rather than the result of a secondary sensitivity to radio- or chemotherapy.


Session II: Chair Mike Kastan

DNA damage response

One common feature of the disorders discussed at this meeting is that cells from these patients show various defects in genotoxic stress responses, and the resultant genomic instability may explain the heightened incidence of cancer. For this reason, DNA damage responses have been extensively studied. The main question discussed at this meeting was whether or not a defect in DNA damage responses also contributes to neurodegeneration.

The ATM protein kinase and other phosphoinositide 3-kinase related kinases (PIKKs) have all been molecularly cloned within the last decade. The large size of both the gene and the gene product presents a major obstacle to straightforward structure/function analysis of the individual domains of these proteins. All PIKK family members contain a highly conserved catalytic domain that bears homology to the kinase domain of PI 3-K. Outside of this region, there is very little sequence homology among family members and it is speculated that these domains may interact with other proteins to facilitate the different functions associated with individual PIKKs.

The ATM protein kinase primarily responds to (double-strand break) dsb-inducing agents, such as ionizing radiation (IR) and radiomimetic agents. The related family member, A-T- and Rad3-related (ATR), responds to IR at a later stage of the response, appears to be required for the maintenance of genome integrity during the normal cell cycle, and serves as the primary activator of the pathways triggered by cellular exposure to ultraviolet light (UV) or inhibitors of DNA replication, such as hydroxyurea (HU). A third PIKK, DNA-PKCS, is critical for the DNA repair process of non-homologous end joining (NHEJ). Cells that are deficient in any of these activities display cell cycle checkpoint defects and increased sensitivity to genotoxic agents.

Recently, the Robert T. Abraham laboratory has cloned and characterized a new PIKK family member, ATX. Kathy Brumbaugh, representing the Abraham laboratory, highlighted the main structural difference between previously identified PIKKs and ATX. Notably, there is a large insert region carboxy-terminal to the kinase domain of ATX that is not present in any of the other PIKKs. ATX is found in both the nucleus and cytoplasm, and like ATM, ATR, and DNA-PKCS, exhibits strong phosphorylating activity toward the Ser-15 site in p53. Exposure of human cells to ultraviolet (UV) light stimulates ATX kinase activity and triggers the appearance of ATX-containing nuclear foci. Cells rendered deficient for ATX function exhibit reduced basal viability and increased sensitivity to killing by UV or IR. Conversely, overexpression of ATX in ATM-deficient fibroblasts complements the radiosensitive phenotype of these cells. Thus, ATX is another PIKK that, like ATR, may signal through partially overlapping pathways with ATM. Given this potential overlap, pharmacologic or genetic strategies that increase the activity of ATR or ATX in AT cells may alleviate some of the disease phenotype.

The G2/M phase checkpoint is the critical place in the cell cycle that ensures a cell has accurately replicated its DNA prior to cell division. ATM and ATR are critical enforcers of this G2 checkpoint. However, studies done by Al Fornace and colleagues have demonstrated that the p38 mitogen-activated protein kinase (MAPK) may also play a role in mediating the G2/M checkpoint after cells are exposed to UV light. Inhibitors of the p38 kinase abrogate the G2/M checkpoint after UV treatment, which reinforces the long-known concept that UV light activates cytoplasmic signaling pathways, such as MAPKs, in addition to damaging DNA. UV light exposure is only one example of an agent that does not simply induce DNA damage; cells exposed to IR, in addition to incurring DNA damage, will generate reactive oxygen species that modify proteins and lipids within the cell. Therefore, any genotoxin or cell stress will generate more than one type of damage, as the primary lesion may be processed or trigger other secondary effects.

Kathleen Dixon presented data consistent with this idea. Current literature suggests that A-T cells have a normal response to UV. In contrast, Dixon's data suggest that UV lesions are processed to generate secondary dsbs, and thus activate ATM. While IR induces dsbs at 1 h, UV results in the appearance of dsbs at 8 h due to disruption of a replicative intermediate. The p34 subunit of replication protein A (RPA) is a critical component of several DNA repair pathways, including post-replicative repair. RPA is phosphorylated and undergoes a mobility shift after cells are exposed to UV light. This UV-induced phosphorylation first appears at 4 h, with a peak at 12 h, and is partially defective in A-T cells. The ATM-dependence of UV-induced RPA phosphorylation is restricted to the later time points. Using a comet assay to detect dsbs, Dixon demonstrated an accumulation of dsbs in normal cells at 24 h after UV light exposure, and that the number of UV-induced dsbs is higher in A-T cells. If repair does not occur, the replication fork may attempt to bypass the UV lesion and result in dsb formation, thus activating ATM.

Dixon also presented results from an in vitro assay system that measures DNA repair. This microhomology-mediated mutagenesis assay involves the addition of a restriction-digested plasmid to cell-free extracts from different cells. Extracts from cells that lack ATM are able to repair the strand break, but with reduced fidelity. This defect is rescued by restored expression of ATM, as nuclear extracts from these cells repair the break with increased fidelity.

In addition to the confusion surrounding the involvement of ATM in UV responses, a discrepancy exists in the literature regarding the role ATM plays in the G2 checkpoint. If AT cells are exposed to IR while they are in the G2- phase of the cell cycle, they do not arrest. However, AT cells that are irradiated at any other point in the cell cycle accumulate in G2, or exhibit a protracted G2 arrest. Mike Kastan presented data that may delineate the roles of ATM in these two situations. It turns out that the method used to identify the G2 phase of the cell cycle is important. Classically, a propidium iodide (PI) stain to measure DNA content is used to evaluate cell cycle distribution. In a PI-stained population, cells in G1 phase have a 2N DNA content, cells in G2/M have a 4N DNA content, and the S phase population is intermediate. Because the PI stain does not discriminate between G2 and M phases of the cell cycle, Kastan and others have used an antibody that recognizes phosphorylated histone H3 to define the G2- phase of the cell cycle. Cells that are exposed to IR and then stained with PI show a radiation dose-independent accumulation in G2; in other words, no difference is seen between exposure of cells to 1 and 10 Gy IR. G2 accumulation is due to the lack of an S phase checkpoint, as an Nbs1 mutated protein (S343A) causes similar accumulation in G2. This G2 accumulation is also ATM- and BRCA1-independent. If histone H3-phosphorylation is used as a measure of G2, higher doses of IR will result in more cells that arrest in G2. This measures the early G2 checkpoint that is ATM- and BRCA1-dependent. Thus, ATM is required to enforce the early G2 checkpoint, but the prolonged G2 arrest is not dependent on ATM. It has been proposed that this prolonged arrest is mediated by the related ATR kinase.

Over the past several years, a number of ATM, ATR, and DNA-PKCS substrates have been identified. The challenge has been to ascertain the functional relevance of phosphorylation of these proteins. A number of presentations at this session outlined either new substrates for these PIKKs, and/or elucidated the significance of phosphorylation at these sites. Kastan and coworkers analyzed two BRCA1 proteins mutated at Ser residues modified by ATM. Specifically, the phosphorylation of Ser-1387 and Ser-1423 are important for enforcing the S and G2 phase checkpoints, respectively. Because the converse is not true, it appears that these specific modifications of BRCA1 result in the activation of distinct downstream signaling pathways.

Robert T. Abraham and colleagues identified the hRad17 protein as both an ATM and ATR substrate. hRad17 bears similarity to the replication factor C group of proteins that are involved in loading the replication processivity factor PCNA onto DNA. Initial results from hRad17 studies suggest that it may load the PCNA-like checkpoint sliding clamp (CSC), composed of hRad1, hRad9, and hHus1, onto damaged DNA. ATM and ATR phosphorylate two Ser residues in hRad17, and cells that express a nonphosphorylatable mutant hRad17 (hRad17 2A) are unable to enforce the G2 checkpoint after IR and undergo increased cell death in response to multiple genotoxins. In addition, hRad17 2A is unable to associate with hRad1, a member of the CSC. Thus, the phosphorylation of these two hRad17 Ser residues is critical for the enforcement of the G2 checkpoint and association of hRad17 with the CSC.

A potential signaling link between A-T and Fanconi anemia was presented. Alan D'Andrea highlighted recent results from their studies of the Fanconi D protein. Cells that lack the FANCD2 protein exhibit radioresistant DNA synthesis (RDS), a well-known hallmark of A-T cells, and are hypersensitive to mitomycin C (MMC). The Fanconi A, C, G, F, and E proteins are in a preexisting nuclear complex, while the Fanconi D protein only associates with this complex after damage. Interestingly, this damage-inducible complex also contains BRCA1. FANCD2 is differentially modified after exposure to MMC or IR. It is mono-ubiquitinated at Lys-561 after MMC treatment but phosphorylated at Ser-222 after IR. A-T cells are competent to mediate the ubiquitination but not the phosphorylation of FANCD2. Interestingly, FANCD2-/- cells that express a S222A FANCD2 are complemented for MMC hypersensitivity, while cells that express a K561R FANCD2, which is unable to be ubiquitinated, are complemented for RDS. These results suggest that Ser-222 phosphorylation of FANCD2 is an ATM-dependent event that prevents RDS, while the ubiquitination of Lys-561 is an ATM-independent step that activates signaling pathways involved in MMC resistance.

A connection between A-T and another disease discussed at this meeting was presented by Martin Lavin. Using a yeast 2-hybrid screen to search for ATM-interacting proteins, Lavin and colleagues identified the BLM helicase. The BLM helicase copurifies and coimmunoprecipitates with ATM. This association was mapped to the helicase region of BLM and an amino-terminal fragment of ATM. They demonstrated that the Bloom's syndrome helicase (BLM) is also an ATM in vitro substrate, but the BLM protein is not reproducibly phosphorylated in response to IR. BLM is, however, phosphorylated in nocodazole-treated cells, which are arrested in mitosis. It is unknown whether this nocodazole-induced phosphorylation is dependent on ATM, and may instead reflect an ATR-dependent modification.

Another critical area of investigation in the DNA damage arena is how responses to genotoxins are initiated. Junjie Chen presented immunofluorescence evidence regarding proteins that bind to IR-induced dsbs. One of the downstream effector kinases of ATM, Chk2, exhibits a diffuse immunostaining pattern in the nucleus. In contrast, antibodies that recognize the phosphorylated form of Chk2 illuminate IR-induced foci, and colocalize with antibodies against the histone H2ax protein, Nbs1, and a p53-binding protein 53BP1. Therefore, Chen proposed a model whereby a dsb is first demarcated by histone H2ax, then recruits ATM or ATR, followed by Nbs1, Chk2, and 53BP1.

Many of the hypotheses generated for mammalian cell genotoxic stress responses are based on paradigms set forth by studies in lower eukaryotic model systems. Ted Weinert presented a model for chromosomal instability in the budding yeast, Saccharomyces cerevisiae. When a test chromosome that contained a fragile site was introduced into rad9- yeast, it was significantly less stable than in wild-type yeast. DNA breaks are induced in the fragile site by limiting replication with HU. His hypothesis is that ScRad9 is involved in the regulation of a stalled replication fork. In the presence of ScRad9, forks resume replication after repair and there is no resultant error, while in the absence of ScRad9, an alternative pathway is activated to cause breaks or chromosome loss. ScRad9 is thought to lie downstream of the ATR ortholog, ScMec1, and upstream of the human Chk2 ortholog, ScRad53 in the pathway that responds to secondary DNA damage after replication stress. Therefore, Weinert's findings are consistent with this model and suggest a specific role for ScRad9 in the regulation of stalled replication forks.

To close this session, Steve Harris described his investigation of the DNA damage response in the hyphal fungus, Aspergillus nidulans. Harris described the cloning and characterization of orthologs to the ATR, Rad26, Rad51, and RecQ helicase (similar to BLM and Werner's helicases) proteins. Thus, A. nidulans is another genetically tractable model system in which genotoxic stress responses may be studied to generate testable paradigms for mammalian cells.


Session III: Chair Fred Alt

DNA damage in the nervous system

This session centered on the central theme that echoed throughout the conference. The atm-/- mice display many of the signs associated with A-T, but do not develop neurodegenerative disease. Fred Alt chaired this session, which highlighted various deficient mouse model systems as well as some basic paradigms in neuronal biology.

Thanks to the gene-targeting efforts ongoing in numerous laboratories, many deficient mouse models that lack factors involved in immune receptor rearrangement and double-strand break (DSB) repair are available. All of these mice exhibit a block in lymphocyte development, also known as severe combined immunodeficiency (SCID), while only a subset show increased sensitivity to IR and/or neuronal defects. The recombinase activating gene (Rag)1 or Rag2 deficiency generates mice with a block in lymphocyte development, but they are otherwise normal. Classical SCID mice are deficient in the expression of DNA-PKCS and exhibit variable sensitivity to IR, but reveal no proliferation defects. Mice deficient in Ku70 or Ku80, which are DNA-PKCS interacting proteins, display IR sensitivity, cell proliferation defects, and neuronal apoptosis. Of all the SCID mice, the DNA ligase IV deficient animals display the most severe phenotype, as neuronal apoptosis leads to late embryonic lethality. The ligase IV lethality can be rescued by crossing the mouse with a p53- or atm-deficient strain, which suggests that an ATM-p53 checkpoint pathway is responsible for the early death of these animals. These ligase IV/p53- or ligase IV/atm-deficient mice still demonstrate a block in lymphocyte development, IR sensitivity, cell proliferation defects, and increased neuronal apoptosis compared to normal mice.

Peter McKinnon led one of the groups that generated the ligase IV-/- mice and crossed them with either atm-/- or p53-/- mice. He presented findings that the ligase IV-/-/atm-/- mice show some cerebellar abnormalities, with some regions displaying loss of Purkinje cells. In addition, the McKinnon group investigated cell death in specific neuronal tissues of atm-deficient mice. Atm-/- mice display no detectable overall neuronal apoptosis in response to IR, but the retinas of these mice reveal enhanced sensitivity to IR-induced death. Thus, it appears that ATM-dependent cell death in the nervous system is a tissue or cell-type specific process, with the retina as an example of a tissue that displays ATM-independent cell death.

These studies of retinal cell death utilized a new assay to detect apoptotic or dying cells. The assay, in situ end labeling plus (ISEL+), is a variation on the traditional TdT-mediated dUTP-X nick end labeling (TUNEL) assay that exhibits higher sensitivity. In the TUNEL assay, the enzyme terminal deoxynucleotide transferase (TdT) adds fluorescently-labelled dUTP onto broken DNA strands. ISEL involves TdT addition of digoxigenin-labelled rather than fluorescently-labelled oligonucleo-tides, which makes ISEL+ ten times more sensitive than TUNEL. Even so, the discussion that surrounded this technique questioned whether it was sensitive enough to detect early apoptotic cells or later stage death. One participant gave a reasonable estimate of how many DNA breaks are necessary to obtain an ISEL+ spot that was in the thousands. Furthermore, the question of whether normal replication-induced dsbs were detected with ISEL+ arose. Cells that incorporate BrdU, which is a measure of cells that are replicating their DNA, are not all ISEL+, so some of the dsbs generated during normal DNA replication are not detected, probably due to relatively efficient repair of these breaks. Finally, if ISEL+ is not a reliable assay for apoptosis, the discussion centered on which assay might allow early detection of apoptosis in tissues. The only other assay mentioned was detecting activated caspase 3, because this enzyme plays a critical role in developmental apoptosis of the central nervous system.

Carrolee Barlow presented evidence on the role of ATM in adult neurogenesis. The brains of A-T patients at birth display no gross abnormalities, but undergo neurodegeneration during the course of their life. Consistent with this fact, atm-/- mice are also born with brains that exhibit no major malformation. On the contrary, they do not develop the neurodegenerative sequelae seen in A-T patients. Barlow and coworkers studied adult neural progenitor cells and identified high levels of ATM protein in dividing neural progenitors. The amount of ATM protein in these progenitors decreases upon differentiation into various neural cell types. Progenitor cells from mice that lack ATM are hyperproliferative and exhibit random genomic instability. Perhaps because of this high rate of proliferation, the atm-/- progenitors do not differentiate into the many required cell types, such as neurons or oligodendrocytes. In addition, loss of ATM renders cells less responsive to stimuli that normally promote neurogenesis. Thus, ATM is required to differentiate progenitor cells into the many neural cell types required for population of the adult brain, and appears to control proliferation of these cell types.

In light of the fact that there is currently no suitable model system to study the A-T neurodegenerative phenotype, perhaps other models of neurodegeneration can provide fuel to generate testable hypotheses based on paradigms in basic neuronal biology. Karl Herrup discussed his recent findings in the field of neuronal apoptosis. As neurons leave the ventricular zone and become post-mitotic, only a low level of adult neurogenesis occurs. Once these neurons have differentiated, re-entry into the cell cycle leads to death. In Herrup's model of Alzheimer's, the neurons contain substantial cell cycle markers (cyclins, PCNA,etc), but the cell cycle cannot be completed and neurons become polyploid. The neurons remain polyploid for months before they succumb to a terminal apoptotic process, which ultimately causes the neurodegeneration in Alzheimer's disease.

Thus, neurons with an intact post-mitotic checkpoint are able to survive. This idea is consistent with data from mice that lack expression of the pRb tumor suppressor protein. Rb loss leads to continued re-entrance into the cell cycle and cell death in the central and peripheral nervous systems. Furthermore, tumors of the CNS are not neuronal in origin because expression of an oncogene in a post-mitotic neuron is lethal. Finally, the loss of contact with surrounding cells leads to cell cycle entry and death. Based on this premise, a discussant proposed that perhaps neurons in A-T patients are unable to enforce checkpoints, neurons re-enter the cell cycle, and this is the cause of the A-T neuronal cell death.


Session IV: Chair Gene Johnson

DNA damage and neural apoptosis

his report describes the consensus outcome of an interdisciplinary workshop held at the National Institutes of Health in April 2001. The purpose of the workshop and this document is to define the terms "spasticity", "dystonia", and "rigidity", as they are used to describe clinical features of hypertonia in children. The definitions presented here were designed to allow differentiation of clinical features even when more than one is simultaneously present. We encourage the development of clinical rating scales based on these definitions, and we encourage research to relate the degree of hypertonia to the degree of functional ability, change over time, and societal participation in children with motor disorders.

The next session explored mechanisms of neuronal apoptosis and what role DNA damage might play in effecting this response. The chair of this session, Gene Johnson, began with a brief description of the different types or stages of cell death seen in neurons. First, cells in the early stage of proliferation or immediately after they become post-mitotic are susceptible to undergoing apoptosis. Second is the classical neuronal programmed cell death that occurs during development. Third, experimentally-induced neuronal apoptosis through genetic manipulation, axotomy, or addition of stressors to mimic a disease process. Finally, the poorly defined chronic neurodegenerative disease, which is the focus of this meeting, results from irreparable cell death. Our limited understanding of the neuronal apoptosis seen in chronic neurodegenerative disease is due to the confines of current model systems as well as the lack of sensitive assays to detect low levels of apoptosis.

The two biochemically-defined pathways of apoptosis include the intrinsic pathway, which comprises factor withdrawl- and stress-induced apoptosis, and the extrinsic pathway, which is signaled through death receptors. Although these two pathways involve the activation of the apaf-1/caspase-9 and the caspase-8 signaling cascades, respectively, they ultimately converge to activate caspase-3 and cleave downstream substrates.

Johnson and colleagues studied nerve growth factor (NGF)-withdrawl-induced apoptosis of sympathetic neurons, which is one of the intrinsic pathways. The biochemical events that occur after NGF withdrawl include increased reactive oxygen species (ROS), and increased c-jun, CyclinD1, and c-myb levels. Bax and BIM, two proapoptotic bcl-2 family members are required for death, while treatment with the caspase inhibitor BAF prevents NGF-withdrawl-induced death. This apoptosis can also be prevented by pretreatment with antioxidants or expression of dominant negative c-jun.

In addition to NGF-withdrawl induced death, Johnson and coworkers investigated araC-induced death. AraC, a nucleotide chain terminator that may induce oxidative damage, is a genotoxic stress that, like NGF-withdrawl, induces apoptosis through the intrinsic pathway. The biochemical profile of araC-induced death is similar to NGF-withdrawl, and induction of the pro-apoptotic Bax is required. An interesting finding came from studies that utilized cephalon 1347 (CEP1347), an inhibitor of mixed-lineage kinase (MLK) kinases. This agent, which blocks BIM induction, is a long-term neuroprotectant in NGF-deprived neurons. However, it is ineffective at blocking araC-induced death. The dependence of both apoptosis pathways on Bax suggests that NGF-withdrawl and araC-induced death pathways converge at some point, but the results from the CEP1347 studies are consistent with a model whereby the upstream pathways are different.

Michael Greenberg studied neuronal apoptosis in a different cellular model system. His model is cerebellar granule neurons, in which brain-derived neurotrophic factor (BDNF) is required for survival. These neurons are transfectable and require the phosphoinositide 3-kinase (PI 3-K)/Akt pathway for survival. The PI 3-K/Akt pathway targets a family of Forkhead transcription factors, which include FKHRL1. Forkhead family members act to induce transcription of several proapoptotic genes. Greenberg and coworkers mapped Akt phosphorylation sites on FKHRL1 and verified the modification of these sites in intact cells. They investigated the mechanism by which Akt phosphorylation of FKHRL1 leads to cellular survival. Phosphorylation of FKHRL1 in cells exposed to BDNF causes a change in subcellular localization. The phosphorylated FKHRL1 is found in the cytoplasm, while a mutated FKHRL1 that cannot be phosphorylated remains strictly nuclear and induces apoptosis. Thus, it appears that Akt promotes cell survival by relocalization of FKHRL1 from the nucleus to the cytoplasm.

Due to the fact that FKHRL1 is a transcriptional activator, a cDNA array approach is underway to identify putative target genes. Thus far, 19 known cDNAs and 16 expressed sequence tags (ESTs) have been identified. Due to a confidentiality agreement with a private company, Greenberg could not divulge the identity of these cDNAs. However, he did report that 5 are apoptosis-related, 9 are cell cycle-related, and 5 are stress-responsive or known to be involved in G2/M arrest.

In the second half of this session, three speakers outlined current and potential future therapeutic approaches for treatment of A-T and other neurodegenerative diseases. Xandra Breakefield discussed the possibility of gene therapy for A-T. A number of factors are considered in the design of gene therapy protocols. After a vector system is identified, safety concerns need to be addressed and a suitable preclinical model established. The most challenging problems associated with ATM gene therapy are the size of the encoding cDNA and the lack of an appropriate animal model system that recapitulates the neurodegenerative phenotype. There are clear limitations on which vectors can be utilized due to the size of the ATM cDNA, and truncation or other manipulation of this cDNA is not possible if kinase activity is to be maintained. Breakefield's group is testing vectors in atm-/- mice and looking for expression in the brain. The lack of neurodegeneration in these mice makes it difficult to predict the potential advantages of restoration of ATM expression in the brain of the mouse. As an initial test of a gene therapy vector for ATM, complementation of T cell function may be a better marker for reconstitution, due to the fact that these mice do exhibit a defect in T cell function.

Allen Mandir and Tom Crawford discussed potential opportunities for developing new therapies for A-T patients. The best-case scenario would be to treat A-T patients in the asymptomatic or "presymptomatic" phase of the disease. In lieu of this early treatment, one would hope to stabilize the course of the disease or reverse the symptoms. When considering gene therapy as a treatment option, it is unclear what would be the consequence of introducing ATM into damaged neurons. In this case, one would want to introduce ATM at the earliest possible stage, prior to the accumulation of neurological damage. Another important question that would need to be addressed is how to deliver the vector for therapeutic benefit. Finally, the measure of therapeutic benefit would need to be established. These could include detection of improvement or stabilization of defined clinical signs, and activities of daily living or quality of life issues.

Yossi Shiloh completed this session with his report of a role for ATM in processes that do not involve DNA damage. Specifically, cytoskeletal abnormalities have been reported in A-T cells in culture. In recent work, Shiloh and colleagues identified highly polymerized F-actin in the perinuclear region of A-T lymphoblasts and A-T fibroblasts. The overall actin fiber organization in AT fibroblasts is also impaired, and this is complemented by ATM expression. Furthermore, A-T cells exhibit impaired cell motility and show impaired activity in a wound-healing assay, which is also corrected by ATM expression.

One additional feature of A-T cells in culture is their increased demand for certain growth factors. Two recent publications, one from Shiloh, MaCauley, and colleagues and one from Peter Glazer, demonstrate reduced insulin-like growth factor-1 receptor (IGF-1R) levels on A-T cells. Interestingly, increased IGF-1R expression on A-T cells can partially complement the poor growth of these cells in culture as well as their radiosensitivity. These findings support a role for ATM in non-genotoxin-induced responses, which could include stress that results from general cellular metabolism or cell culture.


Session V: Chair Mark Mattson

Oxidative stress and neurodegeneration

The next session of this meeting focused on the relationship between oxidative stress and neurodegeneration. Rod Levine discussed a number of experimental methods useful in the analysis of oxidative stress effects. Levine emphasized the importance of detection of more than one marker for oxidative stress because a single marker may give misleading results. One can measure increased antioxidant proteins such as superoxide dismutase (SOD) or glutathione peroxidase, changes in low molecular weight antioxidants such as vitamin C, vitamin E, and glutathione, or assess macromolecular damage to DNA, lipid, or protein. One common oxidative modification of proteins is the addition of a carbonyl group. The amount of carbonyl-bearing proteins is elevated in many human diseases including rheumatoid arthritis, Parkinsons's disease, Alzheimer's disease, ALS, aging, diabetes, and Werner's syndrome. In healthy individuals, levels of carbonyl-bearing proteins dramatically increase during the last 1/3 of life. In contrast, Werner's syndrome patients exhibit these high levels of carbonyl-modified proteins between the ages 10 and 20.

In collaboration with Carrolee Barlow, Levine measured levels of carbonyl-bearing proteins in wild-type and atm-/- mice. Multiple tissues were examined and Levine and Barlow demonstrated no differences in the levels of carbonyl-bearing proteins between these mice. However, there is clear evidence of oxidative stress in atm-/- mice as measured by other markers. The thymus and testes exhibit increased F2 isoprotanes, the brain displays increased nitrotyrosine, and the cerebellum reveals a 6-fold increase in heme oxygenase. Levine emphasized that this was a good example of a case where the evaluation of multiple measures of oxidative stress gave a less biased result.

Levine and Barlow are currently testing a manganese-containing Eukaryon compound, which has both catalase and SOD activity and can cross the blood-brain barrier, in atm-/- mice. One of these compounds, EUK-134, can rescue the longevity defect in Caenorhabditis elegans (C.elegans) mutant mev-1 (succinate dehydrogenase). These compounds have been tested in rodents and exhibit very low levels of toxicity.

Mark Mattson, the chair for this session, next discussed his results from studies of the secondary effects of homocysteine on neurodegeneration. Homocysteine is generated from methionine, and can be converted back to methionine by vitamin B12 or to cysteine by vitamin B6. Another important nutrient, folic acid, plays a major role in keeping homocysteine levels low. Increased levels of homocysteine increase the risk for cardiovascular disease and stroke, and lead to increased DNA damage. Interestingly, a low calorie diet has been shown to decrease the risk of neurodegenerative disease in Alzheimer's and Parkinson's mouse models.

Yossi Shiloh concluded this session with work from his laboratory and others that suggest a role for ATM in the oxidative stress response. Two atm-/- mouse models have been generated that either lack SOD expression or that overexpress SOD. Both mice have a similar phenotype, which suggests that a critical level of SOD is required for a normal response to oxidative challenges. These atm-/-/SOD mice display retarded growth and increased IR sensitivity compared to atm-/- mice, but no enhancement of the neurological phenotype. Thus, it appears that even these mice are unable to recapitulate the neurodegeneration seen in A-T patients.


Future Directions

Future directions for the next wave of research in the crossroads of neurodegenerative disease pathologies and chromosomal instability syndromes were addressed in a closing session. First, Giovanna Spinella from the National Institute of Neurological Disorders and Stroke (NINDS) emphasized that, if the genetic syndromes discussed during this meeting indeed share commonalities, this evidence could be reported to the NINDS and possibilities for special funding would be discussed. Also, in closing remarks, Dave Frohnmayer reminded the group that the Fanconi Anemia Research Fund has funds to support post-doctoral fellows, as well as a database of available cell lines and antibodies.

Gene Johnson and Mike Kastan led the final discussion, which began with a model proposed by Kastan. Given the following evidence presented at this meeting: 1) ATM is a checkpoint protein that responds to genotoxic stress (multiple groups), 2) post-mitotic neurons that re-enter the cell cycle undergo cell death, which is an intrinsic and presumably protective response to proliferative stress in this cell type (K. Herrup), and 3) atm-/- mice exhibit increased proliferation of neural progenitors in the brain (C. Barlow), Kastan proposed that the transition from the mitotic to the post-mitotic (G0) phase in a neuron is a checkpoint. Thus, the general hypothesis proposed by Kastan is that ATM enforces this checkpoint, and neurons that lack ATM re-enter the cell cycle and die. This model predicts the neurodegeneration that occurs in A-T patients is the secondary consequence of loss of this ATM-dependent checkpoint.

Discussion of data that are consistent or at odds with this model followed. The proposed ATM checkpoint would only affect later stage neurons. If this model is true, would other post-mitotic tissues be affected? It is possible that only certain tissues or cell-types are affected, but one other example in A-T might be alpha fetoprotein (AFP) production by hepatocytes. AFP levels are only elevated in the adult human during pregnancy, in certain malignancies, and in A-T. This is an instance where a fetal protein is expressed by an adult tissue, possibly one that has re-entered the cell cycle.

Rather than a stochastic re-entry into the cell cycle, perhaps a genotoxic insult leads to cell cycle entry and the events of checkpoint loss and DNA damage are inseparable. The signal to re-enter the cell cycle in the Alzheimer's model is microglial products, and although it is unclear what this signal might be for A-T cells, genotoxic stress is a prime candidate.

In one suggestion of how to test this hypothesis, a discussant proposed that the atm-/- mouse could be crossed with Harry Orr's transgenic mouse in which SV40 T antigen (TAg) is expressed under the control of a Purkinje cell-specific promoter. When TAg is expressed in the Purkinje cells of these mice, post-mitotic neurons re-enter the cell cycle and undergo apoptosis. These TAg transgenic mice display ataxia due to an apoptotic loss of Purkinje cells. If the model proposed above is correct and TAg expression coerces the cell to resume cycling, the TAg transgenic/atm-/- mice should have an enhanced loss of Purkinje cells and subsequent cerebellar function due to the loss of the ATM-dependent checkpoint. Alternatively, if TAg-driven cell cycle entry leads to ATM-dependent apoptosis, the Purkinje cell loss would be reduced in the TAg transgenic/atm-/- mice.

Another question that arose is whether this model involves a nuclear or cytoplasmic function for ATM or both. The currently available antibodies are not specific enough to pick up nuclear ATM in Purkinje cells, but this does not necessarily mean that Purkinje cells lack nuclear ATM. Clearly, the answer to this question awaits the development of high-quality specific antibodies for ATM.

The recurring theme that sounded throughout this meeting was echoed in the final discussion. ATM is a multi-functional protein, and the model system chosen to investigate ATM functions is important. Currently, a model system to study the neurodegenerative aspects of A-T does not exist, and research in this area would benefit from the generation of such a model system. Ultimately, advances in the mechanism behind A-T neurodegeneration may lead to treatments that alleviate the neuronal pathologies that currently plague A-T patients, and could very well impact on the therapy of other neurodegenerative diseases as well.




Alt, Fred
Auerbach, Arleen
Barlow, Carrolee
Breakefield, Xandra
Brooks, PJ
Brumbaugh, Cathy
Chen, Junjie
Chun, Jerold
Concannon, Pat
Cortes, Marisa
Crawford, Thomas
D'Andrea, Alan
Dixon, Kathleen
Ellis, Nathan
Fornace, Albert
Frohnmayer, Dave
Gallo-Hendriks, Ellie
Gatti, Richard
Geller, Herbert
German, James
Harris, Steven
Heintz, Nat
Herrup, Karl
Jeggo, Penny
Johnson, Eugene
Kastan, Michael
Lavin, Martin
Lawrence, David
Lehmann, Alan
Levine, Rodney
Mandir, Allen
Margus, Brad
Mattson, Mark
McKinnon, Peter
Michael Greenberg
Monnat, Raymond J.
Pandita, Tej
Park, David
Qin, Jun
Reinhart, Peter
Shiloh, Yossi
Spinella, Giovanna
Thompson, Larry
Weinert, Ted


Last updated April 12, 2011