Differential localization of ATM is correlated with activation of distinct downstream signaling pathways

Abstract

ATM, the gene mutated in the genetic disease ataxia telangiectasia (AT), is a well-known protein involved in the DNA double-strand break response, where it plays an important role in sensing damage and signaling to DNA repair machinery and cell cycle checkpoints. However, a number of recent papers, including ours have found that ATM also plays important roles outside of the nucleus, which may explain some of the phenotypic features seen in AT patients. Our research into mechanisms of TSC2 regulation helped uncover a pathway upstream of TSC2 that is regulated by cytoplasmic ATM in response to ROS initiated by ATM activation of LKB1 and AMPK. We found that TSC2 activation results in mTORC1 repression and subsequent induction of autophagy. Elucidation of this stress response pathway provides a molecular mechanism for ATM signaling in the cytoplasm, and lays the groundwork for further studies on how ATM activity is regulated beyond DNA damage in different cellular compartments.

More than 15 years of intense research into the biological functions of the ATM gene have considerably broadened our understanding of the molecular basis behind the pleiotropic disease phenotype observed in patients with ataxia telangiectasia (AT). AT is a systemic disease characterized by neurodegeneration, hypersensitivity to ionizing radiation, immunodeficiency, metabolic defects and an elevated risk of hematopoietic malignancies.1 The ATM kinase has largely been thought to function as a DNA damage sensor. In response to DNA damage, ATM rapidly orchestrates a stress response consisting of signaling to the DNA repair machinery and activation of cell cycle checkpoints to allow repair of damage or induction of apoptosis if the damage is too severe to be repaired.2 However, not all of the pathophysiology of AT can be explained by ATM's role in the double-strand break (DSB) response. For example, cells from AT patients have high levels of oxidative stress, leading to the hypothesis that ATM may play a role in maintaining redox homeostasis. ATM deficiency has also been linked with metabolic disease, in particular insulin resistance.3

Recent studies have begun to elucidate a number of additional functions of ATM both inside the nucleus, such as repressing transcription around sites of damage,4 as well as outside the nucleus. ATM has been reported to localize to various organelles including peroxisomes,5 and centrosomes,6 and has been associated with vesicles involved in endocytosis and protein transport,7 suggesting that ATM may regulate a larger number of biological processes than previously thought.

Our recent paper describes a new pathway for ATM activation in the cytoplasm in response to reactive oxygen species (ROS).8 We demonstrated that ATM activation in the cytoplasm in response to ROS (as determined by phosphorylation of ATM in cytoplasmic fractions and activation of downstream substrates including p53 and Chk2), rapidly engages the LKB1-AMPK pathway to activate TSC2 to suppress mTORC1 signaling. This pathway does not require p53, and is therefore distinct from the cell cycle checkpoint response in the nucleus which is largely p53-dependent. DNA damaging agents have also been reported to induce mTORC1 repression, although the precise mechanism has not been identified. To determine whether DNA damage-induced ATM activation resulted in mTORC1 repression via this rapid cytoplasmic pathway, we analyzed whether AMPK was activated in response to these agents. In both MCF7 cells and MEFs, AMPK activity was unchanged or even decreased, suggesting that this rapid cytoplasmic pathway is distinct from that which regulates mTORC1 in response to double strand breaks. Table 1 illustrates some of the similarities and differences between ATM functions in the cytoplasm and the nucleus. In addition to elucidating an important mechanism of coordinating cell growth pathways with redox homeostasis, our work adds to the growing body of literature regarding DNA-damage independent (“pseudo DNA damage response”) mechanisms of ATM activation. A recent study demonstrating delayed activation of ATM in response to HDAC inhibitor-induced senescence, which was partially dependent on mTOR, further solidifies the molecular linkage between ATM and mTOR.9

mTORC1 is a known repressor of autophagy.10 Autophagy is a catabolic process of recycling cellular components, and is a general response of cells to many types of stress, leading to either enhanced cell survival, or if taken too far, can cause type II programmed cell death.11 We confirmed that cells exposed to pro-oxidizing agents including H₂O₂ were induced to undergo autophagy via the ATM/AMPK/TSC2/mTORC1 pathway, suggesting that induction of autophagy in response to ROS reported by others could also occur via this pathway.12

Interestingly, the earliest reports of ATM in the cytoplasm were mainly in neuronal cells (from both humans and rodents).13,14 Although these cells are non-dividing, basal metabolic flux in these cells is among the highest in the body. It is possible that in these cells the abundant cytoplasmic ATM regulates autophagy as a survival measure in response to stress or to maintain redox homeostasis through turnover of damaged ROS-generating organelles such as mitochondria and peroxisomes. This may contrast with other proliferative somatic cell types, which must retain the ability to regulate cell cycle progression to maintain genomic stability, and which therefore may favor nuclear ATM signaling and if necessary, induction of apoptosis. These divergent pathways as a result of ATM subcellular localization, different mechanisms of activation and cell survival outcomes may explain some of the pleiotrophic phenotypes seen in AT patients and opens interesting new avenues of research.

Figures and Tables

Table 1 Summary of similarities and differences between cytoplasmic and nuclear ATM signaling and cellular outcomes

	Cytoplasmic ATM	Nuclear ATM
Stimulus for activation	Reactive oxygen species	DNA double strand breaks, chromatin structure change, delayed response to replication fork stalls (via ATR)
Autophosphorylation at S1981	Yes	Yes
Activated protein as monomer/dimer	Unknown	Monomer
Signaling pathways	LKB1-AMPK-TSC2-mTORC1	Many pathways, including p53-p21, chk2 checkpoint responses
p53 dependence	Independent	Primarily p53 dependent
Cell survival/death pathway regulated	Autophagy (may be pro-survival or death mechanism)	DNA repair (cell survival) or apoptosis (cell death mechanism)