Sensors and Inhibitors for the Detection of Ataxia Telangiectasia Mutated (ATM) Protein Kinase

Abstract

Recruitment and activation of the ataxia telangiectasia mutated (ATM) kinase regulate multiple cell-cycle checkpoints relevant to complex biological events like DNA damage repair and apoptosis. Molecularly specific readouts of ATM using protein assays, fluorescence, or radiolabeling have advanced significantly over the past few years. This Review covers the molecular imaging techniques that enable the visualization of ATM—from traditional quantitative protein assays to the potential use of ATM inhibitors to generate new imaging agents to interrogate ATM. We are confident that molecular imaging coupled with advanced technologies will play a pivotal role in visualizing and understanding the biology of ATM and accelerate its applications in the diagnosis and monitoring of disease, including radiation therapy and patient stratification.

Graphical Abstract:

INTRODUCTION

Ataxia telangiectasia mutated (ATM) kinase is the major protein in signaling the DNA damage response (DDR) in response to DNA double-strand breaks (DSBs). ATM is recruited and activated by the Mre11-Rad50-Nbs1 (MRN) complex at DNA double-strand breaks. Adaptations in the DDR pathway are hallmarks of cancer, and ATM’s regulatory role in DDR can be advantageous for abnormal cell survival.¹ As a result, a number of inhibitors have been developed to prevent ATM from performing its function. None of the reported ATM inhibitor drugs are used in the clinical settings. Two ATM inhibitors are currently undergoing clinical trial for cancer therapy in combination with other drugs.^2,3 One ATM inhibitor is undergoing clinical trial along with radiation therapy.⁴

The role of ATM signaling in the DDR has been widely studied in a variety of tumors, but insights into its processes are much less defined than those of its cousins, poly-ADP–ribose polymerase (PARP) and ataxia telangiectasia and Rad3 related (ATR) proteins. A focus of research in recent decades has included the distribution of ATM, cellular factors that participate in regulating ATM, and tools to investigate the biological effects of radiation therapy to ATM. The monitoring of ATM activity in real time used traditional protein assay methods as well as fluorescent proteins and in vitro bioluminescence techniques.

In addition, several ATM inhibitors have appeared as potential candidates for cancer therapy. Recently, in vivo imaging agents have emerged from the initial success of therapeutic inhibitors. Described in this Review are the features enabling optical and positron emission tomography (PET) imaging methods to be used for quantitative and diagnostic purposes and to allow direct and indirect imaging and tracking of ATM in cells and in in vivo models. Despite extensive studies over the past two decades, much of how ATM is activated is unknown and debated. There exist today vast opportunities for the development of ATM imaging agents and to address the need to better unravel the role of ATM in DNA damage repair.

TECHNIQUES TO QUANTIFY THE PHOSPHORYLATION OF ATM

ATM regulates cellular responses to DNA DSBs via autophosphorylation at Ser1981. The activation of ATM then signals a phosphorylation cascade to hundreds of other proteins.⁵ The isolation and identification of these phosphorylation events provides important insights into the cellular mechanisms of ATM activation. Much of these processes were observed and elucidated through the use of analytical techniques via immunoblotting, ChIP, immunofluorescence, kinase assay, or mass spectrometry (Figure 1).

Figure 1.

Schematic representation of the varying techniques to detect the kinase activity of ATM.

Immunoblotting.

This technique allows the separation of ATM from other proteins in cell extracts and identification of ATM via primary and secondary antibodies. Immunoblots have led to significant insights into the role of ATM in response to DNA damage. ATM is reported to be responsible for the phosphorylation events that are occurring at Ser15 and Ser46 residues on p53.^6,7

Chromatin Immunoprecipitation (ChIP) to Examine ATM Kinase Function.

The interactions between proteins can be determined by various approaches from cell extracts through coimmunoprecipitation and immunoprecipitation techniques of specific proteins. The technique typically uses protein beads from rabbit and mouse immunoglobulin (IgG) targeting ATM protein.⁸ The region of the DNA that is associated with the ATM is then detected by PCR amplification.

Immunofluorescence Used to Visualize ATM Kinase.

To detect the autophosphorylation of ATM kinase, immunofluorescence techniques use primary antibodies specific to the phosphorylated ATM and a fluorescently labeled secondary antibody. The secondary antibodies amplify the signal that corresponds to activated ATM (Figure 2).⁹

Figure 2.

Visualizing ATM with immunofluorescence. (A) Immunofluorescence of autophosphorylated ATM kinase in cells treated with radiation and topoisomerase inhibitor camptothecin (CPT). Images in the middle column are enlargements of cells, indicated by the white arrows on the left column. Fluorescence indicates site-specific autophosphorylation of ATM kinase. (B) Immunofluorescence is used to determine ATM kinase activity in cells treated with inhibitors of the DNA damage response. ATM kinase is less active in cells without damage (untreated cells) and cells treated with ATM inhibitor (ATMi), as indicated by the lack of phosphorylated ATM. Reprinted in part with permission from ref 9. Copyright 2017, Springer Nature.

ATM Kinase Assay.

In vitro assays have helped identify the pivotal role of activated ATM to phosphorylate other proteins that regulate the cell-cycle arrest, DNA repair, and apoptosis.^10–12 Several mechanistic pathways for the ATM activation have been elucidated through ATM kinase assays. For example, ATM has been shown to dissociate its form from dimer to monomer without autophosphorylation and recruitment of MRN but rather by direct sensing of DNA breaks.¹² The association of either BRCA1 and 53BP1 to ATM has indicated through these kinase assays that these proteins are regulated by ATM.^13,14

Mass Spectrometry.

ATM can be identified through mass spectrometry, in which it is first proteolytically digested into smaller peptides and then separated, ionized, and analyzed. Mass fingerprinting can be subsequently used to identify ATM by correlating the experimental mass-to-charge peaks with the known mass for ATM.^15,16 Mass spectrometry can also be used to quantify and identify known substrates phosphorylated at ATM-specific phosphorylation motifs (Ser–Gln and Thr–Gln) in order to analyze ATM activity.^17,18

DIRECT IMAGING OF ATM KINASE ACTIVITY

To date, there are no known optical reporters for in vivo dynamic and real-time imaging of ATM in mouse models. To investigate ATM kinase activity, in vitro studies typically use the cell imaging microscopy setup with bioluminescence, genetically encoded, fluorescence resonance energy transfer (FRET) reporters, and/or and high-throughput screening (HTS) techniques. The latest study involves the use of a small-molecule radiotracer for PET imaging.

Bioluminescence Imaging (BLI).

Upon ATM-dependent phosphorylation, changes in the light emitted from an ATM substrate are used to assess ATM kinase activity. This can be done by transfecting cells with the DNA of firefly luciferase encoding the luminescent protein. Clone selection is performed by administration of the antibiotic, Geneticin (G418), which specifically selects for positive transfected cells with pEF plasmid vectors containing the sequence of the bioluminescent ATM reporter (ATMR). ATMR protein tracks the activation and recruitment of ATM kinase to repair the DSBs in DNA with the help of a Chk2 substrate, a phosphopeptide-binding domain called forkhead-associated (FHA2) domain, and N- and C-terminal luciferase fragments (Figure 3A). The luciferase fragments are responsible for the blue to yellow/green light released from ATMR in the presence of inactive ATM kinase. When ATM kinase is activated, it phosphorylates the ATMR substrate, causing a conformational change in ATMR and decreasing its bioluminescent activity (Figure 3B). Advantages to this method include its nonintrusive nature, low detection of uninvolved signals, and efficiency.¹⁹ A nuclear localization signal (NLS) attached to ATMR results in ATMR–NLS with improved specificity for ATM localized in the nucleus (Figure 3C).^19,20 In vivo mouse studies demonstrated that the ATM-specific inhibitor KU-55933 was effective at blocking ATM kinase activity, as indicated by the increase in bioluminescence when compared to the vehicle-treated mice.²⁰

Figure 3.

Bioluminescence techniques to image ATM. (A) Schematic diagram of ATMR and ATMR–NLS construct. Additional nuclear localization signal (NLS) in ATMR–NLS improves specificity to ATM kinase. N-terminus and C-terminus split luciferase sequences interact to generate bioluminescence. (B) ATMR phosphorylation of Chk2 and subsequent interaction with the FHA2 domain alters the structure of ATMR and reversibly halts bioluminescence. (C) In single-cell imaging, ATMR–NLS demonstrates increased specificity to ATM due to bioluminescence localization in the nucleus. Reprinted in part with permission from refs 19 and 20. Copyright 2017, Springer Nature, and copyright 2013, Elsevier, respectively.

Genetically Encoded (GE) Fluorescent Imaging.

Tracking the recruitment and retention of ATM kinase at DSBs in live cells can be accomplished using a plasmid with ATM kinase and a fluorescent protein construct such as yellow fluorescent protein (YFP), green fluorescent protein (GFP), or other fluorescent proteins of the same family.^21,22 Fluorescent microscopy allows for the tracking of the YFP-tagged ATM activity throughout the repair process in live cells (Figure 4), and it can be used to determine the density of ATM kinase at various time points. The live imaging aspect of this technique gives it an advantage over immunofluorescence due to real-time tracing and monitoring of ATM kinase to gain insights into its cellular processes.

Figure 4.

Visualizing ATM with genetically encoded fluorescence imaging. A DSB is induced by the addition of ²³⁸U to cells and imaged in real time. White dots represent wild-type YFP–ATM recruitment to the nucleus for DNA repair. Scale bar: 5 μM. Reprinted in part with permission from ref 22. Copyright 2017, Springer Nature.

Fluorescence Resonance Energy Transfer (FRET).

Ting and co-workers developed CFP–YFP-based reporters to track ATM activity via conformational changes that induce energy exchange from CFP to YFP.²³ This is quantified by imaging cells in vitro and measuring the proportion of fluorescent intensities of yellow and cyan emission light, which are directly related to the phosphorylation of ATM kinase (Figure 5). FRET is irreversible but can be overcome by decreasing the affinity of the forkhead-associated (FHA) phosphospecific-binding domain for the p-Thr on the internal Chk2 construct.²⁴

Figure 5.

Imaging ATM with FRET. CFP–YFP-based reporter is composed of cyan fluorescent protein (CFP), yellow fluorescent protein (YFP), a Chk2 substrate with a T68 residue, and an FHA2 phospho-binding domain. Active ATM kinase phosphorylates Chk2, leading to interaction with FHA2 domain and changing thFRET efficiency, measured by the proportion of yellow to cyan fluorescent output. This process can be reversed by dephosphorylation of CFP–YFP reporter. Neocarzinostatin induces cells with DSBs, and fluorescence from CFP–YFP reporters is monitored. Colors closer to red indicate greater phosphorylation and a lower proportion between yellow to cyan light released. Reprinted in part with permission from ref 24. Copyright 2007, Elsevier.

High-Throughput Screening (HTS) of ATM Kinase.

HTS is a method that utilizes both imaging microscopy and large-scale data analyses of ATM. This enables the study of ATM function through the simultaneous processing of multiple biological factors such as protein phosphorylation, protein movement, and changes in protein abundance.²⁵ The phosphorylation and measurement of active intracellular pS1981-ATM foci is associated with DBSs and it is visible within minutes. Images are obtained using microscopy and can be used to calculate the activity of ATM kinases in each cell.^25,26

Positron Emission Tomography (PET).

The ATM inhibitor, AZD1390, comes from the same family of ATM-specific compounds as AZD0156. The structural changes between the two inhibitors allows for AZD1390 to cross the blood–brain barrier (BBB). The brain-penetrating properties of AZD1390 were characterized, and PET imaging was performed in monkeys. The inhibitor targets and blocks ATM-dependent signaling and repair of DNA DSBs in the genome, and nuclear labeling and PET imaging of the AZD1390 inhibitor enable direct tracing of the location of ATM kinase. To accomplish this, AZD1390 was labeled with ¹¹C at the methyl position of its 2-imidazolidinone and compared with non-brain-penetrable ATM inhibitor AZD0156 (Figure 6B). PET imaging with ¹¹C-labeled AZD1390 shows that AZD1390 has a greater ability to cross the BBB than AZD0156 (Figure 6C). ¹¹C-labeled AZD1390 takes advantage of the ability of PET imaging to visualize ATM in vivo on a whole-body level.²⁷

Figure 6.

PET imaging of BBB-permeable ATM inhibitors. (A) Structure of the blood–brain barrier in a human brain. Tight junctions between endothelial cells prevent the crossing of large molecules.²⁸ (B) Development of two BBB-permeable ATM inhibitors: AZD1390 and AZ32.^27,29–31 (C) (left) The asterisks mark the location of ¹¹C on AZD1390 (top) and AZD0156 (bottom). (right) PET images showing the 5 to 123 min postinjection average distribution of [¹¹C]-AZD1390 (left) and [¹¹C]-AZD0156 (right) in cynomolgus macaque brains. Image intensity is color-coded and shown as standardized uptake value (SUV). Except where otherwise noted, this work is licensed under a Creative Commons Attribution 4.0 International License. (Figure 6C, right image). Copyright, 2018, Durant et al., ref 27.

INDIRECT IMAGING OF ATM KINASE ACTIVITY

The activation of ATM by its phosphorylation enables key events at the site of DNA damage. ATM kinase targets include but are not limited to p53, Chk2, BRCA1, 53BP1, and other downstream targets, all of which have been studied using the above-mentioned techniques (kinase assay, ChIP, immunoblotting) to quantify ATM expression.^13,14 In addition, one of the early responses is the induction of γ-H2AX and deoxycytidine kinase (dCK), whose proteins are phosphorylated by ATM. Both indirect markers have been shown to correlate with ATM expression using imaging techniques.

Phosphorylation of H2AX.

Following IR-induced DSB, ATM phosphorylates H2AX into γ-H2AX at residue serine139.³² DDR proteins are subsequently recruited to γ-H2AX to resolve the DSB.³³ Visualization techniques such as dual multiplexed immunoblotting can be used to simultaneously quantify ATM and downstream ATM targets, such as H2AX, through phospho-specific antibodies.³⁴

PET Imaging of dCK.

A PET probe, [¹⁸F]-FAC, was designed to track deoxycytidine kinase (dCK). The activation of dCK enhances the rate of DNA repair and therefore could serve as a biomarker for ATM function (Figure 7A). [¹⁸F]-FDG was not able to stratify between irradiated and nonirradiated in four different types of tumor, whereas [¹⁸F]-FAC showed statistical significance in radiotracer uptake (Figure 7B). It was concluded that the activation of dCK enhances the rate of DNA repair and therefore could serve as a biomarker to measure and image the activity of ATM kinase in response to DSBs (Figure 7B).³⁵

Figure 7.

Indirect visualization of ATM through dCK PET imaging. (A) Immunoblot of 10K + dCK WT cells treated with DMSO vehicle, 10 μM ATM inhibitor, or 10 μM DNA-PKc inhibitor for 1 h and 3 Gy pf irradiation. (B) [¹⁸F]-FAC and [¹⁸F]-FDG are administered to mice for PET imaging after 3 Gy of irradiation of right tumor. Imaged mice have bilateral 10K ± dCK (WT, S74A, or S74E) tumors. Except where otherwise noted, this work is licensed under a Creative Commons Attribution 4.0 International License. Copyright, 2014, Bunimovich et al., ref 35.

FIRST- AND SECOND-GENERATION ATM INHIBITORS

ATM induces cell-cycle arrest and helps to repair DNA. Many cancers are highly dependent on DDR signaling for survival. In the last two decades, a number of ATM inhibitors have been developed to block the autophosphorylation of ATM and downstream protein targets involved in the cell cycle and DNA repair (Table 1). These strategies have been successful, and a number of these inhibitors have entered clinical trials.³⁶

Table 1.

First- and Second-Generation ATM Inhibitors

Compound	Structure	IC50	Notable Properties	References
AZD0156 [CAS no. 1821428-35-6]		0.58 nM	Constructed based on imidazo [4,5-c]quinoline-2-one core Current phase I clinical trial (ClinicalTrials.gov identifier: NCT02588105)	[29] [36]
KU-59403 [CAS no. 845932-30-1]		3 nM	Constructed based on KU-55933 Enhances G2 phase cell cycle arrest	[37]
KU-60019 [CAS no. 925701-49-1]		6.3 nM	Developed based on KU-55933 Current clinical trial (ClinicalTrials.gov identifier: NCT03571438)	[36] [38]
KU-55933 [CAS no. 587871-26-9]		13 nM	ATP-competitive inhibitor Developed based on PI3K and DNA-PK inhibitor, LY294002	[39] [40]
AZ31 [CAS no. 2088113-98-6]		46 nM	Constructed based on 3-quinoline carboxamide core	[30] [31] [41]
CP466722 [CAS no. 1080622-86-1]		6 μM	Reversible and nontoxic ATM inhibition Has a quinazoline core	[42] [43]
GSK635416A [CAS no. N/A]		25 μM	Decreased ATM autophosphorylation Decreased Chk2 phosphorylation	[44]

AZD0156.

AZD0156 is an ATP-competitive ATM inhibitor developed from an imidazo[4,5-c]quinoline-2-one core, with a half-maximal inhibitory concentration (IC₅₀) of 0.58 nM. AZD0156 binds to the ATP site of ATM by interacting with the its catalytic lysine (Lys2717), kinase hinge (Cys2770), and back pocket (Tyr2755). Additionally, AZD0156’s amine moiety situates in an acidic subpocket, and it is stabilized by the three acidic residues (Asp2720, Asp2725, and Asp2889) in the subpocket.²⁹ The preclinical success of AZD0156 is currently tested in a phase I clinical trial (ClinicalTrials.gov identifier: NCT02588105).³⁶

KU-59403.

KU-59403 is developed from a previous scaffold of an earlier ATM inhibitor called KU-55933 (based on a scaffold of PI3K and DNA-PK inhibitor LYS294002) and exhibits an improved IC₅₀ of 3 nM. KU-59403 has an additional cyclic amine and amide moiety when compared to LYS294002.³⁷ Administration of KU-59403 inhibits ATM activity and enhances G2-phase cell-cycle arrest in cancer cells.³⁷

KU-60019.

KU-60019 is a second-generation inhibitor developed from KU-55933. It has an improved IC₅₀ of 6.3 nM. KU-60019 has an additional morpholine and amide moiety, improving its solubility in aqueous environments when compared to KU-55933. Following ionizing irradiation, KU-60019 demonstrates higher efficacy than KU-55933 in blocking activation of the ATM targets Chk2, p53, and H2AX. Akt phosphorylation at the Ser473 position is also inhibited by the administration of KU-60019.³⁸ KU-60019 is currently undergoing a clinical trial with the casein kinase 2 inhibitor CX4945 (NCT03571438).³⁶

KU-55933.

KU-55933 is an ATP-competitive inhibitor developed from the nonspecific PI3K and DNA-PK inhibitor LY294002.³⁹ The oxygen in the morpholine moiety of KU-55933 highlights the importance of hydrogen-bonding at the ATP-binding site of ATM for inhibition. KU-55933 has an IC₅₀ of 13 nM, 200-fold lower than that of KU-58050, which structurally differs only in an oxygen in the cyclic amine side chain.³⁹ KU-55933 downregulates cyclin D1 and subsequently induces G1 phase arrest in cancer cells.⁴⁰

AZ31.

AZ31 is developed from a 3-quinoline carboxamide core, with an IC₅₀ = 46 nM.³⁰ Administration of AZ31 inhibits the autophosphorylation of ATM at Ser1981 and phosphorylation of ATM targets, p53 (Ser15) and KAP1 (Ser824).³¹ The combination of AZ31 with irinotecan leads to decreased phosphorylation of H2AX and RAD50.⁴¹

CP466722.

CP466722 is a reversible and nontoxic ATM inhibitor with a quinazoline core and an IC₅₀ of 6 μM. Transient and reversible ATM inhibition is indicated by the complete loss of ATM inhibition within 1 h of CP466722 administration followed by washoff with cell media. Following the administration of CP466722, phosphorylation of Chk2 (Thr68) and H2AX (Ser139) is blocked. Additionally, CP466722 shows inhibitory effects on Src kinase and inhibits G2-M checkpoint progression.⁴² Further studies show that CP466722 reduces the expression of PD-L1 in cisplatin-resistant non-small-cell lung cancer cells and decreases JAK/STAT3 signaling.⁴³

GSK635416A.

GSK635416A is a direct ATM inhibitor discovered from a kinase inhibitor screening meant to identify novel compounds with radiosensitizing properties superior to known radiosensitizers. GSK635416A has an IC₅₀ of 25 μM. Administration of GSK635416A leads to a decrease in ATM autophosphorylation and Chk2 phosphorylation.⁴⁴

BLOOD–BRAIN-BARRIER-PERMEABLE ATM INHIBITORS

Gliomas exhibit elevated levels of DNA damage response (DDR) where ATM is a key regulator. Therefore, ATM can act as a potential diagnostic and therapeutic target.⁴⁵ However, one significant challenge is the BBB, which blocks the delivery of chemotherapeutics to the brain (Figure 7A).^46,28 Recently, two ATM inhibitors, AZD1390 and AZ32, were successfully designed, synthesized, and shown to cross the BBB and exhibit good permeability. Structurally, AZD1390 was synthesized based on an imidazo[4,5c]quinoline-2-one core, and AZ32 was synthesized based on a 3-quinoline carboxamide core (Table 2).

Table 2.

Blood–Brain-Barrier-Permeable ATM Inhibitors

Compound	Structure	IC50	Notable Properties	References
AZD1390 [CAS no. 2089288-03-7]		0.78 nM	Not a substrate of MDR1 or BCRP Current phase I clinical trials (ClinicalTrials.gov identifier: NCT03423628 and NCT03215381)	[27] [36]
AZ32 [CAS no. N/A]		6.2 nM	Developed based on AZ31 Decreased solubility in aqueous environments	[31]

AZD1390.

AZD1390 was developed as a BBB-permeable analogue of AZD0156 and has an IC₅₀ = 0.78 nM. Compared to AZD0156, AZD1390 has a piperidine moiety in place of a noncyclic tertiary amine side chain, an isopropyl moiety in place of a tetrahydropyran, and an additional fluorine (Figure 7B). AZD1390 inhibits ATM autophosphorylation (Ser1981) and phosphorylation of KAP1 (Ser824). AZD1390 is not a substrate for BBB endothelial cell effiux transporters, MDR1 (multidrug resistance 1), and BCRP (breast cancer resistance protein), indicating that AZD1390 has good BBB permeability, as confirmed by PET imaging with [¹¹C]AZD1390 in cynomolgus macaque brains (Figure 7C).²⁷ The clinical phase I trials of AZD1390 in glioblastoma were made possible through these promising results (NCT03423628 and NCT03215381).³⁶

AZ32.

AZ32 was developed based on AZ31 and has an IC₅₀ = 6.2 nM. In comparison to AZ31, AZ32 has fewer positions for possible hydrogen-bonding, thus decreasing the solubility of this compound in an aqueous environment (Figure 7B). Furthermore, AZ32 has a lower molecular weight than AZ31.³¹ Phosphorylation of ATM downstream targets, KAP1 and p53, is blocked after the administration of AZ32.³¹

MULTITARGET ATM INHIBITORS

A multitarget approach that targets not only ATM but also other members of the phosphatidylinositol 3-kinase-related kinase (PIKK) family, such as ATR, represents potentially powerful methods of inhibiting DNA repair in cancers and improving clinical outcome. A select few inhibitors show broad inhibitory effects on PIKK family proteins, which include ATM, ATR, DNA-dependent protein kinase (DNA-PK), and the mammalian target of rapamycin (mTOR) (Table 3). NVP-BEZ235, undergoing several clinical trials, shows the most promise as a multitarget ATM inhibitor.

Table 3.

Multitarget ATM Inhibitors

Compound	Structure	IC50	Notable Properties	References
Torin2 [CAS no. 1223001-51-1]		10 nM	ATP-competitive inhibitor Also inhibits mTOR, ATR, DNA-PK, & PI3K	[47] [48] [49] [50]
ETP-46464 [CAS no. 1345675-02-6]		25 nM	Potent ATR and mTOR inhibitor and has higher specificity for ATR Decreased ATM autophosphorylation in combination with cisplatin	[51] [52]
NVP-BEZ235 [CAS no. 915019-65-7]		100 nM	Inhibits PI3K, mTOR, DNA-PKc Completed 8 phase I clinical trials (NCT01343498, NCT01482156, NCT00620594, NCT01634061, NCT01195376, NCT01337765, NCT01471847, NCT01248494), 1 phase Ib/ll clinical trial (NCT01495247), and 1 phase II clinical trial (NCT01658436)	[51] [53] [54] [55] [56] [57]
Wortmannin [CAS no. 19545-26-7]		150 nM	Fungal steroid metabolite ATR and DNA-PK inhibitor	[58] [59] [60]
CGK733 [CAS no. 905973-89-9]		10-20 μM	ATM and ATR inhibitor Leads to ublqultlnatlon and degradation of cyclin D1	[61] [62]

Torin2.

Torin2 is a multitarget ATP-competitive inhibitor that inhibits ATM in addition to mTOR, ATR, DNA-PK, and PI3K (phosphoinositide 3-kinase). It has an IC₅₀ = 10 nM. It was developed from Torin1 by replacement of its propionyl-piperazine moiety with fluorine and replacement of the quinoline moiety with pyridine. Torin2 forms a hydrogen bond with Val2240 through its quinoline nitrogen and a second hydrogen bond with Tyr2225 with its pyridine nitrogen in the inner hydrophobic pocket of mTOR.⁴⁷ Administration of Torin2 leads to diminished levels of phosphorylated ATM after irradiation.^48,49 Furthermore, administration of Torin2 inhibits the phosphorylation of the downstream ATM target Chk2.⁵⁰

ETP-46464.

ETP-46464 is an ATM/ATR/mTOR inhibitor with IC₅₀ = 25 nM.⁵¹ This inhibitor is developed based on a quinoline core. ETP-46464, when administered with cisplatin, results in a decrease in ATM autophosphorylation and Chk2 phosphorylation.⁵² Additionally, ETP-46464 demonstrates higher selectivity for ATR over ATM and DNA-PK.⁵¹

NVP-BEZ235.

NVP-BEZ235 is an mTOR/PI3K inhibitor that also inhibits ATM, ATR, and DNA-PKcs and has an IC₅₀ = 100 nM for ATM. It is developed based on an imidazo[4,5-c]quinoline core, which gives this inhibitor the ability to mimic ATP hydrogen-bonding at its target site.^51,53 Following NVP-BEZ235 administration, ATM autophosphorylation at Ser1981 is decreased.⁵⁴ Additionally, NVP-BEZ235 decreases the phosphorylation of ATM downstream targets, Chk2 (Thr68), SMC1 (Ser966), p53 (Ser15), KAP1 (Ser824), and H2AX (Ser139).^55,56 NVP-BEZ235 has undergone eight phase I clinical trials (NCT01343498, NCT01482156, NCT00620594, NCT01634061, NCT01195376, NCT01337765, NCT01471847, NCT01248494), one phase Ib/II clinical trial (NCT01495247), and one phase II clinical trial (NCT01658436).⁵⁷

Wortmannin.

Wortmannin is a multitarget fungal steroid metabolite with an IC₅₀ = 150 nM for ATM. It irreversibly inhibits the catalytic activity of ATM, DNA-PK, and ATR through covalent modification of a homologous lysine residue in their catalytic domains.⁵⁸ Administration of wortmannin decreases ATM autophosphorylation and leads to increased S-phase arrest following DNA damage.^59,60

CGK733.

CGK733 is an ATM and ATR inhibitor containing a thiourea, with an IC₅₀ = 10–20 μM. CGK733 reduces p21^CIP1 phosphorylation in senescent cancer cells, leading to apoptosis.⁶¹ CGK733 also leads to G1 phase arrest by inducing the ubiquitination of cyclin D1.⁶²

DISCUSSION

ATM has been historically difficult to study and express in cells due to its large size of approximately 350 kDa, consisting of 3056 amino acids. Studies have focused on observing and analyzing key players and posttranslational modifications that are involved in recruiting ATM to DSBs. Traditionally, ATM can be purified through chromatin immunoprecipitation and kinase assay using ATM-specific targets. Immunofluorescence techniques allow for direct visualization of ATM formation in the cell nucleus. However, one major limitation of this technique is that cells must be fixed in order to analyze the DSB-induced ATM foci, rendering real-time dynamic imaging of ATM at the DNA damage sites impossible. Quantification and identification of ATM can be achieved through immunoblotting and mass spectrometry, respectively. Although these techniques can be used to purify ATM and quantify the expression of ATM in cells and are easily established and reproduced in the laboratory, they are inadequate for visualizing ATM expression and subsequent cellular pathways in whole cells in vitro.

ATM expression can also be assessed indirectly by quantifying the expression of ATM-specific targets. Many of the recent studies investigate ATM expression indirectly by measuring the levels of other related substrates, such as γ-H2AX, using immunoblotting techniques. PET imaging of dCK can also be used to indirectly image ATM expression in vivo. These methods are suitable when ATM does not need to be targeted directly.

On the other hand, imaging techniques of ATM such as bioluminescence and fluorescently tagged ATM directly enable the study of ATM’s DNA damage responses in real time in living cells. BLI and FRET techniques make use of synthetic molecular targets that are modifiable by p-ATM, which allows for not only the imaging of activated ATM through fluorescent markers but also the localization of ATM expression at the nuclear foci or near DNA damage sites. ATMR generates fluorescence from the interaction between luciferase constructs, whereas fluorescence from FRET constructs is generated by the overlap of the absorption and emission spectra of two fluorophores. Typically, these reporters represent invaluable tools, because they provide unique biological insights into radiation-induced signaling events via direct readout of ATM kinase activity in live cells. Their use is limited in mouse models due to tissue penetration and autofluorescence in the CFP–YFP range. The use of high-throughput screening is advantageous due to the large amount of ATM-specific information that can be obtained in one analysis. Genetically encoded sensors enable fluorescent ATM to be produced in cells and allow for real-time and dynamic live cell imaging, making possible the study of its cellular processes. Genetically encoded ATM sensor imaging uses in vitro conditions and cannot be translated to the clinic.

However, many ATM-specific inhibitors have been developed, which vary widely based on their structure and therefore pharmacokinetic and pharmacodynamic properties. The innate properties of ATM inhibitors make them ideal for transformation into ATM-specific imaging agents, such as what has been shown with [¹¹C]-labeled AZD1390. Changing the methyl substituent to ¹¹C has the profound advantage of being identical in its molecular structure to the inhibitor AZD1390, which has already been shown to be a potent ATM inhibitor. The radiotracer approach to labeling AZD1390 with ¹¹C was the first example of combining a nuclear imaging approach with PET imaging to visualize ATM in vivo with high specificity. The short half-life and the associated logistics that come with the production of ¹¹C could present some challenges when translating to the clinics, as compared to other radionuclides. In the near future, other radioisotopes could be conjugated to ATM inhibitors to utilize their ATM-binding abilities for targeted delivery to cancer and to study their pharmacokinetics and pharmacodynamics. When considering site-specific modifications to enable molecular imaging, it is important to maintain structural integrity with respect to target binding.

Multitarget inhibitors are appealing, because they can target several pathways involving DDR. However, multitarget ATM inhibitors are less suitable for the quantification and imaging of ATM expression alone, because they also exhibit specificity to other closely related targets and tend to have lower binding affinities. We expect that the success of small synthetic imaging agents for the visualization of ATM is broadly dependent on the success of the available ATM inhibitors. Those with favorable pharmacological properties and which are undergoing clinical trials show promise.

CONCLUSIONS

ATM is a valuable target for molecular imaging, especially in cancer. ATM inhibitors that demonstrate low IC₅₀ and good BBB permeability, such as AZD1390, are promising not only as therapeutics for cancer but also as molecules for the development of ATM-specific imaging agents. Additionally, ATM inhibitors mark the possibility of the development of a new class of ATM-specific imaging agents with the potential for combined diagnostic and therapeutic applications, toward the ultimate goal of improved patient care.

ABBREVIATIONS

ATM

ataxia telangiectasia mutated

DDR

DNA damage response

DSBs

DNA double-strand breaks

MRN

Mre11-Rad50-Nbs

PARP

poly-ADP–ribose polymerase

ATR

ataxia telangiectasia and Rad3 related

PET

positron emission tomography

ChIP

chromatin immunoprecipitation

FRET

fluorescence resonance energy transfer

HTS

high-throughput screening

BLI

bioluminescence imaging

ATMR

ATM reporter; NLS, nuclear localization signal

YFP

yellow fluorescent protein

GFP

green fluorescent protein

FRET

fluorescence resonance energy transfer

dCK

deoxycytidine kinase

BBB

blood brain barrier