The synthesis of the non-radioactive 19F version of compound 1 followed methods reported by patent [8], with minor modifications as detailed in the supporting information (Fig. 2; Additional file 1: Fig. S1). The synthesis of nitro-precursor 2 followed the same route as for [19F]1. Full synthetic details are available in the supporting information.

a Synthetic pathway to reference [19F]1 and nitro-precursor 2. b Radiosynthesis of [18F]1 from 2

Radiosynthesis of [18F]1 was achieved through nucleophilic aromatic substitution (SNAr) of 2 (1 mg) with [18F]fluoride (5–12 GBq) dried azeotropically with acetonitrile in the presence of K2CO3 (6 mg) and K2,2,2 (30 mg) reacting for 10 min at 120 °C in DMF (Fig. 2). HPLC purification of [18F]1 was performed using a Luna® 10 µm C18(2) 100 Å 250 × 10 mm semi-preparative column (Phenomenex) in 25:75 acetonitrile/0.1% formic acid aqueous buffer, followed by reformulation in 10% DMSO in 0.9% saline solution for in vitro and in vivo studies. Manual radiolabelling reactions were performed using aliquots of [18F]fluoride (approx. 200 MBq) azeotropically dried with acetonitrile with addition of K2CO3 (1 mg) and K2,2,2 (5 mg) before the subsequent labelling reactions. The radiosynthesis was fully automated on an Eckert & Ziegler Modular-Lab system (full description of setup available in the supporting information, Additional file 1: Figs. S2–S5).

H1299 non-small cell lung cancer cells, either wild type or stably transfected with short hairpin RNA for ATM (shATM) or GFP (shGFP, as a control) expression knockdown, were kindly gifted by Dr Anderson Ryan at the University of Oxford [9]. All cells were cultured at 37 °C and 5% CO2 in Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% foetal bovine serum (FBS) (Gibco, USA), 2 mM L-glutamine, 100 units/mL penicillin, and 100 µg/mL streptomycin (PSG) (Sigma-Aldrich); knockdown cell lines were additionally supplemented with 0.5 μg/mL puromycin dihydrochloride (Merck Life Science, UK). Cells were originally purchased from ATCC and were authenticated by the provider and by STR profiling. Cells were tested regularly for the absence of mycoplasma contamination and were cultured for a maximum of 20 passages after resuscitation from liquid nitrogen storage.

Western blot analysis was performed to compare the total ATM expression between cell models, in addition to expression of phospho-ATMS1981 (pATM), phospho-KAP1S824 (pKAP1), and phospho-CHK2T68 (pCHK2) expression to assess the effect of irradiation and ATM inhibition on ATM catalytic activity. Cells were irradiated (4 Gy; 0.75 Gy/min) or sham-irradiated and incubated at 37 °C for 1 h before harvesting cells for lysis and blotting (full details in supporting information). Where required, inhibitors were added to cells at the stated concentration from DMSO stock solutions (reaching no greater than 10% DMSO final concentration) with incubation for 1 h prior to irradiation/sham-irradiation. The following antibodies and dilutions were used in this work: phospho-ATM S1981 (ab81292, 1:5000, Abcam); phospho-KAP1 S824 (ab133440, 1:10,000, Abcam); phospho-CHK2 T68 (#2661, 1:1000, Cell Signalling Technologies); Total ATM (ab32420, clone [Y170], 1:500, Abcam); Total KAP1 (ab10483, 1:1000, Abcam); Total CHK2 (05-649, 1:1000, Merck Millipore); HRP-conjugated Tubulin (ab197740, clone [YL1/2], 1:4000, Abcam); HRP-conjugated β-Actin (clone 13E5, 1:1000, Cell Signalling Technologies), HRP-conjugated anti-rabbit secondary (HAF008, 1:1000, R&D Systems); and HRP-conjugated anti-mouse secondary (HAF007, 1:1000, R&D Systems). All blots were repeated on at least two separate occasions with new cell lysates for validation.

To assess cellular uptake of [18F]1, aliquots of H1299 shATM or shGFP cells were plated with 1 × 105 cells per well in 24-well plates in 500 μL complete growth medium and incubated overnight to allow cells to adhere. Cell culture medium was removed, cells washed with PBS and media replaced with DMEM (no additives). Cells were irradiated (4 Gy; 0.75 Gy/min) or sham-irradiated and incubated for 1 h before addition of [18F]1, followed by incubation at 37 °C and 5% CO2 for the indicated time. Supernatant cell culture medium was collected, and cells were washed twice with PBS, after which they were lysed with 0.1 M sodium hydroxide (200 μL) for 30 min. The amount of 18F-activity associated with each fraction was determined using a Wizard2 Automatic Gamma Counter (PerkinElmer).

Retention assays followed the same procedure with a 1-h incubation of cell with [18F]1, but after culture medium was aspirated and cells washed with PBS (2 × 500 μL), it was replaced (500 μL DMEM) and cells incubated at 37 °C and 5% CO2 for the desired time. The amount of 18F-activity retained in cells was then determined as above.

Blocking studies followed the same procedure as uptake with 1 h tracer incubation; however, the replacement media contained the blocking agent at the indicated concentrations in no greater than 10% DMSO, and a DMSO control was also run in parallel at the highest DMSO concentration used, followed by incubation at 37 °C and 5% CO2 for 1 h before irradiation.

All animal procedures were performed in accordance with the UK Animals (Scientific Procedures) Act 1986 and with local ethical committee approval. Mice were bred from parent C57BL/6 J mice heterozygous for a targeted mutation in the ATMtm1Awbl allele to produce ATM-null (ATM−/−) mice (Jackson laboratory). Wild-type (ATM+/+) mice from this breeding program were used as controls. In a separate study, tumour xenografts were created in BALB/c nu/nu mice (8 weeks old) via subcutaneous injection of H1299 wild type, shATM knockdown, or shGFP knockdown cells in a 1:1 PBS/Matrigel® formulation on both sides of the hind flank. Total tumour volumes were allowed to reach an average of 600 mm3 (range 330–930 mm3) before imaging and/or biodistribution studies (approx. 6–10 weeks after inoculation).

Animals were anaesthetised by 4% isoflurane gas (0.5 L/min O2) and maintained at approx. 2% and 37 °C throughout the imaging session. C57BL/6 ATM+/+ and ATM−/− mice were administered [18F]1 (83 ng; 0.5–4.6 MBq, molar activity range of 3.2–24.2 GBq/μmol with a standard deviation of 6.5 GBq/μmol) via intravenous bolus injection in the lateral tail vein. Xenograft-bearing BALB/c nu/nu mice were administered [18F]1 (0.5 μg; 0.7–9.6 MBq, molar activity range of 0.62–9.8 GBq/μmol with a standard deviation of 3.0 GBq/μmol) via intravenous bolus injection to the lateral tail vein. PET/CT scanning was completed using a VECTor4CT scanner (MILabs, Utrecht, the Netherlands) for 1 h and was immediately followed by culling, dissection, and gamma counting of selected organs. Tumour tissue was immediately flash frozen in liquid nitrogen and mounted in OCT before producing 10 μm thick sections for autoradiography and H&E staining. Alternatively, frozen tumour tissue was macerated and lysed in RIPA buffer for western blot analysis (full details available in the supporting information). Reconstruction of both CT and PET images was performed with the MILabs reconstruction analysis using a γ-ray energy window of 467–571 keV (background weight 2.5%), 0.8 mm3 voxel size, 8 subsets, and 5 iterations using the manufacturer’s SROSEM reconstruction type. PET images were each registered to CT, attenuation corrected, calibrated by imaging a phantom containing a fluorine-18 standard solution, and analysed using PMod software package (version 3.807, PMOD Technologies).

Statistical analysis was performed using GraphPad Prism v9 (GraphPad Software, San Diego, CA, USA). In each case, a 1-way or 2-way ANOVA was performed with multiple comparisons using Šidák’s or Tukey’s test, with significance denoted as *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, or ns p > 0.05.