Spatial proximity analysis of adenovirus foci and hypoxia
Animal experimentation was conducted under ethical approval from the UK Home Office (Project license 30/3391 and personal license I5916669E). Female CB17-SCID mice (6–8 weeks old) were inoculated subcutaneously in the right flank with 2 × 106 DLD-1 cells in 50 µL. On day 1, when xenografts reached 80–120 mm3, mice received 100 µL of chlodronate liposomes (Chlodronateliposomes.org) intravenously to maximize subsequent adenovirus delivery.54 EnAd-SA-fLuc reporter virus particles (2 × 1010) were injected via the tail vein on days 2 and 4. During steady-state infection, previously shown to be 16 days after the last treatment,12 mice were injected intraperitoneally with pimonidazole at 60 mg·kg−1 and were culled 2 h later. Tumors were explanted, formalin-fixed, and paraffin-embedded.
Eight pairs of serial tissue sections per tumor (4 µm thick, ca. 40 µm apart, 4 pairs for mouse 3) were stained for hexon (horseradish peroxidase/3,3´-diaminobenzidine), pimonidazole adducts (alkaline phosphatase/Fast red), and nuclei (hematoxylin) as previously described.12 Spatial proximity between hexon-positive adenovirus foci and hypoxic regions was analyzed using a custom pipeline in Fiji/ImageJ.55 Briefly, matched serial sections were scanned at 20x (Aperio ImageScope, Leica), aligned, cropped to tissue boundaries, and cleaned of staining artifacts. Binary masks were generated from color-deconvolved signals using intensity thresholding. The pimonidazole mask was skeletonized to render an Euclidean distance map extending 200 µm outward from the core/skeleton of pimonidazole staining. Hexon- and hematoxylin-positive areas, representing nuclei in viable tumor regions, were projected onto this map to extract the normalized pixel fraction as a function of distance from the pimonidazole skeleton. A detailed framework is provided in Supplementary Fig. 1.
Cell lines
DLD-1, SW480, SK-OV-3, HCT116, HT29 (ATCC), and AD293 (Agilent Technologies) were cultured in DMEM (Sigma Aldrich) with 10% (v/v) heat-inactivated fetal bovine serum (FBS, Gibco) at 37 °C and 5% CO2. Cell lines were maintained in a humidified incubator, routinely tested for mycoplasma (Lonza), and authenticated (Source Bioscience).
Hypoxia treatment
Hypoxia incubation (1% pO2) was performed in the In vivo2 400 chamber (Baker Ruskin) using media pre-equilibrated overnight at 1% pO₂. Experiments utilized gas-permeable Lumox® plates (Sarstedt) on a zig-zag rocker (PMR-30, VWR) at five cycles per minute to maintain gas equilibrium. External oxygen calibration of the chamber was performed routinely. Pharmacological hypoxia was induced with 50 µM FG4592 (Roxadustat, Cayman Chemical).
Viruses and infections
EnAd-E1A-cFLAG, used for E1A expression analysis, was generated using a shuttle plasmid containing the E1A open reading frame flanked by NdeI sites in the pColoAd2.4 vector.23 Cloning of this virus is summarized in Supplementary Figs. 3 and 4. EnAd-SA-fLuc, used in experiments tracking viral life cycle, contains a firefly luciferase coupled to the major late transcription unit by a splice acceptor sequence.23 EnAd-CMV-BiTE and EnAd-SA-BiTE express a deca-His-tagged EpCAM/CD3-BiTE under a CMV promoter or integrated into the late transcription unit via a splice acceptor, respectively.25 Ad5-E1A-fLuc is a replication-competent reporter adenovirus encoding firefly luciferase fused to E1A.56 Virus stocks were concentrated and double-purified via cesium chloride gradient centrifugation, then titered by plaque assay on A549 cells as previously described.12
Infections were performed in serum-free media for 2 h. For virus production, inocula were removed, and cells were washed twice with complete medium. The inoculum was set to 125 μL·cm−2 independent of the seeding format, and MOIs were based on titers derived by plaque assay (Supplementary Table 2).
Quantification of infectious adenovirus particles using bioluminescence
A high-dynamic-range method was established to measure infectious viral particles by correlating light emitted by reporter adenoviruses of unknown concentration with plaque assay-tittered standards (Supplementary Fig. 2). Phenol red-free supernatants from virus production assays with EnAd-SA-fLuc were collected, weighed, and cleared by centrifugation. Neat or diluted 25 µL samples, along with a standard series (5.57 × 105 PFU·mL−1 2.72 × 102 PFU·mL−1), were used to infect 1.5 × 104 A549 cells seeded the previous day in 50 µL phenol red-free DMEM + 10% FBS (Gibco) on solid black 96-well plates (Corning). Plates were then spun at 600×gn for 10 min. At 20 to 24 hpi, luminescence was measured using a PolarStar plate reader (BMG Labtech) after injecting 25 µL of a 1.2 mg·mL−1 D-Luciferin solution into the well (Gold Biotechnology). Bioluminescence was recorded over six kinetic cycles at 0.5 s intervals, and virus concentrations were calculated via four-parameter fitting of integrated luminescence (MARS, BMG Labtech). At least three technical replicates were performed for each infection.
Assay for capsid integrity
To analyze the protein composition of adenovirus particles in supernatants, the culture medium was centrifuged (300×gn, 5 min), filtered (0.22 µm), and concentrated using a 300 kDa cut-off ultrafiltration unit (Sartorius). The retentate was washed twice with PBS, and protein concentration was measured using the QuantiPro BCA Kit (Thermo Fisher). Two micrograms of lysate in RIPA buffer (Thermo Fisher) were separated as for immunoblotting, stained with SyproRuby (Thermo Fisher) and imaged on a Biorad Chemidoc imager.
Absolute quantification of viral mRNA and genomic DNA by qPCR
DNA was extracted from infected cell pellets or supernatants using the PureLink genomic DNA Mini Kit (Invitrogen), and total RNA using RNeasy Mini Kit (Qiagen). RNA integrity was routinely tested by capillary electrophoresis (Tape Station, Agilent). For cDNA synthesis, 500 to 800 ng of RNA was transcribed using the QuantiTect Kit (Qiagen). PCR reactions (20 µL) contained 10 ng of cDNA or DNA with 2x qPCRBIO Probe Mix Hi-ROX (TaqMan assays, PCR Biosystems). Fiber, E1A, and E2B 200 bp-standards for absolute quantification were synthesized (IDT Technologies, Supplementary Fig. 7). Fiber detection was used to quantify genome copy number. TaqMan probes were labeled with JOE or FAM and quenched with BHQ1 (Sigma). Measurements were performed on the StepOnePlus cycler (Applied Biosciences). Copy numbers were calculated by linear regression on standard curves (Supplementary Fig. 7). Primer sequences are available in Supplementary Table 3.
To quantify encapsidated genomes, 200 µL of infected supernatants were treated with 100U Benzonase (Millipore) for 3 h at 37 °C. Before DNA extraction and fiber quantification by qPCR, benzonase was inactivated with EDTA (0.1 M final concentration) and heat-inactivation at 80 °C for 10 min.
Flow cytometry
Cells were detached using cell dissociation buffer (GIBCO), stained for 10 min with LIVE/DEAD fixable near-infrared staining kit (1:5000, Thermo Fisher), and then fixed in 2% neutral-buffered formalin for 10 min (Sigma). Antigens were blocked in 2 mM EDTA + 0.5% BSA with Fc-block (1:100, Biolegend) in PBS for 20 min. Cells were then stained with anti-CD46 antibody (1:200, clone TRA-2-10, Biolegend) or the corresponding isotype control for 30 min in PBS + 2 mM EDTA + 0.5% BSA. Stained cells were measured on the FACSCalibur flow cytometer (BD Biosciences). Analysis was performed on cell populations excluding debris, doublets and dead cells.
Real-time monitoring of cytopathic effect
Loss of cell adhesion due to cytopathic effect was monitored in real-time using the xCELLigence RTCA DP instrument (Roche). xCELLigence plates were equilibrated with 50 µL of medium for 1 h to establish a baseline. Then, 10,000 cells were plated in 100 µL of growth medium. Infections were initiated with 50 µL, and impedance (cell index) was recorded every 15 min. Measurements were normalized 4 h before infection.
Immunoblotting
Immunoblotting was performed as described previously, with blot development on X-ray films (GE Healthcare) or the Biorad Chemidoc imager.12 Dot blots were performed to quantify the secretion of deca-His-tagged EpCAM/CD3-BiTE, as described previously.57 Nuclear and cytoplasmic fractionation was performed using the NE-PER™ Kit according to the manufacturer’s instructions (Thermo Scientific). Antibody dilutions and blocking conditions are specified in Supplementary Table 4. Semi-quantitative assessment of band intensity was performed by densitometry using the ImageLab software (Biorad), with intensity normalization to the loading control of the same run.
Cell cycle synchronization and analysis
Cell synchronization was achieved using a double thymidine block: 2 mM thymidine (Sigma) was added for 16 h, followed by an 8-h release, and another 16-h block. For cell cycle analysis, cells were pulsed with 20 µM bromodeoxyuridine (BrdU, Sigma) for 10 min before harvest. Cells were subsequently detached using trypsin, fixed in ice-cold 70% ethanol for 30 min, and treated with 2 M HCl + 1 mg·mL−1 pepsin (Sigma) for 20 min. After two PBS washes, cells were stained with a primary mouse anti-BrdU antibody (1:100, 90 min, BD Biosciences) and a secondary goat anti-mouse-488 Fab-fragment (1:500, 60 min, Thermo Fisher) in DPBS + 2% FBS. After washing, cells were stained with 50 µg·mL−1 propidium iodide and 400 µg·mL−1 RNase A. Flow cytometry was performed on an Attune flow cytometer (Thermo Fisher), and data were processed with FlowJo v10.0.7r2 software (TreeStar Inc., USA).
Assay for promoter activity
In total, 5 × 10⁴ cells were seeded in 24-well plates. The next day, 200 ng of plasmid per reaction (~600–800 pmol) and 2.5 µL Lipofectamine 2000 per 1000 ng DNA were each diluted in 150 µL Optimem (Thermo Fisher), combined, and incubated for 30 min before adding to cells. After overnight incubation, the medium was replaced to start hypoxic exposure. Co-transfection was performed with one twentieth of the transfected plasmid mass of a β-galactosidase-encoding plasmid (SV40-betaGal, Promega) to normalize for transfection efficiency. Cells were washed with PBS, lysed, and subjected to one freeze-thaw cycle before measuring luciferase activity in black 96-well plates with luciferin reagent (Promega). Luciferase expression was normalized to β-galactosidase activity, which was measured by incubating lysates with 2× ONPG buffer (137 mM HNa2PO4 + 63 mM H2NaPO4 + 2 mM MgCl2 + 0.7% (v/v) β-mercaptoethanol + 1.33 mg∙mL−1 ortho-Nitrophenyl-β-galactoside, pH 7.2) at 37 °C for 10 min. Reactions were stopped with 1 M Na₂CO₃, and OD was measured at 410 nm.
Puromycin incorporation assay
Global translation was assessed by pulsing cells with 1 µg∙mL−1 puromycin (Sigma) for 30 min. Pre-treatment with cycloheximide (100 µM) served as a negative control. Lysates were then subjected to immunoblotting.
Reverse transfection of siRNA
siRNA transfections were conducted at a ratio of 10 pmol siRNA to 1 µL RNAiMax (Thermo Fisher) in Optimem (Thermo Fisher). After mixing, the solutions were incubated for 30 min before plating 3 × 10⁴ cells (24-well) or 1.2 × 10⁵ cells (6-well). After overnight incubation, the medium was replaced with complete medium to start hypoxic treatment. siRNA sequences are listed in Supplementary Table 5.
Sequence alignment of adenovirus genomes
Putative HREs were identified by manual scanning of annotated promoters of Ad11p (GenBank: AY598970.1). Corresponding regions from 59 adenovirus genomes (Supplementary Table 1) were aligned using the CLC Genomics Workbench (Qiagen). Nucleotide frequencies of the HRE motif across all genomes were visualized as a sequence logo.
Chromatin immunoprecipitation-qPCR (ChIP-qPCR)
In total, 6 × 106 cells were plated onto 15-cm plates. Chromatin crosslinking was done with 1% methanol-free formaldehyde (v/v) for 10 min with gentle rocking, followed by quenching with 125 mM glycine for 10 min. Cells were washed twice on ice with PBS, scraped into 5 mL of PBS, and pelleted at 500× g. Pellets were resuspended in SDS lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris pH 8.1) with protease inhibitors (1:30, Thermo Fisher).
Sonication was performed at 4 °C using a Bioruptor (Diagenode) for 2:45 min (15-second pulses on/off, high power) followed by benzonase treatment (15 U, Milipore) for 25 min for adequate fragmentation of viral and host chromatin. DNA fragmentation (~300 bp) was confirmed via TapeStation. Protein A agarose beads (40 µL, Millipore) were pre-washed with ChIP diluent and then incubated with sonicated DNA for 1 h at 4 °C on an end-over-end rotator. As previously described,58 15 µL of HIF-1α (PM14), HIF-2α (PM9), HIF-1β (Novus NB100-110), or pre-immune sera (10 µL) were used per ChIP reaction and incubated overnight at 4 °C. After bead addition (90 µL), samples were incubated for 1.5 h at 4 °C with end-to-end rotation, and then pelleted at 380×gn for 8 min. Washes (800 µL) were performed with low- and high-salt buffers and LiCl buffers at 4 °C for 5 min each. The samples were pelleted at 380×gn at 4 °C for 5 min, then washed twice in 800 µL TE. A two-step elution (120 µL buffer each step) was performed at room temperature for 15 min with shaking at 1400 rpm. For ChIP buffer components refer to Supplementary Table 7.
Reverse crosslinking was achieved by adding 12.5 µL of 4 M NaCl and heating at 65 °C overnight. Proteinase K (2 µl of 20 mg·mL−1) was added for 4 h at 45 °C, followed by RNase A (1 µL) at 37 °C for 30 min. Precipitated DNA was purified using MinElute columns (Qiagen) and eluted in 20 µL of DNase-free water. qPCR was performed using the 2× SyGreenBlue Hi-ROX in 20 µL (PCR Biosystems), taking PCR efficiency into account (Supplementary Table 6).
Statistical analysis
All statistical analyses and plots were performed using Prism v.10.6 (Graphpad). Replicates (n) representing biological variance are plotted as single data points, and technical replicates are indicated where relevant. Bars represent the mean ± standard deviation, unless otherwise stated. For statistical assessment, log-normal data were selected to allow the use of parametric tests. Two-sided unpaired t tests assuming unequal variance for conservative significance evaluation were performed (Welch´s correction). Holm–Šidák correction was applied for multiple comparisons.
