Breaking photoswitch activation depth limit using ionizing radiation stimuli adapted to clinical application

The study addresses the central limitation of photoactivated therapeutic systems: their reliance on ultraviolet to near-infrared light, which penetrates only millimetres into tissue, restricting clinical translation. The authors propose using ionising radiation (X-rays, gamma rays, electron beams) as deeply penetrating external stimuli to initiate local energy conversion into low-energy particles and reactive species that can activate photosensitive molecules in situ. They focus on designing a radioswitch based on an azobenzene photoswitch linked to a gadolinium chelate, enabling IR-triggered cis-to-trans isomerisation and downstream pharmacological effects, aiming to merge the spatial reach of radiotherapy with the molecular specificity of photochemistry.

Prior work in stimulus-responsive therapeutics includes endogenous triggers (pH, redox, enzymes) and exogenous triggers (light) enabling precise spatiotemporal control. Traditional photosensitive systems initially used UV light, then evolved to visible/NIR-absorbing compounds and nanoparticles to increase penetration, enabling in vivo small-animal applications and limited clinical uses like photodynamic therapy and photoimmunotherapy. However, depth limitations persist. In radiotherapy, dose enhancement was shown with iodine agents and with radiosensitizing nanoparticles, sometimes engaging immune responses. Emerging IR-activated systems fall into (i) down-converters (e.g., nanoscintillators) producing UV–vis to activate PDT and (ii) ROS/electron-mediated systems causing bond cleavage, DNA damage, or oxidation to release drugs or gases. The authors identify a gap: using IR to induce non-destructive, specific molecular rearrangements typical of photoactivation (e.g., azobenzene isomerisation), rather than radiosensitization or bond scission.

Design and synthesis: An azobenzene derivative bearing a primary amine and para-alkoxy chains was synthesized and coupled via anhydride opening to a modified DOTA chelator, followed by Gd(III) complexation to yield GdAzo (5 steps, 33% overall yield). Photophysical characterisation: UV–vis, 1H NMR, and HPLC confirmed cis/trans signatures (π→π* at 322 nm for cis, 367 nm for trans) and thermal back-relaxation (cis half-life t1/2 = 2.3 h at 37 °C in PBS). IR sources and dosing: Samples (typically 50 µM in PBS/H2O, 200 µL) were pre-irradiated with 365 nm UV to generate cis-rich photostationary state (≈90% cis), then exposed to ionising radiation at clinically relevant doses (2–20 Gy): gamma rays (662 keV), X-rays (≈80 keV mean), or electrons (LINAC, 4.5 MeV). Isomerisation quantification: Absorbance spectra and HPLC (multiple methods) quantified the fraction of trans isomer formed; activation was expressed as change in trans proportion and as radiochemical yields (G-values, µmol/J). Kinetic analysis determined dose-dependent activation constants k in the low-dose regime. Monte Carlo simulation: PENELOPE code simulated secondary electron energy spectra from Gd under different IR modalities to relate primary interactions to secondary low-energy particle generation. Mechanistic studies: Radical scavenger and converter assays (tert-butanol, mannitol, ethanol, sodium azide, DMSO, cadmium perchlorate, sodium selenate) probed the role of specific species. Gas saturation (N2 to remove O2; N2O to convert hydrated electrons to HO•) assessed oxygen dependence and hydroxyl radical involvement. Fenton chemistry (Fe2+/H2O2 with EDTA) chemically generated HO• to test isomerisation without IR. Hydroxyl radical quantification: Coumarin dosimetry measured 7-hydroxycoumarin fluorescence to derive G(HO•) for XR and GR across Gd3+ concentrations (0–2 mM), evaluating Gd’s effect on HO• yields. Concentration effects: Varying cis-GdAzo concentrations assessed linearity and potential catalytic behavior. Biophysical characterisation: Trans-GdAzo aggregation and membrane interaction were studied by fluorescence, relaxivity (MRI), small-angle X-ray scattering (SAXS) to determine micelle formation (CMC ≈0.42 mM) and insertion into DPPC bilayers (d-spacing changes, electron density profiles), and differential scanning calorimetry (DSC). Cell studies: Confocal microscopy (PANC-1) with propidium iodide (PI) assessed membrane permeabilisation after adding cis-GdAzo and applying GR (2 Gy) versus controls; flow cytometry (CCRF-CEM ARAC-8C, Gem-resistant) quantified PI-positive events. Cytotoxicity: Cells exposed to cis-GdAzo ± GR (2 Gy), with or without gemcitabine (0.1 µM), followed by 4-day culture and viability counts (trypan blue). EELS-TEM mapped Gd distribution in cells. Controls included cis-Azo (no Gd) and Dotarem (Gd-chelate without azobenzene). MRI relaxivity assessed detectability.

• Radioswitch activation: Cis-GdAzo undergoes efficient IR-triggered isomerisation to trans-GdAzo across GR, XR, and E sources. HPLC and UV–vis showed no byproducts, and the process is reversible with UV light. Activation efficacy reached 33–69% over 2–20 Gy.
• Kinetics: Dose-related monoexponential behavior at low doses (≤2 Gy). Activation constants upon GR: k(cis-GdAzo) = (2.1±0.2)×10^-1 Gy^-1 versus k(cis-Azo) = (1.83±0.07)×10^-2 Gy^-1 (P<0.0001), indicating the critical role of Gd.
• Yields: At 2 Gy, G-value for trans-GdAzo formation was 81±2 (8.4±0.2 µmol/J) after correcting for thermal relaxation—substantially higher than typical HO• (0.28 µmol/J) or H2O2 (0.073 µmol/J) yields from water radiolysis.
• Source independence: Comparable activation across XR, GR, and E (e.g., at 2/5/10/20 Gy, molecular activation ~30.0/45.5/60.2/68.1% for XR; 33.4/50.3/62.6/68.6% for GR; 36.2/54.5/64.7/71.8% for E), and similar G-values/k, implying secondary low-energy species drive activation rather than primary interactions.
• Monte Carlo simulations: Despite differing high-energy electron outputs among sources, similar relative amounts of low-energy secondary electrons near Gd were produced, supporting a common activation mechanism via low-energy species.
• Mechanism: Hydroxyl radicals (HO•) are necessary and sufficient. HO• scavengers (tert-butanol, mannitol, ethanol) abolished activation unless converting HO• into other oxidants (e.g., NaN3, DMSO increased activation). Electron converters that generate oxidants (Na2SeO4) increased activation; electron scavenging without oxidation had no effect. N2 saturation (removing O2) did not reduce activation, whereas N2O saturation (converting e_aq− to HO•) increased activation consistent with doubled HO• yield. Fenton chemistry (Fe2+/H2O2) replicated activation; H2O2 alone did not.
• Proposed pathway: IR-generated HO• oxidizes the azo bond to a radical cation, rapidly isomerising cis to trans; subsequent reduction yields neutral trans-GdAzo. Catalytic (substoichiometric) pathways were excluded; activation scales linearly with cis-GdAzo concentration, with higher G at lower doses due to reduced non-specific losses.
• HO• yields and Gd effect: Measured G(HO•) ~0.200 µmol/J (XR) and ~0.280 µmol/J (GR/E). Adding Gd3+ increased bulk HO• yields up to 33% (XR, ~200 µM Gd3+) and 20% (GR, ~500 µM), plateauing then decreasing at higher [Gd3+] likely due to recombination. Presence of Gd3+ enabled activation of cis-Azo under GR.
• Biophysics: Trans-GdAzo is amphiphilic, forms micelles above ~0.42 mM; SAXS shows insertion into DPPC bilayers, increasing d-spacing (e.g., 63.7±0.3 Å to 65.9±0.8 Å at 0.1 mol%), and at higher fractions induces mixed micelle–bilayer structures and bilayer disordering.
• Cellular effects: Trans-GdAzo rapidly permeabilised PANC-1 cell membranes (PI uptake within minutes), while cis-GdAzo did not. IR-activated cis-GdAzo (2 Gy GR) significantly increased permeabilisation versus either agent alone. EELS-TEM detected Gd in localized cytoplasmic regions. In Gem-resistant CCRF-CEM ARAC-8C cells, IR-activated cis-GdAzo increased PI-positive events and reduced viability after 4 days; addition of gemcitabine did not further enhance cytotoxicity, consistent with cell collapse rather than enhanced drug uptake. Dotarem control showed no cytotoxicity, excluding pure Gd radiosensitization as the cause.
• Imaging: GdAzo exhibits MRI-detectable relaxivity, enabling potential theranostics.

The radioswitch concept successfully decouples tissue penetration from molecular activation by using ionising radiation to generate local low-energy species (notably HO•) near high-Z gadolinium, which in turn drive azobenzene cis-to-trans isomerisation. This addresses the fundamental depth limitation of conventional photoactivation while preserving molecular specificity and reversibility. The mechanism is robust across clinical IR sources and energies, indicating generalizability to standard radiotherapy settings. The oxygen independence (no reduction under N2) suggests potential utility in hypoxic tumors where photodynamic therapies underperform. Biophysical data show that the activated trans-GdAzo interacts with and perturbs lipid bilayers, explaining rapid cell permeabilisation and cytotoxicity observed in vitro. The theranostic capability via MRI detection of GdAzo further supports a targeted, image-guided approach. Collectively, these results demonstrate a pathway to bring diverse photoswitch-controlled actions into deep tissues, opening new avenues for precision radiotherapy-augmented therapies and on-demand actuation of molecular tools in clinical contexts.

The study introduces and validates a radioswitch strategy that activates an azobenzene-based photoswitch using clinically relevant ionising radiation doses, overcoming the depth limitations of traditional photoactivation. Mechanistically, activation is mediated by hydroxyl radicals generated near gadolinium upon IR, is source-agnostic, and yields substantial isomerisation with high radiochemical efficiency. The activated trans-GdAzo perturbs membranes, enabling rapid permeabilisation and cancer cell killing, and the Gd-chelate confers MRI visibility for theranostic use. Future work should focus on designing radioswitches with longer cis thermal half-lives suitable for in vivo applications, optimizing molecular structures for redox sensitivity and potency at lower concentrations, validating efficacy and safety in animal models (including hypoxic tumors), integrating with radiotherapy planning and imaging, and expanding to other photoswitchable scaffolds to trigger diverse therapeutic or biological functions under IR.

• Thermal stability: The cis-GdAzo thermal half-life is ~2.3 h at 37 °C in PBS, suitable for in vitro studies but insufficient for many in vivo applications; longer-lived systems are needed.
• Concentration requirements: Robust permeabilisation and cytotoxicity required relatively high cis-GdAzo concentrations in vitro; potency and delivery strategies must be improved.
• In vitro scope: Mechanistic and efficacy data are from solution and cell models; in vivo pharmacokinetics, biodistribution, tumor targeting, and safety remain to be established.
• Dose and radical kinetics: At higher doses, non-specific reactions and radical recombination reduce efficiency; optimization of dose rate and microenvironmental conditions may be necessary.
• No synergistic benefit with gemcitabine in resistant cells was observed under tested conditions, indicating the current membrane-disruptive effect leads to cell collapse rather than enhanced drug uptake.
• Dependence on HO• generation near Gd may vary with tissue composition and Gd distribution; precise control of localization will be critical.