NADH-Triggered Photogenerated Carbon Radicals as General Abstracting Agents for Synergism of Apoptosis and Ferroptosis

Abstract

Free radicals, with their high reactivity, offer a promising strategy for tumor treatment by effectively circumventing drug resistance. Free-radical-based cancer therapies include radiotherapy, chemotherapy, and emerging approaches such as chemodynamic therapy (CDT) and photodynamic therapy (PDT). Conventional PDT, widely used in cancer treatment, relies on excited photosensitizers generating reactive oxygen species (ROS) via electron or energy transfer to kill cancer cells. However, the generation of both Type I and Type II ROS critically depends on microoxic and normoxic conditions, and the hypoxic tumor microenvironment severely limits therapeutic efficacy. Enhancing the cancer-cell-killing effect of PDT remains challenging. Herein, we constructed an S-rhodamine photosensitizer activated upon photoirradiation and NADH induction, generating a sulfur-bridged xanthene-based triarylmethyl radical (TAM^•) in an oxygen-independent manner. Acting as a broad-spectrum hydrogen atom abstractor, this radical attacks biomolecules within cancer cells more effectively than ROS by extracting hydrogen atoms from endogenous biological substrates, depleting intracellular NADH and glutathione (GSH), and causing critical DNA damage. These actions induce severe DNA damage and disrupt intracellular homeostasis, ultimately eradicating cancer cells through a combination of ferroptosis and apoptosis. It exhibited potent antitumor activity in the 4T-1 tumor-bearing mouse model. This design strategy for modulating the stability of photogenerated carbon radicals holds significant promise for the development of novel anticancer agents.

Photodynamic therapy (PDT) has emerged as a rapidly developing cancer treatment modality in recent years. Compared to conventional radiotherapy, chemotherapy, and surgery, PDT offers advantages such as noninvasiveness, minimal invasiveness, negligible drug resistance, and controllable treatment areas. − However, the hypoxic nature of the tumor microenvironment poses a significant challenge to the efficacy of both Type-I and Type-II PDT. − The development of enhanced photodynamic therapy strategy is highly in demand. , In 2022, Peng’s team reported a novel photosensitizer, EBSe, specifically targeting RNA. Under photoactivation, EBSe directly and efficiently disrupts RNA, causing its unwinding, and was successfully applied in PDT for hypoxic tumor-bearing mice. Beyond RNA, a variety of biological macromolecules play critical roles in living organisms. NADH (reduced nicotinamide adenine dinucleotide) and its phosphate ester NADPH are crucial coenzymes and key electron carriers within living cells, participating in electron transfer and hydrogen atom transfer (HAT) in numerous cellular redox reactions. − Subsequent studies have reported numerous instances of photocatalytic therapy (PCT) facilitating intracellular NADH/NAD⁺ redox conversion. − The Kim group further developed a bioorthogonally activatable Si-rhodamine-based photocatalyst for NADH/NAD⁺ conversion in vivo, enabling conditionally activated photocatalytic therapy of tumors using organic small molecules. These abundant findings substantiate the generality and feasibility of this approach.

While oxygen radicals are extensively employed in Type-I PDT, recent research focuses on enhanced PDT mediated by other radicals. In 2020, Song et al. developed SiO₂/Ag/AgCl hybrid materials generating photogenerated chlorine radicals (Cl^•) oxygen-independently, inducing cellular oxidative stress and DNA double-strand breaks. In 2023, Peng et al. developed TAF, a photogenerated hydrogen radical (H^•) photocatalyst that, under GSH activation and two-photon NIR excitation, produced ROS while depleting GSH, disrupting cellular redox homeostasis, and inducing ferroptosis and apoptosis. (Scheme a, left) However, the structure–activity relationships (SARs) of H^• and Cl^• remain poorly understood, hindering rational modulation of their properties.

1. (a) Photogenerated H^• and Cl^• in Reported Works and Aryl Engineering for Modulating the Carbon Radical Stability in This Work; (b) Proposed Dual-Lock Mechanism of NADH Photoactivation and Post-Triggered Radical Mediated Oxidative Damage on Biological Substrates.

Compared with oxygen radicals, hydrogen radical, and chlorine radical, carbon-centered radicals exhibit high structural tunability, achieving distinct application-specific objectives. For example, some triarylmethyl radicals show high stability by tuning the substituents on benzene or replacing benzene with pyridine. , Importantly, these stable carbon-centered radicals generated from chemical oxidation or reduction can react with NADPH, heme-containing proteins, and glutathione. Very recently, the Yang group found photogenerated carbon radicals efficiently oxidize NADH by HAT reaction.

Inspired by these advancements, we developed several aryl-substituted S-rhodamine photosensitizers and found that the 3-benzothiophene-substituted S-rhodamine (3-BT) photosensitizer liberated carbon-centered radical activated by photoirradiation and NADH induction. Beyond the generation of ROS, this photosensitizer exhibits dual key roles: first, the reaction depletes NADH and generates triarylmethyl radical with suitable stability; second, functioning as a broad-spectrum hydrogen atom abstractor, the resulted radical attacks biologically active substrates including DNA and GSH in an oxygen-independent manner. It synergistically triggers cell death through the combined induction of ferroptosis and apoptosis, enabling highly efficient cell eradication. 3-BT exhibits potent cytotoxicity with an IC₅₀ as low as 12.5 nM in 4T-1 cells without heavy-atom substitution. It also showed remarkable in vivo antitumor efficacy against 4T-1 tumors in mice. This represents the first report of constructing a photosensitizer molecule by modulating the stability of a carbon-centered radical. This approach provides a valuable reference strategy for the future development of photosensitizer molecules based on novel radical mechanisms.

Results and Discussion

Molecular Design, Synthesis, and Characterization

Four aryl-substituted S-rhodamine photosensitizers (X-Ars) were designed and prepared by introducing thiophene and benzothiophene moieties to the 9-position to modulate the properties (Figure a). The absorption/emission spectra and photophysical/chemical properties of the X-Ars series were characterized. While the fundamental photophysical properties of the X-Ars series were highly similar, compound 3-BT exhibited an unusually high fluorescence quantum yield (Table and Figures S1–S21a). We attribute this to the combined electron-donating capacity of the 3- and 6-position substituents coupled with the greater steric hindrance imparted by the 3-benzothiophene moiety. This combination likely results in a radical intermediate adopting a configuration with a relatively high stability. To investigate the electron-accepting capability of the X-Ars series, cyclic voltammetry was employed to determine and calculate their excited-state reduction potentials (3-BT: E* = 1.33 V vs NHE). These results indicate that the oxidation of NADH by X-Ars photosensitizers is thermodynamically favorable (Figures S3 and S4). The single-crystal structure of 3-BT reveals that the dihedral angle between the 3-benzothiophenyl group and the xanthene scaffold is 65.38°. The intermolecular π–π stacking distance measures 3.35 Å (<3.5 Å), which falls within the typical range for π–π interactions (Figure b,c and Table S1). This structural feature may presumably be correlated with the high molar extinction coefficient and high quantum yield observed for 3-BT.

1.

(a) Design strategy of X-Ars based on stepwise modifications. (b) X-ray crystallographic structures of 3-BT and the dihedral angle between xanthene core and benzothiophene. (c) Crystal packing and the close contact of 3-BT. (d) Total ROS generation of X-Ars (10 μM) in PBS (pH = 7.4) after irradiation of laser (565 nm, 5 mW/cm²) using DCFH-DA as indicator. (e) ¹O₂ generation of X-Ars (5 μM) in PBS (pH = 7.4) after irradiation of laser (565 nm, 5 mW/cm²) using ABDA (50 μM) as trapping agent. (f) O₂ ^•– generation of X-Ars (2.5 μM) in PBS (pH = 7.4) after irradiation of laser (565 nm, 5 mW/cm²) using DHR123 (15 μM) as trapping agent. (g) Cyclic voltammograms of X-Ars (10 mM) in acetonitrile with (n-Bu)₄N⁺PF₆ (100 mM) as the supporting electrolyte, Ag/Ag⁺ as a reference electrode, glassy carbon electrode as a working electrode, and Pt wire as a counter electrode (scan rate = 2 mV/s).

1. Basic Photophysical Properties of X-Ars.

comp.	λabs [nm]	λem [nm]	stokes shift [nm]	ε (104 M–1 cm–1)	φf (%)
2-TH	587	614	27	1.5	0.80
2-BT	589	621	32	1.9	0.01
3-TH	573	600	27	1.9	0.29
3-BT	580	608	28	3.4	18.80

In Vitro Photosensitization and Spectroscopic Properties

Furthermore, we characterized the photosensitizing properties of the X-Ars. ROS generation capabilities of four photosensitizer agents were assessed in PBS (pH = 7.4) using specific molecular probes: 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) for total ROS, 9,10-anthracenediyl-bis­(methylene)­dimalonic acid (ABDA) for singlet oxygen (¹O₂), dihydrorhodamine 123 (DHR123) for superoxide anion (O₂ ^•–), and hydroxyphenyl fluorescein (HPF) for hydroxyl radical (^·OH) (Figures S5–S7 and d–f). The results indicated that the X-Ars series exhibits moderate ROS quantum yields (3-fold higher than Rose Bengal).

Rhodamine dye molecules inherently carry a positive charge, predisposing them toward electron abstraction. Consequently, we investigated whether NADH could serve as an electron donor for the X-Ars series. Under illumination in deoxygenated PBS, the characteristic absorption peaks of both X-Ars and NADH decreased, indicating a redox reaction between them. Time-dependent absorption spectroscopy revealed that 3-BT reacted with NADH at the fastest rate in X-Ars (Figure a–d), and crucially, photoexcitation is required (Figure e). This demonstrates a kinetic advantage for 3-BT in the reaction with NADH. Notably, 3-BT, differing only in its 9-position substituent, exhibited a significant photo-oxidation advantage. We therefore hypothesize that the superior NADH oxidation efficiency of 3-BT stems from the formation of an active radical intermediate, specifically, the xanthene-based trityl radical at the 9-position.

2.

(a–d) Absorption of X-Ars (10 μM) in deoxygenated PBS containing NADH (1 mM) after laser irradiation (565 nm, 5 mW/cm², 1 min). (e) Plots of F–F ₀ of 3-BT at 608 nm upon light irradiation (565 nm, 5 mW/cm²) for different time intervals. (e) Plots of F–F ₀ of 3-PT (1 μM) at 608 nm upon light irradiation (565 nm, 5 mW/cm²) with TEMPO (100 μM) for different time intervals. (g, h) ESR spectra to detect carbon-centered radicals generated by 3-BT (0.2 mM) ((h): triggered by 20 mM NADH) in deoxygenated PBS under light irradiation, using DMPO (10 mM) as a spin trap agent.

Mechanism Validation of 3-BT and NADH Photoreaction

Furthermore, to validate the critical role of this active carbon radical intermediate in the photo-oxidation of NADH, we performed radical quenching experiments. The addition of the radical scavenger 2,2,6,6-tetramethylpiperidinooxy (TEMPO) to the test system significantly inhibited the photo-oxidation of NADH by 3-BT, providing strong evidence that the process is mediated by the active carbon radical intermediate (Figure f). Subsequently, electron paramagnetic resonance (EPR) was employed to detect the generation of free radicals using 5,5-dimethyl-1-pyrroline N-oxide (DMPO) as a spin trapping agent to distinguish paramagnetic signals. In a deoxygenated solution of 3-BT under illumination, EPR spectra indicative of multiple radical adducts were observed. Through spectral simulation and deconvolution, characteristic signals corresponding to carbon-centered radicals (A _N = 15.4678 G A _H = 22.8121 G, g = 2.00651), hydroxyl radicals (A _N = 14.9 G A _H = 14.9188 G, g = 2.00655), and hydrogen radicals (A _N = 16.3476 G A _H = 23.0 G, g = 2.00643) were identified in the system (Figure g). To further investigate the photogenerated carbon-centered radicals, liquid chromatography ion trap time-of-flight mass spectrometry (LCMS-IT-TOF) was utilized to analyze the intermediates and products formed during the reaction. Ultimately, the adduct 3-BT–DMPO, resulting from the trapping of 3-BT radical by DMPO, was successfully detected (Figure S12). However, in the 3-BT + NADH system, a distinct singlet signal emerged, exhibiting characteristic features of a trityl radical without remote hyperfine coupling (Figure h). , Analysis assigned this signal to the central carbon atom at the mesoposition of the xanthene core. Furthermore, the hydrogenated product 3-BT-Re and NAD⁺ were detected through HAT to the trityl radical (Figures S8–S12). These results provide supportive evidence for the plausibility of the proposed reaction mechanism. Although some xanthene-based trityl radicals exhibit remarkably long half-life time, potentially up to 1550 h, our tests on the irradiated X-Ars + NADH system without DMPO indicated that the photogenerated carbon-centered radical is not stable as the conventional trityl radical. Because of this metastable state, the potential for subsequent free-radical-mediated reactions occurs in vivo.

Confocal Imaging and Cell Cytoxicity

Rhodamine dye molecules inherently carry a positive charge, ,− which facilitates electrostatic interactions with the negatively charged inner mitochondrial membrane. Consequently, many rhodamine derivatives lacking strong targeting groups localize to mitochondria. , Colocalization experiments using MitoTracker Green and the probe 3-BT confirmed that 3-BT primarily localizes to the cytoplasmic matrix and mitochondria, demonstrating strong mitochondrial targetting ability, with a Pearson’s colocalization coefficient of 0.826 (Figure a). Subsequent time-lapse imaging of 3-BT revealed a gradual decay in its fluorescence signal. Based on the observed photo-oxidation advantage of 3-BT with NADH, we propose that this photobleaching phenomenon may result from a light-induced interaction between 3-BT molecules and endogenous NADH within the cytoplasmic matrix and mitochondria, leading to fluorescence quenching via a photochemical reaction (Figure b).

3.

(a) CLSM images of HepG2 cells coincubated with 3-BT (1 μM) and MitoTracker Green (scale bar: 10 μm). (b) Time series CLSM images of HepG2 cells incubation with 3-BT (1 μM). (c–f) Cell viability of 4T-1 cells subjected to varying concentrations of X-Ars under light irradiation (565 nm, 5 mW/cm², 10 min) under normoxic condition. (g) Schematic diagram illustrating the physiological functions of the four mitochondrial complexes and ATP synthesis during mitochondrial respiration. (h) Cell viability of HepG2 cells subjected to varying concentrations of 3-BT under light irradiation (565 nm, 5 mW/cm², 10 min) under normoxic and hypoxic conditions.

Building upon the moderate ROS generation capacity and NADH photo-oxidation capability of the X-Ars series, we further evaluated the PDT efficacy of the four molecules against 4T-1 and HepG2 cells using the cell counting kit-8 (CCK-8) assays under normoxic and hypoxic condition. All four molecules exhibited similar cytotoxicity under dark conditions with half-inhibitory concentrations (IC₅₀) around 2 μM. However, following irradiation with 565 nm light (5 mW/cm²) for 5 min, a significant difference emerged. 3-BT exhibited remarkably higher photocytotoxicity against 4T-1 cells, achieving an IC₅₀ as low as 12.5 nM, while the other X-Ars compounds showed IC₅₀ values of 125 nM (Figure c–f). Serving as a key substrate for mitochondrial Complex I, NADH participates in the critical process of oxidative phosphorylation within the mitochondrial respiratory chain (Figure g). Following the same irradiation conditions, we evaluated 3-BT’s cytotoxicity under hypoxia. The results revealed that the cytotoxicity of 3-BT under hypoxia was remarkably similar to that observed under normoxic conditions. This finding strongly indicates that the depletion of NADH and the generation of carbon-centered radicals synergistically contribute to the induction of cell death (Figure h). The much higher photocytotoxicity of 3-BT might be ascribed to the fastest reaction rate with NADH and the relative stability of the carbon-centered radical.

The Phototherapeutic Effect of 3-BT in HepG2 Cells

To further elucidate the mechanisms of cell death during PDT and explore potential cell death pathways, we investigated the 3-BT light-induced apoptosis process and the key malondialdehyde (MDA) and lipid peroxidation (LPO) indicator of ferroptosis using combined Annexin-V FITC/PI apoptosis detection kit, MDA Assay kit, and C11 BODIPY 581/591 LPO detection kit with flow cytometry. Following 3-BT light induction, Annexin-V FITC and PI exhibited double positivity, indicating an increasing proportion of cells under apoptosis, similar to the induction results of the commercially available mitochondrial apoptosis inducer CCCP (Figure a,b). Concurrently, the MDA and LPO signals also showed strong positivity, comparable to the induction results of the established ferroptosis inducer Erastin. These results demonstrate that 3-BT primarily induces cell death through both the apoptosis and ferroptosis pathways during PDT , (Figure c,d).

4.

(a, b) Flow cytometric detection of HepG2 cell apoptosis and ferroptosis induced by CCCP, Erastin, and 3-BT + hv. (c) Fluorescence intensity by flow cytometric detection of HepG2 cell LPO accumulation. (d) Fluorescence intensity by flow cytometric detection of HepG2 cells MDA accumulation. (e) Relative NADH content versus control group. (f) Relative ATP content versus blank and control group. (g) Western blot analysis of the ferroptosis-related protein GPX4 and the apoptosis-related protein Caspase-3 under different conditions. (h) Protein level of Caspase-3. (i) Protein level of GPX4. (j) TEM images of HepG2 cells morphology and subcellular organelle structure under different conditions (green rectangular outline demarcates the enlarged region, while the red dashed line delineates the mitochondrial boundaries, scale bar, 1 μm in original view and 500 nm in enlarged view).

NADH serves as a key hydride-donating substrate for Complex I in the electron transfer chain. A sharp decline in NADH concentration disrupts the cellular redox environment and simultaneously inhibits the mitochondrial respiratory chain, ultimately preventing the synthesis of ATP, the cellular energy currency. ,,,, Given that 3-BT undergoes photoreaction with NADH, we subsequently monitored intracellular NADH levels and ATP energy metabolism using the NADH/NAD⁺ kit and ATP detection kit. Post 3-BT light induction, HepG2 cells exhibited a severe reduction in NADH levels to approximately 20% and a drastic decrease in ATP content to below 5% (Figure e,f). This ultimately led to mitochondrial dysfunction, inhibition of oxidative phosphorylation, and disruption of cellular energy metabolism.

The apoptosis process is accompanied by upregulation of the marker protein Caspase-3, while the ferroptosis process involves downregulation of the marker protein GPX4. Western Blot analysis confirmed that 3-BT synergistically activated apoptosis and ferroptosis death pathways in HepG2 cells (Figure g–i), collectively resulting in high photocytotoxicity. ,

Finally, we examined the PDT process using transmission electron microscopy (TEM). Compared to control cells, HepG2 cells in the dark control group treated with 3-BT displayed abnormal mitochondrial morphology and electron transport, accompanied by endoplasmic reticulum–mitochondria contacts, exhibiting early features of mitophagy. In the light-treated group, we observed complete vacuolization of intracellular mitochondria and partial cellular fragmentation. This demonstrates that 3-BT light induction caused dysfunction in mitochondrial energy metabolism and electron transport, subsequently leading to structural rupture of mitochondria and cell death (Figures j and S13).

H-Abstraction to Damage Biosubstrate, HepG2 3D Spheroids Eradication, and Invasion Ability Evaluation

Given the relatively stable nature and high reactivity of radicals, we hypothesized that the carbon radical 3-BT·, generated upon NADH initiation, might attack biologically relevant substrates typically considered inert such as GSH and DNA. To test this, we performed UV–vis absorption spectroscopy on GSH within the 3-BT/NADH system. Using the DTNB (5,5′-dithiobis (2-nitrobenzoic acid)) assay, we analyzed the formation of the yellow NTB^2– product (absorbance at 412 nm) derived from reduction by GSH. Without NADH, GSH almost remained unreactive during the irradiation of probe 3-BT with GSH, while, in the presence of NADH, GSH was almost depleted (Figure a,b).

5.

(a) Comparison between the NADH-activated enhanced PDT effect and the conventional PDT effect. (b) Absorption spectra of 3-BT (10 μM) before and after 1 min light irradiation in deoxygenated PBS containing 100 μM NADH, 10 μM GSH, and 10 μM DTNB. (c) Color-inverted image of nucleic acid gel electrophoresis (contain 3% agarose) to detect 3-BT (1 μM) and 3-BT­(1 μM)+NADH (100 μM) photo-oxidative damage to DNA ladder makers (5 μL) after exposure to light under hypoxic conditions (5 mW/cm², 5 min). (d) Fluorescence images of HepG2 cells 3D-structured spheroids under 10x objective lens. Red channel (3-BT): λ_ex = 546 nm, λ_em = 590–620 nm, scale bar: 100 μM. (e) Wound healing assay to evaluate the migration ability of HepG2 cells, scale bar = 25 μm.

These results showed that probe 3-BT consumes GSH by radical formation initiated by NADH and subsequently reacting with GSH.

Second, the DNA damaging capability of 3-BT versus 3-BT + NADH was assessed using agarose gel electrophoresis. As shown in Figures c and S14, without NADH, only partial DNA was fragmented, while with NADH assisting, DNA was critically damaged (lane 4 vs lane 3). These results indicate the radical process led to more severe DNA fragmentation than the ROS strategy, which further explains the high phototoxicity of 3-BT. Collectively, the NADH-triggered 3-BT radicals function as effective hydrogen atom abstractors, capable of damaging key intracellular hydrogen atom donors such as GSH and DNA.

To bridge the gap between 2D cell cultures and solid tumors and to better simulate the more densely packed cellular architecture and complex physiological environment inherent to actual tumors, we established HepG2 3D spheroids. These spheroids more closely recapitulate the physiological conditions of tumor cells. We then evaluated the cytotoxic efficacy of 3-BT against these spheroids. Compared to the control group, 3-BT treatment induced significant structural disintegration (“collapse”) of the 3D spheroids. Collectively, these findings demonstrate that 3-BT effectively kills tumor cells and impedes their self-renewal capacity and invasive potential (Figure d). Subsequently, to validate the effect on cellular invasion and migration, Transwell wound healing assays were conducted. The results demonstrated that compared to the PBS group, 3-BT treatment combined with light irradiation significantly suppressed the migration and invasion of HepG2 cells (Figure e).

The Antitumor Therapeutic Effects In Vivo

Female BALB/c mice aged 5–6 weeks were obtained from Beijing HFK Bioscience Company. All mice were housed in designated animal facilities, fed ad libitum, and inspected regularly. All animal studies were approved by the Sichuan University Animal Charity Protection and Treatment Committee and performed in accordance with human care and use of research animals, approval 20240226173. Inspired by the findings from in vitro experiments, we further investigated the PDT efficacy of 3-BT in a tumor-bearing mouse model using BALB/c mice with subcutaneously implanted 4T-1 tumors. Fluorescence imaging revealed that BSA@3-BT was retained at the tumor site for up to 4 h, attributable to the enhanced permeability and retention effect. We subsequently evaluated the antitumor efficacy of 3-BT in mice bearing 4T-1 tumors (initial tumor volume: ∼150 mm³). Mice received an intratumoral injection of 3-BT (100 μM, 50 μL), followed by light irradiation (532 nm, 100 mW cm²) at the tumor site 4 h postinjection (Figure a). Body weight and tumor volume were recorded over 14 days. A slight increase in body weight was observed in all of the groups. In contrast, tumor growth was significantly suppressed in the group treated with 3-BT plus light compared to that of the control group, indicating effective inhibition of tumor growth by 3-BT in vivo. Under the same conditions, the antitumor effect of 3-BT in the absence of light was markedly weaker than that in the 3-BT plus light group, suggesting that light irradiation substantially enhanced the PDT efficacy of 3-BT. At day 15, the mice were sacrificed, and all tumor tissues were excised and weighed. As shown in the results, both tumor volume and weight in the treated group were significantly lower than those in the control group (Figures b-e). Furthermore, TUNEL immunofluorescence staining indicated significant cell death. H&E staining of tumor sections from the 3-BT + Light group on day 15 revealed notable vacuolization and nuclear condensation within tumor cells. These results demonstrate that 3-BT effectively inhibits tumor growth upon light irradiation while exhibiting minimal systemic toxicity, further highlighting the promising application prospects of carbon radicals-mediated HAT-enhanced PDT. (Figures f and S15)

6.

(a) Schematic diagram of vivo treatment. (b) Tumor picture and (c) the corresponding tumor weight with different treatments. (d) Tumor volume with different treatments, respectively. (e) The body weight with different treatments. (f) H&E and Tunel staining primary tumor for different treatment groups (n = 3). Data was presented as mean ± SD. Not significant = n.s. (p > 0.05). Statistical significance (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001). Scale bar = 50 μm.

Conclusions

In summary, through aryl engineering at the 9-position of S-rhodamine, we have identified an oxygen-independent photosensitizer, 3-BT, featuring a triarylmethyl radical intermediate with suitable stability triggered by NADH. Besides ROS generation, the unique photoreactive properties of 3-BT enable a dual mechanism: (1) under light activation, it depletes cytosolic and mitochondrial NADH pools, inducing mitochondrial electron transport chain dysfunction and suppressing ATP synthesis via oxidative phosphorylation. This reduces NADH levels to approximately 20% of normal cellular levels and ATP content below 5%, ultimately leading to cellular “asphyxia”; (2) concurrently, as H atom abstractor, the photogenerated radical in hypoxic environments further depletes GSH and fragments DNA by HAT, which was validated by GSH reactivity and DNA damage experiments with or without NADH. HepG2 cells were eradicated by synergistic ferroptosis and apoptosis. It exhibited potent antitumor activity in the 4T-1 tumor-bearing mouse model. This work not only inspires the rational design of hypoxia-tolerant photosensitizers for HAT-enhanced therapeutic applications but also establishes a novel and reliable strategy for developing efficient antihypoxic tumor therapies targeting oxygen-deprived microenvironments.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacsau.5c01192.

• Detailed synthetic procedures; ¹H NMR, ¹³C NMR, HRMS, absorption, and fluorescence data; and additional information as noted in the text, cell cytotoxicity, and EPR analysis (PDF)

• 3-BT Time series (MP4)

X.Z.: Conceptualization, methodology, formal analysis, investigation, data curation, visualization, writing (original draft preparation); K.–K.Y.: experimental arrangement, guiding data analysis; H.Z.: experimental technical guidance; Y.Z., A.W.: flow cytometry analysis; A.W., X.W.: paper decoration and color matching; M.-Y.W., X.-Q.Y., S.-Y.C., K.L.: writing (review and editing) and project administration.

This work was financially supported by the National Natural Science Foundation of China (22577084, U25A20592) and the Science and Technology Department of Sichuan Province (Nos. 2024ZHCG0151).

The authors declare no competing financial interest.