Introduction

PDT utilizes photosensitizing agents in combination with light irradiation and molecular oxygen to generate cytotoxic singlet oxygen (1O2) and reactive oxygen species (ROS, eg, hydroxyl radicals, superoxide anions, hydrogen peroxide), which mediate oxidative damage to cellular membranes, proteins, lipids, and nucleic acids, ultimately triggering cell death.1 Mechanistically, photoexcitation induces the transition of photosensitizers from their ground state to triplet excited states via intersystem crossing (ISC). The excited-state photosensitizers subsequently engage in two distinct reaction pathways: energy transfer to molecular oxygen, generating singlet oxygen (1O2), or electron/hydrogen transfer reactions with biological substrates, yielding secondary ROS.2

As a precision therapeutic modality, PDT efficacy is governed by three critical parameters: photosensitizer biodistribution, localized light activation, and tissue oxygenation status.3 Clinically approved photosensitizers predominantly exert their effects through singlet oxygen (1O2) generation, which exhibits an ultrashort diffusion radius (10–20 nm) due to its transient lifetime (nanosecond scale).4–6 This spatial constraint necessitates precise subcellular targeting of photosensitizers.7 Present research demonstrates that numerous hydrophobic photosensitizers (such as phthalocyanine PC4 and benzoporphyrin derivative monoacid ring A) can localize in mitochondrial membranes and induce cytochrome C release from the mitochondria into the cytosol, triggering rapid apoptosis.8–10 Some hydrophobic photosensitizers can localize in the plasma membrane and initiate a cascade of signaling pathways that lead to rapid cell necrosis with shorter irradiation times.11 Several photosensitizers localized in lysosomes could lead to the release of lysosomal enzymes into the cytosol to cause slow apoptosis.12,13 These findings underscore the imperative for organelle-selective delivery strategies to optimize PDT outcomes. Consequently, advancing targeted modification approaches for photosensitizer localization represents a pivotal research direction in PDT development.

Beyond photosensitizer localization, tissue oxygen concentration serves as a critical determinant of PDT efficacy, demonstrating dose-dependent sensitivity.14 Progressive hypoxia in advanced tumors substantially impairs PDT outcomes by limiting oxygen-dependent ROS generation, thereby promoting therapeutic resistance and tumor recurrence.15 To address hypoxia-related limitations in PDT, multiple strategies have been investigated over recent decades. Among these, combination therapies with PDT have shown additive or synergistic effects, significantly improving anti-tumor efficacy while reducing therapeutic resistance.16,17 A wide range of clinical and preclinical therapeutic modalities—including chemotherapy, immunotherapy, photothermal therapy, antiangiogenic therapy, and gas therapy—have been explored in combination with PDT.18–23 Notably, gas therapy has gained increasing attention for its high biocompatibility profile.24 It can achieve direct or indirect delivery of O2 to tumor tissues, thereby improving the microenvironment for enhanced PDT efficacy.25 Alternatively, nitric oxide (NO) may react with ROS to generate highly reactive peroxynitrites (ONOO−), amplifying oxidative stress and potentiating PDT effects.26 Notably, the therapeutic action of gaseous agents similarly relies on precise subcellular organelle targeting.27 This review systematically examines strategies and enhancement mechanisms for organelle-specific PDT delivery, explores advanced targeting methodologies in gas-PDT combination therapy, and discusses spatiotemporally controlled gas prodrugs integrated with subcellular targeting strategies to achieve precise therapeutic modulation.

Detailed search strategies are provided in the Supplementary Material.

Subcellular Organelle-Targeted Drug Delivery for PDT Mechanism of PDT

PDT functions through two distinct mechanisms (Figure 1). The first involves photosensitizer-mediated energy transfer to biomolecules, generating free radicals (eg, photosensitizer-substrate complexes and anion radicals) via electron/hydrogen abstraction, termed Type I PDT. The second mechanism entails direct energy transfer to molecular oxygen, producing cytotoxic singlet oxygen (1O2), which characterizes Type II PDT.28

Figure 1 The Mechanism of PDT.

Both pathways induce oxidative damage to essential cellular components—proteins, nucleic acids, and membranes—through ROS. This damage initiates diverse cell death modalities: apoptosis, autophagy, and inflammatory necrosis. Notably, necrotic processes disrupt tumor vasculature, leading to nutrient deprivation and impaired tumor survival.29,30

Concurrently, PDT activates innate immunity by releasing damage-associated molecular patterns, recruits immune cells, and stimulates antitumor immunity, albeit with potential immunosuppressive effects from immune cell apoptosis.31 Thus, PDT exerts antitumor effects via three synergistic pathways: direct tumor cell killing, vascular system destruction, and immunomodulation.32

Organelle-Targeted PDT

The elevated levels of ROS and 1O2 generated during PDT exert antitumor effects through oxidative damage to intracellular macromolecules, ultimately suppressing tumor growth and inducing cellular senescence or death.33 Notably, 1O2—the predominant cytotoxic oxidant produced by clinically approved photosensitizers—exhibits limited diffusion capacity (approximately 20 nm lifetime radius) with differential oxidative susceptibility across subcellular compartments.5,6 Consequently, the organelle-specific localization of photosensitizers critically determines both the spatial patterns of PDT-induced damage and the activation pathways of cell death.34–37 Mechanistic studies identify mitochondria, plasma membranes, endoplasmic reticulum, and lysosomes as primary targets for 1O2/ROS-mediated PDT cytotoxicity.7

Mitochondria

Mitochondria, as multifunctional organelles, regulate cellular energy production, redox homeostasis, and survival pathways, rendering them critical targets under oxidative stress.38 Their heightened sensitivity to 1O2 and ROS enables mitochondria-targeted PDT to selectively eradicate tumor cells while sparing normal tissues. As demonstrated in Figure 2, mitochondrial targeting significantly enhances PDT efficacy compared to non-targeted approaches.39,40 Photoactivation-induced 1O2 generation at mitochondrial sites triggers permeability transition pore opening, cytochrome C release, and subsequent caspase activation. This cascade suppresses cellular respiration, depletes adenosine triphosphate (ATP), and initiates rapid apoptosis or necrosis.39,40 Concurrently, 1O2/ROS-mediated lipid peroxidation disrupts mitochondrial membranes, promoting mitophagy and irreversible organelle damage.41,42

Figure 2 Modification and mechanism of Mitochondrial-targeting PDT. Blue background Orange circle: Mitochondrial-targeted photosensitizer. Red circle: Excited photosensitizer. Red solid arrow: Excitation light. Purple dashed arrow: Path. Targeting mitochondria methods: Photosensitizer itself has mitochondrial targeting ability. Positive charge modification. Increase of lipophilicity. Mitochondrial targeting peptides modification. Mitochondrial targeting protein/carrier modification.

Mitochondrial-targeted drug delivery exploits photosensitizer physicochemical properties.43 For instance, cationic photosensitizers like IR780 (an indocyanine derivative) accumulate in mitochondria via electrostatic attraction to the negatively charged mitochondrial matrix, synergized by the enhanced permeability and retention (EPR) effect for tumor targeting.44,45 Wang et al developed IR780-based nanoplatforms to overcome hypoxic tumor resistance through mitochondrial-specific PDT.46 In an advanced strategy, Sun et al engineered folic acid (FA)-conjugated bovine serum albumin nanoparticles (FA-BCNID-NPs) co-loaded with IR780 and 5-nitro-8-hydroxyquinoline (NQ)-Cu(II) complexes. This system achieved dual targeting (FA-mediated tumor homing and IR780-driven mitochondrial localization), lysosome escape, mitochondrial destruction via PDT/photothermal therapy (PTT), and inhibition of P-glycoprotein drug efflux, demonstrating potent antitumor synergy.47

Mitochondrial-targeted drug delivery can be achieved through multiple strategies, including charge modification (eg, guanidine, biguanide, or triphenyl phosphonium (TPP) conjugation) and lipophilicity enhancement.48,49 Sun et al developed guanidine-functionalized cyclometalated iridium(III) complexes for mitochondrial-specific imaging and PDT, demonstrating potent cytotoxicity across cancer cell types via ROS-mediated mitochondrial apoptosis pathways.50 Shen Jianliang’s team engineered biguanide-modified chitosan (Bi-Ch) to disrupt mitochondrial function at reduced dosages, simultaneously alleviating tumor hypoxia and downregulating multidrug resistance protein 1 (MDR-1) expression to inhibit drug efflux.51 Zhong et al designed TPP-conjugated 5-aminolevulinic acid (5-ALA) derivatives with butoxycarbonyl (Boc) modifications to enhance mitochondrial membrane penetration. Encapsulated within folic acid (FA)-functionalized bovine serum albumin nanoparticles, this Boc-ALA-TPP system promoted cellular internalization via FA receptors and induced mitochondrial oxidative phosphorylation collapse through localized ROS generation, ultimately improving PDT efficacy under hypoxic conditions.52

Mitochondrial-targeting peptides (MTPs), particularly mitochondrial-penetrating peptides (MPPs), offer an alternative strategy with high specificity, low toxicity, and synthetic accessibility.53–55 Positively charged recombinant peptides or engineered nanocages further expand this approach.56 For instance, Wang et al constructed tumor/mitochondria-dual-targeting nanocages by fusing LinTT1 peptide with human heavy-chain ferritin (HFtn) to encapsulate aggregation-induced emission luminogens (AIEgens). These AIEgen-loaded nanocages generated substantial intramitochondrial ROS, inducing mitochondrial dysfunction, apoptosis, and tumor growth suppression.57

However, current mitochondrial targeting strategies face limitations: cationic small molecules may exhibit poor biocompatibility due to excessive positive charges; peptides are prone to enzymatic degradation; and most approaches lack tissue specificity, necessitating supplementary targeting mechanisms.

Lysosomes

Lysosomes are heterogeneous, single-membrane-bound organelles exhibiting diversity in subcellular localization, morphology, dimensions, and enzymatic composition (containing ≥60 hydrolytic enzymes).58 As central regulators of cellular homeostasis, they orchestrate critical processes including apoptosis, autophagy, tumor metastasis, membrane repair, signal transduction, and differentiation.59,60 Notably, cancer cells exhibit elevated lysosomal abundance compared to normal cells,61 positioning lysosomal targeting as a strategic approach to enhance PDT efficacy.62,63 PDT-induced lysosomal rupture releases protons and hydrolases, triggering subcellular dysfunction and autolytic cell death,64,65 with lysosome-targeted PDT demonstrating superior photokilling efficiency.66 This provides a robust theoretical foundation for developing lysosome-directed phototherapy-gas therapy integrated systems.

Lysosome-targeted PDT is achieved primarily through two interconnected mechanisms (shown in Figure 3). The first leverages the inherent endolysosomal trafficking pathway, which serves as the primary route for intracellular transport of membrane-bound components. Nanotherapeutic systems, influenced by particle size and entry mechanisms (including pinocytosis, phagocytosis, and receptor-mediated endocytosis), are naturally directed toward lysosomal compartments. This intrinsic biological process explains why nanoparticles like polyethyleneimine-modified PEGylated nanographene (PPG-Ce6) loaded with chlorin e6 (Ce6) demonstrate enhanced lysosomal accumulation. Subsequent modifications with tumor-targeting ligands such as folic acid further improve cellular internalization and tumor-specific delivery.67–73

Figure 3 Modification and mechanism of Lysosome-targeting PDT. Pink circle: Lysosome-targeting photosensitizer. Red circle: Excited photosensitizer. Purple dashed arrow: Path. Black irregular round shape: Endsome. Brown irregular circle: Lysosome. Black dashed arrow: another path of targeting Lysosome methods (protonation of basic photosensitizers drives their selective accumulation): Lipophilic amine modification (such as: morpholine, tertiary amines, dimethylamine benzene). Increasion of lipophilicity.

The second mechanism capitalizes on lysosomal acidity (pH 4.5–5.5), where protonation of basic photosensitizers drives their selective accumulation. Structural modifications with lipophilic amines—including morpholine derivatives, tertiary amines, and dimethylaminobenzene—enhance lysosomal targeting efficiency. Morpholine-functionalized agents (eg, BODIPY, porphyrins) and engineered nanoparticles like DSPE@M-SiPc (morpholine-substituted silicon phthalocyanine encapsulated in 2-distearoyl-sn-glycero-3-phosphoethanolamine) exemplify this strategy, combining near-infrared imaging capabilities with potent PDT effects. Advanced systems such as diketopyrrolopyrrole (DPP)-based nanoparticles (DPP-NF NPs) further integrate NO donors, achieving lysosomal activation under acidic conditions to synergistically enhance 1O2 generation and photothermal therapy.6,74–77

Compared to lysosomes in normal cells, those in cancer cells exhibit significantly elevated viscosity.78 To exploit this property, Song et al developed a lysosome-targeted bifunctional fluorescent probe for simultaneous viscosity imaging-guided cancer diagnosis and dual-mode PDT. This probe employs a donor-π-acceptor (D-π-A) molecular architecture comprising triphenylamine (electron donor), thiophene (π-bridge), and benzothiazolium salt (electron acceptor). The benzothiazolium ammonium moiety enhances acidic lysosomal accumulation and enables targeted photosensitizer delivery, generating both Type I/II ROS under viscosity-activatable conditions.79

Lysosomal targeting represents a pivotal strategy for enhancing PDT efficacy, as the organelle’s acidic microenvironment and elevated viscosity provide critical parameters for photosensitizer/carrier design. While ROS-induced lysosomal membrane disruption permits subsequent photosensitizer redistribution to other organelles, this translocation occurs at markedly slower kinetics compared to mitochondrial redistribution processes.80

Nucleus

The nucleus, as the central organelle in eukaryotic cells, serves as the primary repository of genetic material and regulates critical cellular processes including proliferation, metabolism, and cell cycle progression. It also functions as the principal site of interaction for therapeutic agents such as chemotherapeutics, nucleic acids, free radicals, and hyperthermia-based treatments.81 Nuclear-targeted PDT demonstrates enhanced efficacy compared to conventional approaches due to direct ROS-mediated damage to nuclear DNA, RNA, and proteins. Furthermore, photoactivated sensitizers can inactivate DNA repair enzymes, arrest cell cycles, and stimulate antitumor immunity (as shown in Figure 4). However, achieving effective nuclear delivery faces substantial challenges: photosensitizers must bypass cytoplasmic barriers (eg, endosomal entrapment, lysosomal degradation) and overcome nucleocytoplasmic transport restrictions to accumulate at nuclear targets.82

Figure 4 Modification and mechanism of Nucleus-targeting PDT. Orange Circle: Nucleus-targeting photosensitizer. Red circle: Excited photosensitizer. Purple dashed arrow: Path. Methods for targeting the cell nucleus: Nuclear localization sequences (eg, SV40 T antigen, adenovirus, TAT peptide). Disruption of nuclear membrane integrity. Passive diffusion (drug diameter should be less than 9 nm or drug molecular weight should be less than 40 kDa).

Nuclear entry is governed by nuclear pore complexes (NPCs)—bidirectional channels spanning the double-layered nuclear envelope. With a central pore diameter of ~70 nm, nuclear pore complexes permit passive diffusion of molecules <9 nm (eg, small gold nanoparticles) or <60 kDa (eg, doxorubicin, paclitaxel).83 For photosensitizers, Liu et al developed an aggregation-induced emission (AIE)-active nuclear-targeted agent (MeTPAE) that inhibits histone deacetylases and induces telomeric DNA damage via precision PDT.84 Light irradiation can also trigger nuclear translocation of certain photosensitizers. Anionic TPPS4 redistributes from lysosomes to nuclei in proliferating cells, a process dependent on cell-cycle status.85,86 Cationic pyridinium zinc phthalocyanine selectively accumulates in nucleoli post-irradiation, while anionic sulfonated and glycine-conjugated Zn-phthalocyanines migrate from lysosomes to nuclei under low-dose irradiation.87

Given the rapid systemic clearance and limited efficiency of passive nuclear transport, active nuclear-targeted drug delivery systems utilizing nuclear localization sequences (NLS) — such as SV40 T antigen, adenovirus-derived peptides, and TAT transduction domains — demonstrate enhanced therapeutic potential.81 Li et al engineered a nucleus-targeting system (PHSA-ICG-TAT) by conjugating TAT peptides with polyethylene glycol 4000 (PEG4000) and human serum albumin (HSA), enhancing both the aqueous solubility of indocyanine green (ICG) and its nuclear localization efficiency. This TAT-functionalized system exhibited superior cytotoxicity compared to free ICG, with combined DNA damage and localized hyperthermia significantly amplifying PDT/photothermal therapy (PTT) efficacy.88 Although amino-rich cationic polymers enable nuclear targeting, their clinical translation is constrained by systemic toxicity.89 Alternative strategies include lectin-mediated glycosylation-dependent nuclear entry, where carbohydrate-protein interactions facilitate nuclear translocation of glycosylated cargos.90

Notably, nuclear membrane destabilization offers another targeting avenue. Wu et al developed polyamine-functionalized polyhedral oligomeric silsesquioxane (POSS) nanoparticles (PPR NPs) incorporating polyethylene glycol (PEG) chains and the photosensitizer rose bengal (RB). Light-induced 1O2 generation first disrupts lysosomal membranes for nanoparticle release. Subsequent irradiation triggers nuclear membrane lipid peroxidation, enabling active nuclear penetration through envelope destabilization, thereby enhancing intranuclear therapeutic accumulation.91

Endoplasmic Reticulum

The endoplasmic reticulum (ER), the largest intracellular organelle, functions as the primary site for protein synthesis, folding, and transport; lipid and steroid biosynthesis; carbohydrate metabolism; and calcium ion storage.92 In cancer cells, excessive protein synthesis often leads to unfolded protein accumulation—a hallmark of malignancy—triggering ER stress-mediated apoptotic pathways.93 During ER-targeted PDT, generated ROS impair ER protein-folding capacity, inducing pathological ER stress that activates pro-apoptotic signaling cascades such as the PERK-eIF2α-ATF4-CHOP axis, ultimately causing cell death.94,95 Under severe ER stress caused by high ROS levels, pro-survival autophagy transitions to autophagy-dependent cell death, a process requiring high-dose PDT.96,97

ER-targeted PDT also induces immunogenic cell death (ICD) through ROS-driven ER stress. This mechanism promotes calreticulin (CRT) translocation to the cell surface, where it acts as an “eat-me” signal by activating dendritic cell maturation, thereby enhancing antitumor immunity.98 Furthermore, ER photodamage initiates paraptosis—a non-apoptotic death mode characterized by cytoplasmic vacuolization and loss of cellular integrity.99 Additionally, ER-generated ROS exacerbate ferroptosis via lipid peroxidation, leveraging the ER’s central role in lipid biosynthesis.100 Collectively, ER stress modulation through PDT enables multimodal tumor suppression via apoptosis, autophagy, immunogenic cell death, paraptosis, and ferroptosis (Figure 5).

Figure 5 Modification and mechanism of ER-targeting PDT. Orange Circle: ER-targeting photosensitizer. Red circle: Excited photosensitizer. Red irregular firework-shaped figure: ROS. Purple dashed arrow: Path. Black horizontal line: Induction, indicating that the above four situations all lead to cell death. Endoplasmic reticulum targeting methods: Targeted nanocarriers (receptors, lipophilic membranes, metabolic enzymes, etc.) for endoplasmic reticulum via small molecule modifications. Modification of endoplasmic reticulum targeting peptides (PA, KDEL, FEHDEL, etc). Modification of endoplasmic reticulum targeting ligands (sulfonylethylenediamine, sulfonamide ligands, etc).

ER-targeted delivery strategies comprise three principal approaches: small molecule-modified nanocarriers utilizing ligand-receptor binding, lipophilic modifications, or metabolic enzyme targeting; peptide-based systems incorporating PA, KDEL, and FEHDEL motifs; and engineered ER ligands such as tosyl ethylenediamine and sulfonamide derivatives.101

Guo et al developed NBS-ER, incorporating p-methylbenzenesulfonamide for ER targeting, which demonstrated reduced dark toxicity and enhanced ER localization compared to the non-targeted NBS-NH2 control. Post-PDT analysis revealed downregulation of ER stress sensors (ATF6, IRE1α), suppression of anti-apoptotic Bcl-xL, and activation of caspase-9 cleavage, confirming ROS-mediated ER stress apoptosis.102 Similarly, You et al designed FAL-ICG-HAuNS nanosystems combining pardaxin (FAL) peptides, indocyanine green (ICG), hollow gold nanospheres, and hemoglobin-loaded liposomes to alleviate hypoxia. This platform induced significant CHOP upregulation and caspase-3 activation under near-infrared irradiation, demonstrating ER stress-driven mitochondrial apoptosis and synergistic PDT/photothermal efficacy.98 Moreover, ER-targeted PDT effectively promoted calreticulin (CRT) exposure—a marker of immunogenic cell death—thereby enhancing dendritic cell maturation, CD8+ T cell proliferation, cytotoxic cytokine secretion, and activating systemic immune responses.

Lipid Droplet

Ferroptosis, an iron-dependent programmed cell death mechanism, is regulated by ferrous ions (Fe2⁺) that catalyze the Fenton reaction—a process converting excess hydrogen peroxide (H2O2) into hydroxyl radicals (·OH) and oxygen (O2) within tumor microenvironments.102 These hydroxyl radicals initiate free radical chain reactions, driving lipid peroxide accumulation in cellular membranes and triggering phospholipid damage that culminates in ferroptosis. Concurrently, depletion of intracellular glutathione (GSH) and inhibition of glutathione peroxidase 4 (GPX4) lead to the production of cytotoxic phospholipid hydroperoxides (PLOOH), causing membrane rupture and irreversible cell death.103

Emerging evidence demonstrates that PDT induces ferroptosis through lethal lipid ROS accumulation.104 Lipid ROS-mediated ferroptosis can originate at cellular membranes or subcellular organelles membranes (endoplasmic reticulum, mitochondria, lysosomes). However, rapid necrosis or apoptosis caused by acute ROS oxidation at these sites often overshadows the slower ferroptotic process.105

Lipid droplets (LDs), highly dynamic organelles serving as lipid storage hubs, play multifaceted roles in membrane synthesis, energy homeostasis, vesicular trafficking, and proteostasis.106,107 Crucially, LDs regulate ferroptosis by enabling photoinduced oxidation of polyunsaturated fatty acid phospholipids (PUFA-PLs) to peroxidized derivatives, establishing them as promising PDT targets for ferroptosis induction.108,109

LDs are structurally composed of a neutral lipid core (primarily triacylglycerols and sterol esters) surrounded by a phospholipid monolayer embedded with regulatory proteins such as perilipins.108,109 The phospholipid hydrophilic heads face the cytosol, while their hydrophobic alkyl chains anchor into the neutral lipid core (Figure 6). This architecture enables hydrophobic interactions as a key mechanism for LD targeting. Fluorescent precursors with tailored hydrophobicity—including BODIPY, coumarin, and benzoxadiazole derivatives—demonstrate inherent LD affinity.110,111 Structural modifications such as long alkyl chain incorporation, styryl additions, or fluorine substitutions further enhance targeting specificity.

Figure 6 Modification and mechanism of LDs-targeting PDT. Deep purple circle: LDs targeted photosensitizer. Red circle: Excited photosensitizer. Purple dashed arrow: Path. Black irregular circle: Lipid Droplets. LDs targeting methods: Targeted modification through hydrophobic interactions. Targeted modification by regulating intramolecular hydrogen bonds. Targeted based on aggregation-induced emission materials. Solid black arrows and corresponding text: LDs regulate ferroptosis by photoinducing the oxidation of polyunsaturated fatty acid phospholipids (PUFA-PLs) into peroxidation derivatives. Meanwhile, the consumption of glutathione (GSH) and the inhibition of glutathione peroxidase 4 (GPX4) lead to the production of cytotoxic lipid hydroperoxides (PLOOH), resulting in membrane rupture and irreversible cell death.

For instance, Peng et al engineered an LD-targeting photosensitizer (BODSel) by integrating a selenomorpholine group into the BODIPY core. This modification optimized lipophilicity-hydrophilicity balance for improved LD localization, while iodine atom incorporation via the heavy atom effect enhanced intersystem crossing efficiency and singlet oxygen generation. The resulting BODSel exhibited low dark toxicity, high biocompatibility, and potent LD-specific PDT efficacy.112

Similarly, Wang et al developed sulfur-substituted coumarin photosensitizers through π-conjugation extension and donor-acceptor optimization. Replacing carbonyl oxygen atoms with sulfur not only amplified intersystem crossing and ROS production but also conferred LD-targeting capability.113

In addition, decreasing the interaction of photosensitizers with water through intramolecular hydrogen bonds (H-bonds) can improve the LD-targeting ability. Peng et al designed the first lipid droplet-targeting type I photosensitizer (MNBS) by substituting the benzophenothiazine structure with a morpholine group in a donor-acceptor (D-A) system to enhance superoxide anion (O2−) production.104 The incorporation of morpholine not only increased hydrophobicity but also improved the H-aggregation tendency, thereby dispersing molecular electrostatic distribution. MNBS accumulated in LDs and induced ferroptosis-mediated PDT, achieving highly efficient antitumor effects under both hypoxic and normoxic conditions.

Novel photosensitizers with AIE features have been designed for LD targeting. For example, Tang’s group synthesized an LD-targeting AIE luminogen (AIEgen) to induce adipocyte apoptosis via type I PDT. They further developed biomimetic AIE photosensitizers (DC@AIEdots) by coating dendritic cell membranes onto nano-aggregated AIEgens, enabling LD targeting and activating photodynamic immunotherapy.112 Zhang et al synthesized an LD-targeting fluorescent material with AIE properties for integrating cancer diagnosis (via LD visualization) and treatment (via apoptosis induction through PDT).113 Hua et al designed three D-A-structured dihydrodibenzo[a,c]phenazine (DHP)-based photosensitizers (DP-CNPY, SMP-CNPY, and DMP-CNPY) by introducing methyl groups into the DHP donor and 2-(pyridin-4-yl)acetonitrile as a strong electron acceptor.114 Among these, SMP-CNPY exhibited LD-targeting ability and strong ROS production, thereby enhancing PDT efficacy under hypoxia.

Li et al reported a series of AIE-active photosensitizers (AIE-Cbz-LD-Cn, n = 1, 3, 5, 7, OMe) through conjugation of quinoline-malononitrile (QM) with carbazole.115 These compounds (AIE-Cbz-LD-C3, C5, and C7) demonstrated LD-targeting specificity, with AIE-Cbz-LD-C7 further inducing lipophagy and ferroptosis in live cells. Liu et al synthesized triphenylamine-based AIE molecules with a D1-D2-π-A structure and intramolecular charge transfer (ICT) properties.116 Modification of D2 with functional groups (phenyl, thiophene, and furan) allowed tuning of donor-acceptor interactions. All molecules displayed high LD-targeting capability and excellent ROS generation efficiency, enhancing PDT within LDs.117

Cell Membrane

The cell membrane, as the primary protective barrier of cells, isolates intracellular components from the external environment and maintains cellular integrity and metabolism. Additionally, the cell membrane regulates substance exchange through pinocytosis, phagocytosis, exocytosis, and protein secretion/transport.7 Intracellular drug accumulation critically depends on cell membrane fluidity, which determines therapeutic efficacy.118 Membrane-targeted drugs bypass transmembrane transport, thereby avoiding drug efflux.119 Furthermore, membrane homeostasis is essential for regulating cellular processes. Disruption or damage to membrane function triggers signaling pathways that induce cell death (eg, apoptosis, necrosis, autophagy, and ferroptosis) (shown in Figure 7). Membrane rupture rapidly releases cellular contents, provoking immunogenic cell death to enhance antitumor immunity and therapeutic outcomes.120 Thus, the cell membrane is a promising target for PDT.

Figure 7 Modification and mechanism of cell membrane-targeting PDT. Orange Circle: Cell membrane targeted photosensitizer. Red circle: Excited photosensitizer. Purple dashed arrow: Path. Targeted strategies for cell membranes: Positive charge modification. Hydrophobic interactions with cell membrane. Lipid analog design: such as Fluorescence strategy. Biospecific recognition strategy: receptor ligand targeting, enzyme driven membrane localization. Amino acid and glycosylation modification strategies. New materials and intelligent design strategies: AIEgens (aggregation induced luminescent materials) and others.

The amphiphilic cell membrane, composed of phospholipids, glycoproteins, and glycolipids arranged in a bilayer, carries abundant negative charges.121 Membrane-targeted PDT can exploit these amphiphilic and electrostatic properties. Sun et al designed two molecular rotor-based self-reporting photosensitizers to disrupt membrane integrity via electrostatic/hydrophobic interactions with phospholipid bilayers during PDT.122 Zhang et al developed a charge-reversible self-delivery chimeric peptide (C16-PRP-DMA) for sustained membrane targeting. This peptide integrates four functional segments: palmitic acid as a lipophilic moiety enabling membrane insertion and self-assembly, a DMA-modified tetra-peptide (RRKK) enhancing membrane affinity via electrostatic interactions while reducing nonspecific in vivo adsorption, polyethylene glycol (PEG) improving biocompatibility and prolonging blood circulation half-life, and protoporphyrin IX (PPIX) generating ROS to directly disrupt membranes and induce rapid necrosis. This design significantly enhanced PDT efficacy.123 Zhang’s group further engineered a self-delivery chimeric peptide for combined low-temperature photothermal therapy (LTPTT) and PDT. Localized ROS generation and mild heating (<45 °C) at the membrane induced rupture and activated antitumor immunity, suppressing metastasis.124

Novel AIEgens-based membrane-targeted photosensitizers have also emerged. Tang et al synthesized a near-infrared (NIR)-emissive AIE photosensitizer (TBMPEI) by cationizing pyridine units. TBMPEI selectively accumulates on membranes, inducing necroptosis via membrane rupture and DNA degradation upon irradiation.125 Qi et al designed AIE-active photosensitizers (DCTPys) that disrupt membranes under mild conditions to trigger necrosis.126

Lipid mimics designed for cell membrane targeting often lack cancer cell selectivity, leading to unintended damage to normal cells. Perfluorocarbons are known to accumulate in the central lipid layer of cell membranes. Based on this property, Chen et al proposed a fluorination strategy for plasma membrane-targeted PDT in cancer cells, disrupting membrane integrity and inducing pyroptosis to enhance therapeutic efficacy.127 Awazu et al employed replication-deficient hemagglutinating virus of Japan (HVJ; Sendai virus) envelopes (HVJ-E) as membrane-targeted drug carriers to deliver photosensitizers, achieving cancer-selective apoptosis and antitumor immunity for amplified PDT effects.128

Specific membrane targeting can be achieved through protein receptors, peptides, amino acids, or glycosylated chlorin compounds. For instance, folate receptors are overexpressed in approximately 40% of human cancers, making them viable targets for folate-modified PDT systems.129 Zhang’s group developed a protein farnesyltransferase (PFTase)-driven plasma membrane (PM)-targeted chimeric peptide based on a K-Ras-derived sequence (KKKKKKSKTKC-OMe). This system initiates lipid peroxidation and membrane rupture at nanomolar concentrations during PDT. Damaged membranes further release damage-associated molecular patterns, activating antitumor immunity to suppress metastatic tumors.130

Simple amino acid modifications can enhance noncovalent interactions between photosensitizers and membranes. Li’s team conjugated protoporphyrin IX (PPIX) to the ε-amine of lysine and modified it with arginine or glutamic acid, enabling single-arginine-level membrane targeting and light-induced membrane disruption.131 Drain et al demonstrated that glycosylated chlorin compounds localized transiently at membranes activate necrosis upon irradiation, despite initial targeting via surface saccharide receptors followed by rapid internalization into the endoplasmic reticulum.11

Organelle-Targeted Gas-Photodynamic Combination Therapy

Despite continuous technological advancements, particularly in subcellular organelle-targeted photosensitizer delivery systems, the efficacy of PDT has been substantially improved. However, challenges such as limited light penetration depth (<1 cm), tumor hypoxia, and ROS quenching have hindered the clinical translation and broad application of PDT.132 To address these limitations, synergistic strategies combining PDT with other modalities (eg, chemotherapy, radiotherapy, photothermal therapy, immunotherapy, sonodynamic therapy, and gas therapy) have emerged as effective solutions.16,23,133,134 Among these, gas therapy utilizing gaseous signaling molecules—such as NO, carbon monoxide (CO), hydrogen (H2), and hydrogen sulfide (H2S)—demonstrates selective pro-apoptotic effects on cancer cells, modulates tumor vasculature, and protects normal tissues within specific concentration and temporal windows.135 Thus, the integration of PDT with gas therapy represents a promising strategy to enhance antitumor efficacy with significant developmental potential.

Gas Therapy and Mechanisms of Action

NO, CO, H2, and H2S are critical gaseous signaling molecules that regulate cellular biology and signaling pathways.136 At specific concentrations, these gases inhibit tumor migration, reverse the Warburg effect in cancer cells to modulate cell death, and exert anti-inflammatory effects.137 Their ultralow molecular weights enable diffusion into tumor interstitium and penetration across biological membranes, thereby targeting deep-seated cancer cells.138 The antitumor mechanisms of these gases are summarized in Table 1. Similar to PDT, gas therapy exhibits low systemic toxicity, minimal off-target effects on normal tissues, and reduced risk of drug resistance.139–141 As a minimally invasive and biocompatible strategy, gas therapy induces biochemical and physiological alterations in tumor microenvironments, achieving precise tumor suppression.

Table 1 Therapeutic Mechanisms of Different Gases

Gas-Photodynamic Combination Therapy and Mechanisms of Action

Although nanocarriers and targeted drug delivery systems enhance the efficacy of PDT, several inherent limitations remain unresolved. For instance, the limited penetration depth of excitation light persists despite optimization through near-infrared irradiation, fiber optics, or light-emitting diodes. Moreover, skin photosensitivity, thermal damage, and localized heating induced by prolonged light exposure cannot be fully addressed by current strategies. Combining PDT with other therapeutic modalities is critical to reduce photosensitizer toxicity and amplify photodynamic effects. Gas therapy synergizes profoundly with PDT through mechanisms such as ROS sensitization, tumor vascular disruption, and enhanced intracellular drug accumulation, as summarized in previous sections. These interactions establish a robust foundation for achieving therapeutic synergy. Consequently, integrating gas therapy with PDT holds significant potential in advancing antitumor treatment paradigms.

Six gaseous agents—oxygen (O2), CO, NO, hydrogen (H2), hydrogen sulfide (H2S), and sulfur dioxide (SO2)—are predominantly employed in gas-photodynamic combination therapy. The mechanisms of gas-PDT synergy include ROS amplification through oxidative stress enhancement, sensitization of drug-resistant cells, inhibition of malignant cell and tissue proliferation, anti-inflammatory modulation, and tumor microenvironment reprogramming, as illustrated in Figures 8 and 9. Subsequent sections critically evaluate gas-PDT combination strategies organized by these mechanistic categories.147–152

Figure 8 The possible main mechanism of Gas-PDT combination therapy.

Figure 9 The enhancement of Gas-PDT combination therapy. Red circle: Photosensitizers. Green circle: Gas prodrugs. Brown circle: Oxygen precursors. Black solid line arrow: Leads to. Blue solid line arrow: Transforms into. Yellow, purple, blue dashed line arrows: Lead to. Black horizontal line: Here refers to Autophagy, Necrosis, Apoptosis ultimately leading to tumor cell death.

Promotes Oxidative Stress O2

The generation of 1O2 or ROS to induce cancer cell death represents a core mechanism of PDT. However, the rapid proliferation of tumor cells creates an imbalance between oxygen supply and consumption, resulting in tumor hypoxia.153 Hypoxia not only restricts ROS production during PDT but also promotes tumor progression, metastasis, and radioresistance. Consequently, tumor reoxygenation strategies are essential to overcome this limitation. Current approaches include endogenous oxygen supplementation; exogenous oxygen delivery; and in situ oxygen regeneration via catalytic reactions.154–158

Endogenous oxygen supplementation modulates the hypoxic microenvironment through tumor vascular normalization, degradation of dense tumor stroma, or hyperthermia-induced oxygen release. Exogenous delivery relies on oxygen nanocarriers such as perfluorocarbons (PFCs), erythrocytes, hemoglobin (Hb), and metal-organic frameworks (MOFs).

Red blood cells (RBCs), the primary oxygen carriers in mammals, exhibit high oxygen-binding capacity, biocompatibility, and immunomodulatory functions. Zhang et al engineered an RBC-derived vehicle co-loaded with hemoglobin, chlorin e6, and sorafenib, which enhances PDT efficacy by boosting intratumoral oxygen/iron levels and triggering ferroptosis.159 To address hemoglobin’s auto-oxidation and nephrotoxicity risks, Wang et al developed a chemiluminescence-driven PDT system using hemoglobin-conjugated conjugated polymer nanoparticles (Hb-CPNs). This system generates ROS via luminol-catalyzed chemiluminescence absorbed by CPNs, eliminating the need for external light. Encapsulation of Hb-CPNs within biomimetic liposomal polymers further improves hemoglobin stability and oxygen-loading capacity, achieving superior therapeutic outcomes.160

PFCs are FDA-approved oxygen carriers that physically dissolve substantial oxygen via weak van der Waals interactions and passively diffuse it into hypoxic tumors. Compared to hemoglobin (Hb), PFCs exhibit twice the oxygen capacity, unique electronic structures, high gas solubility, and chemical inertness, making them clinically validated for oxygen delivery.161–164 Current PFC-based systems include liposomes, nanoparticles, and micelles. Wu et al engineered core-shell nanoparticles emulsified with perfluorotributylamine (PFTBA) and Ce6, augmenting PDT efficacy by leveraging PFCs to capture oxygen from photosynthetic chlorella gel under light.165 Studies confirm that PFC content in nanoplatforms directly correlates with 1O2 yield. Hu et al developed oxygen self-enriching PDT (Oxy-PDT) using PFC nanodroplets loaded with photosensitizers, demonstrating amplified ROS generation and cytotoxicity.166 However, PFCs’ side effects—arterial hypotension, pulmonary injury, thrombocytopenia, and flu-like symptoms—remain understudied.

MOFs, highly ordered porous structures formed by metal ions and organic ligands, enhance photosensitizer delivery via high porosity, surface area, and EPR effects.167,168 Their nanoscale tunability further optimizes tumor accumulation. Zhao et al integrated Ce6 with Zn-MOFs, which disrupt bacterial membranes, amplify ROS production, and enable pH-responsive Ce6 release.169 Despite these advantages, challenges persist in preventing premature drug leakage and improving tumor-targeted oxygen delivery.

In situ oxygen regeneration employs catalytic nanoparticles (eg, CaO2, Cu2O, Pt, Fe) or enzymes to decompose tumor H2O2/H2O into oxygen. Manganese dioxide (MnO2) is widely utilized; Cong et al encapsulated MnO2, paclitaxel, and Ce6 in liposomes, where MnO2 releases oxygen to enhance PDT while synergizing with chemotherapy.170 Catalase (CAT)-based systems, such as Zhang’s Nb@HCC platform (CAT, Ce6, human serum albumin, and nanobody), boost ROS under 660 nm light for ovarian cancer therapy.171 However, CAT’s high molecular weight and instability limit tumor penetration, whereas metal nanoparticles (eg, MnO2) offer cost-effectiveness and stability despite potential cytotoxicity risks.172–178

NO

NO, an endogenous multifunctional signaling molecule, regulates vascular relaxation, neuromodulation, bone metabolism, and tumor progression. Its synthesis originates from precursors such as S-nitrosothiols (SNOs), L-arginine (L-Arg), nitroglycerin, N-nitrosamines, nitro compounds, diazeniumdiolates (NONOates), and metal nitroso complexes. Among these, L-Arg—a natural amino acid—exhibits superior biocompatibility and efficient NO generation via inducible nitric oxide synthase (iNOS) or ROS-driven oxidation, despite challenges like premature release and toxic byproduct formation compared to NONOates and SNOs.179

NO potentiates oxidative stress to amplify antitumor therapy through three interconnected mechanisms: competitive binding to mitochondrial complex IV reduces oxygen consumption and enhances PDT efficacy; reaction with superoxide radicals forms cytotoxic peroxynitrite (ONOO−); and vasodilation-mediated hypoxia alleviation synergizes with PDT. For instance, Li et al combined perfluorocarbon (PFC) nanoliposomes (FI@Lip) with S-nitrosated human serum albumin (HSA-SNO), where HSA-SNO releases NO to inhibit mitochondrial respiration and deplete GSH, restoring 1O2 levels, while PFCs directly deliver oxygen to boost 1O2 generation, collectively enhancing tumor suppression.180 Similarly, Zhang et al designed an arginine-loaded gelatin-coated PCN-224 system (Arg-PCN@Gel), leveraging arginine-derived NO to produce ONOO−, disrupt biofilms, and enhance ROS penetration, achieving targeted cascade PDT against methicillin-resistant Staphylococcus aureus (MRSA).181 Despite these advances, NO-PDT integration faces limitations: NO’s short half-life (<5 s) and restricted diffusion range (20–160 μm) necessitate subcellular targeting strategies (eg, mitochondrial or nuclear localization) to optimize therapeutic precision.

CO

Similar to NO, CO modulates tumor viability by amplifying oxidative stress. CO inhibits cytochrome c oxidase and mitochondrial electron transport, thereby disrupting tumor cell survival and protein synthesis. The therapeutic effects of CO exhibit dual functionality: in normal cells, CO mildly enhances mitochondrial ROS production; in cancer cells, it induces apoptosis through oxygen concentration elevation, excessive ROS generation, and mitochondrial dysfunction.

While low-dose CO inhalation is a potential delivery method, its systemic toxicity and poor targeting limit clinical utility. Current strategies prioritize nanocarrier-mediated delivery of CO-releasing molecules (CORMs), primarily transition metal carbonyl compounds that release CO under exogenous (light, heat, magnetic fields) or endogenous (ROS, acidic pH) stimuli.182–188 For example, Li et al engineered human serum albumin (HSA)-stabilized MnO2 nanoparticles (HIM-MnO2) co-loaded with the CORM MnCO and mitochondrial-targeting photosensitizer IR780. The nanoparticles’ photothermal effect triggers CO release, while IR780’s mitochondrial localization directs CO to impair respiratory function, synergistically inducing tumor cell death.189

H2S

Similar to CO, H2S exhibits biphasic effects. At physiological concentrations, it demonstrates antioxidant, angiogenic, vasodilatory, anti-inflammatory, and anti-apoptotic properties. Conversely, high exogenous H2S concentrations or prolonged exposure arrest the cell cycle, induce intracellular acidification, dysregulate signaling pathways, and trigger mitochondrial dysfunction and apoptosis.135,150

The concentration-dependent effects of H2S lead to variable PDT synergies. Yang et al synthesized FTEP-TBFc NPs via alkyne-azide click chemistry, integrating NIR-II imaging with HPTT/CDT/PDT/GT. Under 808 nm light, the NPs generate fluorescence, heat, and 1O2. Tumor GSH cleaves trisulfide bonds, releasing H2S to suppress COX IV and HSP70, enhancing apoptosis. H2S also inhibits catalase, reducing H2O2 consumption and increasing intratumoral O2 to amplify PDT.190

Ye et al designed acid-responsive 1-JK-PS-FA NPs using JK-2 (H2S donor), a reducible chromophore, NIR775, and FA. FA-mediated targeting and EPR enhance tumor accumulation. The acidic microenvironment triggers JK-2 to release H2S, promoting angiogenesis, alleviating hypoxia, and improving NP uptake. This reprograms the tumor microenvironment, boosting PDT-induced apoptosis and immune activation while suppressing metastasis.191

Similarly, Wang et al developed a novel nanoplatform that enables cascade photo-immunotherapy through tumor-specific responsive reactions.192 Its tetrasulfide-bridged silica-based core releases an AIEgen and manganese carbonyl (MnCO) upon glutathione (GSH) triggering, simultaneously generating hydrogen sulfide (H2S). Under near-infrared light irradiation, the AIEgen mediates photodynamic therapy and triggers the decomposition of MnCO, releasing CO and manganese ions (Mn2⁺). The synergistic action of H2S and CO disrupts mitochondrial integrity, leading to mitochondrial DNA (mtDNA) leakage into the cytoplasm. Meanwhile, Mn2⁺ enhances the activation of the cGAS–STING pathway, promoting type I interferon production and potentiating antitumor immunity.

However, Saenz et al reported H2S-mediated PDT resistance. NaHS pretreatment reduces 1O2 generation via upregulated GSH and catalase, while inhibiting protoporphyrin IX synthesis in ALA-PDT, thereby attenuating cytotoxicity.193

SO2

Endogenous SO2 in a hyperoxidized state induces cellular oxidative stress and damage due to its tetravalent sulfur properties.194 This occurs primarily through two pathways: acting as an oxidant to elevate intratumoral GSH levels or generating free radicals within biological systems.195 Zhang et al developed an injectable hydrogel (TBH) co-delivering the AIE-based theranostic agent TOCAc and the SO2 donor benzothiazole sulfinate (BTS). The TBH gel’s photothermal effect under irradiation triggers localized warming, dissolving and releasing BTS and TOCAc aggregates. Cationic TOCAc selectively targets mitochondrial membranes of cancer stem/non-stem cells, enabling simultaneous type I PDT and photothermal therapy (PTT). Concurrent SO2 release depletes intratumoral GSH, amplifying PDT efficacy and eradicating cancer stem cells to prevent recurrence.140

Similarly, Wang et al synthesized CyI-DNBS by conjugating 2,4-dinitrobenzenesulfonyl chloride (DNBS) to the photosensitizer Cyl-OH. GSH-mediated cleavage releases both SO2 and Cyl-OH, activating PDT under red light.196 Li et al engineered gold nanorod-polydopamine nanocapsules (GPBRs) doped with BTS. Near-infrared irradiation induces photothermal effects, while the acidic tumor microenvironment and hyperthermia synergistically promote SO2 diffusion. SO2-derived ROS (SO3·−) upregulates pro-apoptotic proteins (p53, Bax, caspase-3) and suppresses anti-apoptotic Bcl-2, driving apoptosis.137

Sensitize Drug-Resistant Cells

Multidrug resistance (MDR) remains a critical challenge in oncology, contributing significantly to therapeutic failure and tumor recurrence.197 MDR mechanisms are multifactorial and heterogeneous, encompassing upregulated drug efflux (eg, P-gp overexpression), impaired drug influx, apoptosis resistance, enhanced DNA repair, drug target modifications, and enzymatic detoxification. Current strategies to overcome MDR include tumor vascular normalization, stromal cell reprogramming, DNA repair inhibition, and intratumoral drug concentration escalation.198 Integrating gas therapy with structurally optimized nanomedicine offers a promising approach to reverse MDR synergistically.

Inhibits Drug Excretion

Drug efflux is a major cont