Naphthalimide‐Based Type‐I Nano‐Photosensitizers for Enhanced Antitumor Photodynamic Therapy: H₂S Synergistically Regulates PeT and Self‐Assembly

Abstract

Photodynamic therapy (PDT) relies on a combination of light and photosensitizers (PSs) to achieve local control over cancerous lesions. However, it is subject to limitations, including tumor hypoxia, low tumor targeting, off‐target phototoxicity, and always‐on fluorescence. Here, we propose a design strategy for activated nano‐PSs (N‐PSs) to simultaneously overcome the limitations of PDT, wherein photoinduced electron transfer (PeT) is coupled with an endogenous H₂S‐regulated self‐association process to promote Type‐I photochemical reactions. Using theoretical calculations, spectral analysis, and microscopic imaging, we verified the generation of self‐assembly and occurrence of PeT. And it was also shown that H₂S could synergistically inhibit the PeT and self‐assembly, reflecting by a 21‐fold increase in fluorescence intensity at 635 nm and 35‐fold enhancement of the Type‐I photochemical reaction as inferred from O₂ ⁻ generation. Moreover, the most promising self‐assembled N‐PS, Ts3‐ONB, was found to almost completely inhibit tumor growth in mice under two‐photon excitation through the synergistic regulation of PeT and self‐assembly by endogenous H₂S (V _{14 days} ^{Ts3‐ONB + Light group}/V _{14 days} ^{Control group} ≈ 0.02). As such, the synergistic combination of PeT and self‐assembly is an effective design strategy for developing advanced N‐PSs that can address some current PDT limitations.

Here, we introduce a design strategy for activated photosensitizers (PSs) wherein photoinduced electron transfer (PeT) is combined with a process of self‐assembly that can be regulated by endogenous H₂S—a tumor‐associated biomarker to favor Type‐I photochemical reactions.

Introduction

Photodynamic therapy (PDT) is a non‐invasive and highly effective cancer treatment strategy that is seeing increasing use in the field of clinical oncology.^{[ 1} , ² , ^{3 ]} Photosensitizers (PSs) serve as the vital element in this tactic, which can trigger two different limiting photochemical reactions: Type‐I (electron transfer, less oxygen‐dependent) and Type‐II (energy transfer, largely oxygen‐dependent)^{[ 4} , ⁵ , ⁶ , ^{7 ]} upon appropriate photo‐irradiation conditions, ultimately resulting in irreversible oxidative damage to tumors. However, traditional PSs face limitations that hinder their further clinical application. A key challenge is low tumor targeting, which leads to off‐target accumulation and inevitable damage to healthy tissues. This issue is compounded by the continuous photosensitive activity, resulting in prolonged skin photosensitivity and systemic toxicity even after treatment. Furthermore, the always‐on fluorescence characteristic of most PSs can limit the imaging guided placement of the photo‐excitation light source, thus reducing the potential efficacy of PDT.^{[ 8} , ⁹ , ^{10 ]} Most critically, traditional PSs have been designed to produce singlet oxygen efficiently through Type‐II photochemical reactions.^{[ 11 ]} The oxygen‐dependent nature of Type‐II photochemical reactions drastically diminishes reactive oxygen species (ROS) generation in the hypoxic microenvironment of solid tumor, rendering PDT efficiency greatly reducing in hypoxic condition.

The above issues in principle could be addressed through improved PSs design. One approach that is attracting attention involves the design of nano‐PSs (N‐PSs).^{[ 12} , ^{13 ]} N‐PSs, with their unique nanostructural advantages, show significant potential in the field of PDT: enhancing tumor targeting through size‐dependent EPR effects.^{[ 13 ]} For example, Peng et al. synthesized a case of N‐PS by encapsulating BODIPY derivatives with PEG, that can specifically target tumors and stay for an extremely long time, realizing the efficient tumor treatment effect of multiple phototherapy with a single injection.^{[ 14 ]} Since then, self‐assembly‐based N‐PSs have been reported to favor the precise spatiotemporal regulation of photosensitive activity by taking advantage of the dynamic responsiveness of self‐assembly processes (e.g., pH, thiols, enzymes‐triggered disassembly).^{[ 15 ]} For example, in 2020 Ye and coworkers reported tumor‐targeting and redox‐responsive nano‐assemblies that underwent disassembly to produce active PSs in the presence of glutathione (GSH).^{[ 16 ]}

Furthermore, in general, appropriate regulation of the electron transfer step can be used to improve the effectiveness of Type‐I PDT. One way to control electron transfer is through self‐assembly.^{[ 17 ]} For instance, in 2018 Yoon's group detailed a nanostructured phthalocyanine assembly, wherein regulated electron transfer could be exploited to achieve efficient ROS generation through a Type‐I mechanism.^{[ 18 ]} Also, electronic regulation can be effected in the the excited state via photoinduced electron transfer (PeT).^{[ 19 ]} This mechanism, in particular, has been exploited to prepare “off–on” PSs with improved PDT activity.^{[ 4} , ²⁰ , ^{21 ]} For example, Song's group used aryl nitro compounds to instigate d‐PeT (“d” represents “donor”) with fluorescein thereby developing nitroreductase‐activated photosensitive dyes for PDT‐based treatment of hypoxic tumors.^{[ 22 ]} Separately, Akkaya, et al. used a 2,4‐dinitrobenzene sulfonyl group to favor d‐PeT with a brominated BODIPY derivative thus obtaining a GSH‐activated BODIPY PS for the selective PDT of tumor cells.^{[ 23 ]}

The progress made to date notwithstanding, we considered it likely that the ability to control both PeT and self‐assembly would yield improved “off–on” N‐PSs. To test this hypothesis, we prepared a series of Type‐I naphthalimide‐based N‐PSs (Ts3‐ONB, Ts3‐ONBS, Ts4‐ONB and Ts4‐ONBS), whose self‐assembly is regulated by H₂S—a small molecule associated with the tumor environment (Figure 1).^{[ 24 ]} We verified the synergistic H₂S‐based regulation of PeT and self‐assembly within these nano‐self‐assemblies using inter alia theoretical calculations, spectral analyses, and microscopic imaging. As detailed below, these new N‐PSs exhibited excellent two‐photon activation^{[ 25} , ^{26 ]} with the resulting fluorescence, Type‐I photochemical reactivity, and tumor cell killing being switched via H₂S activation in solution, cells, and mice.

Figure 1.

Synergistic strategy that exploits H₂S‐based regulation of both PeT and self‐assembly. Molecular structures and proposed key chemical and cellular reaction process, including spontaneous nano‐self‐assembly and H₂S‐promoted disassembly.

Results and Discussion

Design and Synthesis of Naphthalimide‐Based Monomers and Nano‐Self‐Assemblies

To design self‐assembly‐based “turn on–turn off” N‐PSs with better performance, the selection of regulatory factors becomes the key point. There are various classical tumor‐associated molecules that can serve as regulatory factors.^{[ 27 ]} For this study, we selected H₂S as the key disassembly trigger since it is one of many important cancers‐associated active molecules, including colon, breast, and ovarian.^{[ 28 ]} With this choice made and considering that the typical characteristic of Type‐I PSs is electron transfer, strong electron‐withdrawing groups—2,4‐dinitrobenzene sulfonyl groups (DNBS) and 2,4‐dinitrophenyl (DNB) groups were introduced into the 3 or 4‐positions of a naphthalimide core for the design of naphthalimide‐based monomers. They were expected to be readily cleaved in the presence of H₂S, and to regulate the electron transfer ability of the entire molecular structure. The choice of the naphthalimide core reflects the fact that it can be excited by a long wavelength (e.g., two photon excitation, 800 nm) pulsed laser^{[ 26 ]} and was expected to act as a PeT donor to the electron deficient dinitroaryl units. Additionally, hoping for its conjugated structure can provide the possibility for the occurrence of self‐assembly^{[ 29 ]} and its easy modification allows the regulation of their self‐assembly behavior. Finally, but more importantly, functionalization to introduce the hydrophilic group N‐(2‐aminoethyl)‐4‐methylbenzenesulfonamide was then expected to provide amphiphilic systems that would easily self‐assembly to provide masked or “turned off” N‐PSs that would disassembly and be “turned on” in the presence of H₂S.

Synthetic details and characterization data for all four new constructs, naphthalimide‐based monomers (Ts3‐ONB, Ts3‐ONBS, Ts4‐ONB and Ts4‐ONBS), as well as controls Ts3‐OH and Ts4‐OH, are provided in the Supporting Information (Scheme S1 and Figures S1–S16). Naphthylimide bearing N‐(2‐aminoethyl)‐4‐methylbenzenesulfonamide group was prepared according to our previous work. To achieve its further functionalization, DNBS and DNB groups were introduced to the hydroxyl groups at the 3 or 4‐positions through nucleophilic substitution reactions, obtaining Ts3‐ONB (Yield 42%), Ts3‐ONBS (Yield 45%), Ts4‐ONB (Yield 42%) and Ts4‐ONBS (Yield 43%), respectively. These molecules served as monomers and were subjected to characterization of the material structure of N‐PSs formed based on self‐assembly behavior in a simulated intracellular environment (Figure 2).

Figure 2.

a) UV absorption spectra and the fluorescence emission spectra of Ts3‐ONB (20 µM) recorded in different solvents using quinine sulfate in 0.1 M sulfuric acid as a standard reference to verify the fluorescence intensity of Ts3‐ONB. b) Frontier molecular orbitals (HOMO and LUMO) of Ts3‐ONB and Ts3‐OH. c) SEM and TEM images of the self‐assembled form of Ts3‐ONB, Ts3‐ONBS, Ts4‐ONB and Ts4‐ONBS, scale bars: 500 and 200 nm; The size distribution of freshly prepared Ts3‐ONB, Ts3‐ONBS, Ts4‐ONB and Ts4‐ONBS (20 µM) nano‐self‐assemblies as determined by DLS at 25 °C (Dulbecco's Modified Eagle Medium, DMEM chosen to simulate the physiological environment). d) Electrostatic potential (ESP) mapping of Ts3‐OH monomers, Ts3‐ONB monomers, Ts3‐ONB monomers, Ts4‐OH monomers, Ts4‐ONB monomers, and Ts4‐ONBS monomers in their ground states (blue, electron poor; red, electron rich); Optimized dimer structures of Ts3‐OH, Ts3‐ONB, Ts3‐ONBS, Ts4‐OH, Ts4‐ONB, and Ts4‐ONBS, and the weak interactions between optimized dimer structures of Ts3‐OH, Ts3‐ONB, Ts3‐ONBS, Ts4‐OH, Ts4‐ONB, and Ts4‐ONBS.

Photophysical Properties of Naphthalimide‐Based Monomers and Nano‐Self‐Assemblies

The photophysical properties of the four naphthalimide‐based monomers of this study (Ts3‐ONB, Ts3‐ONBS, Ts4‐ONB, and Ts4‐ONBS) and the two control compounds (Ts3‐OH and Ts4‐OH) were assessed by means of various spectroscopies, microscopic imaging, and theoretical calculations (Tables S1–S6, Figures S17–S19). Within this set, the fluorescence of naphthalimide‐based monomer Ts3‐ONB is the most thoroughly quenched (Table S2), and it has the highest molar extinction coefficient, so naphthalimide‐based monomer Ts3‐ONB and its control compound (Ts3‐OH) were selected as representative examples. Both Ts3‐ONB and Ts3‐OH absorb light well at 333, 338, 340, and 343 nm (ε _DMSO ^Ts3‐ONB  = 16925 M⁻¹ cm⁻¹, ε _DMSO ^Ts3‐OH  = 8190 M⁻¹ cm⁻¹, ε _H2O ^Ts3‐ONB  = 19625 M⁻¹ cm⁻¹, ε _H2O ^Ts3‐OH  = 6135 M⁻¹ cm⁻¹, Tables S1 and S2).^{[ 26} , ^{30 ]} Furthermore, both compounds exhibit two‐photon absorption cross sections at 800 nm (λ _ex ^two‐photon), leading us to suggest that they could be excited effectively using a long wavelength pulsed laser (Figures S18 and S19).

The fluorescence emission of naphthalimide‐based monomer Ts3‐ONB is quenched in many solvents, including DMSO and H₂O (Figure 2a and Table S2). This and other fluorescence features of naphthalimide‐based monomer Ts3‐ONB were explored using quantitative theoretical calculations (density functional theory, DFT). As can be seen from an inspection of Figure 2b, the lowest unoccupied molecular orbital level (LUMO, −3.09 eV) of the dinitrobenzene moiety (the electron deficient H₂S‐cleavable group) lies between the highest occupied molecular orbital (HOMO)/LUMO level (−6.97 and −2.87 eV) of the relatively electron rich naphthalimide fluorophore. Therefore, following photoexcitation intramolecular electron transfer is favored, leading to quenching of the naphthalimide emission via a classic d‐PeT process.^{[ 19 ]} In contrast to the weak fluorescence of naphthalimide‐based monomer Ts3‐ONB, a strong fluorescence emission (quantum yield of 69.82 ± 0.77% in DMSO; Table S1) is seen for the control molecule Ts3‐OH. In this case, the electron distribution in both the HOMO and, after photo‐excitation, LUMO of naphthalimide‐based monomer Ts3‐OH is predominantly localized on the naphthalene ring moiety (Figure 2b) and no PeT effect is expected for Ts3‐OH.

The relatively weak fluorescence emission of naphthalimide‐based monomer Ts3‐ONB is further diminished as the H₂O content in the initial DMSO solution is increased, with the lowest recorded fluorescence quantum yields being 0.054 ± 0.002% (Figure S20 and Table S2). As the polarity is increased via H₂O addition, regular shaped nanoparticles are formed. Similar nanoparticles are seen in Dulbecco's Modified Eagle Medium (DMEM; chosen to simulate the physiological environment) containing Ts3‐ONB. Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) studies revealed that these regular shaped nanoparticles are nanospheres with a size of about 85 nm (PDI = 0.14, Figure 2c). No evidence of nanoparticle formation was seen in the case of Ts3‐OH. This disparity is attributed to the different electrostatic potentials (ESPs) for the two species in question. Ts3‐ONB monomers display an obvious positive and negative charge distribution in its optimized molecular structure (Figure 2d). It can thus easily form π–π aggregates^{[ 31 ]} and undergo self‐assembly to form naphthalimide‐based nano‐self‐assemblies, which contributes to the observed further quenching of fluorescence. Ts3‐ONBS, Ts4‐ONB, and Ts4‐ONBS with similar structures exhibit analogous self‐assembly features (Figures 2c,d and S17).

H₂S Specific Activation of the Photochemical Reaction of Nano‐Self‐Assemblies in Solutions

The ability of H₂S to activate the photochemical reaction of Ts3‐ONB, Ts3‐ONBS, Ts4‐ONB, and Ts4‐ONBS nano‐self‐assemblies was then investigated using an aqueous solution (pH = 6.0) of Na₂S as a source for H₂S. The data shown in Figure 3a supports the conclusion that appart for Na₂S (100 µM), biothiols such as GSH (10 mM), Cys (100 µM) and Hcy (200 µM) do not generate a fluorescence response when added to the Ts3‐ONB, Ts3‐ONBS, Ts4‐ONB, or Ts4‐ONBS nano‐self‐assemblies.^{[ 32} , ³³ , ³⁴ , ³⁵ , ^{36 ]} This is taken as evidence that these species do not induce disassembly and that these constructs remain as non‐fluorescent nano‐self‐assemblies whose emission is quenched by the synergistic effect of PeT and self‐assembly. The effect was most dramatic in the case of the Ts3‐ONB nano‐self‐assemblies which responded to H₂S more effectively than its congeners (F _Ts3‐ONB /F ₀ = 146, F _Ts3‐ONBS /F ₀ = 89, F _Ts4‐ONB /F ₀ = 15 and F _Ts4‐ONBS /F ₀ = 11, Figure 3a). Kinetic evaluation revealed that Ts3‐ONB nano‐self‐assemblies (20 µM, Figure 3b) reacts with H₂S (100 µM) within 9.0 min to form Ts3‐OH monomers (a species that is not subject to PeT; vide supra) through an aromatic nucleophilic reaction as supported by high‐performance liquid chromatography (HPLC; Figure 3c). Based on the data shown in Figure 3d,e, the fluorescence intensity of Ts3‐ONB nano‐self‐assemblies that can be attributed to Ts3‐OH exhibited a good linear relationship with H₂S concentrations with a detection limit is as low 1.6 µM.

Figure 3.

a) Change in the fluorescence intensity of the nano‐self‐assemblies (prepared through the self‐aggregation of Ts3‐ONB, Ts3O‐NBS, Ts4‐ONB and Ts4‐ONBS (20 µM), respectively) toward Na₂S (100 µM), GSH (10 mM), Cys (200 µM), Hcy (100 µM). b) Time‐dependent fluorescence response seen for Ts3‐ONB (20 µM) nano‐self‐assemblies before and after treatment with Na₂S (100 µM). c) HPLC chromatograms of Ts3‐ONB, Ts3‐OH and Ts3‐ONB + Na₂S obtained by recording the absorbance at 350 nm. The retention time (RT) is indicated. d) Fluorescence spectral changes seen for Ts3‐ONB (20 µM) nano‐self‐assemblies when exposed to different concentrations of Na₂S⁻. (0–16 µM; a mixture of DMSO and H₂O, V _DMSO:V _H2O = 5:1). e) A linear fitting of the fluorescence intensity of Ts3‐ONB (20 µM) nano‐self‐assemblies and Na₂S concentration (0–16 µM). f) Electron spin resonance (ESR) spectra of Ts3‐OH (20 µM) recorded with and without light irradiation (mercury lamp, 93.82 mW cm⁻²). Note: This light source is a fixed light source for the instrument; 2,2,6,6‐tetramethylpiperidine (TEMP) was used a trapping agent for ¹O₂ whereas 5,5‐dimethyl‐1‐pyrroline N‐oxide (DMPO) was used as a trapping agent for O₂ ⁻. g) Change in the fluorescence intensity of singlet oxygen sensor green (SOSG, 5 µM) seen upon photo‐irradiation (two‐photon excitation, 800 nm, 10.49 mW cm⁻²) in the presence of untreated nano‐self‐assemblies Ts3‐ONB, Ts3‐ONBS, Ts4‐ONB and Ts4‐ONBS and control molecules Ts3‐OH and Ts4‐OH (20 µM) recorded between 0–80 s, the test environment: pure H₂O, n = 3. h) Fluorescence emission intensity changes of dihydrorhodamine 123 (DHR 123, 20 µM) in the presence of untreated nano‐self‐assemblies Ts3‐ONB, Ts3‐ONBS, Ts4‐ONB, and Ts4‐ONBS and control molecules Ts3‐OH and Ts4‐OH (20 µM) when subjected to photo‐irradiation (two‐photon excitation, 800 nm, 10.49 mW cm⁻²) for varying times (0–80 s), n = 3. i) Fluorescence emission spectra of DHR 123 recorded upon light irradiation (two‐photon excitation, 800 nm, 10.49 mW cm⁻²) in the presence of untreated, Ts3‐ONB (20 µM) nano‐self‐assemblies, Na₂S‐treated Ts3‐ONB (20 µM) nano‐self‐assemblies or Ts3‐OH (20 µM) for varying time intervals (0–80 s), n = 3. j) Femtosecond transient absorption spectra of Ts3‐OH (0.5 mM in DMSO, λ _ex = 350 nm). k) Changes in the fluorescence intensity of Ts3‐OH (20 µM) seen upon treating with different concentrations of 7,7,8,8‐tetracyanoquinodimethane (TCNQ, 0–16 µM). l) Stern–Volmer plots generated from the fluorescence intensity changes of Ts3‐OH (20 µM) seen in the presence of increasing concentrations of TCNQ (0–16 µM).

The photochemical activity of Ts3‐ONB nano‐self‐assemblies, Ts3‐ONBS nano‐self‐assemblies, Ts4‐ONB nano‐self‐assemblies, Ts4‐ONBS nano‐self‐assemblies, and the control molecules, Ts3‐OH and Ts4‐OH (corresponding to the products generated by reaction with H₂S) were evaluated further in an effort to determine the levels of ROS production. With this goal in mind, the electron spin resonance (ESR) spectra were recorded before and after photo‐irradiation.^{[ 37 ]} The results indicated that only the ESR spectrum of Ts3‐OH has a very strong characteristic peak of O₂ ^·− (Figure 3f), while the ESR spectrum of the nano‐self‐assemblies or Ts4‐OH exhibit only a weak characteristic O₂ ^·− peak (Figure S21) that is negligible in comparison with Ts3‐OH. In addition, there were no obvious characteristic peaks for ¹O₂ in all the nano‐self‐assemblies and the control molecules (Figures 3f and S21). Additionally, using 9,10‐anthracenediyl‐bis (methylene) dimalonic acid (ABDA) and singlet oxygen sensor green (SOSG) as indicators for ¹O₂, none of the nano‐self‐assemblies or control molecules generated an ¹O₂ signal (Figures 3g and S22).^{[ 38 ]} Meanwhile, in addition to the ESR spectral studies, dihydrorhodamine 123 (DHR 123) was used as a general indicator of ROS, including O₂ ^·−.^{[ 39 ]} As shown in Figure 3h, the fluorescence intensity of DHR 123 in the presence of Ts3‐ONB nano‐self‐assemblies exhibited negligible fluorescence changes under 80 s of irradiation (F _Ts3‐ONB /F ₀ = 6.7, Figure 3h). In contrast, a statistically significant change was seen for Ts3‐OH (F _Ts3‐OH /F ₀ = 38.4). This finding is consistent with the suggestion that the ability of the Ts3‐ONB nano‐self‐assemblies to produce O₂ ^·− is inherently limited. No significant change in the fluorescence intensity were observed for DHR 123 when exposed to Ts4‐OH, Ts4‐ONB nano‐self‐assemblies or the Ts4‐ONBS nano‐self‐assemblies (Figure 3h). Importantly, however, after treating the Ts3‐ONB nano‐self‐assemblies with Na₂S aqueous solution, a rapid increase in the fluorescence intensity of DHR 123 was seen upon photo‐irradiation (F _{Ts3‐ONB+Na2S} /F ₀ = 29, Figure 3i). This finding is consistent with what was seen for Ts3‐OH. Furthermore, positive signals near 525 nm appeared in the femtosecond transient absorption spectrum (Figure 3j), which can be attributed to formation of an excited triplet state. Quenching experiments using TCNQ were also performed.^{[ 40 ]} Collectively, they provide support for intramolecular electron transfer occurring within Ts3‐OH following photo‐excitation^{[ 41 ]} (K _Ts3‐OH  = 2.8 ± 0.7 × 10⁵ M⁻¹, Figures 3k,l). These results support a Type‐I PDT process involving O₂ ^·− production by Ts3‐OH via electron transfer from the T₁ state, as well as recovery of O₂ ^·− production by the Ts3‐ONB nano‐self‐assemblies following H₂S treatment.

The Synergistic Regulation of PeT and Self‐Assembly by H₂S in Living Cancer Cells

Next, cell‐based experiments were conducted in an effort to assess in preliminary fashion the biocompatibility of Ts3‐OH and the Ts3‐ONB nano‐self‐assemblies (Figures S23–S25). The fluorescence intensity of both systems showed minimal variation in the presence of ions, amino acids, or variations in pH. When Ts3‐ONB nano‐self‐assemblies encounter nitroreductase and β‐NADPH with similar reducing capacity to H₂S, the fluorescence signal changes can be ignored. Furthermore, the fluorescence intensity of Ts3‐OH and the Ts3‐ONB nano‐self‐assemblies remained stable after 5.0 h of continuous irradiation using an iodine tungsten lamp (82.60 mW cm⁻²). Cell endocytosis inhibition experiments revealed that Ts3‐OH entered cells through free diffusion while Ts3‐ONB nano‐self‐assemblies entered cells through micropinocytosis (Figure S26).^{[ 42 ]}

As the next step in these studies, we investigated the synergistic regulation of PeT and self‐assembly in living cancer cells (HepG2 cell lines). Two‐photon imaging (Figure 4a) revealed a fluorescence signal at 610–650 nm (red channel) that increased in intensity as a function of incubation time (1.0, 3.0, 5.0, 7.0, and 9.0 h). The fluorescence intensity increased by a factor of 6.2 after a 9.0 h of incubation (Figure 4a). The increase in fluorescence is primarily attributed to the near‐quantitative conversion of the Ts3‐ONB nano‐self‐assemblies to Ts3‐OH monomers (which lacks a PeT effect; vide supra) mediated by endogenous H₂S. Consistent with this supposition, after 9.0 h, the nanostructure of Ts3‐ONB nano‐self‐assemblies was no longer observable using electron microscopy (Figure 4b). The putative morphological changes in the Ts3‐ONB nano‐self‐assemblies were further probed using a cell model where H₂S is inhibited by N‐ethylmaleimide (NEM, an inhibitor of H₂S).^{[ 43 ]} In contrast to what was seen in the presence of an H₂S source, in the presence of this inhibitor nanoparticles were still readily observable. These findings are taken as evidence that both PeT and self‐assembly can be regulated by H₂S in living cancer cells. Similarly, the disassembly of Ts3‐ONB nano‐self‐assemblies in the presence of H₂S was also observed by SEM and TEM imaging in vitro (Figure S27).

Figure 4.

a) Two‐photon confocal fluorescence imaging of HepG2 cells after treatment with Ts3‐ONB (20 µM) in the presence of Na₂S (100 µM) and Ts3‐OH (20 µM) as determined at different times (1.0, 3.0, 5.0, 7.0 and 9.0 h), scale bars: 50 µm; Green channel and red channel intensities and corresponding statistical analysis in (a), n = 3. b) Representative bio‐TEM images of HepG2 cells recorded after subjecting to different protocols: not treated (control group), Ts3‐ONB (20 µM, Ts3‐ONB group), NEM + Ts3‐ONB (H₂S‐inhibited group), N‐ethylmaleimide (NEM, an inhibitor of H₂S), scale bars: 2.0 µm in the low‐magnification images and 500 nm in the high‐magnification images. The scale bar in enlargement image is 50 nm. c) Intracellular ROS production by Ts3‐ONB (20 µM) nano‐self‐assemblies and Ts3‐OH (20 µM) in HepG2 cells under normoxic (21% O₂), mild hypoxia (2% O₂) and severe hypoxic (0.1% O₂) conditions as determined using the DCFH‐DA assay under photo‐illumination (two‐photon excitation, 800 nm, 10.49 mW cm⁻², 3.0 min); λ _ex: 488 nm, λ _em: 500–560 nm, scale bars: 20 µm; The corresponding statistical analysis in (b), n = 3. d) Intracellular O₂ ^·− production by Ts3‐ONB (20 µM) nano‐self‐assemblies and Ts3‐OH (20 µM) in HepG2 cells under normoxic (21% O₂), mild hypoxia (2% O₂) and severe hypoxic (0.1% O₂) conditions as determined using a DHE assay under photo‐illumination (two‐photon excitation, 800 nm, 10.49 mW cm⁻², 3.0 min); λ _ex: 559 nm, λ _em: 590–630 nm, scale bars: 50 µm; The corresponding statistical analysis in (c), n = 3. ^ns p > 0.05, ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001 and ^**** p < 0.0001. Note: The same control group was used for comparison between Ts3‐ONB and Ts3‐OH to better highlight the differences in their effects.

2′,7′‐Dichlorofluorescein diacetate (DCFH‐DA) was used as an indicator^{[ 44 ]} to determine the total intracellular ROS produced by Ts3‐OH monomers and the Ts3‐ONB nano‐self‐assemblies through the photochemical reaction in living cancer cells (HepG2 cell line) under photo‐irradiation. As shown in Figure 4c, after irradiation under both normoxic and 2% hypoxic conditions, almost identical bright green fluorescence signals were observed in both the Ts3‐ONB nano‐self‐assemblies group and the Ts3‐ OH monomers group. Under severe hypoxia (0.1% O₂), however, the fluorescence signals were reduced by 42% in the Ts3‐ONB nano‐self‐assemblies group and 31% in the Ts3‐OH monomers group. In contrast, the untreated control group exhibited negligible fluorescence. The specific ROS species generated were then determined using confocal microscopic imaging. Dihydroethidium (DHE) and DHR 123 were used as indicators for O₂ ^·−. Following irradiation, almost comparable fluorescence signals corresponding to O₂ ^·− generation were observed in both the Ts3‐ONB nano‐self‐assemblies and Ts3‐OH monomers groups under normoxic and 2% hypoxic conditions (Figures 4d and S28).^{[ 37} , ^{39 ]} However, under severe hypoxia (0.1% O₂), the fluorescence signals were diminished. In the DHE assay, the signal decreased by 35% in the Ts3‐ONB nano‐self‐assemblies group and by 39% in the Ts3‐OH monomers group. In the DHR123 assay, the reductions were 54% and 55%, respectively. By contrast, negligible fluorescence was observed in the untreated control group. Therefore, we conclude that both the Ts3‐ONB nano‐self‐assemblies and Ts3‐OH monomers can generate O₂ ^·− effectively in the HepG2 cancer cell lines under normoxic and mildly hypoxic conditions.

Both HepG2 and the BRL‐3A cell lines were used to evaluate the cytotoxicity of the nano‐self‐assemblies (Ts3‐ONB, Ts3‐ONBS, Ts4‐ONB, and Ts4‐ONBS) and control monomer molecules (Ts3‐OH and Ts4‐OH). This was done using a standard MTT assay with or without irradiation. No significant cytotoxicity was observed in the absence of irradiation (Figure 5a). Likewise little effect on cell viability was seen after irradiation in the case of Ts4‐OH monomers, and the Ts4‐ONB and Ts4‐ONBS nano‐self‐assemblies. However, Ts3‐OH monomers and Ts3‐ONB nano‐self‐assemblies exhibited significant concentration‐dependent phototoxicity toward HepG2 cells but little effect on BRL‐3A cells viability. For instance, a 74.73 ± 1.48% survival rate for the HepG2 cells was seen when 5 µM of the Ts3‐ONB was tested. This rate dropped to 28.24 ± 1.14% when the concentration was increased to 20 µM (incubation time, 24 h). Conversely, whether it is 5 or 20 µM of Ts3‐ONB, the survival rate of BRL‐3A cells always remains at approximately 100%. As might be expected, the phototoxicity of the Ts3‐ONB nano‐self‐assemblies also proved dependent on the incubation time (Figure S29). With an increase in incubation time under endogenous H₂S, the cytotoxicity induced by Ts3‐ONB (20 µM) nano‐self‐assemblies (28.24 ± 1.14%, survival rate after 24 h) gradually increased and became similar to that induced by Ts3‐OH monomers alone (20 µM, 26.50 ± 0.28%, survival rate after 24 h). In addition, cancer cell lines (4T1 cell lines) and normal cell lines (CHO cell lines) were selected to further evaluate the cytotoxicity of Ts3‐ONB nano‐self‐assemblies and Ts3‐OH monomers. The data indicated that they only exhibited significant phototoxicity to cancer cell lines—4T1 cell lines (Figure S30). Subsequently, calcein‐AM/PI double staining experiments were carried out on living cells using confocal microscopic imaging (Figure 5b).^{[ 45 ]} Treatment with either Ts3‐OH monomers and the Ts3‐ONB nano‐self‐assemblies under conditions of photoirradiation produced a visible red fluorescence, indicating dead cells along with a small amount green fluorescence signal indicating living cells. In contrast, only green fluorescence was observed without irradiation. These results confirm that Ts3‐ONB nano‐self‐assemblies can indeed induce cancer cell death, an effect ascribed to a synergistic process involving PeT and regulation of self‐assembly by endogenous H₂S.

Figure 5.

a) Cell viability of HepG2 cells and BRL‐3A cells incubated with nano‐self‐assemblies (Ts3‐ONB, Ts3‐ONBS, Ts4‐ONB and Ts4‐ONBS) and control molecules (Ts3‐OH and Ts4‐OH) in the absence and presence of photo‐irradiation (two‐photo excitation, 800 nm, 10.49 mW cm⁻²), n = 3. b) Confocal fluorescence imaging of calcein‐AM (2.0 µM, live cell marker) and propidium iodide (PI, 4.5 µM, dead cell marker) stained HepG2 cells after different treatments (two‐photo excitation, 800 nm, 10.49 mW cm⁻²; scale bars 50 µm). Signal intensity statistics data for the Ts3‐ONB (20 µM) nano‐self‐assemblies and Ts3‐OH (20 µM) in (b), n = 3. ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001 and ^**** p < 0.0001.

In Vivo PDT and Biosafety of Ts3‐ONB Nano‐Self‐Assemblies

The PDT effect of Ts3‐ONB nano‐self‐assemblies was further evaluated using 4T1 tumor‐bearing mice in vivo (Figure 6). Mice with similar physiological conditions were chosen for this portion of the study. They were housed under identical environmental conditions. The study consisted of three parallel experiments, with each group divided into Ts3‐ONB nano‐self‐assemblies group, Ts3‐ONB nano‐self‐assemblies + Light group, untreated control group and control + Light group. As can be seen from Figure 6a, the tumors in the Ts3‐ONB nano‐self‐assemblies group almost disappeared (V _{14 days} ^{Ts3‐ONB + Light group}/V _{14 days} ^{Control group} ≈ 0.03, Figure 6a) after being injected into the tail vein of the animals and exposed to light at 1, 3, and 5 days after tumor formation with final monitoring at day 14. The tumors in the control and other groups were found to grow rapidly over this period (Figure 6a). No significant change in body weight throughout the 2 week treatment period was observed (Figure 6a). Hematoxylin–eosin (H&E) staining was performed on the tumor sections and normal organs (heart, liver, spleen, lung, and kidney) taken from mice in the different treatment groups. No histopathological alterations or tissue damage was observed in the post‐treatment heart, liver, spleen, lung, and kidney (Figure S31). However, pronounced tumor nucleolysis and apoptosis were evident in the Ts3‐ONB nano‐self‐assemblies + Light group (Figure 6b). In addition, the blood half‐life, blood routine analysis, and blood biochemical tests were performed respectively to verify the biosafety of Ts3‐ONB nano‐self‐assemblies. The data indicated that the blood half‐life was 0.90 h (Figure S32), and there were no significant differences in blood routine and blood biochemical indexes between the control group and Ts3‐ONB nano‐self‐assemblies group (Figures 6c and S33). In vivo distribution experiments revealed that it is primarily concentrated in the tumor and liver, and it probably metabolizes mostly through the liver (Figure 6d). On the basis of these results, we suggest that Ts3‐ONB nano‐self‐assemblies could emerge as promising candidates for PDT‐based cancer treatment with minimal side effects.

Figure 6.

In vivo photodynamic therapeutic effect seen for the different indicated treatment groups, the concentration of the tail vein injection is 500 µM, n = 3. a) Photographs of harvested tumors; tumor volume changes in different groups, ^**** p < 0.0001; body weight changes seen over the course of 14 days. b) HE staining performed in tumor sections taken from different treatment groups, scale bars: 200 µm. (two‐photo excitation, 800 nm, 22.25 mW cm⁻²). c) Blood routine analysis and biochemical analysis after the tail vein injection of Ts3‐ONB, ^ns p > 0.05. d) Imaging for the metabolic pathway of Ts3‐ONB in tumor‐bearing mice. The fluorescence of Ts3‐OH (the metabolic product of Ts3‐ONB) was used to detect the metabolic pathway.

Conclusion

In summary, we have developed a synergistic strategy for enhancing the Type‐I photochemical reaction of PSs, which involves the synergistic application of a PeT mechanism and nano‐self‐assembly. Using this strategy, we developed a series of nanoscale self‐assembled constructs, namely Ts3‐ONB, Ts3‐ONBS, Ts4‐ONB, and Ts4‐ONBS, as well as two control molecules, Ts3‐OH and Ts4‐OH. H₂S was selected as an activator to control (dis)assembly with the basic structure of the constructs being used to control the Type‐I photochemical reactions that were expected to take place following photo‐excitation. A series of spectral, microscopic imaging, and anti‐tumor experiments were carried out in solution, cells, and murine tumor models. The results obtained provided support for the underlying design premise that synergistic regulation by a combination of endogenous H₂S and internal PeT can be used to promote Type‐I photochemical reactions. These events are expected to lead to controllable PDT therapeutic effects and, indeed, good antitumor efficacy was seen in vitro and in vivo via H₂S‐regulation and photo‐irradiation. And the tumor suppression rate of Ts3‐ONB reached 97%, which was superior to most reported PSs (such as porphyrin, 66%, reported by Xu et al).We thus believe that our proposed synergistic strategy wherein a PeT effect is combined with control over self‐assembly can be used to produce efficient N‐PSs that exploit Type‐I photochemical processes. Work is currently underway in our group to generalize these findings.

Conflict of Interests

The authors declare no conflict of interest.

Supporting information

Supporting Information

Niu H., Wang S., Liu Y., Ma N., Cheng S., Feng B., Jeong H., Yang Y., Wang G., James T. D., Yoon J., Sessler J. L., Zhang H., Angew. Chem. Int. Ed.. 2025, 64, e202512150. 10.1002/anie.202512150

Data Availability Statement

The data that support the findings of this study are available in the supplementary material of this article.