Highly Twisted Conformation Thiopyrylium Photosensitizers for In Vivo Near Infrared‐II Imaging and Rapid Inactivation of Coronavirus

Abstract

Despite significant effort, a majority of heavy‐atom‐free photosensitizers have short excitation wavelengths, thereby hampering their biomedical applications. Here, we present a facile approach for developing efficient near‐infrared (NIR) heavy‐atom‐free photosensitizers. Based on a series of thiopyrylium‐based NIR‐II (1000–1700 nm) dyads, we found that the star dyad HD with a sterically bulky and electron‐rich moiety exhibited configuration torsion and significantly enhanced intersystem crossing (ISC) compared to the parent dyad. The electron excitation characteristics of HD changed from local excitation (LE) to charge transfer (CT)‐domain, contributing to a ≈6‐fold reduction in energy gap (ΔE _ST), a ≈10‐fold accelerated ISC process, and a ≈31.49‐fold elevated reactive oxygen species (ROS) quantum yield. The optimized SP@HD‐PEG_2K lung‐targeting dots enabled real‐time NIR‐II lung imaging, which precisely guided rapid pulmonary coronavirus inactivation.

The relationship between molecular configuration and charge transfer process in NIR‐II chromophores has been revealed, which informed the engineering of an efficient NIR photosensitizer (PS), SP@HD‐PEG_2K. The dramatic improvement of the reactive oxygen species (ROS) quantum yield of the PSs enabled NIR‐II image‐guided in vivo pulmonary coronavirus photoablation for the first time and created a convenient paradigm for the development of NIR heavy‐atom‐free PSs.

Introduction

The sudden outbreak of severe acute respiratory syndrome coronavirus type 2 (SARS‐CoV‐2) resulted in an enormous number of fatalities and economic losses. ^[1] Efficient therapeutic techniques are urgently required to fight against SARS‐CoV‐2 and unidentified coronaviruses. To date, several protease‐responsive probes have been developed to detect SARS‐CoV‐2, ^[2] and various antiviral agents have shown promising activity against SARS‐CoV‐2 via targets such as the spike protein, human angiotensin converting enzyme 2 receptor, RNA polymerase, and 3C‐like protease. ^[3] As an enveloped virus, coronavirus replication is highly dependent on the spike protein, envelope protein, and lipid envelope, which are sensitive to the production of ROS. ^[4] Therefore, photodynamic therapy that can induce in situ ROS generation upon photoirradiation with negligible toxicity, antiviral resistance, and gene mutations could be a feasible option. ^[5]

As the key element of photodynamic therapy, photosensitizers are responsible for the in situ generation of cytotoxic singlet oxygen (¹O₂). Principally, ISC from the lowest singlet excited state (S₁) to the lowest triplet excited state (T₁) dictates the efficiency of ¹O₂. ^[6] According to the Fermi golden rule and perturbation theory, the ISC rate (K _ISC) can be enhanced by decreasing the relative ΔE ${}_{\mathrm{S}{}_1-\mathrm{T}{}_{\mathrm{n}}}$ or intensifying the spin‐orbit coupling of the S₁/T_n states. ^[7] The heavy‐atom effect is one of the most popular strategies for lowering ΔE ${}_{\mathrm{S}{}_1-\mathrm{T}{}_{\mathrm{n}}}$ and promoting SOC constants to facilitate ISC. ^[8] However, dark toxicity and shorter triplet lifetime are great concerns for photodynamic therapy with these heavy atom‐containing molecules. ^[9]

In this regard, heavy‐atom‐free photosensitizers based on compact electron donor/acceptor dyads with orthogonal geometries, including naphthalimide‐perylene, cyanines, BODIPYs, and spiro dyads, ^[10] have been intensively studied to facilitate spin‐orbit charge‐transfer ISC. ^[11] Nevertheless, the separation of the HUMO and LUMO distributions inevitably leads to a hyperchromatic shift and diminished fluorescence, ^[12] and consequently, most heavy‐atom‐free photosensitizers cannot be excited by NIR light beyond 700 nm, suffering from severe reflection, scattering, and tissue absorption, which makes it difficult to perform photodynamic therapy in deep tissues. ^[13] In addition, the relationship between chemical structures and ISC or spin‐vibronic coupling in the NIR‐II chromophores remains ambiguous. ^[14] The basic design concepts of heavy‐atom‐free NIR photosensitizers in image‐guided coronavirus inactivation remain elusive. ^[15]

In this study, we present an approach for increasing ROS generation while maintaining favorable NIR‐II photophysical properties by utilizing a highly twisted conformation. By eliminating the thiophene bridging group from the thiopyrylium‐based NIR‐II dyad H4 ^[16] or by further integrating sterically bulky and electron‐rich moieties, three NIR‐II donor‐(π)‐acceptor dyads (H4, HT, and HD) were prepared (Figure 1). In the absence of thiophene, we found that HT has the potential to generate ROS, which was significantly enhanced by coupling the bulky electron donor 4‐(λ ²‐methyl)‐N,N‐dimethylaniline to the thiopyrylium tetrafluoroborate framework 2. The star dyad HD exhibited the highest ROS quantum yield (Φ_Δ, 30.23 %), which was ≈31.49 times higher than that of H4. Femtosecond transient absorption (fs‐TA) and theoretical calculations revealed that the greatly increased dihedral angle (50.7° and 38.5°) between the donor and acceptor units led to the CT‐dominated nature of HD and ≈6‐fold reduced ΔE_ST (0.03 eV) and ≈10‐fold enhanced k _ISC (12.5×10¹²). In addition, HD exhibited a remarkable coronavirus photodynamic destruction effect and low dark cytotoxicity (EC₅₀=8.65–9.86 nM, SI=5.69×10³–6.49×10³) under 808 nm laser irradiation (0.33 W cm⁻², 1 min) due to its heavy‐atom‐free cationic structure. The lung‐targeting dots SP@HD‐PEG_2K enabled real‐time NIR‐II lung imaging and successfully activated cellular antiviral innate immunity and the mitogen‐activated protein kinase (MAPK) and protein kinase B (Akt) signaling pathways, resulting in enhanced in vivo pulmonary coronavirus photoablation and the alleviation of pneumonia.

Figure 1.

Schematic illustration of the energy output and photophysical properties of H4, HT, and HD. a) The synthetic route of H4, HT, HD, and HD‐PEG_2K. b) The highest occupied molecular orbital (HUMO) and lowest occupied molecular orbital (LUMO) plots of H4, HT, and HD, as well as a schematic illustration of their energy output. Fluo., fluorescence; TD, thermal deactivation. c) In vitro penetration depth in the UV‐VIS/NIR‐I/NIR‐II window using intralipid as tissue phantoms imitator. d) Absorption and emission spectra of H4, HT, and HD. e) Photostability of HD and ICG in phosphate‐buffered saline (PBS) buffer under continuous 808 nm irradiation (1.0 W cm⁻²) and fluorescence intensity were recorded at predetermined time points (0, 5, 10, 20, 30, 40, 50, and 60 min). f) Heat map of normalized 1,3‐diphenylisobenzofuran (DPBF) degradation (415 nm) induced by H4, HT, HD, and HD‐PEG_2K under 808 nm laser irradiation with a power density of 1.0 W cm⁻² and 0.33 W cm⁻², respectively. g) Photothermal heating curves of H4, HT, HD, and HD‐PEG_2K with different concentrations (500 μL) at a laser power of 1.0 W cm⁻². h) and i) Electro paramagnetic resonance of HD‐PEG_2K in ultrapure water following 808 nm laser irradiation (1.0 W cm⁻², 5 min). 5,5‐Dimethyl‐1‐pyrroline N‐oxide (DMPO) and 2,2,6,6‐tetramethylpiperidine (TEMP) were employed as singlet oxygen‐ and spin‐trapping agents, respectively.

Results and Discussion

Molecular Design, Synthesis, and Characterization

A donor‐acceptor type thiopyrylium NIR‐II fluorophore H4 was chosen as the main molecular scaffold due to its large molar extinction coefficient within the NIR region (580–1100 nm), and emission (920–1500 nm) bands with a tail extending into the NIR‐II or NIR‐IIa (1300–1400 nm) region, ensuring effective photon capture and NIR‐II imaging in deep tissues. ^[17] To effectively convert the LE state of thiopyrylium‐based dyads to CT‐domain characters, three NIR‐II fluorescent thiopyrylium photosensitizers H4, HT, and HD, were designed and synthesized by the Knoevenagel condensation reaction of thiopyrylium heterocycle accepter 2 with 2‐formyl‐5‐(4′‐N,N‐dimethylaminophenyl) thiophene, 4‐(dimethylamino) benzaldehyde, and Michler's ketone, respectively (Figure 1a). All desired compounds were fully characterized by ¹H NMR, ¹³C NMR, HPLC, and ESI‐HRMS analytical data (Figures S1–S11).

The fluorescence emission, thermal deactivation and ISC dissipation pathways of the photoexcitation energy can be tuned by different thiopyrylium photosensitizers (Figure 1b). The sterically bulky and electron‐rich Michler's ketone of HD significantly reduced the rate of thermal deactivation, allowing the majority of the photoexcitation energy to be released through ISC and fluorescence (QY=0.044 %, using IR‐26 as a reference, Figure S12) and revealing a potent ROS generation efficiency (Φ_Δ=30.23 %) as measured by DPBF at a power density of 0.1 W cm⁻². There was no significant ROS generation by H4, and a slight increase in Φ_Δ (9.35 %) was observed for HT (Figures 1f and S13). In addition, HD retained its high ROS efficiency even at a power density of 0.33 W cm⁻², whereas the DPBF in H4 and HT was barely quenched (Figure 1f). The type of generated ROS was then determined using electron paramagnetic resonance analysis with TEMP (a singlet oxygen trapping agent) or DMPO (a spin‐trapping agent). As shown in Figures 1h and i, the hyperfine splitting constants of the DMPO‐OH adducts (with α(N)=α(H)=14.9 G) were unambiguously observed after 5 min of irradiation, and TEMPO, a product of TEMP with ¹O₂, was also detected in HD. Thus, the production of both ¹O₂ (type‐II) and HO⋅ (type‐I) was detected. ^[18]

To improve the solubility and biocompatibility of HD, MeO‐PEG_2K‐N₃ was conjugated with HD via a copper‐catalyzed azide‐alkyne cycloaddition to produce HD‐PEG_2K (Figures 1a and S2). HD‐PEG_2K self‐assembled into spherical nanoparticles in an aqueous solution with an average diameter of ≈159.96±36.89 nm, as characterized by transmission electron microscopy, and a hydrodynamic diameter of ≈90.40±0.81 nm, as estimated by dynamic light scattering (Figures 4b and S14). The photothermal effect of HD‐PEG_2K dots was almost eliminated, which may be attributed to the restriction of the free vibration and rotation of Michler's ketone in the aggregate state (Figures 1g and S15). The UV/VIS‐NIR absorption band of HD‐PEG_2K dots in dichloromethane appeared at 500–991 nm, while the emission band under 785 nm laser excitation extended into the NIR‐IIa region (Figure 1d). The molar extinction coefficient of HD‐PEG_2K dots was greater than 1.18×10⁵ M⁻¹ cm⁻¹, exhibiting excellent light harvesting ability (Figures S12 and S16). HD‐PEG_2K dots exhibited excellent photostability in PBS buffer under 808 nm laser irradiation (3.5 W cm⁻², 60 min) (Figure 1e). As shown in Figure 1c, the diluted 1 % intralipid was used to imitate the tissue phantoms ^[19] of rhodamine B and HD‐PEG_2K at varying depths. Figure 1c shows the images collected by the NIR‐I or NIR‐II imaging system. The maximum penetration depth of HD‐PEG_2K in the NIR‐I and NIR‐II windows was found to be ≈3 mm and ≈5 mm, respectively, while it was ≈1 mm for rhodamine B in the visible window (Figure 1c). All these results suggest that HD‐PEG_2k permits efficient singlet‐to‐triplet transformation and holds great promise for noninvasive photodynamic therapy in deep tissue.

Figure 4.

Fabrication and biodistribution of HD‐PEG_2K and SP@HD‐PEG_2K. a) Schematic illustration of the preparation of HD‐PEG_2K dots and SP@HD‐PEG_2K. b) Transmission electron microscopy images of HD‐PEG_2K dots, S. platensis, and SP@HD‐PEG_2K. Scale bar: 1 μm, 3 μm, and 2 μm, respectively. c) The Zeta potential of HD‐PEG_2K dots, S. platensis, and SP@HD‐PEG_2K (n=3). d), e) The representative in vivo NIR‐II images (808 nm excitation, 90 mW cm⁻², 1000 nm LP, 300 ms) of mice (n=3) for 5 min after i.v. administration of HD‐PEG_2K and SP@HD‐PEG_2K, respectively. The signal‐to‐background analysis was carried out by measuring fluorescent intensity profiles along the white dashed line and Gaussian fitting. The corresponding fitted curves were presented next to the NIR‐II images. f) The optical and NIR‐II images of S. platensis, HD‐PEG_2K, and SP@HD‐PEG_2K, respectively. g) The mean fluorescence intensity of the excised organs at different time points (5 min, 1, 3, 9, 24, 48, and 72 h) after i.v. administration of HD‐PEG_2K and SP@HD‐PEG_2K in biologically independent replicates (n=3).

Computational Results and Experimental Tests

To analyze the photophysical processes of the dyads in detail, the broadband fs‐TA spectra of H4, HT, and HD in dichloromethane upon excitation at 800 nm were measured. Figure 2a–c show the fs‐TA plots at different delay times and characteristic kinetic curves. Upon photoexcitation, excited‐state absorption (ESA) immediately appeared in the visible and NIR‐II regions of H4, with two prominent bands at ≈560 and ≈1320 nm. Within a 400 fs timeframe, a major peak at ≈560 nm arose concurrently with a decay of the ESA band at ≈1200 nm (Figure 2a). Notably, H4, HT, and HD had apparent ESA bands at ≈560 nm with long‐lived triplet states, which originated from the singlet ESA band at ≈1200 nm via the ISC process (Figure 2a–c). The singlet ESA intensity of HD and HT was significantly reduced compared to that of H4. These results further demonstrated that S₁ was rapidly deactivated via the ISC process. The kinetic curves within the triplet ESA region revealed a rapid increase in components with lifetimes of 0.7 ps, 0.58 ps, and 0.08 ps for H4, HT, and HD, respectively, reflecting the significantly accelerated ISC process of HD (Figure 2d–f). The k _ISC of H4, HT, and HD was determined to be 1.4×10¹² s⁻¹, 1.7×10¹² s⁻¹, and 12.5×10¹² s⁻¹, respectively. This 10‐fold increase in k _ISC explains why HD performs better photodynamic therapy than H4.

Figure 2.

Experimental and theoretical verification of excited states. a), b), c) Pseudo‐color fs‐TA spectra and the fs‐TA plots of H4, HT, and HD upon photoexcitation with an 800 nm pulse (80 μW) at different pump‐probe delay times. The arrow pointing upwards represents the population of the triplet excited state, whereas the arrow pointing downwards represents the depopulation of the singlet excited state. d), e), f) Kinetic curves at representative singlet (triangle) and triplet (circle) ESA regions, illustrating the dynamics of the singlet excited state formation followed by the singlet‐to‐triplet ISC. g) Summary of the experimental/computational results for H4, HT, and HD. SOC, spin‐orbit coupling. h), i) Computational results of the optimized geometries and corresponding energy levels, as well as the probable ISC channels from the S₁ state to higher‐ or lower‐lying triplet states (T_n). j) Distributions of natural transition orbitals for H4, HT, and HD. T₂ of H4 is dominated by the LE state, but T₂ of HD is dominated by the CT state, which decreases the T₂ energy level from 0.166 eV to 0.030 eV for a facile ISC process. The pulsed femtosecond laser: 800 nm, 1000 Hz, 120 fs, and 80 μW.

First‐principles time‐dependent density functional theory was then performed on the molecular orbitals and energy levels of HD to determine the mechanisms responsible for its improved photodynamic efficacy. In principle, facile ISC requires sizeable SOC constants and a small ΔE _ST. Significant progress is particularly evident with the fact that the S₁→T₂ ΔE _ST (0.03 eV) of HD was significantly smaller than those of H4 and HT (0.166 eV and 0.150 eV, respectively) (Figure 2i), which should facilitate the ISC process. To get a better insight into the excited state characteristics, a natural transition orbital (NTO) analysis of the S₀ geometry was performed. All the samples presented LE‐dominated electronic structures of S₁. The T₂ of H4 and HT exhibited a prominent LE state character, whereas the T₂ of HD exhibited an apparent mixture of LE and CT characteristics. The CT occurrence in HD not only reduces the band gap but also decreases the S₁→T₂ ΔE _ST (Figure 2j). EI‐Sayed's rule states that a change in orbital type is essential to enhance SOC. ^[20] This LE‐dominated S₁ and the mixed LE&CT of T₂ in HD satisfy the change in orbital type and thus favor SOC enhancement. The optimized molecular structure revealed that the dihedral angles of H4 between the thiopyrylium unit and thiophene, and thiophene and N, N‐dimethylaniline were ≈4° and ≈6.2°, respectively, suggesting a coplanar structure. Surprisingly, HD exhibited a twisted structure, as indicated by the dihedral angles of 50.7° and 38.5° between thiopyrylium and N, N‐dimethylaniline (Figures 2g and 2 h). This orthogonal orientation between the donor and acceptor of HD partially disrupted the LE‐dominated conjugated system, resulting in a mixture of ³LE and ³CT for facile S₁ (¹LE)→T₂ (³LE&³CT) ISC. These results imply that the twisted geometry for HD is essential to induce efficient ISC.

HD‐PEG_2K Exerts Potent Antiviral Effects Against Coronavirus In Vitro

Given the excellent ROS efficiency and photophysical properties of HD‐PEG_2K , the photodynamic inactivation potential of HD‐PEG_2K against coronavirus was then investigated. HCoV‐OC43 and HCoV‐229E were incubated with different concentrations of HD‐PEG_2K for 1 h. Then, human embryonic lung fibroblast (MRC‐5) cells were infected with HCoV‐OC43 and HCoV‐229E at a concentration of 2.5×10⁻⁴ PFU/cell for 24 h and collected for further analysis (Figure 3a). Total RNA was extracted from cells, and the level of viral RNA was determined using RT–PCR. To enlarge the selectivity index (SI) and minimize cytotoxicity, we lowered the power of 808 nm laser source to 0.33 W cm⁻¹, which is the clinically permitted laser intensity. As shown in Figure 3b, after a short irradiation period (1 min), HD‐PEG_2K effectively inhibited HCoV‐OC43 viral replication at nanomolar levels (EC₅₀=9.86 nM), which was 5689.66 times lower than that of HD‐PEG_2K in MRC‐5 cells (CC₅₀=56.10 μM, SI=5.69×10³, Figures 3c and e). Similar results were observed with HCoV‐229E, confirming the broad‐spectrum antiviral potential of HD‐PEG_2K (EC₅₀=8.65 nM, SI=6.49×10³, Figures 3b and d). To simulate the process of coronavirus infection, MRC‐5 cells were incubated with HCoV‐OC43 or HCoV‐229E for 2 h before the treatment of HD‐PEG_2K (Figure S17). The results of qPCR analysis indicated that upon 808 nm irradiation, HD‐PEG_2K can also efficiently inhibit the intracellular coronaviruses, holding great potential to mitigate coronavirus infection. In addition, the viability of HD‐PEG_2K against MRC‐5 cells remained up to 70 % even at a high concentration of 20 μM without 808 nm laser irradiation (Figure S18).

Figure 3.

The antiviral activity of HD‐PEG_2K against coronaviruses in vitro. a) The general procedure for in vitro photodynamic inactivation of coronaviruses. b), e) qPCR analyses of RNA expression of HCoV‐229E and HCoV‐OC43 (left), and the TCID₅₀ assay for detecting the live viral HCoV‐229E and HCoV‐OC43 titers (right) in biologically independent replicates (n=3). c), d) The EC₅₀, and selective index of HD‐PEG_2K relative to HCoV‐OC43 and HCoV‐229E. CCK‐8 assays were used to determine the cytotoxicity of HD‐PEG_2K to MRC‐5 cells in biologically independent replicates (n=3). EC₅₀, concentration for 50 % of maximal effect; CC₅₀, cytotoxic concentration; SI, selective index. The left and right Y‐axes of the graphs represent the mean inhibition rate of viral yield and the cytotoxicity of HD‐PEG_2K to MRC‐5 cells, respectively. f) Representative immunofluorescence studies of HCoV‐OC43 infected MRC‐5 cells treated with different concentrations of HD‐PEG_2K under 808 nm laser irradiation (0.33 W cm⁻²) for 1 min (n=3), scale bar: 50 μm. g) The mean gray values of the virus antigen levels in MRC‐5 cells were analyzed by immunofluorescence analysis using ImageJ (n=3). h) Representative western blot images of HCoV‐OC43 infected MRC‐5 cells treated with different concentrations (50 nM and 200 nM) of HD‐PEG_2K w/o 808 nm laser irradiation (0.33 W cm⁻²) for 1 min (n=3). N.C., negative control; V.C., virus control; L, laser irradiation. A two‐sided P value<0.05 was considered statistically significant. *P<0.05, **P<0.01, ***P<0.001.

Immunofluorescence and western blotting assays were then used to evaluate the level of viral protein expression in the infected cells. Infected MRC‐5 cells were subjected to immunostaining with a primary anti‐OC43‐specific polyclonal antibody and a TRITC goat anti‐mouse IgG (H+L) secondary antibody. ImageJ was used to perform an immunofluorescence study on virus antigen levels. As shown in Figures 3f and S19, TRITC‐red fluorescence in the MRC‐5 cells decreased when the concentration of HD‐PEG_2K was increased from 20 nM to 100 nM under 808 nm laser irradiation. In contrast, MRC‐5 cells in the control group where HCoV‐OC43 was not pre‐irradiated by laser exhibited very strong red fluorescence at all tested concentrations. The mean gray values were ≈29.77, ≈24.22, and ≈6.44 per cell at concentrations of 20 nM, 50 nM, and 100 nM, respectively, which were significantly lower than that of the viral control (≈51.58) (Figure 3g). The western blotting analysis confirmed that HD‐PEG_2K ‐mediated photodynamic inactivation can inhibit the invasion and amplification of coronavirus within host cells (Figure 3h).

Next, a viral titer‐median tissue culture infectious dose (TCID₅₀) assay was performed to determine the number of infectious virus particles. As shown in Figure 3b and e, the log₁₀TCID₅₀ of HCoV‐OC43 and HCoV‐229E in HD‐PEG_2K ‐mediated photodynamic inactivation significantly decreased by ≈96.04 % and ≈100 %, respectively, at a concentration of 100 nM. Taken together, our results indicate that 808 nm laser irradiation of HD‐PEG_2K efficiently inactivates various coronaviruses, thereby preventing the infection of host cells and potentially inhibiting the spread of coronaviruses.

SP@HD‐PEG_2K Exhibits Robust Anti‐HCoV Activity In Vivo

The respiratory tract is susceptible to coronaviruses and is considered the primary site of viral replication. ^[21] To specifically deliver NIR‐II photosensitizers, Spirulina platensis (S. platensis) with aqueous channels and junctional pores was used as a vehicle for pulmonary drug delivery. ^[22] As shown in Figure 4a, SP@HD‐PEG_2K was prepared by co‐incubating ultrasound‐pretreated S. platensis suspensions (10 mL) with a series of HD‐PEG_2K dots (1000, 500, 250, 125, 62.5, and 31.25 μg mL⁻¹, 1 mL) for 12 h. The highest encapsulation efficiency of SP@HD‐PEG_2K was calculated to be 27.31 % at a drug/carrier ratio of 1 : 1.25 according to the standard calibration curve (Figures S16 and S20). As shown in Figure 4b, HD‐PEG_2K dots was proved to absorb at the surface of S. platensis by transmission electron microscopy analysis of SP@HD‐PEG_2K . Furthermore, the zeta potential of S. platensis increased from −35.97 to −19.89 mV following the adsorption of HD‐PEG_2K dots (−8.89 mV), indicating the successful drug‐loading process of SP@HD‐PEG_2K (Figure 4c). The release kinetic of SP@HD‐PEG_2K at different pH values was subsequently investigated. As shown in Figure S21, ≈35.48 % of the HD‐PEG_2K dots was released within the first 12 h, and then the dots were released in a sustained manner. In addition, the as‐prepared SP@HD‐PEG_2K inherited the NIR‐II emission characteristic of HD‐PEG_2K (Figure 4f) and exhibited superior photostability in different pH environments (Figure S22), thereby enabling high temporal‐spatial resolution tracking in vivo.

The biocompatibility and biodistribution of SP@HD‐PEG_2K were then investigated. SP@HD‐PEG_2K (300 μmol kg⁻¹) was administered intravenously to C57BL/6 mice (n=3) at a dose of 10.0 μmol kg⁻¹. With the help of S. platensis, the biodistribution of SP@HD‐PEG_2K was significantly altered, and SP@HD‐PEG_2K was delivered to the lung immediately after injection (Figure 4d and e). As shown in Figures 4g and S23–S24, SP@HD‐PEG_2K exhibited a 6.43‐fold higher lung uptake and 1.94‐fold lower liver uptake than those of HD‐PEG_2K , which was also revealed by NIR‐II real‐time monitoring (Figures 4d and e). Furthermore, owing to the superior temporal‐spatial resolution of NIR‐II imaging, lung was outlined and clearly distinguishable from liver with a signal‐to‐background of 6.52 (LP1000, 100 ms, Figure 4e). ICR mice (n=3) were then intravenously administered SP@HD‐PEG_2K (500 μmol kg⁻¹, 200 μL) and PBS (200 μL). No deaths were observed, and the body weight in each group exhibited comparable trends (Figure S25). The hemolysis experiment demonstrated that SP@HD‐PEG_2K didn't cause obvious hemolysis at a high concentration up to 100 μM (Figure S26). Moreover, histological analysis of the major organs (liver, heart, spleen, lung, kidney) revealed that SP@HD‐PEG_2K didn't induce apparent inflammatory lesions or damage compared to the control group (Figure S27). Subsequently, hematologic and hepatologic function analyses were carried out (Tables S2 and S3). There were no significant differences between these groups. Taken together, all the results indicate that the delivery system can efficiently improve bioavailability and biocompatibility, thereby reducing the dosage.

To establish HCoV infection mouse models, 4‐week‐old C57BL/6 mice were intranasally challenged with HCoV‐OC43 at a dose of 1×10⁴ (TCID₅₀). All experiments were carried out in accordance with institutional regulations (Approval number: WP20210408; Approval date: April 15, 2021). HCoV‐OC43 infection induced a significant body weight decrease in a time‐dependent manner, starting at 3 days post‐infection (dpi) (Figure S28). Next, we examined viral replication and pathological changes in mice at different time points (0, 1, 3, 5, and 7 days). The primary organs, including the brain, heart, liver, spleen, lung, and kidney, were collected at 5 dpi for quantitative reverse transcription PCR (RT‐qPCR). On the other hand, hematoxylin and eosin (H&E) staining was carried out to analyze the pathological process. As shown in Figure 5c, as the infection progressed, the alveolar wall thickened moderately to severely, accompanied by inflammatory cell infiltration (black arrow). Furthermore, the endothelial cells of the pulmonary veins were dilated and shed, and some veins displayed atrophy, fibrinoid necrosis, and monocyte and lymphocyte infiltration (yellow arrow), indicating severe pneumonia caused by HCoV‐OC43 infection. Subsequently, in vivo anti‐HCoV activities of SP@HD‐PEG_2K in HCoV‐OC43 infection mouse models were assessed (Figure 5a). SP@HD‐PEG_2K (80 μmol kg⁻¹) in 100 μL PBS solution was administered intravenously 24 h after the challenge with an irradiation time of 1 min per pulse every half an hour lasting 1.5 h (808 nm, 0.33 W cm⁻¹). No significant lung histopathological changes were identified in the SP@HD‐PEG_2K +Laser and negative control groups (Figure 5d). No viral replication was detected in the lungs of mice infected with HCoV‐OC43 at 10 dpi (Figure 5b). However, alveolar atrophy, inflammation, and infiltration were observed in the virus+Laser and SP@HD‐PEG_2K groups_, which were comparable to the viral control group (Figure 5d).

Figure 5.

The in vivo antiviral activity of SP@HD‐PEG_2K. a) Schematic illustration of SP@HD‐PEG_2K‐mediated in vivo viral inactivation. b) The RNA expression of HCoV‐OC43 in the lungs 10 days after different treatments (n=3). V.C., viral control; N.C., negative control; L, laser irradiation. c) Representative H&E staining images of infected lungs were collected on days 0, 1, 3, 5, and 7 post‐HCoV‐OC43 challenge. Black arrows indicate the thickened alveolar wall and inflammatory cell infiltration, while yellow arrows indicate the atrophied alveoli and lymphocyte infiltration. Scale bar: 200 μm (upper) and 50 μm (lower). d) Representative H&E staining images of lungs excised from different treatment groups. Scale bar: 200 μm (upper) and 50 μm (lower). The experiments were carried out in biologically independent replicates (n=3). A two‐sided P value<0.05 was considered statistically significant. *P<0.05, **P<0.01, ***P<0.001.

Transcriptomics Analysis

Encouraged by the positive in vivo antiviral effect, we carried out a transcriptomics study on mice to elucidate the underlying biological molecular mechanism. A genome‐wide analysis was first performed to determine SP@HD‐PEG_2K ‐mediated photodynamic inactivation transcriptional networks. A total of 2416 genes were differentially expressed between SP@HD‐PEG_2K +laser and viral control groups, of which 1419 genes were upregulated and 997 were downregulated following SP@HD‐PEG_2K +808 nm laser irradiation (Figures 6a and c). Next, we performed qRT‐PCR to confirm the differential expression of 14 randomly selected DEGs and the viral RNA expression of HCoV‐OC43. Seven of the identified immune‐related DEGs (IL6, IL1b, Tnfrsf1b, Tnfaip2, Ifitm3, Ifi30, Ifitm1) participated in the regulation of multiple immune‐related physiological activities in the host cell (Figures 6e and f). The remaining 7 identified DEGs were implicated in the cell cycle, MAPK, and AKT signaling pathways (Figures 6b, d, and S29). Previous studies have demonstrated that HCoV‐OC43 infection activated MAPK and AKT signaling pathways to increase viral replication, whereas the inhibition of MAPK and AKT signaling hindered viral proliferation. ^[23] In this study, we found that the marker proteins of both signaling pathways were differentially expressed. Genes associated with antiviral innate immunity, such as IL6, IL1b, Ifitm3, and Ifitm1, were significantly enriched among the differentially expressed genes between uninfected host cells, infected host cells, and SP@HD‐PEG_2K +laser‐treated infected host cells. Moreover, genes implicated in the IL‐17 signaling pathway and oxidative phosphorylation were strongly induced following treatment with SP@HD‐PEG_2K +laser, indicating that these pathways are required for the clearance of HCoV‐OC43 and cause changes in host resistance to HCoV‐OC43 infection. As the virus was cleared by SP@HD‐PEG_2K +808 nm laser, the inflammatory response was significantly reduced compared to the virus‐only group of hosts. The western blotting results were consistent with the RNA‐seq and qRT‐PCR results (Figure S30).

Figure 6.

Transcriptomics analysis (a) Hierarchical clustering heat map of the differentially expressed genes in the five tissue samples (N.C., V.C., Virus+L, SP@HD‐PEG_2K, SP@HD‐PEG_2K+L), the green and red colors correspond to low and high abundance of gene expression, respectively. b) Scatter plot of the 20 most significant GO terms of enrichment in tissues w/o treatment. c) Volcano plot of DEGs. Red dots represent up‐regulated genes, whereas green dots present down‐regulated genes (Fold change>1.5 or Fold change<0.667, FDR<0.05). d) Voronoi plots of the top DEGs enriched in the KEGG pathway. e) qRT‐PCR validation of the RNA sequencing results for immune‐related differentially expressed genes. f) qRT‐PCR validation of the RNA sequencing results in the differentially expressed genes in other essential signalling pathways. N.C., negative control; V.C., virus control; L, laser irradiation. The experiments were carried out in biologically independent replicates (n=3).

Altogether, both in vitro and in vivo studies systematically demonstrated that SP@HD‐PEG_2K +808 nm laser could efficiently inactivate human coronavirus with ROS generation, activate the MAPK and AKT signaling pathways, and activate cellular antiviral innate immunity, resulting in a reduction in the inflammatory cascade and protecting cells from coronavirus infection.

Conclusion

In this study, by revealing the tendency of the excited states to change with the geometry and chemical structure of a series of thiopyrylium dyads, we presented a practical approach for constructing NIR‐II heavy‐atom‐free photosensitizers by directly introducing bulky moieties with strong electron‐donating abilities to the thiopyrylium accepter. We discovered that the donor/acceptor dyad HT generated ROS in the absence of the thiophene spacer and that Φ_Δ was significantly increased to 30.23 % along with an elevated k _ISC when Michler's ketone was used as the electron donor. Based on computational and experimental findings, we hypothesized that the absence of a spacer resulted in a twisted geometry that was further accentuated in our star chromophore HD, leading to the conversion from LE to CT‐domain characteristics. The decreased ΔE _ST and the enhanced SOC constant accelerated the ISC dynamics to produce a sufficient triplet population, resulting in the outstanding Φ_Δ of HD.

Considering the sensitivity of coronavirus to ROS, the modified HD‐PEG_2K was used for photodynamic inactivation and demonstrated excellent photodynamic therapy efficiency (EC₅₀=8.65–9.86 nM, SI=5.69×10³–6.49×10³) against various coronaviruses, including HCoV‐OC43 and HCoV‐229E, using only a mild laser (808 nm, 0.33 W cm⁻², 1 min). In addition, S. platensis was used to modify the pharmacokinetic properties of HD‐PEG_2K . NIR‐II imaging of the resulting SP@HD‐PEG_2K precisely guided the implementation of photodynamic therapy with a high temporal‐spatial resolution, which subsequently inactivated pulmonary coronavirus, prevented pneumonia, and activated innate immune signaling. We believe that our study not only facilitates the design of long‐wavelength photosensitizers but also expands the in vivo applications of photosensitizers and provides a promising tool for in vivo coronavirus inactivation by inhibiting immune microenvironments.

Conflict of interest

The authors declare no conflict of interest.

Supporting information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supporting Information

Liu Y., Gu M., Ding Q., Zhang Z., Gong W., Yuan Y., Miao X., Ma H., Hong X., Hu W., Xiao Y., Angew. Chem. Int. Ed. 2023, e202214875; Angew. Chem. 2023, e202214875.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.