Synthesis of the SCCs.

Ru(bpy)₃²⁺ (bpy = 2,2′-bipyridine) and its derivatives have widespread applications not only in PDT (7, 56, 58⇓–60) but also in photoredox catalysis (61) and photovoltaics (62). A ditopic, Ru(II)-containing donor 1 was obtained by replacing 1 bpy group of Ru(bpy)₃²⁺ with 4,4′:2′,2″:4″,4″′-quaterpyridine. Based on the principles of coordination-driven self-assembly, the geometries of the SCCs are determined by the lengths and angles of the donor and acceptor building blocks (63). The combination of 1 with di-Pt(II) acceptor 2 or tri-Pt(II) acceptor 3 in CH₂Cl₂/CH₃OH (v:v = 1:1) resulted in the formation of [2+2] rhomboidal metallacycle 4 or [6+4] octahedral metallacage 5 (Fig. 1A). The ³¹P NMR spectra of both 4 and 5 revealed sharp singlets with concomitant ¹⁹⁵Pt satellites (δ = 13.62 and 13.32 ppm for 4 and 5, respectively) (Fig. 1 B and C), indicating the formation of a discrete, single product with high symmetry upon self-assembly. The formation of single species is also supported by ¹H diffusion ordered spectroscopy NMR experiments, with diffusion coefficient values of 7.7 × 10⁻¹¹ m²⋅s⁻¹ and 6.9 × 10⁻¹¹ m²⋅s⁻¹ for 4 and 5, respectively (SI Appendix, Figs. S2 and S5). The compositions of 4 and 5 were then established using electrospray ionization time-of-flight mass spectrometry (ESI-TOF-MS). The ESI-TOF-MS spectrum of 4 featured peaks with m/z values of 917.26, 721.59, and 591.15, which could be assigned to the [2+2] assembly with +4, +5, and +6 charges due to the progressive loss of the nitrate anions (Fig. 1D and SI Appendix, Fig. S3). Similarly, a series of peaks corresponding to the [6+4] assembly with +7 to +14 charges were observed for the ESI-TOF-MS spectrum of 5 (Fig. 1E and SI Appendix, Fig. S6). The isotopic distributions of the experimental spectra were also closely matched with the simulated spectra, supporting the proposed stoichiometry for 4 and 5.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Self-assembly and characterization of the SCCs. (A) [2+2] self-assembly to furnish 4 and [6+4] self-assembly to furnish 5. (B and C) ³¹P{¹H} NMR spectra of 4 (B) and 5 (C). (D–E) ESI-TOF-MS spectra of 4 (D) and 5 (E).

Photophysical Properties and ROS Generation of the SCCs.

The steady-state photophysical profiles of the SCCs were summarized in SI Appendix, Table S1. Donor 1 showed a metal-to-ligand charge-transfer (MLCT) absorption band centered at 468 nm (SI Appendix, Fig. S7). The SCCs exhibited similar absorption characteristics to 1, but 4 and 5 had redshifted MLCT bands at 482 and 481 nm, respectively (SI Appendix, Fig. S7). Upon excitation of the MLCT transition, both 4 and 5 exhibit a nonstructure red emission band centered at 672 and 673 nm, respectively, which is redshifted by ∼30 nm relative to 1 (Fig. 2A). The long lifetime of the emission (>400 ns) and enhanced quantum yield under argon atmosphere indicate the emission could be phosphorescent in nature. The redshifted absorption, emission, and large Stokes shift (190 nm) of the SCCs favored further biological studies.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Photophysical properties and ROS generation of the compounds. (A) Luminescence emission spectrum of the compounds. (B) Two-photon absorption cross-sections of the compounds at different excitation wavelengths from 780 to 930 nm. Measurement of ¹O₂ production efficiency via changes in the absorbance by DPBF at 411 nm versus irradiation time (λ_irr = 450 nm) in the presence of compounds in adjusted concentrations. [Ru(bpy)₃]²⁺ was used as the standard. (Inset) Emission spectra of the mixture of 5 and DPBF upon irradiation.

The 2-photon cross-sections of the SCCs from 780 to 920 nm were measured using the 2-photon excitation fluorescence method (64) and compared with 1 (Fig. 2B). Both SCCs displayed their largest 2-photon absorption at 820 nm, with values at 1,969 GM for metallacycle 4 and 5,468 GM for metallacage 5, whereas the precursor 1 showed a 2-photon cross-section of 161 GM under the same conditions. The significant enhancement in 2-photon absorption efficiency observed for the SCCs is attributed to the organization of multiple Ru(II) complexes in a single supramolecular system via self-assembly, and the enhanced charge polarization of the Ru(II) motif induced by the rigid SCC structure (22, 65, 66). The emission spectra of compounds 1, 4 and 5 upon excitation at 820 nm (SI Appendix, Fig. S8) are similar to their linear emission spectra, and their emission intensities show quadratic dependence on the excitation power, confirming that the compounds are 2-photon-active.

Singlet oxygen (¹O₂) is the major ROS generated by Ru(II)-based PSs (7, 55⇓–57). The ¹O₂ quantum yields (Φ_▵) of 1, 4, and 5 were directly determined by assessing the luminescence of ¹O₂ at 1270 nm upon photoexcitation of the compounds (SI Appendix, Fig. S9B). A solution of Ru(bpy₃)²⁺ was used as a reference sensitizer (Φ_▵ = 81.0%). This resulted in Φ_▵ values of 63.3, 75.3, and 85.0% for 1, 4, and 5, respectively. An indirect measurement for Φ_▵ was also performed by monitoring the PS-induced decomposition of green-emitting 1,3-diphenylisobenzofuran (DPBF), giving comparable values (Φ_▵ of 62.9, 76.7, and 86.2% for 1, 4, and 5, respectively, Fig. 2C). The high 2-photon cross-sections and Φ_▵ for the SCCs, especially for the metallacage 5, indicate their great potential as efficient ¹O₂ sensitizers for 2-photon PDT.

Fabrication, Cellular Uptake, and Localization of SCC-Loaded Nanoparticles.

Metallacycle 4 exhibited favorable cellular properties and accumulates in mitochondria (SI Appendix, Fig. S11), whereas metallacage 5 displayed poor cellular uptake due to its limited solubility in aqueous solution (SI Appendix, Fig. S12). To realize its application in physiological conditions and increase its biocompatibility and bioavailability, metallacage 5 was encapsulated using DSPE-mPEG2000 as the matrix. The loading content of 5 in the nanoparticles was determined to be 17.6% (SI Appendix, section 4.1). The formation of metallacage 5 loaded nanoparticles (5-NPs) is directly supported by the observation of spherical micellar nanostructures with a diameter around 230 nm in transmission electron microscopy (TEM) images (Fig. 3B). Dynamic light-scattering (DLS) measurement on the solution of 5-NPs revealed an average size of 260 nm (Fig. 3C), in agreement with the TEM results. The stability of 5-NPs in serum containing cell culture medium was examined by tracking the luminescence emission profile of the mixture (SI Appendix, Fig. S10). No apparent decomposition was observed after 24 h, suggesting 5-NPs have a good stability for biomedical applications.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Fabrication of 5-NPs and their cellular distribution. (A) Schematic illustration of 5-NPs fabrication, cellular uptake, accumulation in the lysosomes, and their applications in 2-photon PDT. (B) TEM image of 5-NPs. (C) DLS result for solution of 5-NPs. (D) ICP-MS quantification of the subcellular distribution of Ru and Pt by A549 cells. (E) Colocalization images of 5-NPs with lysosome dye LTG with corresponding correlation coefficients. (Scale bars: 20 μm.)

The cellular uptake profile of 5-NPs in the A549 lung adenocarcinoma cell line was monitored by confocal laser scanning microscopy (CLSM). Strong red intracellular emission was observed in A549 cells incubated with 5-NPs, with a continuous increase in the first 24 h of incubation (SI Appendix, Fig. S13), indicating successful localization of 5-NPs into A549 cells. The specific subcellular distribution of 5-NPs was examined by a colocalization assay. The red emission from 5-NPs exhibited good overlap with the commercial lysosome dye LysoTracker Green (LTG), with a correlation coefficient (R) of 0.88 (Fig. 3E), indicating that 5-NPs may selectively accumulate in the lysosome after entering the cells. In contrast, poor overlap was observed for the endoplasmic reticulum dye ER-Tracker Green, the mitochondrion dye MitoTracker Green, and the nuclear dye Hoechst 33342 with low-R values (SI Appendix, Fig. S14).

To quantify the cellular uptake and intracellular distribution of 5-NPs, inductively coupled plasma mass spectrometry (ICP-MS) analysis was employed to determine the cellular platinum and ruthenium concentrations. Time-dependent ICP-MS studies confirmed that the intracellular cellular ruthenium and platinum levels reach a maximum at 24 h of incubation, with concentrations of 9.0 mg per million cells and 37.1 mg per million cells for ruthenium and platinum, respectively (Fig. 3D and SI Appendix, Table S2). In an analysis of the subcellular distribution, the lysosomes were found to contain a significantly higher quantity of both ruthenium (74.0%) and platinum (75.2%) compared to the endoplasmic reticulum, mitochondria, and nucleus, in accordance with the colocalization assay (SI Appendix, Table S3). Analyses of the ICP-MS results with 5-NPs revealed that the intracellular molar ratio of ruthenium: platinum is close to 1:2 among all cellular compartments (SI Appendix, Tables S2 and S3), indicating good stability of the metallacage in complex physiological conditions as well as cellular uptake and distribution as an intact ensemble.

(Photo)cytotoxicity toward Monolayer Tumor Cells.

The (photo)cytotoxicity of 4 and 5-NPs was evaluated on monolayer A549 cells using the 3-(4′,5′-dimethylthiazol-2′-yl)-2,5-diphenyl tetrazolium bromide assay and the results are summarized in SI Appendix, Tables S4 and S5. In the absence of light, 5-NPs were found to be strongly nontoxic, with a half-maximal inhibitory concentration (IC₅₀) value greater than 120 μM (based on the concentration of 5, the same for below), whereas light irradiation led to a marked increase in cytotoxicity of 5-NPs (Fig. 4). The cell viabilities were also shown to be sensitive to the light dosage. When exposed to a very low fluence of 6 J⋅cm⁻², 5-NPs were found to be highly cytotoxic, with an IC₅₀ value of 4.83 μM. The low dark cytotoxicity and high (photo)cytotoxicity of 5-NPs led to a (photo)cytotoxicity index (PI) greater than 24.8. The IC₅₀ value was further reduced to 1.89 μM by increasing the light dose to 12 J⋅cm⁻², resulting in a PI value greater than 63.5. In both cases, the morphology of the untreated cells subjected to the same dose of light was found to be unaffected, indicating the noninvasiveness of the light irradiation procedure. In contrast, 4 is toxic in the dark (IC₅₀ = 93.3 μM), likely due to its mitochondrial localization (32, 33). Moreover, the (photo)cytotoxicity of 4 (IC₅₀ = 4.93 μM) was found to be lower than the 5-NPs, resulting in a lower PI for 4 in comparison with the 5-NPs; 5-aminolevulinic acid (5-ALA), a clinically approved PS for PDT, was found to show lower phototoxicity (IC₅₀ = 111 μM) and a lower PI than the 5-NPs under similar conditions. These results suggest the potential of 5-NPs as an efficient PS with low side effects due to their lysosomal accumulation and high ROS production efficiency.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Light and dark cell viability figures of 5-NPs at different concentrations and varied light fluence.

The intracellular generation of ¹O₂ by 2-photon irradiation was confirmed using A549 cells incubated with 5-NPs and 2,7-dichlorodihydro-fluorescein diacetate (DCFH-DA). Upon exposure to a 2-photon laser (820 nm, 6.06 mW, 40 s), intracellular green emission originating from the oxidation of DCFH by ¹O₂ was observed (SI Appendix, Fig. S15), suggesting the generation of ¹O₂ inside the cells by activation of the 5-NPs.

The production of ¹O₂ within lysosomes can lead to uncontrolled lysosomal permeability via massive peroxidation of membrane lipids (67). The integrity of the lysosomes upon 5-NPs-mediated PDT was examined using acridine orange (AO) staining. In acidic organelles, AO displays a red emission, whereas a green emission can be detected in the cytosol and nuclei (68). Upon exposure to 2-photon irradiation, a marked decrease in the red fluorescence from AO was observed for A549 cells treated with 5-NPs (SI Appendix, Fig. S16), indicating 2-photon PDT induced lysosomal disruption by the 5-NPs, whereas cells in the control group that were irradiated with light alone were undamaged as both red and green emission was detected, suggesting no lysosomal damage caused by light exposure (SI Appendix, Fig. S16). The disruption of lysosomal integrity can also lead to the release of lysosomal proteases, such as cathepsin B, from the lysosomes to the cytosol which induces apoptosis (26). The intracellular distribution of cathepsin B before and after 2-photon PDT using the 5-NPs was detected using a Magic Red staining analysis (69). After 2-photon PDT, red fluorescence from Magic Red diffusion from the lysosomes to the cytosol was observed (SI Appendix, Fig. S17), indicating the release of cathepsin B into the cytosol after PDT-mediated lysosomal damage.

An annexin V/propidium iodide double-staining assay (70) was employed to confirm that 2-photon PDT using 5-NPs can efficiently induce apoptosis/necrosis. A549 cells treated with 5-NPs and exposed to light irradiation displayed both green emission by Annexin V and red emission by propidium iodide (SI Appendix, Fig. S18), indicating cellular damage mediated by 2-photon light-activated 5-NPs that lead to apoptosis/necrosis. These results suggest that the cells were in necrosis or late apoptosis after 2-photon PDT.

Two-Photon PDT toward 3D Multicellular Spheroids.

In order to mimic the pathophysiology of solid tumors in vivo, such as the specific hypoxic areas in the tumor center, proliferation gradients, and barriers to drug penetration, 3D A549 MCSs were utilized as a model to evaluate the therapeutic efficacy of 5-NPs-mediated 2-photon PDT. The 3D MCSs were incubated with 5-NPs for 24 h, followed by 1-photon and 2-photon Z-stack imaging microscopy (Fig. 5A and SI Appendix, Fig. S19). The 2-photon excitation displayed an advantage over the 1-photon excitation with a greater penetration depth. The 1-photon Z-stack scanning showed weaker luminescence at a depth of over 120 mm because of the strong tissue absorption of visible light. In contrast, strong red luminescence at every section of depth was observed when 2-photon excitation was employed for scanning due to the deep penetration of the 2-photon excitation wavelength in the therapeutic window for the tissue.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

Two-photon PDT toward 3D MCSs. (A) One- and 2-photon activated Z-stack images of A549 MCSs after the incubation of 5-NPs. (B) Representative confocal images of the MCSs stained with calcein-AM after various treatment. (C) Representative long-term MCSs images after various treatments. (Scale bars: 200 μm.)

The 2-photon PDT efficiency by 5-NPs was examined using a calcein-acetoxymethyl (calcein-AM) cell viability assay. Calcein-AM can distinguish live cells by reacting with the ubiquitous esterase in living cells to generate a green fluorescent product (71). For the treatment group, MCSs coincubated with 5-NPs and calcein-AM were subjected to 2-photon light irradiation, and CLSM was used to image the MCSs (Fig. 5B). No fluorescence signal was observed for the treatment groups, whereas strong green fluorescence was observed for the control group, MCSs coincubated with 5-NPs and calcein-AM but not exposed to light irradiation, and calcein-AM–treated MCSs were only subjected to light irradiation. These results indicate that neither 5-NPs nor 2-photon light irradiation cause cell death in MCSs, and activation of the 5-NPs by 2-photon light irradiation efficiently induces damage in MCSs.

The potency of 5-NPs in inhibiting the growth of MCSs by 2-photon PDT was investigated. The MCSs were incubated with 5-NPs for 1 d before being subjected to 2-photon light irradiation, then kept in the dark, and their volumes were recorded until day 6 (Fig. 5C and SI Appendix, Fig. S21). The volumes of MCSs treated with 5-NPs gradually shrank after 2-photon PDT, and the final volume was 80.4% of the original volume at day 1. In contrast, the control group and MCSs treated with 5-ALA showed marked growth in their volumes over time after the same procedure (186.8 and 181.2% relative to the original volume for the control group and 5-ALA group, respectively). These studies show that 2-photon PDT using 5-NPs can efficiently inhibit the growth of MCSs and achieve better therapeutic efficacy than 5-ALA.

Two-Photon PDT In Vivo.

The efficacy of 5-NPs as PS for 2-photon PDT was further illustrated by in vivo studies in A549 tumor xenografted nude mice. 5-NPs was intratumorally injected into mice, at a dosage of 3 mg/kg, which were irradiated with an 820-nm laser (50 mW, 1 kHz, pulse width 35 fs, 20 s/mm) 2 h after injection (group 1). As controls, group 2 received an injection of physiological saline with laser exposure, group 3 received only the injection of 5-NPs without laser exposure, and group 4 only received physiological saline injection. The effect of the 2-photon PDT treatment on tumor growth was assessed by monitoring the tumor volumes of the mice over a period of 15 d (Fig. 6 A–C). In group 1, the growth of the tumor cells was significantly inhibited, and a decrease in the tumor volume was observed at the end of the treatment, with only a small tumor mass under the skin (65.2% of original volume). In contrast, the tumor cells in groups 2, 3, and 4 grew into large solid tumors after 15 d, with more than 12-fold enhancement in the tumor volume. The body weight of the mice in each group showed a slight increase during the treatment process (Fig. 6D), indicating the noninvasiveness of the treatment regimen. The effectiveness of PDT was further evaluated by hematoxylin and eosin histological examination on tissue sections from the different treatment groups. The tumor tissue in group 1 exhibited obvious destructive damage after the treatment, whereas no morphological change was observed for the tumor tissue in the other groups (SI Appendix, Fig. S22). In addition, the primary organs in group 1 did not reveal any pathological changes compared with other groups (SI Appendix, Fig. S22). These observations demonstrate that the 5-NPs can efficiently damage tumor tissue upon 2-photon light irradiation with low systemic toxicity toward healthy tissue in vivo.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

In vivo 2-photon PDT studies. (A and B) Representative photographs of A549 tumors in mice with different treatments. (C) In vivo tumor growth inhibition curves for mice with different treatments. (D) Average body weights of tumor-bearing mice with 4 different treatments.