Metal-Organic Layer Delivers 5-Aminolevulinic Acid and Porphyrin for Dual-Organelle-Targeted Photodynamic Therapy

Abstract

The efficacy of photodynamic therapy (PDT) depends on the subcellular localization of photosensitizers. Herein, we report a dual-organelle-targeted nanoparticle platform for enhanced PDT of cancer. By grafting 5-aminolevulinic acid (ALA) to a Hf₁₂-based nanoscale metal-organic layer (Hf-MOL) via carboxylate coordination, ALA/Hf-MOL enhanced ALA delivery and protoporphyrin IX (PpIX) synthesis in mitochondria, and trapped the Hf-MOL comprising 5,15-di-p-benzoatoporphyrin (DBP) photosensitizers in lysosomes. Light irradiation at 630 nm simultaneously excited PpIX and DBP to generate single oxygen and rapidly damage both mitochondria and lysosomes, leading to synergistic enhancement of the PDT efficacy. The dual-organelle-targeted ALA/Hf-MOL outperformed Hf-MOL in preclinical PDT studies, with a 2.7-fold lower half maximal inhibitory concentration in cytotoxicity assays in vitro and a 3-fold higher cure rate in a colon cancer model in vivo.

Graphical Abstract

We report a dual-organelle-targeted nanoplatform ALA/Hf-MOL for photodynamic therapy by conjugating 5-aminolevulinic acid (ALA) to a 2D Hf₁₂-based metal-organic nanophotosensitizer (Hf-MOL). Upon endocytosis, ALA is released from Hf-MOL and accumulates in mitochondria for porphyrin IX (PpIX) synthesis, while Hf-MOL is retained in lysosomes. Light irradiation excites both PpIX and Hf-MOL to elicit potent cytotoxicity and antitumor responses.

As a noninvasive modality for cancer treatment, photodynamic therapy (PDT) exerts cytotoxicity by generating reactive oxygen species (ROS) via photoexcited photosensitizers (PSs).^[1] The PDT efficacy depends on both tissue concentrations and ROS generation efficiency of PSs,^[2] but many potent PSs with highly conjugated structures, including porphyrins and phthalocyanines, tend to have limited solubility under physiological conditions, leading to aggregation-induced quenching of photoexcited PSs and non-ideal PDT performance.^[3] Nanotechnology has offered an effective strategy to increase the accumulation and spatial separation of these highly conjugated PSs in tumors to enhance their PDT efficacy.^[4] In particular, we have recently demonstrated nanoscale metal-organic layers (MOLs) comprising PSs as a novel and efficient class of nanophotosensitizers.^[5] The 2-D MOL spatially isolates PSs by metal–oxo secondary building units (SBUs) to prevent self-quenching while maximizing the accessibility of PSs to ground-state oxygen (³O₂) for efficient energy transfer and singlet oxygen (¹O₂) generation.^[6] Since the generated ROS can only diffuse in the submicron range in biological systems,^[7] PDT efficacy strongly depends on the subcellular localization of PSs.^[8]

Subcellular organelles are cornerstones of cells and play vital roles in regulating and maintaining cell functions. Some PSs can localize in specific organelles such as endo/lysosomes, endoplasmic reticulums, and mitochondria.^[9] PDT causes damage to these organelles by excessive oxidative stress, membrane permeabilization, and expression of proapoptotic factors to initiate cell apoptosis.^[10] The destruction of specific organelles can elicit more lethal cell damage.^[11] We posited that nanophotosensitizers with the ability to target multiple organelles could synergistically enhance cell death to provide potent PDT efficacy.

5-aminolevulinic acid (ALA), an endogenous metabolite in the heme synthesis pathway, has received FDA approval for PDT of actinic keratosis and fluorescence-guided visualization of malignant tissues during glioma surgery.^[12] ALA selectively accumulates in mitochondria and produces photosensitizing protoporphyrin IX (PpIX).^[13] However, the low bioavailability and irreversible dimerization of ALA under physiological conditions limit its clinical performance.^[14]

Herein, we report a dual-organelle-targeted nanoparticle platform, ALA/Hf-MOL, for enhanced PDT of cancer. ALA/Hf-MOL was synthesized by grafting ALA to Hf-MOL via Hf-carboxylate coordination. ALA/Hf-MOL enhanced ALA delivery and PpIX synthesis in mitochondria, while trapping the Hf-MOL comprising 5,15-di-p-benzoatoporphyrin (DBP) photosensitizing ligands in endo/lysosomes. Light irradiation at 630 nm simultaneously excited PpIX and DBP to generate singlet oxygen and cause mitochondrial membrane permeabilization (MMP) and lysosomal membrane permeabilization (LMP), respectively, in cancer cells, leading to concomitant release of cytochrome C and cathepsin B, respectively (Figure 1). The dual-organelle-targeted PDT by ALA/Hf-MOL enhanced cell death in vitro and PDT efficacy in a murine colon cancer model in vivo.

Figure 1.

Dual-organelle-targeted PDT by ALA/Hf-MOL. After endocytosis, ALA is released from Hf-MOL via the phosphate concentration gradient and enables PpIX synthesis in mitochondria, while Hf-MOL is trapped in endo/lysosomes. The ROS generated from light irradiation causes membrane permeabilization of both lysosomes and mitochondria to release cathepsin B and cytochrome C, respectively, which synergistically elicits potent cell death.

Hf-MOL was synthesized by heating H₂DBP and HfCl₄ in N,N-dimethylformamide (DMF) at 80 °C with water and propionic acid (PA) as modulators.^[6b] The capping PA was replaced by trifluoroacetic acid (TFA), as confirmed by the decrease of the PA intensity in the ¹H NMR spectrum and the appearance of the TFA peak in the ¹⁹F NMR spectrum (Figure S1). ALA/Hf-MOL was synthesized by carboxylate exchange in DMF at room temperature. The ALA to DBP ratio in ALA/Hf-MOL was 74.5% by ¹H NMR (Figure S1), affording a chemical formula of Hf₁₂(μ₃-O)₈(μ₃-OH)₈(μ₂-OH)₆(DBP)₆(μ₂-ALA)_4.47(μ₂-TFA)_1.53. Transmission electron microscope (TEM) images showed that ALA/Hf-MOL retained the nanoplate morphology with a size of ~200 nm (Figure 2a,b). Dynamic light scattering (DLS) measurements of Hf-MOL and ALA/Hf-MOL showed comparable number-average sizes of 208.4 ± 4.5 nm and 202.4 ± 5.6 nm and polydispersity indices of 0.061 and 0.088, respectively (Figure 2c). Powder X-ray diffraction (PXRD) studies indicated that ALA/Hf-MOL retained the crystalline framework after ALA loading and incubation with phosphate-buffered saline (PBS) (Figure 2d).

Figure 2.

a, b) TEM images of Hf-MOL after TFA modification (a) and ALA/Hf-MOL (b). Scale bar = 200 nm. The coordination of TFA and ALA to Hf₁₂ SBUs is also shown. c, d) Number-averaged sizes (c) and PXRD patterns (d) of Hf-MOL and ALA/Hf-MOL. e) Release profiles ALA/Hf-MOL in different PBS concentrations and pH values (1x PBS ≈ 10 mM phosphate, 0.1 x PBS ≈ 1 mM phosphate). f) SOSG assays showing comparable ¹O₂ generation by Hf-MOL and ALA/Hf-MOL under 630 nm light irradiation (100 mW/cm²). All data are shown as mean ± SD with n=3.

Liquid chromatography-mass spectrometry (LC-MS) analysis indicated that ALA could be released from ALA/Hf-MOL in a pH-independent but phosphate concentration-dependent manner (Figures 2e and S2). As Hf-MOL is stable in PBS (~10 mM phosphate), the release of ALA is likely caused by the substitution of ALA on Hf₁₂ SBUs by phosphate anions in the physiological environment. Furthermore, as the cytoplasm has ~10-fold higher free phosphate concentration (~10 mM phosphate) than the interstitial fluid or plasma (~1 mM phosphate),^[15] ALA/Hf-MOL can retain ALA in the extracellular space but rapidly release ALA upon uptake by cells with a higher phosphate concentration.^[16]

The singlet oxygen sensor green (SOSG) assay showed similar ¹O₂ generation by Hf-MOL and ALA/Hf-MOL under 630 nm light irradiation, indicating that conjugation of ALA to Hf-MOL did not impact ¹O₂ generation from Hf-MOL (Figure 2f). However, cell viability assays revealed that ALA/Hf-MOL with 630 nm light irradiation [denoted ALA/Hf-MOL(+)] exhibited an IC₅₀ value of 0.28 ± 0.13 μM in CT26 cells, which was 2.7-fold lower than that of Hf-MOL(+) (0.76 ± 0.13 μM, Figure 3a). ALA(+) showed negligible cytotoxicity. No dark toxicity was observed in all treatment groups (Figure S3). A physical mixture of ALA and Hf-MOL showed weakly additive phototoxicity with an IC₅₀ value of 0.63 ± 0.09 μM (Figure 3a). As Hf-MOL and ALA/Hf-MOL show comparable cellular uptake and ROS generation (Figure S4), we hypothesize that the synergistic PDT enhancement by ALA/Hf-MOL over Hf-MOL is attributed to improved ALA delivery and PpIX synthesis.

Figure 3.

a) Cell viability assays of Hf-MOL, a physical mixture of ALA and Hf-MOL, and ALA/Hf-MOL in CT26 cells with a total light dose of 90 J/cm² at 630 nm (mean ± SD, n=3). b-d) Schematic showing PpIX quantification (b) in CT26 cells (c) and tumors (d) (mean + SD, n=3). *, p < 0.05; **, p < 0.01; ****, p < 0.0001.

We used LC-MS to quantify PpIX in cultured CT26 cells and subcutaneous CT26 tumors after different treatments (Figures 3b and S5).^[17] ALA/Hf-MOL showed significantly higher PpIX accumulation with 2.3-fold and 1.5-fold higher PpIX content than free ALA in vitro and in vivo, respectively (Figure 3c,d). Flow cytometric analysis of mitochondria isolated from cultured CT26 cells showed that ALA/Hf-MOL treatment led to >4.3-fold PpIX accumulation than free ALA or PBS treatment (Figure S6). These results confirm the improved ALA delivery and PpIX synthesis by ALA/Hf-MOL.

We next used confocal laser scanning microscopy (CLSM) to study the subcellular localization of ALA/Hf-MOL and to observe the dual-organelle targeting in real time. ALA/Hf-MOL-treated CT26 cells showed smeared red signals (PpIX) in addition to distinct puncta (Hf-MOL), suggesting the conversion of ALA to PpIX and retention of Hf-MOL in lysosomes (Figure S7). We then co-stained CT26 cells with LysoTracker Green DND-26 and MitoSOX Superoxide Indicator Red to monitor the real-time status of lysosomes and mitochondria, respectively. Using the built-in 630 nm monochromatic laser in the confocal microscope, we performed in situ PDT while recording real-time images of subcellular organelles (Figures 4a–d, Movies S1–S4). The fluorescence signals of lysosomes and mitochondria were retained in the cells treated with PBS(+) or ALA(+), indicating no cellular damage. Hf-MOL(+) moderately compromised the integrity of lysosomes and mitochondria, but neither organelle reached more than 40% damage after 15 minutes of PDT. In stark contrast, ALA/Hf-MOL(+) induced rapid depolarization of both lysosomes and mitochondria, with fluorescence signals decreasing to <50% within 2 minutes of light exposure (Figure 4e,f).

Figure 4.

a)-d) Real-time imaging of lysosomes (green) and mitochondria (red) in CT26 cells treated with (a) PBS(+), (b) ALA(+), (c) Hf-MOL(+), and (d) ALA/Hf-MOL(+). A dose of 1 μM DBP and a fluence of 100 mW/cm² laser at 630 nm were used. From left to right, the cells were irradiated for 1, 2, 3, and 5 minutes, respectively. All scale bars are 5 μm. e), f) Time-dependent retention (mean ± SD, n=3) of (e) lysosomal and (f) mitochondrial signals in different treatment groups.

To investigate the effect of dual-organelle-targeted PDT on MMP and LMP, we stained mitochondria and lysosomes for viability markers and organelle contents. ^[18] ALA/Hf-MOL(+) caused the strongest depolarization of mitochondrial membrane potential as visualized by increased monomer signals in JC-1 assay and the higher cytochrome C release from mitochondria (Figures 5a,b, S8). The reduced fluorescence of acridine orange (AO) in the Hf-MOL(+) and ALA/Hf-MOL(+) groups indicated the induction of LMP, suggesting the dysregulation of lysosome pH due to photodamage (Figure 5c). The release of cathepsin B from lysosomes was visualized as scattered fluorescence signals in the cytosol (Figure 5d). The ROS level in ALA/Hf-MOL(+)-treated cells was 2.3-fold higher than Hf-MOL(+), which further supported the photosensitization of the synthesized PpIX and the increased oxidative stress from enhanced MMP and LMP (Figures 5e and S9). Apoptotic cells were labeled by Annexin-V, and the plasma membrane permeabilization was stained with propidium iodide (PI). CLSM and flow cytometry studies showed that ALA/Hf-MOL(+) treatment gave a higher Annexin-V⁺/PI⁺ population than Hf-MOL(+) treatment (Figures 5f and S10). ALA/Hf-MOL(+) also upregulated calreticulin (CRT) on cell surfaces for enhanced immunogenic cell death (Figures S11 and S12). Taken together, ALA/Hf-MOL(+) synergistically disrupts lysosomes and mitochondria to enhance cell death, thus providing a novel platform for dual-organelle-targeted PDT.

Figure 5.

a) Mitochondrial potential depolarization by JC-1 assay. Red and green channels indicate J-aggregate and monomer forms of JC-1 molecules, respectively. b) Cytochrome C (green) release from mitochondria. c) LMP visualized by AO assay. d) Cathepsin B (green) release from lysosomes by CV-Cathepsin-B assay. e) 2’,7’-dichlorodihydrofluorescein diacetate assay showing total ROS (green) generation. f) Apoptosis assay by Annexin-V (green) and PI (red) staining. Nuclei were visualized by Hoechst 33342 (blue). The different treatments are shown at the top. A dose of 5 μM DBP and a light dose of 60 J/cm² at 630 nm were used in CT26 cells. All scale bars are 20 μm.

The in vivo antitumor efficacy was then evaluated in subcutaneous CT26 tumor-bearing BALB/c mice. Compared to PBS(+) control group, ALA(−), Hf-MOL(−), and ALA/Hf-MOL(−) groups showed negligible antitumor efficacy with tumor growth inhibition (TGI) values of <2% (Table S1). ALA(+) moderately inhibited tumor growth with a TGI value of 54.1%. ALA/Hf-MOL(+) significantly improved tumor regression with a TGI value of 99.2%, which was higher than the TGI value of 89.6% for Hf-MOL(+) (Figure 6a,b). ALA/Hf-MOL(+) was also significantly more effective than a physical mixture of ALA and MOL plus light irradiation (TGI = 84.8%, Figure S13). These treatments had minimal impact on mouse health (Figures S13 and S14). Hf-MOL(+) eradicated tumors in only 20% mice, while ALA/Hf-MOL(+) eradicated tumors in 60% mice.

Figure 6.

a, b) Tumor growth curves (a) and tumor weights (b) of CT26 tumor-bearing BALB/c mice (n=5) after PDT treatment with PBS(−), ALA(+), Hf-MOL(+), or ALA/Hf-MOL(+) (black and red arrows refer to particle administration and light irradiation of 90 J/cm² at 630 nm, respectively). c) Representative contour plots with adjunct histograms showing functional staining of both lysosome (x-axis, LysoTracker Green DND-26) and mitochondria (y-axis, MitoSOX Superoxide Indicator Red) of treated CT26 tumors by flow cytometry. d) Normalized mean fluorescence intensities (MFI) of lysosomal and mitochondrial staining of viable CT26 tumor cells showing significant dual-organelle damage induced by ALA/Hf-MOL(+) in vivo (n=3). e, f) Representative images of H&E staining (e) and TUNEL staining (f) of excised CT26 tumors (Scale bars = 100 μm). All data are shown as mean + SD. *, p < 0.05; **, p < 0.01; ***, p < 0.001.

We excised the tumors immediately after PDT treatment and stained viable tumor cells with functional lysosome and mitochondria markers to verify the dual-organelle disruption (Figures 6c and S15). Compared to ALA(+), ALA/Hf-MOL(+) showed 5.5-fold and 1.9-fold decreases in viable mitochondria and lysosomes, respectively. (Figure 6d). Hematoxylin and eosin (H&E) staining showed that ALA/Hf-MOL(+)-treated tumors exhibited the least nucleus density (Figure 6e). The effective cancer cell killing was also corroborated by the terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining; ALA/Hf-MOL(+) showed more DNA fragments than Hf-MOL(+) and other control groups (Figures 6f and S16).

In summary, we have designed a dual-organelle-targeted nanophotosensitizer by conjugating ALA to the SBUs of Hf-MOL. ALA/Hf-MOL enhanced ALA delivery and PpIX synthesis in mitochondria while retaining the photosensitizing Hf-MOL in lysosomes. Dual-organelle disruption induced synergistic PDT enhancement for superb anticancer efficacy in vitro and in vivo. Thus, MOLs provide a unique molecular nanomaterial platform to develop dual-organelle-targeted cancer therapy.