Nanoscale Metal–Organic Framework Confines Zinc-Phthalocyanine Photosensitizers for Enhanced Photodynamic Therapy

The performance of photodynamic therapy (PDT) depends on the solubility, pharmacokinetic behaviors, and photophysical properties of photosensitizers (PSs). However, highly conjugated PSs with strong reactive oxygen species (ROS) generation efficiency tend to have poor solubility and aggregate in aqueous environments, leading to suboptimal PDT performance. Here, we report a new strategy to load highly conjugated but poorly soluble zinc-phthalocyanine (ZnP) PSs in the pores of a Hf12-QC (QC = 2″,3′-dinitro-[1,1’:4′,1”;4″,1’”-quaterphenyl]-4,4’”-dicarboxylate) nanoscale metal–organic framework to afford ZnP@Hf-QC with spatially confined ZnP PSs. ZnP@Hf-QC avoids aggregation-induced quenching of ZnP excited states to significantly enhance ROS generation upon light irradiation. With higher cellular uptake, enhanced ROS generation, and better biocompatibility, ZnP@Hf-QC mediated PDT exhibited an IC50 of 0.14 μM and achieved exceptional antitumor efficacy with >99% tumor growth inhibition and 80% cure rates on two murine colon cancer models.

P hotodynamic therapy (PDT) destroys a malignant tumor while sparing surrounding normal tissues by localizing a photosensitizer (PS) in the tumor and irradiating the tumor with visible or near-infrared light to produce cytotoxic reactive oxygen species (ROS). 1−4 The clinical utility of PDT is limited by tissue penetration of light, localization of the PS in the tumor, and the solubility and photophysical properties of the PS. 5,6 For example, clinically used porphyrin-based PSs often cause phototoxicity side effects due to their strong absorption in the visible spectrum and retention in the skin. 7 Phthalocyanine (Pc) PSs present a promising alternative due to their very strong absorption in 650−800 nm and weak absorption in 400−600 nm, allowing for effective treatment of tumors with low PS doses and reduced phototoxicity. 8 Metalation of Pcs with diamagnetic ions (e.g., Zn 2+ , Si 4+ , Al 3+ ) increases triplet state yields and lifetimes to enhance the generation of cytotoxic singlet oxygen ( 1 O 2 ). 9−11 Despite their improved photophysical properties, Pcs have not been widely used for PDT due to their limited synthetic accessibility and their strong tendency to aggregate in biological media. 11 Pcs have been functionalized with ionic or hydrophilic groups in their peripheral positions to increase aqueous solubility or coordinate with bulky metal complexes (axial functionalization) to prevent π−π stacking. 12 However, the introduction of ionic or hydrophilic groups to Pcs can adversely impact their cellular uptake while axial functionalization of Pcs is limited to a few nontoxic high-valent metals such as Si 4+ . 13,14 An alternative strategy to address the solubility and aggregation issues of Pcs is through their encapsulation in or conjugation to liposomes, micelles, or other nanoparticles (NPs). 15−19 In particular, micelles have been widely investigated as a delivery vehicle for lipophilic conjugated Pcs with superb photophysical properties. 20,21 Nanoscale metal−organic frameworks (nMOFs) have recently provided an excellent strategy to deliver porphyrin, chlorin, and bacteriochlorin PSs for PDT. 22−28 With structural tunability, rigidity, and porosity, nMOFs can efficiently load PSs via direct incorporation as bridging ligands, postsynthetic ligand exchange, postsynthetic surface modification, and physical loading into pores. 29−35 These strategies allow isolation or confinement of lipophilic PSs in rigid nMOF structures to reduce aggregation, improve cellular uptake, and reduce photodegradation. 36−41 We hypothesized that nMOFs could also be used to encapsulate Pcs to enhance their PDT efficacy.
Hf-QC was synthesized through a solvothermal reaction between HfCl 4 and H 2 QC in a mixture of N,N-dimethylformamide (DMF), acetic acid, and water at 80°C (Figure 2a Figure S15), which matched well with the modeled structure for Zr 12 -QPDC. 1 H nuclear magnetic resonance (NMR) analysis of digested Hf-QC showed an acetate (OAc) modulator to QC linker ratio of 0.11:1, corresponding to approximately 0.5 missing linkers per SBU ( Figure S11). Thermogravimetric analysis (TGA) of Hf-QC showed a weight loss of 39.4% in the 300−800°C region, matching the expected value of 37.9% for the Hf-QC with a 0.5 linker defect per SBU ( Figure S14). On the basis of these results, Hf-QC was formulated as Hf 12 ZnP@Hf-QC was synthesized by heating a mixture of ZnP and Hf-QC in DMF at 70°C for 24 h. Loading of ZnP in ZnP@Hf-QC was confirmed by the presence of characteristic Ultraviolet−visible (UV−vis) and infrared (IR) peaks for ZnP (Figure 3a, Figure S18). UV−vis spectroscopy and inductively coupled plasma-mass spectrometry (ICP-MS) showed the loading of 13.6 wt % ZnP in ZnP@Hf-QC, corresponding to a ZnP to Hf 12 SBU ratio of 0.68:1. 1 H NMR analysis of digested ZnP@Hf-QC showed that the OAc modulator to QC linker ratio was maintained after ZnP loading ( Figure S19). TGA of ZnP@Hf-QC showed a weight loss of 36.3% in the 300−800°C region, which matched well with the expected value of 34.1% for physical loading of ZnP in the nMOF pores and confirmed the ratio of ZnP to Hf ( Figure S14). On the basis of these results, ZnP@Hf-QC was formulated as (ZnP) 0.68 @ Hf 12 (μ 3 -O) 8 (μ 3 −OH) 8 (μ 2 −OH) 6 (QC) 8.5 (OAc).
TEM and DLS showed that ZnP@Hf-QC retained the hexagonal nanoplate morphology and size (175.8 ± 5.6 nm) of   Hf-QC (Figures 2e and 3c, Figure S15). HRTEM images ( Figure S15) and PXRD patterns (Figure 3b) of ZnP@Hf-QC supported the maintenance of Hf-QC structure after ZnP loading. ZnP@Hf-QC displayed a slightly more negative ζ potential of −24.0 ± 1.5 mV compared to Hf-QC at −22.1 ± 0.7 mV ( Figure S16), consistent with loading negatively charged ZnP into the pores of Hf-QC. The presence of QC and ZnP was confirmed by its characteristic UV−vis and 1 H NMR signals in digested ZnP@Hf-QC (Figures S11, S19). The stability of ZnP@Hf-QC was demonstrated by PXRD and DLS after incubation in phosphate-buffered saline (PBS) or Dulbecco's Modified Eagle Medium (DMEM) at 37°C for 24 h (Figure 3b,c).
We then examined apoptosis and immunogenic cell death of CT26 cells after PDT by CLSM and flow cytometry. CT26 cells treated with ZnP@Hf-QC(+) showed upregulation of phosphatidylserine by Annexin V staining on cell membranes and colocalization of propidium iodine (PI) and Hoechst 33342 (Figure 5b, Figures S26, S27). These results indicated apoptosis and compromised membrane functions for ZnP@  Hf-QC(+) treated CT26 cells, which were absent in control groups. Calreticulin (CRT) staining revealed enhanced immunogenic cell death and surface translocation of CRT signals in ZnP@Hf-QC(+) group (Figure 5c, Figures S30,  S31). Taken together, ZnP@Hf-QC(+) not only killed cancer cells more effectively but also induced immunogenic cell death to expose tumor antigens and danger signals for immune activation. 45,46 We evaluated antitumor efficacy of ZnP@Hf-QC(+) on two subcutaneous murine colon cancer models with CT26 tumors on BALB/c mice and MC38 tumors on C57BL/6 mice. Hf-QC and ZnP@Hf-QC were pegylated before intravenous administration. The mice were injected with PBS, ZnP, Hf-QC, or ZnP@Hf-QC via tail veins at an equivalent ZnP dose of 50 nmol (equivalent Hf dose of 0.88 μmol). Twelve hours post injection, the mice were anesthetized, and tumor areas were irradiated with 700 nm LED with a total light flux of 60 J/ cm 2 (100 mW/cm 2 ).
Compared to PBS(+), Hf-QC(+) had little effect on tumor growth with minimal tumor growth inhibition indices (TGIs) of 17.8% and 7.4% for CT26 and MC38 tumors, respectively. ZnP(+) moderately slowed tumor growth with TGI values of 41.3% and 41.4% for CT26 and MC38 tumors, respectively. ZnP@Hf-QC(+) treatment showed excellent antitumor efficacy with >99% TGIs and 80% cure rates for both CT26 and MC38 tumors (Figure 6a,b, Figures S32, S33, S38, S39). H&E and TUNEL staining revealed severe apoptosis/necrosis and infiltration of inflammatory cells in tumor regions in the ZnP@Hf-QC(+) group (Figure 6c, Figure S43). Several mice in the ZnP(+) and ZnP(−) groups showed weight loss, pulmonary edema, and local liver inflammation ( Figures S35,  S36), likely caused by aggregation of ZnP into large particles in vivo. In comparison, although ZnP@Hf-QC were observed to accumulate in spleens and livers similar to other nanoparticles 47,48 (Figures S36, S37) by tumor tissue sections, mice treated with ZnP@Hf-QC with or without light irradiation showed steady body weights (Figures S41, S42). ZnP@Hf-QC and its aggregate were not observed in lungs and minimal abnormities were observed in the major organs of ZnP@Hf-QC treated mice compared to PBS control ( Figures S34, S40). The different in vivo behaviors between ZnP and ZnP@Hf-QC showed that the nMOF pore loading strategy provides an efficient, safe, and biocompatible approach to deliver PSs with unfavorable solubility and pharmacokinetic properties.
In summary, we developed an nMOF confinement strategy to isolate ZnP PSs and prevent their aggregation and excited state quenching. As a result, the isolated PSs in ZnP@Hf-QC efficiently absorbed light to significantly enhance 1 O 2 generation and efficiently kill cancer cells. ZnP@Hf-QC mediated PDT effectively eradicated/regressed colorectal cancers in two mouse models. The confinement of photosensitizers in nMOF pores provides a new strategy to unleash the potential of poorly soluble, highly conjugated PSs in PDT.