Fluorescence and Magnetic Resonance Dual-Modality Imaging-Guided Photothermal and Photodynamic Dual-Therapy with Magnetic Porphyrin-Metal Organic Framework Nanocomposites

Phototherapy shows some unique advantages in clinical application, such as remote controllability, improved selectivity, and low bio-toxicity, than chemotherapy. In order to improve the safety and therapeutic efficacy, imaging-guided therapy seems particularly important because it integrates visible information to speculate the distribution and metabolism of the probe. Here we prepare biocompatible core-shell nanocomposites for dual-modality imaging-guided photothermal and photodynamic dual-therapy by the in situ growth of porphyrin-metal organic framework (PMOF) on Fe3O4@C core. Fe3O4@C core was used as T2-weighted magnetic resonance (MR) imaging and photothermal therapy (PTT) agent. The optical properties of porphyrin were well remained in PMOF, and PMOF was therefore selected for photodynamic therapy (PDT) and fluorescence imaging. Fluorescence and MR dual-modality imaging-guided PTT and PDT dual-therapy was confirmed with tumour-bearing mice as model. The high tumour accumulation of Fe3O4@C@PMOF and controllable light excitation at the tumour site achieved efficient cancer therapy, but low toxicity was observed to the normal tissues. The results demonstrated that Fe3O4@C@PMOF was a promising dual-imaging guided PTT and PDT dual-therapy platform for tumour diagnosis and treatment with low cytotoxicity and negligible in vivo toxicity.


Synthesis and Characterization of Fe 3 O 4 @C and Fe 3 O 4 @C@PMOF. Fe 3 O 4 @C micro-structure
was fabricated by one-pot solvothermal strategy according to previous report 31 . Fe 3 O 4 @C nanoparticles were dispersed in the DMF suspension of ZrCl 4 in a hydrothermal procedure for 30 min, and then DMF solution of TCPP was added into the mixture. Fe 3 O 4 @C@PMOF was prepared by the in situ self-assembly of PMOF on the surface of Fe 3 O 4 @C to obtain the core-shell nanocomposites. The proposed method was time-saving and efficient compared with the layer-by-layer self-assembly of MOF 32,33 . Therefore, a simple strategy was developed to prepare the Fe 3 O 4 @C@PMOF composite. Transmission electron microscopy (TEM) images of Fe 3 O 4 @C and Fe 3 O 4 @C@ PMOF revealed their well-defined micro-structure with the average diameter of 80 and 95 nm, respectively ( Fig. 1a and b). Moreover, ca 7.5 nm PMOF layer was successfully coated on Fe 3 O 4 @C to form the Fe 3 O 4 @C@ PMOF hybrid nanocomposites. Fe 3 O 4 @C nanoclusters consisted of numerous 10 nm Fe 3 O 4 nanoparticles as illustrated in Fig. 1a and b, different to the solid Fe 3 O 4 structure of ferumoxsil and ferumoxide [23][24][25] . Thus, improved T 2 -MR imaging efficiency is expected because of the altered proton relaxation effect of the nanocluster structure. The composites less than 100 nm pass through the tumor microenvironment easily and remain for a long time before blood clearance 34 . Moreover, the size of Fe 3 O 4 @C@PMOF was suitable for PDT because the diffusion length of singlet oxygen ( 1 O 2 ) was 90− 120 nm in aqueous environment and 20− 220 nm inside cells 28 .
Dynamic light scattering analysis indicated that Fe 3 O 4 @C@PMOF had a relatively narrow size distribution and was well dispersed for real application ( Supplementary Fig. S1). − 15.7 and − 3.39 mV of zeta potentials were observed for Fe 3 O 4 @C and Fe 3 O 4 @C@PMOF ( Supplementary Fig. S2). Therefore, the carbon layer of Fe 3 O 4 @C was oxidized by H 2 O 2 under solvothermal procedure to form abundant carboxylic groups, which were then used to coordinate with Zr 4+ ions. The PMOF layer formed through the coordination between Zr 4+ ions and TCPP as the zeta potential changed from original − 15.7 mV to − 3.39 mV. The near-neutral surface of the nanocomposites makes them excellent candidates for in vivo applications 35 .
The magnetic properties of Fe 3 O 4 @C and Fe 3 O 4 @C@PMOF were characterized by Vibrating Sample Magnetometer (VSM) at the field of ± 20 kOe (Fig. 1c). The saturation magnetization of Fe 3 O 4 @C was 39.8 emu g −1 . The magnetic hysteresis curve was retained in Fe 3 O 4 @C@PMOF with the saturation magnetization of 24.5 emu g −1 . Both Fe 3 O 4 @C and Fe 3 O 4 @C@PMOF were well dispersed, but they were collected easily with external magnet and the solution became transparent (Inset in Fig. 1c). Thus, the great MR imaging potential was revealed from Fe 3 O 4 @C@PMOF.
Powder X-ray diffraction (XRD) patterns of Fe 3 O 4 @C, PMOF, and Fe 3 O 4 @C@PMOF were recorded (Fig. 1d). The peaks observed at 30.1, 35.3, 42.9, 53.5, 57.0, and 62.5° were assigned to (220), (311), (400), (422), (511) and (440) planes of cubic structure of Fe 3 O 4 crystal (JCPDS No.75-1609). The simultaneous existence of the characteristic peaks of Fe 3 O 4 and PMOF in its XRD pattern indicates the successful formation of Fe 3 O 4 @C@PMOF nanocomposites. The formation was also confirmed by Fourier transform infrared spectroscopy (FT-IR) with the characteristic peak at 964.45 cm −1 assigned to the pyrrole ring of the ligand, TCPP ( Supplementary Fig. S3). Thermogravimetric analysis (TGA) results revealed that Fe 3 O 4 @C was highly stable in the tested temperature ( Supplementary Fig. S4 (Fig. 2a). This feature provided efficient photothermal capacity. When PMOF shell was covered, a strong absorption peak emerged at 416 nm for Soret band (Fig. 2b) and four peaks at 517, 554, 583, and 634 nm were observed for Q band as the typical character of porphyrin (inset of Fig. 2b) 16 . Thus, Fe 3 O 4 @C@ PMOF was potential for PDT because of its matched NIR absorption 36 . Single emission peak was observed at 668 nm from Fe 3 O 4 @C@PMOF with 553 nm excitation (Fig. 2c). Strong NIR emission and long Stocks shift led to a high signal/noise ratio for fluorescent image because of the low auto-fluorescence and scattering light from biological tissue 37,38 . The optical property of Fe 3 O 4 @C@PMOF illustrated its PTT and PDT potential and fluorescence imaging capacity.
T 2 -weighted imaging was tested to validate the MR contrast potential of Fe 3 O 4 @C@PMOF nanocomposites. The relaxation rates vary linearly with increased Fe concentration (Fig. 2d). The darkening T 2 MR imaging at different Fe concentrations confirmed the T 2 -weighted MR efficiency (Inset of Fig. 2d). Fe content of Fe 3 O 4 @C@ PMOF was tested by inductively coupled plasma-atomic emission spectroscopy, and the slope in Fig. 2d was the r 2 value, which was 72.6 mM −1 s −1 (r 1 = 1.23 mM −1 s −1 , r 2 /r 1 = 59.0, Supplementary Fig. S5). The r 2 value was higher than that of commercial magnetic nanoparticles (10 nm, r 2 = 59.91 ± 6 mM −1 s −1 ) 39 due to the integration of numerous 10 nm Fe 3 O 4 nanoparticles in a carbon layer for efficient T 2 -contrast effect 21 . The r 2 /r 1 ratio was 59.0, and therefore Fe 3 O 4 @C@PMOF showed the potential for T 2 -weighted MR imaging.
In vitro photothermal and photodynamic properties of Fe 3 O 4 @C@PMOF. PTT efficiency of Fe 3 O 4 @C@PMOF was evaluated by measuring the temperature change under 808 nm NIR laser irradiation (Fig. 3a). After a 5 min of laser exposure, the solutions containing Fe 3 O 4 @C and Fe 3 O 4 @C@PMOF respectively were rapidly heated to higher than 50 °C because of the high NIR absorptivity of Fe 3 O 4 , while the PBS solution showed less photo-generated heating efficiency ( Supplementary Fig. S6). However, when Fe 3 O 4 @C@PMOF solution was exposed to 655 nm laser for 30 min, the temperature gave rise to 29 °C, which was negligible to damage cancer cells ( Supplementary Fig. S7). Infrared thermal photographs of Fe 3 O 4 @C@PMOF solution before and after 808 nm irradiation illustrated the photothermal capacity directly (Fig. 3b).
Singlet oxygen ( 1 O 2 ) is the electronic excited state of molecular oxygen and highly reactive in the oxidization damage of biological tissues 4,28 . 9, 10-anthracenediyl-bis (methylene) dimalonic acid (ABDA) was used as indicator to verify the 1 O 2 generation capacity of Fe 3 O 4 @C@PMOF because ABDA can react with 1 O 2 irreversibly 40 . The reaction was monitored by the decreased ABDA absorption at 379 nm. The absorption spectra of ABDA in Fe 3 O 4 @C@PMOF were recorded at different exposure times (Fig. 3c), and the variation of absorbance was illustrated in Fig. 3d. The rapid decrease of ABDA absorption represented fast 1 O 2 generation by where A refers to the absorption of ABDA at 379 nm, c is fitting parameter, and t is irradiation time. The 1 O 2 generation rate (v) was calculated as 0.055 min −1 for Fe 3 O 4 @C@PMOF. The 1 O 2 quantum yield is calculated with the equation 40 : where h is Plank constant, c is velocity of light, λ is the wavelength of laser, n ABDA is the amount of ABDA consumed by 1   illness and activity changes was observed from the mice, which also showed the same weight trend to the mice in control group within 3 weeks (Fig. 4b).
In vivo dual-modality imaging of mice with Fe 3 O 4 @C@PMOF as probe and clearance study. To prove the in vivo efficiency of dual-modality imaging, Fe 3 O 4 @C@PMOF nanocomposite was injected  intravenously into a 20 g healthy nude mouse. In vivo T 2 MR image was recorded ( Supplementary Fig. S10). The liver region was darkening after being injected for 22 h. The same result was observed in fluorescence imaging simultaneously. The fluorescent spot was also observed in lymph possibly because of the high affinity between porphyrin and lymph node. Thus, the nanocomposites did transfer not only through blood circulation but also participate in the lymph circulation simultaneously. Then, it accumulated in liver. Finally, most of Fe 3 O 4 @C@ PMOF was excreted through excrement within 8 days with the similar metabolic pathway to coproporphyrin 42 .
The images in Fig. 5 illustrated the whole metabolism process of Fe 3 O 4 @C@PMOF in nude mice after intravenously injected at different time.
To verify the accumulation of Fe 3 O 4 @C@PMOF in tumor site, MCF-7 tumor-bearing nude mice were selected as model (Fig. 6a). After being intravenously injected with Fe 3 O 4 @C@PMOF for 22 h, fluorescent signal was localized mainly in the liver region than the other organs. The tumor region was lightened slowly and became the brightest tissue of the mice via enhanced permeability and retention (EPR) effect after 26 h. In vivo T 2 MR image was also recorded simultaneously (Fig. 6c). The dramatic dimming was observed at tumor area and also demonstrated the high tumor uptake of Fe 3 O 4 @C@PMOF, which was the same as the fluorescence imaging result.  To further evaluate the tissue distribution of the nanocomposites, major organs dissected from the mice were harvested and imaged ex vivo at 26 h post injection (Fig. 6b). The highest fluorescent intensity of cancer tissue indicated tumor-targeted delivery for imaging, PDT and PTT. The intestine was also lightened and indicated the potential metabolic pathway that Fe 3 O 4 @C@PMOF was excreted through excrement. To further study the excretion of the nanocomposites, high level of Fe was detected in feces of mice after injected with Fe 3 O 4 @C@PMOF (Fig. 6d). Thus, both MR and fluorescent imaging results and ex vivo fluorescence images of the tissues confirmed the efficient tumor location of our nanocomposites.
In vivo photothermal and photodynamic synergetic therapy. Motivated by its high tumor accumulation, Fe 3 O 4 @C@PMOF was used for in vivo imaging-guided tumor treatment. Nude mice with subcutaneous MCF-7 breast cancer xenografts were selected as model. For in vivo monitoring of the photothermal effect generated from Fe 3 O 4 @C@PMOF, the temperature change of the tumor site was recorded with infrared camera under irradiation of 808 nm laser. To study the in vivo synergetic efficiency of PTT and PDT, MCF-7 tumor-bearing mice were randomly divided into five groups. The group injected with saline was regarded as the negative control. All of the other four groups were injected with Fe 3 O 4 @C@PMOF (10 mg kg −1 ). The injected mice without any irradiation were used as the positive control. After the injection for 26 h, the irradiation was carried out. In PTT group, the mice after injection with Fe 3 O 4 @C@PMOF were irradiated with 808 nm laser for 10 min; the mice in PDT group were subjected to the irradiation of 655 nm laser for 10 min. In the PTT-PDT co-therapy group, the mice were firstly irradiated with 808 nm laser for 10 min, followed by the irradiation of 655 nm laser for 10 min. Upon 808 nm laser irradiation, the temperature of the tumor site in the PTT and PTT-PDT co-therapy groups rapidly increased to higher than 50 °C, which is high enough to ablate the cancer cells. For the negative control groups, the tumor tissues didn't show any significant temperature elevation (Fig. 7a). The injection and irradiation were repeated every two days within 8 days. Tumor sizes and body weights of the mice were monitored every two day after different treatments ( Fig. 7b and Supplementary Fig. S11). The size of tumors was normalized to their initial size. Experimental results indicated that the tumor sizes of mice in both negative and positive control groups became larger and larger. In contrast, the tumor growth of the mice after single PTT or PDT inhibited remarkably within 8 days. The PTT-PDT co-therapy group exhibited the highest therapeutic efficacy compared with that of the single PTT or PDT groups. (Fig. 7b and d).
Apoptotic and necrotic tumor cells were tested to validate the photo-therapy efficiency (Fig. 8b). Intensive necrosis area was markedly stained by eosin in the dominated tumor section of PTT-PDT treatment group. The results clearly demonstrated that the synergetic therapeutic efficacy of PTT and PDT was superior to any single therapy. The mice after PTT and PDT treatment behaved normally and the weight didn't decrease remarkably ( Supplementary Fig. S11). No pathological changes was noticed for mice after 8 days after PTT and/or PDT, as revealed by hematoxylin and eosin (H&E)-stained major organ slices of the mice because of the excellent biocompatibility of the nanocomposites and the remote controllability, improved selectivity, and safety of phototherapy.

Discussion
In summary, we developed Fe 3 O 4 @C@PMOF for fluorescence-magnetic resonance dual-modality imaging-guided photothermal and photodynamic cancer dual-therapy by in situ growth PMOF shell on Fe 3 O 4 @C core for the first time. The Fe 3 O 4 @C@PMOF nanocomposites featured with some unique advantages over common therapy agents, such as high biocompatibility and stability, and simple self-assembly process to form the MOF shell with the abundant carboxylic groups in the periphery of Fe 3 O 4 @C to interact with Zr 4+ ions. Effective photothermal and photodynamic therapy of tumors was achieved by passive tumor targeting and excellent photophysical properties of the nanocomposites. Besides, the improved safety and low bio-toxicity of phototherapy, different to chemotherapy, had no damage to normal tissues because of the controllable and local irradiation of the laser. Both irradiation and emission of photo-therapy and fluorescence imaging were around infrared or NIR region, so high penetration depth achieved the efficient photo-therapy and imaging. Fe 3 O 4 @C@PMOF with low biotoxicity shows promise for future clinical translation as validated by the dual-modality imaging-guiding synergetic therapy. The results demonstrate the availability of as-synthesized MCTPs on tumor and illustrate great potential in tumor diagnosis and treatment.

Methods
Animal experiments. To validate the dual-modality imaging-guided photothermal and photodynamic dual-therapy, we used female BALB/c-nu mice with the body weight of 18-22 g as model. The mice were obtained from the Institute of Hematology & Hospital of Blood Disease, Chinese Academy of Medical Sciences & Peking Union Medical College with the license No. SCXK-2014-0013, Tianjin, China. The mice were housed one per cage in a specific pathogen-free environment and had free access to standard solid pellet food (HFK, Beijing, China) and water. We confirmed that all experimental protocols were approved by the Institutional Animal Care Committee of Nankai University and all methods were carried out in accordance with the relevant guidelines and regulations from the Institutional Animal Care Committee of Nankai University. Synthesis of Fe 3 O 4 @C nanospheres. Fe 3 O 4 @C micro-structure was fabricated by one-pot solvothermal strategy according to previous report 31 . Briefly, 0.08 g of ferrocene was dissolved in 32 mL acetone. 0.4 mL of 30% hydrogen peroxide was added and the mixture solution was transferred to a 50 mL Teflon-lined stainless autoclave. The mixture was then kept at 210 °C for 24 h. After the autoclave was cooled to room temperature, the products were collected by a magnet after ultrasonication for 20 min. Black solid was washed with acetone and ethanol three times and dried in vacuum at 40 °C for 24 h to obtain Fe 3 O 4 @C nanospheres. was added to the suspension of 10 mg Fe 3 O 4 @C nanospheres in 5 mL DMF. The mixture was transferred to a 25 mL three-necked flask and kept at 120 °C with vigorous stirring for 30 min. Then, 100 μ L of TCPP DMF solution (38 μ mol L −1 ) was added drop-wise into the mixture in 2 min. The reaction proceeded for another 3 h at 120 °C. The product was collected magnetically and washed with DMF and ethanol and dried for further use.

Preparation of Fe
In vitro singlet oxygen generation. The singlet oxygen ( 1 O 2 ) generation capacity of Fe 3 O 4 @C@PMOF was tested using 9, 10-anthracenediyl-bis (methylene) dimalonic acid (ABDA) method after laser irradiation. 1 O 2 generated from Fe 3 O 4 @C@PMOF oxidizes ABDA to decrease its UV absorption. PBS (pH 7.4) containing 20 μ mol L −1 Fe 3 O 4 @C@PMOF and 200 μ mol L −1 ABDA was therefore irradiation with 655 nm laser (0.3 W cm −2 ). The absorbance spectra of solution were recorded at different time points. The stability of ABDA to PBS, light, and single Fe 3 O 4 @C@PMOF was also tested as control.
Cytotoxicity and photoxicity of Fe 3 O 4 @C@PMOF. The cytotoxicity and photoxicity of Fe 3 O 4 @C@ PMOF were evaluated by the viability of MCF-7 cells using a standard methyl thiazolyl tetrazolium (MTT) assay. Four groups were tested separately with different treatments. Briefly, the cells were incubated to 96-well culture plates at a density of 5 × 10 3 cells per well in culture medium. Fe 3 O 4 @C@PMOF was introduced to the medium at the concentration between 0 and 400 μ g mL −1 and incubated for 8 h after MCF-7 cells reached 90-95% confluences. Then, the PDT and PTT groups were under 655 nm for 10 min (0.3 W cm −2 ) and 808 nm irradiation for 10 min (1.0 W cm −2 ), respectively. The PDT-PTT dual-therapy group was firstly under 808 nm irradiation for 10 min and then 655 nm for 10 min. N, N'-dimethyl sulfoxide (150 μ L) was used to completely liberate the formazan crystals. The absorbance at 490 nm was measured to calculate the cell viability.
In vitro MR test with Fe 3 O 4 @C@PMOF as probe. In vitro MR imaging with Fe 3 O 4 @C@PMOF as probe was carried out at different Fe concentrations (0.07, 0.14, 0.28, 0.56, 1.12 mM) with a 1.2 T MR Imaging System, Huantong, Shanghai, China. The T 2 value could be tested directly with the 1.2 T MR imaging system, and Fe content of Fe 3 O 4 @C@PMOF was determined by inductively coupled plasma-atomic emission spectroscopy. The slope of the linear fitting equation between 1/T 2 and Fe content was the r 2 value. Images were recorded using a 50 mm animal coil and a 2D gradient imaging sequence. The MR parameters were described as follows: spin-echo T 2 -weighted MR sequence, TR/TE = 5000/64.6 ms, FOV = 50 × 80 mm 2 , matrix = 512 × 256, slice thickness = 0.4 mm, 30.0 °C.
In vivo MR imaging. In vivo MR imaging was performed on the mice or MCF-7 tumor-bearing mice anesthetized with 4% chloral hydrate (6 mL kg −1 ). After intravenous injection of Fe 3 O 4 @C@PMOF solution (20 mg kg −1 ) into the mice, the MR images were recorded by positioning the mice on the animal plate of imaging system. The MR imaging was recorded on a 1.2 T MR imaging system, Huantong, Shanghai, China. Images were obtained using a small animal coil, before and at subsequent intervals following injection with the imaging sequence: TR/TE = 300/32.6 ms; FOV = 50 mm × 80 mm; matrix = 512 × 256; slice thickness = 0.4 mm without gap; 128 coronal or axial slices. To study the contents of nanocomposites in excretion of mice, mice after intravenous injection with Fe 3 O 4 @C@PMOF were housed in metabolic cages to collect their urine and feces. The collected urine and feces were digested by chloroazotic acid and measured by ICP-AES.
In vivo fluorescence imaging. In vivo fluorescence images of nude mice and tumor-bearing mice were recorded after anesthetized with 4% chloral hydrate (6 mL kg −1 ). After intravenous injection of Fe 3 O 4 @C@PMOF solution (20 mg kg −1 ) into the mice, the fluorescence images were recorded with NightOWL LB 983 small animal in vivo imaging system (Berthold Technologies GmbH & Co. KG, Germany) with the optimal wavelength of PMOF (Ex = 550 nm, Em = 660 nm). The data were treated with IndiGO software.
PTT-PDT dual-therapy of tumor-bearing mice with Fe 3 O 4 @C@PMOF as probe. The efficiency of PTT-PDT dual-therapy of Fe 3 O 4 @C@PMOF was tested with tumor-bearing mice (18− 22 g, n = 3 per group) as model after anesthetized with 4% chloral hydrate (6 ml kg −1 ). The mice were separated into five groups. The mice injected intravenously with saline were used as negative control, while the mice injected with Fe 3 O 4 @C@ PMOF (10 mg kg −1 ) without any laser irradiation were used as positive control. The other mice injected with Fe 3 O 4 @C@PMOF were subjected to laser irradiation for PTT, PDT, and PTT-PDT dual-therapy, respectively. The distribution of Fe 3 O 4 @C@PMOF was monitored with fluorescent imaging. When the signal of Fe 3 O 4 @C@PMOF reached zenith at the tumor site, the tumor site of the mice in PDT group were irradiated with 655 nm laser for 10 min (0.3 W cm −2 ), while the mice in PTT group were irradiated under 808 nm laser for 10 min (1.0 W cm −2 ). The mice in PTT-PDT dual-therapy group were firstly irradiated under 808 nm laser for 10 min (1.0 W cm −2 ) and then 655 nm laser (0.3 W cm −2 ) for 10 min, respectively. The weights and tumor sizes of the mice were monitored simultaneously. The tumor volumes were calculated as (width 2 × length)/2 based on the previous report 28 . The main organs of mice were collected after treatment for 8 days. Hematoxylin and eosin (H&E) stained images were used to investigate the biotoxicity. The body weights of the mice were assessed with a counter balance within 8 days.