Accepted Manuscript Tumor acidity-activatable Manganese Phosphate Nanoplatform for Amplification of Photodynamic Cancer Therapy and Magnetic Resonance Imaging Yongwei Hao, Cuixia Zheng, Lei Wang, Jinjie Zhang, Xiuxiu Niu, Qingling Song, Qianhua Feng, Hongjuan Zhao, Li Li, Hongling Zhang, Zhenzhong Zhang, Yun Zhang PII: DOI: Reference:

S1742-7061(17)30529-9 http://dx.doi.org/10.1016/j.actbio.2017.08.028 ACTBIO 5039

To appear in:

Acta Biomaterialia

Received Date: Revised Date: Accepted Date:

18 May 2017 17 August 2017 21 August 2017

Please cite this article as: Hao, Y., Zheng, C., Wang, L., Zhang, J., Niu, X., Song, Q., Feng, Q., Zhao, H., Li, L., Zhang, H., Zhang, Z., Zhang, Y., Tumor acidity-activatable Manganese Phosphate Nanoplatform for Amplification of Photodynamic Cancer Therapy and Magnetic Resonance Imaging, Acta Biomaterialia (2017), doi: http:// dx.doi.org/10.1016/j.actbio.2017.08.028

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Tumor acidity-activatable Manganese Phosphate Nanoplatform for Amplification of Photodynamic Cancer Therapy and Magnetic Resonance Imaging Yongwei Hao1,2,3, Cuixia Zheng1, Lei Wang1,2,3*, Jinjie Zhang1,2,3, Xiuxiu Niu1, Qingling Song1, Qianhua Feng1, Hongjuan Zhao1, Li Li1, Hongling Zhang1,

Zhenzhong Zhang1,2,3* and Yun

Zhang1,2,3* 1. School of Pharmaceutical Sciences, Zhengzhou University, 100 Kexue Avenue, Zhengzhou 450001, P. R. China. 2. Key Laboratory of Targeting Therapy and Diagnosis for Critical Diseases, Zhengzhou 450001, P. R. China. 3. Collaborative Innovation Centre of New Drug Research and Safety Evaluation, Henan Province, 100 Kexue Avenue, Zhengzhou 450001, P. R. China. *Correspondent: Dr. Lei Wang, Prof. Zhenzhong Zhang and Prof. Yun Zhang, School of Pharmaceutical Sciences, Zhengzhou University, 100 Kexue Avenue, Zhengzhou, Henan Province 450001, P.R. China. Tel: 86-371-67781910; Fax: 86-371-67781908; Email: [email protected], [email protected]; [email protected];

1

Abstract Amorphous biodegradable metal phosphate nanomaterials are considered to possess great potential in cancer theranostic application due to their promise in providing ultra-sensitive pH-responsive therapeutic benefits and diagnostic functions simultaneously.

Here we report the synthesis of

photosensitising and acriflavine-carrying amorphous porous manganese phosphate (PMP) nanoparticles with ultra-sensitive pH-responsive degradability and their applications for a photoactivable synergistic nanosystem that imparts reactive oxygen species (ROS) induced cytotoxicity in synchrony with hypoxia-inducible factor 1α/vascular endothelial growth factor (HIF1α/VEGF) inhibitor that suppresses tumor growth and treatment escape signalling pathway. Carboxymethyl dextran (CMD) is chemically anchored on the surface of porous manganese phosphate theranostic system through the pH-responsive boronate esters. Upon the stimulus of the tumor acid microenvironment, manganese phosphate disintegrates and releases Mn2+ ions rapidly, which are responsible for the magnetic resonance imaging (MRI) effect. Meanwhile, the released photosensitizer chlorin e6 (Ce6) produces ROS under irradiation while acriflavine (ACF) inhibits the HIF-1α/VEGF pathway during the burst release of VEGF in tumour induced by photodynamic therapy (PDT), resulting in increased therapeutic efficacy. Considering the strong pH responsivity, MRI signal amplification and drug release profile, the PMP nanoparticles offer new prospects for tumor acidity-activatable theranostic application by amplifying the PDT through inhibiting the HIF1α /VEGF pathway timely while enhancing the MRI effect. Keywords: amorphous porous manganese phosphate, photodynamic therapy, magnetic resonance imaging

2

1. Introduction Combination therapy holds considerable appeal for effective cancer treatment [1, 2]. Photodynamic therapy is a clinically approved non-invasive therapeutic approach that employs a photosensitizer (PS), an appropriate exciting light and oxygen (O2) molecules through generation of cytotoxic reactive oxygen species (ROS) to attack biomolecules (e.g., DNA, biological membrane) inside cancer cells [3]. However, a fundamental challenge in oncology is that many resistance mechanisms and escape pathways ultimately limited the treatment efficacy. Due to the consumption of O2 induced by PDT as well as the inherent inadequate O2 supply for the solid tumors, the PDT would aggravate the hypoxia phenomenon [4-6]. Under hypoxia, stabilization of HIF-1α occurs through inhibition of 4-prolyl hydroxylase activity, an enzyme that requires oxygen to be functional. Upon stabilization, HIF-1α protein was transported into the nucleus where it heterodimerizes with HIF-1 β, forming the active HIF-1 transcription complex [7]. This process finally increased the level of vascular endothelial growth factor (VEGF) because HIF-1α plays a pivotal role in physiological and pathophysiological angiogenesis by directly regulating VEGF, a master regulator of angiogenesis in endothelial cells [8-11]. One previous research has demonstrated that burst release of VEGF following PDT is within 6 hours [12]. Therefore, co-packing interactive therapeutic agents into one system with spatiotemporally synchronized release would make it to synergize within the critical time window for PDT-mediated therapy and vascular regrowth inhibition during the burst of VEGF in tumour. The benefits of co-encapsulation of photodynamic agent and additional agent in one single carrier have been confirmed by many research groups in vitro and in vivo [13-16]. Porous nanomaterials, particularly porous silicon based nanosystems, have been paid great attention because of their large surface area, tunable pore size and volume as well as high loading capacity for drugs, dye agents and photosensitizers (PS) [17, 18]. With the rapid development of imaging approaches, such as magnetic resonance imaging, there is a pressing need for the development of nanomedicine of synergistic drug combination as well as diagnostic application [19, 20]. Very 3

recently, biodegradable manganese-based nanomaterials have been successfully developed for anticancer delivery [21-23]. Manganese was introduced to the therapeutic systems since Mn is one of the necessary elements in human body for metabolism and the biological system can efficiently control its uptake and excretion, showing low toxicity and high biosafety [24, 25]. Moreover, our previous work also demonstrated that MnO2 based nanocarriers with tumor microenvironmentresponsive MRI function can be used for anticancer drug delivery [26, 27]. However, the relaxivity of these systems was not high enough owing to the absence of water molecules coordinated with Mn2+. In order to enlarge the water-accessible surface, hollow nanostructures were introduced, which possess a higher r1 relaxivity [28, 29]. Besides, a hollow pH-responsive manganese phosphate nanosystem for cancer cells targeted MRI and therapeutic agent delivery was investigated in vitro [30]. Despite these efforts, their applications as pH-responsive theranostic platforms by combing imaging functions and therapeutic agents in vivo application also require considerable improvement. Importantly, it was reported that the decomposition of the pHresponsive materials in the amorphous form under the acid environment was accelerated without the lattice energy limitation [31]. Lattice energy is a key parameter for the predication of the stability of ionic compounds [32]. In other words, the separation of manganese ions from phosphate ions if the material was in the crystal form would be difficult due to the lattice energy. Alternatively, the amorphous porous manganese phosphate nanoplatform would be a superior candidate as a theranostic nanosystem for MRI and synergistic drug combination. Here, we report the synthesis of amorphous porous manganese phosphate nanoparticles and their application for synergistic drug combination in the pursuit of amplification of photodynamic cancer therapy. As shown in scheme 1, in such nanoparticles, chlorin e6 (Ce6), a photosensitizer, was loaded for photodynamic therapy [33]. Additionally, the nanoparticles could also enable efficient loading of acriflavine (ACF) for inhibition of HIF-1α/VEGF pathway, therefore increasing the PDT efficacy induced by Ce6 when it was exposed to the 660 nm laser irradiation [34, 35]. In order to minimize the premature drug release, it was highly desired to explore the on-demand drug 4

release strategies through capping the nanoparticles with an intelligent gatekeeper [36]. Carboxymethyl dextran (CMD), a hydrophilic polymer, was chemically anchored on the surface of porous manganese phosphate through the pH-responsive boronate esters because it has been widely used for many biomedical applications [37-39]. The CMD modification is expected to endow photosensitising and ACF-carrying amorphous porous manganese phosphate (PMP) nanoparticles with some merits. On one hand, it could act as a gatekeeper by forming a dense layer around the nanoparticles, which is favorable for minimizing premature drug release. On the other hand, with hydrophilic character, CMD coating would improve the stability and biocompatibility of the system. The enhanced permeability and retention (EPR) effect is an unique phenomenon of solid tumors, which is relating to their anatomical and pathophysiological differences from normal tissues. The reticuloendothelial system (RES), which is enriched in the liver and spleen, can be a major obstacle to tumor delivery of macromolecular drugs relying one EPR effect [40]. Just as PEGylation reduced the rate of RES uptake and increased the circulation half-life of various types of nanoparticles, CMDylation was able to reduce nanomaterials accumulation in reticuloendothelial system (RES) and prolong their blood circulation time, resulting in increased chance of accumulation in the region of interest (ROI) through the EPR effect. Therefore, CMDylation thus benefits EPR-based targeting of drugs to tumors.

2. Materials and methods 2.1. Materials Oleic acid and manganese (II) 2, 4-pentanedionate were obtained from Alfa Aesar (USA). Oleylamine and Acriflavine (ACF) were purchased from Xiya Reagent (Shandong, China). (3Aminopropyl) trimethoxysilane (APTMS) was ordered from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Carboxymethyl dextran sodium salt (CMD) and triethyl phosphate were obtained from Tokyo Chemical Industry (Tokyo, Japan). 4-Formylphenyboronic acid was purchased from Aladdin Industrial Corporation (Shanghai, China). Ce-6 was ordered from J&K Scientific Ltd. (Beijing, Chian). Sodium borohydride was obtained from Sigma-Aldrich, Inc. (St Louis, MO, USA). 5

2.2. Synthesis and modification of PMP The preparation of porous manganese phosphate (PMP) was achieved by following a literature procedure with a minor modification [30]. First, oleic acid (2 mL), oleylamine (3 mL) and methylbenzene (12 mL) were placed into a stainless steel autoclave, then deionized water (0.2 mL), manganese (II) 2,4-pentanedionate (1 mmol) and triethyl phosphate (0.4 mL) were added into the above reaction system. The temperature was kept at 180℃ for 9 h, and the stainless steel autoclave was cooled down rapidly to room temperature with the help of cold water, and then the light brown mixture was added into anhydrous ethanol (10 mL). The suspension was centrifuged at 12 000 g for 15 minutes and the precipitate was washed with anhydrous ethanol for three times. The final product (PMP) was dried in a vacuum drying oven for further use. NH2-PMP was synthesized by dispersing 34 mg of PMP in n-hexane (90 mL) with the help of sonication. APTMS (100 µL) and acetic acid (9 µL) were added into the above solution and the solution was stirred for 5 minutes. The product was obtained by centrifugation at 12 000 g for 10 minutes and washed with anhydrous ethanol for three times and dried in air at room temperature. For the preparation of PBA-PMP, NH2-PMP (100 mg) in methanol (25 mL) was added with 4formylphenyboronic acid (2.0 mg). The mixture was stirred at room temperature for 4 h. Then sodium borohydride (2.3 mg) in cold methanol was added into the reaction system in an ice bath. After the reaction for 30 minutes under the ice bath, the solution was stirred for another 24 h at room temperature. The solution was centrifuged at 12 000 g for 10 minutes and washed three times with anhydrous ethanol. The final product was dried at room temperature for further use. Besides, a model system, in which the starting APTMS alone was used to follow each step reaction by 1H NMR spectroscopy. 2.3. Preparation of Drug-loaded PMP NPs and CMD Modification Firstly, ACF (10 mg) in deionized water (3 mL) and Ce6 (10 mg) in ethanol were dispersed into PBA-PMP aqueous solution (4 mL) with a concentration of 10 mg/mL under magnetic stirring for 4 h in the dark. Then, CMD (40 mg) dissolved in deionized water (1 mL) was added into the

6

above reaction system. The resultant solution was then stirred for another 4 h. In order to remove the free ACF and Ce6, the mixture was dialyzed for 6 h. 2.4. Characterization A zetasizer Nano ZS-90 instrument (Malvern, UK) relying on the dynamic light scattering (DLS) technology and a transmission electron microscope (TEM, FEI Tecnai G2 20, USA) were used for characterizing the nanoparticle size distribution, zeta potential and morphology of the asprepared nanoparticles. Powder X-Ray diffraction (XRD) was carried out by using an X-ray diffractometer (X`Pert PRO MPD, PANalytical, Almelo, Netherlands). X-ray photoelectron spectroscopy (XPS) was applied for the verification of the Mn and P elements (Thermo ESCALAB 250XI, USA). Nitrogen adsorption-desorption isotherms were recorded on a NOVA Touch apparatus (Quantachrome Instruments, Florida, USA). A MicroMR apparatus (0.5 T; Shanghai Niumag Electronic Technology Co., Ltd., Shanghai, China) was employed for evaluating the imaging property and the longitudinal (T1) relaxivity values. The concentration of ACF was measured by an HPLC. HPLC separation was carried out on an AcclaimTM 120 C18 column (4.6 mm×250 mm, 5 µm) maintained at 30 ℃. The isocratic mobile phase was composed of water, methanol and acetonitrile with a ratio of 10:85:5, which was delivered at a flow rate of 1 mL min-1. The measuring wavelength was set at 456 nm, and the injection volume was 20 µL. The concentration of Ce6 was determined at 660 nm by an UV-Vis spectrophotometer (Shimadzu, Tokyo, Japan). The 1H NMR spectra of the APTMS derivative and CMD derivative were determined by using a Bruker Avance-III-HD 400 MHz NMR Spectrometer (Bruker BioSpin Corp., Billerica, MA, USA) The release measurements were carried out in phosphate-buffered saline (PBS) pH 7.4 and acetic buffer solution pH 6.0. The formulation was diluted with the release medium, and was then placed into pre-treated dialysis bags (MW cutoff = 8000 KDa). Each brown bottle was added with the release medium and the dialysis bags were placed into them completely. These bottles were agitated in a shaker at 37 ℃ in the darkness. The samples were taken out at predetermined

7

intervals and replaced with an equal volume of fresh medium. The concentrations of Ce6 and ACF were determined as described before. The cumulative release% was calculated by using the following formula: Cumulative release%= F1/F0×100%, where F0 was the drug content in the formulation placed in the dialysis bag, and F1 was the cumulative released drug amount in the release medium. 2.5. Cell Experiment 2.5.1. Cellular Uptake Given the fluorescent nature of ACF, we employed the fluorescence of ACF to determine the internalization of the free ACF and C-PMP/Ce6/ACF NPs. Briefly, SMMC-7721 cells were seeded in 6-well plates at 3×105 cells per well and incubated for 12 h at 37 ℃. Then, the free ACF and CPMP/Ce6/ACF NPs dispersed in cell culture medium were added at an equivalent ACF concentration of 3.3 µM. At the end of incubation, the drug-contained medium was abandoned and each well was rinsed with PBS. Similarly, the parallel wells in need of irradiation were subjected to a 660 nm laser (UltraFire, China) for 20 s at 3 W/cm2 after being stained with Lyso-Tracker Red (Beyotime Biotechnology, Shanghai, China) per the manufacturer’s instruction. After the treatment, a fluorescence microscope was used to observe the fluorescence. 2.5.2. Cytotoxicity Assay In order to evaluate the cytotoxicity of PMP-based formulations, SMMC-7721 cells were seeded into a 96-well plate (5×103 cells/well) in medium (100 µL) and cultured for 12 h to allow cell adhesion. The cells were then treated with the tested NPs. The concentrations of ACF and Ce6 were 1.0 µM and 1.58 µM, respectively. After incubation, the cell viability was evaluated by the methylthiazolyldiphenyl tetrazolium bromide (MTT) assay. Each experiment was conducted for three independent measurements. 2.5.3. qRT-PCR and ELISA Assay SMMC-7721 cells were seeded in 6-well plates at 3×10 5 cells per well. After incubation for 12 h, the cells were then added with the tested NPs at the relevant concentrations. After incubation for 4 h, the cells in the irradiation groups were irradiated with a 660 nm laser for 20 s at 3 W/cm2. After 8

24 h of incubation, total RNA from cells samples was extracted by TRIzol reagent (CoWin Biosciences, Beijing, China). cDNA was synthesized with Oligo (dT)18 primer and the RT EasyTM I Reverse Transcription Kit (World’s Foregene, Chengdu, China). In order to quantify VEGF expression, qRT‐PCR analysis was performed by using Real time PCR EasyTM-SYBR Green I (World’s Foregene, Chengdu, China), and β-actin was used as the endogenous control. PCR primers for VEGF were as follows: forward, 5’ GAAGGAGGAGGGCAGAATCATCAC3’; reverse, 5’CACAGGATGGCTTGAAGATGTACTC

3’;

for

β-actin

were:

forward,

5’AGAGCTACGAGCTGCCTGAC3’; reverse, 5’AGCACTGTGTTGGCGTACAG3’. qRT‐PCR was performed with a LightCycler@ 96 instrument (Roche, Switzerland). The results were calculated by using the 2−∆∆CT methods. On the other hand, the parallel cells were seeded and subjected to PDT as described above. The cell medium was collected for evaluating the VEGF level by using a VEGF ELISA Kit (MEIMIAN, Jiangsu, China) per the manufacturer’s instruction. 2.5.4. Apoptosis Determination SMMC-7721 cells were seeded in 6-well plates at 3×10 5 cells per well. After incubation for 12 h, the cells were then added with the tested NPs at the relevant concentrations. The concentrations for ACF and Ce6 were 1.0 µM and 1.58 µM, respectively. After incubation for 4 h, the cells were irradiated with a 660 nm laser for 20 s at 3 W/cm2. After incubation for 24 h, the cells were collected for subsequent staining by Annexin-FITC/PI per the manufacturer’s instruction. Finally, the collected cell samples were analyzed by a flow cytometry (BD Bioscience, Franklin lakes, NJ, USA). 2.5.5. ROS Determination Cells were seeded and subjected to PDT as described in section 2.5.4. Alkaline comet assay was carried out for evaluating the damage to cells by the ROS. The collected cells were combined with molten agarose (at 37 °C) and added on the clean glass slides. After being placed in a refrigerator at 4°C for 2 h in darkness, slides were then immersed in pre-chilled fresh lysis solution for 2 h in darkness. Slides were then placed in a horizontal electrophoresis apparatus reservoir, 9

which was filled with alkaline electrophoresis buffer. The electrophoresis was carried out at 60 V for 20 min. Then the slides were washed twice with pure water. Finally the air-dried slides were added with PI and a fluorescence microscopy was used to capture the images. The results were processed by using Comet Score™ software (TriTek, Annandale, VA, USA) to evaluate DNA damage by calculating the % DNA tail [41]. 2.5.6. Western Blotting Cells were seeded and subjected to PDT as described in section 2.5.4. Treated cells were collected for lysis in RIPA buffer containing protease and phosphatase inhibitors. The total protein concentrations were measured by using the BCA Protein Assay Kit (Beyotime Biotechnology, China). For each group, 20 µg of protein was loaded onto 10% SDS-PAGE gel and was electrotransferred from the gel onto the PVDF membrane. Afterwards, the membranes were incubated with a blocking buffer (5% skimmed milk in TBST) and incubated with the primary antibody at 4 ℃ for 12 h. Antibodies were measured with horseradish peroxidase (HRP)-conjugated secondary antibody. Finally, an UVP imaging system (Ultra violet Products Ltd., CA, USA) was used to take the images. 2.6. In vivo imaging and Antitumor Efficacy BALB/c nude mice (four weeks) were purchased from the SJA Laboratory Animal Co., Ltd. (Changsha, China) and all the mice received care in compliance with the criteria of the national Regulation on the Management of Laboratory Animals. The SMMC-7721-bearing mice models were established according to the following method. SMMC-7721 cells (1×10 7) were injected on the right axilla of the five week-old Balb/c nude mice. After implantation for one week, in order to assess the distribution behaviour in vivo after PMP-based NPs injection to the SMMC-7721-bearing mice, NIR fluorescence in vivo imaging (Carestream Health, Inc., USA) was carried out. The instrument was employed to record the NIR emission spectra of mice administrated with two formulations, respectively. The captured images and the corresponding X-Ray images were analyzed by the Bruker Molecular Imaging Software. At the scheduled time points, MRI was conducted on a 3.0 T clinical scanner (Siemens Healthcare Sector, Erlangen, Germany). When the tumor volume reached about 100 mm3, the mice were divided into eight groups 10

randomly. All formulations were intravenously injected into mice via the tail every two days, respectively. The dosage of Ce6 and ACF was 5 mg/kg and 2.5 mg/kg, respectively. At the predetermined time point, the tumors of the PDT groups were irradiated with a 660 nm laser (3 W/cm2, 2 min). The mice were weighted by an electronic balance and their tumor volumes were measured by a vernier caliper. The tumor volumes were calculated as V=0.5×a2×b, where a and b represent the shortest and longest diameter of the tumor, respectively. After five times of treatment were completed, major organs were subjected to the hematoxylin and eosin (H&E) staining. The collected tumors were rapidly frozen at −80 °C. Then, the collected tumor samples were sliced with a cryostat microtome. The tumor sections were incubated with a 1:250 dilution of CD-31 primary antibody (Proteintech, Wuhan, China) at 4 °C overnight, and then they were incubated with Alexa Fluor 488-labelled secondary antibody (1:200) for 1 h at ambient temperature. Images were captured by a Nikon microscope. 2.7. Statistical Analysis The data were presented as the mean ± standard deviation (SD). The differences between two groups and multiple groups were analysed by Student’s t test and one-way ANOVA, respectively. The level of significance was set at probabilities of *p < 0.05, **p < 0.01, and ***p < 0.001.

3. Results and discussion 3.1. Preparation and Characterization of PMP NPs Amorphous porous manganese phosphate (PMP) materials with designable porosity and functionality have promising applications in drug delivery. The successful synthesis was validated by a series of methods. As shown in Figure 1A, the TEM image showed the prepared uniform nanoparticles were round with a size of about 168 nm. TEM image of the PMP NPs at a higher magnification indicated some faint dots dispersed on each particle, indicating the formation of porous structure (Figure 1B). Besides, the EDS also demonstrated the existence of Mn and P elements (Figure 1C). As shown in Figure 1D, the X-ray diffraction (XRD) pattern of the PMP NPs demonstrated that the crystal form of them was amorphous due to the observed broad peaks in 11

the entire tested range. This phenomenon was beneficial because amorphous nanoparticles exhibited good biodegradability, which is also found among the analogous calcium phosphate nanomaterials [42-44]. Generally, Mn 2p states are used to distinguish the difference between Mn2+ and Mn4+. The Mn 2p peaked at binding energies of 641.85 kev and 653.67 kev, corresponding to the Mn 2p3/2 and Mn 2p1/2 spin-orbit peaks, respectively (Figure S1). These data suggested that the Mn atoms have +2 oxidation state, which matched the reported values of Mn (II) oxidation states [45]. As shown in Figure 1E, PMP exhibited a type of V in the International Union of Pure and Applied Chemistry (IUPAC) classification, and its surface area was about 20.58 g/m2. Its pore size was about 3.34 nm (Figure 1F), which is enough to load the small drug molecules because the estimated molecular size of ACF and Ce6 through the ChemDraw Software12.0 is about 1.25 nm and 1.53 nm, respectively. 3.2. Characterization of PMP-based Theranostic NPs The success of one drug delivery system in vivo depends largely upon the stability of the nanoparticles. For this purpose, the CMD was employed to functionalize the surface of the drugloaded NPs [46]. As shown in the FT-IR spectra in Figure 2A, the characteristic peaks at 2924.5 cm-1 and 2853.6 cm-1 corresponding to the -C-H stretching vibration of oleic acid decreased obviously after modification with (3-Aminopropyl) trimethoxysilane (APTMS). In addition, new peaks at 1053 cm-1 and 1636 cm-1 were ascribed to the –C-N stretching vibration and the N-H deformation vibration, respectively, further supporting the successful APTMS modification. Consistent with the phenomenon observed in the inset in Figure 2A, the PMP NPs with hydrophobic surface property could only be suspended in hexane phase, whereas the amine PMP NPs could be well-dispersed in water. After the PBA modification, one new peak at 1384 cm-1 corresponding to the –B-OH stretching vibration further demonstrated the introduction of PBA. Besides, we used a model system, in which the starting APTMS alone was used to follow each step reaction by 1H NMR spectroscopy (Figure S2). For APTMS derivative, final product was vacuum dried to obtain a yellowish solid. 1H NMR spectroscopy was performed with the following

12

parameters: 1H NMR (400 MHz, DMSO-d 6) δ8.38 (s, 2H), 7.99 (d, J = 7.4 Hz, 2H), 7.86 (d, J = 7.4 Hz, 2H), 4.22 (s, 2H), 3.68 (s, 9H), 2.18 (t, J= 7.4 Hz, 2H),1.9 (s, 1H), 1.12-1.41(m, 2H), 0.85 (t, J = 6.4 Hz, 2H). Compared to the 1H-NMR spectrum of CMD (Figure S3), the new peaks at 2.94 ppm, 1.70 ppm and 0.60 ppm of CMD derivative were attributed to the protons of the APTMS derivative. Besides, new peaks at 7.86 ppm and 7.96 ppm were the characteristic resonances of benzene ring. Moreover, the disappearance of the characteristic peaks of –B-(OH)2 and 1,2-diols further confirmed the formation of the boronate esters. Therefore, these results indicated that APTMS derivative was successfully conjugated to the CMD. Subsequently, Ce6 and ACF were introduced into PMP NPs by soaking. Although the surface area of the PMP NPs was not big enough, the loading efficiency of Ce6 was more than 90% when the ratio of the carrier and Ce6 was set at 1:1. This high efficiency of Ce6 may be due to the electrostatic interaction between the Ce6 and the unreacted APTMS on the inner pores of the nanoparticles. In contrast, the loading efficiency of ACF was low. When the ratio of the carrier and ACF was set at 1:1, the loading efficiency of ACF was less than 50%, which may be due to the repulsion from the above mentioned unreacted APTMS. The loading efficiency of Ce6 and ACF reached up to 85.6% and 46.0% respectively when the ratio of carrier/Ce6/ACF was set at a weight ratio of 4:1:1. Recently, the hollow mesoporous nanoparticles have emerged as the drug delivery vehicles because the hollow counterparts proved to be the drug molecules reservoir [36]. However, the weak interactions between the loaded drug molecules and the material matrix made on-demand release difficult. For our system, this high drug loading efficiency may be related to the special feature of the PMP NPs. When these PMP NPs were immediately incubated in the PS/drug solution, the PS/drug molecules were surely to be imported and entrapped in the pores instead of the huge cavity. After being capped with CMD, the surface area decreased from 20.58 m2/g to 10.18 m2/g, and the pore size distribution also decreased to 1.46 nm (Figure 2B), indicating the existence of CMD. As shown in Figure 2C, the TEM image of C-PMP/Ce6/ACF showed that the drug loaded NPs were evidently wrapped around by the high molecular CMD as well as the entrapment of drug

13

molecules. These data indicated that the pores were blocked to some extent due to the PS/drug loading as well as the capping of the gatekeeper CMD. The DLS data showed that hydrodynamic size of C-PMP/Ce6/ACF NPs was around 180.7 nm in the aqueous solution (Figure 2D), and its PDI was about 0.21. The zeta potential was around -22.6 mv (Figure 2E), which is beneficial to make the NPs having good stability in vitro and in vivo. As shown in Figure 2F, those CMDylated nanoparticles exhibited excellent dispersibility in PBS without any precipitation, allowing their further applications in biological systems. Thus, the CMD played a critical role in our nanoplatform, which is consistent with the Resovist (ferucarbotran), a commercialized iron oxide nanoparticles (IONP) formulation modified with CMD [47]. 3.3. In Vitro Release and MRI Evaluation The release behaviours of the drug-loaded PMP NPs were investigated by the dialysis method. As shown in Figure 3A, the release percentages of ACF and Ce6 for C-PMP/Ce6/ACF NPs after incubation for 8 h at pH 7.4 were 49.2% and 47.8%, respectively. We inferred that the CMD modification on the surface partially blocked the pores containing drug molecules, leading to the release of Ce6 and ACF in a slow manner. However, the NPs showed a more rapid release profile at pH 6.0 in comparison to that at pH 7.4, indicating the highly pH-responsive release property. Therefore, the key factors including crystalline form and type of surface functional groups may contribute to the rapid drug release. On one hand, the CMD was easily detached from the nanoparticles because it was introduced through the borate ester bonds between boronic acids and idols, which are easily to break under the acid environment [48, 49]. On the other hand, the amorphization of the PMP NPs was also critical for the separation of manganese ions and phosphate ions without the limitation of the lattice energy, which further promoted the drug release. According to our prior researches, manganese ions could be used as the T1-weighted contrast agents [50]. It is expected that the imaging property would be improved when the acid triggered the manganese ions release. As shown in Figure 3B, the r1 relativity value was measured to be 1.9 mM−1 s−1 and 4.8 mM−1 s−1 at pH 7.4 and pH 6.0, respectively. At pH 7.4, it was inevitable for the water molecules to

14

access the exposed Mn2+ on the surface of the nanoparticles due to the incompletely coating as well as the large surface of the nanoparticles, which lead to the self-imaging property. However, at pH 6.0, the r1 relativity value was measured to be 4.8 mM−1 s−1 because of a lot of released Mn2+ ions, which enlarged the water-accessible surface [51]. The r1 relativity of free Mn2+ was determined with a value of 5.2 mM-1s-1, Therefore, the r1 value of our system at pH 6.0 was comparable to that of free Mn2+, indicating our system under in vitro mimic microenvironment still showed the potential for positive MRI. Besides, the T1-weighted MRI images in Figure 3C also demonstrated the enhanced imaging function of our system in pH 6.0 PBS compared to that in pH 7.4 PBS. In order to test our hypothesis, the TEM image of the C-PMP/Ce6/ACF was further captured as shown in Figure 3D. As expected, fewer nanoparticles and some dissolved PMP matrix/fragments could be traced at pH 6.0. Thus, these data indicated that low pH stimulus could cause the dissolution of PMP NPs, which not only promoted the drug release, but also rendered the amplification of the imaging signal relying on manganese ions. 3.4. Cellular Uptake and In Vitro Efficiency In order to explore the intracellular drug release behaviour of C-PMP/Ce6/ACF NPs, its drug localization was investigated by fluorescence microscopy. As shown in Figure 4A, green fluorescence signal was observed in the SMMC-7721 cells in both the free ACF group and CPMP/Ce6/ACF NPs group. However, the fluorescence of free ACF group was stronger than that of C-PMP/Ce6/ACF NPs at 1 h. One possibility was that ACF is a small molecule that diffuses into cells easily while C-PMP/Ce6/ACF NPs entered in the cells through the energy-dependent endocytosis. As the incubation time increased to 4 h, the fluorescence of C-PMP/Ce6/ACF was stronger than that of free ACF group. Previous study of the interaction between CMD and cancer cells had revealed that CMDylated nanoparticles selectively attached to cancer cells to some extent, which is associated with the fact that the three-dimensional dextran hydroxyl network shell pointed to the endocytotic mechanisms of uptake [52]. Thus, the CMD may also affect the interaction between the drug-loaded nanoparticles and tumor cells. The cellular fluorescence after a 660 nm

15

laser irradiation was further studied. As shown in Figure 4B, the green fluorescence decreased to some extent for both groups due to the quenching effect induced by the light treatment. Moreover, the green ACF fluorescence was separated from the endosomes/lysosomes, suggesting the successful endosomal/lysosomal escape. This phenomenon could be explained by the increased inner osmotic pressure due to a lot of fresh manganese ions and phosphate ions as well as the ROS generation by Ce6 under the irradiation. The survival rate of cells incubated with the C-PMP was almost 100% when the particle concentration was less than 50 µg/ml (Figure S4). Therefore, the CPMP had no obvious cytotoxicity and was suitable as drug carriers. Figure 4C showed the cell inhibition rates of SMMC-7721 cells after treatment with different formulations. As expected, the C-PMP/Ce6/ACF plus irradiation showed the strongest anticancer effect compared to any other drug-loaded PMP NPs. Photodynamic therapy (PDT) is believed to aggravate hypoxic conditions to tumor cells, resulting in the overexpression of angiogenic markers such as vascular endothelial growth factor (VEGF) [53]. The levels of VEGF were determined by using qRT-PCR and ELISA assay, respectively. Compared to other groups, C-PMP/Ce6/ACF NPs plus irradiation could significantly decrease VEGF mRNA by about 40% and effectively inhibited the expression of the VEGF protein in SMMC-7721 cells. These results were attributed to the intracellular fate of CPMP/Ce6/ACF NPs, which possessed high cellular uptake and rapid lysomal escape, as well as synergistic effect of ACF and ROS. On the basis of the excellent results of the cell inhibition, the cell apoptosis was determined. As shown in Figure 5A, the C-PMP/Ce6 NPs in the dark have lower toxicity to SMMC-7721 cells compared with the irradiation group. However, cells apoptosis was induced in more and more cells by the C-PMP/Ce6/ACF NPs plus laser group. Comet assay was further carried out for evaluating the DNA damage induced by the ROS. As shown in Figure 5B and Figure S5, only natural DNA retained round shape, while damaged DNA migrated away from the nucleus, depending on the molecular weight of its fragments. For the groups without irradiation, such as C-PMP/Ce6 NPs and C-PMP/Ce6/ACF NPs, a little damage was observed since most of the DNA remained round. After

16

irradiation, obvious tail could be observed, indicating the damaged DNA migration. These results further confirmed that the enhanced photocytotoxicity of C-PMP/Ce6/ACF NPs resulted from DNA damage by singlet oxygen generation as well as the effect of ACF. In order to better understand the apoptotic signalling pathway, the main apoptotic regulators such as anti-apoptotic protein Bcl-2 and apoptotic protein Bax were evaluated by Western blotting. As shown in Figure 5C, the C-PMP/Ce6 NPs and C-PMP/Ce6/ACF NPs plus the 660 nm irradiation led to the lower Bcl-2/Bax ratios. Overall, there was enough potency for the C-PMP/Ce6/ACF NPs to achieve the enhanced PDT effect in the presence of the 660 nm laser irradiation. 3.5. In Vivo NIR Imaging and MRI The biodistribution of C-PMP/Ce6/ACF NPs was monitored by a near-infrared fluorescence imaging system. As shown in Figure 6A, no significant fluorescent signals were detected at the tumor site in the free Ce6 group because of the lack of targetability. In contrast, strong fluorescence signals were observed in the tumors of mice treated with C-PMP/Ce6/ACF NPs. The nanoparticles accumulation at the tumor site increased with the time extended, indicating that the nanoparticles remained in circulation and could accumulate into the tumor effectively. This feature might be attributed to the hydrophilic CMD, which could potentially escape macrophage capture to some extent. Importantly, after 24 h, C-PMP/Ce6/ACF NPs were still retained in the tumor tissue, which may be due to the special interaction of CMD and the tumor cells. Among all the observed time points, the fluorescence intensity at the tumor site reached the maximum level after 8 h injection. Thus, the T1-weighted MRI image was obtained at the same time point. As shown in Figure 6B and Figure 6C, the appreciable whitening phenomenon could be easily observed at the tumor site compared to that of the control mice. Consequently, the C-PMP/Ce6/ACF NPs could accumulate at the tumor site to some extent and could provide information about the tumor when it was used as a T1 MRI contrast agent. 3.6. In Vivo Antitumor Efficacy Encouraged by the promising results of combination therapy in vitro, we studied the in vivo

17

antitumor efficiency of the C-PMP/Ce6/ACF system. As shown in Figure 7A, the body weights of mice treated with all the PMP-based formulations barely changed during the duration of the study. Besides, the histopathological analysis of major organs sections stained with H&E (Figure S6) showed there was no pathological variation in the major organs, confirming excellent biocompatibility of all the tested nanoparticles. On one hand, the good biocompatibility might be attributed to the confinement and retention of Mn2+ ions in the nanoparticles at pH 7.4. On the other hand, the flexible CMD coating on the surface of PMP NPs could hinder the capture of the nanoparticles by phagocytotic cells, which further attributed to the excellent biocompatibility. Manganese neurotoxicity could occur in humans following exposure to high levels of manganese in the air or water. Manganese neurotoxicity is associated with elevated brain levels of manganese, especially in the human caudate-putamen, globus pallidus, substantia nigra, and subthalamic nuclei [54]. One earlier study showed that the loss of cells in globus pallidus and caudate putamen as well as in frontal cortex of rats was significantly higher (p < 0.05) for manganese phosphate exposure group [55]. Moreover, old age and gender influence the pharmacokinetics of inhaled manganese sulfate and manganese phosphate in rats [56]. However, the disturbance and impairments of biological functions induced by nanoparticles are closely relating to the particle size and surface chemistry [57]. Therefore, despite the inspiring therapeutic performance of C-PMP/Ce6/ACF, its potential long-term toxicity in vivo should be well evaluated prior to finding its way into clinical applications. In order to evaluate the therapeutic efficiency of various groups, the volumes of the SMMC7721 tumors were measured over the therapeutic period. As shown in Figure 7B, the tumors in the control group grew rapidly with a V/V0 value of 8.2 ± 0.6 at the end of the study. When the tumorbearing mice were treated with C-PMP plus irridiation or C-PMP/Ce6, no obvious tumor inhibition was observed. However, the tumor volumes of all the laser irradiation groups decreased after five times injections, which mainly attributed to the effect of the PDT effect. Remarkably, the tumors from the C-PMP/Ce6/ACF plus irradiation group were inhibited continuously and showed the

18

smallest average tumor volume with a V/V0 value of 2.7 ± 0.4, indicating the strongest antitumor effect of the combined therapy. The enhanced efficacy of the C-PMP/Ce6/ACF plus irradiation highlighted the critical function of co-packaging interactive therapeutic agents into one system. The requirement for co-loading to get optimal efficacy is most likely owing to the rapid microvessel damage and shutdown by both ROS and ACF simultaneously. In order to test our hypothesis, microvessel formation was evaluated by immunostaining with anti-CD31 antibody. Representative tumor microvessel densities in each group are shown in Figure 7C. ACF-based formulations significantly impaired the microvessels formation, especially for the C-PMP/Ce6/ACF group. The results suggested that the altered tumor growth was directly correlated with altered microvessel formation apart from the PDT effect. In addition, in order to further evaluate the effect of different therapeutic formulations, hematoxylin-eosin (H&E) staining was carried out to observe the tumor cell morphologies (Figure 7D). As expected, the shrinkage of nuclei, nuclear fragmentation, and decrease of the tumor cells density was observed after the therapeutic process, suggesting the occurrence of cell apoptosis. Among all the therapeutic groups, such phenomena were obvious for the C-PMP/Ce6/ACF plus laser group. These histological assays further demonstrated the superiority of our combination therapeutic nanosystem.

4. Conclusions In summary, this organic-inorganic hybrid theranostic nanoplatform for tumor combination therapy was constructed with the entrapment of photosensitizer Ce6 and ACF in the porous pores of PMP NPs. The resultant C-PMP/Ce6/ACF NPs were featured with distinctive advantages such as ultra pH-responsive drug release, MRI function and rational drug combination exploiting the blockage of the treatment escape signalling pathway. The complementarity and superiority of the combination were confirmed in vitro and in vivo. The present work provides new strategies for the design and construction of the tumor acidity-activatable theranostic platforms for amplification of the photodynamic cancer therapy and magnetic resonance imaging simultaneously.

19

5. Acknowledgments This research was financially supported by the National Natural Science Foundation of China (Nos. 81572991, 81673021, and 81573364), the China Postdoctoral Science Foundation (no. 2014M562002 and 2015T80783), and Outstanding Young Talent Research Fund of Zhengzhou University (1421331073).

6. Supporting Information Details of the experimental results of XPS spectra, cell viability, synthesis route, 1H-NMR spectra, percentages of DNA tail of different groups and representative H&E staining of major organs are available from the supporting information. References: [1] Y. Wang, Y. Xie, J. Li, Z.H. Peng, Y. Sheinin, J. Zhou, D. Oupicky, Tumor-Penetrating Nanoparticles for Enhanced Anticancer Activity of Combined Photodynamic and HypoxiaActivated Therapy, ACS Nano, 11 (2017) 2227-2238. [2] L. Yang, S. Zhang, X. Ling, P. Shao, N. Jia, M. Bai, Multilayer photodynamic therapy for highly effective and safe cancer treatment, Acta Biomaterialia, 5 (2017) 271–280. [3] Z. Zhou, J. Song, L. Nie, X. Chen, Reactive oxygen species generating systems meeting challenges of photodynamic cancer therapy, Chem Soc Rev, 45 (2016) 6597-6626. [4] L. Feng, L. Cheng, Z. Dong, D. Tao, T.E. Barnhart, W. Cai, M. Chen, Z. Liu, Theranostic Liposomes with Hypoxia-Activated Prodrug to Effectively Destruct Hypoxic Tumors PostPhotodynamic Therapy, ACS Nano, 11 (2017) 927-937. [5] S. Gao, G. Wang, Z. Qin, X. Wang, G. Zhao, Q. Ma, L. Zhu, Oxygen-generating hybrid nanoparticles to enhance fluorescent/photoacoustic/ultrasound imaging guided tumor photodynamic therapy, Biomaterials, 112 (2017) 324-335. [6] Q. Chen, L. Feng, J. Liu, W. Zhu, Z. Dong, Y. Wu, Z. Liu, Intelligent Albumin-MnO2 Nanoparticles as pH-/H2O2-Responsive Dissociable Nanocarriers to Modulate Tumor Hypoxia for Effective Combination Therapy, Adv Mater, 28 (2016) 7129-7136. [7] M.M. Vleugel, A.E. Greijer, A. Shvarts, P. van der Groep, M. van Berkel, Y. Aarbodem, H. van Tinteren, A.L. Harris, P.J. van Diest, E. van der Wall, Differential prognostic impact of hypoxia induced and diffuse HIF-1 alpha expression in invasive breast cancer, J. Clin. Pathol., 58 (2005) 172-177. [8] D. Shweiki, A. Itin, D. Soffer, E. Keshet, Vascular endothelial growth factor induced by hypoxia may mediate hypoxia-initiated angiogenesis, Nature, 359 (1992) 843--845. [9] N. Solban, Mechanistic Investigation and Implications of Photodynamic Therapy Induction of Vascular Endothelial Growth Factor in Prostate Cancer, Cancer Research, 66 (2006) 5633-5640. [10] B.Q. Spring, R. Bryan Sears, L.Z. Zheng, Z. Mai, R. Watanabe, M.E. Sherwood, D.A. Schoenfeld, B.W. Pogue, S.P. Pereira, E. Villa, T. Hasan, A photoactivable multi-inhibitor nanoliposome for tumour control and simultaneous inhibition of treatment escape pathways, Nature Nanotechnology, 11 (2016) 378-387. [11] Y. Kim, H.J. Nam, J. Lee, D.Y. Park, C. Kim, Y.S. Yu, D. Kim, S.W. Park, J. Bhin, D. Hwang, H. Lee, G.Y. Koh, S.H. Baek, Methylation-dependent regulation of HIF-1α stability restricts retinal and tumour angiogenesis, Nature Communications, 7 (2016) 10347. 20

[12] S.K. Chang, I. Rizvi, N. Solban, T. Hasan, In vivo optical molecular imaging of vascular endothelial growth factor for monitoring cancer treatment, Clin Cancer Res, 14 (2008) 4146-4153. [13] Y. Hao, L. Wang, Y. Zhao, D. Meng, D. Li, H. Li, B. Zhang, J. Shi, H. Zhang, Z. Zhang, Y. Zhang, Targeted Imaging and Chemo-Phototherapy of Brain Cancer by a Multifunctional Drug Delivery System, Macromol Biosci, 15 (2015) 1571-1585. [14] Y. Yi, H.J. Wang, X.W. Wang, Q.L. Liu, M. Ye, W.H. Tan, A Smart, Photocontrollable Drug Release Nanosystem for Multifunctional Synergistic Cancer Therapy, Acs Applied Materials & Interfaces, 9 (2017) 5847-5854. [15] S.S. Lucky, K.C. Soo, Y. Zhang, Nanoparticles in Photodynamic Therapy, Chemical Reviews, 115 (2015) 1990-2042. [16] D.K. Chatterjee, L.S. Fong, Y. Zhang, Nanoparticles in photodynamic therapy: An emerging paradigm, Adv. Drug Deliv. Rev., 60 (2008) 1627-1637. [17] Y. Yan, J. Fu, T. Wang, X. Lu, Controlled release of silyl ether camptothecin from thiolene click chemistry-functionalized mesoporous silica nanoparticles, Acta Biomaterialia, 51 (2017) 471-478. [18] M.S. Kang, R.K. Singh, T.-H. Kim, J.-H. Kim, K.D. Patel, H.-W. Kim, Optical imaging and anticancer chemotherapy through carbon dot created hollow mesoporous silica nanoparticles, Acta Biomaterialia, 55 (2017) 466-480. [19] M. Gao, F. Fan, D. Li, Y. Yu, K. Mao, T. Sun, H. Qian, W. Tao, X. Yang, Tumor acidityactivatable TAT targeted nanomedicine for enlarged fluorescence/magnetic resonance imagingguided photodynamic therapy, Biomaterials, 133 (2017) 165-175. [20] J.P. Celli, B.Q. Spring, I. Rizvi, C.L. Evans, K.S. Samkoe, S. Verma, B.W. Pogue, T. Hasan, Imaging and Photodynamic Therapy: Mechanisms, Monitoring, and Optimization, Chemical Reviews, 110 (2010) 2795-2838. [21] P. Huang, X.Q. Qian, Y. Chen, L.D. Yu, H. Lin, L.Y. Wane, Y.F. Zhu, J.L. Shi, Metalloporphyrin-Encapsulated Biodegradable Nanosystems for Highly Efficient Magnetic Resonance Imaging-Guided Sonodynamic Cancer Therapy, Journal Of the American Chemical Society, 139 (2017) 1275-1284. [22] H.H. Fan, Z.L. Zhao, G.B. Yan, X.B. Zhang, C. Yang, H.M. Meng, Z. Chen, H. Liu, W.H. Tan, A Smart DNAzyme-MnO2 Nanosystem for Efficient Gene Silencing, Angewandte ChemieInternational Edition, 54 (2015) 4801-4805. [23] I. Brigger, C. Dubernet, P. Couvreur, Nanoparticles in cancer therapy and diagnosis, Adv. Drug Deliv. Rev., 54 (2002) 631-651. [24] J. Li, Z. Zhao, J. Feng, J. Gao, Z. Chen, Understanding the metabolic fate and assessing the biosafety of MnO nanoparticles by metabonomic analysis, Nanotechnology, 24 (2013) 455102. [25] L. Yu, Y. Chen, M. Wu, X. Cai, H. Yao, L. Zhang, H. Chen, J. Shi, "Manganese Extraction" Strategy Enables Tumor-Sensitive Biodegradability and Theranostics of Nanoparticles, J Am Chem Soc, 138 (2016) 9881-9894. [26] Y. Hao, B. Zhang, C. Zheng, M. Niu, H. Guo, H. Zhang, J. Chang, Z. Zhang, L. Wang, Y. Zhang, Multifunctional nanoplatform for enhanced photodynamic cancer therapy and magnetic resonance imaging, Colloids Surf B Biointerfaces, 151 (2017) 384-393. [27] Y. Hao, L. Wang, B. Zhang, H. Zhao, M. Niu, Y. Hu, C. Zheng, H. Zhang, J. Chang, Z. Zhang, Multifunctional nanosheets based on folic acid modified manganese oxide for tumortargeting theranostic application, Nanotechnology, 27 (2015) 025101. [28] T.L. Ha, H.J. Kim, J. Shin, G.H. Im, J.W. Lee, H. Heo, J. Yang, C.M. Kang, Y.S. Choe, J.H. Lee, I.S. Lee, Development of target-specific multimodality imaging agent by using hollow manganese oxide nanoparticles as a platform, Chem Commun (Camb), 47 (2011) 9176-9178. [29] J. Shin, R.M. Anisur, M.K. Ko, G.H. Im, J.H. Lee, I.S. Lee, Hollow manganese oxide nanoparticles as multifunctional agents for magnetic resonance imaging and drug delivery, Angew Chem Int Ed Engl, 48 (2009) 321-324. [30] J. Yu, R. Hao, F. Sheng, L. Xu, G. Li, Y. Hou, Hollow manganese phosphate nanoparticles as smart multifunctional probes for cancer cell targeted magnetic resonance imaging and drug 21

delivery, Nano Research, 5 (2012) 679-694. [31] S. Kumar, B. Linehan, Y.C. Tseng, A new combination approach of CI jet and QESD to formulate pH-susceptible amorphous solid dispersions, Int J Pharm, 466 (2014) 368-374. [32] D.T. Liu, S.Y. Zhang, Z.J. Wu, Lattice energy estimation for inorganic ionic crystals, Inorganic Chemistry, 42 (2003) 2465-2469. [33] Z. Dong, L. Feng, W. Zhu, X. Sun, M. Gao, H. Zhao, Y. Chao, Z. Liu, CaCO3 nanoparticles as an ultra-sensitive tumor-pH-responsive nanoplatform enabling real-time drug release monitoring and cancer combination therapy, Biomaterials, 110 (2016) 60-70. [34] R. Weijer, M. Broekgaarden, M. Krekorian, L.K. Alles, A.C. van Wijk, C. Mackaaij, J. Verheij, A.C. van der Wal, T.M. van Gulik, G. Storm, M. Heger, Inhibition of hypoxia inducible factor 1 and topoisomerase with acriflavine sensitizes perihilar cholangiocarcinomas to photodynamic therapy, Oncotarget, 7 (2016) 3331-3346. [35] M. Broekgaarden, R. Weijer, M. Krekorian, B. van den Ijssel, M. Kos, L.K. Alles, A.C. van Wijk, Z. Bikadi, E. Hazai, T.M. van Gulik, M. Heger, Inhibition of hypoxia-inducible factor 1 with acriflavine sensitizes hypoxic tumor cells to photodynamic therapy with zinc phthalocyanineencapsulating cationic liposomes, Nano Research, 9 (2016) 1639-1662. [36] Q. Feng, Y. Zhang, W. Zhang, X. Shan, Y. Yuan, H. Zhang, L. Hou, Z. Zhang, Tumortargeted and multi-stimuli responsive drug delivery system for near-infrared light induced chemophototherapy and photoacoustic tomography, Acta Biomater, 38 (2016) 129-142. [37] X. Deng, Z. Yin, Z. Zhou, Y. Wang, F. Zhang, Q. Hu, Y. Yang, J. Lu, Y. Wu, W. Sheng, Y. Zeng, Carboxymethyl Dextran-Stabilized Polyethylenimine-Poly(epsilon-caprolactone) Nanoparticles-Mediated Modulation of MicroRNA-34a Expression via Small-Molecule Modulator for Hepatocellular Carcinoma Therapy, ACS Applied Materials & Interfaces, 8 (2016) 17068-17079. [38] D.G. You, V.G. Deepagan, W. Um, S. Jeon, S. Son, H. Chang, H.I. Yoon, Y.W. Cho, M. Swierczewska, S. Lee, M.G. Pomper, I.C. Kwon, K. Kim, J.H. Park, ROS-generating TiO2 nanoparticles for non-invasive sonodynamic therapy of cancer, Scientific Reports, 6 (2016) 23200. [39] E. Kim, J.M. Kim, L. Kim, S.J. Choi, I.S. Park, J.Y. Han, Y.C. Chu, E.S. Choi, K. Na, S.S. Hong, The effect of neutral-surface iron oxide nanoparticles on cellular uptake and signaling pathways, Int J Nanomedicine, 11 (2016) 4595-4607. [40] J. Fang, H. Nakamura, H. Maeda, The EPR effect: Unique features of tumor blood vessels for drug delivery, factors involved, and limitations and augmentation of the effect, Adv. Drug Deliv. Rev., 63 (2011) 136-151. [41] I. Postiglione, F. Barra, S.M. Aloj, G. Palumbo, Photodynamic therapy with 5aminolaevulinic acid and DNA damage: unravelling roles of p53 and ABCG2, Cell Proliferation, 49 (2016) 523-538. [42] W. Chu, Y. Huang, C. Yang, Y. Liao, X. Zhang, M. Yan, S. Cui, C. Zhao, Calcium phosphate nanoparticles functionalized with alendronate-conjugated polyethylene glycol (PEG) for the treatment of bone metastasis, Int J Pharm, 516 (2017) 352-363. [43] G. Balasundaram, M. Sato, T.J. Webster, Using hydroxyapatite nanoparticles and decreased crystallinity to promote osteoblast adhesion similar to functionalizing with RGD, Biomaterials, 27 (2006) 2798-2805. [44] P. Mi, D. Kokuryo, H. Cabral, H.L. Wu, Y. Terada, T. Saga, I. Aoki, N. Nishiyama, K. Kataoka, A pH-activatable nanoparticle with signal-amplification capabilities for non-invasive imaging of tumour malignancy, Nature Nanotechnology, 11 (2016) 724-730. [45] C.V.K. Sharma, C.C. Chusuei, R. Clerac, T. Moller, K.R. Dunbar, A. Clearfield, Magnetic property studies of manganese-phosphate complexes, Inorganic Chemistry, 42 (2003) 8300-8308. [46] C. Fortier, E. Louvier, Y. Durocher, G. De Crescenzo, Tailoring the Surface of a Gene Delivery Vector with Carboxymethylated Dextran: A Systematic Analysis, Biomacromolecules, 16 (2015) 1671-1681. [47] H.R. Neves, R.A. Bini, J. Barbosa, C. Salmon, L.C. Varanda, Dextran-Coated Antiferromagnetic MnO Nanoparticles for a T-1-MRI Contrast Agent with High Colloidal Stability, Particle & Particle Systems Characterization, 33 (2016) 167-176. 22

[48] Y. Hao, C. Zheng, L. Wang, Y. Hu, H. Guo, Q. Song, H. Zhang, Z. Zhang, Y. Zhang, Covalent self-assembled nanoparticles with pH-dependent enhanced tumor retention and drug release for improving tumor therapeutic efficiency, Journal of Materials Chemistry B, 5 (2017) 2133--2144. [49] J.J. Liu, Z. Luo, J.X. Zhang, T.T. Luo, J. Zhou, X.J. Zhao, K.Y. Cai, Hollow mesoporous silica nanoparticles facilitated drug delivery via cascade pH stimuli in tumor microenvironment for tumor therapy, Biomaterials, 83 (2016) 51-65. [50] Y. Hao, L. Wang, B. Zhang, H. Zhao, M. Niu, Y. Hu, C. Zheng, H. Zhang, J. Chang, Z. Zhang, Y. Zhang, Multifunctional nanosheets based on folic acid modified manganese oxide for tumor-targeting theranostic application, Nanotechnology, 27 (2016) 025101. [51] T. Kim, E.J. Cho, Y. Chae, M. Kim, A. Oh, J. Jin, E.S. Lee, H. Baik, S. Haam, J.S. Suh, Y.M. Huh, K. Lee, Urchin-shaped manganese oxide nanoparticles as pH-responsive activatable T1 contrast agents for magnetic resonance imaging, Angew Chem Int Ed Engl, 50 (2011) 10589-10593. [52] Q.Y. Huang, L.L. Zhang, X.Y. Sun, K. Zeng, J. Li, Y.N. Liu, Coating of carboxymethyl dextran on liposomal curcumin to improve the anticancer activity, RSC Advances, 4 (2014) 5921159217. [53] R.L. Lecaros, L. Huang, T.C. Lee, Y.C. Hsu, Nanoparticle Delivered VEGF-A siRNA Enhances Photodynamic Therapy for Head and Neck Cancer Treatment, Mol Ther, 24 (2016) 106116. [54] P.K. Pal, A. Samii, D.B. Calne, Manganese neurotoxicity: a review of clinical features, imaging and pathology, Neurotoxicology, 20 (1999) 227-238. [55] F. Salehi, G. Carrier, L. Normandin, G. Kennedy, R.F. Butterworth, A. Hazell, G. Therrien, D. Mergler, S. Philippe, J. Zayed, Assessment of bioaccumulation and neurotoxicity in rats with portacaval anastomosis and exposed to manganese phosphate: A pilot study, Inhalation Toxicology, 13 (2001) 1151-1163. [56] D.C. Dorman, B.E. McManus, M.W. Marshall, R.A. James, M.F. Struve, Old age and gender influence the pharmacokinetics of inhaled manganese sulfate and manganese phosphate in rats, Toxicology And Applied Pharmacology, 197 (2004) 113-124. [57] L. Yildirimer, N.T. Thanh, M. Loizidou, A.M. Seifalian, Toxicology and clinical potential of nanoparticles, Nano today, 6 (2011) 585-607. Figure legends Scheme 1. (A) Synthesis scheme for C-PMP/Ce6/ACF nanoparticles. (B) Schematic illustration of proposed mechanism of the multifunctional PMP nanoparticles for amplification of photodynamic cancer therapy and magnetic resonance imaging. Figure 1. Characterization of manganese phosphate NPs. (A, B) TEM images of the as-prepared PMP NPs. (C) EDS spectrum of the PMP NPs. (D) XRD pattern of the PMP NPs. (E) N2adsorption-desorption curves of the PBA-PMP NPs. (F) Pore diameter distribution of PBA-PMP NPs. Figure 2. Characterization of the derivatives of PMP and the drug-loaded PMP NPs. (A) FT-IR spectra of PMP and its derivatives. Insets are the photographs of hexane soluble PMP NPs (a) and water soluble PMP-NH2 NPs (b), respectively. (B) N2-adsorption-desorption curves of the C-

23

PMP/Ce6/ACF NPs and the pore diameter distribution. (C) TEM image of C-PMP/Ce6/ACF NPs. (D) Size distribution of C-PMP/Ce6/ACF NPs. (E) Zeta potential of C-PMP/Ce6/ACF NPs. (F) Macroscopic aspect of C-PMP/Ce6 NPs (a), C-PMP/Ce6/ACF NPs (b), C-PMP/ACF NPs (c) dispersion in PBS. Figure 3. In vitro release and T1-weighted MRI evaluation. (A) In vitro drug release. (B) Plots of T11

with C-PMP/Ce6/ACF NPs dispersed in PBS at different Mn concentrations. (C) T1-weighted

MRI images of C-PMP/Ce6/ACF NPs dispersed in PBS at different Mn concentrations. (D) TEM image after the treatment of C-PMP/Ce6/ACF NPs with acid solution for 0.5 h. Figure 4. Cellular uptake and in vitro efficiency. (A) Cellular uptake. (B) Fluorescence image of treated cells. Note: (a) ACF + laser group; (b) C-PMP/Ce6/ACF NPs + laser group. (C) Cell inhibition. (D) Quantitative real-time PCR (qRT-PCR) determination of the levels of VEGF in SMMC-7721 cells treated with different formulations. (E) VEGF protein expression in SMMC7721 cells treated with different formulations. Scale bar: 20 µm. Figure 5. Assay of the altered cell functions. (A) Cell apoptosis of SMMC-7721 cells for 24 h: (a) Control; (b) C-PMP + laser; (c) C-PMP/ACF; (d) C-PMP/ACF + laser; (e) C-PMP/Ce6; (f) CPMP/Ce6 + laser; (g) C-PMP/Ce6/ACF; (h) C-PMP/Ce6/ACF + laser; (B) Comet assay on SMMC7721 cells with different treatments. Scale bar: 20 µm. (C) Western blotting. Figure 6. The ability of PMP-based NPs to deliver theranostic ingredients in vivo. (A) In vivo nearinfrared (NIR) fluorescence images of SMMC-7721 tumor-bearing nude mice at different times after intravenous injection of C-PMP/Ce6 NPs (a) and free Ce6 (b). (B) In vivo T1-weighted MRI images of control mouse (a) and experimental mouse (b). (C) Average MRI signal intensity of each group. Figure 7. Anticancer activities of PMP-based formulations in vivo. (A) Changes in body weights of animals as a function of time. (B) Volumetric changes of the SMMC-7721 tumor with a schedule of multiple doses. (C)

Representative fluorescence images of tumor sections after

immunofluorescence staining. Note: (a) Control; (b) C-PMP + laser; (c) C-PMP/ACF; (d) C-

24

PMP/Ce6+laser; (e) C-PMP/Ce6/ACF; (f) C-PMP/Ce6/ACF + laser. Scale bar: 100 µm. (D) Representative H&E-stained tumor sections from different groups: (a) Control; (b) C-PMP + laser; (c) C-PMP/ACF; (d) C-PMP/ACF + laser; (e) C-PMP/Ce6; (f) C-PMP/Ce6 + laser; (g) CPMP/Ce6/ACF; (h) C-PMP/Ce6/ACF + laser; Scale bar: 50 µm.

25

Scheme 1. (A) Synthesis scheme for C-PMP/Ce6/ACF nanoparticles. (B) Schematic illustration of proposed mechanism of the multifunctional PMP nanoparticles for amplification of photodynamic cancer therapy and magnetic resonance imaging.

Figure 1. Characterization of manganese phosphate NPs. (A, B) TEM images of the as-prepared PMP NPs. (C) EDS spectrum of the PMP NPs. (D) XRD pattern of the PMP NPs. (E) N2adsorption-desorption curves of the PBA-PMP NPs. (F) Pore diameter distribution of PBA-PMP NPs.

26

Figure 2. Characterization of the derivatives of PMP and the drug-loaded PMP NPs. (A) FT-IR spectra of PMP and its derivatives. Insets are the photographs of hexane soluble PMP NPs (a) and water soluble PMP-NH2 NPs (b), respectively. (B) N2-adsorption-desorption curves of the CPMP/Ce6/ACF NPs and the pore diameter distribution. (C) TEM image of C-PMP/Ce6/ACF NPs. (D) Size distribution of C-PMP/Ce6/ACF NPs. (E) Zeta potential of C-PMP/Ce6/ACF NPs. (F) Macroscopic aspect of C-PMP/Ce6 NPs (a), C-PMP/Ce6/ACF NPs (b), C-PMP/ACF NPs (c) dispersion in PBS.

27

Figure 3. In vitro release and T1-weighted MRI evaluation. (A) In vitro drug release. (B) Plots of T11

with C-PMP/Ce6/ACF NPs dispersed in PBS at different Mn concentrations. (C) T1-weighted

MRI images of C-PMP/Ce6/ACF NPs dispersed in PBS at different Mn concentrations. (D) TEM image after the treatment of C-PMP/Ce6/ACF NPs with acid solution for 0.5 h.

28

Figure 4. Cellular uptake and in vitro efficiency. (A) Cellular uptake. (B) Fluorescence image of treated cells. Note: (a) ACF + laser group; (b) C-PMP/Ce6/ACF NPs + laser group. (C) Cell inhibition. (D) Quantitative real-time PCR (qRT-PCR) determination of the levels of VEGF in SMMC-7721 cells treated with different formulations. (E) VEGF protein expression in SMMC7721 cells treated with different formulations. Scale bar: 20 µm.

29

Figure 5. Assay of the altered cell functions. (A) Cell apoptosis of SMMC-7721 cells for 24 h: (a) Control; (b) C-PMP + laser; (c) C-PMP/ACF; (d) C-PMP/ACF + laser; (e) C-PMP/Ce6; (f) CPMP/Ce6 + laser; (g) C-PMP/Ce6/ACF; (h) C-PMP/Ce6/ACF + laser; (B) Comet assay on SMMC7721 cells with different treatments. Scale bar: 20 µm. (C) Western blotting.

30

Figure 6. The ability of PMP-based NPs to deliver theranostic ingredients in vivo. (A) In vivo nearinfrared (NIR) fluorescence images of SMMC-7721 tumor-bearing nude mice at different times after intravenous injection of C-PMP/Ce6 NPs (a) and free Ce6 (b). (B) In vivo T1-weighted MRI images of control mouse (a) and experimental mouse (b). (C) Average MRI signal intensity of each group.

31

Figure 7. Anticancer activities of PMP-based formulations in vivo. (A) Changes in body weights of animals as a function of time. (B) Volumetric changes of the SMMC-7721 tumor with a schedule of multiple doses. (C) Representative fluorescence images of tumor sections after immunofluorescence staining. Note: (a) Control; (b) C-PMP + laser; (c) C-PMP/ACF; (d) C-PMP/Ce6+laser; (e) CPMP/Ce6/ACF; (f) C-PMP/Ce6/ACF + laser. Scale bar: 100 µm. (D) Representative H&E-stained tumor sections from different groups: (a) Control; (b) C-PMP + laser; (c) C-PMP/ACF; (d) CPMP/ACF + laser; (e) C-PMP/Ce6; (f) C-PMP/Ce6 + laser; (g) C-PMP/Ce6/ACF; (h) CPMP/Ce6/ACF + laser; Scale bar: 50 µm.

32

33

In this study, we report the synthesis of the tumor acidity-activatable amorphous porous manganese phosphate nanoparticles and their application for a photoactivable synergistic nanosystem that imparts reactive oxygen species (ROS) induced cytotoxicity in synchrony with hypoxia-inducible factor 1α/vascular endothelial growth factor (HIF-1α/VEGF) inhibitor that suppresses tumor growth and treatment escape signalling pathway. Besides, upon the stimulus of the tumor acid microenvironment, the manganese phosphate nanoparticles finally disintegrate and release Mn2+ ions rapidly, which are responsible for the magnetic resonance imaging (MRI) effect. This nanoplatform is featured with distinctive advantages such as ultra pH-responsive drug release, MRI function and rational drug combination exploiting the blockage of the treatment escape signalling pathway.

34