Abstract

Radiotherapy (RT) is a major modality for cancer treatment, along with surgery and chemotherapy. Despite its therapeutic effect, the recurrence and metastasis of tumors due to the acquired resistance of cancer cells to RT remain significant clinical problems. Therefore, it is imperative to overcome radioresistance and improve radiosensitivity in cancer patients. Here, we synthesized hydroxychloroquine (HCQ)-loaded hollow mesoporous silica nanoparticles (HMSNs) to enable effective inhibition of radiation-induced cytoprotective autophagy and enhance the therapeutic efficacy of RT. HCQ-HMSN-treated HCT116 colon cancer cells showed a 200-fold higher intracellular uptake of HCQ than that of free HCQ – treated cells, thereby effectively inhibiting the radiation-induced autophagy of cancer cells. In vivo imaging and therapy studies of a tumor xenograft model showed preferential accumulation of HCQ-HMSNs in tumor tissues and significant enhancement of RT by inhibiting autophagy in the tumor sites. Histopathology analyses of major organs, blood chemistry profiles, and changes in body weights of mice confirmed the good biocompatibility of HCQ-HMSNs.

Keywords:Radiation therapy, Autophagy, Hydroxychloroquine, Hollow mesoporous silica nanoparticle, Cancer

1. Introduction

Radiotherapy (RT) is a major modality for cancer treatment. Despite its therapeutic effect, the recurrence and metastasis of tumors due to the acquired resistance of cancer cells to RT remain significant clinical problems [1-3]. Therefore, it is urgent to overcome radioresistance and improve radiosensitivity in cancer patients. Studies have shown that cancer cells attain radiation resistance through the cytoprotective autophagy [4,5]. Autophagy is a lysosome-mediated degradation process that protects cancer cells from multiple stresses and represents a therapeutic challenge because it delays apoptotic cell death in response to DNA damage. To mitigate the cytoprotective autophagy process in cancer cells, autophagy inhibitors, such as chloroquine (CQ) and its derivative hydroxychloroquine (HCQ), have proved useful in cancer therapy and are currently under evaluation in clinical trials involving chemotherapy or radiation therapy [6- 8]. For example, Kasper et al. found that a CQ treatment could sensitize HCT116 cancer cells to radiation and enhance the therapeutic effect of in vivo radiation therapy [9]. Autophagy inhibitors can lead to a reduction in the survival rate of cancer cells, especially during the recovery phase following radiation therapy. By increasing the lysosomal pH, autophagy inhibitors block the degradation of the autophagic cargo and sensitize cancer cells to radiation therapy.

However, a major challenge in the use of autophagy inhibitors is that it is difficult to achieve a sufficiently high concentration for autophagy inhibition in target tumors because of
their nonspecific biodistribution in vivo after administration, which limits their clinical applications [10]. Moreover, the acidic tumor microenvironment is known to inhibit the ability of CQ and HCQ to cross cell membranes, thereby significantly reducing their intracellular uptake [6,11]. Recently, HCQ-loaded liposomes have been developed to enhance the accumulation of HCQ in tumors and improve their antitumor effect with chemotherapeutic agents, such as doxorubicin or tyrosine kinase inhibitor [12- 14]. However, there is no report on the development of autophagy inhibitor-loaded nanocarriers to overcome autophagy- related radioresistance during RT even though RT is one of the standard methodologies in cancer therapy, along with surgery and chemotherapy.

Herein, we developed HCQ-loaded hollow mesoporous silica nanoparticles (HMSNs) to enhance the RT of cancers (Fig. 1). HMSNs, which have a large cavity inside the silica shell, have recently gained considerable interest as nanocarrier platforms for cancer imaging and therapy because of their many attractive properties, such as high capacity of drug loading, pore size tunability, bio-degradability, biocompatibility, good stability, and variety of possible surface functionalization [15,16]. HMSNs, unlike liposomal formulations, can be prepared with uniform size distributions on a large scale. In addition, their drug release kinetics could be controlled by modulating their pore sizes [17]. We expected that the encapsulation of HCQ into HMSNs might enhance the efficiency of HCQ delivery into tumor cells. In particular, the enhanced permeability and retention (EPR) effect of the HMSNs might increase the HCQ accumulation in the tumor without affecting the cellular uptake, due to the acidic extracellular pH of tumor microenvironments, while preventing nonspecific side effects. When combined with radiation therapy, HCQ-loaded HMSNs (HCQ-HMSNs) could effectively prevent tumor growth in vivo through autophagy inhibition.

2. Materials and methods
2.1 Materials

Hydroxychloroquine sulfate (HCQ) was obtained from Tokyo Chemical Industry Co., Ltd (Tokyo, Japan). Tetraethyl orthosilicate (TEOS), ammonium hydroxide, cetyltrimethylammonium chloride solution (CTAC), triethylamine (TEA), sodium carbonate, sodium chloride, (3-aminopropyl) triethoxysilane (APTES), dimethyl sulfoxide (DMSO), succinic anhydride, N-(3-dimethylaminopropyl)-N’-ethylcarbodiimide hydrochloride (EDC), and phosphatase inhibitor were purchased from Sigma-Aldrich (St. Louis, MO, USA). Absolute ethanol, a Pierce bicinchoninic acid (BCA) Protein Assay kit, and 4’,6-diamidino-2- phenylindole (DAPI) were purchased from Thermo Fisher Scientific (Waltham, MA, USA). ATTO 680 NHS ester was obtained from ATTO-TEC. Methoxy-PEG-amine (5 kDa) was purchased from NANOCS (New York, NY, USA). A protease inhibitor cocktail was obtained from Roche Applied Bioscience. CCK- 8 cell viability assay kits were obtained from Dojindo (Kumamoto, Japan). Primary antibodies for p62 (610832), light chain 3B (LC3B; 2775), and Tubulin (SC-5286) were purchased from BD Biosciences (San Jose, CA, USA), Cell signaling Technology, and Santa Cruz Biotechnology, respectively.

2.2 Preparation of HCQ-HMSNs

As proposed by Chen et al. [18], 50 nm sized HMSN was synthesized with improved Na2 CO3 etching method. Based on a modified Stöber method [18,19], we first synthesized 50 nm-sized SiO2 nanospheres. Deionized water (DW, 5 mL) and ammonium hydroxide (0.4 mL) were dissolved in absolute ethanol (35.7 mL) and stirred for 10 min at 25°C. TEOS (1 mL) was added and reacted for 1 h at 25°C. Then, SiO2 nanospheres were washed with ethanol three times and dissolved in DW (20 mL). CTAC solution (8.264 mL, 25 wt/v in H2O) and TEA (27.6 µL) were dissolved in DW (11.7084 mL) and stirred for 1 h at 25°C to form the mesoporous silica shell around the SiO2 core. As-synthesized SiO2 solution (10 mL) was added and stirred for 1 h at 25°C. Then, TEOS (0.15 mL) was added, and the mixture was reacted for 1.5 h at 80°C. After cooling the mixture to 50°C, sodium carbonate (954 mg) was added and reacted for 30 min at 50°C. The HMSN pellets were washed with a 1 wt% solution of NaCl in methanol and ethanol, consecutively. The obtained HMSNs were dissolved in absolute ethanol (20 mL) and mixed with APTES (1 mL) to modify their surface with an amine group. The mixture was reacted for 18 h at 88°C in an oil bath. Then, the product was washed with ethanol three times.For the conjugation of near-infrared (NIR) fluorophores on the surface of the HMSNs, 30 mg of HMSNs (10 mg/mL DMSO solution) were mixed with ATTO 680 Renewable biofuel NHS ester (49.7 µg, 60 nmol) and stirred for 30 min at 25°C. Then, the product was washed with DMSO three times. The amines remaining on the HMSNs were succinylated to create carboxylates by reacting succinic anhydride (30 mg) for 18 h at 25 °C. Afterward, the as-synthesized HMSNs in DMSO were mixed with methoxy-PEG-amine (90 mg, 5 kDa) and EDC (11.5 mg) for 2 h at 25°C. Then, the product was washed with DMSO twice and with DW once.For the HCQ loading into the HMSNs, 30 mg of HMSNs (3 mg/mL DW solution) were mixed with HCQ (600 mg) and stirred for 18 h at 25°C. Then, the final product was washed with DW.

2.3 Characterization of the HCQ-HMSNs

The morphology of the prepared HMSNs was observed by transmission electron microscopy (TEM; JEM-F200, JEOL Ltd. Japan). The amount of HCQ loaded was measured through the UV-Vis absorption spectrum at 342 nm (DU730, Beckman Coulter, Brea, CA). For the in vitro drug release test at different pH conditions, 200 µL of HCQ-HMSNs (1 mg/mL, 11.6 wt% HCQ loading) were added into the dialysis tube (D-Tube Dialyzer Minis, MWCO 12- 14 kDa), and then the tube was immersed into 4 mL of sodium phosphate buffer solution (PBS) at pH 7.4 or 5.0 at 37°C. At 0.5, 1, 2, 4, 6 , 9, 12, and 24 h post-incubation, the buffer solutions were replaced with fresh solutions. Then, the drug release profile of the HCQ-HMSNs was monitored by analyzing the UV-Vis absorption spectrum at 342 nm. The experiments were performed three times.

2.4 In vitro cytotoxicity test

The HCT116 colon carcinoma cell line was obtained from the American Type Culture Collection (ATCC, Manassas, VA). HCT116 cells were cultured in RPMI- 1640, which contains 10% (v/v) fetal bovine serum (FBS) and 1% antibiotic-antimycotic, in a humidified 5% CO2 atmosphere at 37°C. HCT116 cells were seeded into 96-well plates at a density of 5 × 103 cells/well and incubated overnight for cell attachment. The cells were treated with different concentrations of free HCQ and HCQ-HMSNs (i.e., 0, 6.25, 12.5, 25, 50, and 100 μM HCQ eq.) for 24 h. For comparison, the cells were also treated with various concentrations of HMSNs without HCQ loading (17.4, 34.8, 69.6, 139.2 and 278.4 μg/mL; 278.4 μg/mL HMSNs corresponding to HCQ-HMSNs at 100 μM HCQ eq.) to check the cytotoxicity of the carrier itself. Then, the cells were washed three times with fresh cell culture medium and further incubated for 48 h. Cell viability was measured by the CCK- 8 Isoxazole 9 solubility dmso assay kit and calculated relative to that of the untreated control cells.

2.5 Analysis of intracellular uptake of HCQ and HCQ-HMSNs

Intracellular uptake of free HCQ and HCQ-HMSNs was analyzed using high performance liquid chromatography (HPLC) according to a previous report [20]. Briefly, HCT116 cells were seeded into 6-well plates at a density of 3 × 105 cells/well. The cells were treated with free HCQ or HCQ-HMSNs at 100 μM eq. for 24 h. Then, after washing the cells, they were harvested and lysed in ice. The samples were diluted to a protein concentration of 5 mg/mL and used for the analysis. Methanol (80 µL) was added to 20 µL lysate solution, followed by mixing for 5 minutes and centrifugation at 14,000 rpm at 4 °C for 10 min. Then, 10 µL of deionized water was added to 70 µL of the supernatant and centrifuged at 14,000 rpm at 4 °C for 10 min. The supernatant solutions (10 µL) were analyzed by HPLC (Waters Alliance HPLC, Waters Corporation, Milford, MA). HCQ was detected at 342 nm absorbance.

2.6 Analysis of Intracellular location of HCQ-HMSNs

To analyze the intracellular location of HCQ-HMSNs, confocal fluorescence microscopy of the HCQ-HMSN-treated cells was performed. HCT116 cells were seeded at a density of 3× 104 cells/well onto Lab-Tak slide II. The cells were treated with ATTO 680-conjugated HCQ-HMSNs (5 µM HCQ eq.) and incubated for 24 hat 37 °C. After washing the cells three times, cell culture medium containing the LysoTracker® Blue DND-22 probe was added to the cells to stain the lysosomes of the cells. After 30 min, the cell culture medium was replaced with fresh media. Fluorescence images of LysoTracker (λex 373 nm, λem 422 nm) and ATTO 680-conjugated HCQ-HMSNs (λex 633 nm, λem 650 nm long-pass filter), and of the cells were observed using confocal fluorescence microscopy (CSLM, ZEISS LSM 780, Jena,Germany).For comparison, HCT116 cells were treated with either the cell culture medium or free ATTO 680 dye (5 µM) under the same conditions mentioned above, and their fluorescence images were obtained.

2.7 Western blotting

HCT116 colon cancer cells were cultured into 6-well plates at a density of 3 × 105 cells/well. The cells were exposed to X-ray irradiation (6 Gy) and incubated overnight. Then, the cells were incubated with HCQ or HCQ-HMSNs at 20 μM for 4 h. The cells were harvested in ice-cold RIPA lysis buffer (50 mM Tris-Cl, pH 7.4, 150 mM NaCl, 1% NP-40, 0.5% Na-deoxycholate, 0.1% SDS, 1 mM EDTA) containing protease inhibitor cocktail and phosphatase inhibitor. Soluble lysate fractions were isolated by centrifugation at 20,000 × g for 20 min at 4°C and quantified using the BCA Protein Assay kit. Samples were resolved by SDS PAGE using equal concentrations of protein and were transferred to PVDF membranes. The membranes were blocked with 5% skim milk and then probed with the indicated primary and secondary antibodies following standard protocols. Immunoblotted proteins were quantified using Image J software (NIH, Bethesda, MD) . The experiments were performed three times.

2.8 Confocal microscopy study for visualization of autophagy inhibition

For fluorescence imaging of autophagy inhibition, GFP-LC3 in the MigRI-based retroviral vector was generously provided by Craig Thompson (MSKCC; NY. USA) and pBabe-mCherry-GFP-LC3 (#22418) was a gift from Jayanta Debnarth through Addgene. HCT116 cells stably expressing GFP-LC3 or mCherry-GFP-LC3 were generated following standard protocols for retrovirus transduction.HCT116 cells stably expressing the fluorophore-tagging autophagy marker protein, GFP-LC3 or mCherry-GFP-LC3 were seeded at a density of 3 × 104 cells/well onto Lab-Tak slide II. On the next day, the cells were exposed to X-ray irradiation (6 Gy) and incubated overnight. Then, the cells were incubated with HCQ or HCQ-HMSNs at 20 μM HCQ eq. for 4 h. Then, the nuclei of the cells were stained using DAPI. Fluorescence images of DAPI (λex 405 nm, λem 446 nm), GFP (λex 488 nm, λem 521 nm), and mCherry (λex 561 nm, λem 647nm) were obtained using an CSLM (ZEISS LSM 780) and quantified using the software ZEN Black (Zeiss, Oberkochen, Germany). Autophagosome signals from mCherry-GFP-LC3 or GFP-LC3 were captured in at least five distinct fields from different regions for an individual experimental set.

2.9 HCQ-HMSNs tumor targeting

An in vivo NIR fluorescence imaging study was conducted by injecting HCT116 cells (5 × 106 cells/0.1 mL HBSS) subcutaneously into the right hind flank of Balb/c nude mice. Six mice with HCT116 tumors ~200 mm3 in size received intravenous injections of the ATTO 680-conjugated HCQ-HMSNs (20 mg HCQ eq./kg, n = 3). The injection dose of HCQ was determined by referring to a previous paper [11]. For comparison, the mice in the control groups received intravenous injections of 100 μL PBS and free ATTO 680 dye-containing PBS (n = 3 per group). Then, NIR fluorescence images (λ ex = 660/20 nm, λem = 710/40 nm) of the mice were obtained using the IVIS Lumina imaging system (Xenogen Corporation-Caliper, Alameda, CA) 5 min and 24 h after injection. For biodistribution analysis, the mice were sacrificed 24 h after injection, and ex vivo NIR fluorescence imaging of the tumors and organs (spleen, kidneys, liver, lungs, and heart) was carried out.

2.10 In vivo anti-tumor effect of HCQ-HMSNs

The antitumor effect of combinations of HCQ-HMSN and RT was explored using HCT116 xenograft colon cancer models. When the tumors reached an appropriate size (~ 100 mm3 ), the mice were randomized into five groups (six mice per group) on day zero: group 1, PBS control (100 μL, two times); group 2, RT (3 Gy) alone; group 3, HCQ (20 mg HCQ eq./kg, two times) + RT (3 Gy); group 4, HCQ-HMSN (20 Women in medicine mg HCQ eq./kg, two times) alone; group 5, HCQ-HMSN (20 mg HCQ eq./kg, two times) RT (3 Gy). PBS (control group), HCQ, and HCQ-HMSNs were injected intravenously via the tail vein of the mice on day 1. X- ray irradiation of tumors was done after 24 h of the drug treatment (i.e., on day 2). Then, PBS, HCQ, and HCQ-HMSNs were injected intravenously once more immediately after the RT. Tumor growth and body weight were measured every day for 14 days. All the mice were sacrificed on day 14. Then, the heart, lungs, liver, spleen, and kidney were collected from the experimental group. The organ samples were sectioned and stained with hematoxylin and eosin (H&E) to monitor pathological changes.

To investigate the effect of the HCQ-HMSNs on the inhibition of radiation-induced autophagy, another set of tumor-bearing mice was prepared (n = 3 per group) for the immunohistochemical staining of the LC3B levels in tumors. The tumor tissues were harvested 24 h after RT (i.e., on day 3) and sliced. The tumor sections were stained with anti- LC3B antibodies (Abcam, Cambridge, MA) to assess the change in the autophagy. For apoptosis analysis, caspase-3 assay of the tumor sections on day 3 was also performed. Slides were digitally scanned in a high-resolution bright field with a Vectra PolarisTM Automated Imaging System (PerkinElmer, Hopkinton, MA), and the resulting images were analyzed with the InForm software (PerkinElmer) for automated quantification.

2.11 Serum biochemical analysis

To analyze the potential toxicity in vivo, eight Balb/c nude mice without tumors were used. PBS (100 μL) or HCQ-HMSN (20 mg HCQ eq./kg) were injected intravenously via the tail vein of the non-tumor-bearing mice twice on day 1 and day 2, respectively. On day 14, serum samples were collected from the mice of the PBS- or HCQ-HMSN-treated group. Alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), albumin (Alb), total protein (TP), A/G ratio, total bilirubin (T-BIL), total cholesterol (T-Chol), triglycerides (TG), glucose, blood urea nitrogen (BUN), and creatinine (Crea) were analyzed at Biotoxtech Co., Ltd. (Cheongwon, Chungbuk, Korea) .All animal experiments were approved by the Institutional Animal Care and Use Committee.

2.12 Statistical analysis

All data arepresented as mean ± S.D. The significant difference between test groups was carried out by Student’s t-test.

3. Results and discussion
3.1 Characterization of HCQ-HMSNs

HMSNs with diameters of ~ 50 nm were synthesized using hard-soft templates (i.e., dense SiO2 and CTAC) with theNa2 CO3 etching method. Based on a modified Stöber method, we first synthesized SiO2 nanospheres with diameters of 48.8 ± 7.32 nm. For forming the mesoporous silica shell on the surface of this SiO2 core, CTAC was used as a soft template at 80°C. It has been documented that, as a cationic surfactant, CTAC plays an important role in the selective silica etching and redeposition mechanism [21]. Followed by Na2 CO3 etching at 50°C, HMSNs were obtained with a hollow core diameter of 43.3 ± 6.15 nm and a shell thickness of 10.2 ± 1.99 nm (Fig. 2a). Afterward, the surface of the HMSNs was modified with amine groups using APTES and was then conjugated with ATTO 680 dye for NIR fluorescence imaging. To enhance the stability of the HMSNs, succinylation and PEGylation were conducted using succinic anhydride and methoxy-PEG-amine (5 kDa), respectively. The zeta-potential values of HMSNs-NH2 , succinylated HMSNs (HMSNs-COOH), and PEG- conjugated HMSNs (HMSNs-PEG) were 41.15, -26.50, and -4.65 mV, respectively, confirming successful succinylation and PEGylation of the HMSN’s surface (Fig. S1a, Supporting Information). The fluorescence spectrum of ATTO 680-conjugated HMSNs also confirmed conjugation of the dyes to HMSNs (Fig. S1b, Supporting Information). The prepared HMSNs (HMSNs-PEG) showed good dispersion stability for at least 24 h when dispersed in FBS-containing cell culture medium (Fig. S2, Supporting Information). The HMSNs was then loaded with HCQ to form HCQ-HMSNs. From the absorbance measurement of the HCQ-HMSNs from UV-Vis spectra at 342 nm (Fig. 2b), the loading content of HCQ in the HMSNs was calculated to be 11.6 wt/wt%. The drug release profiles of the HCQ-HMSNs showed similar release rates at physiological (pH 7.4) and lysosomal (pH 5.0) pH conditions at 37°C (Fig. 2c). Therefore, unwanted excessive release of HCQ from HMSNs in the acidic extracellular region of tumor tissues could be prevented. As aforementioned, the acidic tumor microenvironment is known to inhibit the ability of free CQ and HCQ to cross cell membranes, thereby significantly reducing their intracellular uptake.The similar release rate at the two different pH conditions indicates that diffusion of the loaded HCQs through the pores was the major mechanism of HCQ release from HMSNs.According to the report by Gao et al., doxorubicin (molecular weight: 543.5 Da) can pass through the pores of HMSNs only one by one when the pore diameters are ≤ 6.4 nm.[17] The amount of
doxorubicin released from HMSNs with a pore diameter of 3.2 nm was about 25 % during incubation for 24 h at a pH of 5.0. Therefore we can conjecture that HCQ (molecular weight: 335.8 Da) was slowly released from HMSNs one by one through the pores with diameters of approximately 2.2 nm [18].

3.2 In vitro cytotoxicity and intracellular uptake tests

An in vitro cytotoxicity test was performed to check enhanced delivery of HCQ by using HMSNs and the consequent increase in its therapeutic effect. HCT116 colon carcinoma cells were treated with free HCQ and HCQ-HMSNs at various concentrations for 24 h. No cytotoxic effect was observed in the HCQ-treated cells up to a concentration of 100 μM (Fig. 3a, left graph). However, the viability of the HCQ-HMSN-treated cells significantly decreased to 91.6, 83.4, 49.0, and 21.2% at treatment concentrations of 12.5, 25, 50, and 100 μM HCQ eq., respectively. As HMSNs without HCQ loading did not induce any cytotoxic effect at the tested concentrations (Fig. 3a, right graph), the high cytotoxic effect in the HCQ- HMSN-treated HCT116 cells indicates enhanced delivery of HCQ into the cells. In fact, when the amounts of intracellular uptake of free HCQ and HCQ-HMSNs (100 μM HCQ eq.,) were analyzed by the HPLC method (Fig. 3b), HCQ-HMSN-treated cells showed a 200-fold higher uptake of HCQ than free HCQ-treated cells.HCQ is known to inhibit autophagy by increasing the lysosomal pH of cancer cells. Therefore, intracellular location of the HCQ-HMSNs was observed using confocal fluorescence microscopy. ATTO 680-conjugated HCQ-HMSNs (5 µM HCQ eq.) were treated to HCT116 cells for 24 h. Then, the lysosomes in the cells were stained with Lysotracker probe. According to Fig. 3c, strong fluorescence signals of the HCQ-HMSNs (red) were observed inside the cells, and most of them were merged with Lysotracker. This data indicates that the HCQ-HMSNs were efficiently taken up into the cells via endocytosis and moved to the lysosomes (green), which are the target sites of the HCQ action. When HCT116 cancer cells were treated with free ATTO 680 dye at 5 µM for comparison, no cellular uptake of the free dye was observed (Fig. 3c).

3.3 Cellular autophagy inhibitory effect of HCQ-HMSNs

Next, the inhibitory effect of the HCQ-HMSNs on cellular autophagy was investigated. The well-known autophagy marker proteins, microtubule-associated protein 1 LC3B and
SQSTM1 (p62), were examined by Immunoblot. HCT116 colon cancer cells were irradiated at 6 Gy and further incubated overnight; then, they were treated with free HCQ or HCQ – HMSNs at 20 μM for 4h, respectively. Immunoblot analysis of cell lysates from irradiation itself showed decreasing levels of the cytosolic form of LC3-I and the autophagy receptor p62 compared to those of non-irradiated control cells (r < 0.01; Fig. 4a), indicating that the irradiation induced sufficient autophagy in the cells. As expected, the treatment of the cells with the autophagy inhibitor HCQ increased the levels of LC3-II in autophagosome-bound form (r < 0.01) but did not decrease the p62 levels. These results suggested that the autolysosomes formed by fusion of autophagosome and lysosome were not effectively degraded owing to the autophagy inhibitory effect of HCQ; thus, LC3 -II in autophagosome- bound form was accumulated before lysosomal degradation. It is noteworthy that, upon treatment with HCQ-HMSNs, membrane-bound LC3-II was accumulated more than in the HCQ-treated cells, and the lysosomal degradation of p62 was more effectively prevented than in the HCQ treatment. The ratio of LC3-II to LC3-I was higher in the HCQ-HMSN-treated cancer cells than in the HCQ-treated cells, especially after irradiation (r < 0.05; Fig. 4a, right graph).

Confocal microscopy analysis was performed to observe the inhibition of the radiation- induced autophagy by HCQ-HMSNs (Figs. 4b and 4c). The HCT116 cells that were stably expressing GFP-LC3 or mCherry-GFP-LC3 were used to visualize the formation of autophagosomes and autolysosomes. Similar to the immunoblot results revealing LC3 conversion (Fig. 4a), the levels of GFP-LC3 puncta representing autophagosomes substantially increased in response to X-ray irradiation itself (r < 0.05; Figs. 4b and S3a, Supporting Information) compared to the cytoplasmic diffused form of GFP-LC3 under the normal condition without irradiation. Treatment of the X-ray irradiated cells with either HCQ or HCQ-HMSNs further enhanced the levels of GFP-LC3 puncta, suggesting that LC3 II- conjugated autophagosomes tend to accumulate in the cytoplasm by blocking lysosome functions. The efficacy of HCQ-HMSNs-mediated autophagy inhibition, such as cytosolic accumulation of autophagosomes, was better than that of HCQ, as the size and number of GFP-LC3 puncta in the HCQ-HMSN-treated cells was significantly higher than that in the free HCQ-treated cells (P < 0.01; Fig. S3a, Supporting Information).

Autophagy flux analysis with LC3 fused with the tandem fluorophore proteins mCherry- GFP-LC3 (Fig. 4c) was also performed to further confirm the enhanced inhibitory effect of HCQ-HMSNs on the final step of autophagy (i.e., autolysosome formation). The mCherry (Red) fluorophore is known to be stable in acidic conditions, while GFP signals are easily quenched. Based on these unique features of fluorophores, formation of autolysosomes with low pH lysosomes generates an increasing number of mCherry puncta, whereas most autophagosomes that accumulated in the cytoplasm at a neutral pH were observed to have fluorescent signals with yellow (due to the color merging of mCherry and GFP) puncta for LC3 [22, 23]. In fact, HCT116 cells following X-ray irradiation showed a significant increase in red colored puncta (P < 0.01; Figs. 4c and S3b, Supporting Information), indicating the formation of autolysosomes. Both HCQ- and HCQ-HMSN-treated cells showed an increase in the yellow (mCherry and GFP) puncta in the cytoplasm (P < 0.05), which indicates inhibition of autolysosome formation in the X-ray irradiated cancer cells. The number and size of the yellow puncta were substantially higher in the HCQ-HMSN-treated cells than in the free HCQ-treated cells (P < 0.05). These data confirmed that HCQ-HMSNs inhibit radiation- induced autophagy by suppressing autophagy flux more efficiently than free HCQ. According to previous reports [24,25] and Fig. 3b, the intracellular uptake of drug-loaded HMSNs is much higher than that of free drugs. In addition, a cytotoxicity test in HCT116 cells showed enhanced cytotoxicity in the HCQ-HMSN-treated cells compared to HCQ-treated cells (Fig. 3a). These results with drug release data explain the reason why HCQ-HMSNs block the autophagy flux more efficiently than HCQ.

3.4 Tumor targeting performance of HCQ-HMSNs

To confirm the passive tumor targeting ability of HCQ-HMSNs, a colon cancer model was produced in Balb/c nude mice using the HCT116 cell line. ATTO 680-conjugated HCQ- HMSNs (3 mice, 20 mg HCQ eq./kg) were administered systemically via the mouse’s tail vein. The distribution of HCQ-HMSNs was observed with NIR fluorescence images (λ ex = 660/20 nm, λem = 710/40 nm) 5 min and 24 h after injection. As shown in Fig. 5, HCQ-HMSNs were evenly distributed in the entire body of the mouse right after the intravenous injection (i.e., 5 min). Then, strong fluorescence signals were detected in the tumor sites 24 h post-injection, indicating the accumulation of HCQ-HMSNs in the tumor tissues via the EPR effect. Ex vivo images of the tumors and major organs confirmed the accumulation of HCQ-HMSNs in the tumor tissues. When free ATTO 680 dyes as a hydrophilic model drug were injected into the tumor-bearing mice, NIR fluorescence signals from the body disappeared in 3 h, indicating rapid clearance of the hydrophilic dye from the body without its accumulation in tumor tissues (Fig. S4, Supporting Information). These data support the benefit of HMSNs as a delivery carrier of hydrophilic HCQ molecules for passive tumor targeting.

3.5 Enhanced radiotherapy by HCQ-HMSNs

To evaluate whether HCQ-HMSNs increase the sensitivity of tumors to RT, mouse xenograft tumors were treated with PBS (control group), RT alone, HCQ plus RT, HCQ – HMSN alone, and HCQ-HMSN plus RT. Free HCQ and HCQ-HMSNs were injected intravenously on day 1 and day 2, respectively, and RT was performed on day 2. The RT treatment alone led to inhibition of tumor growth by 41.1% compared to the control group (**P < 0.01). However, the combination of HCQ with RT did not enhance the therapeutic efficacy of RT. Interestingly, the treatment with only HCQ-HMSNs resulted in tumor growth inhibition by 31.1% (*P < 0.05). Notably, the mean tumor size of the HCQ-HMSN plus RT- treated group was greatly reduced by 82.8% and 70.8% compared to the control group and RT-treated group on day 14, respectively (***P < 0.001, Fig. 6a).

After 24 h of RT (i.e., on day 3), 2.3-fold increases in the LC3B levels (*P < 0.05) in tumor sections were observed compared to the control group (Fig. 6b). When RT was combined with HCQ-HMSNs, the LC3B levels were further increased (5.7-fold compared to the control group, ***P < 0.001; 2.5-fold compared to the RT-treated group, **P < 0.01), which is consistent with the in vitro study (Fig. 4a). As mentioned above, autophagy protects cancer cells from multiple stresses and delays apoptotic cell death in response to DNA damage. Therefore, the tumor sections on day 3 were also stained with the caspase-3 assay kit for apoptosis analysis (Fig. 6c). As a result, a significant increase in apoptotic cells was observed in the HCQ-HMSN-treated tumors compared to the tumors in the other groups (*P < 0.05), indicating that the HCQ-HMSNs accumulated in the tumors sensitized cancer cells to RT.

Previous reports showed that achieving a sufficiently high concentration of HCQ for autophagy inhibition in target tumors is difficult because of their nonspecific biodistribution in vivo after administration and the acidic tumor microenvironment [10]. In this study, no therapeutic benefit was observed in the HCQ and RT group, which indicates insufficient delivery of HCQ to tumor tissues at the tested dose. In contrast, tumor growth inhibition by HCQ-HMSN alone and the data in Fig. 6 support the achievement of efficient delivery of HCQ to tumor tissues. Wang et al. also showed that HCQ-loaded liposomes lead to improved accumulation of the drug in the tumors and a better chemotherapeutic effect on B16F10 melanoma compared with the HCQ treatment [12]. In the current study, the combined HCQ- HMSNs and RT treatment lead to significant enhancement in the tumor growth compared with the RT treatment alone owing to the efficient delivery of HCQ to the tumor tissues and the increase in the sensitivity to RT by autophagy inhibition.

3.6 In vivo biocompatibility of HCQ-HMSNs

Next, we assessed the biocompatibility of HCQ-HMSNs by analyzing the histopathology of major organs, blood biochemistry profiles, and body weight changes of the mice (Fig. 7 and Fig. S5, Supporting Information). No significant histological changes of the organs were observed in any of the groups and none of the blood biomarkers was significantly varied by the treatment with HCQ-HMSNs compared with those in the PBS-treated control group. Furthermore, there was no statistical difference in the bodyweight of the mice in each group compared with those of the control group. These results confirmed the biocompatibility of HCQ-HMSNs.

4. Conclusions

Here, we showed that, as radiosensitizers, HCQ-HMSNs effectively inhibit radiation- induced autophagy, thus overcoming radioresistance. The HMSNs enhanced the delivery of HCQ to tumors and significantly inhibited the radiation-induced autophagy and induced cancer cells apoptosis and tumor growth inhibition. Therefore, HCQ-HMSNs are a new effective and safe nanomedicine strategy for combining of conventional RT and the autophagy inhibitor HCQ through systemic delivery to tumor tissues.