Tumor Microenvironment Modulation by ... - ACS Publications

Aug 29, 2017 - ABSTRACT: Due to the rich stroma content and poor blood perfusion, pancreatic ductal adenocarcinoma (PDA) is a tough cancer that can ...
13 downloads 0 Views 3MB Size
Research Article www.acsami.org

Tumor Microenvironment Modulation by Cyclopamine Improved Photothermal Therapy of Biomimetic Gold Nanorods for Pancreatic Ductal Adenocarcinomas Ting Jiang,†,‡,§ Bo Zhang,†,§ Shun Shen,‡ Yanyan Tuo,‡ Zimiao Luo,‡ Yu Hu,*,† Zhiqing Pang,*,‡ and Xinguo Jiang‡ †

Institute of Hematology, Union Hospital, Tongji Medical College, Huazhong University of Science & Technology, Wuhan, Hubei 430022, P. R. China ‡ School of Pharmacy, Fudan University, Key Laboratory of Smart Drug Delivery, Ministry of Education, 826 Zhangheng Road, Shanghai 201203, P. R. China S Supporting Information *

ABSTRACT: Due to the rich stroma content and poor blood perfusion, pancreatic ductal adenocarcinoma (PDA) is a tough cancer that can hardly be effectively treated by chemotherapeutic drugs. Tumor microenvironment modulation or advanced design of nanomedicine to achieve better therapeutic benefits for PDA treatment was widely advocated by many reviews. In the present study, a new photothermal therapy strategy of PDA was developed by combination of tumor microenvironment modulation and advanced design of biomimetic gold nanorods. On one hand, biomimetic gold nanorods were developed by coating gold nanorods (GNRs) with erythrocyte membrane (MGNRs). It was shown that MGNRs exhibited significantly higher colloidal stability in vitro, stronger photothermal therapeutic efficacy in vitro, and longer circulation in vivo than GNRs. On the other hand, tumor microenvironment modulation by cyclopamine treatment successfully disrupted the extracellular matrix of PDA and improved tumor blood perfusion. Moreover, cyclopamine treatment significantly increased the accumulation of MGNRs in tumors by 1.8-fold and therefore produced higher photothermal efficiency in vivo than the control group. Finally, cyclopamine treatment combined with photothermal MGNRs achieved the most significant shrinkage of Capan-2 tumor xenografts among all the treatment groups. Therefore, with the integrated advantages of tumor microenvironment regulation and long-circulation biomimetic MGNRs, effective photothermal therapy of PDA was achieved. In general, this new strategy of combining tumor microenvironment modulation and advanced design of biomimetic nanoparticles might have great potential in PDA therapy. KEYWORDS: tumor microenvironment modulation, cyclopamine, biomimetic gold nanorods, erythrocyte membrane, photothermal therapy, pancreatic ductal adenocarcinomas



INTRODUCTION Photothermal therapy (PTT), in which photon energy is converted to thermal energy upon irradiation, is a minimally invasive hyperthermic treatment method for tumors,1−3 and has aroused extensive attention worldwide. Different types of nanoparticles including those based on gold, carbon, copper, and iron, etc., have been used for PTT for a wide range of tumors and have achieved therapeutic benefits to various extents.1,2,4 Unlike conventional therapeutics, these nanoparticles can display a preferential tumor accumulation based on enhanced permeability and retention (EPR) effect and thus mediate a selective hyperthermia effect upon precise irradiation with near-infrared light. However, as far as we are concerned, research studies about PTT for pancreatic ductal adenocarcinoma (PDA), a common and lethal malignancy resulting in tens of thousands of deaths every year,5 were still limited. The main reasons behind this phenomenon might be due to the poor delivery of photothermal nanoparticles to PDA which was © 2017 American Chemical Society

mainly hindered by the highly desmoplastic response of PDA and the short circulation of nanoparticles used for PTT. PDA is well-known for the highly desmoplastic property and consequently the rich extracellular matrix (ECM).5,6 Besides, tumor vessels in PDA were also very scarce. The dense ECM not only isolated tumor cells into tumor nests, but also compressed tumor vessels to reduce tumor blood perfusion. The compromised tumor perfusion plus the blocking resistance from the dense ECM could reduce the transport of nanoparticles to arrive at the tumor site and the following penetrating process to finally reach tumor cells.7 The highly desmoplastic response of PDA was regulated by many signaling pathways, among which the hedgehog (Hh) signaling pathway played a crucial role.8−12 As known to all, Hh ligands derived Received: July 1, 2017 Accepted: August 29, 2017 Published: August 29, 2017 31497

DOI: 10.1021/acsami.7b09458 ACS Appl. Mater. Interfaces 2017, 9, 31497−31508

Research Article

ACS Applied Materials & Interfaces

Figure 1. Schematic graph of MGNRs preparation and MGNRs delivery in stroma-rich PDA with and without cyclopamine treatment. Without cyclopamine treatment, the vessels are compressed by surrounding ECM components including collagen and fibronectin, which lead to poor blood perfusion and limited MGNRs delivery. In contrast, after treatment with cyclopamine, ECM components were disrupted, and tumor vessels were decompressed, resulting in enhanced blood perfusion and more effective delivery of MGNRs. Thus, the PTT effect induced by irradiated MGNRs could be improved.

and narrow spectral bandwidth, were selected as the model photothermal nanoparticles. As far as we were concerned, it was the first time that tumor microenvironment modulation and advanced design of photothermal nanoparticles were combined to achieve better PTT benefits of PDA. In this study, GNRs were synthesized by a seed-mediated growth method and were further coated with RBC membrane (MGNRs).21 In addition, the photothermal efficiency and long circulation of MGNRs were assessed by PTT tests in vitro and pharmacokinetics and biodistribution studies in vivo, and compared with those of GNRs. Furthermore, the effect of cyclopamine treatment on the tumor microenvironment of PDA including ECM morphology and tumor perfusion was investigated. Finally, the therapeutic benefit of MGNRs plus cyclopamine treatment was evaluated.

from tumor cells could act on the Patched1 receptor in tumorassociated fibroblasts to alleviate the inhibition of the 12transmembrane protein Smoothened (SMO), resulting in activation of the transcription factor glioma-associated oncogene family zinc finger-1 (GLI-1) and expression of ECM-related proteins including collagens and fibronectins.8 Consequently, cyclopamine, a specific inhibitor of the Hh signaling pathway, which was able to significantly inhibit the desmoplastic response, might be utilized to modulate the tumor microenvironment of PDA to help enhance nanoparticle delivery to PDA (Figure 1).13,14 It was well-documented that the circulation time of nanoparticles was a critical factor in dominating tumor nanoparticle accumulation, especially for tumors like PDA with scarce tumor vessels. Therefore, the circulation time of those nanomaterials for PTT was also a crucial issue deserving further exploration to improve PTT efficacy. Red blood cell (RBC) membrane-coated nanomedicines were reported to harbor the properties of RBC such as minimal uptake by phagocytes and long-circulation lifetime and have attracted increasing attention recently.15−18 It was now well-known that RBC membrane modification on nanoparticles could significantly increase the circulation time of nanoparticles compared with PEGylation, a classical method used for avoiding phagocyte uptake and increasing circulation time of nanoparticles. Furthermore, RBC membrane modification would not significantly impact the photothermal conversion efficiency of photothermal nanoparticles.19,20 Therefore, in this study, RBC membrane might be utilized to modify photothermal nanoparticles to elongate their circulation time, and help increase their delivery to PDA (Figure 1). To test our hypothesis mentioned above, modulation of ECM of PDA and optimization of the circulation time of photothermal nanoparticles simultaneously were performed to obtain more PTT benefits of PDA (Figure 1). Human pancreatic cancer cell line Capan-2 xenografts with rich ECM were used as the tumor models. Cyclopamine was used as the modulator of ECM of PDA, and gold nanorods (GNRs), which have aroused extensive attention for PTT because of advantages including small size, high absorption coefficient,



MATERIALS AND METHODS

Materials. Chloroauric acid (HAuCl4·3H2O), cetyltrimethylammonium bromide (CTAB), silver nitrate (AgNO3), sodium borohydride (NaBH4), L-ascorbic acid (AA), and ethylenediaminetetraacetic acid EDTANa2 were purchased from Aladdin (Shanghai, China). 3-(4,5Dimethylthiazol-2-yl)-2,5-diphenyltetrazdium bromide (MTT) assay and 11-mercaptoundecaonic acid (MUDA) were purchased from Sigma-Aldrich (Saint Louis, MO). Roswell Park Memorial Institute (RPMI) 1640 medium, fetal bovine serum (FBS), penicillin− streptomycin solution, and trypsin−EDTA solution were purchased from Gibico. Cylopamine was obtained from Struchem Co., Ltd. (Wujiang, China). Cremophor EL (polyoxyethlene castor oil derivatives) was purchased from BASF (Ludwigshafen, Germany). Propidium iodide (PI) and 3′,6′-di(O-acetyl)-4′,5′-bis[N,N-bis(carboxymethyl)aminomethyl] fluorescein, tetraacetoxymethyl ester (Calcein-AM), were purchased from KeyGEN BioTECH (China). DyLight488-labeled tomato lectin (Lycopersicon esculentum) was from Vector. CD31 goat polyclonal primary antibody was ordered from R&D. Hoechst 33342 was purchased from Beyotime Biotechnology Co., Ltd. (Nantong, China). The microbubble contrast agent was provided by VisualSonics (Vevo MicroMarkerTM, Toronto, Canada). Human-derived pancreatic cancer Capan-2 cell lines were from American Type Culture Collection (ATCC). The cells were cultured in RPMI 1640 medium supplemented with 10% FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin at 37 °C in a humidified atmosphere containing 5% carbon dioxide (CO2). Male Balb/c nude 31498

DOI: 10.1021/acsami.7b09458 ACS Appl. Mater. Interfaces 2017, 9, 31497−31508

Research Article

ACS Applied Materials & Interfaces

solution with different Au concentrations including 21, 42, and 84 μg/ mL were radiated with an 808 nm laser at 0.75 W/cm2 for 5 min and cooled naturally to room temperature. In Vitro PTT Studies. The in vitro cytotoxicity of the MGNRs upon near-infrared (NIR) irradiation was assessed with MTT assays. Briefly, Capan-2 cells were seeded into a 96-well plate at the density of approximately 104 cells/well, and incubated for 24 h until cells reached approximately 80% confluence. After being washed twice with PBS, cells were treated with a varying concentration of MGNRs and GNRs at 37 °C for 3 h. Afterward, cells were radiated at 0.75 W/cm2 with an 808 nm laser for 5 min. For MTT assay, cells were treated with 0.5 mg/mL of MTT in 1640 medium for 4 h and dissolved with dimethyl sulfoxide after removing the medium. Relative cell viability was determined by monitoring the optical densities at 570 nm with a microplate reader (BioTeK SynergyTM, America). For fluorescent imaging, cells were stained by Calcein-AM (2 μmol/L) and PI (4 μmol/L) and photographed using an inverted fluorescence microscope (LEICA DMI 4000 B, Germany). Calcein-AM is a green fluorescence probe labeling living cells, whereas PI is a red fluorescence probe labeling cells with compromised membranes. In order to find out whether cyclopamine enhances the cancer cellkilling effect of PTT treatment, the in vitro cytotoxicity of the MGNRs upon NIR irradiation in the presence of cyclopamine was also assessed with MTT assays. Briefly, Capan-2 cells were seeded into a 96 well plate and incubated as described above. Afterward, cells were incubated with GNR or MGNR solution with Au content 42 μg/ mL in the presence of varying concentration of cyclopamine at 37 °C for 3 h. Then cells were radiated at 0.75 W/cm2 with an 808 nm laser for 5 min and subjected to MTT assay as described above. Cyclopamine Treatment of Tumor Xenograft-Bearing Mouse Models. To develop PDA-bearing mouse models, Capan-2 cells (5 × 106 cells/100 μL) were subcutaneously implanted into the right hind flank region of nude mice.22 Tumor diameters were determined using a caliper, and tumor volumes (V) were approximated using the following formula: V = 0.5 × a × b2, where a represents the maximum diameter, and b represents the minimum diameter. When the tumor diameter reached 4−6 mm, mouse models were subjected to cyclopamine treatment. Cremophor EL, an FDA-approved surfactant with low toxicity used in oral preparations, injections, and external preparations was used to solubilize water-insoluble cyclopamine for oral administration. Cyclopamine was dissolved in the mixture of ethanol, Cremophor EL, and deionized water with the cyclopamine concentration of 10 mg/mL and orally administrated to mouse models once a day at the dose of 50 mg/kg for 3 weeks. The control group received blank vehicles containing an equal dose of ethanol and Cremophor EL. ECM Modulation by Cyclopamine Treatment. When cyclopamine treatment ended, mouse models were sacrificed, and tumor xenografts were sliced for immunofluorescence staining of fibronectin, a marker of ECM,8,23 as previously described.22 To evaluate the fibronectin disruption effect of cyclopamine treatment, the fluorescence intensity of fibronectin in tumor slices from six randomly assigned regions in each tumor (n = 3) was quantified by the ZEN 2012 software.22 Tumor Perfusion Improvement by Cyclopamine Treatment. When cyclopamine treatment ended, a DyLight 488-lectin-labeling experiment was performed to evaluate the effect of cyclopamine treatment on tumor perfusion as previously described.24 Briefly, DyLight 488-lectin was intravenously injected at the dose of 5 mg/kg 1 h before heart perfusion. Tumor xenografts were collected and sliced for tumor vessel staining by CD31 goat polyclonal primary antibody. After the nuclei were stained with Hoechst 33342, the slices were mounted in Dako flurescent mounting medium. The colocalization of CD31 signals and DyLight 488-lectin signals in the tumor sections from six randomly assigned regions in each tumor (n = 3) were captured at 200 × magnification and further analyzed using the ImageJ software (NIH) to assess tumor perfusion. In Vivo Ultrasound Imaging of Tumor Xenografts. When cyclopamine treatment ended, tumor xenografts of mouse models were subjected to in vivo ultrasound imaging with a Vevo LAZR

mice and Balb/c mice aged 6−8 weeks were bought from the Shanghai Slac Lab Animal Ltd. (Shanghai, China). All animal experiments were carried out in accordance with the protocol evaluated and approved by the Experimental Animal Ethics Committee of School of Pharmacy, Fudan University (2014-03-YJ-PZQ-01). Synthesis of Gold Nanorods. GNRs were synthesized by a seedmediated growth method developed by Jana et al. and Nikoobakht et al.21 Briefly, to make the seed solution, 72 μL of 1% HAuCl4 and 20 mL of 0.1 M CTAB were mixed and stirred vigorously. After 60 μL, ice-cold 0.1 M NaBH4 was added, and the mixed solution was stirred for another 2 min. The brown solution was kept at 25 °C for at least 3 h before further use. To make the growth solution, 110 μL of 0.1 M AA was added dropwise to the yellow mixture of 20 mL of 0.1 M CTAB, 200 μL of 0.01 M AgNO3, 286 μL of 1% HAucl4, and 40 μL of 1 M HCl under gentle stirring. After the growth solution turned colorless, 24 μL of seed solution was added, and the solution was kept at 28 °C with a water bath for 4 h to obtain GNRs. RBC Ghost Derivation. RBC ghosts were prepared as previously reported by Zhang et al.15 Briefly, the whole blood was collected from the Balb/c mice by cheek porch puncture with heparin-treated sterile centrifuge tube and was centrifuged at 800 × g for 5 min at 4 °C to remove the serum and the buffy coat. The resulting packed RBCs were then washed with ice-cold PBS containing 1 mM EDTANa2 3 times and spun down at 800 × g at 4 °C. For hypotonic medium treatment, the washed RBCs were suspended in ice-cold 0.25 × PBS for 10 min and were centrifuged at 20000 × g for 4 min at 4 °C. The released hemoglobin was removed, and the resulting RBC ghosts were collected. Preparation of MGNRs. The obtained GNRs were washed with deionized water twice to remove the excess CTAB by centrifugation at 8000 g for 8 min and resuspended in 5 mL of deionized water. The solution was then mixed with 50 μL of 20 mM MUDA−ethanol solution and kept at room temperature overnight for MUDA activation. After centrifugation at 6000g for 8 min, the GNRs were collected and resuspended in 1.2 mL of RBC ghosts. Then the mixture was sonicated in a capped glass vial at a frequency of 53 kHz and a power of 100 W for 2 min. The resulting MGNRs were kept at 4 °C in dark place for later use. Characterization of MGNRs and GNRs. The concentration of the atomic gold in the MGNRs and GNRs solution was determined by inductively coupled plasma mass (ICP-MS, Nexlon 300X). Briefly, the sample was added into 1 mL of aqua regia, mixed at room temperature overnight, and kept at 60 °C for 12 h to remove acids. The sample was subsequently resuspended in deionized water at a final volume of 5 mL, filtered through a 0.22 μm membrane filter unit (Millipore), and subjected to ICP-MS for Au content quantification. The morphologies of MGNRs and GNRs were directly observed under a transmission electron microscope (TEM, JEM-1400plus, JEOL, Japan) operating at 100 kV. The zeta potentials of MGNRs and GNRs were measured using a Malvern Nano ZS (Malvern Instruments, UK) at 25 °C, respectively. The UV−vis absorption spectra of MGNRs and GNRs were recorded by a UV−vis recording spectrophotometer (UV-2401, Japan), respectively. To investigate the stability of MGNRs, MGNRs and GNRs were dispersed in phosphate buffer saline (PBS, 0.01 M, pH = 7.4) or 10% FBS in water and stored at room temperature. The plasmon resonance peak shifts, absorbance at 808 nm, and zeta potentials of MGNRs and GNRs were monitored at preset time points. In Vitro Photothermal Conversion Effect. The in vitro photothermal conversion effects of MGNRs and GNRs were measured by monitoring the temperature changes of MGNRs and GNRs solution when radiated with an 808 nm laser (Changchun New Industries Optoelectronics Technology, Changchun, China) at varying power densities. Briefly, 100 μL portions of MGNR and GNR solution with Au content of 42 μg/mL in 96 well plates were radiated with an 808 nm laser at different power densities including 0.75, 1, and 2 W/ cm2 for 5 min and cooled down naturally to room temperature for 5 min. The solution temperature was recorded every 10 s during the whole experiment. To assess the effect of Au concentration on the photothermal conversion effect, 100 μL portions of MGNR and GNR 31499

DOI: 10.1021/acsami.7b09458 ACS Appl. Mater. Interfaces 2017, 9, 31497−31508

Research Article

ACS Applied Materials & Interfaces

Figure 2. Characterization of MGNRs. (A) TEM images of GNRs (left) and MGNRs (right). (B) Zeta potentials of GNRs before and after RBC membrane coating. (C) UV−vis−NIR spectra of MGNRs and GNRs dispersed in water and 1 × PBS, respectively. The inset is the photograph taken 2 h after dispersion. (D) Zeta potential of MGNRs at 4 °C during 1 week. (E) UV−vis−NIR spectra of MGNRs dispersed in 10% FBS at 37 °C during 3 days. (F) Absorbance of MGNRs in 10% FBS at 808 nm wavelength at 37 °C during 3 days. imaging system (VisualSonics Inc., Toronto, Canada) to directly visualize the effect of cyclopamine treatment on tumor blood perfusion.22 Briefly, after anesthetization with 5% chloral hydrate, mouse models were placed for ultrasound scanning. Videos and images of the tumor region were acquired before, during, and after a bolus injection of 107 microbubbles into the mice through the tail vein. Parameters associated with the perfusion model function from tumor xenografts were acquired for analysis of tumor perfusion by VevoLAB software.25 MGNR Pharmacokinetics and Biodistribution Studies. For pharmacokinetics studies, male Balb/c mice aged 6−8 weeks were randomly divided into the MGNR group and the GNR group. Each mouse was intravenously injected with 150 μL of MGNRs or GNRs suspended in PBS (Au content of 2.5 mg/mL). A 50 μL portion of blood was collected by cheek pouch puncture at varying time points (1 min, 5 min, 15 min, 30 min, 1 h, 2 h, 4 h, 8 h, 24 h, and 48 h) after injection and subjected to Au content determination with ICP-MS as described above. The GNP concentration in the blood was expressed as the percentage of injected dose per milliliter (% ID/mL). The pharmacokinetics parameters including area under the concentration− time curve (AUC0−t), mean residence time (MRT0−t), elimination rate constant (k), clearance (Cl), and the half-life (t1/2) were calculated by DAS 3.0 software. For biodistribution studies, mouse models with or without cyclopamine treatments were intravenously injected with 150 μL of MGNRs suspended in PBS (Au content of 2.5 mg/mL). At 24 h post injection, the mice were sacrificed followed by heart perfusion with saline. Tumors and major organs including hearts, livers, spleens,

lungs, and kidneys were harvested and digested with 4 mL of aqua regia for Au content determination with ICP-MS. The GNP concentration in tissues was expressed as the percentage of injected dose per gram of tissue (% ID/g). In Vivo Photoacoustic (PA) Imaging of Tumor-Bearing Mice. Due to the specific photoacoustic character of MGNRs, the concentration of MGNRs at the tumor site can be indirectly determined by PA imaging using a Vevo LAZR system equipped with a linear array transducer (LZ400, 30 Hz center frequency). To determine the specific absorption wavelength of MGNR, the spectral scanning of MGNR was performed after subcutaneous injection of 150 μL of MGNR dispersion on the back of mice. To reduce the signal disturbance caused by hemoglobin in circulation, PA signals at the wavelength of 685, 705, 750, 780, 808, 850, and 865 nm were all collected for the spectral unmixing process. For PA imaging, Capan-2 tumor-bearing mice with or without cyclopamine treatments were intravenously injected with 150 μL of MGNR dispersion in PBS (Au content of 125 μg). After being anesthetized using isoflurane at a concentration of 1.5−2%, the mice were placed on the working plate with continuous isoflurane inhalation anesthesia for further PA imaging. PA signals of MGNR in tumors were obtained at different time points (2, 8, 12, 24, and 48 h) post injection, and the data were analyzed by Vevo LAB software. In Vivo PTT Efficacy. Capan-2 tumor-bearing nude mouse models were randomized into 4 treatment groups (n = 5): control group, cyclopamine group, laser only group, and cyclopamine + laser group. The cyclopamine group and cyclopamine + laser group were orally treated with cyclopamine solution while the control group and laser 31500

DOI: 10.1021/acsami.7b09458 ACS Appl. Mater. Interfaces 2017, 9, 31497−31508

Research Article

ACS Applied Materials & Interfaces

Figure 3. Photothermal effect of MGNRs in vitro. (A) Temperature changes of MGNR (red) and GNR (black) dispersions in water with Au content of 42 μg/mL upon irradiation with an 808 nm laser at different power densities for 5 min and then cooled naturally to room temperature for 5 min. (B) Temperature changes of MGNRs (red) and GNRs (black) dispersions in water at different Au concentrations (21, 42, 84 μg/mL) upon irradiation with an 808 nm laser at 0.75 W/cm2 for 5 min and then cooled naturally to room temperature for 5 min. (C) In vitro cell viability assays of Capan-2 cells after incubation with different concentrations of MGNRs or GNRs for 3 h followed by irradiation of an 808 nm laser at 0.75 W/cm2 for 5 min. (D) Fluorescence images of dead or live Capan-2 cells after incubation with MGNRs and GNRs of different Au concentrations (21, 42, 84 μg/mL) for 3 h followed by irradiation of an 808 nm laser at 0.75 W/cm2 for 5 min. The live cells are stained with Calcein-AM (green), and the dead cells were stained with PI (red). Original magnification: 100 ×. only group received blank vehicles as described above. After cyclopamine or vehicle treatment for 3 weeks, the laser only group and cyclopamine + laser group were intravenously injected with 150 μL of MGNRs dispersion (Au content of 125 μg). At 2 h post injection, tumor xenografts were radiated at 2 W/cm2 with an 808 nm laser (Changchun New Industries Optoelectronics Technology, Changchun, China) for 5 min. The temperature was recorded using an infrared camera thermographic system (InfraTec, vario CAM hr research, German) during the irradiation. After the irradiation ended, mouse models were returned to the animal house and maintained

under standard housing conditions. Tumor size and body weight of each mouse were measured every 2 days for 3 weeks. When tumor volume reached above 1500 mm3, all mouse models were sacrificed, and the tumors were collected for photography and wet weight recording. Tumor growth inhibition rates based on tumor size (TGIRv) and tumor weight (TGIRw) were calculated by the following W −W V −V formulas, respectively: TGIR v = c V t ; TGIR w = cW t . In these c

c

formulas, Vc and Wc indicated tumor volume and tumor weight in the control group, respectively; Vt and Wt indicated tumor volume and 31501

DOI: 10.1021/acsami.7b09458 ACS Appl. Mater. Interfaces 2017, 9, 31497−31508

Research Article

ACS Applied Materials & Interfaces

significantly by +13.7, +20.6, and +31.3 °C in the MGNR group and +15.1, + 22.4, and +31.9 °C in the GNR group, respectively (Figure 3B). These results indicated that the GNR concentration mainly determined the increase of the temperature with a constant laser power density during PTT and the RBC membrane coating did not significantly affect the photothermal conversion efficiency of GNRs. To explore the photothermal effects of MGNRs on ablating cancer cells in vitro, Capan-2 cells treated with MGNRs/GNRs demonstrated significantly decreased viability (P < 0.01) after irradiation with an 808 nm laser at 0.75 W/cm2 (Figure 3C). Moreover, when the Au concentration of MGNRs or GNRs increased from 21 to 84 μg/mL, the cell viability decreased from 81.11% to 34.94% in the MGNR group and from 84.45% to 49.23% in the GNR group. Compared with GNRs, MGNRs caused generally significantly lower cell viabilities (P < 0.05) at the Au concentration of 42 and 84 μg/mL. This higher PTT efficiency of MGNRs might be due to the better stability of MGNRs in cell culture medium. However, the MGNRs without laser irradiation did not show any significant cytotoxicity at Au content of 84 μg/mL. The in vitro photothermal ablation of cancer cells caused by MGNRs was further confirmed by the fluorescence confocal microscopy (Figure 3D). There were 96.71% and nearly 100% of the cells treated with MGNRs and GNRs at Au concentration of 84 μg/mL stained red while those in the control group were green, indicating that cancer cells were hardly alive after irradiation; these results were consistent with the cell viability results. To further investigate whether cyclopamine increased the sensitivity of Capan-2 cells to PTT or not, an in vitro PTT experiment was carried out. Results showed there were no significant differences in cell viability after PTT in the presence of various concentrations of cyclopamine, indicating that cyclopamine did not affect the sensitivity of Capan-2 cells to PTT (Figure S1). Effect of Cyclopamine on Tumor Microenvironment. To mimic the extensive fibrosis of human PDA, the Capan-2 cell line was chosen for establishing PDA xenograft-bearing mouse models. H&E staining and immunofluorescence staining of tumor slices from Capan-2 xenografts revealed extensive fibrosis with abundant ECM (Figure S2 and Figure 4A), indicating that the Capan-2 xenograft model could represent the real benefit of the Hh pathway blockade strategy. To explore the tumor microenvironment changes induced by cyclopamine, ECM morphology and blood perfusion in tumors were investigated after cyclopamine treatment. In this study, fibronectin was used as the ECM marker due to its high expression level in the Capan-2 tumor xenograft.8 As shown in Figure 4A, in the control group, fibronectin bundles were compact and divided the tumor parenchyma into separate compartments as previously reported.27,28 However, after cyclopamine treatment, fibronectin bundles were effectively disrupted and became fractured, loose, and scattered compared with those of the control group. Additionally, fibronectin expression in tumor tissues significantly decreased to 34.6% in the cyclopamine treatment group as compared with that in the control group (100%). The effect of ECM disruption on tumor perfusion was investigated by the DyLight 488-lectin labeling experiment and ultrasound imaging. As shown in Figure 4B, it was demonstrated that the percentage of DyLight 488+ tumor vessels was increased from 25.0% in the control group to 61.3% in the cyclopamine group, indicating that cyclopamine treatment significantly increased the percentage of functional

tumor weight in the treatment group, respectively. Tumors were imbedded in paraffin and sectioned into slices of 5 μm thickness for hematoxylin-eosin (H&E) staining as previously described.26 For in vivo toxicity assay, the body weight of mouse models was monitored every 3 days during the whole experiment. Major organs including hearts, livers, spleens, lungs, and kidneys were obtained from mouse models at the study end point and sectioned for H&E staining for histological analysis. The H&E staining tumor slices and normal organ slices were observed under the fluorescence microscope (LEICA DMI4000B, Germany). Statistical Analysis. Statistical differences were analyzed with an unpaired Student’s t-test for two groups’ comparison and one-way analysis of variance (ANOVA) for multiple-group comparison. Data are presented as mean ± standard deviation, and P values