lable at ScienceDirect

Biomaterials 69 (2015) 212e221

Contents lists avai

Biomaterials

journal homepage: www.elsevier .com/locate/biomateria ls

AuNP flares-capped mesoporous silica nanoplatform for MTH1detection and inhibition

Wen Gao, Wenhua Cao, Yuhui Sun, Xueping Wei, Kehua Xu, Huaibin Zhang, Bo Tang*

College of Chemistry, Chemical Engineering and Materials Science, Collaborative Innovation Center of Functionalized Probes for Chemical Imaging inUniversities of Shandong, Key Laboratory of Molecular and Nano Probes, Ministry of Education, Shandong Provincial Key Laboratory of Clean Production ofFine Chemicals, Shandong Normal University, Jinan 250014, PR China

a r t i c l e i n f o

Article history:Received 24 April 2015Received in revised form7 August 2015Accepted 8 August 2015Available online 10 August 2015

Keywords:AuNP flaresMesoporous silica nanoparticleMTH1DetectionInhibitionCancer theranostic

* Corresponding author.E-mail address: tangb@sdnu.edu.cn (B. Tang).

http://dx.doi.org/10.1016/j.biomaterials.2015.08.0210142-9612/© 2015 Elsevier Ltd. All rights reserved.

a b s t r a c t

The human mutT homologue MTH1, a nucleotide pool sanitizing enzyme, represents a vulnerabilityfactor and an attractive target for anticancer therapy. However, there is currently a lack of selective andeffective platforms for the detection and inhibition of MTH1 in cells. Here, we demonstrate for the firsttime a gold nanoparticle (AuNP) flares-capped mesoporous silica nanoparticle (MSN) nanoplatform thatis capable of detecting MTH1 mRNA and simultaneously suppressing MTH1 activity. The AuNP flares aremade from AuNPs that are functionalized with a dense shell of MTH1 recognition sequences hybridizedto short cyanine (Cy5)-labeled reporter sequences and employed to seal the pores of MSN to prevent thepremature MTH1 inhibitors (S-crizotinib) release. Just like the pyrotechnic flares that produce brilliantlight when activated, the resulting AuNP flares@MSN (S-crizotinib) undergo a significant burst of redfluorescence enhancement upon MTH1 mRNA binding. This hybridization event subsequently inducesthe opening of the pores and the release of S-crizotinib in an mRNA-dependent manner, leading tosignificant cytotoxicity in cancer cells and improved therapeutic response in mouse xenograft models.We anticipate that this nanoplatform may be an important step toward the development of MTH1-targeting theranostics and also be a useful tool for cancer phenotypic lethal anticancer therapy.

© 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Redox dysregulation in cancer cells results in reactive oxygenspecies (ROS) overproduction, damaging DNA and free bases indeoxyribonucleotide triphosphates (dNTPs) pools [1e3]. The MutTHomolog1 protein, MTH1, can effectively degrade oxidized nucle-otides to prevent their incorporation into DNA, and thereby isimportant for minimizing cancer-associated damage in the dNTPpool as required for cancer cell survival. Experimental evidence hasshown that inhibition of MTH1 activity selectively and effectivelykills cancer cell lines, but is considerably less toxic to normal cells,which validates MTH1 as an “Achilles heel” for cancer cells and anew target for cancer treatment [4,5]. However, the localizationand relative abundance of MTH1 mRNA in different cell types, aswell as the regulation of intracellular release of MTH1 inhibitors,are poorly characterized. Northern blots [6] and reverse transcrip-tion polymerase chain reaction (RT-PCR) [4,5] are commonly used

to detect MTH1mRNA transcripts. However, these methods requiremillions of cells and are not applicable to real-time studies of livesingle cells. Although in situ staining [6,7], green fluorescent pro-tein (GFP) [8], and fluorescence resonance energy transfer (FRET)sensors [8] have been used to detect MTH1 mRNA in living cells,these probes are often difficult to transfect, require additionalagents for cellular internalization, and are unstable in cellular en-vironments. This leads to a high background signal and an inabilityto detect targets. Small molecules such as TH287, TH588, SCH51344and S-crizotinib inhibit the MTH1 catalytic activity [4,5], but thetherapeutic effects may be limited by their insolubility and insta-bility in cellular environments as well as their poor penetration intotumor sites. Moreover, these inhibitors lack selectivity anddistribute indiscriminately into all cells, requiring a high dose tosignificantly affect tumor volume in mouse xenografts. The un-controlled release mechanism makes it difficult for MTH1 in-hibitors to reach their full therapeutic efficacy. Therefore, thedetection and visualization of MTH1 mRNA in living cells andcontrolled release of the inhibitors according to MTH1mRNA levelsremains a challenge.

Nanomaterials have showed their powerful ability to construct

W. Gao et al. / Biomaterials 69 (2015) 212e221 213

multifunctional platforms for intracellular research [9,10]. Goldnanoparticles (AuNPs) conjugated to short, fluorophore-labeledoligonucleotide duplexes can be intracellular “nanoflares” and areof significant interest because of their strong fluorescence signals,low background and probe stability [11,12], which enables thevisualization and quantification of MTH1 mRNA in living cells,without adding external transfection agents. Another motivationfor employing AuNP flares as an MTH1 probe is that they have thepotential of becoming a switch to control the release of MTH1 in-hibitors in response to intracellular MTH1 mRNA levels. Surface-functionalized, end-capped mesoporous silica nanoparticles(MSNs) are promising scaffolds to construct mRNA-responsivecontrolled release systems, because of their unique mesoporousstructures, large surface areas, tunable pore sizes, and goodbiocompatibility, both in vitro and in vivo [13e15]. These end-capped MSNs with “controlled release” properties have been syn-thesized with different pore-blocking caps such as organic mole-cules [16,17], supramolecular assemblies [18,19], nucleotides[14,20] and nanoparticles (NPs) [21,22]. AuNPs in particular havesuccessfully controlled the opening and closing of the poreentrance of MSN through either a reversible pH-dependent boro-nate ester bond [23], an acetal linker [24] or photo-controlledelectrostatic interactions [25]. We hypothesized that the AuNPflares could be employed as caps to encapsulate MTH1 inhibitorswithin the porous channels of the MSNs, which will be selectivelyopened in the presence of MTH1 mRNA targets.

Herein, we present a novel and effective strategy for combinedMTH1 detection and inhibition based on AuNP flares-capped MSNnanoplatform. As shown in Fig. 1, the AuNP flares consisted of AuNPfunctionalized with a dense shell of 20-base MTH1 recognitionsequences hybridized to short cyanine (Cy5) dye-labeled reportersequences via a goldethiol bond. After entrapping MTH1 inhibitors(S-crizotinib) in the mesopores of MSN, the AuNP flares wereimmobilized on the exterior surfaces of the MSN through thelinkage of reporter sequences. In the bound state, the Cy5 fluores-cence was quenched, and the AuNP blocked the pores. In thepresence of MTH1 mRNA, the recognition sequences hybridizedwith this complementary target sequences by forming the longerand more stable duplexes, causing the liberation of AuNP from thereporter sequences and the opening of the pores, which can thenproduce fluorescent signals correlated with the relative amount oftheMTH1mRNA and release of S-crizotinib in anmRNA-dependent

Fig. 1. Schematic illustration of AuNP flares-capped MSN

manner. These AuNP flares-capped MSNs were successfully used todetect and quantify MTH1 mRNA and inhibit MTH1 activity bothin vitro and in vivo. To the best of our knowledge, this is the firsttime that AuNP flares have been employed as pore-blocking caps toconstruct mRNA-responsive controlled release system based onMSNs, which allows a single nanoplatform capable of bothdetecting and regulating MTH1. By targeting MTH1, a novel cancerphenotypic lethal, AuNP flares-capped MSNs have the potential tobe a useful tool for cancer diagnosis and therapy.

2. Materials and methods

2.1. Materials

DNA oligonucleotides were synthesized and purified by SangonBiotechnology Co., Ltd (Shanghai, China). The sequences of theseoligonucleotides are shown in Table S1. Cetyltrimethylammoniumbromide (CTAB), tetraethyl orthosilicate (TEOS), (3-aminopropyl)triethoxysilane (APTES), 1-ethyl-3-(3-dimethylaminopropyl) car-bodiimide hydrochloride (EDC), 3-(4,5-dimethyl-thiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), S-crizotinib and RhodamineB (RhB) were purchased from SigmaeAldrich. Hydrogen tetra-chloroaurate (III) (HAuCl4$4H2O, 99.99%), Trisodium citrate(C6H5Na3O7$2H2O), Sodium dodecylsulfate (SDS), mercaptoethanol(ME), MgCl2, CaCl2, KCl were obtained from Sinopharm ChemicalReagent. Co. Ltd. (Shanghai, China). Deoxyribonuclease I (DNase I)was purchased from Solarbio Science and Technology Co., Ltd.(Beijing, China). Cell culture products, unless mentioned otherwise,were purchased fromGIBCO. DAPI were bought from Beyotime Inst.Biotech, (Haimen, China). All chemicals and solvents used were ofanalytical grade. Water was purified with a Sartorius Arium 611 VFsystem (Sartorius AG, Germany) to a resistivity of 18.2 MU cm.

2.2. Characterization

High resolution transmission electron microscopy (HRTEM) wascarried out on a JEM-2100 electron microscope. N2 adsorp-tionedesorption isotherms were recorded on a MicromeriticsASAP2020 surface area and porosity analyzer. The samples weredegassed at 150 �C for 5 h. The specific surface areas werecalculated from the adsorption data in the low pressure rangeusing the BET model and pore size was determined using the

nanoplatform for detection and inhibition of MTH1.

W. Gao et al. / Biomaterials 69 (2015) 212e221214

BarretteJoynereHalenda (BJH) method. X-ray diffraction (XRD)analysis was carried out on a D/Max 2500 V/PC X-ray diffractometerusing Cu (40 kV, 30 mA) radiation. Absorption spectra weremeasured on a pharmaspec UV-1700 UVevisible spectrophotom-eter (Shimadzu, Japan). Dynamic light scattering (DLS) measure-ments were performed on a Malvern zeta sizer Nano-ZS90. FTIRspectrawere collected on a Nicolet Impact 410 FTIR spectrometer inthe range of 400e4000 cm�1. X-ray photoelectron spectroscopy(XPS) spectrawere performedwith a Phobios 100 electron analyzer(SPECS GmbH) equipped 5 channeltrons, using an unmonochro-mated Mg Ka X-ray source (1253.6 eV). The deconvolution of theXPS peaks was done by an XPS Peak fitting program (version 4.1).Fluorescence spectra were obtained with FLS-920 EdinburghFluorescence Spectrometer with a Xenon lamp and 1.0 cm quartzcells at the slits of 3.0/3.0 nm. All pH measurements were per-formed with a pH-3c digital pH-meter (Shanghai LeiCi DeviceWorks, Shanghai, China) with a combined glasscalomel electrode.Absorbance was measured in a microplate reader (RT 6000, Rayto,USA) in the MTT assay. Confocal fluorescence imaging studies wereperformed with a TCS SP5 confocal laser scanning microscopy(Leica Co., Ltd. Germany) with an objective lens (�40). In vivo im-aging were performed with Caliper IVIS Lumina III (Caliper Co.,USA).

2.3. Preparation of the AuNP flares

First, the 5 nm AuNP were prepared by the modified citratereduction method [26]. 5.9 mg of HAuCl4 was dissolved in 1 mL ofultrapurewater and added into 47mL of ultrapurewater to the finalconcentration of 0.3 mM. The solutionwas incubated with constantstirring at 25 �C for 10min. 0.5mL of 50mM sodium citrate solutionwas added to the reaction mixture to a final concentration of0.5 mM, and the solution was incubated under constant stirring foranother 10 min. 0.67 mL of cooled 0.1 M NaBH4 was added into thestirred mixture; the solution developed an orange-red color. Theresulting nanoparticles were screened for their size and uniformityby HRTEM. The prepared AuNP were stored at 4 �C.

Second, the thiolated recognition strand were mixed with re-porter strand (1:1.2) in phosphate buffered saline (PBS: 137 mMNaCl, 10 mM Phosphate, 2.7 mM KCl, pH 7.4), heated to 75 �C andmaintained 10 min, then slowly cooled to room temperature, andstored in the dark for at least 12 h to allow complete hybridizationand flare duplex formation. Then, the flare duplexes added to asolution of AuNP (3 nM) with a final concentration of 150 nM andshaken overnight. After 12 h, SDS solution (10%) was added to themixture to achieve 0.1% SDS concentration. Phosphate buffer(0.1 M, pH 7.4) was added to the mixture to achieve 0.01 M phos-phate concentration and the NaCl concentration of the mixture wasslowly increased to 0.7 M over an eight-hour period. The resultingAuNP flares were purified from unreacted materials by three suc-cessive rounds of centrifugation (14,000 rpm, 20 min, 4 �C), su-pernatant removal, and resuspension in PBS with a concentrationof 15 nM as stock solution at 4 �C. The AuNP flares was diluted tocertain concentration for use in all subsequent experiments. Theconcentration of AuNP flares was determined by measuring theirextinction at 528 nm (ε ¼ 2.7 � 108 L mol�1 cm�1).

2.4. Quantitation of DNA duplexes loaded on the AuNP flares

The flare duplexes loaded on AuNP were quantitated accordingto the previous protocol [27]. Briefly, mercaptoethanol (ME) wasadded (final concentration 20 mM) to the AuNP flares solution(3 nM). After it was incubated overnight with shaking at roomtemperature, the flares were released. Then the released flareswereseparated via centrifugation and the fluorescence was measured

with a fluorescence spectrometer. The fluorescence of Cy5 labeledflare was excited at 648 nm and measured at 688 nm. The fluo-rescence was converted to molar concentrations of flares by inter-polation from a standard linear calibration curve that was preparedwith known concentrations of flares with identical buffer pH, ionicstrength and ME concentrations. By dividing molar concentrationsof the flares by the original AuNP flares concentration, we calcu-lated that there was 18 ± 1 flares per AuNP.

2.5. Preparation of AuNP flares@MSN nanoplatform

First, amino modified mesoporous silica was synthesized ac-cording to the typical co-condensation method reported previouslywith some modifications [28]. Cetyltrimethylammonium bromide(CTAB, 0.25 g, 0.69 � 10�3 mol) was first dissolved in 120 mLsartorius ultrapure water. NaOH (2 M, 0.75 mL) was then added tothe above solution and the mixture was stirred for 5 min, followedby adjusting the solution temperature to 80 �C. After that, TEOS(1.25 mL, 5.6 � 10�3 mol) was first introduced dropwise to thesurfactant solution, followed by the dropwise addition of APTES(0.25 mL, 1.07 � 10�3 mol). The mixture was allowed to stir for 2 hto give rise to white precipitates (as synthesized NH2-MSN). Thesolid product was filtered, washed with ultrapure water andmethanol, and then dried in air. To remove the surfactant template(CTAB), 0.5 g as-synthesized aminomodifiedmesoporous silicawasrefluxed for 24 h in a solution of 3 mL of HCl (37.4%) and 50 mL ofmethanol followed by extensive washes with ultrapure water andmethanol. Finally, the precipitates were dried for 24 h in vacuum.

Then, AuNP flares@MSN was obtained by coupling the carboxylgroups of the reporter strands and the amino groups on the surfaceof NH2-MSN to form the CONH amide bonds. 3 mL EDC solution(2.8 mM) was added to 500 mL of AuNP flares (3 nM) solution andthe solutionwasmixed and reacted for 30min at room temperatureto activate carboxyl groups. Then the mixture was added to 0.5 mLof a previously prepared NH2-MSN solution (2 mg/mL; MES buffer:10 mM, pH 6.0) with gentle stirring in darkness. The solution wasreacted for 1 h to result in the formation of the amide bond. Afterthat, the precipitates were centrifuged (14,000 rpm, 25 min, 4 �C)and washed with PBS buffer (10 mM, pH 7.4) for three times.

2.6. Binding experiment

50 mg of AuNP flares@MSN was redispersed in 1 mL of hybrid-ization buffer (10 mM PBS, pH 7.4, 100 mM NaCl, and 1 mM MgCl2)and incubated with different concentrations of complementaryDNA targets (0, 2, 5, 10, 25, 50, 100 and 200 nM). After incubationfor 1 h at 37 �C, the fluorescence of Cy5 was excited at 648 nm andmeasured at 688 nm. All experiments were repeated at least threetimes.

2.7. Specificity experiment

The complementary DNA target for AuNP flares@MSN and othertargets (MTH1 single-base mismatched target, survivin target, K-ras target, Galnac-T target, c-myc target, HER-2/neu target) werespiked in 1 mL hybridization buffer containing 50 mg/mL nano-platform, while the DNA targets were 200 nM. All experimentswere repeated at least three times.

2.8. Loading experiment

For RhB loading, 1 mg NH2-MSN was added and dispersed in1 mL RhB solution (0.5 mg/mL). The mixture was stirred for 24 h indarkness to reach the maximum loading. The other procedures forcapping the RhB loaded NH2-MSN were the same as the methods

W. Gao et al. / Biomaterials 69 (2015) 212e221 215

mentioned above. The prepared AuNP flares@MSN (RhB) wascentrifuged, washed several times with PBS buffer (at least 6) to getrid of any RhB physisorbed, and dried under high vacuum. Thesupernatant and all the washings solutions were collected and theamount of RhB washed was measured using fluorescent spectros-copy (lex ¼ 532 nm, lem ¼ 575 nm). The loading of RhB 0.036 mg/mg AuNP flares@MSN) was calculated with the difference betweenthe original and the washed amount of RhB. MTH1 inhibitor (S-crizotinib, 0.5 mg/mL, DMSO solution) was loaded and capped byfollowing the same procedure. The loading content of S-crizotinibwas monitored by UV/Vis spectroscopy at 270 nm calculated to be0.051 mg/mg AuNP flares@MSN.

2.9. Capping efficiency of AuNP flares@MSN (RhB)

To evaluate the capping efficiency, 1 mL as-prepared AuNPflares@MSN (RhB) (50 mg/mL) or 1 mL MSN (RhB) (50 mg/mL) wasplaced at room temperature in darkness and the fluorescence in-tensity of the sample was measure at 0, 6, 12, 24, 36, 48, 60, and72 h, respectively. The percentage of dye leaking from the nano-platform was calculated as follows: (fluorescence intensity of eachsample)/(fluorescence intensity of the loaded RhB).

2.10. Controlled release of AuNP flares@MSN (RhB)

50 mg of AuNP flares@MSN (RhB) was incubated with 200 nMperfectly matched target in 1.0 mL of hybridization buffer (10 mMPBS, pH 7.4, 100 mM NaCl, and 1 mM MgCl2) for 24 h at 37 �C.Single-base mismatched DNA target (200 nM) was added in thecontrol reaction. The release of RhB from AuNP flares@MSN wasalso studied as a function of the amount of the DNA targets (0, 2, 5,10, 25, 50, 100, 200 nM; 2 h at 37 �C). RhB release was monitored byfluorescent spectroscopy at appropriate time intervals and calcu-lated as the same formula above.

2.11. Nuclease assay

Two groups of 50 mg/mL AuNP flares@MSN (RhB) in PBS buffer(10 mM, pH 7.4, 2.5 mMMgCl2, and 0.5 mM CaCl2) were placed in a96-well fluorescence microplate at 37 �C. After allowing the sam-ples to equilibrate (10 min), 1.3 mL of DNase I in assay buffer (2 U/L)was added to one group. The fluorescence of these samples wasmonitored for 12 h and was collected at 2 h intervals during thisperiod. Then 100 nM DNA targets were paralleled added into thetwo samples with incubation for 1 h at 37 �C, the fluorescence ofRhB and Cy5 was measured at appropriate excitation wavelength,respectively.

2.12. Cell culture

Human cervical carcinoma cell line HeLa, Human hepatocellularliver carcinoma cell line HepG2 and human hepatocyte cell line HL-7702 were obtained from the Committee on Type Culture Collec-tion of Chinese Academy of Sciences. Cells were grown in cell cul-ture media and incubated at 37 �C in a 5% CO2/95% air humidifiedincubator (MCO-15AC, SANYO). The cell culture mediumwas RPMI-1640 (2000 mg/L D-Glucose, 300 mg/L L-Glutamine, Hyclone, USA)supplemented with 10% heat-inactivated fetal bovine serum (FBS),100 U/mL penicillin and 100 mg/mL streptomycin (Invitrogen,Carlsbad, CA).

2.13. TEM imaging of cells

HeLa, HepG2 and HL7702 cells were treated with 50 mg/mLAuNP flares@MSN (S-crizotinib) for 1 h and 3 h, respectively. Then,

cells were trypsizined, centrifuged, and fixed in 2.5% glutaralde-hyde in 0.1 M sodium cacodylate buffer (pH 7.4) for 1 h in roomtemperature and rinsed. Cells were then post fixed 1 h in 2%osmium tetroxide with 3% potassium ferrocyanide and rinsed, nextenbloc staining with a 2% aqueous uranyl acetate solution anddehydration through a graded series of alcohol, two changes ofpropylene oxide, a series of propylene oxide/Epon dilutions, andembedded in 100% Epon. The thin (70 nm) sections were cut on aLeica UC6 ultramicrotome, and images were taken on a JEOL 1200EX (JEOL, Ltd. Tokyo, Japan) using an AMT 2k digital Camera.

2.14. Confocal fluorescence imaging

HeLa, HepG2 and HL7702 cells in the logarithmic growth phasewere seeded at an initial density of 5 � 104 cells/dish in 20-mmglass bottom dishes (SPL Life Sciences, Seoul, Korea) and incu-bated for 24 h before adding the test substance. Then these cellswere respectively treated with AuNP flares@MSN (S-crizotinib)(50 mg/mL) for 3 h and 12 h. After the nanoplatform incubation, themedium was removed and washed three times with PBS. The cellnuclei were then stained with DAPI (10 mg/mL) for 30 min indarkness at 37 �C. Cells were finally washed three times with PBSand examined by confocal laser scanning microscopy (CLSM) with405 nm and 633 nm excitation.

2.15. RT-PCR

Total RNA from sorted cells was extracted with the RNeasy MiniKit (Qiagen, Valencia, CA). cDNA synthesis was performed using aniScript kit (Bio-Rad). RT-PCR was carried out with SYBR Green I(Qiagen) on an ABI PRISM 7000 sequence detection system. Rela-tive level of tumormRNAwas calculated from the quantity of tumormRNA PCR products and the quantity of GAPDH PCR products. Theprimers used in this experiment were shown in Table S1.

2.16. MTT assay

Cytotoxicity was measured by using the MTT assay in the log-arithmic phase of cell growth. HeLa, HepG2 and HL7702 cells wereseeded at a density of 5 � 104 cells/well in a 96 well-plate andincubated for 24 h before adding the test substance. Then freshmedium containing increasing concentrations of AuNP flar-es@MSN, AuNP flares@MSN (S-crizotinib) and free S-crizotinib wasadded to eachwell, respectively. After 12 h incubation, mediumwasremoved and replaced with medium containing MTT (0.5 mg/mL).Cells were incubated at 37 �C for another 4 h after which mediumwas removed. DMSO (100 mL) was added to lyse the cells anddissolve the formazan produced. The absorbance at 570 nm of eachwell was monitored using an RT 6000 microplate reader. Viabilitywas calculated based on the recorded data.

In the co-incubation experiment, HepG2 cells (5000 cells/well)and equivalent HL7702 cells were cultured in 96 a well-plate andincubated for 24 h before treatment. Then freshmedium containing50 mg/mL of AuNP flares@MSN (S-crizotinib) and free S-crizotinibwas added to eachwell, respectively. After 12 h incubation, the cellswere treated as mentioned above.

2.17. In vivo imaging of MTH1

All animal experiments were carried out according to the Prin-ciples of Laboratory Animal Care (People's Republic of China) andthe Guidelines of the Animal Investigation Committee, BiologyInstitute of Shandong Academy of Science, China. Male nude mice(6e8 week old, ~20 g) were housed under normal conditions with12 h light and dark cycles and given access to food and water ad

W. Gao et al. / Biomaterials 69 (2015) 212e221216

libitum.Right flank of the mice were injected a suspension of 1 � 106

HepG2 cells in PBS (150 mL). When tumor grew up to an averagevolume of 300e500 mm3, 100 mL of AuNP flares@MSN (S-crizoti-nib) was intravenously injected. The NIR fluorescent signals wereobserved using Caliper IVIS Lumina III at 0, 3, 6, 12 and 24 h timepoints. At 24 h postinjection, tumors and normal organs includingheart, liver, spleen, lung, and kidney from HepG2 tumor-bearingnude mice were collected and visualized with Caliper IVISLumina III. Tumor-to-muscle ratios (T/M) were determined by im-age processing on Living Image Software. Briefly, fluorescence im-ages were background subtracted using a rolling-ball radius of 150.Region-of-interests (ROIs) were defined by drawing a contouraround the tumor (Tumor ROI) and around the entire animal exceptthe tumor (Muscle ROI). T/M was determined by dividing theaverage pixel intensity of the Tumor ROI by the average pixel in-tensity in the Muscle ROI. Values are presented as means ± S.D.calculated for groups of five animals.

2.18. In vivo antitumor efficacy

HepG2 cells (1� 106 cells per 150 mL of PBS) were injected in theright flank of nude mice. When the tumor volume reached to about80e100 mm3 (two weeks), 100 mL PBS, free S-crizotinib (25 mg/kg),AuNP flares@MSN (25 mg/kg), and AuNP flares@MSN (S-crizotinib)(25 mg/kg containing S-crizotinib 1.25 mg) was administeredintratumorly once daily for 30 days. The mouse body weight andtumor size was recorded. The tumor volume (V) was determined bymeasuring length (L) and width (W), and calculated asV¼ L�W�W� 0.52. The relative tumor volumes were calculatedfor each mouse as V/V0 (V0 was the tumor volume when thetreatment was initiated). At termination, tumors and normal or-gans from PBS- and AuNP flares@MSN (S-crizotinib)-treated micewere collected and applied for analyzing the nanoplatform toxicity.

3. Results and discussion

3.1. Synthesis and characterization of AuNP flares-capped MSNnanoplatform

The AuNP flares were designed using 5 nm AuNP because thissize is an efficient quencher [29] and it can be suitable for cappingthe MSN pores. The thiol-modified single-stranded DNA (50-GAAATTCCACGGGTACTTCAAAAAA-(CH2)6-SH-30) was designed asthe recognition sequence to target intracellular MTH1 mRNA, andthe carboxyl-modified single-stranded DNA (50-COOH-(Cy5)TGAAGTACCCG-30) also labeled with Cy5 dye was designed as thereporter sequence. After hybridization, the resulting flare duplexswere immobilized on the surface of AuNP via the thiol groups ofrecognition sequence. High-resolution transmission electron mi-croscopy (HRTEM) images of AuNP and AuNP flares are shown inthe Supporting Information, Fig. S1. The results showed that theborder of the AuNP was clear and that of the AuNP flares wasambiguous. This is indicative of the assembly of flare duplexes onthe surface of the AuNP. The UV/Vis absorption spectra indicatedthat the maximum absorption of the AuNP was at 513 nm and thatit was red-shifted to 528 nm for the AuNP flares, which furtherconfirmed that the AuNP was successfully functionalized with flareduplexes (Fig. S2). Based on the previously establishedmethod [27],each AuNP was calculated to carry 18 ± 1 Cy5-labeled flares tar-geting MTH1. Details of the characterization are provided in Sup-porting Information and Fig. S3.

MSNs were then synthesized according to a previously reportedmethod [28]. The honeycomb-like mesoporous structure wasconfirmed by HRTEM, scanning electron microscopy (SEM) and N2

sorption analysis. As shown in Fig. 2a and Fig. S4a, the MSN particleis spherical with an average diameter of 80 nm. The N2 adsorp-tionedesorption isotherm of MSNs also showed a typical Type IVcurve with a total surface area of 405 m2/g and average porediameter of 2.3 nm with a narrow pore-size distribution (Fig. S5).The surface of the MSN was functionalized with APTES to result inNH2-MSN, which was then reacted with COOH groups on the re-porter sequences to cap the AuNP flares onto the pores (AuNPflares@MSN). The TEM and SEM images verified the successfulcapping with AuNP flares, as distinctive dark, crystalline spots withdiameters of 5e6 nm on the surfaces of the MSN (Fig. 2b, Fig. S4b).These AuNP flares (5e6 nm diameter) were large enough to blockthe 2.3 nm pores of the MSNs and thus inhibit the release of thecargo. Powder X-ray diffraction (XRD) in 30� < 2q < 70� range(Fig. S6) showed similar (111), (200), and (220) diffraction peaks tothose observed for AuNP flares, supporting the presence of AuNPflares on MSN. The successful fabrication of AuNP flares@MSN wasfurther confirmed by dynamic light scattering, Fourier transforminfrared and X-ray photoelectron spectroscopy (see the SupportingInformation Table S2, Figs. S7 and S8).

To evaluate the feasibility of AuNP flares@MSN for the detectionof MTH1 mRNA, we first examined the response of this nanoplat-form to synthetic DNA targets. The addition of the perfectlymatched DNA target resulted in a 7.7-fold increase in fluorescencesignal upon target recognition and binding. In contrast, the signaldid not obviously change in the presence of a single-base mis-matched target and was of comparable magnitude to backgroundfluorescence (Fig. 3a). These results indicated that the nanoplat-form was efficient at signaling the presence of specific targets.Fig. 3b showed that the fluorescence intensity of the nanoplatformincreases with increasing concentration of the DNA targets from0 to 200 nM, thus suggesting that the hybridization of the AuNPflares in the nanoplatform and DNA targets led to fluorescencerecovery and that the fluorescence intensity is associated withconcentration of the DNA targets. The selectivity of the AuNP flaresfor the nanoplatform was also performed and is shown in Fig. S9.The results revealed that the AuNP flares were specifically bound tothe DNA target and generated 6- to 7-fold higher fluorescent signalcompared with other targets.

To investigate the mRNA-induced controlled release property ofthe AuNP flares@MSN nanoplatform in an extracellular environ-ment, rhodamine B (RhB) was used as a model cargo molecule. TheRhB-loaded AuNP flares@MSN (RhB) sample was prepared, and theloading of RhB was determined to be 0.036 mg/mg of AuNP flar-es@MSN as described in Supporting Information (Fig. S10). Toexamine the capping efficiency, the AuNP flares@MSN (RhB) wasfirst placed for 72 h in darkness and the fluorescence intensity inaqueous solution was measured at appropriate time intervals. Asshown in Fig. S11, about 89% RhB was leaked after 72 h for thecontrol MSN without capping, while less than 8% RhB was leakedafter 72 h for AuNP flares@MSN, indicating that the capping strat-egy was successful with good efficiency.

Then release experiments were performed and the releaseprofile of RhB was displayed in Fig. 4a. As can be seen, withouttarget or with a single-base mismatch, the AuNP flares@MSN wastightly capped and showed a negligible release of RhB (curve i andcurve ii). In contrast, the presence of the complementary DNAtarget induced the hybridization between recognition sequenceson the AuNP flares@MSN and DNA target, the opening of the pores,and a remarkable release of the dye (curve iii). Curve iii also showedthat AuNP flares@MSN responded rapidly to the target and within2 h more than 52% of the RhB was released. This target-triggeredpore uncapping could also clearly be seen in the HRTEM (Fig. 4b).

The release of RhB from AuNP flares@MSN was also studied as afunction of the amount of the DNA targets (see inset in Fig. 4a).

Fig. 2. HRTEM micrographs of (a) MSN and (b) AuNP flares@MSN. The inset in (b) is a high-resolution TEM image.

Fig. 3. The nanoplatform responds to synthetic target DNA. (a) Fluorescence spectra of the 50 mg/mL nanoplatform alone (black), with 200 nM target (red), and with 200 nM single-base mismatched target (blue). (b) Fluorescence spectra of the nanoplatform (50 mg/mL) in the presence of various concentrations of DNA target (0, 2, 5, 10, 25, 50, 100 and 200 nM).Inset: calibration curves for the fluorescence intensity versus corresponding target concentrations. (For interpretation of the references to colour in this figure legend, the reader isreferred to the web version of this article.)

Fig. 4. (a) Release of RhB from AuNP flares@MSN in the absence of the target (i), and in the presence of single-mismatched target (ii) and perfectly matched target (iii) at con-centrations of 100 nM. Inset: Percentage of released RhB as a function of the concentration of DNA target after 2 h of reaction. Data are shown as mean ± S.D. of three independentexperiments. (b) HRTEM micrographs of the sample obtained after hybridization with DNA target showing the pores opening.

W. Gao et al. / Biomaterials 69 (2015) 212e221 217

Clearly, the release of the RhB was proportional to the target con-centration. The maximum release was observed at a target con-centration of 100 nM; at higher concentrations the release was

partially inhibited, most likely because excess DNA target wasadsorbed onto the surface of the MSN, inducing partial porecapping.

W. Gao et al. / Biomaterials 69 (2015) 212e221218

Because nuclease resistance is critical for the AuNP flares@MSNwhen used in living cells or in vivo, its nuclease stability wasevaluated with enzyme deoxyribonuclease I (DNase I) [30] underphysiological conditions via analysis of Cy5 or RhB fluorescence.Fig. S12 showed that both Cy5 and RhB fluorescence intensity of thenanoplatform treated with DNase I was not obviously changedcompared with the case without DNase I. However, the fluores-cence intensity of the two solutions was all markedly enhancedafter hybridization with the DNA targets, indicating that AuNPflares@MSN possessed high resistance to nucleases and the releaseof RhB was indeed due to the hybridizationwith the targets insteadof nuclease degradation. The above results confirmed the mRNA-responsive controlled-release mechanism and the feasibility ofAuNP flares@MSN specific detection and targeted inhibition ofMTH1.

3.2. In vitro imaging and inhibition of MTH1

For intracellular applications of AuNP flares@MSN nanoplat-form, a MTH1 inhibitor (S-crizotinib) was chosen as the cargomolecule for a controlled release study in human cancer cells(HepG2 and HeLa, MTH1-positive) and normal liver cells (HL7702,MTH1-negative). The AuNP flares@MSN (S-crizotinib) was endo-cytosed rapidly by these three cell types and is localized mainly inthe cytoplasm and subcellular vesicles as determined by TEM(Fig. 5a). After the positive cancer cells were incubated with AuNPflares@MSN (S-crizotinib) for 3 h, strong red fluorescence wasobserved under confocal laser scanning microscopy (CLSM). Addi-tionally, DAPI nuclear staining and bright-field imaging showed noDNA damage or apoptotic cells within the initial 3 h. Inhibition ofMTH1 by S-crizotinib could selectively induce an increase in DNAsingle-strand breaks leading to DNA damage and cytotoxicity inhuman cancer cells [5]. After 12 h, we observed both HepG2 andHeLa cells undergoing apoptosis with clear nuclear DNA conden-sation and cell shrinkage. Red fluorescent signals were alsoobserved in two cancer cells similar to the case at 3 h. When theHL7702 cells were incubated under the same condition, the fluo-rescent signals became faint and these cells were viable throughoutthe imaging experiments (Fig. 5b). Our experiments indicated thatthe relative expression levels of MTH1 mRNA in HepG2 and HeLa

Fig. 5. (a) Time-course changes in subcellular localization of AuNP flares@MSN (S-crizotinibthree cells after incubated with AuNP flares@MSN (S-crizotinib). Arrows indicate DNA cond

were higher than that in HL7702, which is consistent with real-time reverse transcription-PCR (RT-PCR) results (Fig. S13). Thebinding of MTH1 mRNA triggered the release of S-crizotinib, whichcould effectively induce DNA damage and cell death in cancer cells.Thus, the AuNP flares@MSN (S-crizotinib) nanoplatform can beused to distinguish cancer cells from normal cells as well asselectively release S-crizotinib on the basis of MTH1 mRNA levels.

To further evaluate the therapeutic efficacy of this nanoplat-form, an MTT assay was used for quantitative testing of the viabilityof the three model cells in the presence of AuNP flares@MSN, AuNPflares@MSN (S-crizotinib), and free S-crizotinib. As shown in Fig. 6,the AuNP flares@MSN had no obvious effect on HepG2, HeLa andHL7702 cell viability at concentrations up to 100 mg/mL, whereasAuNP flares@MSN (S-crizotinib) exhibited a statistically significantcytotoxic effect on the two cancer cells but not on the normal cells.Notably, the half-maximal inhibitory concentration (IC50) of AuNPflares@MSN (S-crizotinib) against the two cancer cells was similarand calculated to be ~36 mg/mL (the loaded S-crizotinib concen-trations are ~1.8 mg/mL). These values were 4.4% that of free S-cri-zotinib (IC50 of 41 mg/mL), suggesting that the nanoplatform couldreduce the dosage and greatly improve the therapeutic effect of S-crizotinib. This is mainly because the encapsulation of S-crizotinibwithin AuNP flares@MSN overcomes many severe problemsincluding solubility and poor cellular uptake when administered asfree S-crizotinib. Moreover, AuNP flares@MSN (S-crizotinib) couldprevent unexpected release in normal cells and therefore cause amarked dose escalation within cancer cells. This additional benefitwas validated by the co-incubation experiment of HepG2 andHL7702. As shown in Fig. S14, the mixed cells (HepG2: 0.1 mL,5 � 104/mL, HL7702: 0.1 mL, 5 � 104/mL) still maintained about74.5% of the cell viability after incubation with free S-crizotinib(50 mg/mL), which was higher thanwhen an equal amount of AuNPflares@MSN (S-crizotinib) was added (cell viability, 50.2%). AuNPflares@MSN (S-crizotinib) reduced the retention of S-crizotinib innormal cells and allowed the localized release of highly concen-trated S-crizotinib within cancer cells. These results, plus intracel-lular imaging experiment, further confirmed that thisnanoplatform for the detection and inhibition of MTH1 is effectiveat the in vitro level, therefore could be further extended to in vivoapplication.

) within three types of cells (see arrows). (b) The time course of confocal images of theensation and cell shrinkage of apoptotic cells. Scale bar ¼ 25 mm.

Fig. 6. In vitro viability of (a) HeLa, (b) HepG2, and (c) HL7702 cells incubated with AuNP flares@MSN, AuNP flares@MSN (S-crizotinib), and free S-crizotinib for 12 h.

W. Gao et al. / Biomaterials 69 (2015) 212e221 219

3.3. In vivo MTH1 imaging

We next studied the in vivo MTH1 imaging ability and tumorselectivity of AuNP flares@MSN (S-crizotinib) in a subcutaneousHepG2 tumor model. The AuNP flares@MSN (S-crizotinib) (25 mg/kg) was intravenously administered, and fluorescence images wereacquired at different time points post-injection. Excitation andemission wavelengths were set as 640/670 nm for Cy5. After in-jection, MTH1 mRNA was bound to AuNP flares in the tumor areaand the quenching effect between AuNP and Cy5 was lessened,resulting in fluorescence signal recovery over time. The tumor-to-muscle tissue (T/M) ratios were 1.55 ± 0.32, 2.42 ± 0.21,3.15 ± 0.44, and 2.11 ± 0.23 at 3, 6, 12, and 24 h post-injection,showing high levels of MTH1 in tumor regions (Fig. 7a and b).Decrease at the 24 h time point was probably due to the clearanceof AuNP flares@MSN in vivo. To confirm our hypothesis, ex vivoimaging was performed as shown in Fig. 7c. In addition to accu-mulation in the HepG2 tumors, we found a high level of fluores-cence activity in the livers and kidneys at 24 h, indicating the

Fig. 7. (a) NIR fluorescent imaging of MTH1 mRNA in HepG2 tumor-bearing mice intravenopoints, and fluorescent signals were normalized by the maximum average value. The colorindicate tumors location. (b) Tumor/muscle (T/M) ratio of HepG2 tumor-bearing mouse modmajor organs and tumors excised at 24 h post-injection of AuNP flares@MSN (S-crizotinib). (to the web version of this article.)

hepatic and renal clearances of AuNP flares@MSN. The uptake inother organs was much lower as determined by ex vivo imaging.These imaging results suggested that AuNP flares@MSN (S-crizo-tinib) could specifically accumulate at the tumor site via a passivemechanism and enabled clear visualization of MTH1 mRNA levelsin tumors. Moreover, the ability of AuNP flares@MSN to be elimi-nated from the body would decrease the biological persistence ofthis nanoplatform injected into the body and eliminate its risks oflong-term toxicity.

3.4. In vivo antitumor efficacy

To investigate the in vivo potential of AuNP flares@MSN (S-cri-zotinib) to suppress tumor growth, we established HepG2 tumorsin the flank of 20 mice. Two weeks after tumor engraftment themice were randomized into four treatment groups (n ¼ 5/group).Each group received intra-tumoral injections of either phosphatebuffered saline (PBS), free S-crizotinib (25 mg/kg) [5], AuNP flar-es@MSN (25 mg/kg) or AuNP flares@MSN (S-crizotinib) (25 mg/kg,

usly received AuNP flares@MSN (S-crizotinib). Images were acquired at indicated timebar indicates radiant efficiency (low, 0; high, 3.33 � 108). White circles were used toel. Data are shown as means ± S.D., n ¼ 5 per group. (c) Ex vivo fluorescence images ofFor interpretation of the references to colour in this figure legend, the reader is referred

Fig. 8. In vivo anticancer activity of AuNP flares@MSN (S-crizotinib). (a) HepG2 tumor growth rate in each group after treatment (control groups: 50 mL PBS; therapeutic groups:25 mg/kg; intratumorally daily). Tumor volumes were normalized to their initial size. (b) The effect on tumor growth following 30-day treatment. Images depict representativetumors for each treatment group. (c) Photographs of the major organs excised after 30-day treatment with AuNP flares@MSN (S-crizotinib). No noticeable abnormality was found inthe heart, liver, spleen, lung, or kidney. (d) Body weight curves of HepG2 tumor-bearing mice for each group. All data are shown as mean ± S.D.; n ¼ 5 per group.

W. Gao et al. / Biomaterials 69 (2015) 212e221220

loaded S-crizotinib was 1.25 mg/kg). To study the therapeutic ef-fects, we continued to monitor and record HepG2 tumor growthrates (Fig. 8a). Following sacrifice of the mice after 30 days, thetumors on mice received AuNP flares@MSN (S-crizotinib) treat-ment showed a greater decrease in tumor volume (78%) comparedto free S-crizotinib (42%). In marked contrast, other control groupswith PBS only or AuNP flares@MSNwithout loading S-crizotinib didnot show significant tumor suppression (Fig. 8b). We also examinedthe major organs from HepG2 tumor-bearing nude mice includingheart, liver, spleen, lung and kidney. We did not observe anynoticeable organ damage compared to the control group (Fig. 8c)suggesting that AuNP flares@MSN (S-crizotinib) did not have toxicside effects in HepG2 tumor-bearing nude mice. The body weightdata further confirmed this (Fig. 8d). Collectively, our results veri-fied that the ability to package S-crizotinib with controlled releaseof this inhibitor according to MTH1 mRNA levels allowed AuNPflares@MSN (S-crizotinib) to effectively suppress tumor growth inmice and greatly improve the therapeutic efficacy of S-crizotinib.The AuNP flares@MSN (S-crizotinib) can both detect and inhibitMTH1 and is an ideal platform for image-guided cancer therapy.

4. Conclusion

In summary, we have successfully fabricated a multifunctionalnanoplatform based on AuNP flares-capped MSNs that can detectMTH1 mRNA and simultaneously suppressed MTH1 catalytic ac-tivity both in vitro and in vivo. This nanoplatform signaled thepresence of MTH1 mRNA with the liberation of AuNP flares, whichresulted in the release of S-crizotinib from the pores of the MSNsinto the cytosol. These MTH1 inhibitors specifically induced DNA

damage and cell death in cancer cells. Furthermore, the highspecificity, good nuclease stability, enhanced cellular uptake andMTH1 mRNA-dependent release properties enabled the nanoplat-form to clearly visualize the MTH1 in tumors as well as efficientlyreduce tumor volume by more than 70%. Therefore, the AuNPflares@MSN (S-crizotinib) integrating MTH1 detection and regula-tion in a single nanoplatform is a promising first step toward thedevelopment of MTH1 directed theranostics and is expected to be anovel cancer phenotypic lethal anticancer strategy.

Acknowledgments

This work was supported by 973 Program (2013CB933800),National Natural Science Foundation of China (21227005, 21390411,21305081).

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.biomaterials.2015.08.021.

References

[1] J.S. Dawane, V.A. Pandit, Understanding redox homeostasis and its role incancer, J. Clin. Diagn. Res. 6 (2012) 1796e1802.

[2] S. Oka, M. Ohno, D. Tsuchimoto, K. Sakumi, M. Furuichi, Y. Nakabeppu, Twodistinct pathways of cell death triggered by oxidative damage to nuclear andmitochondrial DNAs, EMBO J. 27 (2008) 421e432.

[3] T.C. Jorgenson, W. Zhong, T.D. Oberley, Redox imbalance and biochemicalchanges in cancer, Cancer Res. 73 (2013) 6118e6123.

[4] H. Gad, T. Koolmeister, A.S. Jemth, S. Eshtad, S.A. Jacques, C.E. Str€om, et al.,MTH1 inhibition eradicates cancer by preventing sanitation of the dNTP pool,Nature 508 (2014) 215e221.

W. Gao et al. / Biomaterials 69 (2015) 212e221 221

[5] K.V.M. Huber, E. Salah, B. Radic, M. Gridling, J.M. Elkins, A. Stukalov, et al.,Stereospecific targeting of MTH1 by (S)-crizotinib as an anticancer strategy,Nature 508 (2014) 222e227.

[6] H. Oda, Y. Nakabeppu, M. Furuichi, M. Sekiguchi, Regulation of expression ofthe human MTH1 gene encoding 8-oxo-dGTPase. Alternative splicing oftranscription products, J. Biol. Chem. 272 (1997) 17843e17850.

[7] D. Yoshimura, K. Sakumi, M. Ohno, Y. Sakai, M. Furuichi, S. Iwai, et al., Anoxidized purine nucleoside triphosphatase, MTH1, suppresses cell deathcaused by oxidative stress, J. Biol. Chem. 278 (2003) 37965e37973.

[8] M.C. Schmidt, R.R. Mccartney, X. Zhang, T.S. Tillman, H. Solimeo, S.W. Lfl, et al.,Std1 and Mth1 proteins interact with the glucose sensors to control glucose-regulated gene expression in saccharomyces cerevisiae, Mol. Cell Biol. 19(1999) 4561e4571.

[9] W. Gao, L. Ji, L. Li, G. Cui, K. Xu, P. Li, et al., Bifunctional combined Au-Fe2O3nanoparticles for induction of cancer cell-specific apoptosis and real-timeimaging, Biomaterials 33 (2012) 3710e3718.

[10] H.-Y. Lu, Y.-J. Chang, N.-C. Fan, L.-S. Wang, N.-C. Lai, C.-M. Yang, et al., Syn-ergism through combination of chemotherapy and oxidative stress-inducedautophagy in A549 lung cancer cells using redox-responsive nanohybrids: anew strategy for cancer therapy, Biomaterials 42 (2015) 30e41.

[11] D.S. Seferos, D.A. Giljohann, H.D. Hill, A.E. Prigodich, C.A. Mirkin, Nano-flares:probes for transfection and mRNA detection in living cells, J. Am. Chem. Soc.129 (2007) 15477e15479.

[12] N. Li, C. Chang, W. Pan, B. Tang, A multicolor nanoprobe for detection andimaging of tumor-related mRNAs in living cells, Angew. Chem. Int. Ed. 51(2012) 7426e7430.

[13] H. Meng, W.X. Mai, H. Zhang, M. Xue, T. Xia, S. Lin, et al., Codelivery of anoptimal drug/siRNA combination using mesoporous silica nanoparticles toovercome drug resistance in breast cancer in vitro and in vivo, ACS Nano 7(2013) 994e1005.

[14] E. Climent, R. Martínez-M�a~nez, F. Sancen�on, M.D. Marcos, J. Soto, A. Maquieira,et al., Controlled delivery using oligonucleotide-capped mesoporous silicananoparticles, Angew. Chem. Int. Ed. 49 (2010) 7281e7283.

[15] A. Calvo, B. Yameen, F.J. Williams, G.J.A.A. Soler-Illia, O. Azzaroni, Mesoporousfilms and polymer brushes helping each other to modulate ionic transport innanoconfined environments. An interesting example of synergism in func-tional hybrid assemblies, J. Am. Chem. Soc. 131 (2009) 10866e10868.

[16] N.K. Mal, M. Fujiwara, Y. Tanaka, Photocontrolled reversible release of guestmolecules from coumarin-modified mesoporous silica, Nature 421 (2003)350e353.

[17] R. Casasús, E. Climent, M.D. Marcos, R. Martínez-M�a~nez, F. Sancen�on, J. Soto, etal., Dual aperture control on pH- and anion-driven supramolecular nanoscopic

hybrid gate-like ensembles, J. Am. Chem. Soc. 130 (2008) 1903e1907.[18] S. Saha, K.C.F. Leung, T.D. Nguyen, J.F. Stoddart, J.I. Zink, Nanovalves, Adv.

Funct. Mater. 17 (2007) 685e693.[19] D. He, X. He, K. Wang, M. Chen, Y. Zhao, Z. Zou, Intracellular acid-triggered

drug delivery system using mesoporous silica nanoparticles capped with T-Hg2þ-T base pairs mediated duplex DNA, J. Mater Chem. B 1 (2013)1552e1560.

[20] N. Li, Z. Yu, W. Pan, Y. Han, T. Zhang, B. Tang, A near-infrared light-triggerednanocarrier with reversible DNA valves for intracellular controlled release,Adv. Funct. Mater. 23 (2013) 2255e2262.

[21] M. Ma, H. Chen, Y. Chen, X. Wang, F. Chen, X. Cui, et al., Au capped magneticcore/mesoporous silica shell nanoparticles for combined photothermo-/chemo-therapy and multimodal imaging, Biomaterials 33 (2012) 989e998.

[22] F. Muhammad, M. Guo, W. Qi, F. Sun, A. Wang, Y. Guo, et al., pH-triggeredcontrolled drug release from mesoporous silica nanoparticles via intracelluardissolution of ZnO nanolids, J. Am. Chem. Soc. 133 (2011) 8778e8781.

[23] E. Aznar, M.D. Marcos, R. Martínez-M�a~nez, F. Sancen�on, J. Soto, P. Amor�os, etal., pH- and photo-switched release of guest molecules from mesoporoussilica supports, J. Am. Chem. Soc. 131 (2009) 6833e6843.

[24] R. Liu, Y. Zhang, X. Zhao, A. Agarwal, L.J. Mueller, P. Feng, pH-responsivenanogated ensemble based on gold-capped mesoporous silica through anacid-labile acetal linker, J. Am. Chem. Soc. 132 (2010) 1500e1501.

[25] J.L. Vivero-Escoto, I.I. Slowing, C.W. Wu, V.S. Lin, Photoinduced intracellularcontrolled release drug delivery in human cells by gold-capped mesoporoussilica nanosphere, J. Am. Chem. Soc. 131 (2009) 3462e3463.

[26] D. Zikich, N. Borovok, T. Molotsky, A. Kotlyar, Synthesis and AFM character-ization of poly(dG)-poly(dC)-gold nanoparticle conjugates, BioconjugateChem. 21 (2010) 544e547.

[27] L.M. Demers, C.A. Mirkin, R.C. Mucic, R.A. Reynolds, R.L. Letsinger,R. Elghanian, et al., A fluorescence-based method for determining the surfacecoverage and hybridization efficiency of thiol-capped oligonucleotides boundto gold thin films and nanoparticles, Anal. Chem. 72 (2000) 5535e5541.

[28] J. Kim, H.S. Kim, N. Lee, T. Kim, H. Kim, T. Yu, et al., Multifunctional uniformnanoparticles composed of a magnetite nanocrystal core and a mesoporoussilica shell for magnetic resonance and fluorescence imaging and for drugdelivery, Angew. Chem. Int. Ed. 47 (2008) 8438e8441.

[29] C. Fan, S. Wang, J.W. Hong, G.C. Bazan, K.W. Plaxco, A.J. Heeger, Beyondsuperquenching: hyper-efficient energy transfer from conjugated polymers togold nanoparticles, Proc. Natl. Acad. Sci. U. S. A. 100 (2003) 6297e6310.

[30] D.S. Seferos, A.E. Prigodich, D.A. Giljohann, P.C. Patel, C.A. Mirkin, PolyvalentDNA nanoparticle conjugates stabilize nucleic acids, Nano Lett. 9 (2009)308e311.