Mesoporous silica nanoparticles for 19F magnetic resonance imaging, fluorescence imaging, and drug delivery† †Electronic supplementary information (ESI) available: Detailed synthetic procedure, experimental procedure and Fig. S1–S7. See DOI: 10.1039/c4sc03549f Click here for additional data file.

We described perfluorocarbon encapsulated in mesoporous silica nanoparticles which enabled dual modal imaging (NIR/19F MRI) and drug delivery.


Introduction
Efficient delivery of drugs to diseased tissues is a major goal in the eld of drug delivery in an effort to reduce adverse effects. 1 Toward this end, various nanoparticle-based drug carriers such as liposomes, polymers, and inorganic materials have been developed. 2 Among these nanocarriers, mesoporous silica nanoparticles (MSNs) have attracted signicant attention owing to their attractive properties such as extremely large surface areas (1000 m 2 g À1 ), tunable pore sizes (1.5-10 nm), and ease of functionalization via various synthetic approaches. 3 Since controlled release of drugs from the pores of MSNs results in prolonged drug efficacy, MSNs serve as ideal materials for drug delivery. 4 To assess the drug efficacy and toxicity of drug carriers, it is essential to monitor the localization of the drug carrier. Accordingly, MSNs modied with imaging agents have been developed. Fluorescence-traceable MSNs are useful for visualizing cellular localization via uorescence microscopy. In particular, near infrared (NIR) uorescent dye-modied MSNs are powerful nanomaterials for visualization in cells and in vivo localization. 5 Magnetic resonance imaging (MRI)-traceable MSNs have also attracted attention in the eld of drug delivery owing to their deep tissue imaging capabilities. MRI is a noninvasive in vivo molecular imaging technique used in both clinical-and research-based elds. 6 Gd 3+ complex-or superparamagnetic iron oxide (SPIO)-loaded MSNs are widely utilized owing to their high sensitivity. 7 Recently, multimodal imaging techniques (NIR/MRI) have gained attention because the combination of NIR and MRI provides detailed information regarding deep tissues and cell localization. 8 Therefore, MSNs that can be traced via multiple techniques (NIR/MRI) are desired. l9 F MRI has attracted signicant attention owing to the high sensitivity comparable to that of 1 H and negligible background signals. 9 19 F MRI contrast agents are suitable for tracking specic biological makers. 10 Although MSNs loaded with 19Ftraceable uorine compounds have been developed, the uorine-containing compounds can diffuse from the pores of the MSNs. 11 In contrast, peruorocarbon (PFC)-encapsulated micelles have emerged as highly sensitive 19 F MRI contrast agents and have been utilized as cell-tracking markers. 12 Although PFC encapsulated micelles with potential for use in drug delivery have been reported, 13 nanoparticles are not suitable for use as drug carriers owing to their low stability in aqueous solutions. Thus, MSNs with PFC-based cores are potentially viable for efficient drug delivery and as traceable drug carriers by 19 In a previous study, we developed novel, highly sensitive 19 F MRI contrast agents termed FLAME (FLuorine Accumulated silica nanoparticles for MRI Enhancement), composed of a PFC core and amorphous silica shell. 14 FLAME has excellent properties such as high sensitivity, feasible surface modications, and biocompatibility. Furthermore, we showed that FLAME was useful for cell labeling and in vivo tumor imaging. 14 In this study, by advancing the silica coverage of the PFC core, we developed 19 F MRI traceable MSNs as drug carriers. The MSNs consisted of the PFC core and an NIR dye modied-mesoporous silica shell, enabling both dual modal imaging (NIR/ 19 F MRI) and drug delivery. The modication of targeting ligands on the MSN surface enhanced the internalization of the MSNs into tumor cells, resulting in adequate drug efficacy due to fast drug release in acidic environments.

Results and discussion
Design, preparation, and physical properties of 19 F MRI-and uorescence-traceable drug delivery carrier For efficient drug delivery and monitoring of drug carriers, we designed a novel drug carrier with dual modal imaging capabilities (NIR/ 19 F MRI), termed mFLAME (mesoporous FLAME, Fig. 1a). mFLAME consisted of a peruoro-15-crown-5-ether (PFCE) core and mesoporous silica shell. PFCE is a highly sensitive 19 F MRI marker owing to its twenty magnetically identical uorine atoms. 12 Mesoporous silica shells are stable in aqueous solutions, and drugs can be loaded into their pores. Furthermore, active targeting to foci can be achieved by modifying targeting ligands on the mFLAME surface. To impart uorescence imaging capabilities, Cy5 dye was covalently modied on a mesoporous silica shell by silica polymerization in the presence of Cy5-conjugated 3-aminopropyltriethoxysilane (APTES).
The procedure used to prepare mFLAME is shown in Scheme S1. † Generally, PFCE requires the use of surfactants for biological applications owing to its extremely low water solubility. 12 We discovered that n-cetyltrimethylammonium bromide (CTAB), which is commonly used for the synthesis of MSNs, was capable of dispersing PFCE in water. Furthermore, the coreshell type nanoparticles that constitute the PFCE core and that of the mesoporous silica shell could be produced from the PFCE emulsions by a sol-gel process.
Characterization of the nanomaterials was carried out using dynamic light scattering (DLS). The z potential and hydrodynamic diameter of the PFCE emulsions were +51.0 mV and 78 nm, respectively. In contrast, those of mFLAME were À21.1 mV and 165 nm, respectively. The DLS data revealed that mFLAME did not form aggregates in the aqueous solution. The increase in size and decrease in z potential as compared to those of the PFCE emulsions were due to the formation of the silica shell. Transmission electron microscopy (TEM) revealed that mFLAME had mesopore and core-shell structures, and the average diameter of the particles was 79 AE 20 nm (Fig. 1b). The N 2 adsorption/desorption isotherms of mFLAME revealed a typical mesoporous structure with a Brunauer-Emmett-Teller (BET) surface area of 715 m 2 g À1 , pore volume of 1.21 cm 3 g À1 , and pore width of 6.7 nm (Fig. 1c). Fluorescence measurements conrmed that mFLAME could be traced by NIR uorescence (Fig. 1d).
Next, 19 F NMR was measured to conrm whether PFCE is encapsulated in mFLAME. The 19 F NMR of mFLAME showed a single sharp peak at À16.4 parts per million (ppm), whose chemical shi was consistent with that of PFCE (Fig. S1 †). The transverse relaxation, T 2 , of mFLAME was 0.211 s, which was almost the same as that of the PFCE emulsion (T 2 ¼ 0.242 s). As such, the silica coating did not affect the 19 F MRI sensitivity of PFCE. 19 F MRI measurements using capillary phantoms revealed strong 19 F MRI signals from the phantoms of mFLAME, and the 19 F MRI signal intensity increased according to the concentration of PFCE (Fig. 1e). In vivo MRI was performed following the injection of carboxylated mFLAME (mFLAME-COOH) (Scheme S2 †) into a living mouse. The 19 F MRI signals of mFLAME were detected in the liver, indicating that mFLAME had sufficient sensitivity for in vivo 19 F MRI applications (Fig. S2 †).

Specic cellular uptake of folate-functionalized mFLAME
To demonstrate the efficient delivery of mFLAMEs for cancer therapy, we focused on the folate receptor, which is a 38 kDa glycophosphatidylinositol-linked membrane protein found on the surface of most solid tumors. 15 Recently, it was reported that folate receptor-mediated uptake could be exploited to facilitate the entry of nanomaterials into cells. 16 Hence, a folate-functionalized mFLAME nanoparticle (mFLAME-FA) was synthesized ( Fig. 2a and Scheme S3 †).
The z-potential value of mFLAME-FA was À45.1 mV, while that of the amino-functionalized mFLAME (mFLAME-NH 2 ) was +18.4 mV. A PEGylated mFLAME (mFLAME-PEG) was also prepared as a control (Scheme S2 †). The absorption spectra of mFLAME-PEG and mFLAME-FA revealed that only mFLAME-FA had absorption peaks at 280 and 362 nm, which were derived from the folate moiety (Fig. 2b). The mean diameters of mFLAME-FA in water 0, 1, 3, 5, and 7 days aer synthesis were determined using DLS; almost no change in particle size was observed for 7 day (Fig. 2c). This result suggested that the mFLAME-FA possessed satisfactory stability in aqueous solutions. The cytotoxicity of the mFLAME nanoparticles towards KB cells was evaluated using the 3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide (MTT) assay; the cell viability was not affected by up to 0.32 mg mL À1 of PFCE (Fig. S3 †), suggesting that the mFLAMEs were biocompatible and safe for in vivo applications.
To further investigate the potential biomedical applications of mFLAMEs, their uptake into specic cells was evaluated via uorescence imaging and 19 F MRI. The expression of folate receptors on the cell surface was conrmed by a small-molecule imaging agent, uorescein isothiocyanate-folic acid (FITC-FA) (Scheme S4 †). As reported previously, 17 folate receptors were overexpressed on KB cells and not expressed in A549 cells (Fig. S4 †). The nanoparticle uptake into cells was visualized using confocal laser scanning microscopy (Fig. 3a). mFLAME-FA was internalized in KB cells. Furthermore, pre-incubation of free folic acid inhibited the specic uptake of mFLAME-FA. In contrast, the uptake of mFLAME-PEG was not observed with KB cells. In addition, almost no uorescence was observed from A549 cells incubated with mFLAME-FA or mFLAME-PEG (Fig. S5 †), indicating that folic acid on the mFLAME-FA surface was efficient for the internalization of the nanoparticle. The spot-like uorescence image also indicated that mFLAME-FA was internalized via endocytosis, and nearly all nanoparticles remained in the endosomes (Fig. S6 †).
Flow cytometry was also carried out in order to investigate the cellular uptake of mFLAME by KB cells (Fig. 3b). KB cells  took up ca. 6.5-fold more mFLAME-FA than mFLAME-PEG. Upon treatment with free folic acid, the endocytosis of mFLAME-FA decreased by more than 75%.
The nanoparticle uptake into cells was also examined by 19 F MRI. Aer mFLAME-FA was incubated with KB cells, the KB cells were transferred to a well of a microtiter plate and 1 H/ 19 F MRI experiments were performed (Fig. 3c). Fig. 3d shows MR images of the microtiter plate, including those of KB cells treated with mFLAME. 19 F MRI signals were observed from mFLAME-FA, while no 19 F MRI signal was observed from the control sample. Long acquisition time is required to acquire high contrast images. However, long acquisition time is not a disadvantage because the time scales of the dynamics of cellular uptake, DOX release, and apoptosis are not very fast. These results demonstrated that mFLAME could be used as a multimodal probe for 19 F MRI and uorescence imaging.

Drug encapsulation and cellular toxicity of drug-loaded mFLAMEs
To examine the potential of mFLAMEs for drug delivery, a chemotherapeutic agent and well-known anti-cancer drug, 18 doxorubicin (DOX), was loaded into mFLAME. The loading amount of DOX in mFLAME was calculated using the difference in the UV spectra before and aer loading. We then evaluated the release proles of encapsulated DOX at different pH values by UV-visible absorption spectroscopy. Encapsulated-DOX was released gradually from mFLAME in a time-dependent manner (Fig. 4a). Interestingly, the release rates from mFLAME at pH 5.0 were faster than those at pH 7.5, which suggested that the surrounding pH affects the electrostatic interactions between mFLAME and DOX. The results indicated that mFLAME may have useful drug release abilities in the endosome and lysosome (pH 5.0-5.5), which have a lower pH than other cellular components, such as cytosol. 19 mFLAME may also be useful for the treatment of cancer, because the pH of cancer cells is lower than that of normal tissues.
Next, DOX-loaded mFLAMEs were evaluated by assessing the viability of KB cells. Fig. 4b shows that DOX-loaded mFLAME-FA had a greater cytotoxic effect on KB cells as compared to that of mFLAME-PEG. In addition, the cytotoxicity of mFLAME-FA was greater than that of free DOX because mFLAME-FA was effectively internalized into KB cells owing to the folate ligand.
The cellular uptake of DOX-loaded mFLAME was analyzed by confocal laser scanning microscopy (Fig. 4c). The uorescence of DOX was similar in KB cells incubated with DOX-loaded mFLAME-FA and in cells treated with free DOX. In contrast, a weaker uorescence intensity was observed from cells incubated with DOX-loaded mFLAME-PEG, indicating that DOXloaded mFLAME-FA exhibited strong cytotoxic effects on KB cells owing to efficient cellular internalization and drug release in the cells.

Conclusions
We successfully developed a novel drug carrier based on a multimodal imaging agent, mFLAME, which consists of a PFCE core and mesoporous silica shell. The silica shell of mFLAME imparted various practical and useful properties such as dispersibility in water, chemical surface modiability, biocompatibility, and efficient drug loading and release capacities. The PFCE core can serve as a highly sensitive 19 F MRI contrast agent owing to the unrestricted mobility of multiple uorine nuclei in the liquid-phase core. Since most MRI traceable-MSNs consist of doped magnetite nanoparticles or Gd 3+ ion complexes for 1 H MRI, it is oen difficult to distinguish the distribution of the drug carrier due to the high background 1 H MRI signals from water and lipids in living bodies. The 19 F MRI detection system based on mFLAME has the potential to overcome this problem owing to the low background signals of 19 F MRI. Furthermore, because of the integrated features of the coreshell structure of 19 F MRI and uorescence imaging agents together with the drug delivery vehicle, mFLAME can be a useful tool for theranostic cancer treatment. By taking advantage of such properties, we demonstrated the dual modal detection of folate receptor-mediated cellular uptake via 19 F MRI and uorescence microscopy. More importantly, drug-doped mFLAME-FA exhibited adequate cellular uptake and drug release in folate receptor-overexpressing tumor cells. mFLAME should be tested in tumor-bearing mice, and simultaneous in vivo analysis of drug efficacy and biodistribution should be conducted in the near future.