Fabrication of functional hollow microspheres constructed from MOF shells: Promising drug delivery systems with high loading capacity and targeted transport

An advanced multifunctional, hollow metal-organic framework (MOF) drug delivery system with a high drug loading level and targeted delivery was designed and fabricated for the first time and applied to inhibit tumour cell growth. This hollow MOF targeting drug delivery system was prepared via a simple post-synthetic surface modification procedure, starting from hollow ZIF-8 successfully obtained for the first time via a mild phase transformation under solvothermal conditions. As a result, the hollow ZIF-8 exhibits a higher loading capacity for the model anticancer drug 5-fluorouracil (5-FU). Subsequently, 5-FU-loaded ZIF-8 was encapsulated into polymer layers (FA-CHI-5-FAM) with three components: a chitosan (CHI) backbone, the imaging agent 5-carboxyfluorescein (5-FAM), and the targeting reagent folic acid (FA). Thus, an advanced drug delivery system, ZIF-8/5-FU@FA-CHI-5-FAM, was fabricated. A cell imaging assay demonstrated that ZIF-8/5-FU@FA-CHI-5-FAM could target and be taken up by MGC-803 cells. Furthermore, the as-prepared ZIF-8/5-FU@FA-CHI-5-FAM exhibited stronger cell growth inhibitory effects on MGC-803 cells because of the release of 5-FU, as confirmed by a cell viability assay. In addition, a drug release experiment in vitro indicated that ZIF-8/5-FU@FA-CHI-5-FAM exhibited high loading capacity (51%) and a sustained drug release behaviour. Therefore, ZIF-8/5-FU@FA-CHI-5-FAM could provide targeted drug transportation, imaging tracking and localized sustained release.

studied because of its exceptional chemical and thermal stability 16,17 . ZIF-8, which is composed of Zn and 2-methylimidazole, has sufficiently large pore apertures for passing drug molecules. In addition, hollow particles are fascinating materials because of their high drug loading capability 18,19 . Thus, hollow ZIF-8 with tunable functionality and controllable morphology fulfils the pre-requisites for an outstanding drug carrier.
To enhance an anticancer drug's therapeutic effects, a drug carrier should be bonded with specific molecules that can distinguish cancer cells from normal cells. FA is a suitable targeting molecule because the folate receptor is often over expressed on many cancer cell surfaces and has been rarely observed on normal cell surfaces 20,21 . Moreover, FA possesses high selectivity and binding affinity to the folate receptor, which has led to wide studies of FA as a stable and effective target-specific agent 22,23 . However, directly modifying the surface of ZIF-8 with FA is difficult. A chitosan (CHI) coating can serve as a good option to overcome this barrier. CHI contains many amino and hydroxyl groups and therefore has been used extensively for coupling with other functional groups 24,25 . Furthermore, nanomaterials can generate hydrogen bonds with CHI molecules through surface hydroxyls when inorganic nanomaterials are dispersed into CHI solutions 26 . Meanwhile, CHI does not induce immune rejections or allergic reactions, making it promising for enhancing the solubility, permeability and stability of inorganic nanomaterials 27,28 . In addition, a CHI coating can prevent quick recognition and elimination of inorganic nanomaterials by the immune system, thereby prolonging their circulation in the body 29 . Because it exhibits such properties, CHI is regarded as a sustainable material for the functionalization of inorganic materials.
Moreover, fluorescent probes are a powerful tool for the real-time visualization and tracking of drugs in living cells. The peripheral carboxyl on 5-carboxylfluorescein (5-FAM) can be easily immobilized onto the CHI, meaning that 5-FAM has the potential to be used as a fluorescent probe for monitoring the drug delivery process 30 . Thus, in this paper, a multifunctional drug delivery system for specific targeted drug delivery and fluorescence tracking imaging is reported; our objective is to minimize side effects and improve the efficiency of cancer treatment (Fig. 1). Herein, we demonstrate the fabrication of hollow MOF microparticles via a one-step solvothermal reaction of Zn(NO 3 ) 2 and 2-methylimidazole without extra template materials. FA-CHI-5-FAM is prepared by bonding the fluorescent molecule 5-FAM and targeting molecule FA to CHI through a reaction between the carboxyl groups of FA, 5-FAM and the amino groups of the CHI chain. 5-Fluorouracil (5-FU), an antimetabolite compound that can prevent tumour-cell pyrimidine nucleotide synthesis, is used as a model drug. Subsequently, hollow ZIF-8-loaded with 5-FU is coated with FA-CHI-5-FAM, resulting in the formation of ZIF-8/5-FU@FA-CHI-5-FAM. We observe that ZIF-8/5-FU@FA-CHI-5-FAM has a high drug loading level (51%). Although the drug loading capacity of ZIF-8/5-FU@FA-CHI-5-FAM does not exceed the carrier-free drug nanoparticles 31-33 , this system is superior to most drug carriers. Meanwhile they can enhance an anticancer drug's therapeutic effects by targeted drug delivery and fluorescence imaging, making them more effective than carrier-free drug nanoparticles. In addition, compared with most of the existing pure organic and inorganic carrier materials, the ZIF-8/5-FU@FA-CHI-5-FAM is not stable in acidic condition, providing excellent biodegradability and low cytotoxicity as drug delivery hosts. All in all, this system can be used to provide a sustained drug release while allowing the targeting of cancer cells and cellular imaging. Furthermore, such systems, which can function in the targeting, imaging, and therapy domains, have the potential to overcome the conventional limitations of cancer diagnosis and therapy.        For the efficient and effective delivery of 5-FU, we designed a multifunctional hollow MOF that exhibits targeting ability and a high drug loading level. Here, the 5-FU encapsulation efficiency on ZIF-8/5-FU@FA-CHI-5-FAM is 51%, which is higher than the 32% encapsulation efficiency for the original ZIF-8. The large specific surface area of the hollow ZIF-8 is reasonably regarded as a cause of the high drug loadings. The BET surface area of the hollow ZIF-8 is as high as 1596 m 2 /g, whereas the BET surface area of solid ZIF-8 is only 136.7 m 2 /g (Supp lementary Figures S9 and S10). Additionally, the controlled release of 5-FU from ZIF-8/5-FU@FA-CHI-5-FAM is conducted in PBS buffer solutions (pH = 7.4 and pH = 5 ) at 37 °C, and the drug delivery profiles are shown in Fig. 9.

Conclusion
In conclusion, a promising drug delivery system based on multifunctional hollow MOF (ZIF-8/5-FU@FA-CHI-5-FAM) is successfully developed for targeted tumour therapy and optical imaging. In this drug delivery system, the pores and hollow cavity of ZIF-8 can be used to load the anticancer drug 5-FU, FA modification of the shell provides the molecular targeting, and the imaging agent 5-FAM is used to monitor the drug delivery process through fluorescence imaging. Therefore, the as-synthesized ZIF-8/5-FU@FA-CHI-5-FAM exhibits excellent receptor-specific targeting effects for MGC-803 cells and shows an outstanding efficacy in killing the cancer cells. Additionally, the nature of the hollow ZIF-8/5-FU@FA-CHI-5-FAM for the storage and release of the 5-FU drug molecule is investigated; hollow ZIF-8 displays a higher drug loading level than the original ZIF-8. Under physiological conditions, the ZIF-8/5-FU@FA-CHI-5-FAM exhibits a sustained drug release for 45 h. Overall, this paper provides an efficacious method to explore MOFs as a new targeted drug delivery system, and we anticipate that this approach represents a promising platform for the efficient treatment of tumours. We are applying such materials as drug delivery systems, and an extension of this method is underway to address some broader biological applications.

Experimental Section
Material and Methods. All the starting reagents and solvents were of A.R. grade and were acquired from commercial sources; they were used directly without further purification. Power X-ray diffraction (PXRD) patterns were collected on a PANalytical Empyrean sharp shadow system X-ray diffractometer at a scanning rate of 2°/min in the 2θ range from 5° to 80°; the diffractometer was equipped with a Cu Kα radiation source (λ = 1.540598 Å). The size and morphology of the as-prepared nanocrystals were investigated using a HITACHI S-4800 200 kV scanning electron microscope (SEM). Transmission electron microscopy (TEM) was performed using a FEI Tecnai G 2 F20 high-resolution transmission electron microscope. Energy-dispersive X-ray (EDX) mapping analysis was carried out on a scanning electron microscope equipped with an EDX apparatus. Photoluminescence (PL) spectra were collected on a Hitachi F-7000 fluorescence spectrophotometer. Fourier transform infrared spectroscopy (FTIR) spectra were recorded with a NEXUS-670 Fourier transform infrared spectrophotometer. Thermogravimetric analysis (TGA) was performed from 40 °C to 600 °C at a heating rate of 10 °C/min using a NETZSCH STA449F3 thermal analyser; the samples were maintained under an N 2 atmosphere during the analysis. A laser scanning confocal microscope (OLYMPUS, IX81) was used to analyse the cell proliferation and the distribution of drug carriers. UV-Vis spectra were recorded in the wavelength range from 200 to 500 nm at room temperature using a TU-1901 diode UV-visible spectrophotometer.
Construction of hollow ZIF-8. The hollow ZIF-8s with excellent crystalline structures were obtained via a mild phase transformation under solvothermal conditions, as previously reported 34 . In a normal procedure, 0.558 g Zn(NO 3 ) 2 · 6H 2 O was dissolved in 15 mL methanol to obtain solution A and 0.616 g 2-methylimidazole was dissolved in 15 mL methanol to obtain solution B. After forming a homogeneous solution, solution A was added dropwise to solution B, followed by ultrasonication for 15 min at room temperature. The product was then separated via centrifugation at 11,000 rpm twice and redispersed in 15 mL methanol by ultrasonication to form solution C. Subsequently, 0.558 g Zn(NO 3 ) 2 · 6H 2 O in 15 mL methanol solution was added to solution C and then transferred to 50 mL Teflon-lined stainless steel autoclaves. The mixture was then hydrothermally treated at 120 °C for 2 h. Finally, the mixture was washed via centrifugation with methanol several times and dried at 30 °C in a vacuum oven.
Preparation of FA-CHI-5-FAM. The FA-CHI-5-FAM conjugates were prepared through a dehydration condensation reaction between the carboxyl groups of the FA, 5-FAM and amino groups of the CHI chain. Briefly, 10 mg FA and 1 mg 5-FAM were dissolved in 50 mL dimethyl sulfoxide (DMSO) with stirring to form solution A. Then, 10 mg CHI in an acetic acid aqueous solution (0.1 M, pH 4.7) was mixed with solution A to product solution B under continuous stirring. Afterwards, 20 mg N-(3-dimethylaminopropyl)-N -ethylcarbodiimide hydrochloride (EDC) was added to solution B, followed by stirring in the dark at room temperature for 16 h to allow the FA and 5-FAM to conjugate onto the CHI molecules. The solution was brought to pH 9.0 with NaOH aqueous solution (1.0 M), and the purified FA-CHI-5-FAM conjugate was obtained by centrifugation, washed with DMSO several times and finally freeze-dried.
The aforementioned synthetic procedure was subsequently used to synthesize CHI-5-FAM. Details of the procedure are provided in the supporting information.
Preparation of ZIF-8/5-FU@FA-CHI-5-FAM. Typically, 0.1 g hollow ZIF-8, 0.1 g of 5-FU and 0.5 g FA-CHI-5-FAM were added to deionized water under ultrasound, and the solution was then stirred for 4 days at room temperature. Subsequently, the product was collected by centrifugation, washed with distilled water two or three times, and then dried under vacuum at 25 °C.
The aforementioned synthetic procedure was used to synthesize ZIF-8/5-FU@CHI-5-FAM and ZIF-8@ FA-CHI-5-FAM. The detailed procedure is described in the supporting information.
Loading efficiency of 5-FU. To determine the encapsulation efficiency, several 5-FU water solutions in the 0-10 μ g/mL concentration range were prepared to obtain a calibration curve. The amount of 5-FU not loaded but contained in the excess of solvent was measured by UV-V is absorption at its wavelength of maximum absorbance (λ max = 265 nm) and subsequently calculated on the basis of the calibration curve. The drug loading efficiency was calculated by the following equation: loading efficiency (%) = (m 1 − m 2 )/m, where m 1 , m 2 and m represent the initial weight of 5-FU, the weight of 5-FU present in the excess of solvent and the weight of the hollow ZIF-8, respectively. After 4 days of soaking, the loading efficiency of 5-FU on hollow ZIF-8 was 51%, which is greater than the 32% loading efficiency of 5-FU on solid ZIF-8.
Cytotoxicity study. The in vitro cytotoxicities of ZIF-8@FA-CHI-5-FAM, ZIF-8/5-FU@FA-CHI-5-FAM and 5-FU were assessed in MGC-803 cells by the MTT method. In brief, MGC-803 cells were added to each well of a 96-well plate and incubated for 24 h. Subsequently, different concentrations of ZIF-8@FA-CHI-5-FAM, ZIF-8/5-FU@FA-CHI-5-FAM and 5-FU (0, 6.25, 12.5, 25, 50 and 100 μ g/mL) were added to the wells and incubated for another 48 h. After the previous nutrient solution was removed, 20 μ L of MTT solution was added to each well and incubated for another 4 h. All media were then removed, and 100 μ L of DMSO was added to each well. The absorbance of each sample was monitored at 570 nm using a microplate reader. The cell viability was expressed as a percentage of the absorbance of the sample well to that of the cell control. All experiments were performed in triplicate, and the results were averaged. Cellular uptake study. MGC-803 cells and HASMC cells were seeded into a 6-well plate. After culturing for 24 h at 37 °C, the cells were washed three times with phosphate-buffered saline (PBS) and blocked in PBS containing BSA (1%) at 4 °C for 20 min. The cells were then incubated with ZIF-8/5-FU@FA-CHI-5-FAM and ZIF-8/5-FU@CHI-5-FAM for 2 h (concentration was 0.1 μ g/mL). After three washings with PBS buffer, cell targeting was detected with a laser scanning confocal microscope (OLYMPUS, IX81). Cell targeting was detected on a laser scanning confocal microscope for luminescence imaging under excitation wavelength of 405 nm.
In vitro drug release study. The release assays were carried out by soaking the samples in phosphatebuffered saline (PBS, pH 7.4) at 37 °C. In short, 0.05 g ZIF-8/5-FU@FA-CHI-5-FAM was introduced into a dialysis bag and then immersed into 10 mL PBS solution in a 50 mL centrifuge tube at 37 °C. At predetermined time intervals, 2 mL of solution was withdrawn, and the amount of 5-FU released from the ZIF-8/5-FU@FA-CHI-5-FAM was examined at 265 nm by recording the UV absorbance of the solution. The cumulative release percentages of 5-FU from ZIF-8/5-FU@FA-CHI-5-FAM were calculated as follows and plotted against time: Cumulative 5-FU release = amount of released 5-FU/amount of total 5-FU × 100%.