Preparation and characterization of a dual-receptor mesoporous silica nanoparticle–hyaluronic acid–RGD peptide targeting drug delivery system

Abstract

A dual-receptor targeting delivery system based on mesoporous silica nanoparticles modified with hyaluronic acid and RGD peptide (MSNs/NH₂–HA–RGD) was developed in the present study, and characterized by TEM, SAXRD, nitrogen adsorption–desorption analysis, DLS, FT-IR, ¹³C NMR and UV-vis. The results showed that MSNs/NH₂–HA–RGD had an ideal monodispersibility, uniform particle size (172.5 ± 10 nm) and well-defined mesoporous structure. Moreover, cellular uptake results showed that MSNs/NH₂–HA–RGD had an ideal dual-receptor mediated endocytosis pathway and could be significantly internalized into ovarian cancer cells. Chlorambucil (CHL), an anticancer drug, was chosen as a model drug to investigate drug loading, in vitro drug release behaviors and cytotoxicity. The results showed that CHL-loaded MSNs/NH₂–HA–RGD exhibited a high drug loading capacity of about 10.1% and pH-sensitive drug controlled release behaviors. The cytotoxicity test showed that CHL-loaded MSNs/NH₂–HA–RGD had a stronger cytotoxicity for ovarian cancer cells than one receptor or no receptor modified MSNs. It is expected that MSNs/NH₂–HA–RGD may be a prospective candidate for targeted delivery of anticancer drugs to cancer foci.

---

1 Introduction

Cancer is a growing worldwide deadly disease for human beings. Traditional drug delivery systems that have been suggested can hardly overcome the toxic side effects of the naked drugs’ leakage before reaching the cancer target.^1–3 In recent decades, maximum therapeutic efficacy has been achieved by using active targeting drug delivery systems, which could significantly improve the uptake of tumor cells by coupling specific tumor-targeting ligands (i.e. hyaluronic acid, antibodies, peptides, etc.)^4–6 to the drug carrier or by triggering the release of anticancer drugs in the tumor site in response to a local stimulus (i.e. pH, enzymes, light, etc.).^3,7 In short, active targeting limits undesired side effects on the normal cells, and reduces the normal doses for therapeutic efficacy.^1,3,6 Thus, active targeting has good potential for cancer treatments.

Various nanocarriers have been explored extensively as potential diagnostic and therapeutic agents for cancer imaging and treatment in targeting delivery systems.^8,9 Among them, mesoporous silica nanoparticles (MSNs), an inorganic nano-carrier, have shown great potential as a targeting drug carrier due to high biocompatibility.^10–13 The textural parameters of synthesized MSNs were readily affected by different reaction conditions (i.e. pH, temperature, template, solvent, multistep addition of TEOS, etc.).^3,12 In particular, MSNs possess some unique features, such as high surface area, tunable particle size as well as pore volume, and facile surface functionalization.^14–16 For MSNs, the facile surface functionalization is helpful for the modification of different targeting moieties,^3,17–19 thereby improving MSNs for targeting cancer cells. All these features make MSNs an ideal carrier and attract high research interest in the application of MSNs as an active targeting delivery system compared to traditional organic lipids or polymer-based nanoparticles.^3,18,20

Hyaluronic acid (HA) has been extensively investigated as biomedical and pharmaceutical materials due to its biocompatibility and degradability.^1,7,21,22 It has been used as an alternative to polyethylene glycol (PEG) for prolonging half-lives of carriers in blood circulation.^3,5 Besides, it is worth mentioning that HA is also a good candidate as a targeting moiety owing to its specific internalization via CD44-mediated endocytosis, a receptor which is over expressed in cancer cells, thereby improving the overall receptor-mediated uptake of HA modified nanoparticles.^3,23–25 Recently, many drug carriers based on HA have been developed. The drug is released once HA is degraded by an acid-sensitive hyaluronidase.^24,26,27 Of various hyaluronidases involved in the degradation of HA, extracellular hyaluronidase-2 showed its ability to degrade HA to units up to 50 saccharides long, and the drug was released once the HA fragments are further degraded into tetrasaccharides by hyaluronidase-1 in cancer cells,²⁸ and could effectively improve the therapeutic efficacy towards cancer cells. However, CD44 is expressed ubiquitously in different cancers, limiting the use of HA-coupled drug carriers for some cancer treatments.^29,30 Therefore, several strategies have been developed to enhance the binding affinity and selectivity of carriers against cancer cells, including introducing integrin-binding peptides to the backbone of HA.^29,31 Integrins are heterodimeric receptors that play an important role in tumor growth.³² Among them, α_vβ₃ and α_vβ₅ (α_vβ_3/5) integrins that are specifically over-expressed in tumor vessels and most cancer cells are effectively targeted by arginine–glycine–aspartic acid (RGD).^33–37 Most importantly, integrins have limited selectivity for targeted action and are very difficult to saturate,^2,29 which can overcome the drawbacks of CD44. Therefore, many active targeting delivery systems based on RGD peptide were developed and appeared as the attractive candidate for cancer therapy by receptor binding.^4,6

In this article, dual-receptor modified mesoporous silica–hyaluronic acid–RGD peptide nanoparticles (MSNs/NH₂–HA–RGD) for targeted delivery to CD44- and integrin-overexpressing cancer cells were developed by combining MSNs, HA and RGD. As expected, MSNs/NH₂–HA–RGD had a more ideal cellular uptake than one receptor or no receptor modified MSNs. Moreover, a model anticancer drug (chlorambucil, CHL) loaded onto MSNs/NH₂–HA–RGD showed high drug loading capacity, an ideal drug release performance and ideal cytotoxicity results. The results suggested that MSNs/NH₂–HA–RGD was an effective drug delivery system and could be a potential strategy for cancer therapy.

2 Experimental section

2.1 Materials

Tetraethyl orthosilicate (TEOS), cetyltrimethylammonium bromide (CTAB), 3-aminopropyltriethoxysilane (APTES), N-hydroxysuccinimide (NHS), 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC·HCl) and N-(3-aminopropyl)methacrylamide hydrochloride (DMAPMA) were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Chlorambucil (CHL) and hyaluronic acid (HA, MW = 190 000) were purchased from Sigma-Aldrich (St. Louis, USA). The RGD-containing peptide (RGD, RhB-CGRGD) was purchased from China Peptides Co., Ltd. (Shanghai, China). All other chemicals and solvents were of analytical reagent grade and used without further purification.

2.2 Synthesis of MSNs/NH₂

Like a typical procedure reported by He et al.,¹² 0.912 g of CTAB was dissolved in 250 mL of distilled water and vigorously stirred at 80 °C for several minutes, and then 1.700 g of KH₂PO₄ and 0.152 g of NaOH were added to adjust the pH of the solution to a neutral condition, consequently accelerating the hydrolyzation and condensation of the silicon source. Immediately, 4.5 mL of TEOS was added dropwise to the CTAB solution. After 1 h, the resultant product was collected by centrifugation for 30 min with a centrifugal force of 13 000g and dispersed in 150 mL of ethanol. To extract CTAB from the nanoparticles, 2 mL of hydrochloric acid (36–38%) was added to the dispersion under ultrasonic agitation for 30 min. Finally, MSNs were dispersed in deionized water and the freeze drying powder was used for measurements.

MSNs/NH₂ was prepared by a post-grafting procedure. Briefly, 0.075 g of MSNs and 0.3 mL of triethylamine were added to 20 mL of anhydrous toluene and stirred at 80 °C for 24 h before adding 1 mL of APTES. The resulted sample (MSNs/NH₂) was filtered, washed with toluene and acetone for three times and dried in an oven at 70 °C for 24 h.

2.3 Preparation of MSNs/NH₂–HA–RGD

In brief, 10 mg of MSNs/NH₂ were sonicated in a neutral phosphate buffer solution for 45 min, and then added into a neutral phosphate buffer solution of HA (3 mg mL⁻¹) which was pre-activated by NHS (0.166 g) and EDC·HCl (0.080 g) for 15 min at room temperature. Finally, pH of the solution was adjusted to 9.0 by using triethylamine and incubated in a shaker bath under a light-sealed environment for 14 h at 38 °C. At the end of incubation, the HA modified MSNs (MSNs/NH₂–HA) were purified by centrifugation for 30 min at −4 °C, followed by washing for three times with deionized water, dialysis and freeze-drying. The obtained sample was referred to as MSNs/NH₂–HA.

10 mL of the MSNs/NH₂–HA solution (8 mM) was mixed with 8.9 mg DMAPMA and 12 mg EDC·HCl at pH 6.0. After 2 h, the same amount of DMAPMA and EDC·HCl were added again as previously described. The solution was dialyzed and freeze-dried to afford MSNs/NH₂–HA–DMAPMA.

11 mg of MSNs/NH₂–HA–DMAPMA was dissolved in 17.5 mL of the neutral phosphate buffer solution. 814 μL of Hepes buffer containing 20 mg of TCEP was mixed with 6.5 mg of RGD-containing peptide (RhB-CGRGD) in another beaker. These two solutions were then mixed, and pH was adjusted to 9.0. The solution was dialyzed and lyophilized after 2 h, resulting in MSNs/NH₂–HA–RGD.

2.4 RhB labeled MSNs/NH₂ and MSNs/NH₂–HA

2.6 mL of thionylchloride was added dropwise to an anhydrous 1,2-dichloroethane solution of RhB (0.02 g mL⁻¹) over 5 min at room temperature, and refluxed for 5 h. The solvent of 1,2-dichloroethane was then removed in a vacuum. After the crude acid chloride of RhB was dissolved with dry dichloromethane, it was added dropwise to a solution of MSNs/NH₂ (0.06 g) in triethylamine (10 mL) at 0 °C under stirring. The reaction was then allowed to keep at room temperature under nitrogen atmosphere for 24 h. The resulting solid was washed with dichloromethane, and then evaporated in a vacuum to remove the residual solvent which finally gave MSNs/NH₂–RhB. RhB was grafted onto the surface of MSN/NH₂–HA by a three-step reaction. Firstly, 0.3 mL of ethylenediamine was added dropwise to the ethanol solution of RhB, refluxed for 4 h, and then the solvent was removed by evaporation, and RhB was purified by silica column chromatography (CH₂Cl₂–MeOH = 20 : 1, v/v). Secondly, the above product was dissolved in 2 mL of DMSO, and then was added dropwise to a DMSO/distilled water solution (1 : 1) of HA, which was pre-activated by EDC·HCl (31 mg) and NHS (22 mg) at room temperature and stirred vigorously overnight. HA–RhB was obtained via dialysis and finally was lyophilized. Thirdly, HA–RhB was grafted onto the surface of MSNs by an amidation reaction.

2.5 Characterization

The morphology, pore size and mesoporous structure of the nanoparticles were characterized by transmission electron microscopy (TEM) (JEM-2100F, JEOL, Japan) at an accelerated voltage of 200 kV. The phase composition was further confirmed by small-angle X-ray diffraction (SAXRD) (D/MAX-RB, RIGAKU, Japan). Nitrogen adsorption–desorption analysis at 77 K was carried out on an adsorption analyzer to study their porosity (ASAP 2020, Micromeritics, Shanghai, China). All samples were degassed for 12 h at 423 K under vacuum before testing. The pore size distribution curve was calculated by using Brunauer–Emmett–Teller (BET) and Barrett–Joyner–Halanda (BJH) model isotherms. Particle diameter and zeta potential of the representative samples were performed by dynamic light scattering (DLS) (Mastersizer 2000, Malvern, UK). The Fourier transform-infrared (FT-IR) spectra were collected and analyzed on a FT-IR spectrometer (Nexus, Thermo Nicolet, USA). The ¹³C NMR solid-state spectrum was recorded on an Avance III spectrometer (Bruker, Switzerland). Ultraviolet-visible (UV-vis) spectroscopy was performed on a UV-2600 (Shimadzu, Japan).

2.6 Drug loading and in vitro drug release study

The model anticancer drug CHL was loaded in the mesoporous silica materials by adding the drug in the MSNs/NH₂ solution. In brief, 10 mg of MSNs/NH₂ were firstly stirred in 8 mL of a CHL–ethanol solution (1 mg mL⁻¹) for 30 min, and then were centrifugated and washed thoroughly with ethanol to completely remove the unloaded and adsorbed CHL molecules. Finally, CHL-loaded MSNs/NH₂ were further modified by HA and RGD. This CHL-loaded sample was referred to as CHL@MSNs/NH₂–HA–RGD. In order to achieve an optimal drug loading efficiency, MSNs were also used as an initial drug carrier to load CHL, and then CHL-loaded MSNs were functionalized with amino, and the rest of the steps remained the same. In addition, MSNs/NH₂–HA–RGD were also used as an initial drug carrier to mix with CHL.

To evaluate the drug loading efficiency of different drug loading methods, the drug in the decanted supernatant and the washing solutions after drug loading, and the drug loss during subsequent chemical modifications of nanoparticles were measured by HPLC. The HPLC analysis was performed on an Agilent 1100 series (Agilent, USA) as follows: Hypersil ODS2 column (4.6 × 250 mm, 10 μm); mobile phase: acetate buffer solution (pH 4.0) : methanol : acetonitrile (25 : 65 : 10, v/v); flow rate of the mobile phase: 0.8 mL min⁻¹; UV detector wavelength: 302 nm. The drug loading efficiency (LE) was calculated according to the following equation:

where W is the mass of the drug in nanoparticles; W₀ is the initial mass of the drug added.

Two buffer solutions with pH 7.4 and pH 5.5 were used to simulate the pH values of normal blood/tissues and tumor environments, respectively. CHL-loaded MSNs/NH₂–HA–RGD were added to 10 mL of release media with 150 U mL⁻¹ hyaluronidase and then put into a pretreated dialysis bag (MW cut-off = 10 000 Da). The solution was shaken at a speed of 100 rpm at 37 ± 0.5 °C under dark conditions. 3 mL of the supernatant was taken out at certain time intervals and the same volume of fresh PBS was added to the dissolution medium. The supernatant was centrifuged at 13 000 rpm for 30 min and analyzed by the HPLC technique. In vitro drug release studies were carried out at pH 7.4 and pH 5.5 at 37 ± 0.5 °C in the presence or absence of hyaluronidase for 48 h. The HPLC analysis was performed as described in the previous section in the drug loading. All release studies were carried out in triplicate.

2.7 Cell evaluation

The ovarian cancer cell line (SKOV-3) purchased from Wuhan Boster Biotech Co., Ltd. (Wuhan, China) was cultured in McCoy’s 5a medium, and the human non-cancer cell line (HOSEpiC) purchased from Guangzhou Jennio Biotech Co., Ltd. (Guangzhou, China) was cultured in RPMI 1640 medium. Cell culture was conducted at 37 °C in a humidified and 5% CO₂ atmosphere containing 10% (v/v) fetal bovine serum, 100 U mL⁻¹ penicillin and 100 μg mL⁻¹ streptomycin, and 0.5% amphotericin. Several different procedures were employed.

2.7.1 Cellular uptake of MSNs/NH₂, MSNs/NH₂–HA and MSNs/NH₂–HA–RGD. To evaluate intracellular localization, the ovarian cancer cell line (SKOV-3) and the human non-cancer cell line (HOSEpiC) were seeded into 96-well plates (5 × 10⁴ cells per well) and incubated overnight at 37 °C. The culture medium was then carefully removed from the wells and replaced with full culture mediums containing 100 μg mL⁻¹ of MSNs/NH₂–RhB, MSNs/NH₂–HA–RhB and MSNs/NH₂–HA–RGD (RhB excitation at 540 nm and emission at 625 nm), respectively. After incubation for 24 h, cells were washed with cold PBS three times to remove the residual nanoparticles, and then fixed with 4% paraformaldehyde for 30 min at room temperature. After washing three times with cold PBS, cells were stained with 5 μg mL⁻¹ of 4′,6-diamidino-2-phenylindole (DAPI, Beyotime, China, Ex at 340 nm, Em at 488 nm) to locate the nuclei. Then, 250 nM LysoTracker® Green DND-26 (Life Technologies, USA, Ex at 504 nm and Em at 511 nm) was added to the cells to locate the lysosomes. Subsequently, the cells were washed with cold PBS, and the uptake of nanoparticles was observed by confocal fluorescence microscopy (CLSM, NikonEclipse C1, Japan).

To understand the mechanism of receptor inhibition on cellular uptake, the ovarian cancer cell line (SKOV-3) was pretreated with plain medium containing an optimal inhibition concentration (40 μg mL⁻¹) of HA, RGD or HA–RGD for 30 min at 37 °C, respectively. The 40 μg mL⁻¹ concentration was measured by a series of experiments (data not shown). After completion of the procedure, the medium was carefully replaced with full medium containing the MSNs/NH₂–HA–RGD solution. Cells were then washed with PBS, and observed by a confocal fluorescence microscope, as described in the previous paragraph.

2.7.2 Cytotoxicity of MSNs/NH₂–HA–RGD, CHL, and CHL-loaded MSNs/NH₂, MSNs/NH₂–HA and MSNs/NH₂–HA–RGD. In vitro cytotoxicity against the ovarian cancer cell line (SKOV-3) and the human non-cancer cell line (HOSEpiC) was assessed by the standard Cell Counting Kit-8 (CCK-8) assay.¹⁷ Briefly, cells in their exponential growth phase were plated in 96-well plates at a density of 1 × 10⁴ cells per well. After culturing for 24 h, the medium was replaced, and then cells were incubated with various concentrations of MSNs/NH₂–HA–RGD nanoparticles (20 μg mL⁻¹, 40 μg mL⁻¹, 80 μg mL⁻¹, 120 μg mL⁻¹ and 160 μg mL⁻¹) without CHL and various concentrations of CHL (20 μg mL⁻¹, 40 μg mL⁻¹, 80 μg mL⁻¹, 120 μg mL⁻¹ and 160 μg mL⁻¹) formulations. At the end of a predefined time of 48 h, 10 mL of the CCK-8 reagent was added to each well and further incubated for another 4 h and the absorbance was read at 450 nm with a microplate reader. The cytotoxicity was expressed as a percentage of cell viability compared with the vehicle control.

3 Results and discussion

The preparation procedure of the mesoporous silica–hyaluronic acid–RGD peptide targeting delivery system is depicted in Scheme 1A. In brief, amine functionalized MSNs by the post-grafting method were firstly attached to HA (190 000 determined by gel permeation chromatography) through the EDC-mediated amidation reaction. Secondly, the surface of MSNs/NH₂–HA was functionalized with RGD-containing peptide using DMAPMA as a linker. Schematic illustration of cellular uptake of CHL@MSNs/NH₂–HA–RGD nanoparticles through a CD44- and integrin-mediated endocytosis pathway, and hyaluronidase-1 triggering the release of the drug in cancer cells is shown (Scheme 1B). The proposed mechanism of interaction^28,29 is that integrins firstly capture the nanoparticles from solution. Secondly, after the CD44–HA and integrin–RGD surface bind in a synergic fashion, extracellular hyaluronidase-2 shows its ability to degrade HA to about 50 saccharide units, followed by CD44- and integrin-mediated endocytosis. Thirdly, the HA fragments are further degraded into tetrasaccharides by hyaluronidase-1 in cancer cells and finally trigger the release of CHL in cancer cells as shown in Scheme 1B.

Scheme 1 (A) The preparation of MSNs/NH₂–HA–RGD; (B) cellular uptake of CHL@MSNs/NH₂–HA–RGD nanoparticles via a CD44- and integrin-mediated endocytosis pathway, and the release of the drug in cancer cells due to the intracellular hyaluronidase.

3.1 Synthesis of MSNs/NH₂–HA–RGD

3.1.1 TEM, SAXRD, DLS and nitrogen adsorption–desorption analysis. Fig. 1A–F shows transmission electron microscopy (TEM) micrographs, dynamic light scattering (DLS) results, small-angle X-ray diffraction (SAXRD) pattern and nitrogen adsorption–desorption analysis. Both the MSNs and MSNs/NH₂–HA–RGD were nanoparticles with ideal monodispersity and uniform particle size (111.7 ± 14 nm and 172.5 ± 10 nm respectively) (Fig. 1A and B), which was also confirmed by the DLS results (Fig. 1C), and showed a well-defined mesoporous structure in accordance with SAXRD measurements (Fig. 1D). Fig. 1E and F show the nitrogen adsorption–desorption isotherms and the corresponding pore size distributions of the MSNs, MSNs/NH₂ and MSNs/NH₂–HA–RGD, and the textural parameters are shown in Table 1. In this study, the nano-sized MSNs were successfully synthesized under a neutral condition. From Table 1, the BET surface area, BJH mesopore size distribution and total pore volume of the nano-sized MSNs are 156.9 m² g⁻¹, 2.9 nm, and 0.6 cm³ g⁻¹, respectively. However, it is noteworthy that the textural parameters of the MSNs were comparable with similar works by He and co-workers.¹² Meanwhile, as shown in Table 1, the textural parameters were gradually reduced after the introduction of an amino group and HA–RGD, indicating that the amino group and HA–RGD were successfully grafted on MSNs.

Fig. 1 TEM micrographs of MSNs (A) and MSNs/NH₂–HA–RGD (B); (C) particle size distributions of MSNs, MSNs/NH₂ in distilled water and MSNs/NH₂–HA–RGD in pH 7.4 PBS solution; (D) SAXRD patterns of MSNs, MSNs/NH₂ and MSNs/NH₂–HA–RGD; (E) nitrogen adsorption–desorption isotherms of MSNs, MSNs/NH₂ and MSNs/NH₂–HA–RGD; (F) corresponding pore size distributions of MSNs, MSNs/NH₂ and MSNs/NH₂–HA–RGD.

Table 1 Nitrogen adsorption–desorption parameters of different formulations

Formulations	BET surface area (m^2 g^−1)	BJH mesopore size distributions (nm)	Total pore volume (cm^3 g^−1)
MSNs	156.9	2.9	0.6
MSNs/NH2	140.3	2.5	0.4
MSNs/NH2–HA–RGD	25.7	1.4	0.2

---

3.1.2 Zeta potential, FT-IR, ¹³C NMR and UV-vis. Zeta potential, FT-IR spectra and ¹³C NMR solid-state spectrum analysis were performed to validate the successful functionalization of HA and RGD (Fig. 2A–D). From Fig. 2A, the zeta potential was reversed from a negative value of −11.9 ± 1 mV to a positive value of +18.9 ± 2 mV after the modification of MSNs by APTES. And the zeta potential of MSNs/NH₂–HA was −3.5 ± 0.3 mV. After grafting of RGD groups by covalent conjugation on the surface of MSNs/NH₂–HA–DMAPMA, the zeta potential was reversed from a value of −2.1 ± 0.4 mV to −5.0 ± 0.6 mV, indicating that the MSNs were successfully modified by HA and RGD. As indicated in Fig. 2B, all as-prepared samples show intensive absorption bands at 1089 cm⁻¹, 719 cm⁻¹ and 554 cm⁻¹, which were assigned as the stretching vibration of the silica structure (Si–O–Si). The band at 2939 cm⁻¹, 2852 cm⁻¹ and 1561 cm⁻¹ occurred due to the hydrolysis and condensation of –Si–(CH₂)₂–NH₂ of APTES. A lower peak at 1561 cm⁻¹ indicated that some amino groups reacted with HA. Meanwhile, the spectrum of the MSNs/NH₂–HA shows a sharp peak at 1419 cm⁻¹, which is attributed to the C=O stretching vibrations of HA on the surface of the MSNs/NH₂.²¹ The characteristic absorption peaks at 1591 cm⁻¹, 1544 cm⁻¹, 1524 cm⁻¹ and 1468 cm⁻¹ in the FT-IR spectrum of MSNs/NH₂–HA–RGD can be assigned to peaks of the benzene-containing RGD peptide. These results indicated that MSNs/NH₂–HA–RGD were successfully prepared by the facile method, which can be confirmed by ¹³C NMR solid-state spectrum analysis (Fig. 2C). 13 ppm (C1), 22 ppm (C2) and 29 ppm (C3) were assigned to the three types of methylene carbons from APTES. 95–110 ppm (C4), 70–95 ppm (C5–C7) and 174 ppm were visible in the spectrum (see the inset chemical structure in Fig. 2C) in accordance with previous reports.²⁶ The peaks of 128 ppm and 123 ppm were attributed to the benzene-containing RGD peptide. Furthermore, UV-vis spectra are shown in Fig. 2D. The characteristic peaks were observed after modification of the RGD peptide, which were further clarified following the disappearance of the characteristic peaks of the RGD peptide and the color change of the MSNs/NH₂–HA–RGD solution after centrifugation (Fig. 2D inset). The results showed that HA–RGD was successfully conjugated on the outside surface of MSNs.

Fig. 2 (A) Zeta potentials of MSNs, MSNs/NH₂, MSNs/NH₂–HA, MSNs/NH₂–HA–DMAPMA and MSNs/NH₂–HA–RGD; (B) FT-IR spectra of MSNs/NH₂, MSNs/NH₂–HA and MSNs/NH₂–HA–RGD; (C) ¹³C NMR solid-state spectrum of MSNs/NH₂–HA–RGD; (D) UV-vis spectra of MSNs, MSNs/NH₂, MSNs/NH₂–HA–DMAPMA and MSNs/NH₂–HA–RGD before and after centrifugation (inset left: digital photograph of MSNs/NH₂–HA–RGD in water; inset right: digital photograph of MSNs/NH₂–HA–RGD after centrifugation).

3.2 Drug loading efficiency

Three different drug loading methods were used to load CHL in order to achieve a relatively higher drug loading efficiency. As shown in Fig. 2A, the zeta potential of MSNs and MSNs/NH₂ was −11.9 ± 1 mV and +18.9 ± 2 mV, respectively. Meanwhile, CHL has a negative charge in ethanol. Therefore, MSNs/NH₂ had a strong electrostatic interaction with CHL. Meanwhile, this loading method, which added the drug in the solution of MSNs/NH₂, could efficiently avoid solubilisation action of CHL in toluene during the amino functionalization of MSNs by reflux. As a result, in the first loading method, when MSNs/NH₂ was used as an initial drug carrier, the CHL-loading efficiency of MSNs/NH₂ was 20.07%. After CHL-loaded MSNs/NH₂ reacted with HA, the CHL-loading efficiency of MSNs/NH₂–HA was 15.7%. At the end of the procedure, loading CHL in MSNs/NH₂ lead to a drug loading efficiency of CHL@MSNs/NH₂–HA–RGD up to 10.1%. In contrast, in the second loading method, when MSNs were used as an initial drug carrier, the drug loading efficiency of MSNs reached 17.48%, whereas the CHL-loading efficiency of MSNs/NH₂ was only 2.9% after these drug-loaded MSNs were functionalized with amine in toluene. After reaction with HA, there was almost no residual drug loaded into MSNs/NH₂–HA nanoparticles by this loading method. When MSNs were modified by an amino group and HA–RGD, the textural parameters were gradually reduced as shown in Table 1. Therefore, in the third loading method, the drug loading efficiency was only 1.3% when mixing CHL with MSNs/NH₂–HA–RGD. In short, HA–RGD conjugated on the surface of MSNs can hinder CHL leakage from MSNs, thus keeping its therapeutic activity and hindering it from reacting with amines during subsequent chemical modifications of nanoparticles. Therefore, the optimal formulation was loading CHL at the stage of MSNs/NH₂ (see Table 2). It is worth mentioning that the drug loading efficiency of CHL@MSNs/NH₂–HA–RGD nanoparticles is comparable to the work of Zhao and co-workers.³

Table 2 Drug loading efficiency of different formulations

Formulations	Drug loading efficiency
MSNs + CHL	17.48%
MSNs/NH2 + CHL	20.07%
MSNs/NH2–HA–RGD + CHL	1.3%

---

3.3 In vitro drug release

Fig. 3 shows the pH-responsive drug release behaviours of CHL from bare MSNs/NH₂ and MSNs/NH₂–HA–RGD at different pH conditions (pH 5.5 and pH 7.4) with or without hyaluronidase. CHL@MSNs/NH₂ showed a very fast controlled CHL release behaviour under different pH conditions (pH 5.5 and pH 7.4), and its cumulative amount was about 54% within 48 h at pH 7.4, but was about 37% within 48 h at pH 5.5 due to the pK_a of CHL of 4.82 ± 0.10 (Fig. 3). However, less than 10% CHL was released from MSNs/NH₂–HA–RGD in the absence of hyaluronidase at pH 7.4 within 48 h, which indicated that HA was successfully conjugated on the outside surface of CHL-loaded CHL@MSNs/NH₂, and subsequently was further functionalized with RGD, and finally the HA–RGD coating interfered with the release pattern of loaded CHL from CHL@MSNs/NH₂–HA–RGD efficiently. Nevertheless, when the pH values of the release media decreased from 7.4 to 5.5 in the presence of hyaluronidase, the cumulative amount of CHL increased significantly to 17% and 37% within 48 h, respectively. Compared with the drug release behaviours of CHL@MSNs/NH₂, the results showed that CHL@MSNs/NH₂ was successfully conjugated with HA and RGD.

Fig. 3 The cumulative release profiles of CHL from bare MSNs/NH₂ in different release media (pH 5.5 and pH 7.4) at 37 ± 0.5 °C and MSNs/NH₂–HA–RGD in different release media (pH 5.5 and pH 7.4) at 37 ± 0.5 °C with or without hyaluronidase (HY refers to hyaluronidase). Data was measured as mean ± standard deviation (n = 5), p < 0.05.

The in vitro drug release behaviors may be a result of synergetic effects of the drug delivery system. Acid-sensitive hyaluronidase showed its ability to degrade HA into small fragments at pH 5.5 and lost the function of enzymatic activity at a neutral pH condition,^24,26 which endows the system with excellent drug release characteristics. It could be noticed that MSNs/NH₂–HA–RGD allowed CHL to be responsively released from the system in tumor tissues and their surroundings (pH 5.5–5.8), but hardly any was released in normal tissues and blood (pH ∼7.4) (Fig. 3). In addition, due to HA and RGD working in a synergistic fashion, CHL@MSNs/NH₂–HA–RGD can improve targeting in tumor tissues and the overall uptake of the system. Thus, the dual-receptor targeting is believed to be a prospective candidate for cancer treatments, which has an ideal pH-responsive drug controlled release behavior in vitro.

3.4 In vitro cell evaluation

3.4.1 Cellular uptake of MSNs/NH₂, MSNs/NH₂–HA and MSNs/NH₂–HA–RGD. Fig. 4A and B show the cellular uptake of MSNs/NH₂–RhB, MSNs/NH₂–HA–RhB, MSNs/NH₂–HA–RGD and no nanoparticles (control group) in SKOV-3 cells and HOSEpiC cells. The fluorescent signals in SKOV-3 cells from RhB-labeled MSNs/NH₂–HA–RGD showed a relatively high uptake capacity, whereas, the signals of MSNs/NH₂–HA–RhB and MSNs/NH₂–RhB were very weak. Meanwhile, RhB signals of MSNs/NH₂–RhB, MSNs/NH₂–HA–RhB and MSNs/NH₂–HA–RGD emitted from the human non-cancer cell line (HOSEpiC) were significantly weak. Those findings indicated that MSNs/NH₂–HA–RGD were readily taken up by SKOV-3 cells and had a possibility to significantly improve selectivity and targeting of cancer cells through CD44- and integrin-mediated MSNs than one receptor or no receptor modified MSNs (p < 0.05). For the uptake kinetics, it can be seen from Fig. 4D that the overall amount of MSNs/NH₂–RhB, MSNs/NH₂–HA–RhB and MSNs/NH₂–HA–RGD in SKOV-3 and HOSEpiC cells occurred as a function of concentration. No significant difference in the fluorescence intensity mean was found for MSNs/NH₂–RhB between SKOV-3 and HOSEpiC cells. A 2.5-fold and 2.2-fold increase in internalization was recorded from the mean fluorescence intensity in SKOV-3 cells compared to HOSEpiC cells when they were incubated with different concentrations of MSNs/NH₂–HA–RhB (100 μg mL⁻¹ and 200 μg mL⁻¹) for 24 h (p < 0.05). In contrast, mean fluorescence intensity of MSNs/NH₂–HA–RGD gave a 2.8-fold and 2.4-fold increase for 100 μg mL⁻¹ and 200 μg mL⁻¹ respectively. Surprisingly, mean fluorescence intensity of MSNs/NH₂–HA–RGD was lower compared with that of MSNs/NH₂–HA–RhB and MSNs/NH₂–RhB. Such a decrease might be due to the high molecular weight and fluorescence quenching after modification.³ Meanwhile, it is noteworthy that the mean fluorescence intensity of MSNs/NH₂–RhB, MSNs/NH₂–HA–RhB and MSNs/NH₂–HA–RGD was weaker and weaker in HOSEpiC cells. The above findings suggested that MSNs/NH₂–HA–RGD expression in SKOV-3 cells, a CD44- and integrin-overexpressing ovarian cancer cell line,^38,39 was accompanied by a significant co-localization with HA and RGD.

Fig. 4 (A) The intracellular localization of MSNs/NH₂–RhB, MSNs/NH₂–HA–RhB and MSNs/NH₂–HA–RGD in an ovarian cancer cell line for 24 h. (B) The intracellular localization of MSNs/NH₂–RhB, MSNs/NH₂–HA–RhB and MSNs/NH₂–HA–RGD in an ovarian epithelium cell line for 24 h. (C) The images of receptor-mediated uptake of MSNs/NH₂–HA–RGD with different inhibitors (HA, RGD and HA–RGD) in the ovarian cancer cell line. (D) The mean fluorescence intensity of MSNs/NH₂–RhB, MSNs/NH₂–HA–RhB and MSNs/NH₂–HA–RGD with different concentrations in SKOV-3 and HOSEpiC cells. (E) The intracellular mean fluorescence intensity of MSNs/NH₂–HA–RGD pretreated with HA, RGD or HA–RGD. All data was measured as mean ± standard deviation (n = 5), **p < 0.05.

The two receptor-mediated endocytosis of MSNs/NH₂–HA–RGD was further evaluated by pretreating SKOV-3 cells with HA, RGD or HA–RGD. When SKOV-3 cells were treated with HA–RGD prior to adding MSNs/NH₂–HA–RGD, the fluorescent signals of MSNs/NH₂–HA–RGD were weaker as compared with the cells pretreated with free HA or RGD (Fig. 4C). The significantly weak signal of fluorescence intensity meant a strongly decreased uptake of MSNs/NH₂–HA–RGD. Compared to the SKOV-3 cells, pretreatment with HA, RGD or HA–RGD caused a decrease in the mean fluorescence intensity by 1.7-fold, 1.5-fold and 19-fold respectively (p < 0.05) (Fig. 4E), showing a significant improvement in increasing the efficacy of uptake in a synergistic fashion.

3.4.2 Cytotoxicity of MSNS/NH₂–HA–RGD, CHL, and CHL-loaded MSNs/NH₂, MSNS/NH₂–HA and MSNS/NH₂–HA–RGD. The in vitro cytotoxicity of CHL@MSNs/NH₂–HA–RGD against SKOV-3 cells and HOSEpiC cells was evaluated by using the CCK-8 assay. As shown in Fig. 5, no significant cytotoxicity was found with MSNs/NH₂–HA–RGD samples without CHL for SKOV-3 cells and HOSEpiC cells in a wide range of sample concentrations from 20 μg mL⁻¹ to 160 μg mL⁻¹. As for SKOV-3 cells (Fig. 5A), CHL, CHL@MSNs/NH₂, CHL@MSNs/NH₂–HA and CHL@MSNs/NH₂–HA–RGD showed a cytotoxic effect on the SKOV-3 cell line for different nanoparticle treatments. Importantly, a significant difference in cytotoxicity was apparent (p < 0.05) for the synthesized CHL, CHL@MSNs/NH₂–HA and CHL@MSNs/NH₂–HA–RGD with an increased CHL concentration, identified as 39%, 33% and 26% at 160 μg mL⁻¹, respectively. This suggested that the two receptor-mediated endocytosis pathway of MSNs/NH₂–HA–RGD may have an important role, and it could significantly improve the intracellular accessibility of the poorly soluble CHL towards SKOV-3 cells compared to one receptor or no receptor mediated endocytosis. In contrast, when CHL@MSNs/NH₂–HA–RGD was further incubated with the HOSEpiC cell line (Fig. 5B), poor cellular uptake and poor release of CHL from CHL@MSNs/NH₂–HA–RGD reduced the efficacy of CHL. Therefore, it had higher cell viability than free CHL even at a largest dose of CHL (160 μg mL⁻¹). It is noteworthy that the SKOV-3 cell viability of synthesized CHL@MSNs/NH₂–HA–RGD was identified as 60%, 53%, 40%, 30% and 26% at 20 μg mL⁻¹, 40 μg mL⁻¹, 80 μg mL⁻¹, 120 μg mL⁻¹ and 160 μg mL⁻¹, respectively, as shown in Fig. 5A. However, the HOSEpiC cell viability was 69%, 65%, 57%, 50% and 45%, respectively, as shown in Fig. 5B. Therefore, it indicated that there was a significant difference (p < 0.05) of cell viability between the two kinds of cells. Briefly, the MSNs/NH₂–HA–RGD had poorer cytotoxicity and could significant improve the anticancer effect of CHL@MSNs/NH₂ governed by HA and RGD predominantly, further indicating that CHL@MSNs/NH₂ was successfully functionalized with HA and RGD, and CHL@MSNs/NH₂–HA–RGD is hopeful for treating cancer.

Fig. 5 Cytotoxicity assessments of CHL, CHL@MSNs/NH₂, CHL@MSNs/NH₂–HA, CHL@MSNs/NH₂–HA–RGD and MSNs/NH₂–HA–RGD without CHL against SKOV-3 cells (A) and HOSEpiC cells (B) at different concentrations after incubation for 48 h. Data was measured as mean ± standard deviation (n = 5).

4 Conclusions

Novel dual-receptor modified mesoporous silica nanoparticles of MSNs/NH₂–HA–RGD were developed and characterized by TEM, SAXRD, nitrogen adsorption–desorption analysis, DLS, FT-IR, ¹³C NMR and UV-vis. The cell evaluation showed that MSNs/NH₂–HA–RGD had a relatively high uptake against CD44- and integrin overexpressing SKOV-3 cells than one receptor or no receptor modified MSNs (p < 0.05), indicating a synergistic effect for improving targeting to cancer cells. Furthermore, CHL was chosen as a model drug and loaded in the initial carrier of MSNs/NH₂, and then was further conjugated with HA and RGD, resulting in CHL@MSNs/NH₂–HA–RGD, which was confirmed by the cumulative release profiles and cytotoxicity assessments. CHL-loaded CHL@MSNs/NH₂–HA–RGD had a relatively higher loading efficiency of 10.1% and had pH-responsive controlled drug release behaviors in vitro. The cumulative amount of CHL loaded MSNs/NH₂ showed a natural ability for controlled release of CHL (54%) within 48 h, which was caused by the well-defined mesoporous structure. In contrast, the MSNs/NH₂–HA–RGD value was 37% at pH 5.5 for a period of 48 h, but the cumulative amount of CHL was only 17% at pH 7.4 (p < 0.05) due to acid-sensitive hyaluronidase. The in vitro cytotoxicity test showed that CHL-loaded MSNs/NH₂–HA–RGD showed a higher cytotoxic effect towards SKOV-3 cells than free CHL, CHL@MSNs/NH₂ and CHL@MSNs/NH₂–HA (p < 0.05). These findings showed that MSNs/NH₂–HA–RGD would be a promising system for targeting cancer cells and has potential for efficient cancer treatments.

Acknowledgements

This research was supported by the National Natural Science Foundation of China (grant number: 51473130 and grant number: 51572206).

Notes and references

1. A. Almalik, S. Karimi, S. Ouasti, R. Donno and C. Wandrey, et al., Biomaterials, 2013, 34, 5369–5380.
2. K. N. Sugahara, G. B. Braun, T. H. de Mendoza, V. R. Kotamraju and R. P. French, et al., Mol. Cancer Ther., 2015, 14, 120–128.
3. Q. Zhao, J. Liu, W. Zhu, C. Sun and D. Di, et al., Acta Biomater., 2015, 23, 147–156.
4. J. A. Park, Y. J. Lee, J. W. Lee, K. C. Lee and G. I. An, et al., ACS Med. Chem. Lett., 2014, 5, 979–982.
5. H. Geng, Q. Zhao, Y. Wang, Y. Gao and J. Huang, et al., ACS Appl. Mater. Interfaces, 2014, 6, 20290–20299.
6. C. Wang, B. Chen, M. Zou and G. Cheng, Colloids Surf., B, 2014, 122, 332–340.
7. M. He, Z. Zhao, L. Yin, C. Tang and C. Yin, Int. J. Pharm., 2009, 373, 165–173.
8. H. Guo, Y. Liu, Y. Wang, J. Wu and X. Yang, et al., Carbohydr. Polym., 2014, 111, 908–917.
9. N. Mezzaroba, S. Zorzet, E. Secco and S. Biffi, et al., PLoS One, 2013, 8, e74216.
10. T. Tamanna, J. B. Bulitta, C. B. Landersdorfer, V. Cashin and A. Yu, RSC Adv., 2015, 5, 107839–107846.
11. H. Moehwald, D. Borisova and D. G. Shchukin, ACS Nano, 2011, 5, 1939–1946.
12. Q. He, X. Cui, F. Cui, L. Guo and J. Shi, Microporous Mesoporous Mater., 2009, 117, 609–616.
13. Q. He, J. Shi, F. Chen, M. Zhu and L. Zhang, Biomaterials, 2010, 31, 3335–3346.
14. N. H. N. Kamarudin, A. A. Jalil, S. Triwahyono, N. F. M. Salleh and A. H. Karim, et al., Microporous Mesoporous Mater., 2013, 180, 235–241.
15. W. X. Mai and H. Meng, Integr. Biol., 2013, 5, 19–28.
16. S. Kim, S. Philippot, S. Fontanay, R. E. Duval, E. Lamouroux, N. Canilho and A. Pasc, RSC Adv., 2015, 5, 90550–90558.
17. Q. He, Y. Gao, L. Zhang, Z. Zhang, F. Gao and X. Ji, et al., Biomaterials, 2011, 32, 7711–7720.
18. C. C. Huang, W. Huang and C. S. Yeh, Biomaterials, 2011, 32, 556–564.
19. L. Jia, J. Shen, Z. Li, D. Zhang and Q. Zhang, et al., Int. J. Pharm., 2012, 439, 81–91.
20. L. Du, Y. Zhang, Y. Du, D. Yang, F. Gao and D. Tang, RSC Adv., 2015, 5, 100960–100967.
21. M. Ma, H. Chen, Y. Chen, K. Zhang and X. Wang, et al., J. Mater. Chem., 2012, 22, 5615–5621.
22. Y. Shen, J. Wang, Y. Li, H. Sun, J. Tu, B. Wang and C. Sun, RSC Adv., 2015, 5, 46464–46479.
23. L. Y. Bourguignon, P. A. Singleton, F. Diedrich, R. Stern and E. Gilad, J. Biol. Chem., 2004, 279, 26991–27007.
24. H. Harada and M. Takahashi, J. Biol. Chem., 2007, 282, 5597–5607.
25. X. Ling, C. Zhao, L. Huang, Q. Wang, J. Tu, Y. Shen and C. Sun, RSC Adv., 2015, 5, 81668–81681.
26. M. Yu, S. Jambhrunkar, P. Thorn, J. Chen and W. Gu, et al., Nanoscale, 2013, 5, 178–183.
27. H. Wang, Y.-Y. Fu, X. Zhang, C. Yu and S.-K. Sun, RSC Adv., 2015, 5, 93041–93047.
28. K. Y. Choi, H. Y. Yoon, J.-H. Kim, S. M. Bae, R.-W. Park, Y. M. Kang, I.-S. Kim, I. C. Kwon, K. Choi, S. Y. Jeong, K. Kim and J. H. Park, ACS Nano, 2011, 5, 8591–8599.
29. S. Ouasti, P. J. Kingham, G. Terenghi and N. Tirelli, Biomaterials, 2012, 33, 1120–1134.
30. X. Y. Yang, Y. X. Li, M. Li, L. Zhang and L. X. Feng, et al., Cancer Lett., 2013, 334, 338–345.
31. N. Tirelli, Y. D. Park and J. A. Hubbell, Biomaterials, 2003, 24, 893–900.
32. W.-J. Song, X.-Q. Liu, T.-M. Sun, P.-Z. Zhang and J. Wang, Mol. Pharm., 2010, 8, 250–259.
33. Y. Akashi, T. Oda, Y. Ohara, R. Miyamoto and T. Kurokawa, et al., Br. J. Cancer, 2014, 110, 1481–1487.
34. A. Li, A. Hokugo, A. Yalom, E. J. Berns and N. Stephanopoulos, et al., Biomaterials, 2014, 35, 8780–8790.
35. J. H. Park, H. J. Cho, U. Termsarasab, J. Y. Lee and S. H. Ko, et al., Colloids Surf., B, 2014, 121, 380–387.
36. R. Varshney, S. Singh, A. K. Tiwari, R. Mathur, S. Singh, P. Panwar, N. Yadav, K. Chutani, B. Singh and A. K. Mishra, RSC Adv., 2015, 5, 54439–54445.
37. A. Prasannan, T. A. Debele, H.-C. Tsai, C.-C. Chao, C.-P. Lin and G.-H. Hsiue, RSC Adv., 2015, 5, 107455–107465.
38. M. H. El-Dakdouki, E. Pure and X. Huang, Nanoscale, 2013, 5, 3895–3903.
39. D. Loessner, K. S. Stok, M. P. Lutolf, D. W. Hutmacher and J. A. Clements, Biomaterials, 2010, 31, 8494–8506.

---