Multilayer fluorescent magnetic nanoparticles with dual thermoresponsive and pH-sensitive polymeric nanolayers as anti-cancer drug carriers

Abstract

A multi-step process was used to synthesize fluorescent folic acid (FA)-conjugated stimuli-responsive magnetic nanoparticles as anti-cancer drug nanocarriers. Sol–gel processing of tetraethyl orthosilicate and fluorescein isothiocyanate-conjugated 3-aminopropyltriethoxysilane was used to synthesize Fe₃O₄@SiO₂–FITC followed by the distillation precipitation polymerization of 2-hydroxyethyl methacrylate and N,N′-methylenebis(acrylamide) to obtain Fe₃O₄@SiO₂@P(HEMA) nanoparticles. Conjugating with FA and RAFT agent, Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles were synthesized via the polymerization of N-isopropylacrylamide and acrylic acid. The core–shell structure of nanoparticles was revealed via TEM. Furthermore, the progression of each step was studied by means of FT-IR and TGA. VSM and XRD were used to show that the synthesized nanoparticles retain their superparamagnetic properties. The synthesized nanoparticles exhibited dual thermo- and pH-sensitive behaviours. These nanoparticles were used as nanocarriers of the anti-cancer drug doxorubicin via controlled release in simulated physiological and acidic conditions. In addition, the synthesized nanoparticles showed a relatively non-toxic nature to HeLa cells, whereas cell viability decreased significantly when cells were incubated with DOX-loaded nanoparticles.

---

Introduction

Recent advances in drug delivery systems have driven the development of multifunctional nanoparticles with combined sensing, diagnostic, and therapeutic functions.^1–5 Among different nanoparticles, magnetic nanoparticles (MNPs) have attracted considerable attention due to their unique size-dependent properties,⁶ facile synthesis,⁷ and great potential as mediators of heat for localized hyperthermia therapy,⁸ as well as the wide range of applications such as drug delivery.⁹ Shielding MNPs not only prevent agglomeration and improve the dispersion stability of MNPs in various systems, but also facilitate further functionalization to use them in drug delivery systems.^10,11 Fe₃O₄ nanoparticles, as an example of MNPs, have received great interest as vehicles for drug delivery because they can be suitably modified to carry drug molecules and can be magnetically guided to the targeted organs or lesion sites inside the body.^12,13 In this field, silica is considered to be an appropriate material for encapsulating magnetic nanoparticles because of its good biocompatibility, excellent physicochemical stability, and ease of functionalization.^13,14 Moreover, fluorescein isothiocyanate (FITC) has been chosen as the fluorescence imaging agent and organic dye, which can be covalently attached to the silica layer and used in MRI/optical dual-modal analysis.¹⁵ In addition, folic acid (FA) is one of the most promising candidates with the potential for cancer-cell specific targeting.¹⁶ In fact, FA has a high affinity for folate receptors (FRs), which are over-expressed in various human carcinomas, e.g. breast, ovary, lung, and kidney.¹⁷ However, insufficient drug release limits the dosages of drugs to levels below the optimum therapeutic value and reduces the efficiency of drug delivery systems.¹⁸ In order to address this issue, stimuli-responsive delivery systems have been explored to improve the bioavailability of drugs.^19,20 Among the different stimuli, thermoresponsive and pH-sensitive systems are used more often due to the difference in temperature and pH values in different tissues and cell compartments. For example, the extracellular environment of a tumor has a lower pH (∼6.8) than blood and normal tissues (∼7.4),²¹ whereas that of late endosomes and lysosomes is even lower (∼5.0–5.5).²² Moreover, temperature is capable of affecting the structure of a thermoresponsive material, resulting in the gradual release of entrapped drug molecules. To date, several thermoresponsive polymers, such as poly(N-isopropylacrylamide) (P(NIPAAM)),²³ poly(2-oxazoline)s,²⁴ and poly(2-hydroxyethyl methacrylate) (P(HEMA)),^25,26 have been used in different systems. In addition, some polymers such as poly(acrylic acid) (PAA)^27,28 and poly(methacrylic acid) (PMAA)²⁹ are well-known pH-sensitive polymers. Among different strategies known for grafting polymers on nanoparticles, the “grafting from” method³⁰ is the most applicable mechanism, which can be combined with living and controlled polymerization mechanisms^31,32 that are widely applied for the synthesis of thermoresponsive polymers. Therefore, the synthesis of nanostructured particles with targeting, imaging and controlled drug release properties could be an efficient method to obtain drug carriers for different drugs.

In this work, we describe the fabrication of novel dual thermo- and pH-sensitive magnetic nanoparticles. These nanoparticles are conjugated with the targeting molecule of FA and organic fluorescent dye of FITC. Stimuli-responsive polymeric layers are synthesized via the combination of distillation precipitation polymerization (DPP) of HEMA and reversible addition–fragmentation chain transfer (RAFT) polymerization of NIPAAM and AA. After a precise characterization of the synthesized nanoparticles via transmission electron microscopy (TEM), Fourier transform infrared spectroscopy (FT-IR), thermogravimetric analysis (TGA), a vibrating sample magnetometer (VSM), X-ray diffraction (XRD) and florescence emission, they were exposed to environments with different temperatures and pH values to be examined as stimuli-responsive nanoparticles. Then, doxorubicin (DOX) as an anti-cancer drug was loaded onto the nanoparticles, and its release percentage was studied in different conditions. Finally, in vitro cellular cytotoxicity was assessed to evaluate the biocompatibility of the synthesized nanoparticles, and the cytotoxic effect of DOX-loaded nanoparticles was also investigated using HeLa cells.

Experimental methods

Preparation of Fe₃O₄@SiO₂–FITC nanoparticles

First, in a 200 mL flask, MNPs (1.000 g) were dispersed in a solution of deionized water/ethanol (20 mL/80 mL) by ultrasonication (35 min) at room temperature. The obtained suspension (6 mL) was transferred to a 50 mL flask and tetraethyl orthosilicate (TEOS) (0.85 mL) was added dropwise for 3 h, and then stirring was continued for 6 h. Then, ammonia solution (25%, 2.6 mL) was added dropwise, and after stirring for 24 h, the synthesized FITC–APTES solution (ESI, Section S2†) (5.6 mL) and TEOS (0.65 mL) were poured dropwise into the flask and the reaction was performed at room temperature in dark conditions for 48 h. After the reaction, the resulting product was reprecipitated with n-hexane (100 mL), followed by centrifugation (10 000 rpm, 30 min) to obtain Fe₃O₄@SiO₂–FITC nanoparticles. Subsequently, the nanoparticles were washed with ethanol and deionized water several times. The Fe₃O₄@SiO₂–FITC nanoparticles were dried for 24 h at 40 °C in a vacuum oven (10 mbar).

Synthesis of Fe₃O₄@SiO₂@P(HEMA) nanoparticles

Distillation precipitation polymerization (DPP)³³ of HEMA in the presence of MBA was used to synthesize a crosslinked P(HEMA) shell on the surface of Fe₃O₄@SiO₂–FITC nanoparticles. Briefly, in a 200 mL flask, Fe₃O₄@SiO₂–FITC nanoparticles (0.100 g) were dispersed in acetonitrile (80 mL) by ultrasonication (40 min) at room temperature. Then, a solution of HEMA (0.64 mL, 0.005 mol), MBA (0.160 g, 0.001 mol) and AIBN (0.016 g, 0.097 mmol) in acetonitrile (80 mL) was added to the reactor. The reaction was performed for 75 min at acetonitrile's boiling temperature, and 40 mL of acetonitrile was evaporated from the reaction flask. After the reaction, the resulting product was centrifuged (10 000 rpm, 30 min) to obtain the Fe₃O₄@SiO₂@P(HEMA) nanoparticles. Subsequently, the nanoparticles were washed with ethanol and deionized water several times. The Fe₃O₄@SiO₂@P(HEMA) nanoparticles were dried for 24 h at 40 °C in a vacuum oven (10 mbar). Monomer conversion was determined to be about 0.26 via the gravimetric method.

Synthesis of RAFT agent-attached Fe₃O₄@SiO₂@P(HEMA) nanoparticles

Fe₃O₄@SiO₂@P(HEMA) nanoparticles contain hydroxyl groups on their surface due to the P(HEMA) nanolayer. These can be used as functional groups to attach acyl bromide-containing RAFT agent (the synthesis method is described in the ESI, Section S3†). To this end, in a 50 mL flask, Fe₃O₄@SiO₂@P(HEMA) nanoparticles (0.160 g) were dispersed in toluene (20 mL) by ultrasonication (40 min) at room temperature, and TEA (2 mL) was added to the reactor, which was then placed in ice water. Then, synthesized RAFT agent (0.3 mL) was added to the reactor dropwise, and after 3 h, the reaction was performed for 24 h at room temperature. After the reaction, the resulting product was centrifuged (10 000 rpm, 30 min) to obtain RAFT agent-attached Fe₃O₄@SiO₂@P(HEMA) nanoparticles. Subsequently, the nanoparticles were washed with ethanol and deionized water several times and dried for 24 h at 40 °C in a vacuum oven (10 mbar).

Conjugating FA on the surface of RAFT agent-attached Fe₃O₄@SiO₂@P(HEMA) nanoparticles

N,N′-carbonyldiimidazole (CDI) (20 mg, 0.12 mmol) was dispersed in THF (5 mL) at 0 °C and synthesized N-glycinylmaleimide (GMI) (ESI, Section S6†) (0.016 g, 0.10 mmol) was added, and then the mixture was stirred for 12 h in a nitrogen atmosphere (Scheme 1). RAFT agent-attached Fe₃O₄@SiO₂@P(HEMA) nanoparticles (0.060 g) were dispersed in the mixture for 2 h at 55 °C. Then, the product was reprecipitated by means of diethyl ether. The obtained nanoparticles were dispersed in DMSO, and after adding synthesized folate–SH (ESI, Section S5†) (0.010 g) and TEA (20 mL), the reaction was performed at room temperature for 24 h. After the reaction, the resulting product was centrifuged (10 000 rpm, 30 min) to obtain RAFT agent- and FA-attached Fe₃O₄@SiO₂@P(HEMA) nanoparticles. Subsequently, the nanoparticles were washed with DMSO and diethyl ether several times and dried for 24 h at 40 °C in a vacuum oven (10 mbar).

Scheme 1 Synthesis route of the intermediate agent of reaction of folate–SH with RAFT agent-attached Fe₃O₄@SiO₂@P(HEMA) nanoparticles.

Synthesis of Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM) nanoparticles

In a 100 mL flask, RAFT agent- and FA-attached Fe₃O₄@SiO₂@P(HEMA) nanoparticles (0.100 g) were dispersed in DMF (15 mL) by ultrasonication (30 min). After stirring for 30 min, a mixture of NIPAAM (0.310 g, 2.65 mmol) and AIBN (1.1 mg, 6.7 μmol) was added to the reactor and polymerization was performed for 72 h at 65 °C. Monomer conversion was determined to be about 0.41 via the gravimetric method. To obtain the Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM) nanoparticles, the reaction mixture was centrifuged for 30 min (10 000 rpm). Moreover, the nanoparticles were dispersed in DMF followed by centrifuging several times and then dried for 24 h at 40 °C in a vacuum oven.

Synthesis of Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles

In a 100 mL flask, Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM) nanoparticles (0.100 g) were dispersed in DMF (15 mL) by ultrasonication (30 min). After stirring for 30 min, a mixture of AA (0.26 mL, 3.75 mmol) and AIBN (1.6 mg, 9.74 μmol) was added to the reactor and polymerization was performed for 72 h at 65 °C. Monomer conversion was determined to be about 0.35 via the gravimetric method. To obtain the Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles, the reaction mixture was centrifuged for 30 min (10 000 rpm). Moreover, nanoparticles were dispersed in DMF followed by centrifuging several times and dried for 24 h at 40 °C in a vacuum oven. The reaction procedure for the synthesis of the final nanoparticles is shown is Scheme 2.

Scheme 2 Synthesis procedure for final nanoparticles used as DOX nanocarriers.

DOX loading

A previously described method³⁴ was used to load DOX as an anti-cancer drug into the dual thermo- and pH-sensitive Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles, which are conjugated with FA. Nanoparticles (0.010 g) were dispersed in the DOX solution (3 mL, 1.0 mg mL⁻¹) via ultrasonication (30 min) and the solution was tuned to the desired pH with subsequent shaking for 48 h in dark conditions. The DOX-loaded nanoparticles were separated by centrifugation, and the loaded drug content was obtained from the drug concentrations before and after loading by means of UV-visible absorption at 480 nm.

In vitro controlled drug release

The DOX-loaded nanoparticles were dispersed in 4 mg mL⁻¹ phosphate-buffered saline (PBS) at two pH values (5.3 and 7.4). 5 mL of the dispersions was put into dialysis bags (molecular weight cut-off of 14 000). Drug release was performed at 25 and 37 °C in 120 mL of PBS solutions. To measure the amount of DOX released at different times, 1 mL of solution was withdrawn and replaced with 1 mL of fresh PBS. The DOX concentrations were determined by UV-visible absorption at 480 nm according to a pure DOX calibration curve.

In vitro cytotoxicity

Cell viability was obtained via the MTT assay of FA-conjugated Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) and drug-loaded nanoparticles with HeLa cells. The cells were seeded in 96-well plates at a density of 1 × 10⁴ cells per well and cells were cultured in 5% CO₂ at 37 °C for 24 h. Then, nanoparticles were added to the cells and incubated for 24 h. After incubation, the culture medium was replaced with fresh medium with subsequent incubation for 24 h. The cytotoxicity was expressed as the percentage of cell viability compared to that of untreated control cells.

Results and discussion

The stimuli-responsive P(NIPAAM-co-AA) brushes were grafted onto the fluorescent FA-conjugated P(HEMA)-coated magnetic silica nanospheres through RAFT polymerization. During this process, sol–gel process was combined with DPP to synthesize Fe₃O₄@SiO₂@P(HEMA) nanoparticles. Then, RAFT agent and FA were conjugated with surface of hydroxyl groups of P(HEMA), and finally, RAFT polymerization was used to synthesize the block copolymer brushes. The process is schematically described in Scheme 2.

Synthesis of multilayer magnetic nanoparticles

The corresponding TEM image of the Fe₃O₄ nanoparticles is shown in Fig. 1a, which exhibited a cubic shape with an average characteristic diameter of 60 nm. To confirm the cubic structure of nanoparticles, a scanning electron microscopy (SEM) image of the Fe₃O₄ nanoparticles is shown in Fig. 2a, in which such a shape is observed. Fe₃O₄@SiO₂–FITC nanoparticles were prepared by a sol–gel process via the simultaneous controlled hydrolysis of TEOS and APTES. As shown in Fig. 1b, these nanoparticles had a spherical shape and a well-defined core–shell structure with a slightly deep contrast core and a light contrast shell, which arose from the different mass contrasts between the magnetite core and the silica shell. Moreover, the average diameter of 160 nm shows a significant increase in the diameter of the nanoparticles. Further confirmation of the success of the sol–gel method was investigated via a SEM image, as shown in Fig. 2b. This change in the shape of the nanoparticles may be ascribed to the amorphous structure of the silica layer, in which a spherical shape is favoured. The P(HEMA) shell with MBA as the crosslinker was coated onto the Fe₃O₄@SiO₂–FITC nanoparticles by means of the in situ DPP of MBA and HEMA by the efficient synergic hydrogen-bond interactions between the amide groups of MBA, as well as the ester groups of the polar HEMA and the hydroxyl groups on the surface of Fe₃O₄@SiO₂–FITC core. The structure of these nanoparticles (Fe₃O₄@SiO₂@P(HEMA)) is shown in Fig. 1c and it shows a nanolayer with a mean thickness of 15 nm for the P(HEMA) shell. It means that the P(HEMA) shell was efficiently coated onto the Fe₃O₄@SiO₂–FITC nanoparticles with the aid of the efficient hydrogen-bonding interactions between the amide groups of the MBA crosslinkers as well as the ester groups of HEMA monomers and the silanol groups on the surface. Because of the functional hydroxyl groups on the P(HEMA) shell, these groups reacted with the RAFT agent synthesized from 2-bromoisobutyryl bromide (ESI, Section S3†). The RAFT-conjugated nanoparticles were used to synthesize P(NIPAAM-co-AA) brushes from the surface. The TEM image of the resultant Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles is shown in Fig. 1d and it exhibits a thicker organic layer than Fe₃O₄@SiO₂@P(HEMA) nanoparticles. The thickness of the organic layer is increased to 35 nm from the 15 nm observed in the Fe₃O₄@SiO₂@P(HEMA) nanoparticles. Because P(NIPAAM-co-AA) has a similar mass contrast with P(HEMA) due to their compositions, it is very difficult to distinguish the grafted P(NIPAAM-co-AA) brushes from the P(HEMA) layer. In addition, the resultant nanoparticles have retained the spherical shape with slight aggregation, originated from the interparticle hydrogen-bonding interactions.

Fig. 1 TEM images of (a) Fe₃O₄, (b) Fe₃O₄@SiO₂–FITC, (c) Fe₃O₄@SiO₂@P(HEMA) and (d) final nanoparticles.

Fig. 2 SEM images of (a) Fe₃O₄ and (b) Fe₃O₄@SiO₂–FITC nanoparticles.

The encapsulation of the Fe₃O₄ nanoparticles by silica was also investigated by FT-IR analysis by comparing the FT-IR spectra of Fe₃O₄ (Fig. 3a) and Fe₃O₄@SiO₂–FITC (Fig. 3b) nanoparticles. The Fe₃O₄ nanoparticles exhibit a strong peak at 585 cm⁻¹, which is related to the stretching vibration of the Fe–O groups.³⁵ The Fe₃O₄@SiO₂–FITC nanoparticles exhibit a strong absorption peak at 1100 cm⁻¹, which is attributed to the asymmetrical vibration of Si–O–Si group in the silica nanolayer.³⁶ The FT-IR spectrum of the Fe₃O₄@SiO₂@P(HEMA) nanoparticles is shown in Fig. 3c. The strong characteristic peak at 1730 cm⁻¹ is assigned to the stretching vibration of the carbonyl of the ester group of P(HEMA) nanolayer, and the peak at 1650 cm⁻¹ is attributed to the carbonyl group of the amide in the MBA crosslinker. On the basis of the abovementioned results, the hydrogen-bond interactions were very efficient to encapsulate the Fe₃O₄@SiO₂–FITC nanoparticles within a P(HEMA) nanolayer via DPP. Thus, the functional P(HEMA) layer with reactive hydroxyl groups can be used to graft RAFT agent and FA onto the surface to achieve polymerization processes and targeting. The FT-IR spectrum of the Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles is shown in Fig. 3d. The new absorption peaks at 2820 and 2770 cm⁻¹ are attributed to the C–H stretching of the –NHC(CH₃)₂ groups together with a strong absorption peak at 1730 cm⁻¹ arising from the C=O stretch of the amide group of the NIPAAM, whereas the absorption peak at 2930 cm⁻¹ is assigned to the C–H symmetric, and the asymmetric stretching of the methyl and methylene groups in the P(HEMA) nanolayer was similar to that of the Fe₃O₄@SiO₂@P(HEMA) nanoparticles.

Fig. 3 FT-IR spectra of (a) Fe₃O₄, (b) Fe₃O₄@SiO₂–FITC, (c) Fe₃O₄@SiO₂@P(HEMA) and (d) final nanoparticles.

TGA was performed for Fe₃O₄ (Fig. 4a) and Fe₃O₄@SiO₂–FITC (Fig. 4b) nanoparticles. According to the results, the Fe₃O₄ nanoparticles show a weight loss of 0.7 wt% between 150and 600 °C, whereas the Fe₃O₄@SiO₂–FITC nanoparticles show a weight loss of 13.8 wt% in the same temperature range, which is attributed to the degradation of the organic fluorescent FITC. The weight losses of the Fe₃O₄@SiO₂@P(HEMA) (Fig. 4c) and the Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) (Fig. 4d) nanoparticles are 47.7 wt% and 81.3 wt%, respectively. The Fe₃O₄@SiO₂@P(HEMA) nanoparticles exhibit a two-stage weight loss process between 150 and 600 °C. The first weight loss is due to the decomposition of the polymer nanolayer, which is continued by the degradation of the organic fluorescent FITC. Considering 8.4 wt% of mass loss related to the FITC (Fig. 4b), 39.3 wt% of the mass loss is attributed to the P(HEMA) crosslinked nanolayer, which shows that it is efficiently coated onto the inorganic magnetic/silica core–shell particles via DPP. According to the previous work by Xu et al.,³⁷ the weight loss of P(HEMA) was used for the calculation of the shell thickness of the P(HEMA) to be around 20 nm, which was consistent with the result from the TEM determination (15 nm). The considerable difference (33.6 wt%) between the mass loss of Fe₃O₄@SiO₂@P(HEMA) and Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles revealed the successful modification of the Fe₃O₄@SiO₂@P(HEMA) nanoparticles via RAFT polymerization of NIPAAM and AA. As seen from Fig. 4d, a three-step weight loss between 150 and 600 °C is evident. The three distinct weight loss steps may be attributed to the hierarchical structure of the crosslinked P(HEMA) nanolayer and the P(NIPAAM-co-AA) hairy-like brushes.

Fig. 4 TGA curves of (a) Fe₃O₄, (b) Fe₃O₄@SiO₂–FITC, (c) Fe₃O₄@SiO₂@P(HEMA) and (d) final nanoparticles.

The magnetic properties of the nanoparticles were studied by VSM at room temperature, as shown in Fig. 5. No significant magnetic hysteresis loops were observed for all the samples, i.e. the remanence did not exist when the magnetic field was removed. This indicates that all the nanoparticles show a superparamagnetic feature originating from the Fe₃O₄ inner core at room temperature.³⁸ The saturation magnetization (Ms) values for the Fe₃O₄, Fe₃O₄@SiO₂–FITC, Fe₃O₄@SiO₂@P(HEMA) and Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles are 80, 16, 9 and 2 emu g⁻¹, respectively. Comparison of the Ms values of the Fe₃O₄@SiO₂–FITC, Fe₃O₄@SiO₂@P(HEMA) and Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles shows the total amounts of the P(HEMA) and P(HEMA)@P(NIPAAM-co-AA) polymeric layers to be around 43.7 and 87.5 wt%, respectively, which are consistent with the results from TGA (47.7 and 81.3 wt%, respectively). With the successive polymerization processes and coating different polymeric layers onto the Fe₃O₄@SiO₂–FITC nanoparticles, the saturation magnetization value is remarkably reduced due to the decrease in the efficient mass content of the magnetite component. However, the magnetism of the final nanoparticles is still strong enough and they can be separated and controlled by an external magnetic field, as depicted in Fig. 5e and f.

Fig. 5 Hysteresis loops of (a) Fe₃O₄, (b) Fe₃O₄@SiO₂–FITC, (c) Fe₃O₄@SiO₂@P(HEMA) and (d) final nanoparticles. In addition, final nanoparticles are dispersed in water: (e) far from and (f) near a magnet.

Fig. 6 illustrates the XRD patterns of the Fe₃O₄, Fe₃O₄@SiO₂–FITC, Fe₃O₄@SiO₂@P(HEMA) and Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles. As shown in Fig. 6a, the Fe₃O₄ nanoparticles show six discernible characteristic peaks in the 2θ region between 10° and 80°.³⁹ According to the results, these six peaks are observed for all the samples (weaker for the final nanoparticles), which indicates that the crystal structure of the magnetite is unchanged and essentially maintained during the polymerization process. The crystal sizes of the magnetite are calculated to be around 42 nm using the Scherrer equation⁴⁰ for all samples, which is further evidence that the crystal structure of the magnetite is unchanged during the polymerization process. However, the peaks' intensities are decreased after each process, which confirms that the magnetic cores are incorporated into all of the samples. Simultaneously, the diffraction peak at 2θ between 21° and 25° can be assigned to the amorphous components in the different samples. The characteristic peak at 2θ = 18.1° for Fe₃O₄@SiO₂–FITC, Fe₃O₄@SiO₂@P(HEMA) and Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles is attributed to the fluorescent FITC, which exists in the silica layer in samples.^39,41

Fig. 6 XRD patterns of (a) Fe₃O₄, (b) Fe₃O₄@SiO₂–FITC, (c) Fe₃O₄@SiO₂@P(HEMA) and (d) final nanoparticles.

As stated before, FITC was chosen as the fluorescent reagent in this work. To study the fluorescence property, the fluorescence emission spectra of the FITC, Fe₃O₄@SiO₂–FITC, Fe₃O₄@SiO₂@P(HEMA) and Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles were recorded in ethanol (Fig. 7). It is shown that the peak position of the fluorescence emission for all samples is about 522 nm. It illustrates that the environments of the polymer and ethanol were nearly the same, and there is no change in the molecular structure of FITC. The fluorescence intensity gradually decreases by adding more nanolayers via sol–gel and polymerization processes. Fig. 7e and f show the images of the fluorescent Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles in normal and UV-light. It is evident that the ethanol solution of final nanoparticles is dazzling green fluorescent under the irradiation of UV light. These results confirm that the fluorescent Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles have been successfully prepared.

Fig. 7 The fluorescence intensity of (a) FITC, (b) Fe₃O₄@SiO₂–FITC, (c) Fe₃O₄@SiO₂@P(HEMA) and (d) final nanoparticles dispersed in ethanol. In addition, the final nanoparticles are dispersed in ethanol: (e) in normal light and (f) UV-light.

Stimuli-responsive properties

The solubility behaviours of the Fe₃O₄@SiO₂@P(HEMA), Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM) and Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles were investigated by UV-vis absorption at 600 nm. To study the thermo-responsiveness of the samples, aqueous suspensions of nanoparticles (1 mg mL⁻¹) were heated from 15 °C to 55 °C at pH = 6.5, and the results are depicted in Fig. 8. According to the results, the UV-vis absorbance decreases with the chain extension processes due to the more hydrophilic nature of NIPAAM and AA than HEMA. Increasing the hydrophilic blocks in the structure leads to greater solubility in water and lower UV-vis absorbance. However, all samples show a thermo-responsive behaviour, and cloud points are calculated as 29, 30 and 33 °C for theFe₃O₄@SiO₂@P(HEMA) (Fig. 8a), Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM) (Fig. 8b) and Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) (Fig. 8c) nanoparticles, respectively. It is evident that more hydrophilic structures shift the cloud point to higher temperatures. The thermoresponsive behaviour of the nanoparticles is originated from the competition between the intermolecular and intramolecular hydrogen bonding interactions. The intermolecular hydrogen bonding interactions below the lower critical solution temperature (LCST) and the intramolecular hydrogen bonding interactions above the LCST determine the thermo-responsiveness of these systems.⁴² In the case of P(HEMA), at a temperature below the LCST, the hydrophilic C=O and –OH groups in the side chains interact easily with water molecules to form intermolecular hydrogen bonds, and subsequently, Fe₃O₄@SiO₂@P(HEMA) nanoparticles exhibit a hydrophilic-like state. When the temperature is increased above the LCST, these intermolecular hydrogen bonding interactions are missed and the nanoparticles show lower solubility.⁴³ In the Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM) nanoparticles, at temperatures below the LCST, the hydrophilic C=O and N–H groups in the side chains of the P(NIPAAM) also interact with water molecules to form intermolecular hydrogen bonds. However, at temperatures above the LCST, the intramolecular hydrogen bonding interactions between the C=O and N–H groups result in a hydrophobic-like structure. Although the interaction between AA and water molecules results from electrostatic interactions, the Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles show such a behaviour with more water-solubility. To visualize the thermoresponsive behaviour of the Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles, we took a photo of a suspension of the nanoparticles at 15 (Fig. 8d) and 45 °C (Fig. 8e). A clear and transparent suspension at 15 °C and an opaque and turbid suspension at 45 °C are the best confirmation of the thermo-responsiveness of Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles.

Fig. 8 The thermoresponsive behaviour of (a) Fe₃O₄@SiO₂@P(HEMA), (b) Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM) and (c) Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles at pH = 6.5. In addition, final nanoparticles are dispersed in water: (d) at 15 and (e) 45 °C.

The pH-sensitive properties of the different samples were investigated at 25 °C at different pH values, as shown in Fig. 9. In the case of the Fe₃O₄@SiO₂@P(HEMA) nanoparticles (Fig. 9a), no evident pH-sensitive property can be seen at pH values from 2 to 10. This shows that after equilibrium in acidic media or in NaOH aqueous solution at pH up to 10, no electrostatic interaction exists between the functional groups of P(HEMA) and the ionized media. However, after equilibrium at higher pH values, the Fe₃O₄@SiO₂@P(HEMA) nanoparticles show better solubility in basic media. This could be due to the ionization of the hydroxyl groups of the P(HEMA) nanolayer. The creation of negative charges favors the formation of ionic bonds with Na⁺ as the counterion and the increase of the ionic strength.⁴⁴ In addition, at higher pH values in aqueous NaOH solution, some pendant groups of P(HEMA) can be hydrolyzed to form AA segments. This can be another reason for the higher solubility of Fe₃O₄@SiO₂@P(HEMA) nanoparticles at pH values higher than 10. In the case of Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM) nanoparticles (Fig. 9b), no more pH-sensitivity can be observed except at higher pH values. As described by other researchers, P(NIPAAM) shows no sensitivity to pH.^45,46 Therefore, such behaviour originates from the pH-sensitivity of the P(HEMA) nanolayer. However, Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles (Fig. 9c) show more hydrophilic nature than Fe₃O₄@SiO₂@P(HEMA) and Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM) nanoparticles at basic and neutral pH values. This is ascribed to the more hydrophilic properties of PAA, which result from the electrostatic repulsion between the carboxyl anions of PAA. As the alkalinity of the solution increases, the carboxylic acid groups dissociate into carboxyl anions, and then the PAA blocks become more and more extended. At acidic pH values, due to the high proton concentration, carboxylic acid moieties cannot dissociate well and the PAA blocks collapse and nanoparticles show a more hydrophobic nature.

Fig. 9 The pH-sensitive behaviour of (a) Fe₃O₄@SiO₂@P(HEMA), (b) Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM) and (c) Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles at 25 °C.

Drug loading and release

Typically, DOX was loaded into the Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanocarriers at pH = 8. The loading capacity was measured to be 0.389 mmol g⁻¹. The relatively high loading content was attributed to the surface water-soluble block copolymer and a crosslinked P(HEMA) shell. The P(NIPAAM-co-AA) contains rich –COOH groups, which can form strong hydrogen bonds with the –OH and –NH₂ groups in the DOX molecules.⁴⁷ To assess the feasibility of Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles as anti-cancer drug carriers, the in vitro release behaviour under a simulated physiological condition (pH = 7.4) and in an acidic environment (pH = 5.3) at temperatures of 25 and 37 °C was investigated (Fig. 10). The results showed that no burst release was observed in all samples. However, at pH = 7.4 and T = 25 °C, only 13.1% of DOX was released after 10 h and no more drug release took place. By increasing the temperature to 37 °C, the drug release percentage increased to 21% after 8 h and the release process was sustained until 24% of DOX was released after 40 h. This could be assigned to the collapsed polymeric nanolayers of nanoparticles because the second temperature was higher than the LCST of the nanoparticles (33 °C), which facilitates a higher degree of DOX release. However, on decreasing the pH of the drug release solution down to 5.3, the DOX release was more rapid, whereas at T = 25 °C, 57.5% of DOX was released after 10 h with sustained release of 59.2% after 40 h. By increasing the temperature to 37 °C, the drug release percentage increased to 75.1% after 8 h and the release process was sustained until 82.1% of DOX was released after 40 h. To describe this phenomenon, it should be noted that DOX is dissolved in water at low pH, but insoluble in an alkaline environment.⁴⁸ At low pH value, DOX became positively charged and soluble and could diffuse more easily from the nanolayers to aqueous media. In other words, the hydrophilic-like behaviour of the nanoparticles at lower pH value is another important reason for the rapid drug release due to collapsed polymeric nanolayers.

Fig. 10 The DOX release behaviour of final nanoparticles.

Drug release kinetics

Controlled release formulations help to realize more and more effective products with pre-determined release properties. Thus, the use of mathematical modelling is an efficient method to predict the release kinetics before the release systems are realized. Although there are some mathematical models to predict release kinetics, zero-order, first-order, Higuchi and Korsmeyer–Peppas models are the most used models.⁴⁹ Thus, the release kinetics was studied according to these models. The release kinetics model and parameters of each model were obtained by linear regression analysis, and the coefficients of correlation (R²) were used to verify the accuracy of the fitting, as shown in Table 1 (also, ESI, Fig. S2–S5†). According to the regression results, no proper data fitting occurred in the cases of the zero- and first-order models. This may be described via the source of these models. The zero-order model is used to model systems with slow release kinetics. However, the stimuli-responsive behaviour of Fe₃O₄@SiO₂@P(HEMA)@P(HEMA)@P(NIPAAM-co-AA) nanoparticles shows a rapid change in solubility in aqueous media at low pH and high temperature values. In addition, drug loading was performed at pH = 8 and 25 °C, and each release condition causes nanoparticles to show more hydrophobic nature with rapid drug release behaviour. Moreover, the first-order model is more applicable to porous matrices, whereas Fe₃O₄@SiO₂@P(HEMA)@P(HEMA)@P(NIPAAM-co-AA) nanoparticles may not be categorized in such systems. In the case of the Higuchi model, a relatively good correlation could be observed between the experimental and linear regression data for drug release at different release conditions. This can be ascribed to the application of this model for water-soluble drugs. However, according to the Korsmeyer–Peppas model, for all conditions, good correlation could be observed between the experimental and linear regression data for drug release with an n value of 0.37–0.41. This suggests that the drug release process was controlled by Fickian diffusion. However, the positive axis intercept showed the probable existence of burst effect.

Table 1 Mathematical models' correlation coefficients and release exponents for final nanoparticles

Kinetics model	Equation^49,50	Parameter	Value
			pH = 5.3		pH = 7.4
			25 °C	37 °C	25 °C	37 °C
Zero-order	Qt = Q0 + K0t	R^2	0.8637	0.8080	0.9639	0.8926
		K0	0.80	1.26	0.46	0.70
First-order		R^2	0.7922	0.6953	0.8239	0.7445
		K	0.022	0.022	0.022	0.022
Higuchi	ft = KHt^½	R^2	0.9769	0.9397	0.9922	0.9856
		KH	7.18	11.47	3.95	6.21
Korsmeyer–Peppas		R^2	0.9950	0.9729	0.9855	0.9937
		aM∞	13.80	19.05	4.90	9.55
		n	0.37	0.40	0.44	0.41

---

Cytotoxicity studies

The in vitro toxicity of the Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles with and without conjugated FA was studied via HeLa cells using an MTT assay with and without DOX loading, as shown in Fig. 11. The cells were incubated with different concentrations of nanoparticles for 24 h at 37 °C. According to the results, DOX-free nanoparticles show low cytotoxicity to HeLa cells at concentration of 5–100 μg mL⁻¹ after 24 h incubation for neat or FA-conjugated nanoparticles. However, the DOX-loaded nanoparticles have relatively high cytotoxicity. In the nanoparticles in which FA is conjugated, cell viability reaches 30.4% at a concentration of 100 μg mL⁻¹ after 24 h. This originates from the toxic nature of DOX, which is released from the nanoparticles gradually.⁵¹ In this case, DOX could passively diffuse through the cell membrane and into the cytoplasm, as well as quickly accumulate in the nucleus. However, when no FA exists in the nanoparticles, cell viability increases to 40.5% and lower toxicity is achieved. This may be ascribed to the behaviour of HeLa cells as folate receptors. In this case, FA-conjugated nanoparticles can specifically bind to folate-receptors on HeLa cells and enhance the capacity of cellular uptake against HeLa.⁵²

Fig. 11 Cytotoxicity of DOX-free and DOX-loaded Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles with and without conjugated FA.

Conclusions

The FA-conjugated fluorescent dual thermo- and pH-sensitive Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles were synthesized via the combination of sol–gel, DPP and RAFT polymerization processes. TEM, FT-IR and TGA were used to reveal the core–shell structure of nanoparticles, characterization of functional groups and progression of each step. The VSM results showed that the superparamagnetic properties of the synthesized nanoparticles are still unchanged, and according to the XRD results, the crystal structure of magnetite is intact. Furthermore, the turbidity results showed that more hydrophilic structures shift the cloud point to higher temperatures and Fe₃O₄@SiO₂@P(HEMA)@P(NIPAAM-co-AA) nanoparticles are the most sensitive to pH stimulus. In vitro drug release was studied via loading DOX as an anti-cancer drug and its controlled release was performed in physiological and acidic conditions. Results showed that DOX was released more rapidly in acidic conditions and at higher temperatures while burst release was observed. In addition, the release kinetics were studied via different methods, and it was concluded that the Korsmeyer–Peppas model is the best one to describe the release kinetics. MTT test results showed that the synthesized nanoparticles are relatively non-toxic to HeLa cells, whereas the cell viability decreased significantly when cells were incubated with DOX-loaded nanoparticles.

Notes and references

1. D.-E. Lee, H. Koo, I.-C. Sun, J. H. Ryu, K. Kim and I. C. Kwon, Chem. Soc. Rev., 2012, 41, 2656.
2. J. Gao, H. Gu and B. Xu, Acc. Chem. Res., 2009, 42, 1097.
3. S.-H. Hu and X. Gao, J. Am. Chem. Soc., 2010, 132, 7234.
4. Z. Xiao, E. Levy-Nissenbaum, F. Alexis, A. Lupták, B. A. Teply, J. M. Chan, J. Shi, E. Digga, J. Cheng, R. Langer and O. C. Farokhzad, ACS Nano, 2012, 6, 696.
5. T. D. Schladt, M. I. Shukoor, K. Schneider, M. N. Tahir, F. Natalio, I. Ament, J. Becker, F. D. Jochum, S. Weber, O. Köhler, P. Theato, L. M. Schreiber, C. Sönnichsen, H. C. Schröder, W. E. G. Müller and W. Tremel, Angew. Chem., Int. Ed., 2010, 49, 3976.
6. A. Demortière, P. Panissod, B. P. Pichon, G. Pourroy, D. Guillon, B. Donnio and S. Bégin-Colin, Nanoscale, 2011, 3, 225.
7. M. B. Gawande, A. Velhinho, I. D. Nogueira, C. A. A. Ghumman, O. M. N. D. Teodoro and P. S. Branco, RSC Adv., 2012, 2, 6144.
8. A. Hervaultab and N. T. K. Thanh, Nanoscale, 2014, 6, 11553.
9. J.-H. Lee, K.-J. Chen, S.-H. Noh, M. A. Garcia, H. Wang, W.-Y. Lin, H. Jeong, B. J. Kong, D. B. Stout, J. Cheon and H.-R. Tseng, Angew. Chem., 2013, 125, 4480.
10. C.-F. Lee, Y.-C. Wang and W.-Y. Chiu, J. Polym. Sci., Part A: Polym. Chem., 2013, 51, 2880.
11. A. M. Schmidt, Macromol. Rapid Commun., 2005, 26, 93.
12. H. Deng and Z. Lei, Composites, Part B, 2013, 54, 194.
13. P. Panahian, M. Salami-Kalajahi and M. Salami Hosseini, Ind. Eng. Chem. Res., 2014, 53, 8079.
14. G. Liu, H. Wu, H. Zheng, L. Tang, H. Hu, H. Yang and S. Yang, J. Mater. Sci., 2011, 46, 5959.
15. Y. Zhang, S. W. Y. Gong, L. Jin, S. M. Li, Z. P. Chen, M. Ma and N. Gu, Chin. Chem. Lett., 2009, 20, 969.
16. S.-J. Yang, F.-H. Lin, K.-C. Tsai, M.-F. Wei, H.-M. Tsai, J.-M. Wong and M.-J. Shieh, Bioconjugate Chem., 2010, 21, 679.
17. M.-K. Yoo, I.-K. Park, H.-T. Lim, S.-J. Lee, H.-L. Jiang, Y.-K. Kim, Y.-J. Choi, M.-H. Cho and C.-S. Cho, Acta Biomater., 2012, 8, 3005.
18. T. Zhou, C. Xiao, J. Fan, S. Chen, J. Shen, W. Wu and S. Zhou, Acta Biomater., 2013, 9, 4546.
19. A. Chatzipavlidis, P. Bilalis, L.-A. Tziveleka, N. Boukos and C. A. Charitidis, Langmuir, 2013, 29, 9562.
20. Z. Su, B. Yu, X. Jiang and J. Yin, Macromolecules, 2013, 46, 3519.
21. J. O. Lee, K. T. Oh, D. Kim and E. S. Lee, J. Mater. Chem. B, 2014, 2, 6363.
22. K. Zhou, Y. Wang, X. Huang, K. Luby-Phelps, B. D. Sumer and J. Gao, Angew. Chem., Int. Ed., 2011, 50, 6109.
23. E. Zeinali, V. Haddadi-Asl and H. Roghani-Mamaqani, RSC Adv., 2014, 4, 31428.
24. L. Tauhardt, K. Kempe, M. Gottschaldt and U. S. Schubert, Chem. Soc. Rev., 2013, 42, 7998.
25. M. Nikdel, M. Salami-Kalajahi and M. Salami Hosseini, Colloid Polym. Sci., 2014, 292, 2599.
26. M. Banaei and M. Salami-Kalajahi, Synthesis of Poly(2-Hydroxyethyl Methacrylate)-Grafted Poly(Aminoamide) Dendrimers as Polymeric Nanostructures, Colloid Polym. Sci., 2015 DOI:10.1007/s00396-015-3559-y.
27. M. Nikdel, M. Salami-Kalajahi and M. Salami Hosseini, RSC Adv., 2014, 4, 16743.
28. P. Panahian, M. Salami-Kalajahi and M. Salami Hosseini, J. Polym. Res., 2014, 21, 455.
29. P. Yang, D. Li, S. Jin, J. Ding, J. Guo, W. Shi and C. Wang, Biomaterials, 2014, 35, 2079.
30. M. Sarsabili, M. Parvini, M. Salami-Kalajahi and P. Ganjeh-Anzabi, Adv. Polym. Technol., 2013, 32, 21372.
31. P. Ganjeh-Anzabi, V. Haddadi-Asl, M. Salami-Kalajahi and M. Abdollahi, J. Polym. Res., 2013, 20, 248.
32. H. Roghani-Mamaqani, V. Haddadi-Asl, K. Khezri and M. Salami-Kalajahi, Polym. Int., 2014, 63, 1912.
33. G. He, J. Zhao, S. Hu, L. Li, Z. Li, Y. Li, Z. Li, H. Wu, X. Yang and Z. Jiang, ACS Appl. Mater. Interfaces, 2014, 6, 15291.
34. P. Du, H. Yang, J. Zeng and P. Liu, J. Mater. Chem. B, 2013, 1, 5298.
35. J. Wei, A. Du, F. Jin, Z. Wang and X. Liu, J. Magn. Magn. Mater., 2013, 340, 70.
36. N. Amirshaqaqi, M. Salami-Kalajahi and M. Mahdavian, Iran. Polym. J., 2014, 23, 699.
37. F. Xu, J. H. Geiger, G. L. Baker and M. L. Bruening, Langmuir, 2011, 27, 3106.
38. B. Liu, D. Zhang, J. Wang, C. Chen, X. Yang and C. Li, J. Phys. Chem. C, 2013, 117, 6363.
39. J. Huang, H. Liu, H. Men, Y. Zhai, Q. Xi, Z. Zhang, J. Zhang, Z. Yin and L. Li, Macromol. Res., 2013, 21, 1021.
40. J. Wang, S. Zheng, Y. Shao, J. Liu, Z. Xu and D. Zhu, J. Colloid Interface Sci., 2010, 349, 293.
41. Y. Li, C. Dong, J. Chu, J. Qi and X. Li, Nanoscale, 2011, 3, 280.
42. X.-D. Xu, C.-S. Chen, Z.-C. Wang, G.-R. Wang, S.-X. Cheng, X.-Z. Zhang and R.-X. Zhuo, J. Polym. Sci., Part A: Polym. Chem., 2008, 46, 5263.
43. J. V. M. Weaver, I. Bannister, K. L. Robinson, X. Bories-Azeau, S. P. Armes, M. Smallridge and P. McKenna, Macromolecules, 2004, 37, 2395.
44. L. Ferreira, M. M. Vidal and M. H. Gil, Int. J. Pharm., 2000, 194, 169.
45. D. Kuckling, H.-J. P. Adler, K.-F. Arndt, L. Ling and W. D. Habicher, Macromol. Chem. Phys., 2000, 201, 273.
46. J. Zhang, L.-Y. Chu, Y.-K. Li and Y. M. Lee, Polymer, 2007, 48, 1718.
47. Z. Xu, D. Wang, M. Guan, X. Liu, Y. Yang, D. Wei, C. Zhao and H. Zhang, ACS Appl. Mater. Interfaces, 2012, 4, 3424.
48. L.-B. Chen, F. Zhang and C.-C. Wang, Small, 2009, 5, 621.
49. S. Dash, P. N. Murthy, L. Nath and P. Chowdhury, Acta Pol. Pharm., 2010, 67, 217.
50. N. Mhlanga and S. S. Ray, Int. J. Biol. Macromol., 2015, 72, 1301.
51. B. Yang, Y. Li, X. Sun, X. Meng, P. Chen and N. Liu, J. Chem. Technol. Biotechnol., 2013, 88, 2169.
52. P. Yu, X.-M. Xia, M. Wu, C. Cui, Y. Zhang, L. Liu, B. Wu, C.-X. Wang, L.-J. Zhang, X. Zhou, R.-X. Zhuo and S.-W. Huang, Colloids Surf., B, 2014, 120, 142.

---

Footnote

† Electronic supplementary information (ESI) available: Materials, preparation of FITC–APTES, synthesis of RAFT agent, synthesis of FA–NHS, synthesis of folate–cysteamine (folate–SH), synthesis of N-glycinylmaleimide (GMI), drug release kinetics, instrumentation. See DOI: 10.1039/c5ra01444a

---