Thermo-sensitively and magnetically ordered mesoporous carbon nanospheres for targeted controlled drug release and hyperthermia application

Materials Science & Engineering C 84 (2018) 21–31

Contents lists available at ScienceDirect

Materials Science & Engineering C journal homepage: www.elsevier.com/locate/msec

Thermo-sensitively and magnetically ordered mesoporous carbon nanospheres for targeted controlled drug release and hyperthermia application

T

⁎

Lin Chena,b, Huan Zhanga,c, Jing Zhenga,c, Shiping Yud, Jinglei Dud, Yongzhen Yanga,b, , ⁎⁎ Xuguang Liua,e, a

Key Laboratory of Interface Science and Engineering in Advanced Materials (Taiyuan University of Technology), Ministry of Education, Taiyuan 030024, China Research Center on Advanced Materials Science and Technology, Taiyuan University of Technology, Taiyuan 030024, China College of Chemistry and Chemical Engineering, Taiyuan University of Technology, Taiyuan 030024, China d Department of Interventional Therapy, the Second Hospital of Shanxi Medical University, Taiyuan 030001, China e College of Materials Science and Engineering, Taiyuan University of Technology, Taiyuan 030024, China b c

A R T I C L E I N F O

A B S T R A C T

Keywords: Ordered mesoporous carbon nanospheres Themo-sensitivie Magnetic Controlled release Hyperthermia Doxorubicin hydrochloride

A multifunctional nanoplatform based on thermo-sensitively and magnetically ordered mesoporous carbon nanospheres (TMOMCNs) is developed for effective targeted controlled release of doxorubicin hydrochloride (DOX) and hyperthermia in this work. The morphology, specific surface area, porosity, thermo-stability, thermosensitivity, as well as magnetism properties of TMOMCNs were verified by high resolution transmission electron microscopy, field emission scanning electron microscopy, thermo-gravimetric analysis, X-ray diffraction, Brunauer–Emmeltt–Teller surface area analysis, dynamic light scattering and vibrating sample magnetometry measurement. The results indicate that TMOMCNs have an average diameter of ~146 nm with a lower critical solution temperature at around 39.5 °C. They are superparamagnetic with a magnetization of 10.15 emu/g at 20 kOe. They generate heat when inductive magnetic field is applied to them and have a normalized specific absorption rate of 30.23 W/g at 230 kHz and 290 Oe, showing good potential for hyperthermia. The DOX loading and release results illustrate that the loading capacity is 135.10 mg/g and release performance could be regulated by changing pH and temperature. The good targeting, DOX loading and release and hyperthermia properties of TMOMCNs offer new probabilities for high effectiveness and low toxicity of cancer chemotherapy.

1. Introduction Chemotherapy is becoming one of the most important cancer treatments as the proliferation and metastasis of cancer cells can be prevented by using chemical drugs. However, as a result of low selectivity, chemical drugs inevitably cause untoward effects on normal cells [1]. For example, the anticancer drug doxorubicin hydrochloride (DOX), which has a broad spectrum anticancer activity, can kill cancer cells by directly embedding in DNA, and also induce serious adverse reactions such as bone marrow suppression and cardiac toxicity because of its low specific selectivity [2]. In order to overcome its shortcoming, pharmacy researchers are focusing on the design of new medicine formulation to improve the efficacy and reduce toxic side effects. Therefore, the concept of drug delivery systems was proposed to deliver DOX

at the right place and suitable time with expected dose, in order to achieve maximum efficacy and minimal side effects [3]. Ordered mesoporous carbon nanospheres (OMCNs) have attracted enormous attention in drug delivery fields because of their good biocompatibility, stability and high specific surface area [4]. As potential candidate drug vehicles, OMCNs show excellent drug loading capacity owing to their high porosity and high pore volume. Drug molecules are adsorbed and stored not only onto the pore surface, but also throughout the interconnected pore structure. Moreover, OMCNs are succeptible to modification with abundant bonding sites by amino, carboxyl and thiol groups, which can effectively govern the diffusion of drugs. To date, several OMCNs-based delivery systems for transporting DOX have been explored [5–7]. For example, Zhou et al. [5] developed an OMCNsbased delivery system with DOX loading capacity of 41.0 mg/g and

⁎ Correspondence to: Y. Yang, Key Laboratory of Interface Science and Engineering in Advanced Materials (Taiyuan University of Technology), Ministry of Education, Taiyuan 030024, China. ⁎⁎ Correspondence to: X. Liu, College of Materials Science and Engineering, Taiyuan University of Technology, Taiyuan 030024, China. E-mail addresses: [email protected] (Y. Yang), [email protected] (X. Liu).

https://doi.org/10.1016/j.msec.2017.11.033 Received 21 July 2017; Received in revised form 23 October 2017; Accepted 24 November 2017 Available online 26 November 2017 0928-4931/ © 2017 Elsevier B.V. All rights reserved.

Materials Science & Engineering C 84 (2018) 21–31

L. Chen et al.

phenol, phosphoric acid and ammonium persulfate (APS) were purchased from Tianjin Guangfu Technology Development Co., Ltd., China. Ferric nitrate nonahydrate (Fe(NO3)3·9H2O) was supplied by Tianjin Fuchen Chemical Reagents Plant, China. Acetic acid, γ-methacryloxy propyl trimethoxyl silane (MPS), sodium hydroxide (NaOH) and ethanol were purchased from Tianda Chemical Reagent Factory, Tianjin, China. All chemicals and reagents were of analytical grade and were used as received without further purification. α-Fe2O3 nanoparticles were prepared by our previous work [16]. Distilled water was used during the experiment.

cumulative release of 70% in 15 h, exhibiting good inhibition for cancer cells. Although OMCNs show favorable DOX loading performance, drug leakage and fast release always take place during their delivery, as the DOX loading ability is only relied on the physical adsorption interactions between OMCNs and DOX, leading to low efficiency and high toxicity. Therefore, lacking of targeted and controlled release properties limits further applications of DOX delivery system based on OMCNs [8]. To improve the targeted and controlled release properties of OMCNs nano-carriers, many efforts have been undertaken to develop multifunctional OMCNs with stimuli responsiveness, which could target and control the DOX release by actively or passively responding to the outside environment, such as magnetic field [9], temperature [10], and pH [11]. Multifunctional OMCNs drug carriers provide new perspectives for drug delivery, controlled release and targeted treatment in the research of cancer chemotherapy [12]. Strategies for developing multifunctional OMCNs drug carriers are proposed as follows: first, introduction of Fe3O4 nanoparticles into OMCNs to impart magnetic targeting properties, which could rapidly realize targeted release of DOX at cancer tissue and induced hyperthermia by an alternating magnetic field [13]; second, grafting of thermo-sensitive polymer layer around the surfaces of magnetic OMCNs (MOMCNs) to bestow thermosensitivity [14], by which a DOX storage gate as well as a release switch is created to response to temperature for the controlled release of DOX upon volume phase transition of thermo-sensitive polymer induced by hyperthermia effect; third, design of a pH-responsive DOX delivery system, in which -NH2 groups are protonated under acidic conditions to become hydrophilic and desalinated under base conditions to become hydrophobic [15]. The peripheral tissue of cancer generally show lower pH value (5.5–6.0) compared with normal cells as more acidic metabolites are produced in the metabolism of cancer cells [15]. Thus, by means of different affinity between DOX carrier and tissue fluid under different pH, the DOX release could be further controlled. Herein, in order to improve the targeted and controlled release of DOX delivery system based on OMCNs, thermo-sensitively and magnetically OMCNs were constructed for targeted and controlled release of DOX and hyperthermia in this work. The DOX-loading system is targeted at cancer tissue by magnetic field at first, and then controlled release of DOX is realized by the endogenous pH stimuli of cancer tissues and the extraneous temperature stimuli that are induced by hyperthermia. It is expected to increase the efficacy of DOX and decrease the side effects, thus overcoming the limitations of conventional cancer treatment methods. First, OMCNs were prepared by soft template method in which α-Fe2O3 nanoparticles were used as catalyst. To obtain a targeted DOX delivery system with hyperthermia effect, MOMCNs were synthesized through impregnation method. Subsequently, silane modified MOMCNs (SMOMCNs) were obtained to introduce C]C double bonds on the surfaces of MOMCNs. Then, thermo-sentitive polymer poly(N-isopropylacrylamide) (PNIPAM), whose lower critical solution temperature (LCST) is quite close to physiological temperature of human being, was grafted by radical polymerization to obtain thermo-sensitive MOMCNs (TMOMCNs). Finally, the loading and controlled release of DOX in vitro and the hyperthermia were investigated, as shown in Fig. 1.

2.2. Thermo-sensitively and magnetically ordered mesoporous carbon nanospheres 2.2.1. Ordered mesoporous carbon nanospheres OMCNs were prepared by soft template method reported in the literature [17], with α-Fe2O3 nanoparticles as catalyst. The procedures were as follows: 0.6 g of phenol was dissolved in 2.1 mL of formaldehyde, and then 15 mL of 0.1 mol/L NaOH solution and certain amount of α-Fe2O3 nanoparticles were added, followed by sonicating for 10 min until the α-Fe2O3 nanoparticles were completely dispersed. After the mixture were reacted at 70 °C for 30 min, 0.96 g of F127 dissolved in 15 mL of distilled water was added and stirred at 66 °C for 2 h, and 50 mL of distilled water was added to dilute the mixture and the resultant mixture was continually stirred at 66 °C for 17 h. After that, 10 mL of above precursor solution and 30 mL of distilled water were put into a 100 mL teflon-lined stainless autoclave, heated at 130 °C for a period of time, and then cooled to room temperature. The products were collected by centrifugation, washed with distilled water and ethanol for several times, and dried for 10 h. The carbonization was performed at 700 °C in N2 atmosphere for 1 h. Finally, the black powder was rinsed by distilled water, purified by removing magnetic components that were produced by α-Fe2O3 nanoparticles and dried to obtain OMCNs. 2.2.2. Magnetic ordered mesoporous carbon nanospheres MOMCNs were prepared using the modified impregnating method as described previously [18]. Theoretical content of the magnetic nanoparticles could be calculated from Fe(NO3)3·9H2O and pore structure properties of OMCNs. If all of the mesopore were impregnated by Fe (NO3)3·9H2O, the mass of Fe(NO3)3·9H2O and the content of magnetic nanoparticle could be calculated according to the following equation:

m(Fe2 (NO3 )3⋅9H2 O) = m(C) ∗Vt (C) ∗ρ(Fe(NO3 )3⋅9H2 O)

(1)

where m(Fe(NO3)3·9H2O) and m(C) present the mass of Fe(NO3)3·9H2O and OMCNs, respectively; Vt(C) is the pore volume of OMCNs (~ 0.64 cm3/g), which is obtained from Table 1; ρ(Fe(NO3)3·9H2O) is the density of Fe(NO3)3·9H2O (1.684 g/cm3); M (Fe(NO3)3·9H2O) and M (Fe2O3) are the mole mass of Fe(NO3)3·9H2O and Fe2O3, respectively (g/mol). When 0.30 g of OMCNs was used to prepare MOMCNs, the dosage of Fe(NO3)3·9H2O is 0.32 g. Therefore, 0.3 g of OMCNs and 0.32 g of Fe(NO3)3·9H2O were dispersed into 5 mL of ethanol and sonicated for 15 min, and 200 μL of 0.2 mol/L HCl solution was added to inhibit the hydrolysis of Fe (NO3)3·9H2O. The dispersion was sealed and stirred at 25 °C for 1 h, then open stirred until ethanol was evaporated completely. After being dried at 60 °C under vacuum for 3 h, the powder was put into a little open glass bottle, which was then placed into a 100 mL teflon-lined stainless autoclave containing 10 mL of 14 wt% ammonia solution. The glass bottle was used to separate the powder from ammonia solution. After sealing, the autoclave was heated to 60 °C for 3 h to hydrolyze ferric nitrate to ferric hydroxide in situ and then cooled to room temperature. The blackish product was filtered and rinsed by water to remove byproducts, and then annealed for 30 min at 500 °C under N2 atmospheres with a heating rate of 10 °C/min to obtain MOMCNs.

2. Materials and methods 2.1. Materials Pluronic poly(ethylene oxide)-poly(propylene oxide)-poly(ethylene oxide) (F127) was supplied by Sigma-Aldrich Company Ltd., China. DOX was obtained from Shanghai Aladdin Biochemical Technology Co., Ltd., China. N-isopropylacrylamide (NIPAM) and N,N′-methylenebisacrylamide (MBA) were purchased from Sinopharm Chemical Reagent Co., Ltd., China. Phosphate buffer (PBS, pH = 7.4) was supplied by Wuhan Doctor Bioengineering Co., Ltd., China. Ammonia solution (25–28 wt%), formaldehyde (37–40 wt%), hydrochloric acid (36%), 22

Materials Science & Engineering C 84 (2018) 21–31

L. Chen et al.

Fig. 1. Scheme of the synthesis, DOX loading and release of TMOMCNs.

2.3. Inductive heating property

Table 1 SBET, pore size and pore volume of OMCNs, MOMCNs and TMOMCNs. Sample

SBET (m2/g)

Pore size (nm)

Pore volume (cm3/g)

OMCNs MOMCNs TMOMCNs

559.10 344.91 74.73

3.43 3.04 3.38

0.64 0.41 0.20

SP-04C high frequency induction heating equipment (Shenzhen Shuangping Double Power Technology Co., Ltd., China) was chosen to measure the inductive heating effect. A 100 mg/mL TMOMCNs PBS dispersion was prepared and placed in the high- frequency induction heating equipment with a heating current of 400 A. The relationships between the temperature of PBS dispersion and heat time were recorded and analyzed. Specific absorption rate (SAR) is one of important parameters for evaluating inductive heating effect of magnetic medium [19]. It represents the rate of the samples absorbing the energy from alternating magnetic field per unit mass (W/g). SAR can be calculated as follows:

2.2.3. Synthesis of thermo-sensitively and magnetically ordered mesoporous carbon nanospheres Silane coupling agents are important organic silicon compounds for the synthesis of organic/inorganic composite materials. In this work, to obtain a thermo-sensitive polymer layer on MOMCNs, MPS was used to introduce C]C on the surfaces of MOMCNs at first [19]. Subsequently, in the presence of initiator APS and crosslinker MBA, monomer NIPAM was grafted on the surfaces of SMOMCNs. In a typical procedure, 2 mL of MPS was added into 60 mL of the mixed solvent of ethanol and water (v/v = 2:1). After the pH was adjusted to 5.0 by acetic acid, 0.30 g of MOMCNs was added, followed by sonication for 10 min. The mixture was transferred to a thermostat water bath and reacted at 65 °C for 4 h under N2 atmosphere. SMOMCNs were obtained by magnetic separation, rinsed by using ethanol and distilled water and dried for 10 h. Then, 0.3 g of NIPAM, 0.045 g of MBA and 0.1 g of SMOMCNs were dispersed in 45 mL of distilled water. The mixture were sonicated for 15 min and transferred to a thermostat water bath, followed by purging with N2 for 30 min. When the temperature was raised up to 65 °C, 1 mL of 1.31 × 10− 5 mol/L APS aqueous solution was added dropwise into the mixture to initiate the free radical polymerization for 10 h. Finally, the product was collected by magnetic separation, rinsed by distilled water and dried for 12 h, to obtain TMOMCNs.

SAR = C ×

ΔT M × Δt m

(2)

where C represents the heat capacity of media solution, the value is 4.186 J/(g k); ΔT and Δt represent changes in temperature (°C) and time (minute), respectively; M and m stand for the total mass of PBS suspension and the mass of TMOMCNs (mg), respectively. 2.4. DOX loading and release DOX loading and release property was analyzed by using U-3900 UV–Visible spectrophotometer (UV–Vis, HITACHI, Japan) with scanning range of 200–600 nm, scan rate of 600 nm/min, the exiting slot of 1 nm. First of all, different concentrations of DOX PBS solution were prepared and standard curve was obtained by recording their absorbance at 480 nm. 2.4.1. Dox loading Effects of loading time on loading capacity were investigated. Typically, 20 mg of TMOMCNs was dispersed into 20 mL of DOX solution in PBS (100 mg/L) and kept shaking in incubator shaker (25 °C 23

Materials Science & Engineering C 84 (2018) 21–31

L. Chen et al.

Japan) and transmission electron microscope (TEM, JEM-100CXII, Japan). Thermo-gravimetric analysis (TGA) of samples was performed at thermal analyzer (NETCHES TG209F3, Germany) under N2 and air atmospheres to characterize their thermo-stability and the grafting ratio of polymer. Crystal structure analysis was carried out by X-ray diffractometer (XRD, RIGAKU D/max-2005 with Cu Kα, Japan). Specific surface area (SBET), pore size and pore volume were recorded by Brunauer–Emmet–Teller (BET) analysis (Quantachrome QUADRASORB SI, USA) at 77 K. The samples were degassed at 150 °C for 6 h under vacuum and pore distribution was calculated by BJH method. Fouriertransform infrared (FTIR) spectra of samples were recorded by FTIR spectrometer (TENSOR 27, Germany) over the range of 400–4000 cm− 1 to detect surface functional groups. Vibrating sample magnetometer (SQUID, Quantum Design Versalab, USA) was used to investigate the magnetic property of the samples under a magnetic field of 20 kOe at 298 K. Hydrodynamic diameter was detected by dynamic light scattering (DLS, Zetasizer Nano-ZS90, UK) to investigate the thermo-sensitivity of the TMMOCNs and the average values were obtained from three parallel measurements.

and 150 rpm) under dark condition. At predetermined time intervals, the suspensions were separated by a magnet for 5 min and then aliquots (50 μL) were taken out and filtered by 0.22 μm aqueous filter membrane. The released free DOX was quantified by the absorbance band of the molecules measured by UV–Vis centered at 480 nm. The loading capacity Qt (mg/g) is calculated by following equation:

Qt =

(C0 − Ct ) × V m

(3)

where C0 and Ct represent the concentration of DOX in the initial solution and the separated solution when the loading time is t (mg/L), respectively. V is the volume of the solution (L), and m is the mass of TMOMCNs (g). Subsequently, effects of DOX concentration on loading capacity were investigated in a similar way with the above study. The initial DOX concentration ranges from 100 to 1000 mg/L and the loading time is 24 h. The equilibrium capacity of DOX loading Qe (mg/g) is calculated according to Eq. (4):

Qe =

(C0 − Ce ) × V m

(4) 3. Results and discussion

where C0, V and m are the same as in Eq. (3), Ce represents the equilibrium concentration of DOX (mg/L).

3.1. The effects of the concentration of α-Fe2O3 nanoparticles and hydrothermal reaction time on the formation of OMCNs

2.4.2. DOX release In vitro release test of DOX from TMOMCNs was evaluated by dialysis method for investigating the pH- and thermo-sensitively controlled release. Typically, 4 mg of DOX-loading nanospheres was redispersed in 4 mL of PBS solution (pH = 7.4), and then filled into a dialysis bag (1000 D). The dialysis bag was immersed into 50 mL of PBS solution with different pH (5.5, 6.2 and 7.4) in a thermostatic oscillator (37 °C or 45 °C, 150 rpm) keeping out of light. The released DOX outside the dialysis bag was sampled by 2 mL at defined time intervals and assayed by UV–Vis spectrometry at 480 nm. Meantime, 2 mL of PBS solution was supplemented. Cumulative release Q is calculated by following equation:

In the preparation of magnetic OMCNs, our original intention was to synthesize them by a facile one-step soft template method with α-Fe2O3 nanoparticles as magnetic source and phenolic resin as carbon source, while an unexpected result was discovered. OMCNs could be produced rapidly in the presence of α-Fe2O3 nanoparticles without formation of magnetic OMCNs, demonstrating a catalytic effect of α-Fe2O3 nanoparticles. Similar results have been confirmed by Cui et al. [20] when they prepared ropelike carbon nanostructures with iron oxide as catalyst. Therefore, effects of α-Fe2O3 nanoparticles on the formation of OMCNs were discussed in this work. The dosage of α-Fe2O3 nanoparticles was adopted at 0, 3.9 × 10− 4 and 7.8 × 10− 4 mol/L to investigate the morphology of OMCNs affected by α-Fe2O3 nanoparticles when the hydrothermal reaction time was 16 h. Results show that no solid products were formed in the absence of α-Fe2O3 nanoparticles. When the dosage of α-Fe2O3 nanoparticles was increased to 3.9 × 10− 4 mol/L, OMCNs formed with irregular morphology and wide size distribution, as shown in Fig. 2(a). It demonstrates that α-Fe2O3 nanoparticles could accelerate the formation of OMCNs. Fig. 2(b) shows FESEM images of OMCNs obtained with 7.8 × 10− 4 mol/L α-Fe2O3 nanoparticles. It can be seen that uniform OMCNs were obtained with an average of diameter of 106 nm, illustrating a regulation effect of α-Fe2O3 nanoparticles. Given that the suitable size of nanodrug carrier for delivery is around 100–200 nm, which favors their metabolism [15], an average diameter of 106 nm of OMCNs could satisfy the requirement of nano-carrier for drug delivery.

i−1

V0 × Ci + V × ∑ Ci − 1 Q(%) =

i=2

m×X

× 100%

(5)

where V0 and V stand for the volume of the solution (50 mL) and every sampling (2 mL), respectively. Ci is the release concentration of DOX at different time (mg/L), m is the total mass of the DOX-loading nanospheres (mg), and X is the DOX loading percentage (%). X(%) = (m′ / m) ∗ 100%, and m′ is the quality of DOX loaded on the nanospheres. 2.5. Apparatus and characterization Morphology and microstructure features of samples were analyzed with field emission electron microscope (FESEM, JEOL JSM-6700F,

Fig. 2. FESEM images of OMCNs with different dosage of α-Fe2O3 nanoparticles: 3.9 × 10− 4 mol/L (a) and 7.8 × 10− 4 mol/L (b) at a reaction time of 16 h.

24

Materials Science & Engineering C 84 (2018) 21–31

L. Chen et al.

Fig. 3. FESEM images of OMCNs with different reaction time: 12 h (a), 20 h (b), and 24 h (c) and TGA curves of OMCNs and MOMCNs (d).

Therefore, 7.8 × 10− 4 mol/L of α-Fe2O3 nanoparticles is regarded as the suitable concentration. Generally, hydrothermal reaction time remarkably affects products' morphology and size. Therefore, the effects of hydrothermal time (12, 20 and 24 h) on the formation of OMCNs at the dosage of α-Fe2O3 nanoparticles 7.8 × 10− 4 mol/L were investigated. Fig. 3(a) shows OMCNs obtained at 12 h with a wide size distribution ranging from 20 to 140 nm, suggesting that the nucleation and growth of OMCNs are simultaneously progressed within 12 h. The OMCNs obtained at 20 h have a similar and uniform morphology (as shown in Fig. 3(c)), which illustrates that the growing up of nucleated OMCNs dominates. When the reaction time was extended to 24 h, OMCNs continually grew and aggregated to bulk carbon materials (Fig. 3(c)). These results demonstrate that uniform OMCNs can be obtained by 16 h of hydrothermal treatment in the presence of 7.8 × 10− 4 mol/L α-Fe2O3 nanoparticles. In 2010, Fang et al. [17] demonstrated a novel low-concentration hydrothermal route to synthesize highly ordered body-centered cubic (Im3m) mesoporous carbon nanoparticles with spherical morphology and uniform size. In their procedure, 20–140 nm OMCNs could be synthesized under 24 h of hydrothermal treatment of phenolic resin and F127 mixture. It can be seen that the presence of α-Fe2O3 nanoparticles could promote and regulate the formation of OMCNs, which shorten reaction time from 24 to 16 h, proving the catalytic effects of α-Fe2O3 nanoparticles. To determine whether α-Fe2O3 nanoparticles were encapsulated by carbon structure, TGA was carried out in air atmosphere, as shown in Fig. 3(d). It can be seen from TGA curve of OMCNs, the amorphous carbon structure is rapidly oxidized by oxygen in air at around 494 °C. At 900 °C, OMCNs were fully converted to gaseous products with a weight loss of 100%. It indicates that there is no iron components encapsulated in OMCNs. This phenomenon may be due to fewer functional groups on the surfaces of α-Fe2O3 nanoparticles, leading to the weaker interaction between phenolic resin prepolymer and α-Fe2O3 nanoparticles.

3.2. Structure characterization 3.2.1. Morphology TEM and FESEM were used to observe morphologies of the products from different stages. Fig. 4(a) shows the TEM image of OMCNs. The body-centered cubic mesostructure with space-group Im3m symmetry is confirmed in Fig. 4(a) inset. The pore size can be measured to be ~3.3 nm. After magnetization, MOMCNs nearly keep the same average diameter of OMCNs (Fig. 4(b)), illustrating that the impregnating of Fe (NO3)3·9H2O could not affect the ordered mesoporous structures of nanospheres. Fig. 4(c) gives the TEM image of MOMCNs, which clearly shows the embedding of metallic nanoparticles in the carbon matrix. To confirm the crystal structure of nanoparticles, XRD analysis of MOMCNs (Fig. 4(d)) was performed. The XRD pattern shows a wide range of diffraction peaks in 10–30°, which is assigned to highly amorphous carbon structure produced by the carbonization of phenolic resin. Diffraction peaks at 30.2°, 35.3°, 43.2°, 53.5°, 57.1° and 62.9° are assigned to iron oxide nanoparticles. By comparing XRD pattern of MOMCNs with those of γ-Fe2O3 (the JCPD card no. 39-1346) and Fe3O4 (the JCPD card no. 88-0866), it is found that the crystal is accordance with Fe3O4 nanoparticles. Therefore, the presence of Fe3O4 nanocrystals with a grain size of ~13 nm is confirmed, as calculated by DebyeSherrer formula (Eq. (6)).

D=

K×λ β × cos θ

(6)

where β is full width at half maximum, K = 0.89, λ = 0.15 nm, θ represents the Bragg diffraction angle (rad). SMOMCNs were prepared by introducing C]C double bond onto the surfaces of MOMCNs with silane coupling agent MPS. As Fig. 4(e) displays, SMOMCNs have an average size of ~125 nm, 21 nm larger than that of MOMCNs. Slight decrease in sphericity and dispersion is also found, demonstrating MPS is grafted onto the surfaces of MOMCNs. In the presence of initiator and cross-linking agent, NIPAM is initiated and polymerized on SMOMCNs to obtain TMOMCNs, whose FESEM image is shown in Fig. 4(f). It can be seen that average particle size 25

Materials Science & Engineering C 84 (2018) 21–31

L. Chen et al.

Fig. 4. TEM image of OMCNs (a), FESEM image of MOMCNs (b), TEM image of MOMCNs (c), XRD pattern of MOMCNs (d), FESEM image of SMOMCNs (e) and FESEM image of TMOMCNs (f).

increases to ~146 nm. The change in size results from the encapsulation of the thermo-sensitive polymer shell. 3.2.2. Surface functional groups FTIR spectrometry was used to investigate the surface chemistry of OMCNs after functionalization (Fig. 5). For MOMCNs and SMOMCNs, the broad bands at 3419 and 580 cm− 1 are attributed to eOH and FeeO vibration, respectively. The weak bands at around 1383 cm− 1 can be assigned to carboxyl O]CeO bands. After silanization, the surface chemistry of SMOMCNs is evident different at 1094 cm− 1 due to SieO vibration. It suggests the introduction of silane agent onto the surfaces of MOMCNs. The surface modification with thermo-sensitive monomer onto SMOMCNs leads to new bands at around 1630 cm− 1, which corresponds to C]O originating from NIPAM [21]. Consequently, the thermo-sensitive polymer is cross-linked on the surfaces of SMOMCNs. 3.2.3. Porous structures As SBET, pore size and pore volume are important parameters influencing the drug loading property of drug-nanocarriers, N2 adsorption isotherms of OMCNs, MOMCNs and TMOMCNs were measured, as shown in Fig. 6(a). Three N2 adsorption–desorption curves are similar.

Fig. 5. FTIR spectra of MOMCNs, SMOMCNs, TMOMCNs and NIPAM.

26

Materials Science & Engineering C 84 (2018) 21–31

L. Chen et al.

Fig. 6. N2 adsorption–desorption curves (a) and pore size distribution (b) of OMCNs, MOMCNs and TMOMCNs.

annealing. After silanization, the weight loss of SMOMCNs is 8.57% at 600 °C, which is 5.59% higher than that of MOMCNs. Fig. 7(b) shows the DTG curves, in which SMOMCNs have two decomposition peak temperatures at 230 and 417 °C, which is attributed to the low stability of silane agent, thus demonstrating grafting of silane agent onto the surfaces of MOMCNs. Compared with that of SMOMCNs, the weight loss of TMOMCNs increases to 14.22%, and decomposition temperatures decrease to 226 and 397 °C, respectively, which is caused by the grafting of polymers onto the nanospheres. In our previous work [19], the weight loss of PNIPAM at 600 °C is 90.28%, therefore, the grafting ratio of PNIPAM can be calculated as follows: x × 90.28% + (1 − x) × 8.57% = 14.22%. The obtained grafting ratio x is 6.92 wt %.

According to International Union of Pure and Applied Chemistry, the adsorption type belongs to the V type and the average pore diameter is ~ 3 nm, as shown in Fig. 6(b), illustrating mesoporous nature of the materials. Moreover, a hysteresis loop of H1 type at higher pressure (P/ P0 = 0.9–0.993) demonstrates their quite uniform pore size distribution [22]. Table 1 lists the SBET, pore size and pore volume of OMCNs, MOMCNs and TMOMCNs. Pore size distributions are calculated by BJH model from the adsorption branches of isotherms. For OMCNs, the pore size is 3.43 nm, which agrees with that observed from TEM image. Compared with that of OMCNs, the SBET of MOMCNs decreases from 559.10 to 344.91 m2/g, and the pore volume reduces by 0.23 cm3/g. It may be caused by the embedding of Fe3O4 nanoparticles into the carbon structure, which occupy part of the pore volume. SBET of TMOMCNs further decreases to 74.73 m2/g, and the pore volume reduces further by 0.21 cm3/g. The N2 adsorption–desorption isotherms were measured under dried state. At this state, PNIPAM crosslinked around the surfaces of MOMCNs may block the pore structure, leading to a lower SBET and pore volume. Nevertheless, the mesoporous structures can provide enough space for DOX loading and release, as will be shown in latter sections.

3.3. Thermo-sensitivity The thermo-sensitivity of TMOMCNs was studied by DLS in the temperature range from 25 to 55 °C, as shown in Fig. 8. It is obvious that the hydrodynamic diameter of TMOMCNs decreases from 289 to 255 nm, suggesting the thermo-sensitivity of TMOMCNs. The differentiated DLS curve demonstrates the hydrodynamic diameter decreases drastically at around 39.5 °C, indicating its phase transition temperature (Lower Critical Solution Temperature, LCST) is about 39.5 °C. This phenomenon can be explained by the fact that when the temperature is lower than LCST, the polymer is hydrophilic owing to hydrogen bond formed between water molecules and amide group. However, when the temperature becomes higher than LCST, a cleavage of hydrogen bond takes place, resulting in a hydrophobic status. Water molecules are squeezed outside of polymer network and the polymer chains shrinkage with a corresponding decrease of the hydrodynamic diameter of TMOMCNs [24]. This result confirms the thermo-sensitivity of TMOMCNs. By comparing the FESEM image of TMOMCNs with their hydrodynamic diameter, it can be found that the size of TMOMCNs characterized by DLS is ~150 nm larger than the observed value by FESEM. This difference is mainly due to their different measuring principles. The size of TMOMCNs is observed under dried state by FESEM, while their hydrodynamic diameter is obtained when they are dispersed in water with a hydration layer around the nanoparticles [24].

3.2.4. Thermal stability and grafting ratio To investigate the thermal stability and get more information about the compositions of different samples, TGA was carried out on OMCNs, MOMCNs, SMOMCNs and TMOMCNs. Fig. 3(d) compares the TGA behavior of OMCNs and MOMCNs under air atmospheres. The as-prepared MOMCNs have a decomposition temperature of 407 °C, which is 87 °C lower than OMCNs, owing to the existence of iron component accelerating the oxidation of carbon. At 900 °C, the residual mass of MOMCNs is 17.52%, which is originated from the oxidation from Fe3O4 nanoparticles to Fe2O3. Therefore, the carbon content can be calculated by Eq. (7) [23], which is 81.95%.

wt %Carbon = 1 − wt %R ×

2M(Fe3 O4 ) 3M(Fe2 O3 )

(7)

where wt%R stands for the amount of remaining ash after combustion in air; wt%Carbon represents the amount of the carbon, M(Fe3O4) and M (Fe2O3) are the mole mass of Fe3O4 and Fe2O3 (g/mol), respectively. Meanwhile, the content of Fe3O4 can be calculated from chemical Eq. (8), which is 16.94%.

4Fe3 O4 + O2 = 6Fe2 O3

3.4. Magnetism and hyperthermia property

(8)

As magnetic property is very important for targeted drug carriers, the magnetic hysteresis loops of MOMCNs and TMOMCNs at 298 K were measured. As shown in Fig. 9(a), when the strength of magnetic field is 20 kOe, the saturation magnetization (Ms) is 11.10 emu/g, much lower than that of bulk magnetic materials as a result of their small size. After grafting thermo-sensitive polymer, the Ms. of TMOMCNs further decreases to 10.15 emu/g, which is attributed to the

Moreover, the TGA measurements of MOMCNs, SMOMCNs and TMOMCNs were conducted under N2 atmosphere with a heating rate of 10 °C/min, as shown in Fig. 7(a). Because MOMCNs were obtained by annealing under 500 °C, they are thermally stable enough and only a weight loss of 2.98% is observed at 600 °C, which is attributed to decomposition of the stable polymer that did not carbonize during 27

Materials Science & Engineering C 84 (2018) 21–31

L. Chen et al.

Fig. 7. TGA (a) and DTG (b) curves of MOMCNs, SMOMCNs, and TMOMCNs in N2 atmosphere.

Fig. 8. Hydrodynamic diameters of TMOMCNs. Fig. 10. The heating curve of TMOMCNs in alternating magnetic field.

decrease of magnetic content in the inorganic/organic hybrid materials. The inset of Fig. 9(a) provides evidence for typical superparamagnetic properties of both MOMCNs and TMOMCNs because of their almost zero coercivity and remanence. Such performance may prevent particle aggregation of nanoparticles and enable their application in biomedical and bioengineering field [9]. Magnetic response property is also confirmed by putting a 0.24 T magnet beside the suspension of TMOMCNs, as shown in Fig. 9(b). Magnetic materials can generate heat when induced by an alternating magnetic field, which could heat cancer tissue to 41–47 °C to kill cancer cells [25]. Therefore, the hyperthermia property of TMOMCNs was measured, during which the PBS suspension of TMOMCNs was placed in an alternating magnetic field (230 kHz, 290 Oe), and the

heating curve was recorded with time, as shown in Fig. 10. With the increase of heating time, the temperature of PBS suspension increases gradually up to 45 °C in 23 min, showing apparent hyperthermia effects. SAR is used to evaluate the ability of a sample to absorb energy from alternating magnetic field per unit mass, which reflects the heating efficiency of the samples. According to Eq. (2), the SAR of TMOMCNs is 58.47 W/g. As SAR is significantly affected by the size and the coating of magnetic nanoparticles, it should be normalized to a frequency of 1 MHz and a magnetic field strength of 100 Oe for comparison with literatures [19]. The normalized value of TMOMCNs is 30.23 W/g. As reported in the literature [26], an SAR of iron oxide nanoparticles in Fig. 9. Magnetization curves of MOMCNs and TMOMCNs (a) and the photograph of magnetic separation for TMOMCNs (b).

28

Materials Science & Engineering C 84 (2018) 21–31

L. Chen et al.

Fig. 11. The adsorption kinetic curves (a) and saturation adsorption curves of DOX (b) by MOMCNs and TMOMCNs at 25 °C.

Table 2 Release parameters of TMOMCNs for DOX. pH

Temperature (°C)

Release time (hours)

Release diffusional exponent (n)

Release rate constant (k × 103)

Correlation coefficient (R2)

5.5 6.2 7.4 7.4

37 37 37 45

20.0 6.0 6.0 20.0

0.258 0.421 0.334 0.180

29.75 17.92 16.48 22.26

0.9081 0.9297 0.9298 0.8496

3.5. DOX loading and release property 3.5.1. DOX loading property As a comparison, the effects of time and DOX concentration on the DOX loading capacity of MOMCNs and TMOMCNs were investigated (as shown in Fig. 11). It can be seen in Fig. 11(a) that the loading equilibrium time of MOMCNs is 12 h with a loading capacity of 60.85 mg/g, while the adsorption equilibrium time of TMOMCNs extends to 24 h with a loading capacity of 57.56 mg/g. This difference in adsorption ability is caused by the grafting of thermo-sensitive polymer around MOMCNs, raising the resistance for DOX transfer and prolonging the adsorption equilibrium time. It can be observed from Fig. 11(b) that the loading capacity of MOMCNs and TMOMCNs increases with increasing equilibrium concentration of DOX. This can be attributed to the increasing driving force of the concentration gradient because the increase in DOX concentration can accelerate the diffusion of DOX molecules onto the drug carrier. The loading capacity of both MOMCNs and TMOMCNs tends to equilibrate when DOX concentration is 500 mg/L, while the equilibrium loading capacity of MOMCNs (284.71 mg/g) is 149.61 mg/g, higher than that of TMOMCNs (135.10 mg/g). Considering that SBET of TMOMCNs is much lower than that of MOMCNs, it can be speculated that the lower loading capacity is attributed to the occupation and block in mesoporous structures of grafting PNIPAM.

Fig. 12. The time-dependence of drug release from TMOMCNs at 37 °C of different pH.

3.5.2. Controlled release property In order to investigate the controlled release property of TMOMCNs, the release tests were performed of pH 5.5, 6.4 and 7.2 separately. The drug loading complex were prepared at DOX concentration of 500 mg/L and loading time of 24 h, with a loading capacity of 135.10 mg/g. Fig. 12 shows the time-dependence of drug release of different pH. It can be seen that the cumulative release increases with the decrease of pH. The cumulative release of TMOMCNs in 20 h increases from 31.26% (pH = 7.4) to 65.29% (pH = 5.5). It illustrates the release behavior of TMOMCNs could be controlled by pH. This pH-responsivity is originated from DOX structures. In weak base environment, DOX is desalinated and becomes hydrophobic, closely integrated with the

Fig. 13. The time-dependence of drug release from TMOMCNs at pH = 7.4.

10–40 W/g could satisfy the requirements of hyperthermia. The higher the SAR is, the higher the heating efficiency is. It illustrates that TMOMCNs have potential in hyperthermia.

29

Materials Science & Engineering C 84 (2018) 21–31

L. Chen et al.

Table 3 The drug-loading capacity, release time and drug release of different drug carriers for DOX. Drug carriers

Drug-loading capacity (mg/g)

Releasing time (hours)

Cumulative release (%)

Release dose (mg)

References

OMCNs Magnetic hollow carbon nanospheres (MnOx-HMCNs) Mesoporous carbon coated CoFe2O4 (CoFe2O4@mC) CoFe2O4 porous hollow microspheres Fe3O4-DOX/pSiO2-PEG core/shell nanospheres TMOMCNs

425.0 375.0 301.2 118.1 16.3 135.1

8/20 8/20 8/20 8/20 8/20 8/20

14.1/− 31.3/39.7 7.6/16.2 65.2/68.2 15.0/18.5 46.4/65.3

59.9/− 117.4/148.9 22.9/48.8 77.0/80.5 2.4/3.0 62.7/88.2

[6] [27] [28] [29] [30] This work

obtained in this work are more suitable for application in magnetic targeted DOX carrier. In summary, by integrating good biocompatibility, DOX loading, controlled release and hyperthermia properties, the thermo-sensitively and magnetically ordered mesoporous carbon nanoshpheres show great potential in cancer chemotherapy with high effect and low toxicity. A serial of direct experiments on cellular and physiological systems will be carried out to confirm their feasibility for practice.

nanocarriers; while under acidic conditions, owing to the protonation of amino group, the hydrophilicity of DOX is improved, thus facilitating the release from the nanocarriers [15]. The thermo-sensitivie release performance was investigated from the release of DOX from TMOMCNs at 37 and 45 °C separately (as shown in Fig. 13). When the temperature is raised from 37 to 45 °C, the cumulative release in 20 h increases from 31.26% to 36.68%, showing temperature-controlled release behavior. This phenomenon is resulted from two aspects: first, when the temperature is higher than the LCST of TMOMCNs (39.5 °C), volume phase transition takes place and the thermo-sensitive polymer shrinks, squeezing DOX out of the polymer network to release; second, owing to the concentration difference, DOX molecules stored inside mesoporous pore gradually diffuse outside of the carrier. Therefore, the release of DOX from TMOMCNs could be tuned by temperature. To reveal of the mechanism of DOX release from the nanocarriers, the Rigter-Peppas model (Eq. (9)) was used to evaluate the relationship of cumulative release Q and release time [6]:

Q = k × tn

4. Conclusion In order to obtain high efficacy and low toxic effect of chemotherapy, a thermo-sensitive and magnetical DOX delivery system based on ordered mesoporous carbon nanoshpheres was developed for DOX loading, controlled release and hyperthermia. OMCNs were prepared by soft template method in combination with hydrothermal treatment, in which α-Fe2O3 nanoparticles were used as catalyst. MOMCNs obtained by impregnation were taken as substrate and NIPAM as thermo-sensitive functional monomer to synthesize TMOMCNs by free radical polymerization. The DOX loading, controlled release and hyperthermia results show that the preparation time of uniform OMCNs was reduced by 8 h in the presence of 7.8 × 10− 4 mol/L α-Fe2O3 nanoparticles. The obtained TMOMCNs have an average size of ~146 nm with a LCST of 39.5 °C, Ms of 10.15 emu/g and normalized SAR of 30.23 W/g, demonstrating apparent thermo-sensitivity and hyperthermia effects. The DOX loading capacity of TMOMCNs is 135.10 mg/g and the release performance can be regulated by pH and temperature: when pH decreases from 7.4 to 5.5, the cumulative release increases from 31.26% to 65.29%; when temperature increases from 37 to 45 °C, the cumulative release increases from 31.26% to 36.68%, and the release is governed by diffusion mechanism. This work provides new possibility for development of thermo-sensitively and magnetically ordered mesoporous carbon nanoshpheres in the low-toxicity and high-efficiency chemotherapy treatment by targeted controlled release of DOX and hyperthermia.

(9)

where Q is the cumulative release (%); k stands for the diffusion rate constant; t is release time (h); n is diffusion index used to explain release mechanism. When n ≤ 0.5, the mechanism of drug release follows the Fickian model, which is diffusional release; when n = 1, it is zero-order release; when 0.5 < n < 1.0, irregular release mechanism dominates [19]. The releases of DOX from TMOMCNs before release equilibrium at different pH and temperature were fitted by Rigter-Peppas model. The results are listed in Table 2. It can be seen that all n values are < 0.5, indicating that drug release is controlled by free diffusion mechanism. Meanwhile, the diffusion rate constant increases with decreasing pH and increasing temperature, which further confirms the tunable release of DOX by pH and temperature. In recent years, several DOX nanocarriers based on mesoporous nanospheres have been investigated and reported. Table 3 compares the loading capacity and release of TMOMCNs for DOX against other mesoporous nanocarriers previously reported in literature. Note that the multifunctional OMCNs or porous carbon drug delivery system have a loading capacity for DOX ranging in 16.3–425.0 mg/g, while TMOMCNs obtained in this work have a DOX loading capacity of 135.1 mg/g, which is at medium level. The cumulative release of different mesoporous carriers for DOX is further compared in the release time of 8 and 20 h, respectively, as shown in Table 3. The release dose represents the DOX release by unit mass of drug carrier, which is equal to the cumulative release multiply loading capacity. According to Table 3, although OMCNs have triple higher drug-loading capacity than TMOMCNs, the release dose in 8 h is almost the same for the two carriers, which manifests TMOMCNs have better release properties. By comparing the release dose of CoFe2O4 porous hollow microspheres and TMOMCNs in 8 and 20 h, it could be found that with the extension of release time the release dose of CoFe2O4 porous hollow microspheres has little increase, while that of TMOMCNs increases from 62.7 to 88.2 mg, showing good controlled release property Furthermore, MnOx-HMCNs and CoFe2O4@mC possess good drug loading and release properties, while the toxicity of Co and Mn is significantly higher than that of Fe3O4 [31], so the TMOMCNs

Acknowledgments This work was supported by the National Natural Science Foundation of China (U1610255 and U1607120), Shanxi Provincial Key Innovative Research Team in Science and Technology (2015013002-10 and 201605D131045-10), Shanxi Provincial Key Research and Development Program (201603D111010 and 201703D321015-1), Higher School Science and Technology Innovation Project of Shanxi (2017114) and Foundation of Taiyuan University of Technology (2015QN101). References [1] E.A. Levine, J.H. Stewart, P. Shen, et al., Intraperitoneal chemotherapy for peritoneal surface malignancy: experience with 1,000 patients, J. Am. Coll. Surg. 218 (4) (2014) 573–585. [2] C.H. Trebelhorn, J.C. Dennis, S.R. Pondugula, et al., Plant-based omega-3 stearidonic acid enhances antitumor activity of doxorubicin in human prostate cancer cell lines, J. Cancer Res. Ther. 2 (2014) 132–143. [3] T. Zhou, X. Zhou, D. Xing, Controlled release of doxorubicin from graphene oxide

30

Materials Science & Engineering C 84 (2018) 21–31

L. Chen et al.

based charge-reversal nanocarrier, Biomaterials 35 (13) (2014) 4185–4194. [4] D. Saha, K.E. Warren, A.K. Naskar, Soft-templated mesoporous carbons as potential materials for oral drug delivery, Carbon 71 (2014) 47–57. [5] L. Zhou, K. Dong, Z. Chen, et al., Near-infrared absorbing mesoporous carbon nanoparticle as an intelligent drug carrier for dual-triggered synergistic cancer therapy, Carbon 82 (2015) 479–488. [6] G. Xu, S. Liu, H. Niu, et al., Functionalized mesoporous carbon nanoparticles for targeted chemo-photothermal therapy of cancer cells under near-infrared irradiation, RSC Adv. 4 (64) (2014) 33986–33997. [7] H. Wang, X. Li, Z. Ma, et al., Hydrophilic mesoporous carbon nanospheres with high drug-loading efficiency for doxorubicin delivery and cancer therapy, Int. J. Nanomedicine 11 (2016) 1793–1806. [8] V.P. Torchilin, Multifunctional, stimuli-sensitive nanoparticulate systems for drug delivery, Nat. Rev. Drug Discov. 13 (11) (2014) 813–827. [9] M.Y. Hua, H.W. Yang, H.L. Liu, et al., Superhigh-magnetization nanocarrier as a doxorubicin delivery platform for magnetic targeting therapy, Biomaterials 32 (34) (2011) 8999–9010. [10] M. Zhou, S. Liu, Y. Jiang, et al., Doxorubicin-loaded single wall nanotube thermosensitive hydrogel for gastric cancer chemo-photothermal therapy, Adv. Funct. Mater. 25 (29) (2015) 4730–4739. [11] X. Wei, Q. Luo, L. Sun, et al., Enzyme- and pH-sensitive branched polymer–doxorubicin conjugate-based nanoscale drug delivery system for cancer therapy, ACS Appl. Mater. Interfaces 8 (18) (2016) 11765. [12] B. Sahoo, K.S.P. Devi, R. Banerjee, et al., Thermal and pH responsive polymertethered multifunctional magnetic nanoparticles for targeted delivery of anticancer drug, ACS Appl. Mater. Interfaces 5 (9) (2013) 3884–3893. [13] Q. Zhang, P. Wang, Y. Ling, et al., Single molecular wells-daws on-like heterometallic cluster for the in situ functionalization of ordered mesoporous carbon: AT1- and T2-weighted dual-mode magnetic resonance imaging agent and drug delivery system, Adv. Funct. Mater. (2017), http://dx.doi.org/10.1002/adfm. 201605313. [14] J. Liu, N.P. Wickramaratne, S.Z. Qiao, et al., Molecular-based design and emerging applications of nanoporous carbon spheres, Nat. Mater. 14 (8) (2015) 763–774. [15] J. Zhu, L. Liao, X. Bian, et al., pH-controlled delivery of doxorubicin to cancer cells, based on small mesoporous carbon nanospheres, Small 8 (17) (2012) 2715–2720. [16] H. Zhang, L. Chen, L. Li, et al., Magnetic porous carbon microspheres synthesized by simultaneous activation and magnetization for removing methylene blue, J. Porous. Mater. 24 (2) (2017) 341–353. [17] Y. Fang, D. Gu, Y. Zou, et al., A low-concentration hydrothermal synthesis of biocompatible ordered mesoporous carbon nanospheres with tunable and uniform size, Angew. Chem. Int. Ed. 49 (43) (2010) 7987–7991. [18] Z. Wu, W. Li, P.A. Webley, et al., General and controllable synthesis of novel

[19]

[20]

[21]

[22] [23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

31

mesoporous magnetic iron oxide@carbon encapsulates for efficient arsenic removal, Adv. Mater. 24 (4) (2012) 485–491. L. Chen, L. Li, H. Zhang, et al., Magnetic thermosensitive core/shell microspheres: synthesis, characterization and performance in hyperthermia and drug delivery, RSC Adv. 4 (87) (2014) 46806–46812. X. Cui, M. Antonietti, S.H. Yu, Structural effects of iron oxide nanoparticles and iron ions on the hydrothermal carbonization of starch and rice carbohydrates, Small 2 (6) (2006) 756–759. Y. Tsuboi, M. Nishino, T. Sasaki, et al., Poly (N-isopropylacrylamide) microparticles produced by radiation pressure of a focused laser beam: a structural analysis by confocal Raman microspectroscopy combined with a laser-trapping technique, J. Phys. Chem. B 109 (15) (2005) 7033–7039. W. Li, D. Zhao, An overview of the synthesis of ordered mesoporous materials, Chem. Commun. 49 (10) (2013) 943–946. Y. Du, W. Liu, R. Qiang, et al., Shell thickness-dependent microwave absorption of core–shell Fe3O4@C composites, ACS Appl. Mater. Interfaces 6 (15) (2014) 12997–13006. J. Rubio-Retama, N.E. Zafeiropoulos, C. Serafinelli, et al., Synthesis and characterization of thermosensitive PNIPAM microgels covered with superparamagnetic γ-Fe2O3 nanoparticle, Langmuir 23 (20) (2007) 10280–10285. B. Mehdaoui, A. Meffre, J. Carrey, et al., Optimal size of nanoparticles for magnetic hyperthermia: a combined theoretical and experimental study, Adv. Funct. Mater. 21 (23) (2011) 4573–4581. J.P. Fortin, C. Wilhelm, J. Servais, et al., Size-sorted anionic iron oxide nanomagnets as colloidal mediators for magnetic hyperthermia, J. Am. Chem. Soc. 129 (9) (2007) 2628–2635. S. Zhang, X. Qian, L. Zhang, et al., Composition-property relationships in multifunctional hollow mesoporous carbon nanosystems for pH-responsive magnetic resonance imaging and on-demand drug release, Nano 7 (17) (2015) 7632–7643. S. Mohapatra, S.R. Rout, R.K. Das, et al., Highly hydrophilic luminescent magnetic mesoporous carbon nanospheres for controlled release of anticancer drug and multimodal imaging, Langmuir 32 (6) (2016) 1611–1620. B. Cai, M. Zhao, Y. Ma, et al., Bioinspired formation of 3D hierarchical CoFe2O4 porous microspheres for magnetic-controlled drug release, ACS Appl. Mater. Interfaces 7 (2) (2015) 1327–1333. F.H. Chen, L.M. Zhang, Q.T. Chen, et al., Synthesis of a novel magnetic drug delivery system composed of doxorubicin-conjugated Fe3O4 nanoparticle cores and a PEG-functionalized porous silica shell, Chem. Commun. 46 (45) (2010) 8633–8635. T. Titma, R. Shimmo, J. Siigur, et al., Toxicity of antimony, copper, cobalt, manganese, titanium and zinc oxide nanoparticles for the alveolar and intestinal epithelial barrier cells in vitro, Cytotechnology 68 (6) (2016) 2363–2368.