Multifunctional nanocarriers of Fe₃O₄@PLA-PEG/curcumin for MRI, magnetic hyperthermia and drug delivery

Nanoparticles (NPs) are finding an increasing number and variety of applications in biomedicine, including detection, imaging and therapeutic treatment [1–7]. In general, these NPs consist of a solid, inorganic core encapsulated by a shell, normally of soft matter, and frequently including functional molecules of a drug or targeting agent [2].

Among the core materials, magnetic nanoparticles (MNPs) are leading candidates for biomedical applications, owing to their superparamagnetic properties [7–14]. For instance, MNPs can experience a force in an applied gradient magnetic field in bioseparation or drug delivery [15]. When localized in a biological environment, MNPs can create inhomogeneities of the local magnetic field, thus influencing the magnetic relaxation rate of water-containing tissues. That is why they have been intensively investigated and are now used as agents for MRI contrast enhancement in clinical diagnostics [16–21]. Probably, one of the most attractive properties of MNPs is the so-called magnetic fluid hyperthermia (MFH), which is based on their capability to adsorb energy from an alternating magnetic field (AMF) to generate heat via their Néel and/or Brownian relaxation dissipation [11,22]. When inserted into tumor tissues, MNPs can kill cancer cells by increasing the local temperature, because cancer cells show stronger thermal cytotoxicity than normal cells [23]. Therefore, MNPs have been intensively studied for their use as potential MFH agents in cancer treatment [24–26]. Among the MNPs, Fe₃O₄ MNPs have been the most intensively studied thanks to their suitable magnetic properties, biodegradability, biocompatibility and ability to be functionalized and drug encapsulated [27].

Functionalization of MNPs is an important step; with a biocompatible coating layer, not only do the MNPs become stabilized and protected from further oxidation but in many cases this process can also impact the magnetic properties of the core [28–31]. Several studies on iron oxides have shown that due to the interaction of the coating material with the NP surface, the magnetization might be decreased or increased depending on the ligands used, as well as the encapsulation procedure [28,30]. Regarding the coating material, beyond silica and carbon, most of the materials that have been studied are various types of polymers [10,14,32–36]. Recent studies have shown that using copolymer is very effective for encapsulating MNPs to disperse them in water [32,33,35]. A hydrophobic monomer such as polylactic acid (PLA) can anchor well on iron oxide NPs, whereas a hydrophilic monomer such as PEG can exhibit a good dispersibility in water. To additionally make the magnetic nanocarrier capable of delivering chemical therapy, several molecular drugs, including folic acid, doxorubicin, dopamine and curcumin have been loaded onto MNPs [37–42]. Among these drugs, curcumin can exhibit great pharmacological potential, including anti-inflammatory, antioxidant, antiproliferative and chemosensitizing, cell cycle arrest potential, and can display apoptotic potential against various cancer cells such as colorectal, breast, pancreatic and colon cancers, therefore renders it great potential in cancer prevention and therapy. Notably, the IC₅₀ of curcumin in healthy cells is significantly higher than that in cancer cells [43]. It was used in this study as a model for drug delivery and release of nanocarriers.

The potential of using an Fe₃O₄-based nanocarrier as both a drug release and hyperthermia agent has been reported in several works [40,41]. MRI and MFH are important on their own [44–47] or are sometimes found in combination [48]. However, few reports have shown experimentally the fabrication of a nanocarrier simultaneously demonstrating the optimal performance of MRI, drug release and MFH. In 2011, although doxorubicin-loaded γ-Fe₂O₃ NPs were evaluated for their ability in diagnosis and hyperthermia, no in vivo experiment was carried out [49].

Saturation magnetization (M_s) is well known to be one of the most important factors in all the biomedical applications of MNPs. The magnetic force, F, in manipulation application is defined as linearly proportional to M_s [15]. The relaxation rate (R₂) in MRI is approximately proportional to (M_s×V_NPs)², where V_NPs is the volume of the NPs [17,18]. The specific absorption rate (SAR) is dependent on magnetization [50] and magnetic anisotropy [51–53]. Therefore, M_s should be carefully evaluated for the applications of MRI and MFH. In addition to M_s of the MNPs, the field strength (H) and frequency (f) of the magnetic field used should also be considered in in vivo MFH experiments. There are some criteria for H×f product for in vivo MFH. For example, the Atkinson–Brezovich criterion requires the H×f product to be lower than 5 × 10⁸ A/(m·s) [54]. However, according to the Dutz–Hergt criterion, this number could be ten-times larger to obtain safety for in vivo experiments [55]. Most authors used a higher field than the Atkinson–Brezovich criterion, while some reports even used a much higher field than the Dutz–Hergt criterion in their studies [56].

In this study, the authors aimed to fabricate functionalized magnetic Fe₃O₄ NPs using amphiphilic PLA-PEG copolymer as a biodegradable coating material. A thorough chemical and physical characterization, including magnetization, magnetic heating performance and nuclear relaxation rate, was carried out. In addition, in vivo MRI and MFH tumor treatment experiments were carried out to show the effectiveness of the nanocarriers. Utilization of the NPs as the carriers for curcumin, a thermally stable anticancer drug [57] with control release ability, is presented to show the true multifunctional potential of the fabricated biomedical nanocarrier. The novelty of this study arises from the trifunctions of the MNPs and the lower field used in the in vivo MFH-mediated drug-release experiments than those in other studies using Fe₃O₄-based NPs.

Materials & methods

Materials

Lactic acid, PEG 2000 and tin(II) 2-ethylhexanoate (Sn(Oct)₂) were purchased from Sigma (MO, USA); FeCl₃, FeCl₂·4H₂O, NH₄OH, C₇H₈, dichloromethane (DCM), CH₃OH and C₂H₅OH and phosphate-buffered saline (PBS; pH 7.4) were purchased from Merck (Darmstadt, Germany); curcumin was provided by PharmaGenica Healthcare Pvt Ltd (Bengaluru, India). Distilled water was used for all experiments.

Synthesis & characterization of chemical, structural & basic magnetic properties

PLA-PEG copolymer was synthesized by ring-opening polymerization reaction between lactic acid and PEG in the presence of catalyst Sn(Oct)₂ [58]. PLA-PEG copolymer (molecular weight of 10 kDa) was dissolved in DCM (1 mg/ml) and stirred for 24 h; then H₂O was added to form a mixture, which was stirred for another 24 h to disperse PLA-PEG copolymer from DCM to H₂O. DCM solvent was then evaporated to obtain a dispersion of PLA-PEG copolymer in H₂O (1 mg/ml).

Fe₃O₄ NPs were synthesized by the coprecipitation method (based on previous work) [59]. Complex Fe₃O₄@PLA-PEG NPs were fabricated from suspensions of readily fabricated Fe₃O₄ and PLA-PEG NPs through the procedure described in Thong et al. [60]. In brief, 20 ml of Fe₃O₄ NPs dispersed in water at the concentration of 3 mg/ml was slowly dropped into 20 ml of 0.3 mg/ml PLA-PEG copolymer solution and stirred for 24 h to obtain Fe₃O₄@PLA-PEG NPs (suspension A). The suspension was stored at room temperature (22–32°C) for more than 90 days and no aggregation was observed by eye. To get a more concentrated suspension, suspension A was introduced to a rotary vacuum evaporator. To prepare Fe₃O₄@PLA-PEG/Cur, 1 ml of 4 mg/ml curcumin dissolved in C₂H₅OH was slowly dropped into 10 ml of suspension A and stirred for 48 h, and then ethanol was allowed to evaporate to obtain Fe₃O₄@PLA-PEG/Cur aqueous suspensions.

Crystal structure was determined by x-ray diffraction (XRD) technique in D8 Advance Bruker equipment. The morphology, size and size distribution of NPs were characterized by a Hitachi S-4800 field emission–scanning electron microscopy (Hitachi S-4800) and transmission electron microscopy (TEM; JEOL JEM 1010), while the hydrodynamic diameter of the nanocarriers was determined from dynamic light scattering measurements using a Nano-Zetasizer (Malvern, Malvern, UK). Chemical bonding of the components in the NPs was verified by Shimadzu Prestige-21 Fourier-transform infrared spectroscopy. The mass contribution of PLA-PEG and curcumin on NPs was determined by thermogravimetric analysis (TGA) with a Shimadzu DTG-60H. Hysteresis loops at room temperature up to 880 kA/m were measured by a homemade vibrating sample magnetometer.

MRI characterization

The longitudinal and transverse relaxation rates, R₁ = 1/T₁ and R₂ = 1/T₂, respectively, were measured on a 7 T Bruker Biospec using a quadrature 300 MHz, 30 mm mouse coil (Animal Imaging Research, LLC, MA, USA). More details about the MRI measurement can be found in the Electronic Supplementary Information (ESI).

The longitudinal and transverse relaxivities, r₁ and r₂, of the nanocarriers were determined, using Equation 1, as a proportional factor from the linear dependencies of the corresponding relaxation rates on iron concentration, C:

where R_1,2⁰ is the relaxation rate of the liquid [30].

MFH measurements

Experiments to determine the MFH effect of the nanosystems were carried out on an RDO induction generator (RDO Induction, NJ, USA) of 5 kW at field amplitudes H ≤8000 A/m and frequency f = 178 kHz. This equipment was also used for the controlled drug-release experiment.

The heating curves of the ferrofluids were measured to determine the SAR, according to Equation 2:

where C_p is the specific heat of the liquid (4.186 J/g°C for water), m_l and m_NPs are the corresponding mass of the liquid and nanocarrier and dT/dt is the initial rate of temperature increase [61].

Temperature change measurements were performed for ferrofluid samples at the concentration of 1.63 mg_Fe/ml at the volume of 1.0 ml. Distilled water was used as the only suspension solvent for the measurements.

In vitro curcumin release

For the passive curcumin release procedure, 10 mg of Fe₃O₄@PLA-PEG/Cur MNPs were dispersed in 30 ml PBS. The dispersion was incubated in a water bath at 37°C for 48 h. At desired time intervals (4 h for the first 12 h and 12 h for the remaining time), a 2 ml sample was withdrawn and replaced with an equal volume of the fresh release medium (PBS, pH 7.4). The curcumin concentration in each taken sample was determined by UV-VIS spectroscopy.

For the first controlled release experiment series by MFH, five samples, each containing 1 ml aliquot of Fe₃O₄@PLA-PEG/Cur aqueous solution, were excited by the same AMF at a frequency f = 178 kHz, H = 5040 A/m, starting from the same temperature of 25°C, and ending after a certain period (5, 10, 15, 20 and 25 min). Another series of experiments were set up with the AMF stimulation at different magnetic fields, at a frequency f = 178 kHz, and amplitudes varying from 3600 to 5040 A/m until the sample temperature reached 37°C or 45°C. After MFH treatment, the concentration of curcumin released by AFM stimulation from the samples was determined by UV-Vis spectroscopy.

Animal experiments

In vivo MRI experiments

MRI monitoring experiments were conducted in sarcoma-bearing Swiss mice. Ten days after tumor incubation, three mice with similar tumor sizes and one normal mouse (with no tumor) were selected for MRI investigation. The mice were numbered as follows:

• Mouse A: normal mouse as biological control;

• Mouse B: noninjected mouse with tumor, cancer control;

• Mouse C: mouse with tumor, injected with 10 μl of 5 mg/ml Fe₃O₄@PLA-PEG ferrofluid directly into the tumor;

• Mouse D: mouse with tumor, injected with 50 μl of 5 mg/ml Fe₃O₄@PLA-PEG ferrofluid directly into the tumor.

Then all the mice were anesthetized with 65 μl of 0.25 mg/ml thiopental (injected through a tail vein). The mice were then taken for imaging in a Philips Gyroscan 3.0 T MRI (Philips, Eindhoven, The Netherlands) at 0, 15, 30, 45 and 60 min after injecting the magnetic fluid into the tumor. Image acquisition was performed in the spin echo mode with a repetition time of 3000 ms and an echo time of 80 ms.

Similar experiments were conducted with other four mice with 5-day-old tumors (among which, one mouse was the biological control) in order to investigate the ability of early tumor diagnosis based on MRI.

In vivo tumor treatment with MFH

The effectiveness of the fabricated nanocarriers in hyperthermia was investigated via the capability of thermal therapy of sarcoma 180, a heterogeneous tumor that is easily directly injected. The animal experiments were performed on mice using Fe₃O₄@PLA-PEG ferrofluid of 5 mg/ml concentration of Fe₃O₄ in water at tumor volume-dependent doses.

Two series of experiments were carried out on Swiss mice with solid tumors of sarcoma 180 were incubated subcutaneously at the stomach site for 5 days and 10 days before MFH experiments. Each experimental series consisted of four groups:

• Group 1: 6 noninjected and nonirradiated mice as a control;

• Group 2: 6 noninjected mice with AMF irradiation;

• Group 3: 6 mice injected with Fe₃O₄@PLA-PEG and without AMF irradiation;

• Group 4: 6 mice injected with Fe₃O₄@PLA-PEG and subjected to AMF.

The treatment efficiency was described by observation of tumor volume V, which was calculated using Equation 3:

where a is tumor length and b is tumor width [62].

A dose of 500 μg ferrofluid/cm³ tumor was injected directly into the tumor site. A field of 5040 A/m and 178 kHz was applied for 40 min, starting at 30 min after the injection of Fe₃O₄@PLA-PEG/Cur NPs into the mice (similar to the authors' previous protocol [42]). Each therapy consisted of three sessions of AMF treatment, with a 48 h break in between.

Results

Characterization

A proposed structure of the NPs is shown in Supplementary Figure 1 in ESI. Detailed characterizations of Fe₃O₄@PLA-PEG and Fe₃O₄@PLA-PEG/Cur were carried out in a previous work [60]. Herein, the authors present only the crucial properties of the systems. First, XRD patterns of all three samples (Fe₃O₄, Fe₃O₄@PLA-PEG and Fe₃O₄@PLA-PEG/Cur) showed six clearly defined diffraction peaks, very similar, independently of the sample, which correspond to the typical (200), (311), (400), (422), (511) and (440) diffraction peaks of spinel structure (Supplementary Figure 2 in ESI). This suggests that the formation of multicompositional NPs, Fe₃O₄@PLA-PEG and Fe₃O₄@PLA-PEG/Cur, does not affect the crystal structure of Fe₃O₄. This may result in the conservation of magnetic properties of the system, which is discussed in the next section. The average size of the core nanocrystal, d_c, estimated from the XRD profile via the Scherrer formula, was ~8.5 nm (Supplementary Table 1 in ESI), which is in good agreement with the size obtained by TEM (8.5 ± 2.6 nm) (Figure 1 & Supplementary Figure 3A in ESI). Field emission–scanning electron microscopy images (Supplementary Figure 3B in ESI) show the morphology of Fe₃O₄, Fe₃O₄@PLA-PEG and Fe₃O₄@PLA-PEG/Cur samples. NPs can be seen as mostly spherical. The Fe₃O₄ MNPs functionalized by PLA-PEG and PLA-PEG/Cur are fairly uniform in size, with an average diameter, d_NPs, of about 10–12 nm, which is a few nanometers larger than the size of the bare Fe₃O₄ MNPs. The hydrodynamic sizes (d_h) of Fe₃O₄, Fe₃O₄@PLA-PEG and Fe₃O₄@PLA-PEG/Cur suspensions measured in water are shown in Supplementary Figure 4 (ESI). Fe₃O₄ NPs present an average hydrodynamic size of 37 nm, which is larger than the size obtained by TEM measurement. The average hydrodynamic size of Fe₃O₄@PLA-PEG rises to 45 nm, while this number of Fe₃O₄@PLA-PEG/Cur NPs is 52 nm. Dynamic light scattering data with narrow size distributions (Supplementary Figure 4 in ESI) and the zeta potential (Supplementary Table 1 in ESI) of the samples with values from -21.2 mV to -40.9 mV show that the samples were well dispersed and stable.

The Fe₃O₄@PLA-PEG and Fe₃O₄@PLA-PEG/Cur NPs were analyzed by TGA (Figure 1D & E) in order to estimate experimentally the mass contribution of nonmagnetic coating materials (i.e., PLA-PEG and curcumin) in the samples. The first weight-loss step in both diagrams around 100°C corresponds to the evaporation of water present in the samples. The next weight-loss step at above 300°C along with the exothermic peak in the differential thermal analysis curve is assigned to the decomposition of the organic components. The mass percentages of the organic components calculated from the initial mass of the precursors (i.e., Fe₃O₄, PLA-PEG and curcumin) of Fe₃O₄@PLA-PEG and Fe₃O₄@PLA-PEG/Cur were 9.1% and 25%, respectively. The values determined by TGA were 13.6 and 24.9%, which were quite compatible to the amount of the original ligands. The higher weight-loss percentage in Fe₃O₄@PLA-PEG/Cur compared with Fe₃O₄@PLA-PEG proves once again the presence of curcumin in Fe₃O₄@PLA-PEG/Cur. The curcumin content in Fe₃O₄@PLA-PEG/Cur NPs was then calculated to be ∼11%. The data also indicate that the mass percentages determined by TGA are in good agreement with those obtained from the initial composition.

Magnetization characteristics

Supplementary Figure 5A (ESI) presents the field dependent magnetisation (M-H) loops recorded at 300 K for the three samples – Fe₃O₄, Fe₃O₄@PLA-PEG and Fe₃O₄@PLA-PEG/Cur – measured for the total mass of the sample (i.e., core + shell). The influence of the shell coating and curcumin loading on the magnetic properties was studied based on changes in the magnetization (M) measured at 880 kA/m. For this, the magnetization of Fe₃O₄ cores (M_c) in Fe₃O₄@PLA-PEG and Fe₃O₄@PLA-PEG/Cur NPs was calculated considering the TGA percentage of Fe₃O₄ mass in the samples to be 74.7 and 70.4 Am²/kg, respectively. The slight decrease in M_c of Fe₃O₄@PLA-PEG/Cur NPs compared with that of Fe₃O₄@PLA-PEG NPs can result from the oxidation of the core in the process of curcumin loading onto the NPs.

The normalized M-H loops, using M_c, are presented in Supplementary Figure 5B (ESI). As clearly shown, the M_c value of the Fe₃O₄ core, after functionalization by PLA-PEG and loading curcumin, was ~10% larger compared to non coated one, which could be explained by the protection of the core from oxidation.

It is also worth noting in Supplementary Figure 5B that the hysteresis loops (i.e., M-H loops) present a very similar shape, with nearly zero coercivity and remanence, supposing a superparamagnetic behavior at room temperature. To confirm the superparamagnetic behavior of these samples, the M-H loops in Supplementary Figure 5B were fitted with a standard Langevin expression presented in Equation 4.

where D is the diameter of the NPs, f(D) corresponds to a log-normal size distribution and L(x) = cotan(x) - 1/x. The second term corresponds to a small paramagnetic contribution due to the atoms on the surface, χ_PM being the paramagnetic susceptibility. As can be seen in Supplementary Figure 5C (ESI), the fittings are really good, and the estimated average sizes are 9.1, 9.6 and 10.0 nm for Fe₃O₄, Fe₃O₄@PLA-PEG and Fe₃O₄@PLA-PEG/Cur samples, respectively. According to TEM and XRD, these values are close to the estimated core size, ∼9 nm. The agreement confirms the superparamagnetic behavior of the MNPs.

High relaxivity r₂

MRI was performed with samples of varying concentrations of Fe₃O₄@PLA-PEG and Fe₃O₄@PLA-PEG/Cur nanocarriers. Figure 2A & B shows the spin echo images of Fe₃O₄@PLA-PEG and Fe₃O₄@PLA-PEG/Cur, respectively, while Figure 2C & D presents correspondingly the longitudinal and transverse relaxation rates, R₁ and R₂, of the two nanocarriers at various concentrations. The white central samples in Figure 2A & B were control samples of distilled water with no NPs. The following sequence images from white to black in clockwise order around the central samples correspond to the samples with concentrations 15, 30, 45, 60, 75 and 90 μg Fe/ml, respectively. There is a clear change in MRI contrast when increasing the sample concentration. Although the MR images became darker with increasing concentration in both samples, a higher MRI contrast enhancement was achieved for the Fe₃O₄@PLA-PEG sample. Similar results were also reported in some recent studies [47,63].

To get quantitative information and to gain more insight into the effect of curcumin loading, ParaVision 5.1 software was used to calculate relaxation rate values from additional measurements of R₁ and R₂ for the Fe₃O₄@PLA-PEG and Fe₃O₄@PLA-PEG/Cur specimens in the concentration range 0–1.6 mM of Fe. As seen in the graphs (Figure 2C & D), linear dependencies are observed in general fitted lines with a high correlation factor (R² ≥0.98) in the whole range of measured concentrations – that is, up to 1.6 mM (90 μg/ml Fe) for the longitudinal relaxation and up to 1.0 mM (56 μg/ml Fe) for the transverse relaxation. It can be observed that the relaxation rate for the Fe₃O₄@PLA-PEG/Cur sample keeps a good linear relationship with concentration for both longitudinal and transverse regimes in the whole dilution range. For the Fe₃O₄@PLA-PEG sample, however, there is an evident change in the evolution of the R₂ relaxation rate with increasing concentration, and the linear behavior is apparently lost after 1 mM. This observation suggests that the presence of curcumin in the core of Fe₃O₄@PLA-PEG/Cur may help reduce the attraction between the MNPs and make the curcumin-loaded conjugates better dispersed than the nonloaded ones at high concentrations, which is important for the final biomedical application.

The obtained relaxivity values, r₁ and r₂ (calculated in the linear concentration region), are collected in Table 1.

MFH studies

The heating efficiency of Fe₃O₄@PLA-PEG and Fe₃O₄@PLA-PEG/Cur NPs after dispersion in water at a concentration of 1.63 mg_Fe/ml was studied using calorimetric methods. The temperature versus time curves were recorded for different AMF (Supplementary Figure 6 in ESI). As depicted, the heating rate rapidly increases with increasing field, and the therapeutic window for hyperthermia therapy, 40–45°C, can be easily reached within around 10 min by modifying the amplitude of the AMF during the treatment. To get a better insight into the hyperthermic effect of the MNPs, the authors calculated the heating efficiency or SAR (Supplementary Table 2 in ESI). The obtained SAR values were calculated for the Fe₃O₄@PLA-PEG/Cur samples measured at a frequency of 178 kHz and Fe concentration of 1.63 mg.ml^-1. The obtained SAR values follow an H² dependence (Supplementary Figure 7 in ESI), as expected according to the linear response theory [66]. Although the H×f values for the experiments were slightly higher than the Atkinson–Brezovich limit [54], many other studies also used fields with higher H×f values [67–69]. The calculation of the intrinsic loss power (ILP), SAR/(f×H²), gave a value of 11.5 nHm²kg^-1 (H = 3600 A/m; f = 178 kHz) for Fe₃O₄@PLA-PEG/Cur sample, which is appreciably greater than those commercial iron oxide-based NPs reported in the literature (e.g., Resovist presents an ILP of 1.5 nHm²kg^-1 [61]). The PLA-PEG and Cur might contribute to improvement of the SAR and ILP value of the NPs. A recent study has reported that SAR and ILP were influenced by size and the structure of iron oxide nanoflowers, which depended on the polyacrylic acid concentration during the synthesis [70].

Effective release by remote AMF-mediated control

The first experimental series for curcumin release efficiency was performed in five samples after applying the same AMF field of 5040 A/m for 5, 10, 15, 20 and 25 min. The release percentages of all analyzed samples are depicted in Figure 3. It is clear that longer exposure time, simultaneously with the higher temperature reached in the solution, resulted in a higher curcumin release. After 20 min under the applied AMF, the curcumin amount loaded in the Fe₃O₄@PLA-PEG/Cur was almost fully released and reached plateau.

On the other hand, the passive curcumin release in an incubated bath at 37°C, as shown in Figure 3, reached a saturation release level of about 80%, only after 24 h. Isothermal AMF release rates for the Fe₃O₄@PLA-PEG/Cur ferrofluid were estimated by varying the field and time for the solution to reach 37°C and 45°C. The experimental conditions are described in Table 2.

The UV-VIS spectra and calculated curcumin released amounts are depicted in Supplementary Figure 8 (ESI). As shown in Table 2, the stronger the magnetic field, the shorter the time required to reach the temperature, and the curcumin released amount depends on both the temperature and the duration to reach to that temperature. This allows the drug release to be controlled by adjusting the magnetic field strength and/or AMF exposure time.

From the inductive heating curve in Supplementary Figure 9 (ESI) & Table 2, the authors could estimate the initial release rate as equal to 40%/h and 51%/h at T = 37°C and 45°C, respectively, which are in agreement with the values of 18%/h and 26%/h observed by Purushotham and Ramanujan [40]. The application of AMF made the release faster and more efficient compared to passive release. The faster rate accompanied by higher percentage release that was observed for the Fe₃O₄@PLA-PEG/Cur nanocarrier as compared with other reports [40,71] might originate from different functionalizing materials and/or drug molecules, namely thermosensitive polymer and doxorubicin [40], PLA and pluronic F127 [72] and alginate [71].

In vivo MRI experiments

MR images were taken for control mouse A and tumor-bearing mice B, C and D. For mice C and D, 10 and 50 μl of Fe₃O₄@PLA-PEG magnetic fluid were directly injected into their tumors. Figure 4A shows the images of the tumor region in the mice right after injection. It seems that no organs or tumors could be identified in these images. As shown in Figure 4B, for mice C and D, there was a distinct dark area, different from the control mouse. The tumors in mice B, C and D can be identified, and a better contrast in the tumor area and the tumor's rim can be seen as well as the size of the tumor (the red arrow) for mice C and D, injected with Fe₃O₄@PLA-PEG magnetic fluid.

Figure 5A presents the obtained MR images for mice C and D at different time periods after injection. As depicted, all tumor positions could be clearly seen at all times and MNPs were observed dispersing in the tumor over time. From 30 to 45 min after injection, the MNPs were evenly distributed and gave the highest contrast for MRI. For observation of smaller tumors, 5-day-old tumors were scanned with an MRI e-scanner in a similar procedure compared with the adult tumors. Figure 5B shows that, with mice C and D, there was a marked dark area, different from the cancer control mouse – mouse B. The contrast in the tumor area and the tumor's rim as well as the size of the tumor (the red circle) were easily observed in mice C and D.

Treatment of the tumor by MFH with Fe₃O₄@PLA-PEG

Figure 6 shows the tumor image and size variations during the treatment for the experimental mice, 5 and 10 days after tumor implantation. For the treated mice with 5-day-old cancer, the tumor volume decreased strongly, so that 50% of them exhibited almost a disappearance of the tumor after three MFH treatments. The effect could be seen right after the first MFH treatment when the tumor surface clearly shrank with the appearance of black eschar covering the wound. After treatment time, the eschar fell off by itself and left a scar behind where fur could grow back normally. For the treated mice with 10-day-old cancer, although the tumor did not disappear totally after treatment, its development was seriously inhibited. The growth after 3 treatments of 5-day-old tumor was 600%, which is equivalent to 37–38% compared with the tumor growth in the control group (1600%; Figure 6B). For the 10-day-old tumor, the treatment effectiveness was lower but still clearly evident as compared with untreated mice (Figure 6).

Observation of regrowth of the tumor after the treatment was performed starting at day 1 after the last MFH therapy. The graphs obtained for the two categories of cancer mice are presented in Supplementary Figure 10 (ESI). For the mice with cancer that was 5-day-old, the tumor continued to decrease for 4–5 days after the last MFH treatment and remained unchanged for the following 2 days after that. Eight days after the treatment, the tumor seemed to develop but at such a slow rate that at day 14 the tumor was almost of the size observed just after the treatment (Supplementary Figure 10A, ESI). For the mice with 10-day-old cancer, Supplementary Figure 10B (ESI) depicts that 4–5 days after the last MFH therapy, the tumor volume slightly decreased, and it started to grow again after the day 5 post-treatment, so that on day 14 it reached 150% growth rate as compared with the size after the last treatment.

Discussion

PLA-PEG is an amphiphilic copolymer in which PEG hydrophilic chains tend to extend outside, while PLA hydrophobic chains arrange inside the micelles. The NPs could be encapsulated via the PEG segment [73] and curcumin is loaded into the NP carrier (Supplementary Figure 1 in ESI). For Fe₃O₄@PLA-PEG/Cur NPs, curcumin molecules might penetrate into the copolymer micelles' center with PLA hydrophobic segments (as observed in the authors' previous Cur/PLA-PEG nanosystems [74] and other reports [58,75,76]). The increase in the hydrodynamic size of the Fe₃O₄ MNPs compared with their size obtained by TEM can be explained by the adsorption of OH^- ions and H₂O molecules through hydrogen bonding in the aqueous solution onto the surface of the Fe₃O₄ MNPs. The hydrodynamic size of Fe₃O₄@PLA-PEG rises due to the water expansion of the hydrophilic PEG component in the outermost shell of the particles. A similar observation has been reported [77,78]. The curcumin present in the core of Fe₃O₄@PLA-PEG/Cur may lead to an increase in the size of the core, and therefore the overall size of NPs when in water (Supplementary Table 1 in ESI).

The ~10% increase in the net magnetization of the NPs after coating and drug loading shows an important role of surface functionalization of NPs. A similar increase in the net magnetization of the NPs has been reported in the literature for different kinds of coatings [79–81] and attributed to an enhancement of the surface magnetic ordering of the MNPs due to the coating [29]. In addition, the obtained M_c^TGA values become closer to the saturation magnetization of bulk Fe₃O₄ (∼92 Am²/kg), which, as explained before, is especially important for biomedical applications. High magnetization could result in good MRI and MFH effects.

In the in vitro MRI experiments, the r_1,2 values of the Fe₃O₄@PLA-PEG/Cur samples being smaller than those of Fe₃O₄@PLA-PEG can be explained by the reduction in magnetization upon curcumin encapsulation from 74.7 and 70.4 Am²/kg (Supplementary Figure S5 in ESI), which in turn decrease the transverse relaxivity of the MNPs [3]. It is worth noting that the change in the r₂/r₁ ratio for the case of Fe₃O₄@PLA-PEG/Cur (Table 1) can be related to the effect of curcumin adsorption on the Fe₃O₄ NP surface, which resulted in a hydrophobicity change of the whole coating layer [82].

Notably, r₂ values of both Fe₃O₄@PLA-PEG and Fe₃O₄@PLA-PEG/Cur samples are significantly higher than those obtained for commercial Resovist, ferumoxytol or other Fe₃O₄-based systems [8,30,33,83]. The high value of the r₂/r₁ ratio also indicates that the T₂ contrast mechanism is dominant for the investigated nanocarriers. The in vivo MRI results reconfirm that the MNPs enhance the contrast in MRI observation. The authors performed the in vivo MRI experiment in the 5-day-old tumor to determine if the MNPs could help distinguish the tumor from its surroundings at its early stage. From Figure 5B, one can note that the MNPs allow a more accurate assessment of the size, shape and part of the internal structure of the tumor at the early stage of cancer as well as an easier assessment of the presence or absence of a tumor when it appears inside the body. The MRI study on a 5-day-old tumor can also illustrate the MRI contrast enhancement for tumor observation, even at its early stage. In Figure 5A & B, a 10 μl dose of the NPs shows better differentiation of the rim and the inside area of the tumor than a 50 μl dose does. Therefore, a 10 μl dose (1.8 mg Fe/kg body weight) is better for MRI contrast enhancement. Some other studies used higher doses for MRI in vivo experiments (5 mg/kg [36,84]), and Chen et al. used an even higher dose of 30 mg/kg [85].

As indicated from the heating curves measured at f = 178 kHz and H_AC = 5040 A/m (Supplementary Figure 8 in ESI), the Fe₃O₄@PLA-PEG and Fe₃O₄@PLA-PEG/Cur ferrofluids reached the temperature range for cancer hyperthermia treatment after a few hundredths of seconds. That makes them both potential candidates for low-field hyperthermia therapy, combined with heat-activated drug-release chemotherapy. The AMF-mediated drug-release profile clearly indicates that the heat generation of the MNPs activated by AMF serves as a driving force to release curcumin from the hydrophobic core of the system in a much faster and more effective way as compared with passive release. Although this phenomenon has also been reported by other researchers, most of the research involved thermosensitive components in the NPs [40,86–88]. The low passive rate achieves nearly zero losses of curcumin at the beginning, and then high active rate release, within a few minutes, allows a remotely controlled targeted delivering.

The biodistribution of NPs depends on many factors such as chemical composition, size, shape and surface modifications [89]. In a previous report, by ex vivo MFH experiments and atomic absorption spectroscopy analysis, the authors of the current study found that when intravenously injected into mice, the amounts of Fe₃O₄@PLA-PEG MNPs are distributed in organs in the following order: liver > lung > spleen > tumor > kidney [90]. More important, a dose that is sufficient for hyperthermia of the tumor would overheat the liver. Therefore, in this study, to evaluate the cancer treatment ability of Fe₃O₄@PLA-PEG MNPs by MFH, the MNPs were directly injected into the mice's tumors. The product H×f of the magnetic field used in these in vivo experiments (8.97 × 10⁸ A/(ms)), although slightly higher than that in the Atkinson–Brezovich criteria (5 × 10⁸) [10], is lower than that in the Dutz–Hergt criteria (5 × 10⁹) [55] is also much lower than that of many published Fe₃O₄-based NPs [56]. Compared with 19 nm cubic Fe₃O₄ NPs at the higher H×f = 1.21 × 10⁹ A/[ms], Fe₃O₄@PLA-PEG MNPs show a similar effect of tumor growth inhibition [86].

The observed efficiency of MFH therapy is also in agreement with several other reports on similar treatments such as via direct injection of MNPs into breast tumors [91] or brain tumors [92].

Conclusion

We have developed a multifunctional magnetic nanosystem with a core-shell structure consisting of ~9 nm Fe₃O₄ MNPs surrounded by amphiphilic copolymer PLA-PEG coating (Fe₃O₄@PLA-PEG) that allows us to reduce the coating mass. Besides, the coating layer helps improve the magnetic properties of the nanosystem by increasing the saturation magnetization of the NPs (∼10%). The high relaxivity ratio (r₂/r₁ = 357.6) indicates that Fe₃O₄@PLA-PEG MNPs can effectively work as T₂ contrast agents for MRI. In addition, the high heating efficiency obtained under AMF (SAR = 82.2 W/g at H = 5–40 A/m; f = 178 kHz) reveals that Fe₃O₄@PLA-PEG MNPs are also very promising for magnetic hyperthermia. In addition, these nanocarriers can be loaded with anticancer drug curcumin (Fe₃O₄@PLA-PEG/Cur) and release the drug under the application of an external AMF in a fast, facile and efficient way (nearly 100% release after 20 min). In vivo studies with mice bearing sarcoma 180 cancer reveal that the mice injected with Fe₃O₄@PLA-PEG and treated with AMF showed clear tumor growth inhibition lasting several days after the hyperthermia treatment. Finally, in vivo MRI experiments clearly showed an improvement in the contrast in the tumor area for those mice injected with Fe₃O₄@PLA-PEG MNPs. Therefore, our study indicates that Fe₃O₄@PLA-PEG/Cur is a promising multifunctional nanosystem, providing both efficient cancer detection (MRI) and treatment (drug delivery and magnetic hyperthermia) capacities.

Future perspective

The combination effect of chemotherapy using curcumin or other cancer drugs with magnetic hyperthermia should be conducted. The heat dose, the heat map of magnetic hyperthermia are vital for the translation of this modality of treatment to clinics and hence will be studied.