Water-dispersible polyphosphate-grafted Fe₃O₄ nanomagnets for cancer therapy

Abstract

We report the development of a new class of water-dispersible polyphosphate-grafted Fe₃O₄ nanomagnets (PPNMs) by a facile soft chemical approach. The grafting of polyphosphate with Fe₃O₄ nanoparticles is evident from Fourier transform infrared (FTIR) spectroscopy, thermogravimetric analysis (TGA), dynamic light scattering (DLS) and zeta-potential measurements. X-ray diffraction (XRD) and transmission electron microscopy (TEM) analyses reveal the formation of highly crystalline Fe₃O₄ nanoparticles with an average size of about 10 nm. These nanoparticles show good colloidal stability, strong magnetic field responsivity and protein resistance characteristics. Induction heating studies confirm localized heating of these superparamagnetic PPNMs with good intrinsic loss power under an AC magnetic field (AMF). The drug loading and release behavior of the PPNMs was explored using doxorubicin hydrochloride (DOX) as a model drug. The decrease in fluorescence intensity and increase in surface charge of drug-loaded PPNMs strongly suggest the conjugation of DOX with the PPNMs. The cell viability and hemolysis assays suggest that the PPNMs do not have adverse toxic effects for further in vivo use. Specifically, high loading affinity for DOX with sustained release, substantial cellular internalization and self-heating capacity makes these novel magnetic nanoparticles suitable for drug delivery and hyperthermia therapy applications.

---

Introduction

Over the last few decades, significant attention has been focused on magnetic nanoparticles (MNPs) as potential tools for various biomedical applications such as drug delivery, magnetic hyperthermia and magnetic resonance imaging (MRI).^1–5 In particular, Fe₃O₄ nanoparticles have received extensive attention due to their unique properties, such as superparamagnetism and low toxicity. Practical in vivo usage of these particles requires a size below 100 nm as well as their colloidal stability and biocompatibility in the physiological medium. However, bare Fe₃O₄ nanoparticles prepared by co-precipitation, even though they have an average size of 10 nm, are polydisperse and tend to aggregate and adsorb proteins quickly in water or physiological media.⁶ In addition, for most biomedical applications the significant challenge is to avoid unwanted uptake of MNPs by the reticuloendothelial system (RES). In order to overcome the above drawbacks, Fe₃O₄ nanoparticles are often surface engineered by coating with various biocompatible, stimuli-responsive organic and inorganic functional molecules.^7–14 These coating molecules not only stabilize nanoparticles but also provide terminal functional groups that can be utilized for attaching targeting moieties as well as drugs. Further, the use of such stimuli-responsive nanoparticles in cancer therapy can increase drug accumulation at target sites, decrease toxicity and avoid under- or overdosing.^15,16

In general, the stability of the bonding between functional molecules and Fe₃O₄ nanoparticles is crucial for biomedical applications. The main drawback of surface-engineered MNPs currently in use is that the polymer coatings are not covalently bound to the MNP surface and will easily be detached. Moreover, the polymer coatings further cause a significant increase in the overall hydrodynamic diameter of the particles and thereby affect their biodistribution as well as clearance. Thus, further functionalization with targeting ligands increases the complexity. The polymer shell around the MNP core influences the magnetic relaxivity dramatically.¹⁷ Therefore, the covalent attachment of nonpolymeric layers that would improve the stability of MNPs under physiological conditions without affecting particle size drastically or undermining magnetic properties (e.g., relaxivity and saturation magnetization) while still providing sites for attaching targeting ligands (e.g., antibodies) would be useful. Thus, there is an emergent interest in developing biocompatible magnetic nanocarriers with suitable surface functionality for biomedical applications.

Recently organosilane, carboxylate and phosphonate anchored layers grafted on MNPs have proven to be better anchoring groups without adversely affecting the properties of the MNPs. Yee et al. reported the binding of alkane phosphonic acids onto the surface of ferric oxide particles either by one or both oxygen atoms of the phosphonate group.¹⁸ Sahoo et al. demonstrated the conjugation of alkyl phosphonates and phosphate to Fe₃O₄ nanoparticles via the formation of P–O–Fe bonding.¹¹ Recently, our research group reported the preparation of water-dispersible and biocompatible Fe₃O₄ nanoparticles using sodium hexametaphosphate (a cyclic phosphate molecule) as a surface passivating agent.¹⁹ The formation of these water-dispersible Fe₃O₄ nanoparticles has been attributed to the presence of bioactive phosphate molecules on their surface. This work further propelled us to investigate the use of linear phosphate molecules in the stabilization of Fe₃O₄ nanoparticles. Other research groups also reported the strong interaction between the inorganic core and the phosphonic moiety, and the most interesting moiety seems to be triphosphate, a type of polyphosphate.^11,18–21 Food and Drug Administration (FDA)-approved sodium tripolyphosphate (STPP) is widely used as a preservative for food and also as a polyanion cross-linker in polysaccharide-based drug delivery systems.²² Chitosan/tripolyphosphate nanoparticles (CS/TPP) have been used as an alternative to chitosan to encapsulate peptides, proteins, pDNA and siRNA.^23–25 In a review article, Rao et al. have demonstrated the essential roles of polyphosphates in the virulence of major diseases such as dysentery, tuberculosis, and anthrax, as well as in apoptosis, in the proliferative aspects of cancer, in osteoporosis, and in aging.²⁶ These polyphosphates are hydrolyzed into simpler phosphates, which in moderate amounts are nutritious. Being a linear molecule, tripolyphosphate has a higher affinity for metal ions. It binds strongly to metal cations as both a bidentate and tridentate chelating agent. Even though STPP is extensively used in the preparation and stabilization of chitosan nanoparticles, its use as a coating material in the preparation of aqueous-stable Fe₃O₄ nanoparticles is hardly reported.

Herein we report the preparation of polyphosphate-grafted Fe₃O₄ nanomagnets (PPNMs) for drug delivery and hyperthermia treatment of cancer. It has been observed that the use of tripolyphosphate as a surface passivating agent provides excellent biocompatibility and colloidal stability to particles. PPNMs showed good magnetic field responsivity and excellent self-heating efficacy under an AC magnetic field (AMF), which is essential for their hyperthermia application. Further, the high affinity of these nanoparticles towards positively charged DOX and their pH-triggered release and substantial cellular internalization makes them suitable for drug delivery. This study may lead to the design of novel combinatorial therapy in which the synergy between nanoparticle hyperthermia and chemotherapeutic agents can provide effective treatment with minimal side effects.

Experimental

Materials

Ferrous chloride tetrahydrate (FeCl₂·4H₂O), ferric chloride hexahydrate (FeCl₃·6H₂O), 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT), doxorubicin hydrochloride (DOX) and bovine serum albumin (BSA) were purchased from Sigma-Aldrich. Dulbecco’s modified Eagle medium (DMEM) and fetal calf serum (FCS) were obtained from Invitrogen, USA and Himedia Laboratories, India, respectively. Ammonia (25%) and dimethyl sulfoxide (DMSO) were procured from Thomas Baker Chemical Pvt. Ltd, India and SD Fine Chemicals, India, respectively. Sodium tripolyphosphate was purchased from SRL Pvt. Ltd, India. All chemicals are of analytical grade and used without further purification. The acetate buffer (AB) pH 4 and 5, and phosphate-buffered saline (PBS)-pH 7.4 were prepared using standard protocols.

Synthesis of polyphosphate-grafted Fe₃O₄ nanomagnets

Polyphosphate-grafted Fe₃O₄ nanomagnets (PPNMs) were prepared by one pot co-precipitation of Fe²⁺ and Fe³⁺ ions in basic medium. In brief, FeCl₂·4H₂O (1.988 g) and FeCl₃·6H₂O (5.406 g) were dissolved in 80 ml of water in a round-bottomed flask and the temperature was slowly increased to 70 °C in a nitrogen atmosphere under constant mechanical stirring at 1000 rpm. The temperature was maintained at 70 °C for 30 min and then 30 ml of ammonia solution (25%) was added instantly to the above reaction mixture, which was kept under stirring for another 30 min at the same temperature. Then, a 10 ml aqueous solution (0.2 g ml⁻¹) of sodium tripolyphosphate was added, the temperature was slowly raised to 90 °C, and the mixture was reacted for 60 min under stirring. The obtained precipitates were then thoroughly rinsed with water and separated from the supernatant solution using a permanent magnet. For comparative purposes, bare Fe₃O₄ MNPs were prepared in a similar method without using polyphosphate as a coating material.

Characterization

X-ray diffraction (XRD) patterns were recorded on a Philips powder diffractometer PW3040/60 with Cu Kα radiation. The crystallite sizes were estimated from the X-ray line broadening using the Scherrer formula. The infrared spectra were recorded in the range 4000–400 cm⁻¹ on a Fourier Transform Infrared Spectrometer (FTIR, Bomem, MB series). The transmission electron micrographs were taken with a Philips CM 200 TEM. Thermal analysis of samples was performed under an argon atmosphere with a scanning rate of 10 °C min⁻¹ using TGA, Setaram Instrumentation. DLS measurement was performed using a Malvern 4800 Autosizer employing a 7132 digital correlator for the determination of hydrodynamic diameters. The zeta potential measurements were carried out on a Zetasizer nano series, Malvern Instruments. A colloidal stability assay was carried out by measuring the absorbance at a wavelength of 350 nm for different time intervals using a JASCO V-650 UV-visible spectrophotometer. The field-dependent magnetization and zero field cooled-field cooled (ZFC-FC, at an applied field of 100 Oe) measurements were carried out with a Physical Property Measurement System (PPMS, Quantum Design). The concentration of phosphorus in the samples was measured using an inductively coupled plasma optical emission spectrometer (ICP-OES, Activa, Horiba Jobin Yvon).

The heating ability of the PPNM suspension was obtained from the time-dependent calorimetric measurements using an induction heating unit (Easy Heat 8310, Ambrell). 1 ml of an aqueous suspension PPNMs of desired concentration was taken in an Eppendorf tube with suitable arrangements to minimize the heat loss. AMFs of 0.251, 0.293 and 0.335 kOe at a fixed frequency of 265 kHz were used to evaluate the specific absorption rate (SAR). The SAR was calculated using the following equation:²⁷

where C is the specific heat of the solvent (C = C_water = 4.18 J g⁻¹ °C), ΔT/Δt is the initial slope of the time-dependent temperature curve and m_Fe is the mass fraction of Fe in the sample. The rise in temperature was also monitored using a high resolution infrared camera (Thermal Imager Testo 875-1), and analyzed by thermography software (Testo IR Soft Software, version 3.1). The concentration of Fe in the PPNM suspension was obtained using a phenanthroline spectrophotometric method.²⁸

The anticancer agent DOX was used as a model drug to estimate the drug loading and release behavior of the PPNMs. In order to investigate the interaction of DOX with PPNMs we performed zeta potential and fluorescence spectroscopic studies. Aqueous dispersions of different quantities of PPNMs (0, 20, 40, 60, 80, 100, 120 and 140 μg) were added to 1 ml of a DOX solution (10 μg ml⁻¹), and this was mixed thoroughly by shaking at room temperature for 15 min. The fluorescence spectra of the supernatant (obtained after magnetic sedimentation of DOX-loaded PPNMs) were then recorded using a Hitachi F 2500 fluorescence spectrophotometer. The fluorescence intensities of the supernatants (washed drug molecules were also taken into consideration for calculations) against that of a pure DOX solution were used to determine the loading efficiency. The encapsulation efficiency (w/w%) was calculated using the following relation:

where I_DOX is the fluorescence intensity of the pure DOX solution, I_S is the fluorescence intensity of the supernatant and I_w is the fluorescence intensity of washed DOX (physically adsorbed DOX molecules). The drug loading content was determined as follows:

For the release study, we have quantified the amount of DOX–PPNM according to the loading efficiency. The loading was carried out, by incubating 0.5 ml of an aqueous solution of DOX (1 mg ml⁻¹) with 1 ml of the aqueous suspension (pH ∼ 7.5) of PPNMs (5 mg ml⁻¹) for 1 h in the dark (however, no decrease in fluorescence intensities was observed after 15 min of interaction). Drug-loaded samples were separated from the free-standing drug molecules through magnetic separation and carefully washed with Milli-Q water. The pH-triggered drug release studies were carried out under reservoir–sink conditions. The drug-loaded PPNMs (5 mg) were immersed into 5 ml of the respective release medium (AB-pH 4, AB-pH 5) and then put into a dialysis bag. The dialysis was performed against 200 ml of PBS-pH 7.4 under continuous stirring at 37 °C to mimic the cellular environment. 1 ml of the external medium was withdrawn and replaced with fresh PBS at fixed time intervals to maintain the sink conditions. The amount of DOX released was determined by measuring the fluorescence emission at 585 nm (excitation wavelength: 490 nm) using a plate reader (Infinite M1000, Tecan-I control, Switzerland) against the standard plot prepared under similar conditions. Each experiment was performed in triplicate and the standard deviation is given in the plot.

The biocompatibility of the PPNMs to mouse skin fibrosarcoma (WEHI-164) cell line was investigated by MTT assay. Cells (0.25 × 10⁶) were seeded overnight in Petri dishes (P-60) containing 4 ml Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal calf serum (FCS) and antibiotics (100 U ml⁻¹ penicillin and 100 μg ml⁻¹ streptomycin) in a humidified atmosphere of 5% CO₂ at 37 °C. Then different concentrations of PPNMs were added to the cells and these were incubated for another 48 h in culture conditions. Then, the media containing PPNMs was carefully removed and the cells were further incubated with 0.5 ml of MTT solution (0.5 mg ml⁻¹) in culture conditions for 2 h. The supernatant was aspirated and 1 ml of DMSO was added to each culture dish to solubilize the MTT crystals. The crystals were thoroughly dissolved and further diluted with DMSO (1 : 10). 200 μl of the above solution from P-60 culture dishes was transferred to 96-well plates and the absorbance was measured in a microplate reader (Tecan infinite 200 PRO, Switzerland) at 544 nm. The cell viability was calculated by comparing the absorption of treated cells to that of the control, which was defined as 100%.

The cellular uptake of DOX-loaded PPNMs was studied by confocal laser scanning microscopy (CLSM) using the WEHI-164 cell line. For CLSM imaging, cells (0.5 × 10⁶) were seeded on glass coverslips and cultured overnight. The cells were then treated with DOX–PPNMs (0.5 μM DOX) for 3 h under culture conditions, followed by washing with PBS. The cells were mounted on a glass slide in cell mounting medium (Invitrogen, USA) containing DAPI for nuclear staining. These cells were then imaged by CLSM (LS510 Meta, Carl Ziess, Germany). The excitation source used was an Ar ion laser (488 nm for DOX and 364 nm for DAPI) and the emission window was set at 575–615 nm and 430–480 nm for DOX and DAPI, respectively.

Protein–particle interaction and hemolysis assay studies were performed to evaluate the protein resistance characteristics (with BSA) and hemocompatibility (with human whole blood) of PPNMs respectively, as reported elsewhere.^2,27

Results and discussion

Fig. 1 shows (a) XRD patterns of bare MNPs and PPNMs, and (b) a TEM image of PPNMs (inset shows its HRTEM image). The XRD pattern shows the formation of a single-phase cubic inverse spinel Fe₃O₄ structure with lattice constant a = 8.378 Å, which is very close to the reported value for magnetite (JCPDS Card no. 88-0315, a = 8.375 Å). The presence of sharp and intense diffraction peaks confirms the formation of highly crystalline particles. The average crystallite size was found to be around 10 nm from X-ray line broadening. These sizes are comparable to the results obtained from TEM. The TEM image of PPNMs clearly shows the formation of almost spherical nanoparticles of about 10 nm in size. It can be clearly seen that these PPNMs have a good distribution, with a lower percentage of agglomeration with respect to bare MNPs (the TEM image of bare MNPs is shown in Fig. S1, ESI†). From the HRTEM image of the PPNMs (inset of Fig. 1b), the average lattice spacing (d_hkl) was measured to be ∼0.30 nm, which corresponds to the (220) plane of inverse spinel Fe₃O₄.²⁹ Furthermore, the selected area electron diffraction (SAED) patterns of bare MNPs and PPNMs can be indexed to highly crystalline reflections such as (220), (311), (400), (422), (511) and (440) of the cubic inverse spinel Fe₃O₄ structure, which is consistent with the XRD results (a typical SAED pattern of PPNMs is shown in Fig. S2, ESI†).

Fig. 1 (a) XRD patterns of bare MNPs and PPNMs, and (b) a TEM image of PPNMs (inset of (b) shows the HRTEM image of PPNMs).

FTIR spectroscopy is a useful tool to identify the functional groups present on the surface of nanoparticles. In the present case, it is applied to find out whether there are any vibrational bands owing to the presence of phosphate groups on the surface of the MNPs. FTIR spectra of STPP and PPNMs along with their peak assignments are shown in Fig. 2. It has been observed that the IR bands for STPP are well-resolved, while those of the PPNMs are rather broad and less intense. In the spectrum of STPP, the characteristic bands attributed to phosphate vibrations appear at 1215 cm⁻¹ (P=O stretching), 1160 and 1130 cm⁻¹ (asymmetric and symmetric stretching vibrations in the PO₂ group), 1078 cm⁻¹ (asymmetric and symmetric stretching vibrations in the PO₃ group), and 918 and 715 cm⁻¹ (asymmetric and symmetric stretching of the P–O–P bridge).³⁰ These bands were absent in the spectrum of bare Fe₃O₄ MNPs (Fig. S3, ESI†), whereas they appear in the spectrum of PPNMs with a slight shifting of band position in the range of 700–1250 cm⁻¹, with bands at 710, 900, 1026, 1087 and 1167 cm⁻¹. Further, the intense peak at around 575 cm⁻¹ in the PPNMs can be attributed to the Fe–O stretching vibrational mode of Fe₃O₄.²⁷

Fig. 2 FTIR spectra of STPP and PPNMs with their peak assignments (inset shows the expanded FTIR spectrum of PPNMs in the range of 400–1500 cm⁻¹).

Fig. 3 shows (a) TGA-DTA and (b) pH-dependent zeta potential plots of PPNMs. The TGA analysis of PPNMs showed a total weight loss of 5.46% up to 400 °C (Fig. 3a), whereas that of bare Fe₃O₄ MNPs was 2.5% (Fig. S4, ESI†). The higher weight loss for the PPNMs further confirmed the presence of organic molecules on the surface of the Fe₃O₄ nanoparticles. Furthermore, the TGA plot of the PPNMs revealed a three-step thermal decomposition profile. An initial weight loss of about 1.82% with an endothermic peak at ∼110 °C is associated with the removal of residual water and physically absorbed hydroxyl and STPP from the surface of the MNPs. The second and third weight loss steps (3.64%) were observed due to decomposition of polyphosphate moieties present on the surface. These two stages of weight loss, with endothermic shoulders at about 180 and 275 °C, may be attributed to either a bilayer coating of STPP or chemical conjugation of phosphate molecules onto the surface of the Fe₃O₄ particles by two different forms. The bilayer coating of STPP can be ruled out as the introduction of second layer of phosphate polyanions onto the negatively charged PPNMs is not electrostatically favorable. Thus, it is assumed that polyphosphate molecules of STPP were conjugated onto the surface of the Fe₃O₄ nanoparticles by two different forms of chemical bonding between the oxygen atoms of polyphosphate and Fe₃O₄.¹⁸ In one case, the polyphosphate moiety is bonded through its three oxygen atoms, resulting in a stronger bridging geometry (tridentate) and a higher desorption temperature. In the second case, only two oxygen atoms participate in the bonding (bidentate), resulting in weaker bonding and a lower desorption temperature. From ICP-OES analysis, the quantity of STTP molecules present on the surface of the PPNMs was found to be 126 mg g⁻¹ of particles.

Fig. 3 (a) TGA-DTA and (b) pH-dependent zeta potential plots of PPNMs (the inset of (b) shows a possible schematic representation for the chemical conjugation of STPP onto the surface of Fe₃O₄ nanoparticles).

After the successful functionalization of polyphosphate moieties onto the surface of nanomagnets, it is important to check their colloidal stability for practical applications. The colloidal stability of the PPNMs was assessed from the changes in light scattering intensity as well as extinction changes with time. DLS measurement indicates that these particles generate an aqueous colloidal dispersion with a mean hydrodynamic diameter of 43 nm (Fig. S5, ESI†). The higher value of hydrodynamic diameter observed by DLS as compared to TEM arises from the presence of hydrated organic moieties and the inherent polydispersity in the distribution.^31,32 However, the light scattering intensity and polydispersity index hardly vary with time, revealing their good aqueous colloidal stability. Further, the insignificant change in the absorbance of the PPNM suspension in aqueous media even up to 72 h indicates the good colloidal stability of the nanomagnets (Fig. S6, ESI†). These nanoparticles are hydrophilic in nature due to the hydrogen bonding association of surface functional groups and water.

The electrostatic contribution to the stabilization of the particles is evident from zeta potential measurements. The variations in the zeta potentials of PPNM suspensions with varying pH values (0.05 mg ml⁻¹) are shown in Fig. 3b. From zeta potential measurements, no isoelectric point (pH of zero-point charge) was observed for PPNMs in the pH range of 2–12, whereas that of bare Fe₃O₄ was 6.67 (Fig. S7, ESI†). Moreover, the high negative surface charge of PPNMs (at pH > 4) makes them colloidally stable in the aqueous phase. The high negative charge on these particles arises from the grafting of phosphate groups on MNPs. In addition, the electrostatic repulsive force originating from the ionization of the phosphate groups provides additional stability to the particles. Furthermore, the high negative zeta potential of PPNMs in physiological medium (−30 mV) could decrease the possibility of their combination with hemoglobin, which would play a significant role in improving their stability and blood compatibility. Based on the TGA and zeta potential measurements, a possible schematic representation for chemical conjugation of STPP onto the surface of Fe₃O₄ nanoparticles is shown in the inset of Fig. 3b.

In order to assess the potential of PPNMs for targeted drug delivery and hyperthermia, further studies were performed on their magnetic and thermo-magnetic properties. The field-dependent magnetization plots of PPNMs at 5 and 300 K are shown in Fig. 4. These PPNMs exhibit superparamagnetic behavior without magnetic hysteresis and remanence at 300 K, whereas ferrimagnetic behavior, with a coercivity of about 80 Oe and a remanence of 11.8 emu g⁻¹, is observed at 5 K. The appearance of hysteresis at lower temperature points to the magnetic ordering of the sample at lower temperature. This transition from superparamagnetic behavior to ferro/ferrimagnetic behavior below a particular temperature, i.e. the blocking temperature (T_B), is usually observed in MNPs.³³ The ZFC-FC plot (top inset of Fig. 4) shows that the blocking temperature (T_B) of the PPNMs is 75 K at an applied field of 400 Oe. The magnetization values of PPNMs were found to be 58.2 and 52.2 emu g⁻¹, at 20 kOe for 5 and 300 K, respectively. Further, bare Fe₃O₄ prepared by a similar route gave a magnetization of 67.6 emu g⁻¹ at 300 K (Fig. S8, ESI†). The magnetization of PPNMs is reduced by ∼22% as compared to the bare Fe₃O₄, although TGA showed only about 5.46% weight loss. The magnetization of bare MNPs could possibly arise from the clustering behavior of the uncapped MNPs, which will contribute to the increase in magnetization due to exchange coupling and dipolar interaction among the surface ions.³⁴ However, in the case of PPNMs the presence of bulkier non-magnetic polyphosphate groups on the surface will suppress the contribution from exchange coupling as well as dipolar interactions among particles. However, the retention of superparamagnetic properties at room temperature with good magnetic field responsivity (bottom inset of Fig. 4) makes these nanoparticles suitable for hyperthermia and drug delivery applications.

Fig. 4 Field-dependent magnetization plots of PPNMs at 5 and 300 K (top inset shows ZFC-FC plots of PPNMs in an applied field of 400 Oe, and bottom inset shows photographs of the PPNM suspension in the presence and absence of a permanent magnet of field strength ∼2.5 kOe).

The effect of magnetic field strength and particle concentration on the heating ability of PPNMs were evaluated. Fig. 5 shows the temperature vs. time plots of (a) 2.3 mg ml⁻¹ of Fe at different applied fields and (b) different concentrations of Fe at an applied field of 0.335 kOe. The temperature vs. time plots of the PPNM suspension show a time-dependent gradual increase in temperature under applied AMF. It has been observed that a magnetic field of 0.251 kOe at a fixed frequency of 265 kHz is enough to raise the temperature of the magnetic suspension of 2.3 mg ml⁻¹ to 42–43 °C (hyperthermia temperature) within 20 min. At this temperature, various cellular damaging mechanisms such as apoptosis, protein denaturation and DNA cross-linking may occur to destroy the cancer cells.^35,36 Further, the required hyperthermia temperature was achieved much faster with increasing field strength, which is apparent as the heat generation/dissipation (P) is proportional to the square of applied AC magnetic field (inset of Fig. 4a) as follows:²⁷

where μ₀ is the permeability of free space, χ₀ is the magnetic susceptibility, H is the magnetic field amplitude, f is the frequency and τ_eff is the effective relaxation time. In heating of superparamagnetic Fe₃O₄ nanoparticles under AMF, a rise in temperature is mainly associated with the combined effect of Néel and Brownian relaxation losses.^27,37 The Néel and Brownian relaxation losses are associated with the magnetic moment rotations within the particles and with the entire particles, respectively. The relaxation times are given by the following equations:

τ_N = τ₀e^KVM/k_BT

where τ_B is the Brownian relaxation time, τ_N is the Néel relaxation time, τ₀ ≈ 10⁻⁹ s, K is the anisotropy constant, V_M is the volume of the Fe₃O₄ nanoparticles, k_B is Boltzmann’s constant, T is temperature, η is viscosity and R_H is the hydrodynamic particle radius. Since the aqueous suspension of Fe₃O₄ nanoparticles is semiconducting in nature with high resistivity (ρ = 10² Ohm cm), heat generation under AMF will be negligible from eddy current loss.³⁸ Further, heat generation due to hysteresis loss will be negligible for superparamagnetic nanoparticles (as the particle size is less than the critical diameter/single domain).³⁸

Fig. 5 Temperature vs. time plots of (a) 2.3 mg ml⁻¹ of Fe at different applied fields (inset shows the linear relationship between SAR and applied AMF) and (b) different concentrations of Fe at an applied field of 0.335 kOe (inset shows the IR thermogram of 1 mg ml⁻¹ of Fe at an AMF of 0.335 kOe).

The use of Fe₃O₄ nanoparticles in hyperthermia therapy depends on their heating ability, which is expressed in terms of the specific absorption rate (SAR). The SAR values of PPNMs were found to be 38, 54.3 and 74.0 W g⁻¹ of Fe with applied fields of 0.251, 0.293 and 0.335 kOe, respectively. Since SAR values are dependent on the field strength and frequency, we have calculated the system-independent intrinsic loss power (ILP) as follows:³⁹

where H is the field strength and f is the frequency. The ILP values obtained were 0.35, 0.37 and 0.39 nH m² kg⁻¹, with applied fields of 0.251, 0.293 and 0.335 kOe respectively for 2.30 mg ml⁻¹ of the sample, which itself shows normalization irrespective of the different applied field. Further, the rise in temperature is also found to be dependent on the concentration of particles in the suspension. However, the heating efficacy (i.e., SAR) decreases with increasing Fe concentration (even though the time required for reaching hyperthermia temperature decreases). This may be due to the decrease in the Brownian contribution to hyperthermia and increase in magnetic dipole–dipole interactions between nanoparticles in suspension, as a result of the increase in local concentration.⁴⁰ The ILP values obtained in the present study are in the range of those reported for commercially available ferrofluids.³⁹ The rise in temperature of the PPNM suspension under AMF is also visualized from the IR thermogram (centre bright circle), and this further demonstrates the localized heating of PPNMs (inset of Fig. 5b). This is highly advantageous for in vitro hyperthermia. Therefore, these PPNMs can be used as an excellent heating source for hyperthermia treatment of cancer.

The negatively charged surfaces of PPNMs make them attractive vehicles for the delivery of electrostatically bound drug molecules. DOX, a cationic drug used in chemotherapy, is chosen as a model drug to estimate the drug loading and release behavior of PPNMs. The high affinity of DOX for negatively charged species is well reported in the literature.^2,40–43 We have performed zeta potential and fluorescence spectroscopic studies to explore the binding of DOX with PPNMs. The zeta potential of a 1 ml aqueous suspension of PPNMs (100 μg) is increased from −32.0 mV to −9.5 mV upon interacting with 10 μg of DOX (Fig. S9, ESI†). This change in surface charge value can be attributed to the binding of cationic DOX with negatively charged nanomagnets through electrostatic interactions, thus partially passivating the surface. The interaction of DOX molecules with PPNMs was also apparent from the variation in fluorescence intensity of the supernatant liquid after removal of DOX-loaded PPNMs through magnetic separation (Fig. 6). The fluorescence intensity of the supernatant liquid drops with increasing concentration of PPNMs, due to the increase in available surface sites for drug conjugation. From the encapsulation efficiency plot (inset of Fig. 6), the encapsulation efficiency is found to be strongly dependent on the weight ratio of PPNMs to DOX (no significant increase in encapsulation efficiency is observed beyond a 1 : 14 ratio). An encapsulation efficiency (w/w) of 64% was obtained upon interacting 0.5 ml of aqueous solution of DOX (1 mg ml⁻¹) with 2.5 ml of the aqueous suspension of PPNMs (5 mg) for 1 h in the dark. Further, the drug-loaded PPNMs offered a DOX loading content of 6.4%. It is worth mentioning that PPNMs still reserve water solubility and good dispersibility after being coupled with DOX molecules (a TEM image of DOX-loaded PPNMs is shown in Fig. S10, ESI†).

Fig. 6 Normalized fluorescence spectra of 1 ml of DOX (10 μg) upon reaction with different quantities (0, 20, 40, 60, 80, 100, 120 and 140 μg) of PPNMs. The inset shows the encapsulation efficiency obtained from decrease in fluorescence intensities.

The release of drug from the DOX–PPNM system (Fig. 7) shows a time-dependent release of drug molecules, and the release rate increases with decreasing pH value. The initial stage of drug release is characterized by a rapid release of drug, followed by a slow and steady release of drug. It has been noted that about 54.6 and 60.5% of loaded DOX molecules were released from the DOX–PPNM system at pH 5 vs. pH 7.4 and pH 4 vs. pH 7.4, respectively, after 48 h under reservoir–sink conditions at 37 °C (Fig. 7). The release of DOX could be attributed to the weakening of the electrostatic interactions between cationic DOX and partially neutralized phosphate groups on the surface of the nanomagnets at acidic pH. Thus, the pH-triggered DOX release occurs across the range of pH values usually found in the intracellular matrix (pH ∼ 4), or the local environment in some cancerous tissues, thereby enabling targeted therapeutics by passive release at clinically relevant sites.

Fig. 7 pH-dependent drug release profile of DOX–PPNM in a cell-mimicking environment (reservoir pH 5/pH 4 and sink pH 7.4) at 37 °C.

In vitro experiments were also conducted to determine the effects of these PPNMs on cell viability. From the MTT assay, it has been observed that PPNMs have a negligible effect on the viability of WEHI-164 cells even at concentrations as high as 1.5 mg ml⁻¹ (Fig. 8a). This result indicates that PPNMs have minimal inherent toxicity on WEHI-164 cells. However, DOX and DOX–PPNM caused significant reduction in the proliferation of WEHI-164 cells (Fig. 8b). The relatively low cytotoxicity of DOX–PPNM compared to pure DOX can be ascribed to the sustained release behavior of the drug from the DOX–PPNM system (the loaded drug is expected to be released slowly over the experimental period⁴⁴). In addition, the percentage of hemolysis was found to be around 3% upon incubation of 0.5 mg of PPNMs, which indicates their good hemocompatibility. We also investigated the interaction of PPNMs with BSA protein in physiological medium (0.01 M PBS, pH 7.4). The PPNMs do not show any significant change in zeta potential (Table S1, ESI†) even after incubation with BSA for 2 h, revealing their protein resistance characteristics in physiological medium.

Fig. 8 Viabilities of WEHI-164 cells incubated in a medium containing different concentrations of (a) PPNMs and (b) DOX and DOX–PPNM at 37 °C for 48 h.

The high loading affinity of PPNMs for DOX and their sustained release and cytotoxicity effect further impelled us to explore the cellular uptake of DOX–PPNM. Fig. 9 shows the CLSM images of WEHI-164 cells after incubation with DOX–PPNM under culture conditions. A significant uptake of DOX–PPNM was clearly observed from the red fluorescence image arising from DOX emissions, suggesting that the drug-loaded nanoparticles were internalized in the cells. The blue fluorescence image shows emission from nuclei stained with DAPI. The merged image of DOX and DAPI fluorescence (as is seen from the magenta colour) clearly indicates that DOX–PPNM is mainly localized in the cytoplasm. This study demonstrates that the use of these nanocarriers as drug delivery vehicles could significantly enhance the accumulation of the DOX at target cancer cells, leading to a high therapeutic efficacy. Specifically, the present study demonstrates the formation of stable, highly crystalline, biocompatible, protein resistance polyphosphate-grafted nanomagnets with good magnetization and excellent intrinsic loss power. This makes them amenable for applications in hyperthermia treatment of cancer, as well as pH-triggered release of DOX for combination therapy involving hyperthermia and chemotherapy.

Fig. 9 CLSM images of WEHI-164 cells after incubation with the DOX–PPNM and DAPI at culture conditions.

Conclusions

In summary, the synthesis of hydrophilic multifunctional polyphosphate-grafted nanomagnets of average size 10 nm is reported. XRD and TEM analysis confirmed the formation of highly crystalline single-phase Fe₃O₄ nanostructures. Detailed structural analysis by FTIR, TGA, DLS and zeta potential confirmed the successful grafting of nanomagnets with polyphosphate moieties. These superparamagnetic nanomagnets exhibit good colloidal stability and biocompatibility along with self-heating efficacy under an AC magnetic field. The drug loading efficiency of the PPNMs and their pH-triggered release were also investigated, using DOX as a model drug. Cellular imaging experiments reveal that these nanomagnets can be easily internalized by cells. These multifunctional nanomagnets may have potential benefits for the enhancement of in vivo therapeutic efficacy by decreasing the systemic toxicity of antitumor drugs, both through protection of the drug throughout circulation and by means of magnetic fields to target the nanomagnets to the infected area. Further, the high loading affinity for DOX with their sustained release and localized heating ability under AMF makes these novel magnetic nanoparticles suitable for cancer therapy.

Acknowledgements

The authors thank Dr B. N. Jagatap, Director, Chemistry Group and Dr V. K. Jain, Head, Chemistry Division, BARC for the encouragement and support. The authors also thank Dr R. S. Ningthoujam, BARC for facilitating the use of the induction heater. The authors are grateful to Head, Analytical Chemistry Division, BARC for providing ICP-OES facility. The authors also thank Mr Manjoor Ali and Ms Vasumathy Pillai for confocal laser scanning microscopy studies.

Notes and references

1. J. Cheon and J.-H. Lee, Acc. Chem. Res., 2008, 41, 1630.
2. K. C. Barick, S. Singh, N. V. Jadav, D. Bahadur, B. N. Pandey and P. A. Hassan, Adv. Funct. Mater., 2012, 22, 4975.
3. O. D. Jayakumar, R. Ganguly, A. K. Tyagi, D. K. Chandrasekharan and K. Nair, J. Nanosci. Nanotechnol., 2009, 9, 6344.
4. J. Zhu, J. Wang, X. Wang, J. Zhu, Y. Yang, J. Tian, W. Cui, C. Ge, Y. Li, Y. Pan and H. Gu, J. Mater. Chem. B, 2015, 3, 6905.
5. J. Zhu, B. Zhang, J. Tian, J. Wang, Y. Chong, X. Wang, Y. Deng, M. Tang, Y. Li, C. Ge, Y. Pan and H. Gu, Nanoscale, 2015, 7, 3392.
6. C. Boyer, M. R. Whittaker, V. Bulmus, J. Liu and T. P. Davis, NPG Asia Mater., 2010, 2, 23.
7. J. Majeed, L. Pradhan, R. S. Ningthoujam, R. K. Vasta, D. Bahadur and A. K. Tyagi, Colloids Surf., B, 2014, 122, 396.
8. Y.-K. Peng, C. N. P. Lui, T.-H. Lin, C. Chang, P.-T. Chou, K. K. L. Yung and S. C. E. Tsang, Faraday Discuss., 2014, 175, 13.
9. M. Das, D. Mishra, P. Dhak, S. Gupta, T. K. Maiti, A. Basak and P. Pramanik, Small, 2009, 5, 2883.
10. D. Portet, B. Denizot, E. Rump, J. J. Lejeune and P. Jaller, J. Colloid Interface Sci., 2001, 238, 37.
11. Y. Sahoo, H. Pizem, T. Fried, D. Golodnitsky, L. Burstein, C. N. Sukenik and G. Markovich, Langmuir, 2001, 17, 7907.
12. S. Chandra, K. C. Barick and D. Bahadur, Adv. Drug Delivery Rev., 2011, 63, 1267.
13. X. L. Liu, H. M. Fan, J. B. Yi, Y. Yang, E. S. G. Choo, J. M. Xue, D. D. Fan and J. Ding, J. Mater. Chem., 2012, 22, 8235.
14. J. Xie, K. Chen, H. Y. Lee, C. Xu, A. R. Hsu, S. Peng, X. Chen and S. Sun, J. Am. Chem. Soc., 2008, 130, 7542.
15. N. Bertrand, J. Wu, X. Xu, N. Kamaly and O. C. Farokhzad, Adv. Drug Delivery Rev., 2014, 66, 2.
16. Z. Xiao, C. Ji, J. Shi, E. M. Pridgen, J. Frieder, J. Wu and O. C. Farokhzad, Angew. Chem., 2012, 124, 12023.
17. K. Yan, H. Li, P. Li, H. Zhu, J. Shen, C. Yi, S. Wu, K. W. Yeung, Z. Xu, H. Xu and P. K. Chu, Biomaterials, 2014, 35, 344.
18. C. Yee, G. Kataby, A. Ulman, T. Prozorov, H. White, A. King, M. Rafailovich, J. Sokolov and A. Gedanken, Langmuir, 1999, 15, 7111.
19. P. Sharma, S. Rana, K. C. Barick, C. Kumar, H. G. Salunke and P. A. Hassan, New J. Chem., 2014, 38, 5500.
20. C. Tudisco, F. Bertani, M. T. Cambria, F. Sinatra, E. Fantechi, C. Innocenti, C. Sangregorio, E. Dalcanale and G. G. Condorelli, Nanoscale, 2013, 5, 11438.
21. K. Löster, S. Seidel, D. Kirstein, F. Schneider and F. Noll, J. Immunol. Methods, 1992, 148, 41.
22. R. Palmeira-de-Oliveira, A. Palmeira-de-Oliveira, C. Gaspar, S. Silvestre, J. Martinez-de-Oliveira, M. H. Amaral and L. Breitenfeld, Int. J. Pharm., 2011, 421, 130.
23. S. Vimal, S. A. Majeed, G. Taju, K. S. Nambi, N. S. Raj, N. Madan, M. A. Farook, T. Rajkumar, D. Gopinath and A. S. S. Hameed, Acta Trop., 2013, 128, 486.
24. K. Konecsni, N. H. Low and M. T. Nickerson, Food Chem., 2012, 134, 1775.
25. S. Tripathy, S. Das, S. P. Chakraborty, S. K. Sahu, P. Pramanik and S. Roy, Int. J. Pharm., 2012, 434, 292.
26. N. N. Rao, M. R. Gómez-Garcia and A. Kornberg, Annu. Rev. Biochem., 2009, 78, 605.
27. K. C. Barick and P. A. Hassan, J. Colloid Interface Sci., 2012, 369, 96.
28. B. Ding, S. Xia, K. Hayat and X. Zhang, J. Agric. Food Chem., 2009, 57, 2938.
29. K. C. Barick, M. Aslam, Y. P. Lin, D. Bahadur, P. V. Prasad and V. P. Dravid, J. Mater. Chem., 2009, 19, 7023.
30. V. Zamora-Mora, M. Fernández-Gutiérrez, J. S. Román, G. Goya, R. Hernández and C. Mijangos, Carbohydr. Polym., 2014, 102, 691.
31. J. Wan, W. Cai, X. Meng and E. Liu, Chem. Commun., 2007, 5004.
32. J. E. Wong, A. K. Gaharwar, D. Muller-Schulte, D. Bahadur and W. Richtering, J. Nanosci. Nanotechnol., 2008, 8, 4033.
33. M. Mikhaylova, D. Y. Kim, N. Bobrysheva, M. Osmolowsky, V. Semenov, T. Tsakalakos and M. Muhammed, Langmuir, 2004, 20, 2472.
34. S. Chandra, S. Mehta, S. Nigam and D. Bahadur, New J. Chem., 2010, 34, 648.
35. L. S. Goldstein, M. W. Dewhirst, M. Repacholi and L. Kheifets, Int. J. Hyperthermia, 2003, 19, 373.
36. C. S. S. R. Kumar and F. Mohammad, Adv. Drug Delivery Rev., 2011, 63, 789.
37. A. Tomitaka, T. Koshi, S. Hatsugai, T. Yamada and Y. Takemura, J. Magn. Magn. Mater., 2011, 323, 1398.
38. R. Ghosh, L. Pradhan, Y. P. Devi, S. S. Meena, R. Tewari, A. Kumar, S. Sharma, N. S. Gajbhiye, R. K. Vatsa, B. N. Pandey and R. S. Ningthoujam, J. Mater. Chem., 2011, 21, 13388.
39. M. Kallumadil, M. Tada, T. Nakagawa, M. Abe, P. Southern and Q. A. Pankhurst, J. Magn. Magn. Mater., 2009, 321, 1509.
40. S. Rana, N. V. Jadhav, K. C. Barick, B. N. Pandey and P. A. Hassan, Dalton Trans., 2014, 43, 12263.
41. S. Nigam, K. C. Barick and D. Bahadur, J. Magn. Magn. Mater., 2011, 323, 237.
42. E. Munnier, F. Tewes, S. Cohen-Jonathan, C. Linassier, L. Douziech-Eyrolles, H. Marchais, M. Soucé, K. Hervé, P. Dubois and I. Chourpa, Chem. Pharm. Bull., 2007, 55, 1006.
43. F. A. D. Wolf, K. Nicolay and B. D. Kruijff, Biochemistry, 1992, 31, 9252.
44. K. C. Barick, S. Nigam and D. Bahadur, J. Mater. Chem., 2010, 20, 6446.

---

Footnote

† Electronic supplementary information (ESI) available: TEM images, SAED pattern, FTIR, TGA, DLS, UV-visible, zeta-potential and magnetization plots, and protein–particle interaction data (Fig. S1–S10, Table S1). See DOI: 10.1039/c5ra16343a

---