Figure 2a shows the X-ray diffraction pattern of dried MNPs used for MFWL and MFWI fluid preparation. The similarity between patterns indicates that stability procedures did not change the crystal structure of MNPs. The crystal structure obtained from the pattern analysis indicates impurity-free single-phase cubic spinel ferrite with no extra peaks other than the ferrite phase. For calculating the crystallite size (DXRD), the most intense peak (311) and Scherrer formula (Eq. 3) was used (Patterson 1939):
$${D}_{\mathrm{XRD}}=\frac{0.9\uplambda }{B\mathrm{cos}\theta },$$
(3)
where B denotes full-width half maxima of diffraction peak, λ is X-ray wavelength, and \(\theta \) is the Bragg’s angle. The crystallite size of surfactant-stabilized particles (MFWL) was found to be higher (11.0 ± 0.5 nm) as compared to charge stabilized particles (8.3 ± 0.5 nm). The procedure of charge grafting might have removed some of the Fe ions from its surface, but that affects physical size only. The slight variation in observed particle size may be due to the probability of size variation from batch to batch, which is common for chemical process, especially when NH4OH is used as a base ion for the precipitation. This was also confirmed by calculating the lattice parameter ‘a’ using the analytical method (Suryanarayana and Norton 1998). The lattice parameter was found to be 0.8436 ± 0.0005 nm for MFWL particles and 0.8360 ± 0.0005 nm for MFWI particles. X-ray diffraction peak positions for both MNPs were indexed as (220), (311), (400), (422), (511), and (440).
Magnetic response as a function of magnetic field for both fluids is shown in Fig. 2b. The magnetic characterization experiments were performed at room temperature (25 °C). The obtained data were fitted with Langevin function (Parekh et al. 2008) for superparamagnetic non-interacting particles to obtain the magnetic parameters like the magnetic size of particles (Dm), size distribution (σ), domain magnetization (Md), and saturation magnetization (Ms). Domain magnetization of both fluids was kept the same as 450 kA/m. Dm was found to be 10.3 nm for MFWL and 8.2 nm for MFWI. The magnetic size is important to know because the reduced magnetic size can lower the heating response during the induction heating. The magnetic size distribution σ was found as 0.53 and 0.50, respectively, for MFWL and MFWI fluid. The obtained saturation magnetization was much lower for MFWI (15 Gauss) as compared to 145 Gauss for MFWL. This difference in magnetization can be attributed to a lower concentration of MFWI (11.6 mg/mL) as compared to MFWL (217 mg/mL). The inset Fig. 2b shows the size distribution curve for both the fluids. The dotted line shows the critical diameter of 9.2 nm above which the particles are dominated by the Brownian mechanism (Rosensweig 2002). It is seen that almost 60% particles in MFWL sample is above this critical diameter, whereas for MFWI sample it is 42%.
Figure 3a and c shows the TEM image of the sample MFWL and MFWI, respectively. The particle morphology, as seen from both the images, is spherical in shape. In addition, these particles are well separated from each other. The size of these particles was estimated using ImageJ software. Nearly 300 particles from various TEM images were considered for plotting the histogram which is shown in Fig. 3b and d, respectively, for MFWL and MFWI samples. The histogram is fitted with the log-normal distribution curve (red line) that fits with the mean size of 10.2 ± 0.24 nm with a size distribution of 0.18 for the MFWL sample and 7.6 ± 0.04 nm with a size distribution of 0.14 for the MFWI sample. The size distribution in MFWI is narrow as compared to that of MFWL. Also, the TEM size is observed less for MFWI as compared to MFWL, which matches with the results obtained from the XRD measurement.
In the thermogravimetric experiment, the particles showed a total weight loss of 32.8% for surfactant-coated particles (MFWL) and 16% for charge grafted particles (MFWI) as shown in Fig. 4a and b. The first weight loss of approximately 2% occurring below 100 °C in both particles is due to the evaporation of physi-adsorbed water molecules attached to the particle surface. The major weight loss in the case of MFWL happening at a high temperature indicates the chemical attachment of surfactant on the particle surface, and the two-step loss indicates the bilayer arrangement of surfactant (Pradhan et al. 2007a; Mahdavi et al. 2013). The particles of MFWL showed major weight loss occurring in two steps, at 248 °C and 344 °C, while, for MFWI major weight loss occured in three steps at 209 °C, 287 °C, and 357 °C. In the case of MFWL sample, The first weight loss of 20.2% was due to the removal of secondary layer or extra surfactant, whereas the second loss of 28.7% was due to the removal of the chemically adsorbed primary layer of surfactant. In the case of MFWI, the removal of adsorbed double diffuse layers (Cao and Wang 2011) of ions from the particle surface resulted in the multiple transitions between 209 and 357 °C. After 400 °C the transition becomes stable, indicating no phase change above this temperature.
Figure 4c shows the FTIR spectra of surfactant lauric acid, the lauric acid-coated MNPs, and charge grafted MNPs. The peak observed at 1698 cm−1 in the case of lauric acid spectrum shifted to 1704 cm−1 for MFWL, indicating the C=O stretch of the carboxylic head of the surfactant molecule and its covalent bonding to the particle surface. The characteristic iron oxide Fe–O stretching vibration occurs at 567 cm−1 for MFWL and at 563 cm−1 for MFWI (Pradhan et al. 2007a). The band at 940 cm−1 corresponding to O–H bending vibration in the spectra of lauric acid vanished in the spectra of MFWL, and the presence of a weak band at 1704 cm−1 indicated strong adsorption and close packing of surfactant layers on particles (Lenin and Joy 2016). The band at 1422 cm−1 confirmed the presence of nitro group in the spectra of MFWI.
Figure 5a and b shows the log-normal distribution fit to the number distribution of both aqueous magnetic fluids. The hydrodynamic size measured was 29.3 ± 0.3 and 24.6 ± 0.3 nm for MFWL and MFWI, respectively. The standard deviation of the distribution curve (σ) was found to be 0.24 and 0.18, respectively, for MFWL and MFWI. Hydrodynamic size is not the actual particle size as it contains the thickness of multiple layers of surfactant as well as diffuse layer of solution (water in our case) around the particle. Basically, DLS derives the hydrodynamic size of the particles using the concept of diffusion of the particles in the medium. The results show a higher value of hydrodynamic diameter (Dh) in both fluids as compared to the crystallite size obtained from the X-ray diffraction pattern indicating the possibility of multiple layers of surfactant coating. This larger size can also be due to the formation of small aggregates stable in solution even after dilution (Odenbach 2002).
Hyperthermia without cells
The heat released due to magnetic fluids placed under the AC magnetic field is given by Eq. 4 defined as
$$ P_{{{\text{adsorbed}}}} = \mu_{0} \pi \cdot \chi^{\prime\prime} \cdot f \cdot H^{2} , $$
(4)
where μ0 is the permeability of free space, χ″ is the imaginary component of susceptibility, f is a frequency, and H is the applied field strength (Rosensweig 2002). This heat dissipation is caused mainly by the relaxation of magnetic moments termed as Neel relaxation and Brownian relaxation. In Neel relaxation, there are rapid changes in the particle's magnetic moment due to opposition by the particle’s crystalline structure, resulting in heat generation when exposed to the AC magnetic field. In Brownian relaxation, the heat is generated from the physical rotation of particles within the medium when particles attempt to realign themselves with the changing magnetic field. These relaxations also depend upon the size of magnetic particles (Odenbach 2002). The heating efficiency is quantified by the parameter SAR, which is equal to the dissipated power divided by the magnetic material density, showing that the SAR varies linearly with the product of frequency, f, and the square of the magnetic field (H2). From the induction heating data of the rise in temperature with respect to time, the SAR value was calculated using Eq. 5:
$$ {\text{SAR}} = C_{p} \cdot \frac{\Delta T}{{\Delta t}} \cdot \frac{1}{{\varphi_{{{\text{magnetic}}}} }} $$
(5)
where Cp denotes the specific heat capacity of the magnetic fluids given by
$$ C_{p} = m_{{{\text{particles}}}} *C_{{p - {\text{particles}}}} + m_{{{\text{carrier}}}} *C_{{p - {\text{carrier}}}} , $$
(6)
where \(\frac{\Delta T}{\Delta t}\) denotes the slope of the graph between temperature rise and induction heating time. φmagnetic is the weight fraction of the magnetic content of particles. Cp for carrier and particles was taken as 4.187 and 0.67 J/g-K, respectively. For compensating the errors in measurement of SAR for non-adiabatic system, hyperthermia data were fitted with Box-Lucas Eq. 7 Kallumadil et al. 2009) given as follows:
$$ T(t) = A(1 - e^{ - Bt} ) $$
(7)
where the rise in temperature, T with respect to time, t is given as a function of saturation temperature, A, and curvature of the heating curve, B. The product A × B is equivalent to the initial heat rise rate, \(\frac{\Delta T}{\Delta t},\) used in the SAR formula.
Figure 6a and b showd the temperature rise versus time and corresponding SAR values for MFWL diluted in the water, while Fig. 6c and d shows the same results for the MFWI sample. The experiments were repeated three times to check the reproducibility of results, and the variation in data is shown as error bar in the graphs. These fluids were diluted in deionized water to 10 mg/mL concentration, and the magnetic field was varied from 1.7 to 10 kA/m. The rise in temperature, as well as corresponding SAR, was found to be higher for the larger magnetic field for both the fluids, as evident from the relation between SAR and magnetic field. This increase was greater for MFWL as compared to MFWI, indicating a better hyperthermic response of MFWL. This might be due to the slightly larger magnetic size of MNPs of MFWL compared to those of MFWI as heating would be more for large particle size due to Brownian relaxation. The increasing SAR with increasing particle size has been reported in the literature for γ-Fe2O3 nanoparticles (Purushotham et al. 2009) and Fe3O4 nanoparticles (Parekh et al. 2018). This is due to the fact that larger magnetic size will have a better response towards magnetic field direction as compared to the smaller size particles because the magnetic moment of particle is a function of particle size,
especially when it is in the nanometer range. Thus, the SAR shows variation with particle size.
In addition to this, it has also been reported (Rosensweig 2002) that the SAR should be dominated by the Brownian relaxation mechanism as compared to the Neel relaxation mechanism. The size distribution curve obtained from magnetic measurement for MFWL and MFWI samples (Inset Fig. 2b) indicated that almost 60% particles in MFWL sample is above the critical size of 9.2 nm, whereas in the case of MFWI, it is 42%. Larger the number of particles dominated by the Brownian relaxation mechanism more the induction heating response. Hence, the observed higher SAR value for MFWL as compared to MFWI is contended.
Figure 7a and b shows the temperature rise versus time and corresponding SAR values for MFWL diluted in the water while varying magnetic particle concentration. Figure 7c and d shows the same for the MFWI sample. The magnetic field was kept fixed at 10 kA/m for both the fluids and for all dilutions. From this figure, it is clear that the initial rise in temperature is faster as concentration increases, but when we calculated SAR, which takes care of particle concentration by normalizing the value of heating rate with particle concentration (as mentioned in Eq. 5), the value remained same. This indicates that SAR is independent of concentration. Usually, the SAR is a function of magnetic field, frequency, particle composition, its size, shape, concentration, etc. When all other parameters are fixed except the particle concentration, then there is a possibility that with increasing or decreasing concentration, the particle–particle interaction or particle–carrier interaction play its role in creating the aggregation of the particles. In both cases, the SAR will be a function of particle concentration. Under such a situation, the system becomes more complex and non-predictable. Moreover, both conditions led to aggregation mechanism, which becomes more toxic to the cells. When SAR is independent of particle concentration, that means that the system is very stable upon dilution, and in order to increase the heating temperature, one can increase only the magnetic field strength. This is a very important inference; otherwise, it will be difficult to fix the concentration and magnetic field for induction heating experiments for in vitro cell line experiments. In the case of MFWI, the small variation in SAR with concentration may be due to the possibility of the formation of small aggregates. The possible reason for this may be the disturbance of charge layers around the particle surface because the dilution of magnetic fluids was carried out using distilled water having neutral pH. However, this increase is within an error bar. Moreover, the induction heating is also performed only on the water (without dispersing MNPs) substantiated that heating was due to the MNPs and not because of radiative heating via the high current flowing through the coil. As reflected in Fig. 7a and c, the temperatures rise in water was less than 1 °C even after prolonged heating of the coil.
These experiments, related to hyperthermic response, with the variation of magnetic field and concentration, help to decide the concentration of magnetic fluids for in vitro experiments and develop future magnetic fluid hyperthermia based in vivo therapeutic strategies on cancer patients. Overall, the experiments aimed to synthesize magnetic fluids with greater hyperthermic response at the lowest possible magnetic field and concentration at a given frequency.
In vitro experiments
To study the cytotoxic effect of the magnetic fluids, MTT assay was performed, and IC50 value was obtained. This assay provides a half-maximal inhibitory concentration, i.e., the IC50 value of the test compound required for 50% inhibition of viable cell numbers in vitro. The effect of different concentrations of MFWL and MFWI on cell viability via MTT assay is shown in Fig. 8a and b, respectively. Further, fitting Hill’s equation to the dose–response curve (Hill 1910; Sebaugh 2011), the assay revealed IC50 values of 0.271 mg/mL and 0.206 mg/mL for MFWL and MFWI, respectively. A similar cytotoxicity study of lauric acid-coated magnetite particles on murine microglial BV2 cells was performed by Calatayud et al. (2017), who reported approximate 30% cell death after 24 h incubation with MNPs at a concentration of 0.1 mg/mL by performing TPB assay. However, using polyacrylic acid-coated MNPs under a similar experimental scenario, the same group observed almost 100% viability of BV2 cells. Intriguingly, using the lauric acid-coated particles, Pradhan et al. (2007b) did not find any detrimental effect on human cervical cancer cells HeLa and mouse fibroblasts L929 at 0.1 mg/mL concentration; however, they observed approximate 8% cell death at 0.2 mg/mL concentration using sulforhodamine B assay. The IC50 of MFWL in the present study is in good agreement with the results of Freitas et al. (Freitas et al. 2008), who reported 50% cell inhibition at 0.254 mg/mL using lauric acid-coated γ-Fe2O3 on human melanoma cells via MTT assay. The variation in the cell death percentage may be attributed to the difference in methods of MNPs synthesis, lauric acid coating, cells’ physiologic features, types of cell lines used, as well as cell viability assay utilised. With regard to MFWI fluid, till date no report is available on its in vitro cytotoxic effect. However, a single in vivo study revealed toxic effect of ionic MnFe2O4 MF causing macrophage apoptosis, lymphocytes’ DNA damage, and severe inflammatory response in the peritoneal cavity of mice (Lacava et al. 1999).
Our experiments related to induction heating on the water revealed that the heat generated was due to MNPs, and it was not radiative heating (Fig. 7a and c). Therefore, we proceeded to evaluate the effect of induction heating hyperthermia on HeLa cells at 15.3 kA/m and 330 kHz after choosing the near IC50 values of 0.25 and 0.2 mg/mL for MFWL and MFWI, respectively. Looking at the concentration-dependent heating response from Fig. 7a and c, it is seen that more than 1 mg/mL concentration is required in both the samples to achieve the hyperthermia temperature of 45 °C. However, the IC50 value is very less than this concentration. The magnetic field used for this experiment was 10 kA/m. Since the cell line experiments were carried out with the cell culture media, the IC50 dilution was also prepared in cell culture media for induction heating experiment. Additionally, as it is inferred from the above measurements that magnetic field strength can be used to increase the temperature, the heating response at IC50 concentration was investigated at 15.3 kA/m field.
Figure 9 shows the heating response for both the fluids as a function of time for IC50 concentration, 0.25 and 0.2 mg/mL for MFWL and MFWI, respectively, prepared in cell culture media and investigated at 15.3 kA/m field and 330 kHz frequency. It is seen that IC50 dilution in cell culture media is able to achieve the heating temperature of 45 °C. The possible reason to achieve the required temperature even with the lower MNPs concentration can be due to the presence of ions and proteins in the cell culture media, which also takes part in the heating mechanism and contribute to the induction heating. Such type of study is also reported by Chanteau et al. (2009).
Although MTT assay is based on the metabolic response of cells and being less laborious and quick to perform to study cell death, TPB assay was performed for hyperthermic study to have an absolute visualization of the cell death under a microscope. Live and dead cells were observed under the microscope, and the viability of cells was calculated using the ratio of live cells to the total (live + dead) cells which was normalized against the control cells.
Figure 10a and b shows the cell viabilities obtained after MFH utilizing MFWL and MFWI, respectively. Reduction in cell count, as well as morphologic alteration observed after MFH by MFWL and MFWI, are represented in Fig. 10c1 to c4 and d1 to d4, respectively. Due to hyperthermia alone, i.e., without MF, after 30- and 60-min’ sessions, cell death was approximately 7 and 15%, respectively. Cells under magnetic filed usually show cell death up to 5% that may be due to mechanical stress faced by cells as a part of experimental procedures. However, under alternating magnetic field, approximately 85% of cells showed viability even when the induction heating time was increased to 60 min. This suggests that the cells well endured the magnetic field and heat-shock without major population loss. It would also be interesting to analyze the expression of heat-shock proteins such as HSP27/HSP70 and HSP90 as these proteins are known to overexpress in cancer under the hyperthermic condition and provides resistance to damage caused by temperature rise, thereby the therapeutic efficacy of chemotherapy (Grimmig et al. 2017). In addition, another reason for cell death under the influence of magnetic field could be the generation of reactive oxygen species in the cells as reviewed by Wang and Zhang (2017).
Considering MFWL, 24 h treatment without hyperthermia resulted in an approximate 38% cell death, whereas MFWI caused around 55% cell death. Further, MFWL, after 24 h of MF treatment and subsequent 30 min HT led to about 45% cell death that was augmented to around 68% after 24 h MF treatment and 60 min HT session. On the other hand, MFWI was comparatively toxic than MFWL that resulted in an approximate 68% and 80% cell death under the same hyperthermic experimental conditions. All the groups in both fluids showed significant cell death compared to control groups (p < 0.05; p < 0.01). This high rate of cell death due to MNPs without induction heating might have occurred due to MNPs aggregation in the media (Eberbeck et al. 2010), leading to a change in MNP concentration. Another reason could be the toxicity exhibited by the surfactant (Pradhan et al. 2007b). Furthermore, a higher rate of cell death observed due to charged MF as compared to lauric acid-coated MF can be attributed to a change in membrane potential due to the presence of charge on the MNPs leading to disruption of cells’ plasma membrane, thereby inducing cell death. Membrane potential studies are warranted to understand the mechanism of cell death caused by charged magnetic fluids.
With regard to the effect of surfactant on MFH, our results corroborate with the findings of Calatayud et al. (2017) who observed an approximate 44% cell death using polyacrylic acid-coated iron oxide MNPs on microglial BV2 cells after 30 min of hyperthermia, however, at a higher frequency 560 kHz and field 23.9 kA/m. Similarly, Hedayatnasab et al. (2020), reported 40 and 60% cell death at 31.47 kA/m and 47.24 kA/m fields, respectively, using polycaprolactone (PCL) coated cetyl trimethyl ammonium bromide (CTAB)-modified iron oxide particles on human liver carcinoma cells HepG2 at a concentration of 0.1 mg/mL, while maintaining the hyperthermic window of 43 to 46 °C. Interestingly, they incubated the cells with MNPs only for 4 hours before MFH compared to 24 h incubation of ours. Previously we reported the effect of MFH on HeLa cells using lauric acid-coated Mn–Zn ferrite fluid under the same frequency and field that resulted in 55 and 60% cell death after 30 and 60 min treatment, respectively, at 0.35 mg/mL concentration (Bhardwaj et al. 2020). Enhanced cell death of 70 and 76% after 30 and 60 min of treatment, respectively, was observed when hyperthermia was performed using MF of 0.75 mg/mL concentration (Parekh et al. 2020). Comparing the hyperthermic effect of different MF types on cell viability in the present study and our previous reports, the MFWI was found to be more toxic at a lower concentration, followed by MFWL. One of the limitations of our study was the observed cytotoxicity due to MF alone without induction heating. MFs leading to effective killing of cancer cells after induction heating are desirable. Overall, our study results suggest that the outer stabilizing layer, MF concentration, and hyperthermia duration are essential parameters for designing and planning future MFH based therapy against cancer.