Comparison of magnetic nanoparticle and microwave hyperthermia cancer treatment methodology and treatment effect in a rodent breast cancer model

Abstract

Purpose: The purpose of this study was to compare the efficacy of iron oxide/magnetic nanoparticle hyperthermia (mNPH) and 915 MHz microwave hyperthermia at the same thermal dose in a mouse mammary adenocarcinoma model. Materials and methods: A thermal dose equivalent to 60 min at 43 °C (CEM60) was delivered to a syngeneic mouse mammary adenocarcinoma flank tumour (MTGB) via mNPH or locally delivered 915 MHz microwaves. mNPH was generated with ferromagnetic, hydroxyethyl starch-coated magnetic nanoparticles. Following mNP delivery, the mouse/tumour was exposed to an alternating magnetic field (AMF). The microwave hyperthermia treatment was delivered by a 915 MHz microwave surface applicator. Time required for the tumour to reach three times the treatment volume was used as the primary study endpoint. Acute pathological effects of the treatments were determined using conventional histopathological techniques. Results: Locally delivered mNPH resulted in a modest improvement in treatment efficacy as compared to microwave hyperthermia (p = 0.09) when prescribed to the same thermal dose. Tumours treated with mNPH also demonstrated reduced peritumoral normal tissue damage. Conclusions: Our results demonstrate similar tumour treatment efficacy when tumour heating is delivered by locally delivered mNPs and 915 MHz microwaves at the same measured thermal dose. However, mNPH treatments did not result in the same type or level of peritumoral damage seen with the microwave hyperthermia treatments. These data suggest that mNP hyperthermia is capable of improving the therapeutic ratio for locally delivered tumour hyperthermia. These results further indicate that this improvement is due to improved heat localisation in the tumour.

• Cumulative equivalent minutes
• hyperthermia
• iron oxide
• microwave
• nanoparticle

Introduction

In the past 40 years, great strides have been made in understanding thermal dosimetry, treatment planning and the design of equipment used in medical hyperthermia [1]. Magnetic nanoparticles (mNP), in combination with an alternating magnetic field (AMF) are able to generate significant localised heating as well as a yet unspecified level of individual cell damage. Pre-clinical tumour studies, using intratumoral mNP delivery, have been promising, especially when combined with radiation and chemotherapy (see Petryk et al. in this issue, pp. 845–51). One of the primary reasons hyperthermia has not achieved widespread clinical use and acceptance is its inability to achieve localised tumour heating and the lack of an inherent increase of tumour cell sensitivity to heat [2–4]. The experiments presented here suggest that magnetic nanoparticle hyperthermia (mNPH) has the potential to become a more effective method of heat delivery to large tumour masses. The ability to treat individual (metastatic) cancer cells is also an area of active investigation with great promise.

Biological effects of hyperthermia

The primary molecular targets of traditional thermal therapy are a diverse assortment of proteins. Alterations in enzymes, structures in the cytoskeleton, proteins associated with the plasma membrane, and proteins associated with DNA repair have all been observed in response to elevated temperatures [5]. The production of heat shock proteins, which assist in the refolding of damaged proteins, and the inhibition of DNA synthesis are also characteristic of cells exposed to heat [6,7]. Both apoptosis and necrosis can be observed in tissue, depending on the thermal dose [8]. Factors which influence the sensitivity of cells to hyperthermia include cell nutrient levels, as well as pH and oxygen concentration. It is also well known that heat increases the cytotoxic effect of many chemotherapeutic agents [9–11].

MHz Microwave hyperthermia

Microwave-induced hyperthermia results from the rotational motion of polar molecules, primarily water, in response to an oscillating electric field. Rotational energy is converted into heat energy through frictional losses in response to the motion of the molecules [12]. Microwave antennae for the delivery of localised medical hyperthermia were developed in the late 1970s and used for the first time in clinical medicine approximately 10 years later [12,13]. Although significant clinical tumour response has been achieved using microwave therapy, significant side effects and inhomogeneous treatment effects have also been reported [14–16].

Magnetic nanoparticle hyperthermia

When exposed to an AMF, mNP are believed to generate heat by one or more of the following mechanisms: 1) magnetic hysteresis, 2) eddy currents, 3) Brownian motion, and 4) Néel paramagnetic switching [17,18]. Heating properties of nanoparticles depend greatly on their composition, size and microstructure [19]. For mNP, eddy currents are not a significant contributor to tumour heating [19,20]. The two main particle types for iron oxide-based mNPH are superparamagnetic mNP (Brownian and Néel mechanisms) and ferromagnetic mNP (hysteresis losses) [21]. The mNP utilised in this study have a ferromagnetic core composed of Fe₃O₄ crystals and are coated with biocompatible hydroxyethyl starch with a hydrodynamic radius of 110 nm [18,22]. At 450 Oe (35.8 kA/m), 165 kHz these mNP demonstrate a specific absorption rate (SAR), in terms of Fe mass, of 151 (W/g Fe), as calculated by the following equation: in which C is the specific heat capacity and is the initial heating slope observed and mNPc is the concentration of Fe (g Fe/g medium) [23,24].

Thermal dose

It is well understood that the biomedical effects of heat are a function both of time and temperature. Therefore, accurately predicting and prescribing treatment for various animal species and tissue is difficult without a means of comparing the thermal histories. Sapareto and Dewey proposed a method to normalise hyperthermia treatments conducted in different settings by describing the biological effect in terms of cumulative equivalent minutes at 43 °C (CEM) [25]. The CEM relationship is: , where ‘t’ is equal to the time interval at a specific temperature ‘T’, R equals 0.25 when temperatures are below 43 °C and 0.45 when temperatures are above 43 °C (rodent) [26]. The total thermal dose is equivalent to the summation of these values.

It is important to note that the CEM relationship, for thermal dose equivalency in tissue, has only been assessed for conventional forms of hyperthermia therapy such as microwave-, ultrasound- and RF-based platforms. It has not been assessed or proven for a modality such as mNPH where the heat source(s) are contained within and immediately outside of cells (cell membrane and interstitium) and the achieved temperature, on the macro-nano scale are not known. Furthermore, it is possible that the mNP heat mechanism of action is not based on heating alone [27,28]. In the following experiments, the validity of the CEM relationship for mNPH was explored by comparing ‘traditional’ 915 MHz hyperthermia with mNP hyperthermia. Although not specifically examined in this work, the intracellular location of mNP, as well as the grouping of the mNP within the cells, may result in improved biological (therapeutic) effects beyond those expected with tissue-level applications of hyperthermia. Although reports of improved mNP heating and/or cytotoxicity following intracellular mNP-initiated hyperthermia have been published, the issue remains unresolved and controversial [24,27–37]. In this work, the mNP were activated within 10 min of injection into the tumour. Previous studies from our group suggest that mNP are largely extracellular at this time [31]. A qualitative TEM-based evaluation of tumours indicates that the majority of mNP are extracellular at the time of AMF initiation (Figure 1).

Materials and methods

Breast cancer cells

Mouse mammary adenocarcinoma (MTGB) cells were grown in Alpha MEM media (Mediatech, Manassas, VA). The Alpha MEM media was modified with the addition of FBS 10% (HyClone, South Logan, UT), penicillin-streptomycin 1% (HyClone), and L-glutamine 1% (Mediatech).

Animal model

Syngeneic, MTGB tumours were grown in the flanks of female C3H mice (Charles River Laboratories, Wilmington, MA) aged 6–8 weeks. Cultured MTGB cells were treated with 0.25% trypsin in EDTA (HyClone). Cells were then suspended in unmodified Alpha MEM media, with a 50 µL sample taken for trypan blue assay evaluation. Cells were stained with trypan blue (Hyclone), in a 1:1 ratio and counted using a haemocytometer (Fisher Scientific, Pittsburg, PA). Cells were pelleted through centrifugation and resuspended in unmodified Alpha MEM media at 10 million cells per mL. Then 1 million cells were injected with a 25 gauge needle into the right rear flank of the mice. After 2–3 weeks, the tumours reached a volume of 150 mm³± 40 mm³, at which time they were treated in the method described below. Tumours were measured with digital calipers in three planes. Volumes were calculated using the measured perpendicular diameters ‘d’ of the ellipsoidal tumours and the equation: . Following treatment, tumours were measured every other day until the volumes were three times the pretreatment volumes. The length of time from treatment until the tumour reached three times its pretreatment volume was the primary study end point for study efficacy.

mNP injection and dosimetry

The mNPs used in this study were bionised nanoferrite (BNF) nanoparticles (MicroMod, Rostock, Germany), suspended in water without any surfactants, and heated via magnetic hysteresis when an AMF is applied.

The mNP were suspended at a total mNP concentration of 42 mg/mL (28 mg of Fe/mL). Intratumoral injections, in four equal quadrants, were performed using two needle tracks, at a dose of 7.5 mg of Fe per cm³ of tumour. The average total dose of Fe per mouse was 1.2 ± 0.2 mg Fe per mouse (0.05 ± 0.006 mg Fe per gram body weight). AMF activation was performed 10 min following injection.

Administration of AMF

The AMF field was generated by a water cooled, whole body circular coil (Fluxtrol, Auburn Hills, MI; Figure 2). The 5.0-cm long coil is comprised of 8-mm square tubing with five turns resulting in an internal diameter of 3.6 cm. The coil was powered by a Huttinger TIG 10/300 generator (Freiburg, Germany), which produces an AMF field of 165 kHz and 450 Oe (35.8 kA/m). Both the coil and the generator were cooled with water at 30 °C (chiller TKD250, Tek-Temp Instruments, Croydon, PA). Mice were treated under anaesthesia using 1–3% isoflurane gas and 95% O₂ with an average rectal temperature of 37.5 °C ± 0.5 °C.

Administration of microwave

The microwave applicator consisted of an open-ended pair of coaxial conductors, driven by a 915 MHz microwave generator and cooled by circulating water. The applicator was sized to fit the mouse flank tumour (Figure 3). A water-based tissue-equivalent coupling gel, which accommodated tumour geometry variations, was placed between the applicator and the tumour surface [38]. Mice were treated under anaesthesia with an average rectal temperature of 35.7 °C ± 1.5 °C.

Temperature recording and thermal dose

Tumour and mouse core (rectal) temperatures were measured throughout the treatment at the centre of the tumour. CEM values were calculated continuously in real-time. FISO fibre-optic probes (FISO, Quebec, Canada) and FISO Evolution software assessed temperatures at a rate of 1 Hz and continuously updated the measured temperatures and CEM values. Treatments were terminated when the thermal dose reached CEM60.

The measured thermal histories for tumours treated with mNPH and 915 MHz were similar (Figure 4). The majority of tumours for both treatment types achieved a thermal dose of CEM60 within 20 min of treatment (12 tumours for mNPH and 11 tumours for 915 MHz hyperthermia). Some tumours took longer than 20 min to achieve a thermal dose of CEM60 (three for mNPH and six for 915 MHz hyperthermia). Minor tumour geometry and/or mNP biodistribution variations resulted in slightly different heating rates (relationship of heating time and temperature). These differences did not meaningfully affect treatment efficacy. Some variation between treatment groups was also apparent both at the initiation and conclusion of treatment. The surface applicator used for 915 MHz microwave hyperthermia was water cooled. This design results in the initial tumour temperatures being slightly lower for the microwave-treated tumours than for the mNP treated tumours. Additionally, mNPH treated tumours retained an elevated temperature slightly longer than 915 MHz treated tumours once the field was removed.

Efficacy treatment groups

Four groups were evaluated for treatment efficacy (time to tumour regrowth) including: (1) mNP + AMF (n = 6), (2) 915 MHz microwave (n = 8), (3) AMF alone, and (4) no treatment (n = 6). Groups 1, 3 and 4 are included in an accompanying manuscript (this volume, pp. 845–51). All tumours receiving hyperthermia were treated to CEM60. Animals treated with 915 MHz microwave hyperthermia or AMF alone received an intratumoral injection of phosphate buffered saline (PBS), at a volume equivalent to the prescribed volume of mNP. AMF control animals were exposed to 450 Oe (35.8 kA/m), 165 kHz for 30 min.

Efficacy end points and statistics

Regardless of the study arm, all mice were sacrificed when the tumour volume reached three times treatment volume. This information was analysed for statistical significance with the two-tailed, two-sample t test (ttest2 function) present in Matlab software (version R2011a, Mathworks, Natick, MA).

Histological evaluation

To determine the histopathological effects of the mNPH and microwave treatments, a subset of tumours (mNPH = 9, 915 MHz microwave hyperthermia = 9) were evaluated histopathologically 24 h following treatment. The tumours were fixed in neutral buffered 10% formaldehyde and processed for standard histological slide preparation. Histological sections were cut 4 μm apart and stained with haematoxylin and eosin (H&E). Qualitative and quantitative histological analysis and photomicroscopy were performed using conventional light microscopy and the NIH open source ‘Image J’ quantification software (National Institutes of Health, Bethesda, MD).The tumour treatment effects (live versus necrotic) was determined by tracing the total tumour area and the area of necrosis. The level of normal tissue injury beneath the tumour was determined by establishing zones 0.75, 1.5 and 2.25 mm beneath the tumour (Figure 5). Each zone was morphologically evaluated for the presence of oedema, haemorrhage and muscle necrosis.

Results

mNP-treated tumours had a 31% increase in regrowth delay (21 versus. 16 days) as compared to microwave-treated tumours at the same thermal dose (Figure 6). Student’s t tests demonstrated statistical significance of p = 0.09.

The mNP treatments demonstrated large regions of necrosis as well as focal necrosis and haemorrhage. Although tumours were largely necrotic 24 h following treatment, small regions of morphologically viable cells were noted near the tumour boundary. Although unclear from microscopic observation, these regions of viability were likely associated with low levels of mNP coverage. Moderate peritumour oedema was also noted in a few mNPH-treated tumours. However, peritumour muscle necrosis and haemorrhage was not observed (Figure 7).

Table I. Summary of post-treatment tissue damage immediately beneath the tumour. Tissue was evaluated for the presence of oedema, muscle necrosis and haemorrhage. Zones were 0.75 mm thick. Zone 1 began at the deep edge of the tumour. If no damage was observed, ‘normal tissue’ was reported.

Mouse	Zone 1	Zone 2	Zone 3
mNPH 1	Oedema	Oedema	Normal muscle
mNPH 2	Oedema	Normal muscle	Normal muscle
mNPH 3	Oedema	Normal muscle	Normal muscle
mNPH 4	Oedema	Normal muscle	Normal muscle
mNPH 5	Oedema	Normal muscle	Normal muscle
mNPH 6	Normal muscle	Normal muscle	Normal muscle
mNPH 7	Normal muscle	Normal muscle	Normal muscle
mNPH 8	Oedema	Oedema	Normal muscle
mNPH 9	Oedema	Normal muscle	Normal muscle
915 MHz microwave 1	Oedema and haemorrhage	Oedema and Necrotic muscle	Necrotic muscle
915 MHz microwave 2	Oedema	Normal muscle	Normal muscle
915 MHz microwave 3	Haemorrhage	Necrotic muscle and Oedema	Normal muscle
915 MHz microwave 4	Oedema and Necrotic muscle	Normal muscle	Normal muscle
915 MHz microwave 5	Necrotic muscle	Necrotic muscle	Normal muscle
915 MHz microwave 6	Oedema and Necrotic muscle	Oedema and Necrotic muscle	Normal muscle
915 MHz microwave 7	Necrotic muscle	Necrotic muscle	NA
915 MHz microwave 8	Necrotic muscle	Necrotic muscle	NA
915 MHz microwave 9	Necrotic muscle	Necrotic muscle	Necrotic muscle

Tumours treated with 915 MHz microwave hyperthermia demonstrated distinct regions of necrosis and cell viability. Viable cells were typically seen in the superficial and lateral tumour region. Additional histological evaluation of the surrounding normal tissue noted the presence of peritumour oedema, mild inflammation and underlying muscle necrosis (Figure 8).

Histological evaluation of the mNP and microwave-treated tumour demonstrated distinctly different patterns of necrosis and viability (Table 1). Although the measured thermal dose for mNPH and microwave treatments was approximately the same, the presence of scattered focal zones of cell damage within an individual tumour following mNPH suggests that the mNP distribution and resulting tumour heating was not uniform (Figure 9). This information highlights the importance of mNP dose and biodistribution in the tumour with respect to delivered thermal dose, treatment effect and safety.

Figure 1. TEM image of mNP associated with murine mammary adenocarcinoma (MTGB) tumour cells. This section was acquired 5 min following delivery of 7.5 mg Fe/cm³ directly into the MTGB tumour mass. mNP (black specs) are located between cells (interstitial space), attached to the exterior aspect of the cell plasma membrane or in membrane associated vesicles. Scale bar = 2 μm. Qualitative evaluation of tumours (n = 4), harvested at the prescribed treatment time, indicate that the mNP distribution at the initiation of AMF is dominated by extracellular mNP.

Figure 2. This figure demonstrates the techniques used to deliver the alternating magnetic field (AMF) to a tumour bearing mouse. The AMF field was generated by a water cooled, whole body circular coil which produced an AMF field of 165 KHz and 450 Oe (35.8 kA/m). The location of the flank-based tumour within the coil is indicated by the red encircled dot. This region demonstrates a near uniform, homogenous AMF. The tumour temperatures and core temperature of the mouse were measured throughout the treatment using real time fibre-optic thermometry. The thermal history of the treatment was continuously calculated and displayed for the tumour and body core throughout the experimental period.

Figure 3. The microwave applicator used for tumour treatments consist of an open-ended pair of coaxial conductors, which was driven by a 915 MHz microwave generator, cooled by circulating water and sized to fit over the flank tumour of the mice. Tissue-applicator coupling gel was placed between the applicator and the tumour surface. Tumour temperatures and core temperature of the mice were measured throughout the treatment, with thermal history (CEM) calculated in real-time.

Figure 4. This graph demonstrates the tumour heating characteristics for the mNPH and microwave treatment modalities. All tumours received a thermal dose equal to 60 min at 43 °C (CEM60). Minor tumour geometry and/or mNP biodistribution variations resulted in slightly different heating rates (relationship of heating time and temperature). These differences did not meaningfully affect treatment efficacy.

Figure 5. This figure represents the histological parameters used to assess a treated mouse tumour and associated normal tissue deep to the tumour. Zones 1, 2 and 3 were used to determine and quantify the morphological tissue response at various depths beneath the tumour following treatment. Each zone had a thickness of 0.75 mm. Zone 1 started at the deep edge of the tumour. A quantitative cross-sectional area method was used to determine the relative amount of viable versus non-viable tumour area. The presence of normal tissue/muscle necrosis, peritumoural oedema and/or haemorrhage was categorised for each of the three zones for each tumour.

Figure 6. Kaplan-Meyer curve demonstrating survival/tumour regrowth for each of the four treatment groups (no treatment, AMF, mNPH and 915 MHz microwave hyperthermia). Tumours were treated at 150 mm³± 40 mm³. Animals were removed from the study when tumour volumes reached a three-fold treatment volume increase. Treatment with mNP hyperthermia resulted in a slightly more effective treatment than 915 MHz microwave hyperthermia (p = 0.09, 5 days).

Figure 7. These photomicrographs acquired 24 h post mNPH treatment, demonstrate the morphological changes observed in two different tumours which received an identical thermal dose (CEM60). Both tumours show extensive uniform necrosis. The lower tumour demonstrates significant oedema in the space between the tumour and underlying muscle (zones 1 and 2). Although minor inflammation was seen, including mNP-containing macrophages, there was no haemorrhage or necrosis in the overlying skin or peritumoral tissue.

Figure 8. Photomicrographs of two mouse mammary tumours 24 h post 915 microwave treatment. These photomicrographs show similar effects consisting largely of generalised central tumour necrosis. The superficial tumour is spared, likely due to applicator cooling. The high magnification (10×) photomicrographs of muscle tissue deep to the tumour demonstrate varying degrees of muscle necrosis, haemorrhage and inflammation. The majority of the changes are seen in zones 1 and 2. See Figure 5 for information on tissue damage quantification techniques.

Figure 9. A representative MTGB tumour with histological sections taken 5 min after mNP injection. These H&E and Prussian blue photomicrographs demonstrate regional heterogeneity in mNP distribution. The 110 nm diameter mNP were injected in four tissue quadrants, with a total of 7.5 mg of Fe per cm³ tumour, 5 min prior to tumour removal and processing. Regions indicated on the low magnification images are shown in the high magnification images. H&E and Prussian blue. (A, B) 10 × magnification, (C, D, E, and F) 100 × magnification.

Figure 10. This graph demonstrates a comparison of necrotic tumour area 24 h following mNPH and microwave treatment (CEM60). Histological evaluation of tumours treated with 915 MHz microwave hyperthermia show 75% of the tumour area is necrotic in comparison to mNPH (79% necrotic). Statistical significance between groups was not found (p = 0.55). Error bars represent standard deviation. All tumours were evaluated 24 h following treatment.

Discussion

mNP delivered via direct injection to the tumour volume was distributed to the majority of the tumour primarily along fascial planes, providing a controlled, effective and reproducible thermal dose. While the mNPH and microwave hyperthermia treatment effects and thermal dose are similar for our mouse tumour model, the data suggest that mNPH is slightly more effective (p = 0.09) from a tumour regrowth analysis standpoint. Since the AMF exposure was completed shortly following mNP delivery, the degree of cellular uptake was small. Therefore, the generated heat is primarily extracellular, although further investigation and quantification of mNP uptake or distribution changes during AMF exposure is necessary. Whether the molecular targets of nanoparticle-based hyperthermia are different, or can be made different, using the intracellular potential of mNP, remains unclear and will require additional studies utilising well documented mNP location parameters [39].

Although histologic evaluation of the treated tumour tissues indicated relatively minor differences in the level of tumour necrosis/treatment efficacy for mNPH and 915 MHz microwave hyperthermia (Figure 10), significantly less peritumour normal tissue damage was observed in tumours treated with mNPH. Since normal tissue effects limit the tolerated dose of all cancer therapies, our results suggest that the improved geometric confinement of mNP hyperthermia to the tumour volume may offer significant advantages over conventional local hyperthermia techniques, where the energy waves are not greatly confined by various tumour and non-tumour tissue boundaries.

While our study suggests that when mNP are primarily extracellular at the initiation of treatment, that mNPH may not be significantly more efficacious than ‘traditional’ hyperthermia of the same thermal dose, it does indicate that mNPH may be more effective at delivering a uniform thermal dose, to a target volume. As the specific geometry of each clinical tumour is unique, this study also indicates that mNP may be better at sparing the sensitive and often dose-limiting normal tissue which surrounds a tumour. If mNP can be further targeted to tumour and tumour cells through tumour antibody-directed mNP, and/or static magnetic fields, the therapeutic ratio may be further improved [40]. Furthermore, the development and use of mNP imaging prior to activation will not only allow for real-time determination of mNP distribution, but will also allow for improved treatment planning [41,42]. If accurate and reproducible mNP treatment planning, based on mNP SAR, mNP concentration, location and field strength can be realised, the dependence on invasive thermal measurements may be reduced or even eliminated. This situation would enable more cost and time efficient treatments without sacrificing safety or efficacy.

Conclusion

These experiments demonstrate a similar therapeutic effect in a mouse mammary adenocarcinoma model for locally delivered intratumoral mNPH at 7.5 mg Fe/g tumour and 915 MHz microwave hyperthermia at the same thermal dose (CEM60). This finding suggests that, for acutely activated, locally delivered mNP, the CEM relationship remains a valid method for prescribing a tumour treatment and delivering a quantifiable and meaningful thermal dose. Similar studies have not yet been conducted for intracellular mNPH. Additionally, the accurate measurement of intracellular mNP, following AMF activation, remains challenging. With these factors in mind, it remains possible that significant thermal effects in tumours can be achieved at measured temperatures and thermal doses previously believed to be too low to result in tumour cytotoxicity. Following treatment, histopathological examination of tumour tissue and the surrounding tissue show reduced normal tissue damage for mNPH as compared to 915 MHz hyperthermia. As such, we propose mNPH has the potential for improved clinical safety and efficacy, as compared to conventional (e.g. microwave, radiofrequency) local hyperthermia treatment techniques.

Declaration of interest

This work was supported by the Dartmouth Center of Cancer Nanotechnology Excellence (National Institutes of Health National Cancer Institute grant no. 1U54CA151662-01). A.A.P. and A.J.G. gratefully acknowledge support from the Thayer School of Engineering Innovation Fellowship.

Acknowledgements

We would like to thank Rendall Strawbridge for his assistance with the animal model and Louisa Howard for her TEM imaging contributions.