Molecular Imaging and Biology

, Volume 15, Issue 3, pp 299–306

Functionalized Magnetonanoparticles in Visualization of Intracranial Tumors on MRI

Authors

    • Jane and Terry Semel Institute for Neuroscience and Human Behavior, David Geffen school of MedicineUniversity of California
  • Whitney Pope
    • Department of Radiology, David Geffen School of MedicineUniversity of California
  • Gary Mathern
    • Department of Neurosurgery, David Geffen School of MedicineUniversity of California
  • Rex Moats
    • Department of Radiology, Children’s Hospital of Los AngelesUniversity of Southern California
  • Andrew Frew
    • Department of Neurosurgery, David Geffen School of MedicineUniversity of California
  • Mark Mandelkern
    • Department of PhysicsUniversity of California
Research Article

DOI: 10.1007/s11307-012-0601-z

Cite this article as:
Akhtari, M., Pope, W., Mathern, G. et al. Mol Imaging Biol (2013) 15: 299. doi:10.1007/s11307-012-0601-z

Abstract

Purpose

The development of nonradioactive and targeted magnetonanoparticles (MNP) capable of crossing the blood–brain barrier (BBB) and of concentrating in and enhancing the contrast of intracranial tumors on magnetic resonance imaging (MRI).

Procedure

Nonradioactive 2-deoxy-d-glucose (2DG) was covalently attached to magnetonanoparticles composed of iron oxide and dextran and prepared for intravenous (tail) injection in the naïve rats and mouse models of glioma. MR images were acquired at 3 and 7 T.

Results

2DG-MNP increased tumor visibility and improved delineation of tumor margins. Histopathology confirmed that 2DG-MNP crossed the BBB and accumulated within brain parenchyma.

Conclusion

Nonradioactive 2DG-MNP can cross an intact BBB on and improve visualization of tumor and tumor margins on MRI.

Key words

MagnetonanoparticlesTargetedGlioma2-DeoxyglucoseMRIContrast

Introduction

Gliomas are the most frequent and deadly malignant primary brain tumors in adults, with an annual incidence of approximately five to six cases per 100,000 [1]. More than half of all cancer deaths from primary brain tumors are caused by gliomas. Classification by grade includes low- and high-grade gliomas, LGG and HGG, respectively. Even LGG are often fatal as they progress to HGG in most patients. HGG survival rates are approximately 50 and 25 % after 1 and 2 years, respectively.

Tumor contrast enhancement is of primary importance in the diagnosis and determining treatment response of gliomas. Most (>90 %) low-grade gliomas are radiographically non-enhancing [2]. High-grade gliomas are infiltrative, and margins may be poorly defined on magnetic resonance imaging (MRI) due to small size or the presence of intact blood–brain barrier (BBB) and blood–tumor barrier (BTB), and therefore may not show enhancement following gadolinium chelate injection. Absence of clear margins may contribute to partial surgical resection and recurrence of the tumor. Tumor contrast enhancement is the basis for assessing response to therapy per the Macdonald and Revised Assessment in Neurooncology (RANO) criteria. For example, if the degree of contrast enhancement increases significantly (≥25 %) during therapy, the tumor is deemed to have progressed, often warranting a change in therapy. With clinically available MRI contrast agents, which do not cross the intact BBB and BTB, contrast enhancement is not entirely specific for tumor, as any process that disrupts the BBB will result in enhancement. For instance, radiation changes that lead to contrast enhancement may be mistaken for tumor progression (so-called pseudoprogression). Administration of antiangiogenic therapy can diminish contrast enhancement without reflecting underlying changes in tumor burden (so-called pseudoresponse) [3].

Standard (T2/FLAIR) and advanced MRI techniques such as perfusion- and diffusion-weighted MRI as well as MR spectroscopy have been used to better differentiate intracerebral tumors and to assess tumor response to therapy [4, 5]. In addition, several radioactive ligands have also been used in positron emission tomography (PET), sometimes co-registered with MRI, for tumor characterization. Among these are 2-deoxy-2-[18F]fluoro-d-glucose (18FDG), and amino acid tracers such as O-(2-18 F-fluoroethyl)-l-tyrosine, 3,4-dihydroxy-6-18 F-fluoro-l-phenylalanine, and 11C-Methionine [6]. However, many radioactive ligands have short half-lives and require on-site synthesis, making their availability limited. Repeated exposure of patients and caregiver to radiation is an additional concern.

Recent advances in nanotechnology have led to the design and synthesis of various nonradioactive magnetonanoparticles (MNP), which change the magnetic environment of the surrounding protons and alter T1 and/or T2 signal in a concentration-dependent manner once in tissues [711]. Recent studies on 2-deoxy-d-glucose (2DG)-conjugated gold nanoparticles (AuNP) by Aydogan et. al. [12] “strongly suggest enhanced uptake of the AuNP-DG over the unlabeled AuNP by the highly glycolytic cancer cells in vitro and indicate that AuNP-DG could serve as a CT contrast agent with tumor targeting capability”. In recent studies by Akhtari et al. [7], 2DG-MNP was injected in the naïve rat, and histology showed that the contrast agents crossed the intact BBB and accumulated in the brain parenchyma. In their studies, 2DG-MNP was also injected during electrical stimulation of the whiskers in naïve rats, and subsequent contrast enhancement on MR images showed good agreement with published 14C-2DG autoradiography data. Quantitative comparison of MRI contrast enhancement due to 2DG-MNP in rodent models of epilepsy also agreed with published data on 14C-2DG autoradiography, and the pattern of uptake also agreed with that of 18FDG-PET. The authors also compared MRI contrast enhancement due to 2DG-MNP particles to that of plain (unconjugated) MNP in an animal model of epilepsy and showed that functional neuroanatomy was visible with 2DG-MNP but not with plain particles in the rat [7].

We have developed a nonradioactive platform technology with the use of nanotechnology to covalently attach 2DG to MNP visible on MRI [7]. These particles cross the BBB and localize to tumor tissues. MR images show that 2DG-MNP provides clear differential enhancement of tumors and their margins, which could help to overcome some of the limitations of the currently used response criteria. The development of these MRI contrast agents which can cross the BBB and accumulate in the tumor could advance our ability to identify small tumors, gliomas, and tumor infiltration of adjacent normal brain.

Some of the experimental methods and results presented here are adopted from previous publications by this group on different applications of this technology and are intended to maintain self containment in this manuscript.

Methods

Particles

2DG was covalently conjugated to the plain particles via a glycerine linker [7, 9, 13]. The particles were characterized using transmission electron microscopy (TEM) with energy-dispersive X-ray spectroscopy (EDAX; Philips, Germany), atomic force microscopy, as well as zeta potential (Zetasizer 300HSA, Malvern Instruments, Malevrn UK).

Animals

All procedures were approved by the University of California, Los Angeles, Institutional Animal Care and Use Committee. A total of six animals were used in this study. Three nude–nude mice were used for xenograft studies. A mouse was used in the phosphatase and tensin homolog (PTEN)-deleted model of glioma. One naïve rat was used to demonstrate presence of these contrast agents in brain tissues. A second naïve rat was used to demonstrate contrast enhancement at a clinical MRI field strength of 3 T.

Dosing and Route of Administration

Contrast agent Injection

All animals were denied food for at least 12 h before being injected with contrast agent (3.5 mg Fe/kg; mg Fe refers to the weight of the hydrous iron oxide core of the magnetonanoparticles) intravenously via the tail vein.

MRI Animal Studies

Scanners

The 7 T Bruker Biospin MRI scanner was used with the following scan parameters: T2, 6,000 repetition time (TR)/10–120 echo time (TE), 1–12 Spin Echo (SE), 148 × 148–192 × 192, 25 slices, 0.7–1 mm slice thickness, 0 mm interslice distance, 2–4 cm field of view (FOV). Acquisition time was 19 min. The specific sequences and parameters are described below for each experiment. Each animal had at least one baseline (pre-contrast) and one post-contrast MR scan, as previously described [7].

The 3 T Siemens Trio MRI scanner was used with the following parameters: T2 sequences, 5,000 TR/12–59 TE, 6 echos, 192 × 192, 8 slices, 2 mm slice thickness, 0 mm interslice distance, 4 cm FOV, 10 averages, 19 min acquisition time. Images were acquired in one resting naïve rat, before and after contrast agent injection, to assess contrast enhancement at clinical MRI field strengths [7].

Resting Brain Activity Studies

(1) Baseline MR images (7 T, T2, TR 6,000 ms, TE 10–120 ms, 12 SE) were obtained in one naïve rat under isoflurane anesthesia. Following injection of 2DG-MNP (3.5 mg Fe/Kg), MR images were obtained 2.5 h post-injection. This rat was euthanized an hour after conclusion of the scans, its brain tissue was removed and prepared for histology [7]. (2) A second naïve rat was anesthetized with intramuscular ketamine cocktail (0.9 ml) and anesthesia-maintained with subsequent injection of 0.2 ml every 15 min, with total dose of no more than1.5 ml. After baseline images were obtained on a 3 T human MRI scanner, 2DG-MNP (3.5 mg Fe/kg) was injected and scans were repeated at 20-min intervals for 120 min, as previously described [7].

Tumor Imaging Studies (3 Mice)

  1. (1)

    A nude mouse with U87 RLUC xenograft (5 weeks post-implant) was anesthetized with isofluorane gas, and baseline MR images (7 T, single echo T2, TR 6,000 ms, TE 50 ms) were obtained. The mouse was then injected with 2DG-MNP (3.5 mg Fe/kg), and MR images were obtained at 3, 5, and 24 h after the injection.

     
  2. (2)

    A second nude mouse with U87 RLUC xenograft (2 weeks post-implant) was anesthetized with isofluorane gas and baseline MR images (7 T, single echo T2, TR 6,000 ms, TE 50 ms) were obtained. The mouse was then injected with 2DG-MNP (3.5 mg Fe/kg), and MR images were obtained 5 h after the injection.

     
  3. (3)

    A third mouse (120 days old) with PTEN-deleted genetic model of glioma was anesthetized with isofluorane gas, and baseline MR images (7 T, T2, TR 6,000 ms, TE 10–120 ms, 12 SE) were obtained. The mouse was then injected with 2DG-MNP (3.5 mg Fe/kg), and MR images were obtained at 2, and 7 h after the injection.

     

Definition of NCE

The paramagnetic property of this class of contrast agents causes a decrease in T2 (and T1) proton relaxation time on MRI, leading to decreased T2 ( and increase T1) signal intensity of the tissues in which these agents are present; hence, “negative contrast enhancement” (NCE) is used to describe T2 signal changes. This decreased signal intensity is similar to darker tissues on film in 14C-2DG autoradiography, as described previously [7].

Quantitative Measurement of NCE

Standard Bruker software was used to measure T2 values on the corresponding regions of interest (e.g., hippocampus, cortex, midbrain, and entorhinal cortex) on multi-echo scans with relationship, \( A=c{e^{{-\tfrac{t}{\mathrm{T}2}}}} \), where A(ti), c, and T2, denote signal intensity at echo times ti, amplitude (NCE), and relaxation search parameters, respectively [7].

Histology

The brain of the resting naïve rat was removed after perfusion with 4 % solution of paraformaldehyde and fixed in a paraffin block. Diaminobenzidine-enhanced Perl’s iron staining was performed as described previously [7].

Tumor Surgery

Sulfamethoxazole (5 mg/ml)/Trimethoprim (1 mg/ml) was administered in drinking water for 2 days prior to and 5 days following intracranial surgery. Animals were anesthetized with isofluorane gas (5 %). After insuring sufficient depth of anesthesia to prevent response to foot-pinch and ear-bar placement without reflex paw movement, the animals were placed in a stereotaxic frame, resting on a heating pad covered by a sterile drape. After clipping the scalp fur, the skin of the incision sites were disinfected with betadine, the analgesic Carprofen (5 mg/kg) was administered subcutaneously, and the head and neck areas were covered with a sterile drape. The eyes were protected with ophthalmic ointment. A 2-cm skin incision was made, and the periosteum was scraped away from the region of interest. Tissue from the periosteum was removed with sterile cotton tipped applicators and sterile saline. The stereotactic apparatus was used to calculate coordinates (2 mm lateral, 0.5 anterior, and 3.3 down from bregma-lambda level) to drill one hole to allow the injection into the striatum. A dentist’s drill was used to open the hole for injection. Once the hole was drilled, the dura was identified and cut. A 10-μl Hamilton syringe fitted with a 26 gauge needle that was clamped to the stereotactic frame was loaded with the appropriate cell lines. A total of 2.5 μl of the cell line (U87 RLUC, approximately 100,000 cells/1 μl) was stereotactically injected at the coordinates listed above at 1.0 μl per min. The needle was left in place for 4 min and was then slowly withdrawn. The skin was then closed using 3–0 nylon sutures. The animals were removed from the stereotactic apparatus and allowed to recover. The animals were returned to sterile cages after full recovery. Sutures were removed 7–10 days post surgery. Animals were housed in autoclaved cages with sterilized water, bedding, and irradiated diet. They were housed at humidity 30–70 % and 12 h/12 h light/dark cycle.

PTEN-Deleted Mouse Glioma Model

PTEN-deleted mouse model of glioma was generously provided by Dr Hong Wu at the David Geffen School of Medicine.

Toxicology

Nanotechnology Characterization Laboratories at the National Cancer Institute evaluated the toxicology profile as described in prior publication by this group [7].

Statistics

The ratios and T2 values of pre-contrast (pr) and post-contrast (po) contrast NCE, as well as the corresponding standard deviations were calculated for tumors using:
$$ \delta \rho ={{\left[ {{{{\left( {\left( {\frac{{\partial \rho }}{{\partial \mathrm{pr}}}} \right)\delta \mathrm{pr}} \right)}}^2}+{{{\left( {\left( {\frac{{\partial \rho }}{{\partial \mathrm{po}}}} \right)\delta \mathrm{po}} \right)}}^2}} \right]}^{{\frac{1}{2}}}}={{\left[ {{{{\left( {\left( {\frac{1}{\mathrm{pr}}} \right)\delta \mathrm{po}} \right)}}^2}+{{{\left( {\left( {\frac{\mathrm{po}}{{\mathrm{p}{{\mathrm{r}}^2}}}} \right)\delta \mathrm{pr}} \right)}}^2}} \right]}^{{\frac{1}{2}}}} $$
[14], with ρ = (pr − po) / pr, and where δpr and δpo denote the corresponding standard deviation.

Results

Figure 1 shows the TEM image of particles. Monodispersity of particles was established through TEM and FEM; the average diameter of the particles was approximately 10 nm with an average iron oxide core diameter of approximately 3 nm. The zeta potential of these particles indicated stability and was −2.4 mV; the pH of these particles was 7.4 in saline.
https://static-content.springer.com/image/art%3A10.1007%2Fs11307-012-0601-z/MediaObjects/11307_2012_601_Fig1_HTML.gif
Fig. 1

a, b Transmission electron microscopy. a shows single particles (arrow) and aggregates (arrowhead) which occur during the drying of the unstained sample on grid and b shows the morphology and size (scale at the bottom of panel) of individual 2DG-MNP particles (negative staining with 1 % uranyl acetate). Most particles measured approximately 10 nm.

Figure 2a shows the post-contrast MR image obtained in the resting naïve rat. Figure 2b–d shows the uptake of contrast agent (red arrows) in the presence of intact BBB; the iron particles are present in intra (cytosomal)—as well as extra (interstitial)—cellular brain tissues (×40) (Fig. 2b–d) [7].
https://static-content.springer.com/image/art%3A10.1007%2Fs11307-012-0601-z/MediaObjects/11307_2012_601_Fig2_HTML.gif
Fig. 2

ad MRI of a resting naïve rat, 1 h after injection with 2DG-MNP (3.5 mg/kg, iv, tail) is shown in a. DAB-enhanced Perl’s iron stain of this brain shows 2DG-MNP crosses the BBB and localizes in intra- and extra-cellular brain regions (red arrows) in b somatosensory cortex, c dentate gyrus, and d hippocampus CA3. Magnification × 40.

Figure 3a shows baseline T2-weighted MR image of a PTEN-deleted genetic model of glioma in mouse. Figure 3b shows tumor, and tumor margin visualizations are significantly improved following the tail vein administration of 2DG-MNP (3.5 mg Fe/kg), (Fig. 3b). Tumor T2 values decreased by 10.2 (±1.2) % (Table 1) and the margin T2 values decreased by 22.4 (±1.8) %. Figure 3c-e shows the volumetric reconstruction of brain (a–c) and tumor (insets) tissues shown in Fig. 3b. The tumor measured 13 mm3 in volume.
https://static-content.springer.com/image/art%3A10.1007%2Fs11307-012-0601-z/MediaObjects/11307_2012_601_Fig3_HTML.gif
Fig. 3

a Pre-contrast and b post-contrast (2DG-MNP) images of a mouse spontaneous glioma model. Multi-echo T2-weighted images show only subtle signal change in a, making tumor difficult to detect. In b, the margins of the tumor are well delineated by the contrast enhancement. ce 3D reconstruction of tumor volume following contrast administration is shown (ce represent three views of the same tumor). Contrast administration improves volumetric measurements due to improved delineation of tumor margins. This may provide substantial benefit in the measurement of non-enhancing tumor, currently a controversial subject in clinical neuro-oncology.

Table 1

Examples of T2 and NCE (amplitude) values and the respective standard deviations are provided for images in Figs. 3 and 4, respectively

Figure number

Pre-contrast (±SD)

Post-contrast (±SD)

3 (T2)

51.0 ± 0.9

45.8 ± 0.6

4 (NCE)

169 ± 7

99 ± 10

Figure 4a shows a baseline MRI scan of a mouse 5 weeks after U87 xenograft. Figure 4b–d show MRI scans obtained at 3, 5, and 24 h after tail vein injection of 2DG-MNP (3.5 mg Fe/kg). Figure 4b, c shows enhancement of solid viable tumor after contrast agent administration. Contrast administration improves visualization of tumor margins but also demonstrates uptake within the central aspect of the tumor, indicating viable tissue, rather than necrotic or cystic changes as might be suggested by the pre-contrast images. Figure 4d shows that contrast enhancement clears within 24 h of administration. The comparison of pre-contrast to post-contrast (Fig. 4b) tumor images showed 41.4 (±6.4) % NCE (Table 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs11307-012-0601-z/MediaObjects/11307_2012_601_Fig4_HTML.gif
Fig. 4

ad MR image of U87 xenograft, a no contrast; and with 2DG-MNP contrast at b 2 h; c 6 h; and d 24 h post-contrast injection. Although tumor appears cystic or necrotic in a with central T2 hyperintenstiy, relatively homogeneous uptake of contrast by tumor in b and c reveal a solid tumor with well-defined margins. The contrast clears (95 %) the brain in 24 h.

Figure 5a shows a baseline MRI scan of a mouse 5 weeks after U87 xenograft. Figure 5b shows MRI scans obtained 5 h after tail vein injection of 2DG-MNP (3.5 mg Fe/kg). Figure 5b clearly shows tumor and margin enhancement; the leading edge of the tumor is well demarcated after contrast enhancement, potentially resulting in improved detection of tumor invasion of adjacent normal brain. The comparison of pre-contrast to post-contrast (Fig. 5b) tumor images showed 41.9 (±12.3) % NCE.
https://static-content.springer.com/image/art%3A10.1007%2Fs11307-012-0601-z/MediaObjects/11307_2012_601_Fig5_HTML.gif
Fig. 5

a, b MRI scans of U87 xenograft in mouse brain, a before and b after 2DG-MNP contrast enhancement. The interface between tumor and normal brain is difficult to detect in a, but clearly delineated by a well define rim of low signal following contrast administration (b). 254 × 190 mm (96 × 96 DPI).

Scaling at 3 T Human Clinical MRI Scanner

Figure 6a, b shows pre-contrast T2-weighted MR images (TE = 47 ms) of the rat brain. Figure 6c, d shows selected slices 1 h post 2DG-MNP contrast. Signal intensity measurements of cortex showed 40 (±4) % NCE; midbrain showed 39 (±3) % NCE. Results show that 2DG-MNP produces significant contrast enhancement at MRI field strengths that are approved for human use [7].
https://static-content.springer.com/image/art%3A10.1007%2Fs11307-012-0601-z/MediaObjects/11307_2012_601_Fig6_HTML.gif
Fig. 6

ad shows T2-weighted MR images of the brain of a resting naive rat at 3 T. Baseline images (pre-contrast) of two consecutive slices are shown in panels a and b; the corresponding post-2DG-MNP contrast images are shown in panels c and d. The latter images show NCE in the dentate gyri (white arrows) and cortex (red arrow) consistent with the NCE observed at 7 T (Fig. 3b). NCE in the midbrain is in good agreement with midbrain uptake of 14C-2DG (Hosokawa et al. 1996).

Toxicology

These contrast agents had favorable toxicology profile. They were not sufficiently toxic to determine an IC50 value and were minimally toxic in all evaluated cell lines [7].

Discussion

The results of this study show that these nonradioactive contrast agents can delineate intracranial tumors on MRI. As was shown previously [7], these results also show that 2DG-MNP cross the blood–brain barrier, and that the NCE observed in tumors and their margins after 2DG-MNP injection is not simply due to blood flow changes.

The magnetic moment (per individual particles in MNP compared to individual chelated gadolinium molecule) of these particles is approximately 5,000 times that of gadolinium chelate, hence much lower tissue concentrations are needed for contrast enhancement.

Current methods of tumor segmentation can be hampered by poor contrast between tumor and adjacent brain; this is particularly true for tumors that do not show gadolinium enhancement. Therefore, increased conspicuity of tumor margins could increase the accuracy of volumetric measurements. This is also particularly relevant in the setting of non-enhancing or poorly enhancing tumor following treatment with antiangiogenic drugs such as bevacizumab, in which it is difficult to distinguish the anti-permeability effects (resulting in diminished contrast enhancement and diminished visualization of tumor burden) from true antitumor response. Currently, only qualitative assessments of non-enhancing tumor burden are made according to the RANO guidelines, even though oncologists agree that quantitative measurements of tumor burden are highly desirable.

This study involves only a limited number of animals, and much more research is necessary to demonstrate that 2DG-MNP, or MNPs conjugated with other ligands, could eventually be used to enhance tumor visualization with MRI in humans. Our in vivo studies have not shown any grossly observable toxicity due to these conjugated MNPs and extended in vitro toxicology studies on 2DG-MNP has also shown a favorable profile. Studies are ongoing to determine in vivo toxicology, as well as pharmacokinetics of these particles, their metabolism, and elimination [7]. Studies on plain iron-dextran magnetonanoparticles of similar size have shown “no pathological brain cell or myelin changes after direct delivery of the particles to the brain” of animals [10] and favorable safety profile in human studies [15]. Pharmacokinetic studies by other investigators [16, 17] on this class of particles (e.g., iron-dextran composition) have shown that their iron content becomes sequestered through lysosomes and that the kidneys mainly excrete their dextran content [16, 17].

The ability of 2DG-MNP to identify intracranial tumors and their margins, as well as the level of their functional activity [7] with standard MRI, would have several advantages. Most importantly, the prevalence of MRI technology worldwide, the lack of need for additional specialized equipment, and the relatively long shelf life of these contrast agents, which are nonradioactive, will increase accessibility for imaging compared to PET. Additional advantages are the excellent anatomical resolution of MRI compared with nuclear modalities and that MRI studies with these contrast agents can be safely repeated when indicated.

Acknowledgments

The authors are greatly indebted to Dr. Gevorg Karapetian and Dr. Ira Harutyunyan for their invaluable help and input with acquisition of the MR images. The authors are also indebted to Dr. Leonard Rome of UCLA School of Medicine Dean’s Office for his foresight and support of this project. This Work was supported by generous grants from Stein/Oppenheimer Award, Jane and Terry Semel Institute for Neuroscience and Human Behavior Chairman’s, Opportunity funds, and Weil Fund, as well as, Office of the Dean, David Geffen School of Medicine, Davis Fund.

Conflict of Interest

The technology used in these studies is owned by the University of California, and Dr. Akhtari has received royalty payments from the University of California during the past 3 years. None of the other authors has any conflict of interest. We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Copyright information

© World Molecular Imaging Society 2012