YS, MI and IM are employees of Otsuka Pharmaceutical Co., Ltd. SM and TI have received donations for developing new magnetic resonance imaging technology from Otsuka Pharmaceutical Co., Ltd. YS and MI have stocks of Otsuka Pharmaceutical Co., Ltd. YS, MI and IM are registered as inventors of a related patent currently pending (WO 2007/132806). Patent application name: Magnetic resonance contrast medium using polyethylene glycol and magnetic resonance image pick-up method. International application number: PCT/JP2007/059849. International publication number: WO 2007/132806. The patent has been registered at this stage in Japan (Patent number: 5435940), Taiwan (Patent number: I400091) and Europe (Patent number: EP2020244B1). In addition to the patent information, the authors would like to describe related stockholding of authors in the competing interests section as follows. There are no further patents, products in development or marketed products to declare. This does not alter our adherence to all the PLOS ONE policies on sharing data and materials, as detailed online in the guide for authors.
Conceived and designed the experiments: YS MI IM. Performed the experiments: YS SM. Analyzed the data: YS SM. Contributed reagents/materials/analysis tools: YS SM TI. Wrote the paper: YS SM. Directed the study: TI MI.
Morphological imaging precedes lesion-specific visualization in magnetic resonance imaging (MRI) because of the superior ability of this technique to depict tissue morphology with excellent spatial and temporal resolutions. To achieve lesion-specific visualization of tumors by MRI, we investigated the availability of a novel polymer-based tracer. Although the 13C nucleus is a candidate for a detection nucleus because of its low background signal in the body, the low magnetic resonance sensitivity of the nucleus needs to be resolved before developing a 13C-based tracer. In order to overcome this problem, we enriched polyethylene glycol (PEG), a biocompatible polymer, with 13C atoms. 13C-PEG40,000 (13C-PEG with an average molecular weight of 40 kDa) emitted a single 13C signal with a high signal-to-noise ratio due to its ability to maintain signal sharpness, as was confirmed by
Magnetic resonance imaging (MRI) is a powerful tool for non-invasive exploration of internal body structures with excellent spatial and temporal resolution. Its power is derived from the widely distributed, robust proton signaling from the enormous amounts of water in the body, which enables rapid and precise depiction of morphology. On the other hand, the detection of lesion sites with a specific tracer, in which the modalities of positron emission tomography (PET) and single-photon emission computed tomography (SPECT) play an important role, is also a powerful tool in imaging diagnostics, especially in the field of tumor detection. However, no tracer-based method for tumor detection using MRI is presently in wide use, possibly because only a few strategies are currently available. Therefore, the further development of tracer-based strategies with general versatility is anticipated.
The 13C nucleus appears to be a promising candidate for detection in tumor imaging because its background signal is much lower than that of protons. However, 13C spectroscopic imaging will not be possible until the inherently low sensitivity of 13C nuclei is resolved. The dynamic nuclear polarization (DNP) technique, which enables marked enhancement of the MR signal from 13C nuclei with the induction of a hyperpolarized state in the nuclei in the liquid phase
In this study, we examined an alternative method for tumor detection using 13C spectroscopic imaging. Our method utilizes a 13C-enriched polymer to overcome the low sensitivity of the 13C nucleus rather than the hyperpolarization phenomenon. It is well recognized that macromolecules selectively extravasate into tumors because of the enhanced leakiness of tumor capillaries
Among the candidate agents for use as a 13C tracer, we identified polyethylene glycol (PEG), which has a unique chemical structure consisting of a repetition of -CH2CH2O-, as a compound satisfying these conditions. On the basis of the determination of its molecular weight in previous rodent studies
13C-enriched PEG using [1,2-13C]ethylene oxide (Cambridge Isotope Laboratories, MA) as a raw material was custom designed for use in this study. 13C-PEG40,000 and 13C-PEG7,000 were synthesized by Meisei Chemical Works, Ltd. (Kyoto, Japan) and DJK (Yokohama, Japan), respectively. The polydispersity value for the 13C-PEG40,000 was 1.47.
NMR spectroscopy was conducted using a Bruker Avance III 400 spectrometer with a 9.4 T magnet. Measurements were performed with 1H irradiation to achieve 1H-13C nuclear Overhauser effect (NOE) and proton decoupling (WALTZ16). When a degree of the NOE effect was investigated, NMR measurements of 12C-based PEG40,000 were carried out with and without the 1H irradiation for NOE. Concentration was set at 50 mg/mL for the experiment. 5 mM of [1-13C]alanine was added as an internal standard.
For comparison of the 13C-NMR spectrum of PEG with those of other hydrophilic polymers, we measured intrinsic 13C signals of non-13C-enriched polymers (∼1% of carbon atoms in natural abundance) since no 13C-enriched polymer was available except for 13C-PEG40,000. PEG40,000 was purchased from NOF corporation (Tokyo, Japan); Dextran40,000, poly-L-lysine ∼50,000, PEG500,000, and Dextran200,000 from Wako (Osaka, Japan); and bovine serum IgG from Sigma. Concentrations of the smaller 12C-polymers (PEG40,000, Dextran40,000 and poly-L-lysine ∼50,000) were 5 mg/mL while the larger 12C-polymers (PEG500,000 and Dextran200,000) and IgG were 10 mg/mL, with phosphate-buffered saline (PBS), pH 7.4, including 20% D2O. All measurements were performed using a 5-mm test tube at 24°C and with a 1.36 s acquisition time, a 2.0 s relaxation delay, and a 30° flip angle (8.5 µs for 90° pulse width), and consisted of 3200 scans. 1 mM of [1-13C]alanine was used as an internal standard. Calculation of signal-to-noise ratio was conducted using Bruker TOPSPIN 2.1 software, with estimation of signal half-width manually performed using the software with an expanded plot.
Evaluation of T1 (for
Amino-terminal PEG40,000 (SUNBRIGHT MEPA-40T, NOF) and tetramethyl rhodamine (TMR) succinimidyl ester mixed isomer (Invitrogen) were used for labeling of PEG40,000 with tetramethyl rhodamine (TMR-PEG40,000). 50 mg of amino-terminal PEG40,000 and 5 mg of TMR succinimidyl ester were mixed for 20 h at room temperature in 0.5 M bicarbonate buffer, pH 8.3, and purified by Superdex75 gel filtration column. The collected TMR-PEG40,000 fraction was concentrated using Amicon Ultra centrifuge filters (cut-off Mw 10,000 Da) before use. The labeling efficiency was ∼100% as estimated by comparison of concentration of TMR with that of PEG, which were determined by intensities of absorbance and 13C-NMR signal, respectively.
C26 (murine colon adenocarcinoma 26 cell line) and Miapaca2 (human pancreas cancer cell line) were obtained from a cell line collection of Otsuka Pharmaceutical Co., Ltd. Caki2 (human Caucasian kidney carcinoma cell line) was purchased from the European Collection of Cell Cultures. T24 (human bladder cancer cell line) was supplied by the Cell Resource Center for Biomedical Research, Institute of Development, Aging and Cancer at Tohoku University. Miapaca2 and T24 were genetically authenticated using STR analysis (Promega, October 2012). C26 cells were maintained in RPMI1640, T24 and Caki2 cells in McCoy's 5A, and Miapaca2 cells in D-MEM/Ham's F-12 with 10% fetal bovine serum (FBS), 50 U/mL of penicillin, and 50 µg/mL of streptomycin (a penicillin/streptomycin mixture) in a humidified atmosphere containing 5% CO2 at 37°C.
Male BALB/c and BALB/c nude mice were purchased from Japan SLC, Inc. All procedures were performed in accordance with the guidelines of Science Council of Japan (Guidelines for Proper Conduct of Animal Experiments) and the guidelines of the Animal Care and Use Committee of Otsuka Pharmaceutical Co., Ltd, and approved by the company committee (OA0057). All surgery was performed under isoflurane anesthesia to relieve pain. All mice were housed in a specific pathogen free facility under the standard conditions recommended in the guidelines and used for experiments before the tumor size became unnecessarily large to minimize suffering. We subcutaneously inoculated 1×106∼107 C26 cells or human cancer cells (Miapaca2, Caki2, T24) suspended in Hank's Balanced Salt Solution (HBSS) into the backs of BALB/c mice or nude mice, respectively. The resulting animal models were not subjected to experimentation until the tumor volume had exceeded ∼100 mm3. The total number of mice used for this study was 59.
TMR-PEG40,000 was intravenously injected into C26 tumor-bearing mice at a dose of 250 mg/kg. The mice were sacrificed 120 h post-injection and tumors were extirpated for use in the preparation of frozen sections. Extirpated tumors were embedded in an OCT compound followed by immediate freezing using liquid nitrogen. Frozen sections were cut at a 5-µm thickness using a microtome, fixed with acetone on slide glasses, stained using 1 µg/mL of Hoechst 33342 (Dojin, Japan), and embedded using Fluoromount (Diagnostic BioSystems, USA) with coverslips. The reproducibility was checked by conducting three independent experiments.
Fluorescence microscopy of the frozen sections was performed using a KEYENCE BZ-9000 system (Osaka, Japan) with a 100× Nikon Plan Apo oil lens with a 1.4 numerical aperture.
13C-PEG was intravenously injected into BALB/c mice at a dose of 92 mg/kg. After 5 µl of blood had been collected from the tail vein at each sampling time, the samples were immediately added to 495 µl of PBS containing 10 mM of EDTA and then centrifuged at 18,000×
13C-PEG40,000 was intravenously injected into the tail vein of C26-tumor bearing mice at a dose of 92 mg/kg. After mice were sacrificed and dissected at various periods after intravenous injection, liver, kidney, spleen, pancreas, and tumor tissue were homogenized in phosphate-buffered saline (PBS). Supernatants were collected after centrifugation for 13C-PEG40,000 quantification using NMR spectroscopy. Concentration of 13C-PEG40,000 in the supernatant was estimated from analysis of the 13C-NMR spectra by comparison of the signal intensity of 13C-PEG with that of 1 mM [1-13C]alanine.
13C-CSI data were obtained using the same field of view (FOV) used for 1H imaging (100×50 mm2) in the coronal plane without slice selection. Free induction decay (FID) data were acquired under 1H irradiation to achieve 1H-13C NOE and 1H decoupling, with a 250-ms repetition time (TR) and 64 acquisitions for 16×8 phase encoding steps. Data processing by 3D Fourier transformation with zero-filling and magnitude calculation allowed for achievement of 64×32 power spectra and subsequent construction of 2-dimensional PEG images with 64×32 matrices by integration of the peak areas. Each 13C image was independently created using an 8-bit scale under the assumption that a pixel of the lowest intensity in a 13C image had an intensity of 0 and a pixel of the highest intensity an intensity of 255. The total acquisition time of one data set was 34 min. The reproducibility was checked by several independent experiments for each tumor model (n = 6 for C26 tumor model, n = 4 for Miapaca2 model, n = 3 for Caki2 and T24 models, respectively).
The molecular weight setting used for 13C-PEG synthesis was based on 3 previous findings. First, the accumulation of macromolecules in tumor requires a prolonged period of circulation
13C-PEG40,000 exhibited a very characteristic 13C-NMR signal, specifically a single, strong, and sharp signal at 69.6 ppm (
(A) 13C spectrum of 1 mg/mL of 13C-PEG40,000 (25 µM). 1 mM [1-13C]alanine was used as an internal standard.
The strong signal of PEG largely depended on the Nuclear Overhauser Effect (NOE) from protons to carbon atoms as shown in
Interestingly, signal sharpness was maintained among PEG molecules of much larger mass, specifically those with an average molecular weight of 500,000 Da (signal half-width, 2.0±0.07 Hz, n = 4), indicating that the flexibility of each carbon atom in a PEG molecule is maintained, regardless of its mass in aqueous solution (
To further examine the utility of 13C-PEG40,000 for tumor detection
(A) Concentration change in 13C-PEGs of different molecular weights (black circles, 13C-PEG40,000; gray squares, 13C-PEG7,000) in blood circulation of mice after i.v. injection of 92 mg/kg. The concentration at the first sampling point after injection (2 min) was set at 100%. Data presented are the mean and standard deviation values calculated from the results of 3 independent experiments. (B) A graph showing the time dependence of the absolute amount of 13C-PEG40,000 accumulated in C26 tumor and normal tissues 9, 24, 48, 96, 120, 192, and 264 h after intravenous injection of 92 mg/kg in 3, 6, 6, 5, 6, 4, and 4, mice, respectively.
The % injected dose/g value largely depends on tumor weight because the denominator is the tumor weight. However, as the tumor weight continued to increase post-injection in this tumor model, the % injected dose/g value could have led to a misunderstanding of the tendency for 13C-PEG40,000 to accumulate in tumor tissue. Thus, time-dependent accumulation of 13C-PEG40,000 in tumor tissue was not evaluated by % injected dose/g but rather by the absolute amount of 13C-PEG40,000 accumulated in the tumor. As shown in
The long-term retention in tumor indicates that PEG molecules are associated with tumor cells after extravasation into tumor interstitium. Indeed, localization of endocytosed fluorescently labeled PEG40,000 (TMR-PEG40,000) to lysosomes was observed as a dot-like manner in the
To validate the potential of 13C-PEG40,000 in 13C spectroscopic imaging for detecting tumors subcutaneously transplanted into mice, 13C chemical shift imaging (CSI) was performed to obtain biodistribution images of 13C-PEG with a custom-made coil. Representative images from each experiment are shown in
The upper and lower panels show a gradient-echo image and the corresponding 13C image, respectively. Each image represents the FOV of 100 mm in length and 50 mm in width. No additional contrast enhancement was performed for any 13C images, and thus all faithfully represent raw data with respect to intensity. Intensity scale bars represent linear change in intensity from 0 (black) to 255 (red). An arrowhead in each gradient-echo image shows the tumor position. Blue and red lines in 13C images represent the rough position of each tissue, with the positions of the tumors highlighted by a red line. Br, H, L, K, T, and Bl in each 13C image indicate the brain, heart, liver, kidney, tumor, and bladder, respectively. (A) Biodistribution image of 13C-PEG7,000 in a C26 tumor-bearing mouse 1 h after intravenous injection. (B) 13C image of endogenous fat created from the same data set used in the 1 h data in (C). The broad and intense endogenous fat signal shown in
Subsequent
Success in visualizing tumors using 13C-PEG largely depends on the sharp and strong 13C signal of PEG. As described in the section of 13C spectral features of PEG, it is generally recognized that PEG exhibits enhanced structural flexibility and hydrophilicity
12C-based PEGs have been widely tested in the modification of therapeutic and diagnostic agents to improve their stabilities in the blood circulation
The most characteristic feature of the 13C-PEG strategy is the lack of temporal limitations. Because of this property, we are able to wait not only until 13C-PEG sufficiently accumulated in tumors, but also until its concentration in circulation adequately declined after administration of the tracer. This is the most beneficial feature of using a stable isotope. This property would enable diagnoses to be made in renal and urinary systems, in which hyperpolarized and/or radiolabeled tracers may work insufficiently because of the substantial background of the tracer itself. Therefore, the 13C-PEG strategy could be utilized to supplement methods that use hyperpolarized and/or radiolabeled tracer strategies, which could introduce a novel application of MRI into the field of tumor detection.
In this study, we have succeeded in clearly visualizing the tumors of tumor-bearing mice by using 13C-enriched PEG in 13C spectroscopic imaging. The 13C-PEG strategy relies on the strong and sharp 13C nuclear magnetic resonance signal of 13C-PEG. This strategy has provided a novel usage of PEG as a tracer.
(TIF)
(TIF)
(DOC)
(DOC)
We thank C. Matsumura and T. Yamada for NMR data collection. We also thank S. Nakazato, K. Onishi and T. Imaoka for their helpful advice on the preparation of tumor-bearing mice. We are grateful to S. Matsumoto and H. Kodama for the preparation of frozen sections and their helpful advice on fluorescent-dye staining. We appreciate R. Kojima and M. Matsumoto for their technical support on data collection, and T. Sogabe and H. Kinoshita for critical discussion of the data. Finally, we thank K. Machida for critical reading of the manuscript.