Molecular imaging of mesothelioma by detection of manganese-superoxide dismutase activity using manganese-enhanced magnetic resonance imaging

Abstract

Malignant mesothelioma (MM) is a fatal malignancy with a rapidly increasing incidence in industrialized countries because of the widespread use of asbestos in the past centuries. Early diagnosis of MM is critical for a better prognosis, but this is often difficult because of the lack of disease-specific diagnostic imaging. Here, we report that manganese-enhanced magnetic resonance imaging (MEMRI) represents a promising approach for a more selective mesothelioma imaging by monitoring a high-level expression of manganese-superoxide dismutase (Mn-SOD), which is observed in many MM. We found that most human MM cells overexpressed Mn-SOD protein compared with human mesothelial cells and that NCI-H226 human MM cells highly expressed Mn-SOD and augmented Mn accumulation when loaded with manganese chloride (MnCl₂). The cells showed marked T₁-signal enhancement on in vitro MRI after incubation with MnCl₂ because of the T₁ shortening effect of Mn²⁺. H226 subcutaneous tumor was preferentially enhanced compared with a lung adenocarcinoma cell tumor and another human MM cell tumor in MnCl₂-enhanced T₁-weighted MR image (T₁WI), correlating with their respective Mn-SOD expression levels. Moreover, in a more clinically relevant setting, H226 xenografted pleural tumor was markedly enhanced and readily detected by MEMRI using manganese dipyridoxyl diphosphate (MnDPDP), a clinically used contrast agent, as well as MnCl₂. Therefore, we propose that MEMRI can be a potentially powerful method for noninvasive detection of MM, with high spatial resolution and marked signal enhancement, by targeting Mn-SOD.

Malignant mesothelioma (MM) is an asbestos-related malignant tumor originating from mesothelial cells lining pleural and peritoneal cavities. There is a long latency period (more than 30 years) between the asbestos exposure and the onset of the disease, and the number of patients is now rapidly increasing in industrialized countries because of the widespread use and installation of asbestos fibers in the past centuries. Epidemiological studies indicate that the incidence is rapidly increasing worldwide.1

Prognosis of MM is poor, as the disease is refractory to chemotherapy and radiation therapy. It has been suggested that diagnosis of MM in its early stage may improve its prognosis.2 However, it is often difficult to make an early diagnosis of MM because of the lack of reliable diagnostics using disease-specific biomarkers. In addition, some invasive and time-consuming examinations are necessary, such as direct thoracoscopic biopsy, to identify the disease. For diagnostic imaging of MM, chest x-ray, computed tomography (CT), and magnetic resonance imaging (MRI) have been widely used, not only for detection of the tumor but also for its staging.3 High-resolution CT has emerged and has proven to be a more reliable method for MM detection compared with chest x-ray and conventional CT because of its ability to detect minute morphological abnormalities in pleura.4 18-Fluoro-2-deoxyglucose positron emission tomography has been shown as a useful method for differentiating between benign and malignant lesions.5 Although rapid progress has been made in this field, these currently available modalities have a limitation because they lack the selectivity for imaging MM. Therefore, the development of a reliable mesothelioma imaging modality using biomarkers specific for MM has been eagerly awaited.

We previously reported that intracellular amounts of manganese-superoxide dismutase (Mn-SOD)6 were significantly altered in human MM cells compared with mesothelial cells.7 Others have also reported similar results.8–12 These results suggested that Mn-SOD is a reliable molecular marker for MM, prompting us to target Mn-SOD in the development of a more selective molecular imaging for MM. MRI studies using a paramagnetic contrast agent have given us a clue toward addressing this issue; divalent manganese ion (Mn²⁺) has long been known as an MRI contrast agent because of its ability to alter the longitudinal relaxation of water.13 Its use with MRI has seen rapid development over the last decade, particularly in neuroscience. The method is referred to as manganese-enhanced MRI (MEMRI), providing a unique opportunity to study neuronal activation and architecture.14, 15

Based on the fact that Mn is bound to Mn-SOD as a cofactor for enzymatic activity16 and makes superior MRI contrast, we hypothesized that MM cells with a high level of Mn-SOD expression would show augmented cellular Mn accumulation and, consequently, be enhanced by MEMRI. In this study, we tested the possibility of MEMRI as a molecular mesothelioma imaging modality using preclinical animal models. Our data provide evidence that MEMRI is a potentially selective mesothelioma imaging method with high spatial resolution and marked signal enhancement by targeting Mn-SOD.

CT: computed tomography; Fat-Sup: fat suppression preparation; FOV: field of view; Gd-DTPA: gadopentetate dimeglumine; MEMRI: manganese-enhanced magnetic resonance imaging; MM: malignant mesothlioma; MnCl₂: manganese chloride; MnDPDP: manganese dipyridoxyl diphosphate; Mn-SOD: manganese-superoxide dismutase; NA: number of acquisitions; RARE: rapid acquisition with relaxation enhancement; ROIs: regions of interest; ROI_out: outlines of ROIs; RT-PCR: reverse transcription polymerase chain reaction; SE: spin echo; ST: slice thickness; T₁WI: T₁-weighted MR image; TE: echo time; TR: pulse repetition time.

Material and Methods

Cells

The human pleural mesothelial cell line MeT-5A, human MM cell lines (NCI-H28, MSTO-211H, NCI-H226, NCI-H2052 and NCI-H2452) and human lung adenocarcinoma cell lines were obtained from ATCC. HMMME, ACC-MESO-1 and ACC-MESO-4 were obtained from RIKEN BRC cell bank. They were grown in RPMI 1640 medium (Sigma-Aldrich, St Louis, MO) supplemented with 10% fetal bovine serum (JRH Biosciences, Lenexa, KS), 50 U/mL penicillin and 50 mg/mL streptomycin (Invitrogen, Carlsbad, CA). H226R cells were established from NCI-H226 by transfection with pDsRed2-C1 (Clontech, Mountain View, CA), drug selection using G418 at a concentration of 0.5 mg/mL and cloning of the survival cells. Fluorescence emitted from H226R cells was observed with a fluorescence microscope (Leica, Wetzlar, Germany).

Real-time PCR

Total RNA was isolated from the cell lines using the SV total RNA isolation system (Promega, Madison, WI). Quantitative reverse transcription polymerase chain reaction (RT-PCR) was carried out using TaqMan Gene Expression Assay (Applied Biosystems, Foster City, CA). Briefly, total RNA (1 μg) was reverse-transcribed using the superscript first-strand synthesis system for RT-PCR kit (Invitrogen) according to the manufacturer's instructions. Primers and probes obtained from Applied Biosystems were used. PCR amplification was conducted with 10-μL reaction mixture using Premix Ex Taq (Takara, Otsu, Japan). Real-time PCR was carried out in a 96-well plate using Mx3000P (Stratagene, La Jolla, CA, USA) at 95°C for 30 sec, followed by 40 cycles of 95°C for 15 sec, 60°C for 1 min and 72°C for 30 sec. The relative quantitative method was used for the quantitative analysis. Actin gene was used as an endogenous control.

Western blot

Cell lysate was prepared by 50 mM Tris-buffered saline (TBS) with 0.5% TritonX-100. Total protein content was determined by DC protein assay kit (BioRad, Hercules, CA). Ten to 20 μg of the protein was separated on a 12.5% polyacrylamide gel and transferred to Immobilon-P membrane (Millipore, Billerica, MA). For detection of Mn-SOD, the blot was blocked by TBS BLOTTO A (Santa Cruz Biotechnology, Santa Cruz, CA) at room temperature for 2 hr and incubated with primary antibody at room temperature for 2 hr. As a primary antibody, anti-human Mn-SOD rabbit polyclonal antibody (Assay Designs, Ann Arbor, MI) was used at a concentration of 0.2 μg/mL. The membranes were washed three times with TBS containing 0.05% Tween 20, incubated with an anti-rabbit IgG conjugated with horseradish peroxidase (GE Healthcare, Little Chalfont, UK). For loading control, an antibody against actin (Santa Cruz Biotechnology) and anti-goat IgG conjugated with horseradish peroxidase (Santa Cruz Biotechnology) were used. The blots were developed with ECL plus western blotting detection system (GE Healthcare). Images were obtained by Chemi-Smart 5000 (Vilber Lourmat, Marne-la-Vallée, France).

⁵⁴Mn loading study

Cells were seeded in a 24-well dish at a density of 1 × 10⁵ per well 18–24 hr before the uptake study. They were incubated with standard medium containing ⁵⁴MnCl₂ for 0.5 hr or 24 hr. Radiolabeled cells were then washed and lysed, and cellular ⁵⁴Mn was measured (n = 3) with a gamma counter (Aloka, Tokyo, Japan). The counts were normalized to the protein content of the lysate determined using a Bio-Rad DC protein assay kit.

In vitro MEMRI

MnCl₂ was dissolved in the medium at a concentration of 0.1 mM. The cells were incubated with the medium containing MnCl₂ or not for 30 min at 37°C under 5% CO₂.17 After incubation, the medium was removed carefully by washing phosphate-buffered saline twice. The cells (0.5–1.0 × 10⁸) were harvested, transferred to a 96-well PCR tube and pelleted by centrifugation. Cells were imaged after settling into a pellet by gravity. The volume of the cell sample was approximately 80 mm³. The MRI acquisitions were performed in a 7.0-T, 40-cm bore magnet (Kobelco and Jastec, Kobe, Japan) interfaced to a Bruker Avance-I console (BioSpec, Bruker Biospin, Ettlingen, Germany). A 72-mm inner-diameter birdcage coil (Bruker Biospin) was used for measurement of cell samples. Sample temperature was maintained at room temperature (approximately 23°C). Measurements were performed in the following order: T₁WI using conventional spin echo (SE) sequence, multiecho SE imaging for T₂ calculations and inversion recovery imaging for T₁ calculations. For T₁WI, two-dimensional (2D) single-slice T₁WI was obtained using conventional SE sequence with the following parameters: pulse repetition time (TR) = 350 msec, echo time (TE) = 9.57 msec, matrix size = 256 × 256, field of view (FOV) = 51.2 × 51.2 mm², slice thickness (ST) = 1.5 mm, fat suppression preparation (Fat-Sup) = on and number of acquisitions (NA) = 8. Slice orientation was horizontal. For this image, the nominal voxel resolution was 200 μm × 200 μm × 1500 μm. Total acquisition time for T₁WI was 11 min 56 sec. For multiecho imaging, 2D single-slice multiecho imaging was performed for T₂ map calculation with the following parameters: TR = 3,000 msec, TE = 10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 110 and 120 msec, number of echoes = 12, matrix size = 256 × 256, slice orientation = horizontal (same slice orientation as T₁-weighted imaging), FOV = 51.2 × 51.2 mm, ST = 1.5 mm, Fat-Sup = on and NA = 1. For this image, the nominal voxel resolution was 200 μm × 200 μm × 1500 μm. Total acquisition time for multiecho imaging was 12 min 48 sec. For inversion recovery, 2D single-slice inversion recovery MRI was obtained using rapid acquisition with relaxation enhancement (RARE) read out for T₁ map calculation with the following parameters: TR = 10,000 msec, TE = 10 msec, inversion time = 51, 100, 200, 400, 800, 1,600, 3,200 and 6,400 msec, matrix size = 128 × 128, slice orientation = horizontal (same slice orientation as T₁-weighted imaging), FOV = 51.2 × 51.2 mm², ST = 1.5 mm, Fat-Sup = on, rapid acquisition with RARE factor = 4 and NA = 1. For this image, the nominal voxel resolution was 200 μm × 200 μm × 1500 μm. Total acquisition time for inversion recovery MRI was 49 min 27 sec. Quantitative T₁ maps were calculated by nonlinear least-squares fitting using inversion recovery MRI. In addition, quantitative T₂ maps were calculated by nonlinear least-squares fitting using multiecho imaging. Regions of interest (ROIs) were defined as precipitated cell regions. All calculations and analyses were performed using the MRVision image analysis software (version 1.5.8; MRVision.) on Linux PC (Red Hat Linux, Mountain View, CA).

In vivo MEMRI

In vivo MRI was acquired using the same 7-T Bruker BioSpec system together with 2-ch phased array RF coil for subcutaneous tumor or volume RF coil for intrapleural tumor. Mice were anesthetized with 1.5–2.0% isoflurane and placed in a prone position, and anesthesia was maintained at this level. In pleural tumor models, mice were ventilated. During the experiment, a warm air-flow over the animal was used to maintain body temperature at 37.5°C. Respiratory rate was maintained at 20–40 breaths per min and monitored throughout the experiment. Before T₁WI acquisitions, scout images were acquired to localize the tumor.

Subcutaneous tumor

T₁WI and T₁ map acquisitions were performed in the following order: preadministration (control), Gd-enhanced and Mn-enhanced experiment. First, one T₁WI and T₁ map were acquired as a control before Gd administration to check for tumor necrosis. Second, T₁WI and T₁ map acquisitions were repeated four times every 16 min after Gd-DTPA (gadopentetate dimeglumine) administration. Gd-DTPA (150 μmol/kg, Bayer) was diluted to 50 mM with saline and injected intravenously to evaluate tumor vasculature. Third T₁WI and T₁ map acquisitions were repeated five times every 32 min after starting Mn infusion. There were 90-min intervals between Gd injection and Mn injection, and Mn infusion was started after confirming washout of Gd. MnCl₂ administration was performed as described in a previous report.18 Before the administration, a 100-mM solution of MnCl₂ (MnCl₂-4H₂O; Sigma-Aldrich) was made with distilled water and diluted to 50 mM by saline to match the osmolality to that of blood. We slowly infused 380 μmol/kg of MnCl₂ as a 50-mM osmotic pressure-controlled solution at a rate of 0.15–0.2 mL/hr through the tail vein using a syringe pump (KDS-100; KD Scientific, Holliston, MA). For T₁WI for subcutaneous tumors, 2D single-slice T₁WI was obtained using conventional SE sequence with the following parameters: TR = 300 msec, TE = 9.57 msec, matrix size = 256 × 256, FOV = 40.0 × 40.0 mm², ST = 1.0 mm, Fat-Sup = on and NA = 4. Slice orientation was transverse. For these images, nominal voxel resolution was 156 μm × 156 μm × 1000 μm. Total acquisition time was 5 min 7 sec. For T₁ map calculation, 2D single-slice Look-Locker T₁ mapping images were obtained with the following parameters: TR = 10000 msec, TE = 10 msec, α = 20°, τ = 400 ms, matrix size = 64 × 64, FOV = 40.0 × 40.0 mm², ST = 1.0 mm, Fat-Sup = on and NA = 1. Slice orientation was transverse. For these images, nominal voxel resolution was 625 μm × 625 μm × 1000 μm. Total acquisition time was 10 min 40 sec. ROIs of subcutaneous tumor were defined by the following methods: (i) outlines of ROIs (ROI_out) were encircled by hand on T₁WI, (ii) Gd-enhanced area was excluded from the ROI_out over the threshold level (+2 SD of mean signal intensity of the ROI_out in the control image).

Pleural tumor

First, one T₁WI was acquired as a control. Second, T₁WI acquisitions were repeated six times every 20 min after MnCl₂ (50 mM, 380 μmol/kg, 0.15–0.2 mL/hr)18 or MnDPDP (Teslascan™; GE Healthcare; 10 mM, 100 μmol/kg, 0.15–0.2 mL/hr) administration. For T₁WI for intrapleural tumors, 2D multi-slice T₁WI was obtained using a respiratory-gated SE sequence with the following parameters: TR = 600 msec with respiratory gating (artificial ventilation), TE = 9.57 msec, matrix size = 256 × 256, FOV = 40.0 × 40.0 mm², ST = 1.0 mm, Fat-Sup = on and NA = 8. Slice orientation was transverse. For these images, nominal voxel resolution was 156 μm × 156 μm × 1000 μm. Total acquisition time was approximately 20 min.

Animal models

Mouse protocols were approved by the Institutional Animal Care and Use Committee of the National Institute of Radiological Sciences, Japan. The subcutaneous tumor model was formed in the lower back of a nude mouse (Clea Japan, Tokyo, Japan) by subcutaneous injection of 5 × 10⁶ of MSTO-211H or PC14 and 1 × 10⁷ of NCI-H226 (n; see figure legends), in 200 μL of Hanks' balanced salt solution (Invitrogen). The intrapleural tumor model was developed by intrapleural injection of 4 × 10⁶ of H226R in 100 μL of Hanks' balanced salt solution (Invitrogen) in nude mice (n = 5 for MnCl₂, n = 4 for MnDPDP).

In vivo optical imaging

Fluorescence emission from intrapleural tumors was monitored by IVIS® lumina optical imaging system (Xenogen, Alameda, CA). Tumor-bearing mice were anesthetized by inhalation of 2.5% isoflurane and placed in a light-tight chamber. We selected the filter set for DsRed (excitation: 500–550 nm, emission: 575–650 nm, background excitation: 460–490 nm) and set the constant imaging parameters [exposure time: 2 sec, binning: medium, lens aperture (f/stop): 2, field of view: 12.5 cm, floor lamp level: high]. In surgically opening the pleural cavity, fluorescence from the implanted tumor was imaged by fluorescence stereoscopic microscope MZ16F (Leica).

Immunohistochemistry

The tumor samples were fixed in 4% paraformaldehyde phosphate-buffered solution (Wako, Osaka, Japan), embedded in OCT compound and sectioned. Immunohistochemistry was done using the rabbit ABC staining system (Santa Cruz Biotechnology) in accordance with the manufacturer's instructions. We used the same antibody against Mn-SOD as used in Western blotting as a primary antibody. Normal rabbit IgG (Santa Cruz Biotechnology) was used as negative control. Sections were observed using a light microscope (Olympus, Tokyo, Japan).

Statistical analyses

All data are presented as mean ± SD. Statistical analyses were performed using Prism 5 (version 5.0; GraphPad Software, La Jolla, CA). A probability value of less than 0.05 was considered significant for each analysis. We compared pairs by Student's t-test for the RT-PCR study and for the mean T₁ values of H226 tumors and PC14 tumors. Comparisons among three groups were analyzed using one-way ANOVA with Tukey's post hoc test for the Mn loading study. Two-way ANOVA with Bonferroni post hoc test was applied to compare normalized MRI signal intensity for subcutaneous tumors.

Results

Mn-SOD expression and Mn accumulation in human MM cells

We initially evaluated the amounts of Mn-SOD protein in MeT-5A human mesothelial cells and in eight human MM cells by Western blot analysis. All MM cells except MESO-1 cells expressed higher levels of Mn-SOD protein than MeT-5A cells. Notably, H226 highly expressed Mn-SOD protein (Fig. 1a). We examined Mn-SOD mRNA expression in MeT-5A cells and five of the MM cells by quantitative RT-PCR analysis (Fig. 1b). In accordance with the result of Western blot, H226 cells had the highest amount of Mn-SOD mRNA among the cells tested, with the level being about 15 times higher than that in MeT-5A (p < 0.01). H2452 had 8.4-fold higher Mn-SOD mRNA than MeT-5A (p < 0.05). MSTO-211H had 6-fold higher Mn-SOD mRNA than MeT-5A, albeit without a statistically significant difference. The level of Mn-SOD mRNA in H2052 was slightly higher than that in MeT-5A, and MESO-1 had almost the same level as MeT-5A (Fig. 1b). We assessed cellular Mn accumulation in MeT-5A and four MM cells by incubation with radiolabeled ⁵⁴MnCl₂ to investigate the effect of Mn-SOD on Mn accumulation (Fig. 1c). Mn-SOD highly expressing-H226 cells showed 6% higher ⁵⁴Mn accumulation compared with MeT-5A under 0.5-hr incubation with ⁵⁴MnCl₂. When incubated for 24 hr, the cells showed significantly higher ⁵⁴Mn accumulation than that in MeT-5A, with the level being about two times higher (p < 0.01). H2452, which had the second highest expression of Mn-SOD, showed a higher uptake at 24-hr incubation compared with MeT-5A, although this was not statistically significant.

Figure 1. Mn-SOD expression and Mn accumulation in human mesothelial and MM cells. (a) Western blot analysis of Mn-SOD protein. Actin was used as loading control. (b) Quantitative analysis of Mn-SOD mRNA by real-time RT-PCR. Relative ratio is shown relative to the amount of Mn-SOD mRNA in MeT-5A cells. Data were obtained from four independent experiments. Mean ± SD. * p < 0.05, ** p < 0.01, compared with MeT-5A. (c) Mn loading study. The ratios of Mn accumulation relative to that in MeT-5A cells at each incubation time (0.5 or 24 hr) are shown. The values were calculated from triplicate samples and shown as mean ± SD. ** p < 0.01, compared with MeT-5A.

In vitro MEMRI of human MM cells

Figure 2a shows a T₁-weighted MR image (T₁WI) of the cell pellets of H226 and MSTO-211H when incubated with a standard medium supplemented with or without 0.1 mM MnCl₂. We chose these cells because they were both MM cells capable of consistently forming xenograft tumors but had distinctly different Mn-SOD levels. Both MM cells were enhanced by MnCl₂ but with a different efficiency. Consistent with the results of the Mn loading study, H226 was markedly enhanced with 0.1 mM MnCl₂ while the signal enhancement of MSTO-211H cells was minimal in comparison. Figure 2b shows the T₁ and T₂ values of those cells. The decrease of T₁ in H226 was larger than that in 211H when loaded with 0.1 mM MnCl₂ (58 vs. 76%, relative to the value with no Mn supplementation). T₂ was very slightly decreased in H226, whereas there was no change in 211H by Mn supplementation.

Figure 2. In vitro MEMRI of H226 and MSTO-211H cells. (a) T₁WI of the cell pellets with or without 0.1 mM MnCl₂ supplementation. Incubation time was 30 min. T₁WI of MnCl₂ solutions is also shown. (b) T₁ and T₂ values of the MnCl₂ solutions, H226 and MSTO-211H cell pellets are shown. Mean ± SD. Mean and SD were calculated from two independent experiments. The values of MnCl₂ solution were obtained in a single experiment.

In vivo MEMRI of subcutaneous tumors of human MM cells with MnCl₂

We implanted H226 cells and MSTO-211H cells subcutaneously in the same nude mice. Mn-SOD expression was evaluated in those tumors by immunohistochemistry using a specific antibody against Mn-SOD. The study revealed that H226 subcutaneous tumors expressed a higher level of Mn-SOD than MSTO-211H tumors in vivo and in vitro (Fig. 3a). When both tumors reached a similar size, T₁-weighted MRI and quantitative T₁ maps were acquired. At the beginning of the MRI experiments, we used Gd-DTPA to evaluate tumor vasculature and exclude the area where Gd was highly accumulated because of disruption of intratumoral microvasculature. In most cases, both tumors were adequately and almost equally enhanced by Gd-DTPA. After Gd was washed out, we scanned for T₁WI with systemic MnCl₂ administration. Although MSTO-211H tumor was slightly enhanced by MnCl₂, H226 tumor was preferentially enhanced in T₁WI compared with MSTO-211H tumor (Fig. 3b). It appeared that a portion of the tumor, rather than the whole area, was enhanced with high signal intensity in most of the cases. Figure 3c shows the ratio of signal changes in the tumors. The ratio in H226 was increased from the start of Mn administration, reaching about 1.7-fold increase at 60 min of Mn loading, then decreasing slowly, whereas the kinetics of MSTO-211H, although similar to H226, reached only a 1.3-fold increase at the peak. The ratio in H226 was significantly increased compared with that in MSTO-211H at 60, 90 and 120 min after Mn loading. Quantitative T₁ measurement showed that mean T₁ in H226 tumors was less than 1,000 msec and less than that in MSTO-211H (approximately 1,350 msec) after Mn administration (Fig. 3d).

Figure 3. MEMRI of H226 and MSTO-211H subcutaneous tumors. (a) Immunohistochemistry of Mn-SOD expression in the subcutaneous tumors using anti-Mn-SOD antibody (Anti-Mn-SOD Ab). IgG was used as negative control. Scale bars: 100 μm. (b) Typical T₁WIs and T₁ maps of the subcutaneous tumors with MnCl₂ administration. In the images, administered contrast agents and time are shown in white letters (Gd: Gd-DTPA, Mn: MnCl₂). Yellow letters indicate inoculated cells. Circles with high MR signal between tumors in the images were used as water sample for standardization. Corresponding T₁ map images are also shown. (c) Change of normalized signal intensities of the tumors after systemic Mn administration (n = 4 for each group). The starting point of MnCl₂ administration was considered as 1. The black bar indicates the administration time of MnCl₂. ** p < 0.01, * p < 0.05 (two-way ANOVA). (d) T₁ values of MSTO-211H and H226 subcutaneous tumors. Mean ± SD.

Comparison between MM cells and lung adenocarcinoma cells

We further conducted comparison studies between MM cells and lung adenocarcinoma cells, because it is often difficult to distinguish between MM and lung adenocarcinoma that has metastasized to the pleural cavity.19 Western blot analysis revealed that Mn-SOD levels in four human lung adenocarcinoma cells (PC14, H441, H1975 and H1650) were lower than that in H226 cells (Fig. 4a). We compared H226 cells with PC14 cells for tumor enhancement in vivo by MEMRI. In accordance with their Mn-SOD levels, T₁ maps showed that T₁ of the H226 xenografts was shortened in comparison with PC14 xenografts by Mn administration (Fig. 4b). After Mn administration, the mean T₁ value in PC14 was 1417 ± 430 msec, whereas it was 853 ± 75 msec in H226 tumors (p = 0.02) (Fig. 4c).

Figure 4. Comparison between human MM and lung adenocarcinoma cells by in vivo MEMRI. (a) Western blot analysis of Mn-SOD protein in human adenocarcinoma cells (PC14, H441, H1975 and H1650) and H226 MM cells. (b) T₁ maps of PC14 and H226 subcutaneous tumors. (c) T₁ values of these subcutaneous tumors before and after MnCl₂ administration (n = 5 for each group). Mean ± SD.

In vivo MEMRI of intrapleural tumors with MnCl₂

We performed MEMRI of intrapleural H226 tumors with MnCl₂. We established H226 cells stably expressing red fluorescence protein, named H226R cells, to easily evaluate tumor formation in the pleural cavity by the fluorescence (Fig. 5a). They were injected into the pleural cavity of nude mice, and tumor formation was confirmed by in vivo optical imaging (Fig. 5b). Then, the tumor-bearing mice were subjected to MEMRI. The xenografted tumor in the pleural cavity was obscure before Mn administration. In contrast, the tumor was enhanced specifically relative to the surrounding tissue even at 20 min after MnCl₂ injection, and the signal increased continuously until 60 min. The tumor was visualized in T₁WI with superior contrast to the neighboring muscle over 100 min, and especially at 60 min (Fig. 5c). The relative signal intensity in the tumors was higher than that in neighboring muscle, reaching approximately 1.5-fold increase of signal in tumor (Fig. 5d). After the MRI study, we confirmed that the Mn-enhanced tumor originated from the inoculated H226R cells by red fluorescence (Fig. 5e).

Figure 5. Detection of intrapleural H226 xenografted tumors by MnCl₂. (a) H226 cells stably expressing DsRed protein (H226R). (b) Detection of tumor arising from intrapleurally inoculated H226R cells by in vivo fluorescence imaging. The yellow arrow indicates a tumor in the pleural cavity. The pseudocolor scale bar indicates signal intensity. (c) Typical T₁WIs of H226R intrapleural tumor with systemic MnCl₂ injection. Times of MnCl₂ administration are indicated in the images. R: right; L: left. The tumors were found in the area encircled by yellow dotted line in the “Mn 60 min” image and showed distinct contrast to the surrounding normal tissues during 20–100 min of MnCl₂ administration. The mouse in (c) was the same as in (b). (d) Relative change of signal intensities of the tumors and neighboring muscle after systemic MnCl₂ administration (n = 4). The starting point of MnCl₂ administration was considered as 1. (e) The tumor enhanced by MnCl₂ was derived from H226R cells. Upper images are T₁WIs with systemic MnCl₂ administration. The areas encircled by yellow dotted line in the upper images indicate a tumor enhanced by MnCl₂. Lower images were obtained by stereoscopic microscope using white light or a filter set for red fluorescence. Yellow arrows indicate the tumor corresponding to the tumor indicated by yellow dotted line in the upper images. Scale bar in the “white light” image: 5 mm. The tumor in (e) was different from the tumor in (c).

In vivo MEMRI of intrapleural tumors with MnDPDP

To investigate the feasibility of this method in a clinically relevant situation, we examined whether the intrapleural tumors were enhanced with MnDPDP, a Mn-chelating contrast agent that is clinically used for liver MRI. Figure 6a shows that H226R pleural tumors were selectively enhanced by MnDPDP and readily detected, in clear contrast with the neighboring muscle. The relative signal intensity in neighboring muscle increased very slightly, while, in contrast, the signal in tumor increased by approximately 1.4 fold relative to before Mn administration (Fig. 6b).

Figure 6. Detection of intrapleural H226 xenografted tumors by MnDPDP. (a) Typical T₁WIs before (left) and 60 min after (right) MnDPDP administration are shown. White arrows indicate the tumors. Note that liver was also enhanced because MnDPDP is a contrast agent for liver MRI. Li: liver; R: right; L: left. (b) Relative change of signal intensities of the tumors and neighboring muscle after systemic MnCl₂ administration (n = 4).

Discussion

The focus of our research was to develop a more reliable imaging modality for MM by targeting biomarkers characteristic of the disease. We previously reported that human MM cells had a higher amount of Mn-SOD protein compared with mesothelial cells when assayed by ELISA.7 In this study, we further evaluated the levels of Mn-SOD mRNA and protein in human MM cells and mesothelial cells and confirmed that Mn-SOD is overexpressed in most human MM cells compared with mesothelial cells. Mn-SOD is an antioxidant enzyme located in mitochondria and probably plays a key role in oxidative stress and tumor cell apoptosis. It is conceivable that because tumor cells, in general, are metabolically active and, thus, produce a high level of the oxidant affecting cell viability, Mn-SOD is possibly induced to counteract the oxidative stress.12 However, multiple studies on Mn-SOD expression in human tumors have shown variable results; upregulation of Mn-SOD protein is observed in many types of human tumors, but in some tumors, the enzyme is downregulated.12 Nevertheless, many studies, including ours, showed that Mn-SOD protein and its activity were consistently elevated in human MM compared with nonmalignant mesothelium or mesothelial cells. These strongly suggest that Mn-SOD is a reliable diagnostic marker for MM, warranting our approach using Mn-SOD as a target for imaging MM.

We conducted the Mn loading study to investigate whether Mn is accumulated in MM cells that highly express Mn-SOD. It was observed that H226 cells, which showed the highest expression of Mn-SOD, had the highest level of accumulation among the cells tested. In addition, H2452 cells, with the second highest expression of Mn-SOD, also showed a higher Mn uptake than MeT-5A. These observations suggested that Mn accumulation is increased in MM cells highly expressing Mn-SOD. There is an analogous example in another metal-binding protein, where cells overexpressing iron-binding ferritin protein enhanced iron accumulation.20–22 Furthermore, consistent with the loading study, in vitro and in vivo MEMRI studies showed that H226 cells were more effectively enhanced by MnCl₂ than MSTO-211H cells, in line with their respective Mn-SOD expressions. Interestingly, Yang et al.23 also suggested that Mn²⁺ may be applicable for monitoring Mn-SOD expression by their MRI study using a brain injury model, further strengthening the notion that MEMRI monitors intracellular Mn-SOD activity.

Although previous studies and our current study have shown overexpression of Mn-SOD in MM, various levels of Mn-SOD expression have been observed in human MM cell lines and in patient samples.7, 11 For example, the Mn-SOD level in H226 was estimated to be about 10-fold higher than that in MSTO-211H cells.7 In this article, we presented the possibility of detection of MM with a high level of Mn-SOD by MEMRI. However, the sensitivity of MEMRI has yet to be determined. Further investigations are needed, and more optimized protocols should be developed to detect MM with lesser expression of Mn-SOD.

It has been suggested that Mn-SOD would be an additional diagnostic marker for differential diagnosis between MM and adenocarcinoma metastasized to pleura because the latter expressed a remarkably low level of Mn-SOD.8–12 Consistent with those previous reports, our study showed that Mn-SOD levels in lung adenocarcinoma cells were lower than those in MM cells. Furthermore, tumor enhancement of H226 MM cells was greater than that of PC14 lung adenocarcinoma cells, which were used for the pleural metastasis model,24 raising the possibility that MEMRI could be useful for the differential diagnosis between MM and metastatic adenocarcinoma. These data strengthen the utility of MEMRI for imaging of MM.

We also demonstrated that in a more clinically relevant setting, H226 tumors in the pleural cavity showed marked signal enhancement in T₁WI after the systemic injection of MnCl₂. Signal enhancement was observed even at 20 min of MnCl₂ injection, and the tumor was clearly visualized, with high signal intensity, with distinct contrast to the surrounding tissues such as neighboring muscle and lung. These results ensure the feasibility of the method for imaging MM in the pleura, which is a major site of the disease.

Although MnCl₂ is used as an MRI contrast agent for the gastrointestinal tract in human,25 systemic administration of MnCl₂ at the concentration used in this study is restrained in human applications because of its toxicity.26 Therefore, we conducted MEMRI of the pleural tumors using MnDPDP, an Mn-releasing contrast agent used for human liver imaging.27 We showed that H226R tumors were selectively enhanced with MnDPDP as well, suggesting that our strategy is clinically translatable. In addition, the development of safer Mn administration and a sophisticated drug carrier for delivering Mn²⁺ exclusively to the lesion will add to the advantages of our method. In summary, MEMRI has proven to be useful for molecular imaging for MM and holds promise for a more reliable diagnosis of this fatal disease.

Acknowledgements

We thank Winn Aung for assistance with in vivo fluorescence imaging and Sayaka Shibata for skillful assistance with the MRI experiments.