| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | RSS | TABLE OF CONTENTS |
|
|
|
||||||||
|
||||||||||
Basic Science Investigation |
1 Department of Radiation Oncology, Duke University Medical Center, Durham, North Carolina; 2 Department of Radiology, Duke University Medical Center, Durham, North Carolina; and 3 Center for In Vivo Microscopy, Duke University Medical Center, Durham, North Carolina
Correspondence: For correspondence or reprints contact: Mark W. Dewhirst, DVM, PhD, Room 201 MSRB, Research Dr., Duke University Medical Center, Box 3455 DUMC, Durham, NC 27710. E-mail: dewhirst{at}radonc.duke.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONReferences |
|---|
Key Words: tumor hypoxia • PET • Cu-ATSM
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONReferences |
|---|
Cu-ATSM is a promising hypoxia marker for PET, which has been shown to be predictive of radiotherapy treatment outcome in several small clinical series (10,11). Copper has several positron-emitting isotopes (60Cu, 61Cu, 62Cu, and 64Cu) with various half-lives, ranging from 9.7 min for 62Cu to 12.7 h for 64Cu. This allows a wide choice of isotopes, so that half-life (t1/2) and radiation energy level can be tailored to particular needs. The synthesis of Cu-ATSM is relatively easy with a 1-step reaction between copper and the substrate (12). Cu-ATSM has small molecular weight and high cell membrane permeability; thus, it can diffuse easily from bloodstream to surrounding cells. Intracellular Cu-ATSM has been shown to be bioreduced and trapped in viable cells under low cellular partial pressure of oxygen (PO2) (12–14).
The use of Cu-ATSM as a hypoxia imaging agent was first reported by Fujibayashi et al. in 1997 (14) in a rat ischemic heart model. Subsequent studies on Cu-ATSM uptake and retention have been performed on tumor cells both in vitro (8,15) and in vivo (9–11). However, none of those studies directly correlated the spatial distribution of Cu-ATSM with that of tumor hypoxia in vivo, particularly in multiple tumor types.
In this study we aimed to validate the ability of 64Cu-ATSM as a hypoxia marker on both macro- and microscales and to measure spatial correlations between 64Cu-ATSM uptake and tumor hypoxia level, as assessed independently. Several immunohistochemical hypoxia markers were used to compare with 64Cu-ATSM in terms of the hypoxia specificity. The uptake of 64Cu-ATSM was compared with that of immunohistochemical markers. EF5 (2-[2-nitro-1H-imidazol-1-yl]-N-[2,2,3,3,3-pentafluoropropyl]acetamide), a well-established hypoxia marker, can be bioreduced by cells in proportion to the level of cellular PO2 (16). Fluorescent staining of EF5 binding can be quantified and directly linked to the oxygen level (17). We also used pimonidazole (18) and carbonic anhydrase IX (CA-IX) (19) as additional hypoxia markers for robust comparisons in selected cases. Three different tumor lines were used in this study: R3230 adenocarcinoma (R3230Ac), 9L glioma (9L), and fibrosarcoma (FSA). We have previously studied hypoxia distribution and oxygen transport in these 3 tumor lines and found that R3230Ac and FSA are similar in their degree of hypoxia and 9L tumor is more hypoxic (20,21). On the basis of this information, we evaluated the ability of 64Cu-ATSM to measure hypoxia in these tumor lines.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONReferences |
|---|
30 µL) of 64Cu-ATSM was taken for thin-layer chromatography and high-performance liquid chromatography (HPLC) analysis. HPLC analysis showed high radiochemical purity, typically >96% yield of 64Cu-ATSM in the final product.
Animal Models
Fischer 344 rats (120–130 g) were purchased from the National Cancer Institute, Frederick Cancer Center. R3230Ac was originally obtained from the American Type Culture Collection. Rat FSA was kindly provided by Dr. Joan Bull (M.D. Anderson Cancer Center, Houston, TX), and 9L cells were obtained courtesy of Dr. Kenneth Wheeler (Wake Forest University, Winston Salem, NC). For tumor implantation, small pieces of tumor tissue (1–2 mm3 each) were freshly obtained from tumor-bearing donor rats and transplanted subcutaneously in the neck of recipient animals. After the tumors had reached 1–2 cm in diameter (usually 3 wk after implantation), animals were used for imaging experiments. All procedures involving animals were performed in compliance with the guidelines of the Duke University Institutional Animal Care Committee.
PET and Experiment Protocol
64Cu-ATSM PET was performed on a microPET R4 system (Concorde Microsystems, Inc.). The performance evaluation of the microPET R4 system can be referred to the report by Knoess et al. (22). Animals were intubated and ventilated with an isoflurane/air mixture (1% isoflurane) using a customized forced-air ventilator (tidal volume, 12 mL/kg). A tail vein catheter was placed for the injection of radiopharmaceuticals and other reagents. An electrocardiograph (ECG) was recorded and real-time monitored through a Labview software-controlled, customized ECG system. Body temperature was maintained at around 37°C using a regulated water-heated blanket surrounding the whole body, which was monitored through a rectal temperature probe during PET.
Animals were first given EF5 solution (30 mg/kg, intravenously, obtained from Dr. Cameron Koch, University of Pennsylvania, Philadelphia, PA) 2 h before PET scans (or 3 h before sacrifice) via the tail vein. For some animals, pimonidazole solution (60 mg/kg; Chemicon) was administered by intraperitoneal injection 1 h before PET scans. The animal was placed on a support cradle in a prone position and advanced into the scanner. The tumor was centered inside the scanner using a laser alignment system. A 15-min 68Ge transmission scan was performed for attenuation correction. After the transmission scan, 64Cu-ATSM (22.2–31.4 MBq per rat) was administered through the tail vein catheter as a bolus injection. A 1-h dynamic PET scan was subsequently performed.
After the PET scan, a single dose of 100 µL Hoechst 33342 (5 mg/kg, no. B2261; Sigma) was administered intravenously. Five minutes after injection of Hoechst 33342, the skin covering the tumor was separated and removed carefully without significantly disturbing the tumor position. The tumor was exposed with the base still connected to the host tissue. A 23-gauge needle was inserted carefully into the tumor along the axial scanning direction as a reference marker for cryosection orientation. Two cross lines were also marked on the tumor surface indicating axial and transverse directions. The tumor was then removed with reference needle attached and snap-frozen in liquid nitrogen for autoradiography and immunohistochemistry. Other organ tissues—including brain, heart, liver, kidney, and muscle—were also collected and snap-frozen along with tumor for biodistribution study.
Immunohistochemistry and Fluorescence Imaging
The frozen tumor samples were cryosectioned at 14-µm thickness. The section orientation was determined from the reference needle and tumor surface cross-line markers to ensure that the cutting was perpendicular to the axial direction and on the same transverse section as the PET. Slides were counted from the initiation of cutting and numbered consecutively so that depth could be recorded. For each tumor, 3 locations along the axial line were chosen and at least 6 consecutive slides (2 for autoradiography and 4 for EF5 staining) were cut for each location.
EF5 was used as a hypoxia marker in all 3 tumor lines. EF5 immunostaining was done following a protocol developed by Dr. Cameron Koch's laboratory (23). Briefly, cryosectioned tumor sections were fixed in 4% formaldehyde, blocked in tissue-blocking reagent, and incubated with the EF5 antibody ELK3-51 conjugated with Cy3 dye (75 µg/mL, obtained from Dr. Cameron Koch). One section was treated without antibody to provide endogenous tissue fluorescence as a negative control. Another adjacent section was incubated with antibody and treated with 0.5 mmol/L EF5 as a "competing stain" to assess nonspecific binding of antibody (24). Stained tumor sections were kept in 1% formaldehyde phosphate-buffered saline (PBS) and imaged within 3–4 d. After imaging, sections were stained with hematoxylin and eosin (H&E) to obtain histologic structure.
To validate unusual EF5 cross-correlations with 64Cu-ATSM PET and autoradiography obtained with FSA, 2 other hypoxia markers, pimonidazole and CA-IX (an endogenous hypoxia-regulated protein), were also used. Tumor sections were fixed and incubated with either murine antipimonidazole primary antibody (Chemicon) or anti–CA-IX mouse monoclonal antibody clone M75 (a gift from Dr. Egbert Oosterwijk, Department of Urology, University Hospital Nijmegen, Nijmegen, The Netherlands) for 1 h. A fluorescein isothiocyanate–conjugated antimouse secondary antibody was added after primary antibody incubation. The stained slides were kept in cold 1% formaldehyde PBS for microscope imaging.
Fluorescent images for EF5 and Hoechst staining were made using a fluorescence microscope (Axioskop 2plus; Zeiss) equipped with a cooled charge-coupled device digital camera (Retiga1300R; Q-Imaging). A tetrarhodamine isothiocyanate (TRITC) fluorescence filter was used for EF5-Cy3 imaging (excitation maximum [EX] = 552 nm; emission maximum [EM] = 570 nm), and a 4',6-diamidino-2-phenylindole (DAPI) filter was used for Hoechst 33342 imaging (EX = 347 nm; EM = 483 nm). An automatic microscope stage and controller (MAC5000; Ludl Electronic Products) with 0.025 µm per step size was used for the whole tissue slide scan. Tissue slides were scanned for Hoechst 33342 fluorescence to reveal the functional vessels and perfused tissues, followed by a second scan on the same slide for either EF5, pimonidazole, or CA-IX. Images acquired from stage scanning were stitched and overlaid for direct comparison of dual fluorescent images. Image acquisition and control was accomplished using MetaMorph imaging software (version 6.2; Molecular Devices Co.).
Electronic Autoradiography
Electronic autoradiography was performed on a storage phosphor system (Packard Bioscience). After cryosectioning tissue, slides were exposed to a storage phosphor screen for 48 h at 4°C. The phosphor screen was then processed and read in the phosphor system to visualize the distribution of 64Cu-ATSM in tumor and other organs. After autoradiography, all slides were H&E stained to obtain corresponding histology information for each.
Carbogen-Breathing Experiments
To investigate whether changing the oxygenation in tumors would influence the uptake of 64Cu-ATSM in FSA tumors, carbogen-breathing experiments were conducted. Rats were intubated with an 18-gauge catheter and ventilated with 1% isoflurane-mixed carbogen gas (95% O2/5% CO2, research grade; Machine & Welding Supply Co.) by a customized ventilator. Each rat was put on carbogen breathing for 30 min before injection of 64Cu-ATSM and EF5. They continued to breathe carbogen during the 1 h PET scan. After imaging and the injection of Hoechst 33342, the tumor was collected and snap-frozen. Autoradiography and immunohistochemistry were conducted on the tumor sections as described.
Image Analysis
Raw PET projection data (list-mode data) were binned into 95 time frames: 55 frames for the first 20 min (30 x 10-s frames, 20 x 30-s frames, 5 x 60-s frames) and 40 frames for the next 40 min (40 x 60-s frames). Images were reconstructed by the ordered-subsets expectation maximization algorithm provided by Concorde Microsystems, with attenuation and decay correction applied.
Regions of interest (ROIs) were drawn around the viable tumor area (regions within the tumor with very low or absent activity were presumed to be necrotic). Three 2-dimensional ROIs in transverse sections and one 3-dimensional volumetric ROI were drawn and used for each tumor. Dynamic uptake curves representing these ROIs were recorded from the 95 time frames. The standardized uptake value (SUV) was measured as the decay-corrected radiotracer concentration normalized to the injection dose and animal weight.
The tumor-to-muscle ratio (T/M ratio) was calculated as well by comparing the radioactivity level in tumor and that in muscle at 1 h after injection. One SUV and one T/M value at 1 h after injection were obtained for each animal by averaging the measurement values from the 4 ROIs of that animal.
The autoradiography and EF5-stained images were taken from consecutive tumor sections and compared section by section. Three pairs of sections from autoradiography and EF5 staining were evaluated for each tumor. To calculate the 64Cu-ATSM uptake area fraction and the EF5 binding area fraction, a threshold was needed to separate the positive uptake area in autoradiography images or hypoxia area in EF5-stained microscopic fluorescent images. Because of the large fluorescent dynamic range, it is hard to determine a threshold visually from a histogram (Fig. 1D). To avoid too much involvement of subjective selection of the threshold, an automatic threshold scheme, Otsu thresholding method (25) built in ImageJ software (ImageJ 1.34 version, National Institutes of Health; http://rsb.info.nib.gov/ij), was used to threshold images. An example of thresholding and segmenting images into positive uptake/staining area and tissue background area was shown in Figure 1. Another thresholding scheme, adaptive automatic thresholding in MetaMorph software, was also tried in select cases to be compared with the Otsu method, and similar segmentation results was obtained. Viable tissue area for each section was measured from the H&E staining in the same slide. The 64Cu-ATSM uptake area fraction (or EF5 binding area fraction) was then calculated by dividing the positive uptake area (or binding area) by the total viable tissue area from autoradiography images (or fluorescent images). Differences between the 2 area fractions were compared statistically. For spatial correlation analysis, autoradiography and EF5 microscopic images were scaled to the same size according to image calibration and aligned by the Turbo-Registration package in ImageJ software. A scatter plot was generated with the x- and y-coordinates representing region-averaged (3 x 3) pixel values in each paired autoradiography image and EF5 image. Each point in the scatter plot corresponds to a specific position within the tumor. Correlation coefficients were calculated on the basis of scatter plots for each of the paired images.
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONReferences |
|---|
|
|
|
|
|
|
To confirm the unexpected EF5–64Cu-ATSM pattern seen with FSA, comparison was made with 2 other hypoxia marker drugs, pimonidazole and CA-IX. Figure 6 shows the CA-IX staining and pimonidazole staining together with 64Cu-ATSM autoradiography in FSA tumor (Figs. 6A–6D). In both cases, the same complementary pattern of hypoxia immunostaining was seen, compared with 64Cu-ATSM autoradiography. Pimonidazole staining was also tested in R3230Ac tumors (Fig. 6F), and it showed a close correlation between 64Cu-ATSM uptake and pimonidazole binding (spatial correlation coefficient r = 0.73, P < 0.001), similar to what was seen with EF5.
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONReferences |
|---|
Our study showed that 64Cu-ATSM has a relatively high T/M ratio within a short period of time compared with other potential hypoxia markers. At 1 h after injection in the rat mammary tumor R3230Ac, 64Cu-ATSM exhibited a T/M ratio of 2.29 ± 0.22. In contrast, Yang et al. reported a T/M ratio of 1.52 ± 0.17 for FMISO 1 h after injection for the 13762 rat mammary tumor (26), and Barthel et al. reported 1.8 for 18F-fluoroetanidazole (FETA) in the EMT6 mammary tumor (27). 18F-EF5 showed a T/M ratio of 0.9 in 9L tumor at 1 h after injection (28). In the same 9L tumor, we observed a much higher T/M ratio, 3.43, with 64Cu-ATSM within the same time frame.
The biodistribution results showed early uptake followed by washout of 64Cu-ATSM in the brain. However, there was still relatively higher 64Cu-ATSM uptake in the brain compared with muscle tissue. In R3230Ac tumor-bearing animals, for example, the level of 64Cu-ATSM seen in the brain at steady state is very close to that in the tumor (data not shown). The high brain uptake will make it difficult to use 64Cu-ATSM in differentiating a brain tumor from surrounding normal brain tissue. An additional complication was suggested by the study on Cu-ATSM retention mechanisms by Obata et al. in 2001 (13). They found different 64Cu-ATSM retention mechanisms in tumor cells and brain tissue. 64Cu-ATSM in tumor cells was reduced mainly in the microsome/cytosol and was more dependent on bioreductive enzymes, whereas, in mouse brain, it was reduced primarily in the mitochondria (13). Both high brain uptake of 64Cu-ATSM and uncertainty regarding its retention mechanisms limit the use of Cu-ATSM for brain tumor imaging.
A direct spatial correlation between the distribution of 64Cu-ATSM uptake and that of tumor hypoxia was analyzed in 3 tumor lines in this study. Lewis et al. demonstrated hypoxia-dependent uptake of 64Cu-ATSM in EMT6 mammary adenocarcinoma cells in vitro and heterogeneous uptake of 64Cu-ATSM in EMT6 tumors in vivo (8), but no evidence was provided for an in vivo spatial correlation between 64Cu-ATSM uptake and hypoxia. In 2001, Lewis et al. correlated the uptake of 64Cu-ATSM with hypoxia in 9L glioma in vivo using the oxygen needle probe method (9). However, the 1-dimensional nature of needle electrode–based oxygen measurement does not allow a direct spatial comparison between the tissue uptake of 64Cu-ATSM and the hypoxia level in vivo, particularly given the fact that hypoxia in tumors is usually heterogeneous (29). Dehdashti et al. assessed the 60Cu-ATSM PET in patients with non–small cell lung cancer (10) and cervical cancer (11) and reported a correlation between 60Cu-ATSM uptake and tumor response to radiotherapy. However, there was still no attempt to perform image correlation with location of PO2 measurement in their studies.
Recently, O'Donoghue et al. reported that the uptake of 64Cu-ATSM at 4 h after injection does not reflect the level of hypoxia in an R3327-AT anaplastic rat prostate tumor model, as assessed by pimonidazole immunostaining and oxygen needle probes (30). However, imaging at 16–20 h after injection of 64Cu-ATSM showed a good correlation with the distribution of tumor hypoxia. Their study displayed a temporal evolution of 64Cu-ATSM uptake from 0.5 to 20 h after the injection of radiopharmaceutical. This uptake rate was also dependent on tumor type.
Our study further demonstrates that retention of 64Cu-ATSM is dependent on tumor type. One difference is that O'Donoghue et al. measured the uptake at both 0.5 and 20 h after isotope injection (30), whereas we examined the uptake at 1 h after injection. The choice of the 1-h time point was based on the in vitro uptake study of Lewis et al. (8), in which they measured the uptake of 64Cu-ATSM at different time points until 1 h after the administration of 64Cu-ATSM. 64Cu-ATSM is highly permeable to cell membranes and can reach almost 80% of steady-state uptake in 30 min in vivo according to our study. Therefore, measuring the SUV and T/M ratio at 1 h after injection is quite representative. Although the report of O'Donoghue suggests an interesting approach to solve confounding biodistribution issues, the method of waiting 20 h after drug administration is not clinically practical. Although delayed imaging is realistic when using 64Cu-ATSM (t1/2 = 12.7 h), the ß-emission from 64Cu also subjects patients to significant radiation dose. 60Cu (t1/2 = 23.7 min) and 62Cu (t1/2 = 9.7 min) do not subject patients to extra ß-radiation and are therefore preferred for clinical diagnostic applications. However, the short half-lives of these isotopes preclude delayed imaging.
A possible reason for the low correlation between 64Cu-ATSM uptake and hypoxia distribution in FSA could be the differing redox statuses of the tumor types. FSA tumors may have a lower-than-average redox potential with high concentrations of electron donors. This redox status might be related to the tumor metabolism rate. We recently published results showing that, even though FSA has a comparable level of hypoxia to R3230Ac, it has much higher glucose consumption rate (29). A relatively low redox potential in FSA normoxic cells could lead to 64Cu-ATSM reduction and trapping.
The higher perfusion level surrounding normoxic cells compared with hypoxic cells could also contribute to the higher uptake of 64Cu-ATSM in normoxic FSA cells. Our results from carbogen-breathing experiments have indicated the influence from blood perfusion on the 64Cu-ATSM uptake. We have known that carbogen can enhance the tumor blood flow and, hence, decrease the hypoxia level in FSA-bearing animals. This is consistent with our decreased EF5 binding and the higher initial uptake (first 1 min) we observed in the dynamic uptake curves (data not shown). However, this decrease in hypoxia was not accompanied by the expected decrease in 64Cu-ATSM uptake. A trend of higher 64Cu-ATSM SUV and increased uptake area fraction was observed along with the increased perfusion level. This indicates that the blood perfusion may play a role in 64Cu-ATSM uptake in carbogen-treated FSA tumor. Future research needs to focus on this and other potential mechanisms that may underlie the spurious retention pattern of this marker in FSA.
|
CONCLUSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONReferences |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. Liu, A. Hajibeigi, G. Ren, M. Lin, W. Siyambalapitiyage, Z. Liu, E. Simpson, R. W. Parkey, X. Sun, and O. K. Oz Retention of the Radiotracers 64Cu-ATSM and 64Cu-PTSM in Human and Murine Tumors Is Influenced by MDR1 Protein Expression J. Nucl. Med., August 1, 2009; 50(8): 1332 - 1339. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Skovgaard, M. Kjaer, J. Madsen, and A. Kjaer Noninvasive 64Cu-ATSM and PET/CT Assessment of Hypoxia in Rat Skeletal Muscles and Tendons During Muscle Contractions J. Nucl. Med., June 1, 2009; 50(6): 950 - 958. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. C. Riedl, P. Brader, P. B. Zanzonico, Y. S. Chun, Y. Woo, P. Singh, S. Carlin, B. Wen, C. C. Ling, H. Hricak, et al. Imaging Hypoxia in Orthotopic Rat Liver Tumors with Iodine 124-labeled Iodoazomycin Galactopyranoside PET Radiology, August 1, 2008; 248(2): 561 - 570. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Takasawa, R. R. Moustafa, and J.-C. Baron Applications of Nitroimidazole In Vivo Hypoxia Imaging in Ischemic Stroke Stroke, May 1, 2008; 39(5): 1629 - 1637. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Z. Wong, J. L. Lacy, N. A. Petry, T. C. Hawk, T. A. Sporn, M. W. Dewhirst, and G. Vlahovic PET of Hypoxia and Perfusion with 62Cu-ATSM and 62Cu-PTSM Using a 62Zn/62Cu Generator Am. J. Roentgenol., February 1, 2008; 190(2): 427 - 432. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. A. Torigian, S. S. Huang, M. Houseni, and A. Alavi Functional Imaging of Cancer with Emphasis on Molecular Techniques CA Cancer J Clin, July 1, 2007; 57(4): 206 - 224. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | RSS | TABLE OF CONTENTS |
| JOURNAL OF NUCLEAR MEDICINE TECHNOLOGY | THE JOURNAL OF NUCLEAR MEDICINE |
