|Home | About | Journals | Submit | Contact Us | Français|
Cu-ATSM, a hypoxia imaging agent, has been shown to be predictive of response to traditional cancer therapies in patients with a wide range of tumors. It is known that the environment of the tumor results in a myriad of physiological consequences, including hypoxia, alterations in metabolism and proliferation. In an effort to better characterize the relationships between Cu-ATSM and other prominent radiopharmaceuticals, this current study was undertaken to compare the regional distribution of 64Cu-ATSM with 18F-FMISO, 18F-FDG and 18F-FLT in 9L tumors.
Taking advantage of the different half-life of 18F (t½ = 110 min) in comparison to 64Cu (t½ = 12.7 h), we undertook a dual tracer autoradiography study in 9L tumors. Four groups were examined: (a) 18F-FMISO, 2 h p.i. and 64Cu-ATSM 10 min p.i. (b) 18F-FMISO, 2 h p.i. and 64Cu-ATSM 24 h p.i. (c) 18F-FDG, 1 h p.i. and 64Cu-ATSM 10 min p.i., and (d) 18F-FLT, 1 h p.i. and 64Cu-ATSM 10 min p.i‥ Small animal PET imaging was performed in 9L tumor-bearing rats with imaging on concurrent days comparing 64Cu-ATSM with 18F-FMISO and 18F-FLT.
It was shown that the regional distribution of 18F-FMISO and 64Cu-ATSM showed an excellent correlation when the 64Cu-ATSM had been allowed to distribute for either 10 min (R2 = 0.84) or 24 h (R2 = 0.86). The regional comparisons between 64Cu-ATSM (10 min) and 18F-FDG (1 h) resulted in a very poor correlation (R2 = 0.08) between the regional uptake of the two agents. The comparison between 18F-FLT and 64Cu-ATSM showed a strong relationship (R2 = 0.83) between the two tracers. The small-animal PET images for the distribution comparisons between 64Cu-ATSM and 18F-FMISO and 18F-FLT were in agreement with the data generated from the autoradiography studies.
The data show that it is important to remember that a number of different metabolic situations can exist when considering the relationship between regions of high glucose uptake, proliferation and hypoxia.
The non-invasive imaging of hypoxia is of significant importance, given that the onset of hypoxia in malignant tissues is associated with and influenced by a myriad of complicated physiological processes [1–5]. Copper(II)-diacetyl-bis(N4-methylthiosemicarbazone), or Cu-ATSM, is a hypoxia imaging agent that shows rapid delineation of tumor hypoxia (< 1 hour) in high tumor to background tissue ratios (tumor to blood ratios 2.0) [6, 7]. In clinical studies Cu-ATSM has been shown to be predictive of response to traditional cancer therapies in patients with rectal, lung, and uterine cervix cancer [8–11]. In these same studies concurrent imaging with [18F]fluoro-2-deoxy-d-glucose (18F-FDG) showed no predictive value.
It is obvious that the macroenvironment of the tumor results in a wide range of physiological consequences, including hypoxia, alterations in metabolism and proliferation. An understanding of the relationships among these complicated interactions is of importance to both the basic scientist and the clinician. The relationship between hypoxia, proliferation and metabolism has partly been explored in clinical studies. A comparison between 18F-fluoromisonidazole (18F-FMISO) and 18F-FDG uptake in humans demonstrated that some hypoxic tumors can have modest glucose metabolism, whereas some highly metabolic tumors are not hypoxic, showing discordance in tracer uptake that was tumor type specific . 18F-FDG uptake has been shown to correlate with tumor proliferative rates in lymphomas  and NSCLC . However, Buck et al showed that 18F-fluorothymidine (18F-FLT) correlated significantly better with the proliferative activity of lung tumors than did 18F-FDG .
In an effort to better characterize the relationships between Cu-ATSM and other validated radiopharmaceuticals, this study was undertaken to compare the regional distribution of 64Cu-ATSM with 18F-FDG (metabolism) and 18F-FLT (proliferation). A comparison was also performed with 18F-FMISO, given that the retention mechanisms for these two hypoxia tracers are different. By exploitation of the different half lives of 18F (t½ = 110 min) and 64Cu (t½ = 12.7 h), we undertook a dual tracer autoradiography study in a simple model of cancer, the 9L gliosarcoma model, in order to better understand the regional distribution of each tracer in comparison to each other.
Unless otherwise stated, all chemicals were purchased from Aldrich Chemical Company, Inc. (Milwaukee, WI). All solutions were prepared using distilled, deionized water (Milli-Q®; >18 MΩ resistivity). High specific activity 64Cu was produced on a CS-15 biomedical cyclotron at Washington University School of Medicine via published methods. 64Cu-ATSM was synthesized by methods previously described , and 18F-FMISO was synthesized according to procedures by McCarthy, Dence and Welch . 18F-FLT was produced by a modified version of Machulla et al . All radiopharmaceuticals were used in >98 % radiochemical purity, as determined by radio-TLC and radio-HPLC.
All animal experiments were conducted in compliance with the Guidelines for the Care and Use of Research Animals established by Washington University’s Animal Studies Committee. The 9L cells were a generous gift from the Brain Tumor Research Center, University of California (San Francisco). All studies were performed on 150–165 g female Fisher 344 rats (Charles River Laboratories, Wilmington MA). The Fischer rats were implanted subcutaneously with 1 × 107 9L gliosarcoma cells in the right flank and were used at day 14 of tumor growth. On the day of study the tumors were approximately 1.5 – 2 g, and on postmortem examination had varying degrees of necrosis. In all cases, rats were anesthetized with 1–2% isoflurane (with room air) for administration of radiopharmaceuticals.
All electronic autoradiography was performed on an InstantImager Electronic Autoradiography System from Packard Instrument Co. (Meriden, CT). Four groups of 9L-bearing animals (n = 5 each group) were examined, in order to compare the regional distributions of the fluorinated tracers 18F-FMISO, 18F-FLT and 18F-FDG. The four groups were (a) 18F-FMISO, 2 h post injection and 64Cu-ATSM 10 min post injection (b) 18F-FMISO, 2 h post injection and 64Cu-ATSM 24 h post injection (c) 18F-FDG, 1 h post injection and 64Cu-ATSM 10 min post injection, and (d) 18F-FLT, 1 h post injection and 64Cu-ATSM 10 min post injection. All rats received an i.v. injection of 1.0 – 1.2 mCi of the required 18F tracer, followed by 64Cu-ATSM. The administration of 50 – 60 µCi of 64Cu-ATSM occurred after 110 min for (a) and 50 min after for (c) and (d). For (b), each rat (n = 5) received an i.v. injection of 240 µCi of 64Cu-ATSM, followed 22 h later by 18F-FMISO. After euthanasia, the tumors were excised and frozen whole in Miles, Inc., Tissue-Tek Embedding Medium (Elkhart, IN). Slices (1 mm thick) were mounted and placed in the InstantImager® in order to visualize the distribution of radioactivity. The slices were not washed at any time and were not treated with any preservative. Autoradiography of the slices initially visualized the 18F localization (1 or 2 h distribution) and was repeated 24 h later, in order to visualize the 64Cu-ATSM (10 min or 24 h distribution). Standard radioisotopic mixtures were also made and imaged in order to ensure that >95% of the initial image was generated by the 18F and that only 64Cu activity remained in the 24-h images.
To generate a comparison between the focal uptake of 64Cu-ATSM and the 18F-agents, a ‘template’ was generated from the 18F image by use of the equipment software to yield the net % max uptake in each defined region. Briefly, the template generated by the software consists of an x–y grid that contains regions of interest (ROI’s) that have been determined by threshold levels defined automatically. The system uses three user-specified parameters to locate and create regions of interest based on separation, sensitivity and background. As a control in this study, regions of interest were also defined manually. The manual operation was performed where every individual tumor slice was first analyzed for the absolute number of counts. The ROIs based on separation were defined as percentages of the overall number of counts within the tumor ROI that showed the highest concentration of activity with fixed levels of sensitivity and background for individual slices. For both the manual and automatic analysis of the autoradiographs, identical templates and data were obtained.
The template generated for each 18F image was overlaid on the complementary 64Cu-ATSM image (i.e., the same tumor slice after 18F decay) in order to generate the 64Cu distribution data. As a control, a template was also generated on the 64Cu image and then overlaid on the complementary 18F image to generate the 18F distribution data. The data presented (net % max) are defined as [a/b ×100 %], where a = number of counts/mm2 in a defined ROI within a tumor slice and b = number of counts/mm2 in the ROI with the highest number of counts from the same tumor slice. These analyses were performed on 10 tumor slices for each study group. Each tumor slice had at least 10 ROIs defined within its boundaries.
All small-animal PET imaging was undertaken on the microPET-Focus 220  (Concorde MicroSystems Inc., Knoxville, TN) and were coregistered with CT images from a MicroCAT II System (ImTek Inc., Knoxville, TN). Isoflurane (1–2%) was used as an inhaled anesthetic to induce and maintain anesthesia during imaging. Drugs were administered via the tail vein. Two separate studies were performed in 9L bearing rats (n = 4 per group). (A) On Day 1, 150 µCi of 64Cu-ATSM were injected into the animals and imaged 10 min post-injection. The animals were allowed to waken and on Day 2 (28 hours after 64Cu-ATSM) were injected with 1.5 mCi 18F-FMISO and imaged at 2 h post-injection; (B) On Day 1, 150 µCi of 64Cu-ATSM were injected into the animals and imaged 10 min post-injection. The animals were allowed to waken and on Day 2 (28 hours after 64Cu-ATSM) were injected with 1.5 mCi 18F-FLT and imaged at 1 h post-injection. All PET scans consisted of a single static 5-min collection. The image registration between CT and PET images was accomplished by using a landmark registration technique, AMIRA image display software (AMIRA, TGS Inc, San Diego, CA). The registration method proceeds by rigid transformation of the microCT images from landmarks provided by fiducial markers directly attached to the animal bed.
Shown in Figure 1A and 1B are representative sample slices from the autoradiographic studies comparing distribution of 18F-FMISO (2 h) and 64Cu-ATSM (10 min or 24 h). Figures 2A and 2B demonstrate the correlation of the regional uptake of 64Cu at both time points with the 2-h distribution of 18F-FMISO. It is evident, both through visual analysis and quantitative comparison, that the regional distributions of 18F-FMISO and 64Cu-ATSM show an excellent correlation when the 64Cu-ATSM has been allowed to distribute for either 10 min (R2 = 0.84) or 24 h (R2 = 0.86). It is also evident from Figure 1B that Cu-ATSM at 24-h results in a more intense uptake within small discreet regions of the tumor volume, more so than is seen with the corresponding 18F-FMISO image.
Figure 1C shows the regional comparisons between 64Cu-ATSM (10 min) and 18F-FDG (1 h). Visually, it appears that there are regions within the tumor with uptake similar for both tracers but also regions in which there is a complete discrepancy of uptake, where uptake of Cu-ATSM is not mirrored by uptake of 18F-FDG. The quantitative analysis of these images (Figure 2C) confirms this, as demonstrated by a poor correlation (R2 = 0.08) between the regional uptake of the two agents. The uptake of 18F-FLT and 64Cu-ATSM appear visually identical (Figure 1D), which is confirmed by quantitative analysis (R2 = 0.83) (Figure 2D).
The small-animal PET images for the distribution comparisons between 64Cu-ATSM and 18F-FMISO and 18F-FLT are given in Figure 3A and 3B, respectively. As with the autoradiography studies, similarities in regional uptake are noted between the two tracers even after 10 min of Cu-ATSM distribution. The images comparing 64Cu-ATSM and 18F-FLT are also in agreement with the data generated from the autoradiography studies. The intense uptake of tracers is on the tumor edge (the tumor had a histologically confirmed necrotic center).
One of the most popular agents for the PET imaging of hypoxia has been 18F-FMISO [12, 20, 21]. Although Cu-ATSM has been validated for use in vitro and in vivo in multiple tumor models  and it is a very effective PET agent for clinically delineating many hypoxic human malignancies [8–11], there has been a concern that its ability to delineate hypoxia may be tumor-dependent. In 2005, it was shown in vitro that the hypoxia-selectivity of 64Cu-ATSM was found to be cell line dependent. It was demonstrated that there was variation in the 64Cu cellular accumulation, with uptake in normoxic cells being anywhere from two to nine times lower than that in hypoxic cells, depending upon the cell line . This study was followed by an intricate in vivo comparison of 64Cu-ATSM and 18F-FMISO in tumor-bearing rats with direct pO2 tumor measurements, autoradiography and fluorescent microscopy . In rats bearing the R3327-AT rat prostate tumor, there was a poor correlation between the intra-tumoral distribution of 18F-FMISO and 64Cu-ATSM except at later times (16–20 hr post injection). However, in the same study 18F-FMISO and 64Cu-ATSM images were also acquired in nude rats bearing xenografts derived from the human squamous cell carcinoma cell line, FaDu. For the FaDu tumor model, the early and late 64Cu-ATSM PET images were similar and were in general concordance with the 18F-FMISO scans. Following these discrepancies, a recent publication showed that the ability of Cu-ATSM to delineate hypoxia was suspect primarily in prostate tumors in a manner that that was directly related to the expression of fatty acid synthase . It was, therefore, important to continue to validate the use of 64Cu-ATSM against 18F-FMISO in this 9L gliosarcoma model. A previous study compared the autoradiographic distributions of 64Cu-ATSM with a well-established hypoxia marker drug (EF5) in 9L tumors; there was a close correlation between 64Cu- ATSM uptake and hypoxia in 9L tumors .
In the current study a comparison was made between the distribution of 18F-FMISO (2 h) and early and late distribution time points for 64Cu-ATSM (10 min and 24 h). Our data show that the distribution of 64Cu-ATSM at both 10 min and 24 h correlates highly with the 2-h distribution of 18F-FMISO. The similarity of the image quality of the radionuclides has been demonstrated (Figure 1A and 1B) and it is, therefore, interesting that the 64Cu-ATSM images show a more focal (localized) distribution of activity. Observed differences in distribution may be attributed to differences in retention mechanisms for the agents and the poor target:background ratios associated with the nitromidazoles. The distribution of 64Cu-ATSM at 24 h does have a greater correlation with 18F-FMISO than the 10 min distribution; however, these differences are not significant and confirm that examination of 64Cu-ATSM at earlier time points is suitable for the delineation of hypoxia within tumors.
18F-FDG is utilized widely in many aspects of PET diagnostic medicine, including cancer diagnosis and diseases of the brain and heart. The mechanism of 18F-FDG trapping follows the well-documented glucose biochemical pathway. The direct measurement of glucose metabolism with 18F-FDG also yields valuable information about tumor localization and quantitation. A previous dual-tracer study with 64Cu-ATSM and 18F-FDG in VX2 tumors (implanted into Japanese white rabbits) showed that the major accumulation of 64Cu-ATSM was observed around the outer rim of the tumor masses, which consisted mainly of active hypoxic cells, and 18F-FDG was distributed more widely with highest levels in the inner regions where pre-necrotic cells were mainly observed . An additional study compared the intratumoral distribution of 64Cu-ATSM and 18F-FDG in four mice-tumor models (LLC1, Meth-A, B16 and colon26) with the immunohistochemical staining of proliferating cells (Ki67), blood vessels (CD34 or von Willebrand factor), and apoptotic cells (terminal deoxynucleotidyltransferase-mediated dUTP nick end labeling method) . It was revealed that there are regions within the tumor that contain cells of different phenotypes that can be distinguished by the use of 64Cu-ATSM and 18F-FDG .
This current study demonstrates that 9L tumors contain regions of high metabolism that are not very hypoxic and vice versa. It is reasonable and possible for regions of high 18F-FDG uptake to also show both high and low 64Cu-ATSM uptake. Tumor cells in close proximity to the vasculature would presumably have high 18F-FDG uptake, whereas cells remote from the vascular bed could be hypoxic. It is, therefore, possible that both conditions could exist within the same tumor. These data are in agreement with the previously reported animal  and clinical data [8–10, 12].
18F-FLT has been proposed as a PET tracer for cell proliferation [28, 29]. It is retained in proliferating cells after phosphorylation by thymidine kinase (TK1). 18F-FLT is taken up by cells and phosphorylated by TK1, which leads to intracellular trapping within the cell. The retention of FLT within the cell provides a measure of cellular TK activity, an enzyme that is closely tied to cellular proliferation. It is apparent that the relationships between proliferation, hypoxia and metabolism are complicated. Studies that compare proliferation with hypoxia markers have led to a wide range of results. In an effort to characterize the distribution of hypoxia and proliferation in human squamous cell carcinoma of the cervix via an immunohistochemical approach, it was shown in 1997 that there was adirect relationship between hypoxia measured by pimonidazole binding and proliferation markers . In a clinical imaging study in 2006, the relationship between 18F-FMISO and tumor markers of hypoxia, proliferation, and angiogenesis showed a weakly positive correlation between both 18F -FMISO (R2 = 0.64) and 18F-FDG uptake (R2 = 0.75) and the proliferative marker Ki67 in peripheral tumor, with no correlation seen in the central tumor .
The results in this current study would suggest, at least in the 9L tumor model, that regions of hypoxia as delineated by 64Cu-ATSM also demonstrate increased levels of proliferation as shown by the kinetics of 18F-FLT. This is in contrast to the observation by Tanaka et al that abundant positive nuclear staining of Ki67 in tumor cells was observed in the regions of high 18F-FDG uptake, and was hardly observed in regions of 64Cu-ATSM accumulation , and, similar to microvessel density, the number of Ki67 positive cells increased with 18F-FDG uptake but decreased with 64Cu-ATSM localization. The reasons for this discrepancy on the macro-level can be related to the fact that hypoxia might develop in different tumors through mechanisms unrelated to oxygen supply. Freyer et al., have demonstrated in spheroids that the oxygen consumption rate may increase as much as 5-fold in cells undergoing active proliferation [32, 33]. It has also been shown that angiogenesis occurs in tumors of a very small size (100 cells) , and it is reasonable to assume that high oxygen consumption rates in tumors could stimulate angiogenesis by the lowering of pO2. Moreover, in well-developed tumors, regions with high proliferative rates may create a locoregional hypoxia that is not directly related to the degree of vascularity, perfusion, or oxygen supply.
This present study highlights that a number of different situations can exist when considering the complex relationships between regions of high metabolism, proliferation and hypoxia. The 9L tumor exhibits many of the physiological attributes that could be apparent in the clinical situation: tumors with regions of varied metabolism, proliferation and hypoxia which may, or may not, be related to each other but are not discernible due to the resolution restrictions of the imaging modality. Specifically, the 9L tumors contain regions of high metabolism (high FDG uptake) that are not very hypoxic (low Cu-ATSM retention) and vice versa. Also seen are regions that demonstrate increased uptake of both tracers and vice versa. In regard to proliferation, regional hypoxia within the 9L tumor strongly correlates to increased levels of proliferation as shown by the kinetics of 18F-FLT. It is apparent that Cu-ATSM is a clinically relevant PET agent that has enormous value in the imaging of oncological hypoxia [6–11] and it has clearly been shown with Cu-ATSM that the 9L tumor has regions of hypoxia within its margins. These regions, at the resolution of the autoradiography instrumentation, are also highly proliferative but have varied levels of glucose metabolism.
This work was supported by the United States Department of Energy (DE-FG02-87ER60212). The production of copper radionuclides at Washington University is supported by a grant from the National Cancer Institute (1 R24 CA86307). We wish to thank Debbie Sultan, Todd Perkins and Thomas Voller for production of 64Cu, John Engelbach, and Lynne Jones for technical assistance, and Bill Margenau for cyclotron support. We also thank Dr. Joseph B. Dence for his editing of this manuscript.
This work was supported by a grant from the U. S. Department of Energy (Grant DE FG02-87ER60512). The production of copper radionuclides at Washington University is supported by a grant from the National Cancer Institute (Grant 1 R24 CA86307).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.