Imaging of hypoxia in mouse atherosclerotic plaques with ⁶⁴Cu-ATSM

Abstract

Introduction

Cardiovascular disease is the leading cause of death in the United States. The identification of vulnerable plaque at risk of rupture has been a major focus of research. Hypoxia has been identified as a potential factor in the formation of vulnerable plaque, and it is clear that decreased oxygen plays a role in the development of plaque angiogenesis leading to plaque destabilization. The purpose of this study is to demonstrate the feasibility of copper-64 labeled diacetyl-bis (N⁴-methylthiosemicarbazone) (⁶⁴Cu-ATSM), a positron-emitting radiopharmaceutical taken up in low-oxygen-tension cells, for the identification of hypoxic and potentially unstable atherosclerotic plaque in a mouse model.

Methods

⁶⁴Cu-ATSM PET was performed in 21 atherosclerotic apolipoprotein E knockout (ApoE^−/−) mice, 6 of which were fed high-fat diet (HFD) while the others received standard-chow diet (SCD), and 13 control wild type mice fed SCD. 4 SCD ApoE^−/− mice and 4 SCD wild type mice also underwent ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG) positron emission tomography (PET) imaging one day prior to ⁶⁴Cu-ATSM PET.

Results

⁶⁴Cu-ATSM uptake was increased in the aortic arch in SCD ApoE^−/− mice (average aortic arch/muscle (A/M) standardized uptake value [1] ratio 7.5–30 minutes post injection: (5.66 ± 0.23) compared to control mice (A/M SUV ratio 7.5–30 minutes post injection (3.87 ± 0.22), p < 0.0001). HFD ApoE^−/− mice also showed similarly increased aortic arch uptake on PET imaging in comparison to control mice. Immunohistochemistry in both HFD and SCD ApoE^−/− mice revealed noticeable hypoxia by pimonidazole stain in atherosclerosis which was co-localized to macrophage by CD68 staining. Autoradiography assessment demonstrated the presence of hypoxia by ⁶⁴Cu-ATSM uptake correlated with pimonidazole uptake within the ex vivo atherosclerotic aortic arch specimens. A significant increase in ¹⁸F-FDG uptake in the SCD ApoE^−/− mice in comparison to controls was also observed at delayed time points.

Conclusion

This pre-clinical study suggests that ⁶⁴Cu-ATSM is a potential PET tracer for hypoxia imaging in atherosclerosis.

1. Introduction

Cardiovascular atherosclerosis is the leading cause of death worldwide despite primary and secondary prevention. Cardiovascular disease impacts the public broadly and affects an estimated 80 million people in the United States [2]. Each year more than a million people in the United States experience a sudden cardiac event (acute coronary syndrome or sudden cardiac death) and an additional 700,000 experience stroke [3]. Diabetes mellitus, another leading cause of morbidity and mortality in Western nations, is strongly associated with accelerated atherosclerosis – 86% of patients with newly diagnosed diabetes have atherosclerosis and as blood glucose levels increase, the incidence of clinical or preclinical atherosclerosis also increases [4].

At present, there is a lack of an effective approach to monitor non-obstructive asymptomatic but potentially unstable atherosclerosis. Although ¹⁸F-Fluorodeoxyglucose (¹⁸F-FDG) PET has been shown to detect vascular atheroma [5], it provides limited value for evaluating the coronary arteries due to the confounding effects of myocardial uptake of the radiotracer. Cu-ATSM was the first bis(thiosemicarbazone) complexes demonstrated to exhibit hypoxia selectivity [6–8], and has been extensively investigated for characterizing hypoxic tumors [9–14] in addition to a few studies addressing its potential application for imaging ischemic or hypoxic myocardium [15–17]. If successful, copper-64 labeled diacetyl-bis (N⁴-methylthiosemicarbazone) (⁶⁴Cu-ATSM) imaging would overcome this limitation.

In this study, we have assessed the feasibility of imaging the development of hypoxia in an apolipoprotein E knockout (ApoE KO or ApoE^−/−) mouse model of atherosclerosis using the PET imaging agent ⁶⁴Cu-ATSM. Lack of ApoE results in severe hypercholesterolemia and spontaneously developed atherosclerotic lesions. Therefore, this mouse model is widely used to study atherosclerosis [18, 19]. Using ⁶⁴Cu-ATSM to target hypoxic cells within plaque is of interest to us because this imaging radiopharmaceutical has been used extensively in preclinical models and human subjects to assess tumor hypoxia in oncology but its use in cardiology has so far been limited to assessing myocardial ischemia [8, 16, 17, 20]. However, there is growing evidence that the combination of increased oxygen demand together with impaired oxygen diffusion capacity also results in the presence of severe hypoxia (<1% oxygen) in macrophage-rich zones into the lesion [21–24]. Leppanen et al. [25] suggests that over time macrophages within the plaque core become ATP-depleted and severely hypoxic, contributing to their death and formation of a necrotic core that increases angiogenesis and plaque destabilization. The direct presence of hypoxia in the macrophage-rich center of human carotid atherosclerosis has been demonstrated by Sluimer et al. [23], and there is some preliminary data to suggest that hypoxia may be the cause for increased glucose utilization, and thus, increased ¹⁸F-FDG activity in atherosclerosis [26]. A recent study successfully assessed hypoxia in rabbit advanced atherosclerosis using ¹⁸F-fluoromisonidazole (¹⁸F-FMISO) PET imaging [27]. Our study is to investigate the feasibility of ⁶⁴Cu-ATSM for the identification of hypoxic and potentially unstable atherosclerotic plaque in a mouse model. In comparison to ¹⁸F-FMISO, ⁶⁴Cu-ATSM has two major advantages: higher cellular uptake and quicker washout from normoxic tissue [28]. The more efficient uptake and washout kinetics of ⁶⁴Cu-ATSM in hypoxic and normoxic cells offers the possibility of a faster and more selective means of detecting hypoxia by PET imaging. Although different schemas in the retention mechanism of Cu-ATSM were proposed in the literature, but, to date, there is a consensus that in vitro Cu-ATSM undergoes bioreductive trapping under hypoxic conditions [8, 29–31]. As presented by Fujiyabashi et al. [8, 17], in theory Cu-ATSM Cu(II) is a lipophilic molecule, with high membrane permeability and low redox potential, and is reduced only in hypoxic cells and involves electron transfer from hyper-reduced Complex I (ubiquinone oxidoreductase) using NADH as a two-electron donor. Upon the intracellular reduction, Cu-ATSM becomes trapped irreversibly and thus the radioactivity accumulates in these areas. In normoxic cells, Complex I is incapable of reducing Cu-ATSM. In addition, we will use the hypoxia-specific dye pimonidazole and the pan macrophage marker CD68 to assess for co-localization of hypoxia in macrophages. Pimonidazole is reduced in cells with low oxygen tension to form protein adducts that can be detected by immunostaining [32]. We hypothesized that there would be high PET image contrast between hypoxic and normoxic tissues after the administration of ⁶⁴Cu-ATSM, and that the presence of hypoxia by ⁶⁴Cu-ATSM uptake will correlate with pimonidazole uptake within the ex vivo atherosclerotic arch specimens. We also hypothesized that pimonidazole uptake would co-localize to macrophages (CD68) on immunohistochemistry.

2. Materials and Methods

General

All animal studies were performed under a protocol approved by the Animal Studies Committee at our institution. Ten-month-old wild type (C57Bl/6, n=7) mice weighting ~30g were purchased from Charles River Laboratories (Wilmington, MA) and fed standard-chow diet (SCD). One-year-old ApoE^−/− mice (n=9) spontaneously expressing atherosclerosis were fed SCD for one year and had roughly the same weight as the wild type mice prior to ⁶⁴Cu-ATSM PET imaging. A subset of ApoE^−/− mice (n = 4) and wild type mice (n = 4) also underwent ¹⁸F-FDG PET imaging in addition to ⁶⁴Cu-ATSM PET imaging one day later. Additional SCD wild type mice (n = 6) and ApoE^−/− mice fed either with SCD (n = 6) or a high-fat Western type diet (n = 6; HFD, containing 21.2% fat and 0.2% cholesterol; Harlan Laboratories, Inc., Indianapolis, IN) underwent longitudinal imaging. These mice underwent PET imaging in 2-month intervals for a total of 10 months. This was done in order to assess the uptake of ⁶⁴Cu-ATSM at various stages of plaque development. The experimental arms are given below in Table 1. All chemicals, unless otherwise stated, were purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO). Water was distilled and then deionized (18 MΩ/cm²) by passing through a Milli-Q water filtration system (Millipore Corp., Milford, MA).

Table 1.

Experimental cohorts

Preliminary Mouse Study (N = 16)
Number of Mice	Species/Diet	Age of Imaging	Radiopharmaceuticals
4	ApoE−/−, SCD	54 –62 (day 1)	18F-FDG
9	ApoE−/−, SCD	54 –62 (day 2)	64Cu-ATSM
4	C57Bl/6, SCD	43 (day 1)	18F-FDG
7	C57Bl/6, SCD	43 (day 2)	64Cu-ATSM
Longitudinal Mouse Study (N = 18)
6	ApoE−/−, HFD	14, 23, 32, 41	64Cu-ATSM
6	ApoE−/−, SCD	14, 23, 32, 41	64Cu-ATSM
6	C57Bl/6, SCD	14, 23, 32, 41	64Cu-ATSM

Copper-64 was prepared as previously described [33, 34]. Radioactivity was counted with a Beckman Gamma 8000 counter containing a NaI crystal (Beckman Instruments, Inc., Irvine, CA). EM Science Silica Gel 60 F254 thick-layer chromatography (TLC) plates (10 × 5 cm) were purchased from EMD Millipore Corporation (Billerica, MA). Radio-TLC plates were analyzed using a BIOSCAN System 200 imaging scanner (Bioscan, Inc., Washington, DC).

Diacetyl-bis (N⁴-methylthiosemicarbazonate) [H₂-ATSM] was labeled with ⁶⁴Cu using the methods developed in house. Briefly, 10 mg of ATSM powder was dissolved in 1 ml of DMSO to give an overall concentration of 10 mg/mL. A total of 1 μL (10 μg) of this solution was then added to 1 μL (~92.5 MBq) of ⁶⁴CuCl₂ in 200 μl of 1 M NaOAc or 3 M HCl buffer, at a concentration of 0.46 MBq/μL. This solution was then stirred for 3 to 5 minutes and then allowed to sit briefly at room temperature, after which it was loaded onto a C₁₈ Sep-Pak® Light (Waters Corporation, Milford, MA), which had been pre-conditioned with 5 mL of ethanol and 5 mL of water. After loading the sample, 10 mL of water was passed through to remove the DMSO and any unreacted ⁶⁴Cu. The labeled complex was then eluted in 350 μL of ethanol (following a 150 μL ethanol elution for the void volume). The purity of the labeled material was determined by radio-TLC using silica gel plates with ethyl acetate as the mobile phase. Radiochemical yield was 75–85% and purity > 95%.

Dynamic Small Animal PET Imaging

Small animal imaging studies were carried out on a Siemens Inveon (Siemens Medical Solutions USA, Inc., Malvern, PA) or Focus 220 microPET scanner (Concorde MicroSystem Inc., Knoxville, TN). Isoflurane (1–2%) was used to induce and maintain anesthesia during imaging. Each animal received ~3.7 MBq of activity in 100 μL saline via lateral tail-vein injection. Mice were imaged in pairs in a supine position in a specially designed bed. Imaging was performed using 0–30 minute dynamic scan sessions for ⁶⁴Cu-ATSM or 0–60 minute dynamic scan sessions for ¹⁸F-FDG. PET data was reconstructed using standard methods with the maximum a posteriori probability (MAP) algorithm. ⁶⁴Cu-ATSM images were reconstructed in 10 second intervals up to 2 minutes, 30 seconds intervals up to 5 minutes, and 5 minute intervals up to 30 minutes. ¹⁸F-FDG scans were reconstructed in 2–5 seconds intervals up to 1 minute p.i., 10–30 seconds intervals up to 4 minutes p.i., 60–120 second intervals up to 10 minutes p.i. and 5 minute intervals up 1 hour. CT was performed prior to each PET study for attenuation correction and anatomical co-localization of the PET data.

Image Analysis

Images were evaluated by measuring the radioactivity concentration of the aortic arch and non-target (thigh muscle) of each mouse on the co-registered PET/CT images using the Inveon Research Workplace (IRW; Siemens Medical Solutions USA, Inc., Malvern, PA).

Regions of interest (ROIs) were drawn around the entire aortic arch (~15 mm³) or a region of ~50 mm³ on the thigh muscle on multiple consecutive, transaxial image slices. Time activity curves (TACs) in kBq/cm³ were generated by plotting the ROI values over time.

Standardized Uptake Values (SUVs) were calculated by dividing the decay-corrected activity per unit volume of tissue (Bq/cm³) by the injected activity per unit of body weight (Bq/g), as described by the following equation:

$$\text{SUV}=\frac{\text{radioactivity}\text{concentration}(Bq/c{m}^3)}{\text{injected}\text{dose}(Bq)/\text{body}\text{weight}(g)}$$

The aortic arch-to-leg muscle (A/M) SUV ratios were calculated and compared.

Statistical Methods

Graph generation and statistical analysis were performed using GraphPad Prism version 6 (GraphPad Software Inc., La Jolla, CA). Differences between groups were evaluated by Student’s t-test (two groups) or one-way ANOVA (three or more groups) using Fisher’s least significant difference method for post-hoc test. Data were reported as mean ± S.E.M. (standard error of mean) unless otherwise indicated and differences at the 95% confidence level (P < 0.05) were considered statistically significant.

Immunohistochemistry

4 mice from each subset (wild type, ApoE^−/− fed SCD or HDF) were sacrificed to collect ex-vivo endarterectomy specimens for histopathological assessment and pimonidazole immunohistochemistry. The animals were injected with the hypoxia-reactive reagent pimonidazole hydrochloride (Hypoxyprobe^TM-1 Kit, Hypoxyprobe, Inc, Belmont, MA) 60–90 minutes before euthanasia. Specimens were embedded in paraffin and sectioned transversely. Serial 5 μm sections of even intervals were placed on slides for immunohistochemistry and histology with hemotoxylin and eosin (H&E) stains. Aortic arch specimens and leg muscle tissue specimens were stained for the presence of hypoxia using an antipimonidazole antibody (Natural Pharmacia Inc., Belmont, MA) and mounted in solution containing 4′-6-diamidino-2-phenylindole (Vector Laboratories, Palo Alto, CA). Although pimonidazole may react with reactive oxygen species [35], the antipimonidazole antibody only recognizes hypoxia derivatives [36].

Rat anti mouse CD68 antibody (AbD Serotec, Bio-Rad Laboratories, Hercules, CA) which recognizes mouse macrosialin for the identification of CD68 was used to stain atherosclerosis-associated macrophages [37–42] in the paraffin-embedded aortic arch of mouse, and the result was co-localized to the pimonidazole staining. Fluorescein isothiocyanate [43] and Cy3 fluorescent labeled antibodies were used as the secondary fluorescent antibodies, respectively. The slides were incubated with DAPI (4, 6-diamino-2-phenylindole) (Fluoro-Gel II with DAPI, electron microscopy sciences, Inc., Hatfield, MA) to show cell nuclei. Negative control staining was performed by replacing the primary antibody with matching isotype control followed by same fluorescent labeled secondary antibody. Immunohistochemistry slides were imaged by Leica TCS SP5 confocal laser scanning microscope (Leica Microsystems Inc., Buffalo Grove, IL).

Autoradiography

Specimens from 4 SCD ApoE^−/− mice and 4 wild-type control mice were collected for autoradiography. The animals were injected with pimonidazole hydrochloride 60–90 minutes before euthanasia. Harvested fresh tissues of aortic arch (with heart attached) were fixed in 4% Paraformaldehyde (PFA) for ~12 hours and cut transversely into a series of adjacent sections on a cryostat (Leica) at 20 μm thickness and attached to adhesive glass slides (CFSA 1X, Leica Microystems, Buffalo Grove, IL). Both the aortic arch of the control mice and the heart of the same ApoE^−/− mice were used as negative controls. The slides used for autoradiography were covered by adhesive tape (CryoJane Tape Transfer System^®, Leica Microsystems, GmbH, Nussloch, Germany) and exposed to a phosphor imaging plate (GE Healthcare Life Sciences, Pittsburgh, PA) for 12 hours at −20°C, and the plates were scanned using phosphor imager plate scanner (Storm^® 840, Amersham Biosciences Corp., Piscataway, NJ)). The resulting images were processed using ImageQuant 5.2 (Molecular Dynamics, Inc, Sunnyvale, CA) and ImageJ (v1.48, public domain) software. After decaying for at least 3 days, the adjacent slides were treated with pimonidazole and CD68 staining for the presence of hypoxia and macrophages. Due to the different resolution of the immunohistochemistry image obtained from the confocal microscopy (< 5μm) and the autoradiograph obtained from the plate scanner (50 μm), the autoradiograph were scaled up 11.2 times with bilinear interpolation to match the size of the immunohistochemistry images in the same ROI for comparison.

3. Results

⁶⁴Cu-ATSM was distributed rapidly in the blood pool and quickly excreted; thus activity was found in the kidney and liver at all time-points post injection likely due to clearance. ROIs were drawn on the PET and CT co-registered images over the aortic arch and an area of leg muscle on each mouse after administration of ⁶⁴Cu-ATSM or ¹⁸F-FDG (Figure 1).

Figure 1.

Representative small animal PET images from SCD ApoE^−/− and wild type (control) mice. Summed images of coronal (top) and transverse (bottom) slices 7.5–30 minutes post injection (p.i.) of ⁶⁴Cu-ATSM (left) and 40–60 minutes p.i. of ¹⁸F-FDG (right) with ROIs of aortic arch (in blue circles) drawn on each mouse.

7.5 minutes post injection of ⁶⁴Cu-ATSM, the radioactivity in the aortic arch and leg muscle was stable (Figure 2A), while the difference in time activity curves (TACs) of aortic arch between the SCD ApoE^−/− mice and control mice increased over time from 15 minutes post injection of ¹⁸F-FDG until the end of imaging (Figure 2B). There was no significant difference in activity in leg muscle (background) (Figure 2, A and B). The average SUVs of the aortic arch 7.5–30 minutes post injection of ⁶⁴Cu-ATSM were significantly higher in the SCD ApoE^−/− mice (n = 9) than the wild type mice (n = 7): 7.80 ± 0.41 and 5.45 ± 0.52, respectively (p = 0.004), as were average aortic arch to muscle (A/M) SUV ratios of the SCD ApoE^−/− mice and the wild type mice: 5.66 ± 0.23 and 3.87 ± 0.22, respectively (p < 0.0001) (Figures 3A). For the mice injected with ¹⁸F-FDG, the A/M SUV ratios between SCD ApoE^−/− mice and control mice became statistically significant at about 40 minutes post injection of ¹⁸F-FDG (Figure 3B). The 40–60 minutes p.i. average A/M SUV ratio difference between SCD ApoE^−/− mice (n = 4, 8.07 ± 0.96) and control mice (n = 4, 4.55 ± 0.21) after ¹⁸F-FDG injection (Figure 4) was also significant (p < 0.005), as was the average A/M SUV ratio difference (5.57 ± 0.38 V.S. 3.67 ± 0.61 ) 7.5–30 minutes after ⁶⁴Cu-ATSM for the same subset of mice.

Figure 2.

Time activity curves (TACs) of ⁶⁴Cu-ATSM (A) and ¹⁸F-FDG (B) show increased difference in aortic arch activity between the SCD ApoE^−/− and wild type mice injected with ⁶⁴Cu-ATSM in comparison to ¹⁸F-FDG injection. Results are presented as mean ± S.E.M. in the graph. The numbers of ApoE^−/− and control mice were 9 and 7 for ⁶⁴Cu-ATSM imaging, and 4 and 4 for ¹⁸F-FDG imaging, respectively.

Figure 3.

A/M SUV ratios observed with small animal PET imaging after injection of ⁶⁴Cu-ATSM (A) and ¹⁸F-FDG (B) in SCD ApoE^−/− mice and control mice. ⁶⁴Cu-ATSM injection yields significant difference in the ratios at a faster rate in comparison to ¹⁸F-FDG injection, which increases over time from 7.5 minutes p.i. to the end of imaging. Results are presented as mean ± S.E.M. in the graph. The numbers of ApoE^−/− and control mice were 9 and 7 for ⁶⁴Cu-ATSM imaging, and 4 and 4 for ¹⁸F-FDG imaging, respectively.

Figure 4.

Comparison of A/M SUV ratios ± S.E.M. for the same subset of SCD ApoE^−/− mice (n = 4) and control mice (n = 4) 7.5 – 30 minutes p.i. of ⁶⁴Cu-ATSM and 40–60 minutes p.i. of ¹⁸F-FDG. Significant differences were observed between ApoE^−/− mice and control mice for both ⁶⁴Cu-ATSM and ¹⁸F-FDG imaging.

Immunohistochemistry of the aortic arch in ApoE^−/− mice showed the presence of atherosclerosis whereas immunohistochemistry of the aortic arch of wild type mice did not. Pimonidazole uptake was found in the aortic arch of both SCD and HFD ApoE^−/− mice (Figure 5 and Supplemental Figure 1). There was large overlap of the fluorescent areas corresponding to the CD68 macrophage staining co-localized to the pimonidazole adducts, demonstrating co-localization of hypoxia to macrophages (Figure 5, Supplemental Figure 1). Wild type aortic specimens did not demonstrate hypoxia by pimonidazole staining (Supplemental Figure 2). The DAPI staining for nuclei delineated and confirmed the cellular components of the plaques imaged.

Figure 5.

Detection of pimodiazole adducts co-localizes with macrophages within atherosclerotic plaque of aortic arch. Aorta were harvested from 40 week old ApoE^−/− mouse on western diet for 33 week that received pimodiazole 2 hours prior to sacrifice. (A) Representative image of hematoxylin and eosin stained section from aortic arch. (B) Magnification of region of interest from (A). (C) Immunofluorescence image from serial section of aortic arch shown in (A). PIMO is pimonidazole. Green is mAb 4.3.11.3 staining pimodiazole adducts (internal elastic lamina is autofluorescent), red is CD68 staining for macrophages, blue is DAPI nuclear stain, and yellow is the overlay of pimonidazole and CD-68. Representative of at least three animals.

Autoradiographs from different transverse slides confirmed that SCD ApoE^−/− mice had higher uptake of ⁶⁴Cu-ATSM in comparison to the aortic arch of control mice as well as the heart of the same ApoE^−/− mice (Figure 6A). The hot spots in the scaled-up autoradiograph (Figure 6B) regionally co-localized to the immunohistochemistry image of the adjacent slide (Figure 6, C–E).

Figure 6.

(A) Representative autoradiographs obtained from transverse slices of the aortic arch in SCD ApoE^−/− mice (top) and control mice (middle) and heart in SCD ApoE^−/− mice (bottom) 30 minutes p.i. of ⁶⁴Cu-ATSM. (B) The section in the yellow square in (A) were scaled-up 11.2 times with bilinear interpolation. The bright spots in scaled-up autoradiograph were co-localized in shape to the (C) pimonidazole and (D) CD-68 strongly-positive area and their merged area. PIMO is pimonidazole. Green is mAb 4.3.11.3 staining pimodiazole adducts (internal elastic lamina is autofluorescent), red is CD68 staining for macrophages, blue is DAPI nuclear stain, and yellow is the overlay of pimonidazole and CD-68.

The results of our longitudinal studies comparing ⁶⁴Cu-ATSM uptake in SCD and HFD ApoE^−/− mice at various time points suggest that both groups of mice developed hypoxic atherosclerosis longitudinally over time. There was no significant difference in A/M SUV ratios between the three groups of mice at 14 weeks old, suggesting that no advanced atherosclerosis had been developed in the ApoE^−/− mice at that age. However, ⁶⁴Cu-ATSM uptake in the both SCD and HFD fed ApoE^−/− mice increased significantly from 14 to 23 weeks (p < 0.05, Figure 7) with both SCD and HFD ApoE^−/− mice showing significantly higher A/M SUV ratios at 23 weeks (p < 0.01) in comparison to the wild-type control mice fed with standard chow. The difference between SCD and HFD ApoE^−/− mice progressively increased over time until 32 weeks old and began to decrease until end of the study (41 weeks old). Although HFD ApoE^−/− mice always had higher A/M SUV ratios than the SCD ApoE^−/− mice, this difference was only statistically significant at 32 weeks.

Figure 7.

Longitudinal study of ⁶⁴Cu-ATSM uptake in ApoE^−/− mice at the same age fed either with SCD or HFD. ⁶⁴Cu-ATSM uptakes were assessed by A/M SUV ratios as mean ± S.E.M of n = 6. *p < 0.01, **p < 0.001, and ***p < 0.001 are groups vs. 14 weeks groups. ^#p < 0.05, ^##p < 0.01 and ^####p < 0.0001 are SCD ApoE^−/− groups vs. SCD wild-type groups at the same age. ^{^^}p < 0.01 and ^{^^^^}p < 0.0001 are HFD ApoE^−/− groups vs. SCD wild-type groups at the same age. ⁺p < 0.05 is HFD ApoE^−/− group vs. SCD ApoE^−/− group at the same age. The numbers of ApoE^−/− HFD, ApoE^−/− SCD and wild-type SCD mice were all 6.

4. Discussion

The overall goal of the current study was to investigate the use of ⁶⁴Cu-ATSM PET imaging for determination of hypoxia in atherosclerotic plaque. This strategy has the potential to aid physicians in stratifying patients into high and low risk populations. We have made significant strides in demonstrating that the hypoxic cell PET imaging agent ⁶⁴Cu-ATSM can be used for the imaging of hypoxic plaque, and have collected preliminary data to support further study to fully validate the feasibility of imaging the extent of hypoxia in mouse models of atherosclerosis. We also conducted a longitudinal study investigating the role of diet and age in this model, and thus form the basis of our future preclinical and clinical studies.

⁶⁴Cu-ATSM is currently an approved Food and Drug Administration Investigational New Drug, thus this agent is already in use in a Phase II multicenter trial of oncologic patients at our institution [44]. Retrospective review of 8 sequential PET-CT examinations in cervical cancer patients imaged with ⁶⁴Cu-ATSM from their study revealed ⁶⁴Cu-ATSM uptake was only present in regions that corresponded to atherosclerosis on CT in the iliac arteries of one diabetic patient who died from myocardial infarction [45]. These anecdotal findings raised the suggestion the ⁶⁴Cu-ATSM is a more discriminating agent in identifying potentially unstable atherosclerotic lesions.

It is well established that macrophages play a key role in atherosclerosis [46]. In the initial stages of atherogenesis, modified lipoproteins recruit monocytes and T cells into macrophages and internalize the modified lipoproteins, resulting in the accumulation of highly oxygen-consuming lipid-loaded macrophages (foam cells) in developing lesions before apoptosis [47]. The present of plaque hypoxia in macrophage-rich areas with neovascularization has been demonstrated in rabbit advanced atherosclerosis [27]. Despite relatively low resolution for autoradioagraphy, the brightest area in the autoradiograph, which indicated the highest ⁶⁴Cu-ATSM uptake, regionally correlated to the pimonidazole and CD68 positive areas, the elevated ⁶⁴Cu-ATSM was accumulated in the hypoxic macrophage-rich core of the atherosclerotic plaques in the aortic arch of ApoE^−/− mouse.⁶⁴Cu-ATSM has time-dependent and cell-dependent spatial distribution and retention kinetics, which result in various correlation between the uptake of Cu-ATSM and immunofluorescent markers of hypoxia in different tumor cell lines [48–53]. There are reports of mildly (pO₂ < 2% O₂) [54] and severely hypoxic retention threshold (2.0–5.0 mmHg and even lower [55]; 0.5% and 0.1% O₂ [56]; versus 40–60 mmHg in normal tissue [57]) of ⁶⁴Cu-ATSM in different tumor types and acquisition time after injection. In a recent review article by Colombié et al. [58], the authors attribute the low correlation between Cu-ATSM uptake and hypoxic distribution in some tumors to the differing redox status of them. These tumor types might have a lower than average redox potential with high concentrations of electron donors leading to reduction and trapping of Cu-ATSM in both hypoxic and normoxic areas, i.e. human prostate tumor which overexpresses the fatty acid synthase (FAS) because this enzyme requires a large amount reductive species (NADPH) as cofactor for function [58, 59]. In atherosclerosis, although the redox potential caused by ATP-consumption in the lipid-loaded macrophages does not tend to be unusually low, we are still uncertain about the exact oxygen retention threshold of ⁶⁴Cu-ATSM in hypoxic plaque. It be either higher or lower than pimonidazole (pO₂ ≤ 10mm Hg), resulting in smaller or larger spatial distribution of ⁶⁴Cu-ATSM than pimonidazole.

Although ¹⁸F-FDG has been a well-described agent for in vivo imaging of atherosclerosis in animal [42, 60] and human subjects [5] and there is strong experimental and clinical evidence supporting that the increased ¹⁸F-FDG uptake in atherosclerotic plaque is associated with the abundance of macrophages, it is a relatively non-specific tracer with many practical limitations [61] because it provides complementary information on metabolic function as opposed to hypoxia. ¹⁸F-FDG vascular uptake can also present in vascular smooth muscle cells and endothelial cells [26, 62]. Moreover, because of the long time course of ¹⁸F-FDG in arteries, a longer ¹⁸F-FDG circulation time is always required for vascular ¹⁸F-FDG PET imaging protocol, preferably at least 90 minutes [63]. One other hand, the rapid blood pool clearance of ⁶⁴Cu-ATSM allowed PET imaging within 30 min of tracer injection, a significant advance over ¹⁸F-FDG. Our finding that the SUV and A/M SUV ratios between HFD and SCD ApoE^−/− with ⁶⁴Cu-ATSM imaging differ from a previous study using ¹⁸F-FDG as the PET imaging agent where ApoE^−/− mice receiving SCD had stable and low ¹⁸F-FDG SUVs while the ApoE^−/− mice receiving HFD showed progressively higher ¹⁸F-FDG uptake up to the end of their study at 32 weeks [42]. This difference may be secondary to the different retention mechanisms between ⁶⁴Cu-ATSM and ¹⁸F-FDG. It implies that the either the oxygen tension (<1% oxygen [64]) is too low to be distinguished by ⁶⁴Cu-ATSM imaging or alternatively, this finding may also suggest that the relative effect of the ApoE^−/− mouse model genetics is stronger than the effect of diet on hypoxia progression in atherosclerotic plaque, or even both reasons takes effects together.

Further studies could be performed using logistic regression analysis to compare immunoreactivity scores of pimonidazole-determined hypoxia in the plaques as well as other immunohistochemistry markers of hypoxia (HIF 1α, HIF 2α and VEGF) to the continuous variable of SUV, and to the A/M SUV ratios post injection of ⁶⁴Cu-ATSM. HIF-1α and HIF-2α are often used as markers of hypoxia. The expression of VEGF is low in normal vessel wall and up-regulated by hypoxia, inflammatory mediators, and certain growth factors. The human study from Slumer et al. in 2008 [23] demonstrated increasing expression of these hypoxia markers over the development of atherosclerosis. Nevertheless, another study using an atherosclerotic murine model found that all three markers of hypoxia correlated negatively with SUV_mean of ¹⁸F-FDG [42]. The discordant finding was explained by the limited number of lesions investigated in the human study, smaller lesions in mice compared to human, and potentially dissimilar gene expression of molecular markers of hypoxia between and mice and human, etc. In future studies, we will validate whether these hypoxia markers will correlate positively with the uptake of ⁶⁴Cu-ATSM to provide more data to explain this discrepancy.

One limitation in the study is the relatively poor spatial resolution of the small animal PET and coregistered CT scanner. The diameter of mouse aortic arch is about 1 mm [65], which is at the same level of the microPET resolution (1–2 mm). As a result the effect of partial volume and spillover may deteriorate the imaging quality and cause underestimation the resultant radioactivity. In further studies we will increase the size of animal model from mouse to rabbit, and apply post-processing techniques to correct for partial volume effects. The tiny size of the mouse aortic arch, which may easily generate shift and deformation during obtaining adjacent sections for autoradiography and immunohistochemistry, together with the dramatic different resolution of the autoradiography plate scanner (50 um) and the confocal microscope (< 5 um), brings up technical difficulty to correlate the autoradiograph and immunohistochemistry images in pixel level.

In order to justify the activity in the thigh muscle as the background activity for uptake normalization purposes, we collected the samples of aortic arch (with heart attached), blood pool and leg muscle from at least three SCD ApoE^−/− mice and control mice after sacrificing (30 minute post injection of ⁶⁴Cu-ATSM). For the ApoE^−/− mice, the %ID/gram of aortic arch with heart (4.02 ± 0.69) was 1.5 time higher than that of blood (2.72 ± 0.18), and 5.1 times higher than that of muscle (0.79 ± 0.11). However, for the control mice the %ID/gram of aortic arch with heart (2.34 ± 0.56) showed a similar value with blood (2.30 ± 0.60), and was 3.9 times higher than the muscle (0.61 ± 0.12). These findings suggest that activity of tracer in muscle is the same between the groups.

5. Conclusion

This study demonstrates the utility of ⁶⁴Cu-ATSM for imaging hypoxia in atherosclerotic plaque, permitting image contrast between hypoxic and normoxic tissues 7.5–30 minutes post injection. ⁶⁴Cu-ATSM imaging may assist in evaluating the progression of hypoxia over time in an ApoE^−/− mouse model.

Advances in Knowledge and Implications for Patient Care.

While studies in humans are necessary for conclusive data, in the long term, a ⁶⁴Cu-ATSM PET imaging strategy could help facilitate the study of plaque biology in human patients.