Use of Small Mammals for Definition of Biodistribution

The first step in the evaluation of any radiopharmaceutical involves determination of the normal biodistribution of a radiotracer over time. These initial screening studies are usually performed in mice, rats, or rabbits to minimize cost. A measured dose is injected intravenously, and the percentage of uptake per gram of tissue is determined for critical organs.^4–7 Uptake in the heart, the target organ, is compared with adjacent background structures at multiple time points after injection. Blood clearance can be easily derived by serial sampling of blood for gamma well counting. Tissue clearance curves can be derived either by sacrificing animals at different time points or by performing serial imaging. Segregation of tissue into cell fractions can provide information on tracer localization in specific cellular fractions.⁸ This information may offer insight into the mechanism of uptake and retention. The smaller hearts in these animals lend themselves to microautoradiography.

Use of Small Mammals for Evaluation of Specific Disease Processes (See Chapter 11)

A limited number of natural biological models are currently available in smaller mammals for evaluation of specific cardiovascular disease processes. The Syrian hamster model is an example of a small model for evaluation of radiotracer kinetics under cardiomyopathic conditions. Other researchers have developed rabbit models of congestive heart failure.^9,10 Investigators have also created a chronic volume-overloaded condition by mechanically disrupting the aortic valve, instigating aortic insufficiency.¹¹ It is more difficult to create a small-animal model of regional ischemia or infarction. However, this has been accomplished by placing a ligature or small clamp around a proximal coronary artery. In the mouse and rat, this occlusion can be done blindly with a small needle and suture through a thoracotomy.^12,13 Unfortunately, this approach generally necessitates disruption of both venous drainage and arterial inflow. Regarding rat models of infarction, the use of Lewis rats may be preferable over other strains because of the improved survival and greater uniformity of myocardial injury.¹⁴ Several investigators have used a rabbit model of regional myocardial ischemia.^15,16 In the rabbit, it is possible to selectively isolate and occlude a proximal coronary artery. Fujita et al. recently established a chronic rabbit model of myocardial infarction without the need for endotracheal intubation, which significantly improves survival.¹⁶ Production of graded ischemia in smaller animals is nearly impossible. The use of radiolabeled microspheres for independent assessment of regional myocardial blood flow is more difficult in these preparations. In addition, these preparations do not permit regional arterial-venous balance measurement for the independent assessment of regional metabolism. However, with the advent of dedicated small-animal imaging systems, quantitative evaluation of radiotracers in these physiologic small-animal models will become more feasible.^17,18

Use of Genetically Engineered Small Mammals for Evaluation of Specific Biological Processes and Radiotracers (See Chapter 45)

The ability to selectively alter gene expression permits the evaluation of the significance of certain gene products for structure-function studies of cardiac proteins and their role in cardiovascular disease. These transgenic models facilitate the evaluation of targeted radiotracers in relevant biological models. Hundreds of mutant mouse strains and also a few mutant rat and rabbit strains have been generated. The number of genetically engineered mouse lines for cardiovascular research is continuously growing. Recent reviews have provided a summary of important new genetically modified animals with cardiovascular disease.^19–27 Transgenic mice are most frequently used for this purpose, since mice breed rapidly, the maintenance costs are lower, and our knowledge of mouse genetics is far more advanced. Although the small size of these animals leads to significant challenges in the evaluation of radiotracers, there have been significant advances in small-animal SPECT and PET instrumentation that make imaging of mice feasible.^17,18

Approaches for creation of transgenic animals are nicely reviewed by Williams and Wagner in their application for evaluation of integrative biology.²⁸ However, the use of transgenic animals will be equally useful for the development and testing of biologically targeted radiotracers. Once targeted radiotracers, are established, they could also be used to evaluate new transgenic strains in conjunction with radiotracers which provide information about physiologic processes (i.e., perfusion, metabolism). Thus, the use of radiotracer imaging would also permit phenotyping of transgenic animals.

There are two basic approaches to mouse genomic manipulation: random chromosomal integration, which can be used for addition of an exogenous transgene, and homologous recombination of foreign DNA, which leads to targeted mutation of an endogenous gene.²⁸

Random chromosomal integration is based on addition of DNA into fertilized oocytes in order to generate “gain-of-function” mutations, in which the transgene is overexpressed.²⁸ Another common application of random chromosomal integration of transgenes is to identify transcriptional control elements that respond to physiologic stimuli. In this application, the coding region of a so-called reporter gene is linked to the regulatory elements of interest. By definition, reporter genes encode biologically innocuous proteins that are easily detected. The goal of this approach is to avoid perturbation of the cell in which the reporter is expressed while assessing the function of transcriptional regulatory sequences that are attached to the reporter gene. For application with radiotracer imaging, these reporter genes are linked with a radiolabeled “reporter probe” that can be detected in vivo in mice with microPET or microSPECT imaging systems.

Gambhir et al. reported methods for monitoring reporter gene expression by using herpes simplex virus type I mutant thymidine kinase (HSV1-sr39tk) as a PET reporter gene and 9-(4-[¹⁸F]-fluoro-3 hydroxymethylbutyl) guanine ([¹⁸F]-FHBG) as a PET reporter probe in animal models.²⁹ Wu et al. applied this technology using microPET imaging to noninvasively monitor cardiac reporter gene expression in rats.³⁰ More recently, Wu et al. applied this approach for the noninvasive assessment of myocardial response to cell therapy using embryonic cardiomyoblasts expressing HSV1-sr39tk and/or firefly luciferase.³¹ The location, magnitude, and survival duration of the transplanted cells were monitored noninvasively using PET and bioluminescence optical imaging.

Gene targeting via homologous recombination in embryonic stem cells is frequently used to create “loss-of-function” mutations, known as knockouts. Targeted inactivation can be accomplished by introducing a positive selection marker, which will disrupt gene structure. A conditional gene targeting approach is preferred, since this approach provides cell-type-specific and/or inducible gene targeting. One conditional gene targeting approach involves Cre/loxP-mediated targeted mutagenesis.³² This approach is based on the ability of Cre recombinase to recognize a unique nucleotide sequence (loxP site), allow the introduction of mutations in the gene of interest, and by the controlled expression of Cre also control the expression during different time points and avoid germline mutations that may be lethal. Alternatively, temporal or cell-type-specific transcriptional control can be accomplished using a tetracycline-responsive promoter.³³

We are at the beginning of a new era for radiopharmaceutical testing with the development of transgenic models for many cardiovascular disease processes. However, the full impact of testing radiopharmaceuticals in transgenic models still remains undefined.

Use of Large Mammals for Evaluation of Specific Disease Processes

Dogs, pigs, and sheep are the large mammals most commonly used for the evaluation of cardiovascular physiology. These models have been used for many years, and the methodologies of their use are well established. The higher cost of large-animal models almost precludes their use for simple biodistribution studies; rather, they are mostly used to evaluate radiotracer kinetics in specific disease processes. In the evaluation of radiotracers, dogs and pigs have been most commonly used. Important interspecies differences in radiotracer uptake have been observed over the years. The most notable example is the uptake of technetium-99m (^99mTc) 1,2-bis(dimethylphosphino)ethane. This ^99mTc-labeled perfusion tracer showed favorable uptake characteristics in rats, dogs, rabbits, and monkeys; however, in phase I trials, no appreciable uptake was noted in the hearts of pigs or humans. Although there was transient uptake of this radiotracer by human heart, trapping was not observed because of a reductive mechanism that is prominent in human myocardium.³⁴

Canine models have been used for many years in fundamental cardiac physiologic studies. All of the important physiologic conditions (ischemia, infarction, stunning, and even hibernation) have been modeled in dogs, and all of the currently approved radiotracers have been evaluated in canine models. Thus, the characteristics of new tracers can be easily compared with existing data on older tracers. However, it is becoming prohibitively expensive to use dogs in these types of studies.

The dog has a dominant left coronary system. In almost all cases, the left circumflex coronary leads to the posterior descending coronary artery; very rarely does the right coronary artery provide any of the blood supply to the left ventricle. The coronary vascular tree in the dog is also highly collateralized. Therefore, some investigators claim that the dog is the ideal model of chronic coronary artery disease in humans, because patients with longstanding critical coronary disease frequently are highly collateralized. In addition, the dog is very tolerant of ischemic injury. In contrast, pigs and sheep have limited native collaterals and tolerate ischemic insults less reliably. When conducting radioisotope experiments in the dog, it may sometimes be necessary to tie off superficial epicardial collaterals to minimize the nearly instantaneous opening of epicardial collateral circuits. Dogs can also be trained to walk on a treadmill, providing an effective model for evaluation of pharmacologic stress agents.³⁵

The pig has a coronary circulation that is more similar than that of the dog to human coronary circulation in that both the right and left coronary artery supply the left ventricle with blood. Pig coronary circulation systems have less native coronary collaterals, a characteristic akin to normal coronaries in humans. Therefore, the pig may be a better model of acute coronary occlusion in humans, because under conditions of acute occlusion, coronary collaterals may not have yet developed. In recent years, porcine models of ischemia, infarction, stunning, and hibernation have been established. However, sheep may provide the best model for studying chronic infarction and left ventricular remodeling.³⁶

In all of these large-animal models, acute coronary occlusion may be created surgically³⁷ or with a balloon angioplasty catheter under fluoroscopic guidance.^38,39 In acute and chronic models, partial occlusions can be created with the aid of a hydraulic occluder or partially occluding ligature. A better way to achieve chronic partial or total occlusion may be placement of an ameroid occluder.^40,41 An ameroid occluder is a small ring that is surgically placed around a vessel. This device slowly swells over 3 to 4 weeks. A complete or partial coronary occlusion is thus created very gradually, allowing for the development of coronary collaterals.

Large-animal models offer important advantages over studies in smaller animals. First, studies in larger animals allow for more complete instrumentation and regional interventions.^42–47 Second, the distribution of a radiotracer can be easily evaluated with standard radionuclide imaging equipment.^48,49 Finally, independent regional measures of flow, function, and metabolism are much easier to obtain in larger animals.⁵⁰

Many physiologic issues can be addressed by using acute animal models. However, to evaluate changes over an extended period of time, chronic models must be used. Chronic models also offer the ability to evaluate radiotracer behavior under conscious conditions. Animals are usually surgically instrumented and allowed to recover for at least 1 week after surgery. After full surgical recovery, an ischemic insult is implemented or stress is applied while the animal is conscious.^35,51

Dynamic Imaging

Pharmacodynamics can be evaluated noninvasively with serial planar imaging, SPECT, and dynamic PET immediately after injection of a radiotracer. Dynamic imaging sequences allow determination of blood clearance and clearance of a radiotracer from the heart and other critical noncardiac structures. All of these imaging approaches can be easily applied to both open-chest and long-term models.

Planar Imaging

Planar imaging offers some advantages for evaluation of radiotracer pharmacodynamics, including high temporal and spatial resolution. Planar imaging of the heart is complicated by superimposition of activity in adjacent extracardiac structures. This can be obviated in open-chest models by isolation of the heart from other structures by using a flexible lead sheet. It is critical that the shielding material is placed directly under the heart, thereby completely isolating the beating heart from extracardiac structures. Time-activity curves can be derived for determination of blood and regional myocardial pharmacodynamics. In this type of kinetic analysis, placement of the regions of interest is critical. We have found that moving the myocardial region of interest one pixel toward or away from the blood pool can greatly alter the myocardial clearance curves. To facilitate optimal placement of the myocardial region of interest, we generate a reference image by subtracting an early blood-pool image from a later image (Fig. 3-1). An initial left ventricular blood-pool image can be derived by reconstruction of the left ventricular phase of the first-pass list mode acquisition. A myocardial perfusion image can be derived by summing the last 1 minute of the dynamic image sequence. The initial blood-pool image is then subtracted from the final perfusion image, providing an index myocardial image that is minimally contaminated by blood pool. The myocardial clearance curves generated by using this planar imaging approach can provide useful information on myocardial uptake and clearance of radiotracer in carefully controlled experimental models of regional ischemia or infarction. This model allows for selective arterial and venous sampling for simultaneous determination of regional metabolism or radiotracer extraction. Figure 3-2 shows a dynamic image series and corresponding clearance curves that can be derived from it.

Figure 3-1 Generation of myocardial uptake and clearance curve from serial planar images. An early blood-pool image (left) can be subtracted from the last perfusion image (middle) to define the myocardial regions within the image that were relatively devoid of blood-pool contamination. A typical subtraction image is shown (right). Regional myocardial uptake and clearance can be derived from regions of interest placed on the subtraction image.

Figure 3-2 A, Dynamic lateral technetium-99m (^99mTc)-tetrofosmin image sequence. One-minute images were acquired during the first 10 minutes after injection. The initial image (upper left) reflects right ventricular (superior) and left ventricular (inferior) blood-pool activity. The apex of the heart points toward the left lower corner of each image. The liver is seen below the heart. The rib spreader produces an attenuation defect in the superior aspect of the liver. A large, dense anteroapical defect is seen (arrow). B, The tissue clearance curves for ^99mTc-tetrofosmin (in four dogs) are shown. Activity of ^99mTc-tetrofosmin is expressed as a percentage of the nonischemic myocardial activity at the time of sacrifice. The clearance of ^99mTc-tetrofosmin in a nonischemic region is shown relative to hepatic and lung clearance.

(From Sinusas AJ, Shi QX, Saltzberg MT, et al: Technetium-99m-tetrofosmin to assess myocardial blood flow: Experimental validation in an intact canine model of ischemia, J Nucl Med 35:664–671, 1994. Reprinted with permission of the Society of Nuclear Medicine.)

In 1979, it was recognized that distortion of global and regional left ventricular geometry could cause artifact defects in perfusion images obtained with thallium-201 (²⁰¹Tl) or intracoronary administration of ^99mTc-labeled albumin microspheres.⁵³ Sinusas et al. subsequently demonstrated that changes in regional function may confound analysis of planar ^99mTc-sestamibi images.⁴⁴ Thus, changes in regional myocardial thickening may lead to misinterpretation of myocardial tracer uptake or clearance because of partial volume effects.⁴⁴ Partial volume effects refers to the underestimation of count density from a structure that is thinner than twice the resolution of the imaging system (usually 12 to 20 mm for a gamma camera).⁵⁴ In models of regional ischemia, systolic thinning may occur in ischemic or postischemic stunned regions. Differences in wall thickness would be most prominent at end systole. The potential impact of these partial volume errors are illustrated in Figure 3-3. We found that analysis of end-diastolic images may compensate somewhat for the partial volume effect associated with regional dyskinesis.

Figure 3-3 Images showing that regional left ventricular dysfunction creates an artifactual perfusion defect. Technetium-99m labeled sestamibi was injected during baseline conditions before coronary occlusion and reperfusion. Shown are lateral planar images obtained at baseline (left) and during coronary artery occlusion (middle). A large anteroapical defect (arrow) is seen both on summed (SUM) and end-systolic (ES) images during coronary artery occlusion. This artifact is less evident on the end-diastolic (ED) image. The corresponding baseline and occlusion quantitative profiles are superimposed on the right. White dots represent the circumferential profile derived from each baseline image. Black-and-white dots represent the corresponding occlusion profile. Changes in defect scores are also shown.

(From Sinusas AJ, Shi QX, Vitols PJ, et al: Impact of regional ventricular function, geometry, and dobutamine stress on quantitative ^99mTc-sestamibi defect size, Circulation 88:2224–2234, 1993. Reprinted with permission.)

Dynamic planar imaging has been used by several investigators to establish the clearance kinetics of ²⁰¹Tl,^{55 99m}Tc-sestamibi,^{56 99m}Tc-tetrofosmin,⁴⁵ and ^99mTc-teboroxime.⁵⁷ Serial planar imaging permits nontraumatic serial assessment of changes in regional activity. However, dynamic planar imaging is limited by camera resolution, partial volume effects, and potential motion artifacts and may be influenced by background scatter and tissue cross-talk.

Dynamic planar images can also be obtained in animals as small as a mouse using high-resolution pinhole collimators.⁵⁸ Figure 3-4 illustrates a series of planar pinhole images in a mouse injected with a ^99mTc-labeled compound targeted at the α_vβ₃ integrin, and corresponding tissue clearance curves. Using an external radioactive reference source (~10 μCi), the regional counts within selected organs can be converted to μCi and expressed as a percentage of injected dose.

Figure 3-4 Dynamic pinhole images in a mouse injected with ¹¹¹In-labeled RP748, a radiotracer targeted at the α_vβ₃ integrin. A, Dynamic gamma camera images of RP748 clearance in ApoE^−/− mouse at various time points after tracer administration, acquired with large field of view gamma camera and 1-mm pinhole aperture. Circles represent different regions of interest. B, Time-activity curve derived from dynamic image series, demonstrating rapid renal clearance of the tracer.

(From Sadeghi MM, Krassilnikova S, Zhang J, et al: Detection of injury-induced vascular remodeling by targeting activated α_vβ₃ integrin in vivo, Circulation 110:84-90, 2004. Reprinted with permission.)

Single-Photon Emission Computed Tomography

High-quality tomographic images can be acquired to evaluate the biodistribution of a radiotracer. Imaging by SPECT provides better separation of organs than planar imaging. Dynamic SPECT imaging has only recently become feasible with the development of multidetector cameras. Several investigators have used dynamic SPECT imaging for kinetic modeling.^59–61 Stewart et al. performed dynamic tomographic imaging with ^99mTc-teboroxime in closed-chest dogs by using a single-photon ring tomographic system.⁶¹ These investigators demonstrated monoexponential clearance of ^99mTc-teboroxime with this approach. By using a triple-detector SPECT system, Smith et al. demonstrated the feasibility of acquiring dynamic 10-second tomographic data with a continuous acquisition protocol.⁶⁰ Analysis of wash-in of ^99mTc-teboroxime in myocardial tissue measured with this dynamic SPECT approach, in conjunction with compartmental modeling, provided an index of regional myocardial blood flow. Smith et al. also demonstrated that the use of attenuation correction in conjunction with dynamic SPECT was critical for compartmental analysis of ^99mTc-teboroxime. Figure 3-5 provides an example of a series of high-quality images and corresponding clearance curves derived from dynamic SPECT in a dog. The potential advantage of kinetic modeling with SPECT over PET is the greater availability of SPECT equipment and reduced cost. However, the accuracy of the dynamic SPECT approach is not established.

Figure 3-5 Dynamic SPECT technetium-99m (^99mTc)-teboroxime images were acquired (20 seconds per image) over 30 minutes in an open-chest dog after creation of a partial occlusion of the left anterior descending region. A, Short-axis ^99mTc-teboroxime images are shown from mid-ventricle over time. A perfusion defect is seen in the anteroapical region. Blood-pool and myocardial clearance curves were generated from this dynamic image set. B, Maximum counts from ischemic, nonischemic, and blood-pool regions are plotted over time. Differences in regional ^99mTc-teboroxime clearance are evident.

Pinhole SPECT systems have been developed to evaluate regional properties of radiopharmaceuticals in small animals.^62,63 Unfortunately, the ten-fold gain in resolution of these systems is at the expense of a 100-fold loss in sensitivity.⁶⁴ This limitation has been overcome by using multidetector SPECT systems.⁶² Optimal reconstruction of the pinhole images will probably require more computer-intensive, three-dimensional, maximum-likelihood expectation-maximization (ML EM) algorithms.⁶⁴ Although these systems cannot compete with the high resolution of microautoradiography, they may provide an alternative to macroautoradiography or well counting. Hirai et al. demonstrated the feasibility of serial SPECT with pinhole collimators for estimation of flow and metabolism in vivo in rat hearts.⁶⁵ Wu et al. performed myocardial perfusion imaging of mice using ^99mTc-sestamibi and demonstrated the feasibility for measurement of perfusion defect size from pinhole SPECT images (Fig. 3-6).³¹ Other investigators recently demonstrated the feasibility of electrocardiographic (ECG)-gated pinhole SPECT imaging in mice⁶⁶ and rats⁶⁷ for the evaluation of regional left ventricular function by applying standard clinical image analysis software to reconstructed ECG-gated microSPECT images.

Figure 3-6 Pinhole ^99mTc-sestamibi SPECT images of a mouse heart. Short-axis (first two rows), vertical long-axis (third row), and horizontal long-axis (fourth row) slices are shown from a myocardial perfusion study in a normal mouse. Image quality is excellent, with the left ventricle well resolved and the right ventricle clearly visible.

(From Wu MC, Gao DW, Sievers RE, et al: Pinhole single-photon emission computed tomography for myocardial perfusion imaging of mice, J Am Coll Cardiol 42:576–582, 2003. Reprinted with permission of the American College of Cardiology.)

High-speed and high-sensitivity SPECT cameras have recently become available that employ cadmium zinc telluride (CZT) detectors and novel schemes for reconstruction and collimation.^68,69 The novel design of these new systems significantly increases sensitivity and reduces imaging time while providing higher spatial resolution than the conventional Anger camera approach.^70,71 These systems will permit dynamic SPECT imaging, facilitating evaluation of the biodistribution and uptake and the clearance kinetics of new radiotracers under evaluation. They also offer the potential for absolute quantification of radiotracers. The increased sensitivity is at the cost of the size of the field of view, but for experimental imaging, these systems may be ideally suited for dynamic imaging of rabbits or focused imaging of larger animals. These systems should expand the horizons of SPECT imaging in experimental animal models.

Positron Emission Tomography (See Chapter 11)

Dynamic PET imaging in large-animal models plays an important role in evaluation of the kinetics of positron-emitting radiopharmaceuticals. However, this chapter will not review this topic. Technetium-94m (^94mTc), a positron emitter with a 53-minute half-life, can be used to evaluate the biodistribution and pharmacokinetics of ^99mTc radiopharmaceuticals for SPECT.^72,73 Performing PET in animal models with Tc-labeled compounds allows in vivo evaluation of regional myocardial uptake and clearance kinetics. Figure 3-7 provides an example of ^94mTc-sestamibi PET images. Imaging with PET offers great advantages in the evaluation of SPECT tracers: It provides high temporal and spatial resolution and reliable attenuation correction. Labeling of SPECT tracers with ^94mTc allows the direct comparison of SPECT radiotracers with PET compounds.

Figure 3-7 Technetium-94m (^94mTc)-sestamibi and nitrogen-13 (¹³N)-ammonia PET images in a 66-year-old woman with previous posterolateral wall myocardial infarction. Shown are transaxial ^94mTc-sestamibi images (top row) and ¹³N-ammonia images (bottom row). Greater hepatic uptake is seen on the ^94mTc-sestamibi images.

(From Stone CK, Christian BT, Nickles RJ, Perlman SB: Technetium 94m-labeled methoxyisobutyl isonitrile: Dosimetry and resting cardiac imaging with positron emission tomography, J Nucl Cardiol 1:425–433, 1994. Reprinted with permission of the American Society of Nuclear Cardiology.)

Recently, several microPET imaging systems have become commercially available and are capable of performing dynamic PET imaging in mice and rats.⁷⁴ However, accurate kinetic modeling with microPET requires a reliable arterial input function, which is difficult to obtain in mice. Some investigators have attempted to derive an input function using high-sensitivity detectors placed over an extracorporeal circuit, while others are designing precision devices for serial blood sampling.

Serial microPET imaging was recently applied in rats following myocardial infarction and transient ischemia, as well as in control rats, to (1) evaluate the biodistribution and uptake and the clearance kinetics of a new ¹⁸F-labeled pyridazinone analog (18F-BMS-747158-02) under a range of physiologic conditions, and (2) define the changes in the uptake ratio of heart to adjacent organs (Fig. 3-8A-B). Serial microPET imaging was also used to define the regional uptake and clearance of normal myocardium and myocardium following permanent (see Fig. 3-8C) and transient (see Fig. 3-8D) coronary occlusion.

Figure 3-8 Dynamic PET imaging. A, Examples of ¹⁸F-BMS-747158-02 (¹⁸F-BMS) PET images of chest of healthy rat at 15, 45, and 115 minutes after tracer injection and ¹³N-ammonia PET image at 10 minutes in coronal view. Example of regions of interest for measurement of tracer activity is displayed in white box. B, Ratio of ¹⁸F-BMS uptake between myocardium and surrounding organs. C-D, Mean values are shown for tracer activity in induced defect and normally perfused myocardium after permanent (C) and transient (D) occlusion. Tracer activity in defect increased over time after reperfusion in transient coronary occlusion.

(Modified from Higuchi T, Nekolla SG, Huisman MM, et al: A new ¹⁸F-labeled myocardial PET tracer: Myocardial uptake after permanent and transient coronary occlusion in rats, J Nucl Med 49:1715–1722, 2008. Reprinted with permission.)

Miniature Detectors

Myocardial radiotracer kinetics can also be derived in large open-chest animal preparations with the use of miniature radiation detectors. In many studies by Okada et al., miniature cadmium telluride radiation detectors were used to evaluate myocardial uptake and clearance of several radiopharmaceuticals, including ²⁰¹Tl- and ^99mTc-sestamibi.^56,75,76 These probes allow continuous monitoring and display of regional myocardial activity.⁷⁷ High-quality myocardial clearance curves derived by using a cadmium telluride detector are shown in Figure 3-9. The use of these probes can be complicated by changes in myocardial thickness and underlying blood pool beneath the probe and in the stability of the instrumentation. In some studies, probes have been placed on the myocardium in combination with a device to measure wall thickening. This system allows online correction of changes in counts associated with cyclic wall thickening. More recently, Stewart et al. measured myocardial clearance of ^99mTc-teboroxime with a 1-inch collimated sodium-iodine (Tl) probe.⁵⁷ The first component of ^99mTc-teboroxime washout derived by using this probe was related to microsphere blood flow. This miniature probe approach allows continuous monitoring of regional myocardial activity under different ischemic conditions or during pharmacologic stress. Despite the potential limitations of these miniature probes, much insight into the in vivo behavior of ²⁰¹Tl-labeled and newer ^99mTc-labeled compounds has been derived from probe studies.

Figure 3-9 Typical decay-corrected myocardial time-activity clearance curves derived from miniature cadmium telluride probes for an ischemic left circumflex (LCX) and control left anterior descending (LAD) region in a dog with a critical coronary stenosis. Both probes were positioned on the epicardial surface of the left ventricle. In this example, there was minimal and equal washout from the two regions.

(From Okada RD, Glover D, Gaffney T, et al: Myocardial kinetics of technetium-99-m hexakis-2-methoxy-2-methylpropyl-isonitrile, Circulation 77:491–498, 1988. Reprinted with permission.)

Serial Myocardial Biopsy

Investigators have obtained serial myocardial biopsy specimens after injection of a radiotracer to evaluate changes in absolute myocardial activity over time.^78–80 This approach is the most accurate method to serially determine changes in absolute myocardial counts per gram. However, serial myocardial biopsy results in myocardial injury, necessitates the analysis of very small pieces of tissue, and has a potential sampling error.

Postmortem Imaging

Imaging of excised organs or tissues can be very useful to establish the final distribution of a radiotracer that has been injected in vivo under a specific physiologic condition. Some investigators have acquired high-resolution, tomography-like images by placing slices of the heart on the surface of a gamma camera (Fig. 3-10).⁴³ Images acquired by using this method are similar to those obtained using macroautoradiography. The gamma camera approach enables the simultaneous acquisition of high-resolution images by using multiple energy windows. Investigators have used this method to compare the distribution of perfusion and metabolic tracers labeled with different radioisotopes. Effective application of this method necessitates sectioning the heart in slices of uniform thickness, which requires an automated slicing device. Alternatively, a complete three-dimensional set of short-axis images of the intact heart can be acquired ex vivo with a SPECT camera (Fig. 3-11). This eliminates the need for precision slicing of the heart and avoids the problems of attenuation and motion. To simulate the in vivo geometry of the heart, hearts are generally stuffed with gauze, frozen in a distended state, or filled with inert dental molding material. All methods of postmortem imaging eliminate partial volume errors associated with cardiac deformation. Images derived using this method can be easily quantified and readily compared with postmortem histochemical stains.

Figure 3-10 Comparability of technetium-99m-sestamibi gamma camera ex vivo slice images (left) and macroautoradiographs (right). Shown are images from two dogs that underwent 3 hours of left anterior descending coronary artery occlusion followed by 3 hours of reperfusion. ^99mTc-sestamibi was injected during coronary occlusion (top) in one dog and after 90 minutes of reperfusion in the second dog (bottom).

(From Sinusas AJ, Trautman KA, Bergin JD, et al: Quantification of area at risk during coronary occlusion and degree of myocardial salvage after reper fusion with technetium-99m-methoxyisobutyl-isonitrile, Circulation 82:1424–1437, 1990. Reprinted with permission.)

Figure 3-11 Postmortem ex vivo technetium-99m-sestamibi SPECT images oriented in standard short-axis format. ^99mTc-sestamibi was injected during inotropic stress in the presence of a partial stenosis of the left anterior descending coronary artery. Short-axis images (5 mm thick) are shown from apex to base. The right ventricle (RV) is seen on the left and the left ventricle (LV) on the right of each image. A mild perfusion defect is seen in the anteroapical region (arrows).

Autoradiography

Microautoradiology and macroautoradiography have been critical in the development and evaluation of radiopharmaceuticals.^11,43,81,82 One of the major limitations of this approach is that each animal provides data for only a single experimental time point, possibly two if dual-isotope autoradiographs are acquired. However, these approaches also offer several advantages.

Microautoradiography

Microautoradiography permits high-resolution assessment of the distribution of radioactivity within an organ. This approach allows several isotopes to be compared directly if they can be separated on the basis of differences in half-life. The appropriate use of this technique requires preparation of tissue standards for each isotope in exponentially increasing activity concentrations.^82,83 Tissue standards are necessary to convert autoradiographic intensity into tracer activity. Digitized autoradiographs can be quantitatively analyzed if the appropriate standards are used; the details of this approach have been reported elsewhere.⁸¹ Figure 3-12 provides an example of digitized dual-isotope microautoradiographs and quantitative profiles comparing the myocardial distribution of ²⁰¹Tl- and ^99mTc- teboroxime in a rabbit model of acute coronary occlusion. Figure 3-13 provides another example of dual-isotope microautoradiography using ^99mTc-sestamibi and an ¹¹¹In-labeled agent (¹¹¹In-RP748) targeted at the α_vβ₃ integrin in a rat following occlusion and reperfusion resulting in nontransmural infarction. Uptake of ¹¹¹In-RP748 demonstrates increased expression and activation of the α_vβ₃ integrin in angiogenic vessels within the infarct region as defined by ^99mTc-sestamibi.

Figure 3-12 A, Dual-isotope microautoradiographs showing technetium-99m (^99mTc)-teboroxime (left) and thallium-201 (²⁰¹Tl) (right). Myocardial activity distributions in the same rabbit heart during coronary artery occlusion. Microautoradiographs were obtained from 30-μm thick short-axis slices. Note the increased contrast and clearer delineation of normal and low-flow zones produced by ^99mTc-teboroxime compared with ²⁰¹Tl-simultaneously injected. B, A representative pair of dual-isotope ^99mTc-teboroxime and ²⁰¹Tl profiles across the defect. Normal-to-defect contrast was increased and normal-to-defect transition was sharper for ^99mTc-teboroxime compared with ²⁰¹Tl.

(From Weinstein H, Reinhardt C, Leppo J. Teboroxime, sestamibi and thallium-201 as markers of myocardial hypoperfusion: comparison by quantitative dual isotope autoradiography in rabbits. J Nucl Med. 1993;34:1510–1517.)

Figure 3-13 A, Dual-isotope ^99mTc-sestamibi (early) and (¹¹¹In-RP748) (late) microautoradiographs (autorad) from a short-axis section (10-μ thick) from rat heart following 45 minutes of occlusion and 3 hours of reperfusion. The early ^99mTc-sestamibi microautoradiography (left) demonstrates a perfusion defect, and the late ¹¹¹In-RP748 microautoradiograph (middle) demonstrates focal retention of ¹¹¹In-RP748, which is targeted at the α_vβ₃ integrin (a marker of angiogenesis) within the infarct region. A color-fused image is shown (right) demonstrating clear localization of ¹¹¹In-RP748 with the perfusion defect.

Digital microautoradiography can also be performed with ¹⁸F-labeled PET radiotracers using 20-μm thick frozen sections placed in a specimen imaging system capable of detecting either beta or gamma radiation.^7,84,85 Higuchi et al. used a specimen imaging system to obtain digital microautoradiographs in rat hearts to determine the time course of an [¹⁸F]galacto-RGD compound for targeted imaging of the α_vβ₃ integrin following myocardial infarction (Fig. 3-14).⁷

Figure 3-14 A, Representative autoradiographs of [¹⁸F]-galacto-RGD uptake in the heart after ischemia and reperfusion with three slices (apex, middle, and base) studied at different time points. B, The bar graph depicts the changes in the myocardial [¹⁸F]galacto-RGD uptake ratio (versus remote myocardium) after coronary occlusion and reperfusion as defined by autoradiography.

(From Higuchi T, Bengel FM, Seidl S, et al: Assessment of α_vβ₃ integrin expression after myocardial infarction by positron emission tomography, Cardiovasc Res 78:395–403, 2008. Reprinted with permission.)

The resolution of autoradiographs depends on the energy level of the tracers and the thickness of the sections.⁸⁶ Hearts are typically cut into sections that are 10 to 50 μm thick. For quantitative analysis, a high-precision cryotome is required, and the thickness of the specimen must be uniform. Spatial resolution is improved if the specimen is firmly pressed onto the film. This technique is relatively economical because a gamma well counter is not required.

Macroautoradiography

Lower-resolution macroautoradiographs can be more easily obtained in larger animal hearts.⁴³ This approach does not even require a cryotome. Thin slices (2 to 5 mm) can be obtained with an inexpensive meat slicer. These myocardial slices are laid flat on cardboard, covered in plastic wrap, and inverted on x-ray film. Macroautoradiographs obtained in this manner can provide information on the spatial distribution of a radiotracer under experimental conditions of low-flow ischemia, myocardial stunning, or infarction. As shown in Figure 3-15, macroautoradiographs can be directly compared with postmortem stains to relate radiotracer uptake to myocardial viability.

Figure 3-15 Postmortem dual-perfusion maps (left) and technetium-99m (^99mTc)-sestamibi macroautoradiographs (right) from a dog undergoing 3 hours of left anterior descending coronary artery occlusion followed by 3 hours of reperfusion. ^99mTc-sestamibi was injected 90 minutes into the reperfusion period. Shown are three slices oriented with the right ventricle on the left and the anterior wall on the bottom. The risk area is stained brick-red with triphenyltetrazolium chloride (TTC). The infarct area is the pale, unstained region within the risk area. Defects seen on the unenhanced macroautoradiographs with superimposed overlay (right) correlate closely with the area of infarction defined by TTC staining. No significant ^99mTc-sestamibi activity is seen in the central necrotic region, whereas some activity is seen in the perinecrotic area. This approach allows for direct comparison of macroautoradiographs with postmortem stains to evaluate radiotracer uptake to myocardial viability.

(From Sinusas AJ, Trautman KA, Bergin JD, et al: Quantification of area at risk during coronary occlusion and degree of myocardial salvage after reperfusion with technetium-99m-methoxyisobutyl-isonitrile, Circulation 82:1424–1437, 1990. Reprinted with permission.)

Postmortem Tissue Well Counting

Postmortem tissue well counting is the most widely applied approach for assessment of myocardial uptake and clearance of a radiotracer. This technique is applied in larger-animal models in which specific physiologic states can be modeled. In general, the radiotracer under evaluation is injected during a specific physiologic condition; the activity retained in the myocardium is determined by doing gamma well counting of the excised tissue. If hearts are excised within minutes of injection, the initial myocardial distribution of a radiotracer can be determined. Hearts excised hours after injection provide information about the relative late retention of a radiotracer.

The initial uptake of a tracer is dependent on flow or delivery of the radiotracer to the tissue and the extraction of the tracer by the tissue. Retention of a tracer may be influenced by perturbations in regional metabolism, cellular viability, or flow. Frequently the distribution of a radiotracer under evaluation is compared with independent measures of regional myocardial blood flow determined using the radiolabeled microsphere technique.

Tissue is counted in a gamma well counter, which has excellent energy discrimination and permits simultaneous evaluation of as many as six radiotracers if each one has gamma emissions at distinguishable energy ranges. Activity spillover from the energy window of one radiotracer to another can be corrected with a matrix derived from counting pure specimens of each radiotracer. This approach has been described in detail by Heymann et al.⁸⁷ and has been applied by many other investigators.⁸⁸ Identification of ²⁰¹Tl redistribution was accomplished using this approach (Fig. 3-16).⁸⁹

Figure 3-16 Relationship of myocardial thallium-201 (²⁰¹Tl) activity and radiolabeled microsphere regional myocardial blood flow as determined by gamma well counting. ²⁰¹Tl- was injected in the presence of a fixed coronary stenosis. The myocardial distribution of ²⁰¹Tl was compared with the strontium-85-microsphere–determined blood flow at the time of radiotracer injection in animals sacrificed at 10 minutes (upper left), 2 hours (upper right), and 4 hours (lower left) after radiotracer injection. The left ventricle was cut into 1-gm to 3-gm pieces for gamma well counting. Regional myocardial ²⁰¹Tl activity and flow were expressed as a percentage of nonischemic region. The solid line represents the line of regression; the dashed line represents the line of identity. Linear regression analysis was performed only on the 20% to 100% flow range. The graph in the lower right shows the superimposed regression lines for each subset of animals. The regression lines become progressively lower between 10 minutes, 2 hours, and 4 hours, with a ²⁰¹Tl excess relative to the ischemic flow region. This is consistent with an relative increase in ²⁰¹Tl in the ischemic region relative to the nonischemic region and represents ²⁰¹Tl redistribution.

(From Pohost GM, Okada RD, O’Keefe DD, et al: Thallium redistribution in dogs with severe coronary artery stenosis of fixed caliber, Circ Res 48:439–446, 1981. Reprinted with modification and permission.)

Gamma well counting in postmortem tissue has obvious advantages and disadvantages. A major disadvantage is that the myocardial distribution of a radiotracer can only be determined at discrete time points. In addition, a given animal can be examined only once. Finally, the resolution of this approach is dependent on how finely the heart is sectioned, usually in 0.3-gm samples. The tissue counting approach is very reproducible and avoids interference of background activity. To facilitate comparisons of results between animals, myocardial radiotracer activity is usually normalized. One way to normalize activity is to express activity in each segment as a percentage of a predefined normal region.⁴² This method enables evaluation of the myocardial activity within an ischemic or infarcted region relative to a normal region in a manner analogous to the interpretation of clinical images. Alternatively, regional myocardial activity can be normalized by the computed average activity for each heart, as proposed by Yipintsoi et al.⁹⁰ A less desirable approach normalizes activity in each myocardial segment according to the segment with the highest activity.⁹¹

Evaluation of Tracer Kinetics and Kinetic Modeling

Regional myocardial uptake of a diffusible tracer is dependent on flow and myocardial extraction. The simplest approach to kinetic modeling assumes a monoexponential clearance. Under these conditions, dynamic image data are fitted to a monoexponential equation, A = A₀ e^−kt, where A₀ represents initial activity, K the decay constant in the elapsed time. Myocardial clearance half-time can be derived from the slope of the linear regression of the natural logarithmic transformation of the raw counts over time (t_1/2 = ln 2/k). To further simplify the analysis, the initial data from 0 to 2 minutes after injection may be excluded to avoid potential blood-pool contamination. More sophisticated compartmental modeling is often required. The dynamic data can be derived from both planar imaging and SPECT; this topic is covered in detail in other sections of the book.

Myocardial Infarction

It has been well established in canine models of coronary occlusion that myocardial necrosis progresses from the subendocardium to the subepicardium in a wavefront manner.^37,92 Advancing myocardial necrosis can be halted by coronary reperfusion. In canine models, permanent cellular injury begins after approximately 40 minutes of occlusion. The extent of necrosis depends on the area at risk and the duration of occlusion. In dogs, the maximum degree of infarction occurs after about 6 hours of occlusion. However, the extent of infarction may be affected by heart rate and left ventricular loading conditions. Thus, in anesthetized models, the method of anesthesia may affect infarct size. Myocardial flow following reperfusion varies and is critically dependent on timing after reperfusion, because a period of hyperemic flow is often seen following release of an occlusion.

Connelly et al. reported that complete coronary occlusion in the rabbit produced an expanding necrotic wavefront from the subendocardium.⁹³ This effect is similar to the phenomenon first described by Reimer and Jennings in the canine model of occlusion.⁹² In the rabbit, mild subendocardial necrosis can be detected by nitroblue tetrazolium staining as early as 15 minutes after occlusion; 1 hour of coronary occlusion consistently produces completed transmural infarction.

Coronary occlusion can be produced by various noninvasive and invasive methods.⁵² Noninvasive closed-chest approaches include sustained coronary occlusion with angioplasty catheters^38,39; selective coronary embolization⁹⁴; and deployment of intracoronary copper coils,⁹⁵ which promote thrombosis. Some of these approaches permit subsequent reperfusion. More invasive surgical approaches involve acute occlusion with a surgical ligature^37,92 or more gradual occlusion after implantation of an ameroid occluder.⁹⁶

Small-rodent models of surgical myocardial infarction are now widely used in the evaluation of the pathophysiology of infarction and postinfarction remodeling^97–101 and have only recently been used for evaluation of targeted radiotracers following infarction.^12,13,18,102 These small-animal imaging studies will be facilitated by the availability of high-resolution microSPECT and microPET imaging systems. Larger experimental models of coronary occlusion and reperfusion are more frequently used to assess the behavior of radiolabeled perfusion tracers under conditions of reduced flow and in the presence of myocardial necrosis. The myocardial uptake and retention of a perfusion tracer is critically dependent on the timing of injection relative to reperfusion.^79,103 Appropriate use of experimental occlusion-reperfusion models allows analysis of the uptake and clearance characteristics of a radiotracer in relation to cellular viability independent of flow.⁴³ These models have also been used to evaluate sympathetic innervation of the heart after coronary occlusion and reperfusion with iodine-123 (¹²³I) meta-iodobenzyl guanidine³⁹ or to examine myocardial metabolism with radiolabeled tracers of aerobic or anaerobic metabolism.^38,104–106

A model of chronic congestive heart failure can be produced by repeated coronary embolization^94,107 or by surgery.¹⁰⁸ Animals with extensive infarction producing extensive myocardial necrosis can be created by repetitive transmyocardial direct current shocks.¹⁰⁹

Myocardial Stunning

Myocardial stunning is characterized by a condition of postischemic mechanical dysfunction that persists after reperfusion despite the absence of irreversible injury. In the dog, a coronary occlusion lasting less than 20 minutes does not result in any myocardial necrosis; however, it produces regional dysfunction that may persist for hours.^110,111 Myocardial stunning can also be produced by repeated brief (5- or 10-minute) coronary occlusions.¹¹² In other species with fewer coronary collaterals, much briefer periods of coronary occlusion result in myocellular necrosis. Myocardial stunning can also occur after a prolonged partial coronary occlusion,¹¹³ in association with nontransmural infarction,¹¹⁴ or with exercise-induced ischemia.¹¹⁵

We recently developed a long-term canine model of ameroid-induced gradual subtotal coronary occlusion resulting in regional myocardial dysfunction (Fig. 3-17). Although resting flow was normal in these animals, on PET imaging with nitrogen-13 ammonia, coronary flow reserve was markedly impaired in response to adenosine. This model matches the clinical scenario reported by Vanoverschelde et al. in which reversible left ventricular dysfunction was identified in patients with viable collateralized occluded coronary arteries with preserved resting flow.¹¹⁶ In our canine model, dysfunctional regions with impaired coronary flow reserve also demonstrate increased fluorine-18 deoxyglucose uptake, similar to the clinical condition reported by Camici et al.¹¹⁷ These observations support the contention that the chronic regional dysfunction may be caused by repetitive stress-induced ischemia in the setting of normal resting flows and impaired flow reserve. This chronic model should prove useful in the evaluation of radiolabeled fatty acids such as [¹²³I]β-methyl-iodophenyl pentadecanoic acid (BMIPP), which are reportedly “memory” markers of reversible ischemic injury.¹⁰⁶

Figure 3-17 Long-term canine model of ischemic dysfunction. PET, perfusion, and metabolic imaging were performed in dogs 5 weeks after implantation of an ameroid occluder. Images from PET with nitrogen-13 ammonia (top row) demonstrate normal resting flow. Nitrogen-13 ammonia images during adenosine stress (middle row) demonstrate reduced flow in the anterior wall (between arrows). Increased accumulation of fluorine-18 deoxyglucose (arrow) is seen in the area of impaired flow reserve (bottom row). A coronary angiogram at 4 weeks showed 90% stenosis of the left anterior descending coronary artery at the site of the ameroid occluder. Magnetic resonance imaging and serial echocardiography showed profound regional dysfunction; however, postmortem examination revealed no significant necrosis in this region. This animal model seems to be analogous to the condition observed clinically in patients with collateralized viable myocardium subtended by a subtotal occlusion. Regional dysfunction can be attributed to repeated myocardial stunning.

All radiopharmaceuticals directed at evaluating myocardial metabolism should be evaluated in one of these pure experimental models of myocardial stunning. These models allow evaluation of a radiotracer under conditions of predominantly altered metabolism. As in infarct models of occlusion-reperfusion, the timing of radiotracer injection during conditions of post-reperfusion stunning is critical because of a transient period of hyperemic flow. The reperfusion hyperemia observed following a brief period (<15 minutes) of ischemia tends to be more intense but shorter in duration.

Myocardial Hibernation (See Chapter 36)

The term myocardial hibernation has generally been applied to patients with coronary artery disease who have chronically depressed left ventricular function in the absence of myocardial infarction that improves after coronary revascularization. It is unclear whether chronic left ventricular dysfunction in humans represents an adaptive response to chronic reduction in coronary blood flow.¹¹⁸ Impairment of left ventricular function under these conditions has been regarded as a protective mechanism by which the heart down-regulates its energy requirements, thereby preventing irreversible myocardial injury.

Short-term canine and porcine studies have successfully demonstrated sustained (1- to 5-hour) matching of reduced flow and function.^113,119,120 We also developed a short-term canine model of sustained low-flow ischemia (Fig. 3-18).^42,47,121 A hydraulic occluder is placed around a proximal coronary artery, allowing for the creation of a variable stenosis. The distal coronary artery is cannulated with a non–flow-obstructing catheter, which permits accurate regulation of the stenosis severity on the basis of measurement of distal coronary artery pressure and transstenotic pressure gradient. The degree of ischemic dysfunction can be assessed by measuring regional myocardial thickening with epicardial Doppler thickening probes.

Figure 3-18 Heart surgical preparation and instrumentation of a model of partial coronary occlusion and pacing-induced demand ischemia. A, Left atrial catheter; B, atrial pacing wire; C, Doppler thickening crystals; D, Doppler flow probes; E, hydraulic occluder; F, distal coronary artery catheter; and G, coronary venous catheters.

(From Shi CQX, Sinusas AJ, Dione DP, et al: Technetium-99m-nitroimidazole [BMS 181321]: A positive imaging agent for detecting myocardial ischemia. J Nucl Med 36:1078–1086, 1995. Reprinted with permission of the Society of Nuclear Medicine.)

Changes in regional metabolism can also be assessed in this model by selective venous sampling for arterial-venous balance measurements. Additional application of atrial pacing or intravenous administration of inotropic agents can provide a model of stress-induced ischemia as it may occur in patients with critical coronary artery disease.⁴⁷ Many investigators have used this type of low-flow model to evaluate radiolabeled tracers of perfusion, myocardial metabolism, and tissue hypoxia.^42,47,89,122 Using this experimental model of sustained low flow, the perfusion tracer ^99mTc-NOET (bis[N-ethoxy, N-ethyl dithiocarbamato]-nitrido ^99mTc) demonstrates substantial redistribution,¹²² as was previously observed for ²⁰¹Tl⁸⁹– and ^99mTc-sestamibi.⁴⁶

The condition of chronic low-flow perfusion-contraction matching has been more difficult to model experimentally. Recently, investigators succeeded in producing regional dysfunction over extended periods in pigs and dogs.^123–125 In each of these studies, the severity of regional dysfunction was out of proportion to the reduction in regional myocardial blood flow. In a clinical study, Vanoverschelde et al. suggested that chronic regional myocardial dysfunction in the presence of relatively preserved resting coronary flow may be attributed to repetitive stunning associated with intermittent stress-induced ischemia.¹¹⁶ The presence of ²⁰¹Tl rest redistribution in patients with reversible left ventricular dysfunction supports the idea that true myocardial hibernation with chronic resting reduction in flow can and does occur.¹²⁶ Whereas a resting defect may be attributed to partial volume errors, the presence of ²⁰¹Tl redistribution suggests differential myocardial clearance that is most likely secondary to reduced resting myocardial flow. Additional imaging studies in animal models of chronic hibernation are needed to establish the pathophysiology of reversible ischemic dysfunction in the presence of coronary artery disease. In an excellent recent review, Canty and Fallavollita suggest that chronic stunning and chronic hibernation represent the extremes of a continuum.¹²⁷ They present the recent experimental studies that demonstrated that there is a progression from chronic stunning with normal flow to hibernating myocardium with reduced resting flow. The application and evaluation of new metabolic radiotracers in these models will be critical for advancing our understanding of this important clinical condition. These low-flow models also will be invaluable in assessing new radiolabeled perfusion tracers.

Dilated Cardiomyopathy

The value of animal models in the evaluation of cardiomyopathies was recently reviewed by Yarbrough and Spinale.¹²⁸ Although several landmark discoveries regarding the pathophysiology of cardiomyopathies have arisen from small-animal models, the proof of concept must ultimately be reproduced in animal species that more closely mimic human physiology, function, and anatomy.¹²⁸ While small-animal models allow the identification of potential novel imaging targets, larger-animal models facilitate the direct translation of radiotracer imaging approaches to applications in patients with cardiomyopathies.

Many canine species are known to have idiopathic dilated cardiomyopathies.⁵² This is seen almost exclusively in the larger breeds of dogs and is characterized by biventricular cardiac dilation. Dilated cardiomyopathies have also been seen in domestic cats, and there is a cardiomyopathic strain of Syrian hamsters. Several investigators have developed experimental models of congestive heart failure associated with ventricular dilation. Chronic rapid atrial or right ventricular pacing in the dog or pig produces dilated cardiomyopathy.^129–131 Chronic mitral insufficiency also results in dilated cardiomyopathy. This can be accomplished surgically or noninvasively by cutting chordae tendineae of the mitral valve with a urologic forceps.^88,132 Other investigators have created a volume-overloaded state by creating aortic insufficiency.

Magid et al. developed a rabbit model of chronic aortic insufficiency.¹³³ The regurgitant fraction and resulting left ventricular dilation and hypertrophy produced by this rabbit model are similar to those reported in severe aortic insufficiency in human beings. Lu et al. used this model to evaluate the value of ¹¹¹In-labeled antimyosin antibody Fab fragment imaging for identification of myocyte necrosis associated with chronic aortic insufficiency.¹¹ These types of experimental studies help to generate hypotheses about the potential role of radionuclide imaging in the evaluation and management of patients with acute or chronic aortic insufficiency.

Atherosclerosis (See Chapter 44)

Several recent reviews have summarized the many models of atherosclerosis that are available for testing of radiotracers targeted at critical components of the atherosclerotic process.^134,135 Animal models of atherosclerosis generally involve use of a high-fat diet in combination with endothelial injury. Vascular injury can be produced by direct physical injury, chemical injury, or immunologic injury. Some animals develop atherosclerosis as a naturally occurring disease: The Watanabe heritable hyperlipidemic rabbit, for example, has a natural LDL-receptor deficiency and provides a natural model for homozygous familial hypercholesterolemia.¹³⁶ The St. Thomas’ Hospital (STH) strain of rabbits resembles human triglyceridemia and combined hyperlipidemia and also develops advanced atherosclerotic lesions.¹³⁷ The most commonly used animal model of atherosclerosis is the high-cholesterol-diet-fed rabbit.⁵² The pig is also commonly used as a model of atherosclerosis. Vallabhajosula et al. developed a novel porcine model of complex coronary atherosclerosis that employs balloon injury followed by an injection of cholesterol esters into the coronary arteries of healthy pigs, creating coronary lesions (15 mm in length).¹³⁸ These investigators demonstrated that the uptake of [¹⁸F]fluorodeoxyglucose (FDG) in the coronary lesions was dependent on the progression of the lesion. Within the first 2 to 3 months after the induction of lesion, FDG uptake was seen in some of the lesions. But after 4 months, most of the lesions were clearly identified by FDG. To reduce the problems associated with use of this large species, miniaturized breeds of swine have been developed. Nonhuman primates are considered to provide a model of atherosclerosis that is most similar to the atherosclerosis seen in humans. Limitations of the nonhuman primate include expense and the rapid development of atherosclerotic lesions.

The transgenic/knockout animal models have greatly enhanced our understanding of atherosclerosis. The most widely used mouse models of atherosclerosis are apolipoprotein E–deficient (ApoE^−/−), LDL receptor–deficient (LDLr^−/−), and ApoE*3Leiden (E3L).²⁶ These models may differ strikingly from one another in their response to specific experimental manipulations. The ApoE^−/−, mouse has become an established model of hypercholesterolemia and spontaneous atherosclerotic lesion development.^139,140 ApoE mice demonstrate lesions of all phases of atherosclerosis throughout the arterial tree.¹⁴¹ Williams et al. more recently demonstrated that the brachiocephalic lesion that develops in ApoE-deficient mice is characteristic of the unstable plaque seen in human patients.¹⁴² The brachiocephalic lesions that develop in the ApoE-deficient mice are characterized by large plaques with a thin fibrous cap and increased lipid core. These are all features associated with plaque rupture. Lesions in this region are also characterized by multiple buried caps, suggesting repeated rupture, and support a higher rupture rate of the brachiocephalic region compared with other vascular segments in these mice. The ApoE mouse fed a high-fat diet develops lesions similar to those in humans. Therefore, the ApoE model provides an excellent approach for evaluation of radiotracers targeted at unstable components of the atherosclerotic plaque. Figure 3-19 illustrates the gross atherosclerotic lesions that develop in ApoE-deficient mice.

Figure 3-19 Illustration of the gross atherosclerotic lesions that develop in ApoE-deficient mice after prolonged feeding on a high-fat diet. A, Aorta isolated within animal demonstrates location of the brachiocephalic artery (open arrow), which is a common site for atherosclerotic lesions that have many of the characteristics of unstable plaque. B, Entire aorta dissected free, demonstrating atherosclerosis throughout the aorta. C, Aorta opened for better visualization of the endocardial lesions that are present in the aortic arch. Arrows indicate location of atherosclerotic plaques.

(Figure provided courtesy of Paula J. Silva, Bristol-Myers Squibb Medical Imaging, Billerica, MA.)

The LDLr^−/− mouse model develops atherosclerosis, particularly following feeding with a lipid-rich diet. E3L mice have a mutant form of the human apoE3 gene and develop atherosclerosis when fed a high-cholesterol diet. These E3L mice are more sensitive to lipid-lowering drugs then either the ApoE^−/− or LDLr^−/− mice.²⁶

Mouse models have proved useful in the study of atherosclerosis, but differences in anatomy, lipid metabolism, and gene expression complicate translation of experimental results obtained in mice to humans.²⁶

Models of atherosclerosis have been used to evaluate a number radiotracers targeted at the vulnerable plaque.^143,144 As outlined earlier, plaques vulnerable to rupture typically have a large, necrotic lipid core and an attenuated fibrous cap that are significantly infiltrated by macrophages and lymphocytes. There is emerging evidence that apoptosis contributes to the instability of the atherosclerotic lesion. Apoptosis of macrophages contributes substantially to the size of the necrotic core,¹⁴⁵ whereas apoptosis of smooth muscle cells results in thinning of the fibrous cap.¹⁴⁶ Annexin V can be radiolabeled with ^99mTc and has been previously used for the noninvasive detection of apoptosis in myocardial infarction and inflammatory myocardial diseases. Kolodgie et al. applied annexin V planar in vivo and ex vivo imaging in experimental atherosclerotic lesions in hypercholesterolemic rabbits for noninvasive detection of atherosclerosis.¹⁴³ They demonstrated that annexin V targeted apoptotic macrophages within the atherosclerotic plaque and may provide an attractive imaging agent for the noninvasive detection of unstable atherosclerotic plaques. Narula et al. also demonstrated the feasibility of noninvasive visualization of experimental atherosclerotic lesions with a mouse/human chimeric antibody Z2D3 F(ab′)₂ directed against proliferating smooth-muscle cells.¹⁴⁷ Narula et al. subsequently demonstrated that ¹¹¹In-Z2D3 antibody imaging could be used for noninvasive assessment of the rate of smooth-muscle cell proliferation after coronary angioplasty.¹⁴⁸ Other investigators using a rabbit model of atherosclerosis demonstrated that radiolabeled MCP-1 may also be a useful tracer for imaging monocyte/macrophage-rich experimental atherosclerotic lesions.¹⁴⁴