Michael R. Lewis, PhD
Current medical imaging modalities include radiography, computed tomography (CT), magnetic resonance imaging (MRI), planar gamma scintigraphy, single photon emission computed tomography (SPECT), and positron emission tomography (PET). Radiography and CT provide only anatomic information, whereas MRI can provide anatomic information as well as some functional information, for example changes in blood flow, when functional MRI (fMRI) is performed in a relatively strong magnetic field. In contrast, scintigraphy, SPECT, and PET do not provide anatomic information but rather functional, physiologic, and metabolic information in both static and dynamic modes. Therefore, scintigraphy, SPECT, and PET are considered to be molecular imaging modalities, utilizing radiopharmaceuticals to probe molecular processes or targets that contribute to function, physiology, and/or metabolism.
Characteristics of Radiopharmaceuticals
Radiopharmaceuticals are radioactive drugs or compounds used in nuclear medicine for the diagnosis or therapy of catastrophic diseases, including cancer. In nuclear medicine, approximately 95% of radiopharmaceuticals are used for diagnosis, while the other 5% are used for therapy. However, radiopharmaceutical therapy is projected to grow exponentially over the next five to ten years. Moreover, as radioactive and non-radioactive biological or targeted therapies for cancer are developed, the need for molecular imaging correlates for these therapies becomes increasingly important for treatment planning. With a couple of rare exceptions, radiopharmaceuticals have no pharmacologic effects, since they are used in so-called "tracer" quantities; that is, administered in minute mass amounts below their respective pharmacologic thresholds.
Ideal characteristics of molecular imaging radiopharmaceuticals include radioactive decay mode, emission energies, physical half-life of the radionuclide label, biological half-life of the targeting vector, high target to non-target tissue uptake ratios, and availability. The decay mode of the radionuclide incorporated into the radiopharmaceutical should result in emission of low to medium energy gamma rays or x-rays for SPECT or emission of positrons for PET. Desirable energy characteristics of radioactive emissions are those that can be readily absorbed by a scintillation detector, as well as easily collimated for imaging focus and shielded for personnel safety. For example, the "workhorse of nuclear medicine," technetium-99m, emits a 141 keV gamma ray that is easily absorbed and detected with conventional thallium-doped sodium iodide crystals approximately 1 cm thick, allowing efficient two-dimensional imaging by planar scintigraphy or three-dimensional imaging by SPECT. In SPECT, the head or multiple heads of a gamma camera are rotated about the subject, allowing multiple acquisitions to be obtained at different angles, which are then reconstructed into a semi-quantitative three-dimensional image. In PET, positron annihilation radiation, emitted as two 511 keV photons at an angle of 180 ± 0.5°, is detected by rings of scintillation detectors wired in coincidence circuits, such that two opposing 511 keV emissions detected within a few nanoseconds are recorded as a "true" or "coincidence" event. "Random" events, that is, the detection of single photons, are discarded, and photons of lower energy than 511 keV are stopped by lead septa covering the detectors in the PET scanner gantry. By back-projection along the coincidence lines of true events, the location of radioactive emissions within the subject can be reconstructed in three dimensions and, unlike SPECT or planar scintigraphy, determined with absolute quantitation.
The physical half-life of the radionuclide with which a radiopharmaceutical is labeled should be long enough to formulate the radiopharmaceutical and accomplish the molecular imaging procedure, yet short enough to minimize the overall radiation dose to the patient. Charged particle emissions such as β- and α particles should be avoided if at all possible, because these particles are cytotoxic and contribute to an unwanted radiation burden to the patient. In general, the physical half-life of the radionuclide should be matched well to the biological half-life of the radiopharmaceutical. Like the radionuclide half-life, the biological half-life of the radiopharmaceutical targeting vector should be long enough to accomplish the molecular imaging procedure, but short enough to minimize the radiation dose to the patient. Furthermore, the biological properties of the radiopharmaceutical should result in a high target to non-target tissue ratio. Factors that contribute to high target to non-target ratios include rapid blood clearance, rapid localization in target tissue, and rapid clearance from metabolizing and excreting tissues such as liver, kidneys, and the gastrointestinal tract.
Physiologic Imaging: Metabolic Processes
The first forms of molecular imaging targeted metabolic processes; that is, these modalities employed radiopharmaceuticals with biological activity or mimicking such activity. The earliest example was thyroid scintigraphy, using technetium-99m pertechnetate or radioiodide to image such diseases as hyperthyroidism or thyroid cancer. Pertechnetate is an iodide mimic that is readily taken up by the sodium/iodide symporter in thyroid tissue. Iodine-123 sodium iodide is also used for thyroid scintigraphy, as it has decay characteristics similar to pertechnetate, but with the advantage that it is trapped by organification in the synthesis of iodine-containing thyroid hormones. This organification results in higher target to non-target ratios and superior imaging contrast compared to pertechnetate. While iodine-123 represents standard of care in diagnosing human thyroid diseases, it is cost prohibitive in veterinary medicine, being approximately twenty times the price of pertechnetate. An important use of technetium-99m pertechnetate or iodine-123 sodium iodide is to identify patients who should benefit from iodine-131 sodium iodide therapy. When such therapy is indicated, patients are usually imaged with a small, non-therapeutic dose of iodine-131, in order to confirm efficacious biodistribution and thyroid uptake before therapy. Moreover, the use of technetium-99m or iodine-123 prior to the use of iodine-131 is an example of the close and intractable relationship between molecular imaging and therapy in nuclear medicine.
Another common example of physiologic imaging is the targeting of osteoblastic activity for bone scintigraphy. Metabolically stable technetium-99m phosphorous compounds (technetium-99m MDP, HDP, or EDP) mimic phosphate in the hydroxyapatite matrix of bone and localize to areas of increased and abnormal osteoblastic activity, as is common in osteosarcoma and bony metastases of prostate and mammary cancer. Furthermore, the biodistribution and bone uptake of these technetium-99m imaging agents is essentially identical to that of the therapeutic drug samarium-153 ethylenediaminetetramethylene phosphonate (samarium-153 EDTMP). Analogous to the management of thyroid diseases, a routine bone scan with technetium-99m MDP, HDP, or EDP can predict which patients should benefit from samarium-153 EDTMP therapy. Diagnosis and treatment of bone tumors with radiopharmaceuticals again represents the important relationship between molecular imaging and therapy.
PET imaging with fluorine-18 fluorodeoxyglucose (FDG) has been approved by the United States Food and Drug Administration (FDA) for the non-invasive diagnosis, staging, and restaging of at least fourteen human cancers. FDG is a glucose mimic that is taken up by glucose transporters, such as GLUT 1, and phosphorylated by hexokinase. However, phosphorylated FDG is not a substrate for the next enzyme involved in glucose metabolism, phosphoglucose isomerase. Phosphorylated FDG cannot be utilized in further glycolysis, is not a substrate for glucose transporter-mediated efflux, and is negatively charged and unable to diffuse across the plasma membrane of cells. Thus, FDG becomes metabolically trapped in tissues with high rates of glucose metabolism. While initially considered as a potentially efficacious brain and heart imaging agent, FDG-PET and FDG-PET/CT have emerged as highly important molecular imaging modalities for human cancer. Unfortunately FDG-PET is currently cost prohibitive in veterinary medicine, over ten times the cost of scintigraphic procedures, and this imaging modality is further limited by the lack of PET scanners in veterinary clinics. However, leading physicians in the development and use of FDG-PET predict that the cost of this procedure will continue to decrease until it is comparable to that of scintigraphy in the near future. At least one experimental evaluation of FDG-PET has been performed in canine lymphoma, and it is expected that this modality will find increasing routine use in veterinary oncology, as cost and availability improve.
Future Directions in Molecular Imaging: Molecule on Molecule
Cancer cells have a wide variety of biological molecules, such as antigens, receptors, and oncogene products, present in much higher concentrations than in normal tissues. This characteristic affords the opportunity to develop radiopharmaceuticals that bind to these molecules with high affinity, specificity, and selectivity. Thus, advances in this area allow molecular imaging to be based on molecular recognition, as opposed to physiologic and metabolic processes. The ability to probe the molecular characteristics of tumor cells can provide important prognostic information that impacts treatment management. For example, estrogen receptor status in mammary cancer and androgen receptor status in prostate carcinoma can be determined by molecular imaging with radiolabeled estrogen and testosterone derivatives, respectively. In addition, the substitution of an imaging radionuclide with a therapeutic radionuclide in the same targeting vector allows the molecular characteristics of cancer to be exploited for treatment.
When designing radiopharmaceuticals for molecular imaging, it is important to consider aspects of signal source, scaling, and signal-to-noise issues. If the molecular target and signal source is DNA, there are typically only two molecules per cell, presenting an insurmountable obstacle to achieving adequate signal-to-noise for detection using current technology. Furthermore, imaging DNA provides no information on whether a given sequence or gene is active. Gene expression can potentially be detected by molecular imaging at the messenger RNA level. However, most mRNAs are present in a range of fifty to a few thousand copies per cell, which would require an imaging agent with very high specificity and sensitivity. Gene products such as proteins typically have concentrations of a few hundred to millions of molecules per cell. Therefore, most "molecule on molecule" imaging has focused on protein targets, particularly cell surface proteins, for which intracellular barriers do not exist. The highest capacity target for imaging based on molecular interaction is in the area of protein function, where the potential for massive signal amplification exists. The physiologic imaging of thyroid, bone, and glucose metabolism, described above, represents three examples of imaging protein function, as opposed to protein imaging by molecular binding.
In 1984 Niels Jerne, Georges Köhler, and César Milstein were awarded the Nobel Prize in Physiology or Medicine, in part for their "discovery of the principle for the production of monoclonal antibodies." This discovery ushered in a new era of molecular medicine, as monoclonal antibodies were widely thought to be the ultimate "magic bullets" for non-invasive imaging of cancer and other diseases. In the area of radiopharmaceutical therapy, two monoclonal antibody products became the first FDA-approved biological radiotherapeutics, Zevalin® labeled with yttrium-90 and Bexxar® labeled with iodine-131, for treatment of recurrent, refractory B-cell lymphoma. Prior to radioimmunotherapy with these drugs, treatment planning sessions are routinely performed with indium-111 Zevalin® and with small doses of Bexxar®, respectively. However, monoclonal antibodies have long residence times in the blood pool, compromising imaging contrast, especially in the case of Zevalin®. Dehalogenation of Bexxar® in normal tissues partially offsets the problem of low imaging contrast, but at the cost of dehalogenation in tumors as well. Tumor dehalogenation of iodine-131 can compromise not only imaging results, but also the efficacy of radioimmunotherapy and potentially lead to hypothyroidism associated with release of free iodide. In solid tumors, radiolabeled antibodies suffer from a number of major problems, including long circulation times, slow accretion into tumor masses, uneven tumor penetration, and antigenic heterogeneity, each of which, or a combination of several, typically results in poor imaging contrast.
An alternative to the use of radiolabeled antibodies for molecular imaging is the utilization of tumor-associated receptor binding peptides. The small size of these peptides leads to rapid biodistribution, rapid uptake in tumors, and rapid excretion of the residual dose by renal filtration and urinary elimination. Somatostatin receptor-expressing tumors have received considerable attention since the FDA approval of indium-111-labeled octreotide (OctreoScan®) for the diagnosis of pancreatic, carcinoid, lung, and other cancers by planar scintigraphy or SPECT. Copper-64-labeled octreotide has also been evaluated for PET imaging of these cancers in human patients, resulting in superior molecular imaging quality and the detection of more tumors than OctreoScan®. However, limitations of tumor receptor-targeting peptides include low absolute tumor uptake compared to antibodies and relatively rapid washout from tumors, limiting the ability to use these peptides particularly for translation of molecular imaging to targeted radiotherapy.
Antibody pretargeting is an alternative approach in which an unlabeled antibody-receptor conjugate or fusion protein is first administered and allowed to accumulate in tumors, and then radionuclide imaging or therapy is given in the form of a small effector molecule that binds rapidly with high affinity to the antibody-receptor construct at the tumor site. When successful, this process results in immediate accumulation of radioactivity in tumors, causing substantial improvements in tumor-to-blood ratio, imaging contrast, and normal organ uptake. Thus, pretargeting combines the desirable properties of higher tumor uptake of antibodies with more rapid biodistribution of radiolabeled peptides, with extremely fast whole-body clearance of radioactivity. On the other hand, challenges in the use of pretargeting include immunogenicity of some agents, complex dosing protocols, and patient compliance.
The widespread focus on radiolabeled antibodies and peptides for imaging cell surface molecules has resulted in most tumor-associated antigens and receptors being extensively studied. In contrast, a wealth of largely unexplored tumor targets reside inside cells, but imaging of these molecules can be quite challenging, owing to often low target concentrations and numerous intracellular barriers to be surmounted in order to reach those targets. One area of intracellular activity in cancer that has received increasing attention is apoptosis, the major mechanism by which chemotherapeutic drugs kill tumor cells. Indium-111-labeled annexin V has been used to image apoptosis in organ transplant rejection and in cancer, by targeting phosphatidyl serine exposed on the plasma membrane. A technetium-99m-labeled, membrane-penetrating peptide containing a capsase-3 substrate sequence was used to image a mouse model of Fas ligand-induced apoptosis, an important pathway in the action of chemotherapy. In our laboratory, we have focused on imaging inhibitors of apoptosis, notably the bcl-2 proto-oncogene in B-cell lymphoma. Overexpression of bcl-2 correlates with higher relapse rates, shorter disease free intervals, and poor overall survival in people with aggressive B-cell lymphoma. Thus, molecular imaging of bcl-2 may aid in the identification of patients at risk for relapse or conventional treatment failure. Such patients might benefit from alternative molecular therapies that act to down-regulate bcl-2 directly or indirectly. Unfortunately there are no high affinity ligands for bcl-2 protein suitable for radiopharmaceutical development. Therefore, we have used peptide nucleic acid (PNA), a DNA-like molecule with high biological stability, to image bcl-2 expression at the mRNA level. A major drawback to PNA-based imaging agents is that PNA is taken up very poorly by cells. Because approximately 87% of human B-cell lymphomas express somatostatin receptors, we have coupled an indium-111-labeled bcl-2 antisense PNA to a somatostatin receptor-binding octreotide derivative for intracellular delivery. We have successfully used this bcl-2 antisense radiopharmaceutical to image human B-cell lymphoma tumors in mice by SPECT and PET. These studies have led to the translation of bcl-2 mRNA imaging to canine B-cell lymphoma patients. Our hypothesis is that dogs with naturally occurring B-cell lymphoma will demonstrate tumor-specific uptake of the molecular imaging agent that correlates negatively with response to chemotherapy. After demonstrating somatostatin receptor imaging of canine B-cell lymphoma with an indium-111-labeled octreotide derivative, we have conducted fourteen imaging sessions with the bcl-2 mRNA-targeted PNA derivative. Planar scintigraphic findings have included true positive and false negative results in naïve patients, as well as true negative and false positive images of dogs in complete remission and true positive images of dogs in partial remission after six weeks of chemotherapy. It should be noted, however, that these results are based on nuclear medicine criteria, not bcl-2 molecular status. Molecular biology studies of excisional biopsy specimens are currently underway to determine whether positive and negative scans correlate with overexpression of bcl-2 mRNA and protein. These studies are designed to evaluate the functional validity of this experimental molecular imaging modality.
Molecular imaging with radiopharmaceuticals has made considerable progress from imaging physiologic and metabolic processes to non-invasive detection of molecular recognition both at the surface and in internal compartments of tumor cells. The increasing sophistication with which molecular imaging has been employed offers numerous possibilities for understanding tumor biology and cancer treatment. While much focus has been placed on laboratory models and human patients, these efforts are likely to result in increasing use of molecular imaging agents and modalities in veterinary oncology.
1. Bernier DR, Christian PE, Langan JK. Eds., Nuclear Medicine: Technology and Techniques. St. Louis: Mosby, Inc.; 1997.
2. Saha GB. Fundamentals of Nuclear Pharmacy. New York: Springer-Verlag; 1998.
3. Welch MJ, Redvanly CS. Eds., Handbook of Radiopharmaceuticals: Radiochemistry and Applications. Hoboken, NJ: John Wiley and Sons, Inc.; 2003.