Single Photon Emission Computed Tomography Tracer

Abstract

Single photon emission computed tomography (SPECT) is the state-of-the-art imaging modality in nuclear medicine despite the fact that only a few new SPECT tracers have become available in the past 20 years. Critical for the future success of SPECT is the design of new and specific tracers for the detection, localization, and staging of a disease and for monitoring therapy. The utility of SPECT imaging to address oncologic questions is dependent on radiotracers that ideally exhibit excellent tissue penetration, high affinity to the tumor-associated target structure, specific uptake and retention in the malignant lesions, and rapid clearance from non-targeted tissues and organs. In general, a target-specific SPECT radiopharmaceutical can be divided into two main parts: a targeting biomolecule (e.g., peptide, antibody fragment) and a γ-radiation-emitting radionuclide (e.g., ^99mTc, ¹²³I). If radiometals are used as the radiation source, a bifunctional chelator is needed to link the radioisotope to the targeting entity. In a rational SPECT tracer design, these single components have to be critically evaluated in order to achieve a balance among the demands for adequate target binding, and a rapid clearance of the radiotracer. The focus of this chapter is to depict recent developments of tumor-targeted SPECT radiotracers for imaging of cancer diseases. Possibilities for optimization of tracer design and potential causes for design failure are discussed and highlighted with selected examples.

1 Introduction

Single photon emission computed tomography (SPECT) and positron emission tomography (PET) are valuable molecular imaging modalities as both are capable of detecting minute amounts of radioactive tracer [232, 272]. Clinical PET is currently about 2–3 orders of magnitude more sensitive than SPECT, has a better spatial resolution, and offers superior quantification. Nowadays, many nuclear imaging centers possess PET or PET/CT scanners. However, the large infrastructure that is needed for the production of β⁺-emitting radioisotopes (e.g., ¹⁸F, ¹¹C, ⁶⁴Cu,⁶⁸Ga, ⁴⁴Sc) makes PET an expensive technology. At the moment, an approved clinical grade generator for PET radioisotopes is only available for ⁶⁸Ga. Hence, for routine application SPECT is still the state-of-the-art nuclear imaging modality because it is less expensive and can make use of a broader array of suitable and available radionuclides (Table 7.1). Importantly, SPECT imaging is a useful technology for monitoring targeted radionuclide therapy employing radioisotopes that emit—concomitantly with the therapeutic radiation—γ-rays of suitable energies for SPECT (e.g., ¹⁷⁷Lu, ^188/186Re, ⁶⁷Cu, ¹³¹I, ²¹³Bi) [10].

Table 7.1 Selection of radioisotopes for SPECT imaging (and therapy)

Generally, SPECT radiopharmaceuticals can be classified according to their biodistribution characteristics. There are those whose tissue distribution is determined exclusively by their chemical and physical properties and those whose distribution and accumulation are determined by their specific interaction with a biological target that is expressed at the site of interest (e.g., tumor-associated receptor) [24, 157]. Herein we focus on the development and (pre) clinical application of target-specific radiotracers. A target-specific SPECT radiopharmaceutical can be divided into two main parts: a targeting biomolecule and a γ-radiation-emitting radionuclide [157]. In the case of using radiometals as the radiation source, a bifunctional chelator is needed as an additional component of the radiopharmaceutical. Thus, a metallic radioisotope is coordinated by a suitable chelating agent that is conjugated to the targeting agent via a linker entity (Fig. 7.1). In a rational design of a SPECT tracer, the single components have to be critically evaluated in order to achieve a balance among the demands of an adequate target binding and a rapid excretion.

Fig. 7.1

Schematic representation of a biomolecule (targeting vector molecule) that is conjugated to a radiometal. The bifunctional chelator that is conjugated to the target-specific biomolecule is necessary for coordination of the radiometal. Two strategies are possible to conjugate the biomolecules to the chelator and the radionuclide: pre-labeling and post-labeling. After radiolabeling, the target-specific radiobioconjugate was intraveniously administered, binds to the tumor cell surface-associated target (e.g., receptor) and allows visualization of target-expressing malignant lesions via SPECT

The majority of diagnostic radiopharmaceuticals currently available in nuclear medicine make use of metallic radioisotopes. For SPECT imaging, technetium-99m is the most widely applied radioisotope because of its ideal physical decay properties and easy availability by a generator system (Table 7.1). Indium-111 is another SPECT radioisotope frequently used in the clinics where it is often employed as a surrogate for yttrium-90 analogs since ⁹⁰Y that is used for therapeutic purposes is a pure β⁻-emitter. In contrast, clinical application of gallium-67 is relatively rare. Non-metallic radionuclides used for SPECT are basically the isotopes of iodine. Iodine-123 is the preferred isotope for imaging purposes (Table 7.1) due to its dosimetry and imaging characteristics that are superior to iodine-131 and iodine-125.

A targeting biomolecule serves as a “carrier” for specific delivery of the radionuclide to the target-expressing cells of interest. Such biomolecules could be biomacromolecules like specific antibodies (or antibody fragments) or small-molecular-weight molecules (e.g., peptides, vitamins, nucleosides). Each class of targeting agents has its pros and cons for its use in diagnostic nuclear medicine and for a potential translation to therapeutic applications. Peptide-based radiopharmaceuticals represent by far the largest group of tumor-targeted nuclear imaging agents currently in use.

During tracer development, the first steps are based on chemistry and molecular biology methods such as peptide syntheses, conventional or combinatorial chemistry, and phage display techniques for preparation, identification, and isolation of high-affinity binders to a particular receptor. Determination of the tumor-targeted radiotracer’s stability in vitro and its ability to bind with high affinity to the target structure on cultured cancer cells are the first requirements in this early development stage. The in vitro evaluation is followed by investigations in vivo using an adequate animal model, typically tumor-bearing small rodents. It is important to recognize that radiolabeled tumor imaging agents display different biodistributions and pharmacokinetics in animal models compared to humans due to a different metabolism, differences in the volume of distribution, and potential cross-reactivity of the targeting entity with normal tissues expressing the target receptor or antigen in humans [42]. Significant variability in the tissue distribution of radiotracers might occur among different animal models (e.g., mice vs. rats) or different animal strains (e.g., nude mice vs. normal mice). However, small rodents have emerged as generally the most useful and cost-effective animal models for developing and evaluating radiotracers and to test new experimental approaches to increase their localization in tumors. Postmortem biodistribution studies allow the detection and quantification of a cumulated activity portion in targeted and non-targeted tissues, and thus the determination of the radiotracer’s pharmacokinetic profile. Collection of blood and tissue samples for identification of metabolites at different time points after radiotracer application provides information about the radiotracer’s circulation time and it is in vivo stability. By increasing the availability of small-animal SPECT and SPECT/CT scanners in recent years, the process of radiotracer development has been significantly improved and accelerated while the number of test animals required has been reduced. Thus, a wide variety of targeted SPECT radiotracers are currently being developed and preclinically tested for imaging of various tumor types expressing one or more of the most relevant receptor types [247].

The focus of this chapter is to present general aspects for the design of SPECT tracers followed by specific examples of recent SPECT imaging agents based on biomacromolecules like antibodies, antibody fragments, proteins as well as other small-molecular-weight biomolecules such as peptides, vitamins, or nucleosides referred to as vector molecules. The examples demonstrate possibilities for optimization of the tracer design by tuning single components of these imaging agents. Finally, potential causes for failures in SPECT tracer design are discussed.

2 General Aspects for the Design of SPECT Tracers

The ideal SPECT tracer exhibits excellent tissue penetration, high affinity to the target structure, specific uptake and retention in the target cells, and rapid clearance from non-targeted tissues and organs. In addition, it is highly stable in vivo, easy to prepare, and safe for human application. These aspects are crucial because injected radiotracers that are not stable, not bound to the target, or not rapidly excreted create high background signals resulting in low tumor-to-background contrast, false positive results, and unnecessary radiation dose burden to the patient [10].

In the case of metallic radioisotopes, a bifunctional chelator is needed that is covalently linked to a biomolecule [157]. Since the stability of the radiometal complex is a critical aspect for the success of a radiopharmaceutical, it is important to choose an ideal chelating system that allows the formation of radiometal complexes of high thermodynamic stability and kinetic inertness [24]. Among the SPECT isotopes currently in use, technetium-99m is still the workhorse of diagnostic nuclear medicine. It is used in the majority of diagnostic scans conducted each year in hospitals worldwide. The preferred use of ^99mTc-radiopharmaceuticals reflects the ideal nuclear properties of the isotope and, until recently, the convenient availability from commercial ⁹⁹Mo/^99mTc-generators.

Technetium is a transition metal that presents a major challenge with respect to designing radiopharmaceuticals with favorable in vivo properties. In order to link the radionuclide to a targeting molecule, [^99mTc]pertechnetate with Tc in the oxidation state +VII that is eluted from the ⁹⁹Mo/^99mTc-generator must be reduced to build a complex with an appropriate bifunctional chelating system, most commonly in the oxidation state +I, +III or +V. The ^99mTc(V)-oxo and ^99mTc(V)-organohydrazino cores are most extensively studied (Fig. 7.2). The ^99mTc(V)-oxo-core generally adopts a square-pyrimidal geometry with the π-bonding oxo-group in the apical position. The core is stabilized by σ- and π-donating groups where amino, amido, and thiolate ligands as well as tetradentate ligands of the N_xS_4-x-class have been investigated [24, 157, 211]. A prominent example of a tetradentate chelator is the peptide-based chelator mercapto-acetylglycylglyclgylcine (H₅MAG₃) [150].

Fig. 7.2

The most frequently used ^99mTc-complexes for radiobioconjugates. a ^99mTc(V)oxo core, b ^99mTc(V)dioxo core, c ^99mTc(V)organohydrazino nicotinamide (HYNIC) core, and d ^99mTc(I)-tricarbonyl core. M = ^99mTc, L = co-ligands

An alternative approach is the use of the ^99mTc(V)-organohydrazino nicotinamide (HYNIC) core that was first introduced by Abrams et al. 20 years ago [1, 250]. The advantages of this system are the facile functionalization of targeting entities via amide linkage. It has therefore been used for ^99mTc-labeling of a variety of high, medium, and small-molecular weight biomolecules [65, 66, 83, 107, 161, 235, 274, 279, 296]. Since the HYNIC chelator can only occupy one or two coordination sites on the metal center, co-ligands such as tricine are needed to complete the coordination sphere of ^99mTc [74, 160, 219]. The possibility for selection of appropriate co-ligands is advantageous for an easy modulation of the hydrophilicity and pharmacokinetics of the ^99mTc-HYNIC-derivatized biomolecules. However, the presence of multiple species in solution due to different bonding modalities of the HYNIC moiety and co-ligands might be problematic for a commercial development, because of the increasing regulatory hurdles and the requirements of fully characterized products.

Another, less frequently employed approach is the use of a ^99mTc(V)-dioxo-core coordinated by nitrogen ligands that form octahedral complexes with the oxygens trans to each other [129]. The group of Nock and Maina made successful use of a tetraamine chelator for ^99mTc-radiolabeling of several peptide-based biomolecules forming monocationic polar complexes with the ^99mTc(V)-dioxo core [171, 205, 207]. The advantages of this radiolabeling strategy include its easy formation at ambient temperature, its high stability in the biological milieu, and considerable hydrophilicity.

A completely different ^99mTc-radiolabeling strategy has been introduced by the development of the tricarbonyl technique which offered new opportunities for the design of ^99mTc(I)-radiotracers [6,7,8,9, 75, 76, 240, 241]. The water-soluble ^99mTc(I)-tricarbonyl precursor’s aqua ligands are readily exchanged allowing the coordination of preferentially tridentate chelators that can be modified to provide complexes with cationic, neutral, or anionic overall charge [92, 174, 194, 216, 240, 251]. In addition, the tricarbonyl radiolabeling strategy is also accessible for the preparation of stable radiometal complexes using β-particle-emitting rhenium isotopes (^186/188Re, Table 7.1). Hence, the production of isostructural compounds with the “matched pair” ^99mTc/^188/186Re for diagnostic and therapeutic purposes has become feasible thanks to the tricarbonyl strategy, a feature which is often not fulfilled with Re(V) complexes [195].

Radiolanthanides (e.g., ¹⁷⁷Lu) and lanthanide-like isotopes as well as indium and gallium are used in the oxidation state +III. They can generally be coordinated by polyaminopolycarboxy chelating systems. Coordination numbers of lanthanides are typically between seven and ten whereas coordination numbers of eight or nine are most common in Ln(III) complexes with polydentate chelators. The 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA) chelator emerged as particularly useful for lanthanide coordination of therapeutic radiopharmaceuticals because of the formation of metal complexes of extremely high thermodynamic stability and kinetic inertness (Fig. 7.3). In addition, the hydrophilic acetate chelating arms of DOTA favor a fast clearance of radiotracers from the blood and non-targeted organs and tissues. Despite the similarities of the SPECT radioisotopes gallium-67 and indium-111 they are different in size, coordination number, and charge density. Ga³⁺ has a small ionic radius (0.65 Å) and the coordination number is six whereas the ionic radius of In³⁺ is larger (0.92 Å) and it is seven- or eight-coordinated in its complexes. The structural differences among Ga and In complexes might have an influence on the overall tissue distribution of one and the same bioconjugate as recently exemplified with a somatostatin analog [114]. A higher tumor uptake and a lower kidney retention have been reported for ⁶⁷Ga-DOTATOC compared to that of ¹¹¹In-DOTATOC. Whereas DOTA appears to be an ideal chelator for coordination of lanthanide (radio)isotopes like Lu³⁺ or In³⁺, its coordination cavity is not ideal for Ga³⁺ as it is too large. On the other hand, there is a perfect fit between the size of Ga³⁺ and the coordination cavity formed by the N₃O₃ donor atoms of the macrocyclic 1,4,7-triazacyclononane-1,4,7-triacetic acid (NOTA) chelator [157]. Consequently, a higher thermodynamic stability constant has been found for Ga-NOTA complexes compared to those of Ga-DOTA-complexes [67]. In some cases, open chelating systems are more favorable than macrocycles because they are capable to form radiometal complexes at ambient temperature which is particularly important for temperature-sensitive targeting agents. Examples are variable versions of diethylenetriamine pentaacetic acid (DTPA, CHX-A″-DTPA, etc., Fig. 7.3). DTPA is one of the most commonly employed acyclic ligands in radiochemistry useful for coordination of ¹¹¹In, ⁶⁷Ga, and radiolanthanides. For ¹¹¹In-complexation DTPA emerged as the ideal chelating agent [173].

Fig. 7.3

The most frequently used macrocyclic (DOTA, NOTA) and acyclic (DTPA) chelators for complexation of radioisotopes for SPECT imaging (e.g., ¹¹¹In, ⁶⁷Ga) and combined therapy/SPECT imaging (e.g., ¹⁷⁷Lu)

In addition to the bifunctional chelator’s function for stable coordination of the radiometal, the linker entity is important for conjugation with the biomolecule and might influence the overall pharmacokinetics of the radiopharmaceutical. By affecting the biomolecule’s lipophilic or hydrophilic characteristics the linker system can serve for controlling its in vivo behavior. Thus, the nature of a bifunctional chelator in terms of geometry, lipophilicity, and overall charge plays a crucial role in determining the biodistribution of (tumor-) targeted radiopharmaceuticals [24].

Functionalization of amino acid side chains (e.g., lysine, cysteine) with chemically reactive probes of bifunctional chelators is a largely uncontrolled random process that results in a heterogeneous mixture of conjugates modified at variable sites. A considerable advantage of small-molecular weight biomolecules (e.g., peptides and vitamins) is the fact that derivatization with a bifunctional chelating agent can be governed by specific chemical reactions that yield a single, clearly defined species. In contrast, loss of binding affinity is of concern during the process of antibody derivatization because modification of the Fab region (antigen-binding site) can possibly have deleterious effects on the target binding of the protein. Both loss of binding activity to the target and overlabeling effects are highly undesired processes because they result in unwanted background radiation and unspecific accumulation of the antibody radioconjugates in the liver. For this reason, recent endeavors were undertaken for the development of site-specific derivatization via enzymatic reactions that are selective for a particular amino acid [128, 185] or sugar residue [34] at a specified site of the antibody.

Since small-molecular-weight molecules are usually stable at a broad range of temperatures and pH values, the radiolabeling procedure is mostly smooth and quantitative. In contrast, proteins are generally sensitive to elevated temperatures. Thus, commonly applied methods for radiometal-labeling of proteins are time-consuming due to the low reaction temperature applicable. To overcome this drawback, pre-labeling strategies have been proposed allowing the preparation of radioimmunoconjugates within a shorter period of time while preventing the risk of affecting the antibody’s scaffold under possibly harsh conditions needed for direct radiolabeling strategies [153, 302].

3 Peptide-Receptor Radionuclide Imaging

Since receptors for regulatory peptides are overexpressed in a variety of human cancers, it is a prominent strategy to use radiolabeled analogs of these physiologically occurring peptides for tumor-targeted nuclear imaging. Advantages of using peptides are their good tissue penetration, a fast clearance, and minimal immunogenicity [247]. Small peptides of usually less than 40 amino acid residues are easily accessible through solid-phase peptide synthesis. Their tolerance toward bulky modification and resistance toward harsh chemical conditions that are sometimes inevitable during radiolabeling procedures are further advantages. Importantly, a formulation of a radiolabeled peptide consists of identical molecules with a well-defined structure. Clearly, the most outstanding example of success in the field of peptide-based diagnostic and therapeutic nuclear medicine has been the use of somatostatin analogs for targeting the somatostatin receptor [144]. Somatostatin-derived tracers designed to image somatostatin receptor subtype 2 (sst2)-expressing tumors have enjoyed almost two decades of successful preclinical development and extensive clinical application. This example has paved the path for further exploration of radiolabeled peptides targeting other tumor-associated receptors such as gastrin-releasing peptide receptors, neurotensin receptors, or cholecystokinin receptors [29, 225].

3.1 Somatostatin Analogs

The prototypes of radiolabeled peptides for SPECT imaging are the somatostatin analogs commonly labeled with ¹¹¹In or ^99mTc. Somatostatin receptors are overexpressed on neuroendocrine tumors including pituitary adenomas, pheochromocytomas, paragangliomas, neuroblastomas, and medullary thyroid cancers. From the five subtypes of somatostatin receptors belonging to the G-protein coupled receptors, subtype 2 is the most widely overexpressed form in neuroendocrine tumors. In the beginning of their development, somatostatin analogs suffered from rapid degradation in vivo. Such limitations have been overcome by stabilization strategies through the development of synthetic peptides. Peptides of high chemical stability became accessible by introduction of D-amino acids or other unnatural amino acids at known cleavage sites, cyclization, or modification of C- and N-termini via amidation, reduction, alkylation, or acylation [247]. The clinically approved ¹¹¹In-labeled DTPA⁰-octreotide (OctreoScan) has proven to be a successful and versatile molecular imaging agent (Figs. 7.4 and 7.5). The most frequently used DOTA-coupled, somatostatin-based peptides are [DOTA⁰, Tyr³]-octreotide and [DOTA⁰, Tyr³, Thr⁸]-octreotate usually referred as DOTATOC and DOTATATE (Fig. 7.4). These analogs have also been successfully employed for therapeutic purposes when radiolabeled with particle-emitting radioiosopes (e.g., ¹⁷⁷Lu, ⁹⁰Y). Several sst2-binding somatostatin analogs are currently used in the clinic. Further research projects are focusing on the development of new and improved somatostatin analogs with a broader receptor subtype affinity profile. Such compounds would extend the range of targeted cancer candidates and increase the net tumor uptake when several receptor subtypes are expressed on the same tumor cell [112, 184].

Fig. 7.4

Chemical structures of DTPA- and DOTA-modified somatostatin analogs for targeted diagnosis and therapy of somatostatin receptor-positive cancer diseases. a DTPA⁰-octreotide, b DOTA⁰-Tyr³-octreotide (DOTATOC), c DOTA⁰-Tyr³-Thr⁸-octreotate (DOTATATE), d [^99mTc]Tc-EDDA-HYNIC-TOC (^99mTc-Tektrotyd)

Fig. 7.5

Scintigraphy [multiple planar spot views, anterior a and posterior b of the head/neck (1), thorax (2), and abdomen (3)] performed 24 h after injection of ¹¹¹In-octreotide (OctreoScan; Covidien, Petten, the Netherlands) in a patient diagnosed with a well-differentiated endocrine carcinoma (carcinoid) with lymphogenic spread in the abdomen and supraclavicular and multiple metastatic lesions in the liver and lungs. The images have been kindly provided by J.J.M. Teunissen (MD, Ph.D.), Erasmus Medical Center, Rotterdam, The Netherlands

Although [¹¹¹In]In-DTPA-(D-Phe¹)-octreotide proved to be reliable for the detection of neuroendocrine tumors (NET), the potential clinical advantage of ^99mTc-labeling in comparison to radiolabeling with ¹¹¹In led to the development of ^99mTc-labeled somatostatin analogs. [^99mTc]Tc-EDDA-HYNIC-TOC (^99mTc-Tektrotyd) is a radiopharmaceutical indicated for diagnosis of tumors with overexpression of SSTR, especially subtype 2 (SSTR2) [15, 90, 143, 285].

The complex formation requires the use of a coligand such as tricine or ethylenediamine diacetic acid (EDDA). The coordination mode of HYNIC with ^99mTc is not known exactly. Usually, a monodentate “end-on”-N=N- coordination is proposed (Fig. 7.4).

The generally high kidney uptake of radiometallated peptides due to their reabsorption in the renal proximal tubules is a drawback for peptide-based tumor targeting as it may lead to reduced contrast and quality of diagnostic imaging and damage radiosensitive kidneys if applied for therapeutic purposes [104]. Thus, several strategies to reduce tubular reabsorption of peptidic radiotracers have been investigated. One strategy relies on the chemical modification of the peptide with entities or overall charges that would potentially reduce renal uptake. A successful example of such modification is given by the work of Schwaiger and co-workers who developed ¹²⁵I-somatostatin analogs modified with carbohydrate entities [248, 305]. Glycation modified the physicochemical behavior of the radiotracers in that pharmacokinetics were significantly improved as shown by reduced hepatic uptake and biliary excretion and a rapid clearance from the circulation via the kidneys without increasing renal accumulation of radioactivity. Another approach is based on the administration of additional substances for potential inhibition of peptide reabsorption. In this respect the co-infusion of the cationic amino acids lysine and arginine is the most prominent example since this combination successfully reduced renal accumulation of radiolabeled somatostatin analogs in preclinical studies [31, 61, 294] and in patients [111, 229].

Originally, it was proposed that peptide agonists that are efficiently internalized into receptor-expressing cancer cells would be the best candidates for tumor imaging [48]. However, the two somatostatin analogs ¹¹¹In-DOTA-sst2-ANT and ¹¹¹In-DOTA-sst3-ODN-8 showed extremely high tumor accumulation despite being receptor antagonists [100]. It could be shown in vitro that a more than 15-fold increased number of binding sites per cell were accessible for antagonists compared to their agonist analogs and in addition slow ligand dissociation from the receptor was determined. These findings attracted the attention of many research groups and led to the development of further sst2-binding somatostatin-based antagonists. The studies confirmed that high-affinity somatostatin receptor antagonists that poorly internalize in tumor cells exhibit improved tumor-targeting characteristics than corresponding agonists. The fact that this phenomenon was found not only for sst2-selective compounds but also for sst3-selective compounds suggests that this phenomenon is valid for more than just one particular receptor [47].

3.2 Bombesin Analogs

The bombesin receptor family comprises four receptor subtypes whereof the gastrin-releasing peptide (GRP) receptor or bombesin receptor subtype 2 (BB2) has been studied most thoroughly [268, 269]. The impetus for targeting the GRP receptor is based on the fact that a variety of human tumors overexpress GRP receptors including prostate, breast, and small cell lung cancers [106, 177, 187, 256]. The development of ^99mTc- and ¹¹¹In-bombesin analogs has been the focus in recent years in many research laboratories (Table 7.2) [20, 206, 288]. De Barros and colleagues presented a BBN-kit using HYNIC and EDDA as chelating source for ^99mTc [59]. A detailed summary was published by Moreno et al. in 2016 [188]. The tricarbonyl technique, which was developed in view of the opportunity to use ^99mTc and ¹⁸⁸Re as matched pair for diagnosis and therapy [6, 242], has been employed most extensively for radiolabeling of bombesin analogs [85, 86, 91,92,93, 145, 226, 230, 251, 266, 267, 317]. A drawback of this strategy is, however, the fact that the most ^99mTc/¹⁸⁸Re-tricarbonyl-based bombesin derivatives are predominantly cleared via the hepatobiliary excretion pathway because of the tricarbonyl’s inherent lipophilicity [64]. Increasing the hydrophilicity of radiolabeled GRP-targeting peptide conjugates is necessary because accumulation of radioactivity in the liver and intestinal tract would compromise their capacity to effectively image solid tumors and metastatic lesions in the abdomen. This has been accomplished, for example, by introduction of “innocent” peptide sequences such as polylysine, polyglycine, or polyaspartic acid residues [159]. Additionally, it was shown that the introduction of a polar serylserylserine spacer into ^99mTc-tricarbonyl pyrazolyl bombesin analogs resulted in a longer retention time of the radiotracer in the tumor tissue compared to analogs with more lipophilic linker entities consisting of β-alanine or triglycine [12]. Based on the promising results experienced with somatostatin analogs conjugated to carbohydrates [230, 248, 305], glycation of bombesin tracers was approached with the aim to increase their overall hydrophilicity [251]. In this respect Garcia et al. tested three different bombesin analogs in vitro and in vivo. One of the derivatives was modified with a linker bearing a lysine that was coupled to the glycomimetic shikimic acid at the ε-amino group. Another bombesin derivative was glycated via an Amadori rearrangement and the third compound was a bombesin analog derivatized with an azido-glucose that was connected to an alkyne-functionalized linker entity via the Cu-catalyzed click reaction (Table 7.2). The introduction of polar carbohydrates had no negative effects on the in vitro stability and the internalization or efflux profile of the radiotracers in cultured tumor cells. In contrast, these modifications led to a significant reduction in abdominal radiotracer accumulation, a clearly higher tumor uptake, and thus improved tumor-to-background ratios in vivo. The best results were obtained with the bombesin analog modified via a “click” reaction that contained a triazole-coupled glucose entity. The tissue distribution could be clearly ameliorated as demonstrated via SPECT/CT imaging studies where the tumor uptake was shown to be increased (Fig. 7.6).

Table 7.2 Sequence of various bombesin analogs

Fig. 7.6

SPECT/CT images of PC-3 tumor-bearing mice 1.5 h after injection of a ^99mTc(CO)₃-(N^αHis)Ac-βAla-βAla-[Cha¹³,Nle¹⁴]BBS(7–14)-NH₂ (control compound) and b ^99mTc(CO)₃-(N^αHis)Ac-Ala(^NTG)-βAla-βAla-[Cha¹³,Nle¹⁴]BBS(7–14)-NH₂, (^NTG = N-linked triazole-linked glucose). T = tumor, L = liver, I = intestines [251]

On the other hand accumulation in the liver was significantly reduced. Despite the higher kidney uptake found for the carbohydrated bombesin analogs at early time points after injection, this decreased rapidly with time indicating that the radiotracers were not trapped in the renal tissues. By this example, the strategy of radiotracer glycation has been demonstrated as a potent method to increase the overall hydrophilicity of a tracer and thus to improve the tissue distribution.

Based on the advantages of using trivalent radiometals for preparation of site-directed diagnostic/therapeutic radiopharmaceuticals [98, 268], interest has been sparked into the synthesis and biological evaluation of trivalent radiometalated bombesin derivatives using radioisotopes such as ¹¹¹In or ¹⁷⁷Lu (Table 7.2) [37, 62, 121, 127, 265]. One such example is the bombesin analog referred to as DOTA-AMBA useful for both diagnostic and therapeutic purposes [119, 146, 168]. Also, a so-called pan-bombesin analog has been designed with the special characteristic of displaying high affinity to all three bombesin receptor subtypes possibly allowing a broader field of application [320].

The majority of research efforts into the design of bombesin-based radiotracers have been performed by using GRP receptor agonists. Such bombesin analogs undergo receptor-mediated endoctytosis enabling residualization of the attached radiometal within the targeted tumor cell. However, ^99mTc-demobesin-1 is a potent antagonist, which clearly exhibited high affinity to the GRP receptor even though significant internalization into PC-3 prostate tumor cells was not observed. This radiotracer allowed imaging of PC-3 tumors in mice with higher tumor-to-background contrast compared to the best available agonist analog [178, 205]. Furthermore, an improved quantification of the beta cell mass after pancreas visualization was accomplished with ^99mTc-demobesin-4 combined with a beta cell imaging with ¹¹¹In-exendin-3 in rodents [289]. Thus, endeavors were directed also toward the development of bombesin antagonists. Recently, superior imaging properties of the ¹¹¹In-radiolabeled bombesin antagonist RM1 over the agonist ¹¹¹In-DOTA-AMBA have been demonstrated [172]. Whether or not bombesin antagonists are also favorable over agonists for therapeutic purposes remains to be investigated.

Dual receptor-targeted probes based on BBN and RGD-peptides combining an integrin α_vβ₃ and GRPR-targeting peptide were developed, e.g., for breast cancer imaging as alternative to ultra sound [49, 125, 126]. They found that this imaging technic is a useful alternative to US but cannot replace US. Furthermore, a ^99mTc-labeled RGD-BBN peptide was used to image lung carcinoma (Lewis lung carcinoma (LLC), U87MG human glioma, and PC-3 human prostate cancer cells) in a small-animal model. It was possible to detect subcutaneous and pulmonary metastatic Lewis lung carcinomas and to distinguish tumor from inflammation using ^99mTc-RGD-BBN [164]. Other attempts dealt with the connection of shepherdin (79–87) to BBN as an inhibitor of the survivin-Hsp90 interaction and Hsp90 ATPase activity [86]. To use the therapeutic impact, ^99mTc has to be located at the nucleus of the cells. For this purpose, BBN was connected to the TAT (49–57) peptide and radiolabeled with ^99mTc with N₂S₂ as chelating moiety [237]. Cell binding and proliferation were tested showing an efficient internalization to PC-3, MDA-MB231, and MCF7 cells with a significant reduction in the cell proliferation.

3.3 Neurotensin Analogs

Neurotensin (NT) is a linear tridecapeptide that can be found in the central nervous system and in peripheral tissues. Among the three NT receptors (NTR), NTR1 has been found in several neuroendocrine tumor types. Of special interest are exocrine pancreatic carcinomas that overexpress NTR1 with an incidence of 75–88% [224]. Thus, several studies focused on the development of NT analogs for radiolabeling with SPECT radionuclides such as ^99mTc [94, 95, 170] and ¹¹¹In [11, 63]. Similar to other small neuropeptides, neurotensin is rapidly metabolized in plasma by endogenous peptidases. Thus, neurotensin analogs which are stabilized at one or more of the three potential cleavage sites were developed. In this respect, the research group of Maina and Nock developed several ^99mTc(V)-neurotensin analogs, referred to as ^99mTc-demotensin, employing amino acid substitutions and/or reduction in the amide bond Arg⁸/Lys⁸-Arg⁹ to the corresponding amine [170, 207]. Garcia and co-workers reported the biological evaluation of neurotensin analogs in which two of the three cleavage sites have been stabilized [39, 95, 169]. These interventions allowed the preparation of neurotensin analogs of high plasma stability, affinity to the NTR1 in the nanomolar range, and significant tumor uptake in preclinical and clinical studies. A promising candidate is the ^99mTc radiolabeled peptide (N^α-His)Ac-Arg-(N-CH₃)-Arg-Pro-Tyr-Tle-Leu (^99mTc-NT-XII), which has been stabilized at the cleavage sites 8–9 and 11–12. Other than in the case of bombesin derivatives (see Sect. 7.3.2), the introduction of a glycomimetic entity (shikimic acid) coupled to the side chain of an additional lysine residue did not result in an improved tissue distribution of the radiotracer. Although the expected lower kidney and liver uptake could be achieved, both the receptor affinity and the tumor uptake were unfavorably reduced. Recently, the group of Garcia reported the evaluation of a ^99mTc(CO)₃-neurotensin analog, ^99mTc-NT-XIX, modified at all three cleavage sites [94, 95]. Despite a slight decrease in receptor affinity and a lower rate of internalization, the in vitro and in vivo stability of this novel radiopeptide has been significantly increased (Table 7.3).

Table 7.3 Stability and affinity of different radiolabeled NT analogs [94]

This example of a triple-stabilized neurotensin analog demonstrates the importance of the radiotracer’s metabolic stability to increase its accumulation in the tumor tissue which was—in the case of ^99mTc-NT-XIX—even able to compensate a slightly lower receptor-binding affinity. The clearly improved tumor-to-background contrast of ^99mTc-NT-XIX over ^99mTc-NT-XII could be visualized by SPECT/CT imaging (Fig. 7.7). Thus, the development of neurotensin ^99mTc-radiotracers, where single-amino acids have been substituted for peptide stabilization, is an example for optimization of a radiotracer’s tissue distribution by increasing its in vivo stability.

Fig. 7.7

SPECT/CT images of HT-29 tumor-bearing mice 1.5 h after injection of a ^99mTc-NT-XII and b ^99mTc-NT-XIX. T = tumor, L = liver, I = intestines [94]

3.4 Other Peptide-Based Radiotracers

Beyond somatostatin and GRP receptor targeting with bombesin and neurotensin analogs, many other regulatory peptide receptors are overexpressed on a variety of tumor types. Thus, peptide analogs in various stages of preclinical or clinical development include derivatives of cholecystokinin-2 (CCK-2) [60], glucagon-like peptide-1 (GLP-1) [138], neuropeptide Y (NPY) [322], and Arg-Gly-Asp (RGD) peptides [246] among others. CCK-2 receptors are expressed in medullary thyroid cancers. Initial gastrin-ligands for CCK-2 receptor targeting comprising a DTPA-DGlu-chelator showed unfavorable tumor-to-kidney ratios of radioactivity accumulation and were therefore not developed further. New gastrin derivatives lacking the glutamate-moiety showed excellent CCK-2 receptor affinity and lower renal retention in a rat AR42J tumor model [103]. Recently, it was found that GLP-1 receptors are highly overexpressed in virtually all insulinomas and gastrinomas [138]. Metabolically more stable GLP-1 congeners referred to as exendin-3 and exendin-4 have been derivatized with a DTPA or DOTA chelating system for radiolabeling with ¹¹¹In or lanthanide radioisotopes. Remarkable tumor targeting was found in a human patient while employing ¹¹¹In-DOTA-exendin-4 [51]. NPY analogs are of interest because of the frequent overexpression of NPY receptors in a variety of tumor types including breast cancer. A recent article reports on the synthesis and evaluation of a large number of NPY analogs where a DOTA-derivatized compound radiolabeled with ¹¹¹In performed as a potent radiotracer [322]. However, the in vivo studies with this tracer showed only a low tumor uptake whereas radioactivity retention in the kidneys was extremely high. RGD peptides that do not belong to the group of regulatory peptides are of particular interest for targeting integrin receptors such as the α_νβ₃ integrin. This integrin subtype is strongly expressed on activated and proliferating endothelial cells during tumor angiogenesis and metastasis but is not readily detectable in resting endothelial cells and most normal organs. Thus, a variety of RGD-peptide analogs for targeting α_νβ₃ integrins have been developed and the promising potential of RGD-based radiotracers for SPECT radio imaging has been shown [246]. To enhance binding affinity for the α_νβ₃ integrin, various multivalent cyclic RGD-based peptides have been developed. All oligomeric peptide probes bound more strongly to the target cells than the monomeric RGD peptide in an integrin α_νβ₃-positive U87MG xenograft mouse model (Fig. 7.8) [260, 299].

Fig. 7.8

Chemical structure of DOTA-3PEG₄-RGD dimer [158, 259]

Through RGD peptides the advantage of multivalent tumor-targeting agents over monovalent agents has been demonstrated. Most likely, the employment of the multimer-strategy also improves tumor-targeting properties of non-RGD-based peptides. Accordingly, investigations of divalent and multivalent peptides are ongoing for targeting of many of the tumor-associated receptors mentioned above, among those imaging agents for targeting the CCK-2 receptor [271] and somatostatin receptor [312]. Also, the strategy of using dual tumor-targeting agents that combine targeting ligands for two different receptors (e.g., integrin and GRP receptor) might improve the radiotracer’s diagnostic utility and applicability [162, 163].

4 Antibodies and Antibody Fragments

Another approach of nuclear imaging is the use of radiolabeled antibodies that target-specific cell surface antigens. Radioimmunoimaging has been traditionally developed in parallel with radioimmunotherapy for the evaluation of the antibodies’ targeting properties and for dosimetry. Common tumor-associated targets for radioimmunoimaging (and -therapy) are epidermal growth factor receptors (EGFR) [281, 309], the carcinoembryonic antigen (CEA) [123], the prostate-specific membrane antigen (PSMA) [152], cluster of differentiation antigens (e.g., CD20), the pancarcinoma antigen (TAG-72), and the HER2 receptor among others. In addition, a number of angiogenesis markers—protein antigens expressed either on blood vessels or in the adjacent matrix of vessels—have been characterized as targets for selective delivery of antibodies to the tumor neovasculature [36]. Examples are the fibronectin extra-domain B (EDB) [204], the integrin α_vβ₃ [217], the vascular endothelial growth factor (VEGF) [38], and annexin A1 [209].

Potential concerns for radioimmunodiagnosis and strategies for optimization have been summarized in several review articles [41, 42, 293]. The main disadvantage of antibodies, namely their immunogenicity, could be largely overcome by the application of humanized antibodies that evade the immune system and are resistant to degradation. However, the slow vascular clearance (days to weeks) of antibodies as a consequence of their high-molecular-weight (IgG antibodies: ~150 kDa) and the low tissue penetration are generally disadvantageous for radioimaging because of the resulting low target-to-non-target contrast at early time points after administration. Although it is generally accepted that antibodies are not the preferred biomolecules for nuclear imaging, the application of antibody fragments for SPECT has been successfully exemplified. Similar to peptides, antibody fragments are rapidly cleared from the blood and from non-targeted tissues. The results thereof are higher tumor-to-background ratios compared with intact antibodies and a lower radiation absorbed dose in non-targeted tissues and organs. A reduced percentage of injected doses of radioactivity in the tumor tissue and higher radiation doses in the kidneys are also consequences of the reduced size of antibody fragments [42].

Efforts have been directed toward the development of antibody fragments such as F(ab′)₂, F(ab′) and single-chain Fv (scFv) fragments to achieve faster clearance from the blood and in addition a better tumor penetration [313, 314]. Application of high-affinity scFv resulted in a relatively high tumor uptake combined with a rapid blood clearance and hence favorable targeting ratios [27]. Multimers of antibody fragments may result in improved tumor localization compared with monomeric species as a result of higher affinity and slower blood clearance [134].

Another approach to achieve improved pharmacokinetics is the pretargeting strategy. Pretargeting involves an initial targeting agent, which itself can be bound by secondarily injected agents. Secondary agents are either quickly clearing radiotracers that bind the initial agent with high affinity [102, 151, 213, 249, 273] or “chase” reagents that clear an unbound radiolabeled antibody in circulation [135]. The pretargeting approach is, however, not commonly applied for SPECT. In contrast, this strategy is much more favorable for radioimmunotherapy in order to reduce the radioactive dose burden to the bone marrow and thus to avoid potential hemotoxicity of long circulating antibodies labeled with particle-emitting radioisotopes.

Radioimmunoimaging is of particular interest to evaluate a potential application of antibodies for targeted radionuclide therapy by interchanging a diagnostic with a therapeutic radioisotope of similar chemical characteristics (e.g., ¹¹¹In and ⁹⁰Y) or using a therapeutic radionuclide that emits concomitantly with therapeutic radiation also diagnostic γ-rays of a suitable energy for SPECT (e.g., ¹⁷⁷Lu, Table 7.1). The most prominent example of an antibody employed for radioimmunotherapy is ibritumomab tiuextan (Zevalin), a ⁹⁰Y-radiolabled monoclonal anti-CD20 antibody for the treatment of non-Hodgkins lymphoma. Its ¹¹¹In-radiolabeled counterpart is usually administrated prior to therapy for detection of receptor-positive malignant tissue via SPECT imaging and for dosimetry.

4.1 Targeting Fibronectin Extra-Domain B: Antiangiogenic Antibody Fragment L19

Angiogenesis is an underlying process in many human diseases, including cancer. An established target in this respect is the extra-domain B of fibronectin (EDB), a domain of 91 amino acids, which is typically inserted in fibronectin molecules at sites of tissue remodeling but not in fibronectin molecules under normal conditions. Thus, the expression of EDB has been shown in malignant tumors but not in healthy tissues [315]. The Neri group has isolated a number of human monoclonal antibodies to EDB [43, 204, 215]. The human antibody fragment, scFv(L19) displayed subnanomolar affinity to EDB and has been shown to efficiently localize on tumoral neovasculature in animal models [68]. Importantly, the ¹²³I-labeled dimeric L19 antibody fragment L19(scFv)₂ has been evaluated for targeting primary tumors and metastatic lesions in cancer patients through immunoscintigraphy [236]. This clinical study was performed with 20 patients whereof the majority had colorectal or lung cancer. It could be demonstrated that the antibody ¹²³I-L19(scFv)₂ selectively accumulated in malignancies and allowed distinguishing among actively growing and quiescent lesions. Another Phase I/II clinical immunoscintigraphy study used ¹²³I-L19(scFv)₂ in patients with head and neck squamous cell carcinoma [32]. It was observed that for head and neck scintigraphy, iodinated antibodies have severe disadvantages. Although the thyroid gland was protected by competitive application of non-radiactive iodide, there were substantial artifacts in this area in all cases as a result of the uptake of liberated free iodide that was always present to a certain degree. Since dehalogenases are present in the salivary glands, free iodide also gave a high background in the 4 h postinjection scintigraphy in the parotid and submandibular glands as well as in the minor salivary glands of the oral and nasal mucosa. Although the ¹²³I-L19(scFv)₂ is probably less suited as a diagnostic imaging modality for head and neck cancer, L19(scFv)₂ offers a general potential to be used as a tumor-targeting agent for both diagnostic and therapeutic purposes. Because neovasculature and tissue remodeling are required for the growth of all aggressive solid tumors, imaging approaches that use angiogenesis markers can be used for different types of cancer. An advantage of this strategy might be the fact that noninvasive imaging of angiogenesis via EDB fibronectin targeting allows the discrimination between quiescent and actively growing lesions.

5 Vitamin-Based Radiotracers

The use of small-molecular-weight targeting compounts is favorable to surmount the drawbacks of long circulation times and thus poor tumor-to-background contrast as well as possible immunogenicity encountered with antibodies. In this respect the application of vitamins as targeting agents provides several advantages: vitamins are small in size, inexpensive, relatively easily amenable for chemical modification, and non-immunogenic. Rapidly dividing cancer cells have an increased demand for certain vitamins such as folates, vitamin B12 (cobalamin), biotin, and riboflavin. These B-group vitamins are required for cell survival and proliferation because they act as co-enzymes of biochemical reactions that are essential for the synthesis of amino acids and for nucleotide bases [233]. The most thoroughly investigated vitamin to be used as tumor-targeting agent is folic acid. The utility of folic acid conjugates has been widely exemplified in a variety of (pre)clinical studies for targeting the folate receptor (FR) that is overexpressed on a wide variety of cancer types [165]. Also, it has been demonstrated that vitamin B₁₂ has the potential to be used as cancer-targeting agent whereas only few studies have focused on the applicability of biotin for direct tumor targeting [233]. Since vitamins are indispensable for sustaining life, it is unlikely that a mutational arrest of vitamin uptake would occur with concomitant failure of vitamin-mediated diagnosis or therapy. This is a distinct feature of vitamins and an advantage for their application as tumor-targeting agents. Thus, using vitamin-based imaging agents is attractive and the strategy holds promise to also be used for therapeutic purposes.

5.1 Folic Acid Conjugates

Folic acid and folates (reduced forms) are water-soluble vitamins of the B-complex group. Humans cannot synthesize folates and hence must necessarily obtain them from food. Although only small quantities of folates are required, these vitamins are vital for various biochemical reactions including those for the synthesis of RNA and DNA, amino acid metabolism, and gene regulation. Cellular uptake of folates is accomplished by either carrier systems or the high-affinity folate receptor (FR). The FR is a glucosylphosphatidylinositol (GPI)-anchored protein that is frequently overexpressed in a variety of tumor types including cancers of the breast, ovaries, cervix, endometrium, lungs, kidneys, colon, and brain [13, 212]. In normal organs and tissues, FR-expression is highly restricted to only a few sites where it is located on the apical side of polarized epithelia in the lung, the placenta, and the choroid plexus of the brain and in the proximal tubule cells of the kidneys [13, 212, 304]. Thus, folic acid can be used as a molecular “Trojan horse” for selective delivery of attached probes to FR-positive cancer cells [165]. During the last decades, a variety of folic acid conjugates of radioisotopes useful not only for SPECT imaging (^99mTc, ¹¹¹In, ⁶⁷Ga) has been developed and evaluated (Fig. 7.9) [16, 79, 130, 131, 196,197,198, 264]. Biodistribution studies of radiofolates in mice showed a specific uptake in FR-positive tumor (xeno)grafts, whereas unspecific radioactivity in background tissues was rapidly cleared in particular if the derivatives displayed hydrophilic properties. In the kidneys, however, high radioactivity retention was observed as a consequence of the specific binding of radiofolates to FRs expressed in the proximal tubule cells. This process results in unfavorably low tumor-to-kidney ratios of radiofolates in general. Clinical application of the two most promising candidates, ¹¹¹In-DTPA-folate [179, 261, 300] and ^99mTc-EC20 [87, 149, 223], also known as ^99mTc-etarfolatide or FolateScan, revealed the same phenomenon in humans that was previously found in tumor-bearing mice [181]. ^99mTc-EC20 is currently undergoing Phase 1 and 2 clinical trials in several institutions within the US and Europe. Mostly, renal cell carcinomas, pituitary adenomas, ovarian, and breast carcinomas were investigated. While imaging of malignant tissue could be successfully achieved, high radioactivity uptake was found in the kidneys of patients where the FR is expressed to approximately the same level as in mouse kidneys [212]. Healthy volunteers were enrolled in another phase I clinical study to assess the pharmacokinetics in terms of safety and radiation dosimetry. The activity of ^99mTc-EC20 at 5 min postinjection was largest in the bone marrow, followed by the liver and kidneys, and decreased in all organs/tissues within 1 day without appreciable retention [310]. In a phase II multicenter study, 43 patients with advanced ovarian cancer were imaged with ^99mTc-EC20 for lesion detection before treatment with vintafolide [189]. In a 2009-reported clinical study, ^99mTc-EC20 was used to determine inflammatory diseases like rheumatoid arthritis in comparison with healthy humans. It could be illustrated that imaging with ^99mTc-EC20 is more sensitive for these diseases compared to physical examinations [180].

Fig. 7.9

Chemical Structures of the vitamin folic acid. a EC20 (M = ^99mTc). b His-folate (M = ^99mTc, ¹⁸⁸Re). c DOTA-folate (M = ¹¹¹In, ¹⁷⁷Lu). d ^99mTc-HYNIC-folate. e DOTA-folate with albumin-binding entity (M = ¹⁶¹Tb, ¹⁷⁷Lu) f and ¹²⁵I-folate (g)

In an attempt to improve the low tumor-to-kidney ratio of radiofolates, it was hypothesized that application of antifolates (e.g., pemetrexed) could increase the “appetite” of the tumor cells for folates and thus lead to an increased accumulation of folic acid conjugates. This hypothesis was confirmed in vitro [190]. However, in mice that were treated with antifolates, radiofolate uptake in tumor xenografts was not increased. While approaching this hypothesis, injection of pemetrexed was accomplished at different time points prior to the radiotracer. None of the experiments revealed an increased tumor accumulation of radioactivity, however, surprisingly administration of pemetrexed short before the radiofolate resulted in a significant reduction in kidney uptake [190]. The result was a tremendous increase in the tumor-to-kidney ratio of radioactivity. This effect could be reproduced with a variety of folic acid conjugates radiolabeled with various radionuclides (^99mTc, ¹⁸⁸Re, ¹¹¹In, ¹²⁵I, ¹⁴⁶Tb, ¹⁷⁷Lu) and in mouse models bearing different tumor (xeno)grafts (KB, IGROV-1, SKOV-3; M109) [191,192,193, 195, 199,200,201, 222]. The clearly superior SPECT imaging quality of mice that received pre-dosed pemetrexed could be impressively demonstrated while using ¹¹¹In-radiolabeled DTPA-folate (Fig. 7.10). This example demonstrates a pharmacological intervention by a non-radioactive substance that results in an improved tissue distribution of the radiotracer compared to the results obtained after radiotracer administration alone.

Fig. 7.10

SPECT/CT of mice injected with a ¹¹¹In-DTPA-folate only and b in combination with pre-dosed pemetrexed

Another approach to reduce the kidney uptake deals with the modification of the backbone of the folate conjugate. For this purpose, an albumin-binding entity was connected to enhance the circulaton time of the radioconjugate in the blood associated with an increase in the tumor-to-kidney ratio. In this regard, other radionuclides for imaging (⁴⁴Sc, ⁶⁴Cu, and ⁶⁸Ga for PET) and therapy (¹⁷⁷Lu, ⁴⁷Sc) were applied to survey the therapeutic eligibility [78, 197, 200, 264]. Other approaches dealing with the combination of bombesin (1–14) with a DOTA-FA as theragnostic agent labeled with ¹⁷⁷Lu for breast cancer showed a higher tumor uptake as the DOTA-FA alone in T47D-tumor-bearing mice combined with a high renal clearance [14]. HYNIC is known to bind Tc efficiently. Thus, approaches with FA-HYNIC conjugates [167] using a click chemistry approach for the linkage of the Tc-core and FA [108] and a multimerization of FA [109] were presented. However, the best tumor-to-organ ratio was found with the PEGylated monomer using KB tumor-bearing mice. Nanocarriers like PAMAM were extensively studied as well. A PAMAM-DTPA-conjugate was presented showing a high tumor accumulation (13.34%ID) combined with an uptake in the liver (9.48%ID) and the heart (6.88%ID/g) after 3 h pi in KB-bearing mice [321]. The PAMAM-HYNIC-conjugate showed a substantial accumulation in the tumor (10.61%ID/g), but the uptake in liver with 69.34%ID/g and spleen with 14.43%ID/g combined with a blood content of 7.68%ID/g was high after 3 h pi in BALB/c mice with 4T1-breast cancer [203]. Other nanocarriers based on poly(ethylene glycol)-poly(lactic-co-glycolic acid) were used with a particle size of 104–128 nm mean showing a high tumor uptake (21.3%ID/g), but also a high uptake in liver, kidneys, spleen, and blood of > 13%ID after 3 h pi in SKOV-3-bearing tumor mice. Efforts to develop nanoprobes as ^99mTc-imaging agents were made by the use of nanographene oxide (nGO). Thus, nGO was modified with PEG and further functionalized with FA ready for labeling with ^99mTcO₄⁻ and SnCl₂. A biodistribution study using mice demonstrated a long residence time in the blood; a high accumulation in liver, spleen, lung, and kidneys; and a comparatively low uptake in the tumor [77].

Humanized IgG1 antibodies like farletuzumab that specifically recognizes the folate receptor alpha (FRα) were radiolabeled with ¹¹¹In and additionally with a fluorescent dye to use this radiopharmaceutical for the intraoperative detection of ovarian cancer lesions. ¹¹¹In-farletuzumab-IRDye800CW showed optimal tumor-to-blood ratios of 3.4–3.7 at protein doses up to 30 μg in mice with intraperitoneal IGROV-1 tumors and can be blocked by coinjection of an excess of unlabeled farletuzumab [113].

5.2 Vitamin B₁₂ Conjugates

The earliest studies of radiolabeled vitamin B₁₂ (cobalamin) using cobalt radioisotopes (⁵⁷Co, ⁵⁸Co, ⁶⁰Co) showed radioactivity accumulation in peripheral, actively growing tumors with highest accumulation in sarcomas [33, 88, 89]. Other studies used radioiodinated arylstannylcobalamin conjugates showing enhanced uptake into renal carcinomas in nude mice when compared with other healthy tissues and organs [307]. Collins et al. developed ¹¹¹In-DTPA-analogs of cobalamin (DTPA cobalamin analogs = DACs) and tested them in preliminary biodistribution experiments in mice with CCL8 sarcomas and in pigs [53]. The overall biodistribution of DACs showed tumor uptake and high radioactivity accumulation in healthy organs that were almost identical to previous studies performed with ^57/60Co-radiolabeled vitamin B₁₂. The same group reported the first patient study performed with ¹¹¹In-DTPA-adenosylcobalamin for cancer imaging [54]. ¹¹¹In-DTPA-adenosylcobalamin was found to be effective for detection of high-grade aggressive tumors in humans with the most successful results in patients with breast cancer and high-grade lung, colon, thyroid, and sarcomatous malignancies [55]. However, the most significant uptake of these cobalamin derivatives was found in the liver, kidneys, and spleen followed by radioactivity accumulation in several glands. Vitamin B₁₂ is bound to soluble transport proteins in circulation, namely, transcobalamin I (TCI), intrinsic factor (IF), and transcobalamin II (TCII) whereof the latter is the principle vitamin B₁₂ binding protein [252,253,254]. TCII-cobalamin binds to TCII-receptors that are ubiquitously expressed in cells for effective acquisition of this important vitamin. Originally, vitamin B₁₂-mediated tumor targeting was thought to be dependent on undisturbed interaction of cobalamin with these main transport systems and tumor uptake were believed to be mediated via up-regulated TCII-receptors [25, 233]. Later, it was hypothesized that selective TCII non-binders would lead to improved tissue distribution. Various cobalamin derivatives comprising a (pyridine-2-ylmethylamino)acetic acid (PAMA) chelator for coordination of the ^99mTc-tricarbonyl core were developed with different spacer lengths [C-2 to C-6, i.e., (-CH₂-)_n, n = 2–6]. ^99mTc(CO)₃-PAMA-cobalamin derivatives with a spacer length of C-5 or longer displayed TCII binding affinity whereas those with shorter spacer lengths (C-2 to C-4, Fig. 7.11) were identified as TCII non-binders, but displayed retained interaction with IF and TCI [298]. The results of biodistribution studies in tumor-bearing mice performed with ^99mTc(CO)₃-PAMA-C5-cobalamin and ^99mTc(CO)₃-PAMA-C6-cobalamin were similar to previously evaluated ¹¹¹In-DTPA-adenosylcobalamin tracers [298]. In contrast, data of ^99mTc(CO)₃-PAMA-cobalamin derivatives with spacer lengths shorter than C-5 showed a significantly improved tumor-to-blood and tumor-to-kidney ratio of radioactivity. Thus, abolished interaction of the radiolabeled cobalamin tracer with TCII resulted in decreased accumulation of the radiotracer in the blood and in organs and tissues that would otherwise be predestined to have high cobalamin uptake such as kidneys and diverse glands (Fig. 7.12).

Fig. 7.11

Chemical Structure of ^99mTc-PAMA-C4-cobalamin, a TCII non-binder vitamin B₁₂ derivative

Fig. 7.12

Whole-body SPECT/CT scans of B16F10 tumor-bearing mice, 24 h after injection of a ^99mTc(CO)₃-PAMA-C6-cobalamin (TCII binder) and b ^99mTc(CO)₃-PAMA-C4-cobalamin (TCII non-binder)

^99mTc(CO)₃-PAMA-C4-cobalamin (Fig. 7.11) was selected as the most favorable candidate because it displayed the highest tumor-to-blood and tumor-to-kidney ratios in animal experiments. These findings suggest that the transport of cobalamin derivatives into malignant tissue is not dependent on the transport protein TCII but rather mediated via TCI. By this example, it could be demonstrated that variation of the radiotracer’s linker length could have a tremendous impact on the overall tissue distribution of a radiotracer and thus, on its successful application. Excellent results achieved in preclinical studies paved the path toward a clinical application of cobalamin-targeted radioimaging in patients using the TCI-selective organometallic ^99mTc-vitamin B₁₂ derivative.

5.3 Other Vitamin Targeting Agents—Pretargeting

It is likely that carriers and receptors of vitamins other than folates and vitamin B₁₂ could be used for tumor-targeted nuclear imaging purposes. Among the vitamins of the B-group, it was suggested that cancer cells also overexpress a biotin receptor that could, however, not yet be identified [186, 233, 311]. Additionally, a possible reason for the generally little interest in biotin as a direct tumor-targeting agent could be the fact that renal filtration and reabsorption of biotin and its conjugates lead to high renal uptake of radioactivity in the kidneys. Recently, it was shown that vitamin C (ascorbate) conjugated nanoparticles could be delivered into the brain presumably via the sodium-dependent ascorbic acid transporter SVCT2 whose RNA was found in the choroid plexus epithelium [234]. The SVCT2 carrier was found on rat glioma cells (C6 and F98) and on mouse fibroblasts (NIH/3T3). This study introduced the perspective of using the SVCT2 transporter for brain targeting through the choroid plexus where it is selectively expressed. There might also be a potential to use this vitamin C transporter for nuclear imaging purposes of cancer diseases in the future.

Efforts have been done to use biotin as component in the (strept)avidin–biotin-pretargeting system as their strong affinities (K_d = 10⁻¹⁵ M) allow the in vivo radiolabeling of high-molecular-weight compounds like proteins, antibody fragments, or antibodies for imaging and radiotherapy [257]. The problems of slow distribution kinetics of these compounds in association with insufficient tumor-to-organ ratios combined with a slow blood clearance should be avoided with this system in contrast to the application of directly radiolabeled biomacromolecules. For this purpose, the biomacromolecule (e.g., antibody) was connected to the (strept)avidin and the radionuclide was connected to the biotin or in the opposite way (Fig. 7.13).

Fig. 7.13

Pretargeting concept using a radiolabeled biotin-conjugate and a (strept)avidin-functionalized antibody. Step 1: distribution of the functionalized antibody to the target (tumor cells); Step 2: clearance of the unbound antibody from the blood; Step 3: administration of the biotin-radioconjugate and binding to the antibody (remaining biotin-radioconjugate will be excreted rapidly)

In general, the pretargeting system consists of three steps [35]. First, a non-radiolabeled but avidin-functionalized antibody was administered intravenously allowing a slow distribution and the binding to the target site over 2 to 3 days in vivo. Afterward in the second step, a clearing agent was given that helps to remove the remaining unbound antibody from the body whereas the antibody reaches the highest uptake in the targeted tissue. In the third step, a radiolabeled conjugate was administered, which consists of the biotin bearing the radiolabel. This radioconjugate binds with a high affinity to the (strept)avidin-functionalized antibody given previously and exhibits fast distribution properties in vivo. Due to the fast blood clearance of the biotin-radioconjugate through the kidneys, a high tumor-to-background ration will be achieved combined with a protection of the liver. Importantly, the antibody itself has to be located at the surface of the cell during the time of the radioconjugate injection. It should not be internalized.

The first approach to use the avidin–biotin-system for radiolabeling purposes was accomplished by Hnatowich and co-workers [118]. A biotin-conjugated antibody and a DTPA-coupled avidin were used and the radiolabeling was done with ¹¹¹In. They showed that the target/non-target radioactivity ratios were significantly improved with respect to the conventional radiolabeling procedures. Later, Rosebrough discovered the pharmacokinetics and biodistribution of radiolabeled avidin, streptavidin, and biotin [231] using rabbits and dogs. He found that both ¹¹¹In-DTPA-biotin and ¹¹¹In-DTPA-biotin–avidin have a high excretion rate (<5% circulation in the blood after 1 h) in contrast to the ¹¹¹In-labeled DTPA-biotin–streptavidin-conjugate (>30% after 6 h). In this regard, ~80% of the dose of ¹¹¹In-DTPA-biotin–avidin was found in the liver after 6 h. In contrast, ¹¹¹In-DTPA-biotin–streptavidin had only ~5% liver accumulation 6 h after injection. They and others stated that the blood clearance of the radiolabeled streptavidin and avidin differed markedly due to the difference in net charge exhibited at physiological pH [124]. Based on these facts, attempts were done to decrease the uptake of streptavidin in the liver [110, 308].

A multitude of biotin–chelator-conjugates not only for ^99mTc [136, 148] but also for ¹¹¹In [18] and ⁶⁸Ga [280] or dual modality imaging probes [70] as well as for pretargeted radioimmunotherapy with ⁹⁰Y, ⁶⁷Ga, ¹⁷⁷Lu, and other therapeutic radionuclides were developed [155, 183]. Additionally, biotin conjugates for biorthogonal click chemistry were designed (see, e.g., [139, 218]); examples are shown in Fig. 7.14. A phase II study with 25 patients bearing a metastatic colon cancer using ⁹⁰Y-DOTA-biotin-conjugate and a NR-LU-10 antibody/streptavidin-conjugate showed relatively disappointing results in terms of therapeutic efficacy and toxicity, but beneficial information was obtained concerning normal tissue tolerance to low-dose-rate irradiation [140]. A deeper insight to the pretargeting concept in general is published by Liu [156].

Fig. 7.14

Selected radioconjugates with biotin skeleton containing chelating systems based on macrocycles (a) and (b); for ^99mTc-labeling (c) and (d); for bioorthogonal click reactions (e) and (f)

6 Intracellular Targets

6.1 ^99mTc-Carbohydrate Complexes

The most frequently used radiotracer for nuclear imaging purposes is currently the glucose analog 2-[¹⁸F]fluorodeoxy glucose ([¹⁸F]FDG). This PET tracer is taken up by tumor cells mainly by facile diffusion through the glucose transport protein 1 (Glut1). In the cell interior [¹⁸F]FDG is phosphorylated by the enzyme hexokinase yielding [¹⁸F]FDG-6-phophate which cannot escape the cell anymore. Thus, this trapping mechanism results in accumulation of radioactivity in metabolically active (cancer) cells [270]. The clinical relevance of [¹⁸F]FDG promoted the development of inexpensive and readily available ^99mTc-labeled glucose analogs.

Most of the derivatives reported in the literature were ^99mTc(V)-glucose complexes [56, 58, 154, 210, 227, 282, 301, 319]. However, these ^99mTc-tracers did not match the criteria and features of [¹⁸F]FDG, such as active transport via Glut1 and phosphorylation via hexokinase. Later, Dapueto and colleagues could show a higher uptake of ^99mTc-IDAG and ^99mTc-AADG into the tumor of melanoma bearing C57BL/6 mice with tumor-to-muscle ratios of 12.1 ± 3.73 and 2.88 ± 1.40 [57].

Endeavors have been undertaken by the group of Schibli and others to design organometallic glucose and glucosamine analogs using the matched pair ^99mTc/^186/188Re [26, 71, 214, 239]. Later, Lin et al., Zeltchan et al. and Khan et al. reported new ^99mTc-labeled glucose derivatives bearing a ^99mTc(CO)₃ core [132, 318]. Both tracers were subjected to cell studies either with Chinese hamster ovary cells CHO, human breast adenocarcinoma MCF-7 cells, and murine sarcoma S180 cells. Following animal studies revealed a substantial uptake into the tumor tissue, but also in excretion organs like liver and kidneys. In addition, the ^99mTc-thioglucose derivative prepared by Zeltchan et al. accumulates in the brain.

Biological characterization has been reported from a variety of organometallic ^99mTc(CO)₃-glucose complexes, derivatized at the C-1, C-2, C-3, and C-6 positions with various chelating systems. These compounds were tested for their ability to be internalized into Glut1 expressing cancer cells, HT29, and in addition, it was investigated on whether or not they would be phosphorylated via the hexokinase reaction. Unfortunately, all of the complexes tested appeared not to be recognized and transported via Glut1. The authors stated the likeliness of ^99mTc(CO)₃-glucose complexes being sterically too demanding for recognition at the extracellular binding site and/or transportation via Glut1. Also, other than [¹⁸F]FDG, the organometallic glucose derivatives were not phosphorylated by hexokinase. Orvig and his collaborators reported several new approaches of organometallic carbohydrate complexes. Among others, they synthesized N-hydroxybenzylamino-deoxyglucose derivatives (Fig. 7.15) and carbohydrate-appended hydroxypyridinone derivatives [26, 80,81,82, 275]. However, most of these compounds revealed neither to be hexokinase substrates nor inhibitors. Although basic cell data of these carbohydrate radiometal complexes is lacking, it is likely that they are not taken up via the Glut1 transporter or other specific transport mechanisms and thus would fail to accumulate in cancer cells in vivo.

Fig. 7.15

[¹⁸F]FDG a two C-2 functionalized glucose derivatives of a N-(2′-hydroxybenzyl)-amino-chelating system [26] b and an imino diacetic acid chelator [239] c radiolabeled with ^99mTc-tricarbonyl (M = ^99mTc) and a ^99mTc-tracer based on deoxyglucose dithiocarbamate d [154]

6.2 Radiolabeled Nucleoside Analogs for Targeting Human Thymidine Kinase

In mammalian cells, salvage pathway phosphorylation of thymidine is catalyzed by two different thymidine kinases (TK): the cell-cycle regulated cytoplasmic TK1 and the constitutively expressed mitochondrial TK2. The human TK1 (hTK1) activity is known to fluctuate with cellular DNA synthesis, the activity being high in proliferating and malignant cells and low or absent in quiescent cells, whereas TK2 activity is low in both dividing and quiescent cells [202]. Since the activity of hTK1 is often dramatically increased in cancer cells, interest has been sparked in targeting this enzyme by radioactive thymidine analogs for selective imaging of proliferating cancer cells. In the cell interior nucleosides are rapidly phosphorylated to nucleotides, which renders them unable to penetrate biological membranes and thus they are “trapped” inside the cells. Thymidine and thymidine analogs labeled with PET radioisotopes such as [¹¹C]methyl-thymidine, 5-[⁷⁶Br]bromo-2′-fluoro-2′-deoxyuridine, and 3′-[¹⁸F]fluoro-3′-deoxythymidine ([¹⁸F]FLT) are either under development or already in use as proliferation marker [40, 96]. However, due to the high costs for the production of PET radioisotopes and the unfavorably short half-lives of PET isotopes, the use of SPECT radioisotopes ^99mTc or ¹¹¹In would be more advantageous. Schmid et al. focused on the preparation of radiometal labeled thymidine complexes functionalized at position N3 with a DO3A-chelator suitable for radiolabeling with ¹¹¹In or lanthanide radioisotopes [243]. However, cellular uptake of the thymidine metal complexes in DoHH2 and HL60 cells failed. Clearly, there is an interest to develop thymidine derivatives suitable for radiolabeling with ^99mTc. Celen et al. reported the preparation and evaluation of a ^99mTc(V)-MAMA-propyl-thymidine complex as a potential probe for in vivo visualization of tumor cell proliferation via SPECT [46]. However, this ligand could not be phosphorylated because it was too bulky. The group of Schibli focused on the development of thymidine analogs labeled with the organometallic ^99mTc-tricarbonyl-core (Fig. 7.16) [69, 276, 277]. The design of organometallic ^99mTc-derivatives could be favorable as these complexes were sterically less demanding than previously prepared thymidine radiometal complexes. Those organometallic thymidine derivatives were systematically evaluated regarding the influence of the spacer length between the thymidine and the chelating system, the overall charge of the complex after radiometal coordination and the uptake in human neuroblastoma SKNMC cells. From these studies, it was concluded that neutral and anionic complexes are more readily accepted as substrates than cationic complexes.

Fig. 7.16

Chemical structures of nucleoside-based SPECT tracers. a 5-[¹²⁵I]Iodo-2′-deoxyuridine ([¹²⁵I]IdUrd) [72, 255], b 5-[¹²⁵I]iodo-4′-thio-2′-deoxyuridine ([¹²⁵I]ITdU, R = H) and 5-[¹²⁵I]iodo-1-(4′-thio-β-arabinofuranosyl)-uracil ([¹²⁵I]ITAU, R = OH) [283] and c ^99mTc(CO)₃-thymidine derivatives (n = 2, 3, 5 or 10) [69]

Moreover, modeling experiments suggested that the flexibility of a longer spacer between the thymidine molecule and the organometallic core further improves the ability of the complex to be accommodated in the binding site of the enzyme. Cellular uptake was higher for complexes with log P values greater than one but still about 6-fold lower than for the ³H-thymidine control compound. Although some of the organometallic thymidine complexes were identified as enzyme substrates, the low and often almost absent permeability of the thymidine metal complexes through the cellular membrane remains a major hurdle for these compounds.

Iodinated thymidine analogs (e.g., 5-iodo-2′-deoxyuridine (IdUrd)) were used in another strategy as cell proliferation markers for nuclear imaging purposes and potential therapeutic application. However, the imaging quality was found to be impaired by the tracer’s rapid in vivo degradation. Pre-application of 5-fluoro-2′-deoxyuridine (FdUrd) was tested with the aim to block thymidine synthesis and thus trigger the tumor uptake of [¹²⁵I]IdUrd [72]. Indeed, as a result of FdUrd pre-dosing [¹²⁵I]IdUrd incorporation into glioblastoma cells and tumors was increased and thus, the tumor-to-background contrast slightly improved. The same research group reported a beneficial effect of combining the administration of [¹²⁵I]IdUrd with unlabeled IdUrd to increase the rate of DNA incorporation of [¹²⁵I]IdUrd in malignant gliomas [73]. Apparently, the C–N-glycosidic bond of IdUrd is too labile in vivo which leads to metabolites that display reduced tumor affinity. In an attempt to increase the radiotracer’s in vivo stability the tracer has been chemically modified by fluorination of the sugar moiety at different positions (3′ or 2′-substitution). However, the preparation of fluorine-stabilized iodinated thymidine analogs with retained cellular uptake, cytosolic phosphorylation, and selectivity for hTK1, appears to be quite challenging [97, 182]. A strategy for stabilizing the C–N-glycosidic bond without interfering with the cytosolic thymidine kinase has been carried out by the replacement of the furanose ring oxygen with sulfur for preparation of 5-[¹²⁵I]iodo-4′-thio-2′-deoxyuridine ([¹²⁵I]ITdU) and 5-[¹²⁵I]iodo-1-(4′-thio-β-d-arabinofuranosyl)uracil ([¹²⁵I]ITAU) (Fig. 7.16) [283]. ITdU exhibited high resistance to the glycosidic bond cleavage reaction provoked by thymidine phosphorylase, while maintaining affinity to nucleoside kinases. Also, the increased in vivo radioiodination stability and rapid DNA incorporation of ITdU resulted in a preferential uptake of radioactivity in the proliferating organs making this tracer a promising tumor-imaging agent. A comparative study of six 5-iodonucleosides revealed that the in vivo proliferation-imaging potential of nucleosides might be estimated by their in vitro affinity for TK1 and their C–N-glycosidic bond stability [284]. However, since these iodonucleosides have not been examined with regard to the important step of the nucleoside transport activity, further investigations would be necessary to allow a clear statement which radiotracer would be the most suitable for imaging of tumor cell proliferation.

By the examples of nucleoside derivatives and conjugates of carbohydrates, it was demonstrated that the development of radiotracers for intracellular targets might be problematic if bulky metal chelates are employed since cellular uptake of these radiotracers via transmembrane-spanning carriers or passive diffusion could be hindered.

6.3 Radioiodinated meta-Iodobenzylguanidine (MIBG)

Finally, we would like to highlight a long-serving but still frequently used tumor-imaging agent with an intracellular target. Meta-Iodobenzylguanidine (MIBG), a catecholamine analog, is suitable for radiolabeling with radioactive iodine (e.g., ¹²³I) for the purpose of SPECT imaging of neuroendocrine and carcinoid tumors, a subtype of neuroendocrine tumors [133]. Radiolabeled MIBG was first synthesized at the University of Michigan as early as 1980 [306]. It localizes through the physiologic nor-epinephrine reuptake mechanisms with uptake into catecholamine storage vesicles of adrenergic nerve ending and the cells of the adrenal medulla. Carcinoid tumor cells share the common characteristic of a sodium-dependent ATP/Mg²⁺ neuronal pump mechanism in their cell membranes that allows the accumulation of nor-epinephrine and MIBG where MIBG is not significantly metabolized. Initially, ¹³¹I-labeled MIBG was used for the detection of neuroendocrine tumors such as pheochromocytomas, but later its application has been extended also to scintigraphic visualization of neuroblastoma and carcinoid tumors [44, 45, 52, 105, 258, 297].

Although both [¹²³I]MIBG and [¹³¹I]MIBG can be used for the purpose of radionuclide imaging, ¹²³I has dosimetry and imaging characteristics superior to ¹³¹I and thus, it is the preferred radionuclide for SPECT imaging (Fig. 7.17). In contrast, ¹³¹I is preferred for therapy due to the emission of β-particles and dosimetric considerations [120].

Fig. 7.17

SPECT images a and SPECT/CT overlay b of a patient with a neuroendocrine tumor (pheochromocytoma) in the upper thorax. Accumulation of [¹²³I]MIBG in the malignant tissue is indicated with red arrows. The images have been kindly provided by N. Schäfer, (MD, Ph.D.), University Hospital, Zurich, Switzerland

To develop an MIBG analog with improved uptake in tumors, no-carrier-added [¹³¹I]MIBG has been developed. The methodology for producing high specific activity (no-carrier-added) [¹³¹I]MIBG was originally described in 1993, but only recently it has been developed for clinical application. With this method, nearly every molecule of MIBG contains an ¹³¹I-radiolabel, whereas prior methods provided a mixture of the ¹³¹I-tracer and the non-radioactive compound (with ¹²⁷I), wherein only 1 of 2,000 molecules of MIBG contained radioactive iodine. As a result of the high specific activity achieved by the no-carrier-added radiolabeling method, the mass of the MIBG administered can be reduced and thus undesired side-effects caused by the non-radioactive MIBG, such as hypertension during infusion could be minimized. The only concern of the no-carrier-added [¹³¹I]MIBG has been that normal tissues and organs with relatively low levels of nor-epinephrine uptake might absorb more radioactivity because of the lack of competitive inhibition of radiotracer uptake by the non-radioactive MIBG.

7 Glutamate-Ureido-Based Inhibitors of Prostate-Specific Membrane Antigen (PSMA)

Prostate-specific membrane antigen (PSMA) is a metallopeptidase expressed in epithelial cells of the prostate and highly overexpressed in 95% of advanced prostate cancers. PSMA is also known as glutamate carboypeptidase II (GCPII), folate hydrolase 1 (FOLH1), and N-acetyl-L-aspartyl-L-glutamate peptidase I (NAALADase). PSMA undergoes constitutive internalization, has no known ligand, is not specific to the prostate gland, and is expressed in other normal (e.g., salivary glands, duodenal mucosa, proximal renal tubular cells, and neuroendocrine cells in the colonic crypts) and neoplastic (e.g., transitional cell carcinoma, renal cell carcinoma, colon carcinoma, and endothelial cells of neovasculature) tissues [262].

PSMA can be selectively targeted using radiolabeled ligands based on urea-linked dipeptides, e.g., Glu-urea-Lys. These small-molecules binding the PSMA can be radiolabeled with γ-emitters like ^99mTc and ^123/131I or positron emitters like ⁶⁸Ga [3, 5, 22] and ¹⁸F for diagnosis [99] as well as with their theranostic counterparts such as ¹⁷⁷Lu (γ-and β-emitter) or ²²⁵Ac (α-emitter) for therapy [4, 122, 141, 221, 245]. Reviews summarizing the theranostic role of PSMA ligands for molecular imaging and targeted molecular radiotherapy have been published recently [2, 17, 137, 142, 220]. Here, the discussion is focused on the appropriate glutamate-ureido-based tracers labeled with ¹²³I or ^99mTc.

7.1 ¹²³I- and ¹³¹I-Labeled PSMA Radioligands

The use of low-molecular-weight PSMA ligands both for diagnosis and therapy began with the development of the ¹²³I- and ¹³¹I-labeled PSMA inhibitors MIP-1072 and MIP-1095 [115, 116, 174]. Both tracers have very similar structures (Fig. 7.18) and comparable pharmacokinetics [50].

Fig. 7.18

Chemical structures of MIP-1072 (a) and MIP-1095 (b)

MIP-1095 can be used for SPECT imaging ([¹²³I]I-MIP-1095), PET imaging ([¹²⁴I]I-MIP-1095), and endoradionuclide therapy PSMA([¹³¹I]I-MIP-1095-RLT). The first evaluation of [¹²³I]I-MIP-1095 for SPECT imaging in patients was described by Barrett et al. in 2013 [23]. Zechmann et al. reported dosimetry and therapy results using matched pair [¹²⁴I]I-MIP-1095 PET imaging and [¹³¹I]I-MIP-1095 PSMA-RLT [316]. Afshar-Oromieh et al. evaluated toxicity and antitumor activity after single and repeated PSMA-targeting radioligand therapy of metastatic prostate cancer with [¹³¹I]I-MIP-1095 [3, 5].

7.2 ^99mTc-Labeled PSMA Radioligands

Several studies have been published dealing with the preclinical development of ^99mTc-labeled PSMA radiotracers [166, 287]. The most promising agents among all ^99mTc-labeled PSMA radiotracers are [^99mTc]Tc-MIP-1404, also known as Trofolastat, and [^99mTc]Tc-MIP-1405 (Fig. 7.19). Both compounds have been developed by Molecular Insight Pharmaceuticals (MIP) [117, 175]. The MIP structures contain single-amino-acid chelators (SAACs). Functionalized polar imidazole rings have been introduced in order to reduce lipophilicity and hepatobiliary excretion. ^99mTc labeling follows the technetium tricarbonyl approach.

Fig. 7.19

Chemical structures of [^99mTc]Tc-MIP-1404 (a) and [^99mTc]Tc-MIP-1405 (b)

In preclinical studies [^99mTc]Tc-MIP-1404 showed a fast clearance from the kidneys and non-target tissues while retained to the LNCaP tumor over 4 h. Furthermore, [^99mTc]Tc-MIP-1404 also displayed the highest tumor-to-blood and tumor–to–skeletal muscle ratios [101, 117, 175]. [^99mTc]Tc-MIP-1404 (Trofolastat) is the first PSMA imaging low-molecular-weight molecule that has been used in a phase-III clinical trials for noninvasive imaging of prostate cancer (European Union Trials Register EudraCT 2012-001864-30; ClinicalTrials.gov NCT02615067 [proSPECT-AS]).

A clinical database search from April 2013 to May 2017 yielding 93 patients with histologically confirmed cancer in whom SPECT/CT with [^99mTc]Tc-MIP-1404 had been performed for primary whole-body staging before therapy [244, 245]. The authors claimed that imaging with this tracer has a high accuracy and low interobserver variability in the diagnosis of PC and allows detection of lymph node and bone metastases in a significant proportion of as yet untreated PC patients.

Recently two additional radiotracers have been introduced, [^99mTc]Tc-PSMA-I&S [228, 303] and [^99mTc]Tc-EDDA/HYNIC-iPSMA, where iPSMA stands for Lys(Nal)-Urea-Glu [84, 238]. The chemical structures are shown in Fig. 7.20. In a comparative analysis Lawal et al. evaluated the diagnostic sensitivity of SPECT/CT versus PET/CT in prostate carcinoma imaging using [^99mTc]Tc-EDDA/HYNIC-iPSMA in comparison with [⁶⁸Ga]Ga-PSMA-11 [147].

Fig. 7.20

Chemical structures of [^99mTc]Tc-EDDA/HYNIC-iPSMA (a) and [^99mTc]Tc-PSMA-I&S (b)

8 Sentinel Lymph Node (SNL) Localization

Detection of the sentinel lymph node (SNL), the first node to receive lymphatic flow as well as metastatic cells from the primary tumor site, is currently employed for planning the therapeutic treatment of cancers such as breast cancer and skin melanoma. Commonly, ^99mTc-labeled colloids in combination with blue dyes are used to detect the sentinel lymph node. The labeled colloids are injected beneath the skin surrounding the primary tumor or alternatively directly into the tumor. After entering the lymphatic vessels, colloids use the same drainage pathway as potential metastatic cancer cells and are eventually retained in the first draining lymph node (sentinel node) by phagocytosis or mechanical trapping. However, colloidal radiopharmaceuticals such as ^99mTc-sulfur colloid, ^99mTc-antimony trisulfide, and ^99mTc-labeled albumin microcolloids have disadvantages, like slow elimination rates from the injection site and migration to secondary nodes [208]. Thus, non-colloidal particles have been developed as alternatives such as ^99mTc-labeled human serum albumin and ^99mTc-labeled dextran.

A fluorescence-labeled ^99mTc-HSA derivative (indocyanine green ICG-^99mTc-NanoColl) was used to improve the surgical accuracy of laparoscopic lymph node (LN) dissection by integration of molecular imaging and intraoperative image guidance [290]. The authors could demonstrate that multimodal ICG-^99mTc-NanoColloid, in combination with a laparoscopic fluorescence laparoscope, can be used to facilitate and optimize dissection of SLNs during robot-assisted laparoscopic prostatectomy.

^99mTc-Tilmanocept (Lymphoseek), a tailor-made ^99mTc-radiopharmaceutical for SLN diagnosis based on dextran has been approved by the FDA in 2013 [19, 21, 263, 278, 286, 292, 295]. It is a macromolecule (18 kDa, 7 nm size) composed of a dextran backbone and multiple subunits of DTPA and mannose (Fig. 7.21). ^99mTc-Tilmanocept belongs to the class of receptor-binding radiopharmaceuticals. The mannose residues serve as ligands for receptors expressed on myeloid cells for recognition and binding. ^99mTc-labeling of the macromolecule is performed via diethylenetriamine pentaacetic acid (DTPA). The structure of the Tc chelate is not exactly known.

Fig. 7.21

Schematic representation of ^99mTc-Tilmanocept for lymphatic mapping and Sentinel Lymph Node (SNL) localization. The unlabeled DTPA-mannosyl-dextran has a molecular weight of 35,800 g/mol and a molecular diameter of 7.1 nm. The final amino (NC), mannose (NM), and DTPA (ND) densities were 23, 55, and 8 mol per dextran [292]

Extending the diagnostic use of Tilmanocept an ¹¹¹In-labeled derivative was described to target mannose receptor expression on macrophages in atherosclerotic plaques of apolipoprotein E-knockout mice [291].

9 Optimization of SPECT Tracer Design and Potential Reasons for Failure

The design and development of a nuclear imaging probe independent of PET or SPECT comprises an appropriate biomolecule as targeting vector, a site for conjugation that does not interfere with the biomolecule’s binding affinity to the tumor-associated target, a suitable linker length, and a radioisotope that matches with an appropriate biomolecule. For stable coordination of metallic radioisotopes, the choice of a suitable chelator is crucial. There are several possible strategies to optimize SPECT tracers with regard to their specificity to and selectivity for the targeted malignant tissue while minimizing their uptake in healthy tissues and organs. Variation of the radionuclide, modification of the bifunctional chelator, introduction of linker entities of variable spacer length for stabilization or modulation of the overall tracer characteristics, alteration of the radiolabeling technique and manipulation of the radiotracer’s blood, and normal tissue clearance by variation of the biomolecule’s overall size (e.g., antibodies versus antibody fragments or peptides). Finally, optimization of the tissue distribution of radiotracers might also be accomplished by a combination with non-radioactive substances whereof the most prominent example is the application of positively charged amino acids (e.g., lysine) that blocks renal uptake of radiolabeled Fab fragments of antibodies [28, 30] and peptides [61, 229, 294].

During the course of about two decades of (pre)clinical research with tumor-targeted SPECT tracers several reasons for potential failures of SPECT imaging agents could be identified (Table 7.4). Based on the data obtained with nuclear imaging agents that initially failed, new strategies to optimize the design and utility of SPECT tracers are currently being developed.

Table 7.4 Potential reasons for failure of tumor-targeted nuclear imaging

10 Summary and Conclusion

A variety of approaches for the design and improvement of SPECT tracers have been discussed herein. Each class of targeting agents, antibodies, peptides, and non-peptide-based small-molecules such as vitamins has its pros and cons for application in diagnostic nuclear medicine. In principle, it would be ideal to use SPECT tracers that accumulate specifically in malignancies and that are rapidly cleared via kidneys allowing high tumor-to-background contrast of radioactivity already short after administration. Such optimal characteristics are, however, not always easy to achieve.

The recent observation that somatostatin and bombesin analog antagonists provide superior characteristics over agonists with regard to their tumor accumulation is an unexpected finding that is not yet completely understood. Using oligomeric ligands to improve binding and targeting properties of radiolabeled peptides over their monomeric counterparts appears to be a more rational design that could be successfully proven, for example, with RGD-based analogs. Recently, vitamin-based radio imaging agents have been developed that are selectively accumulated in tumor cells. In the case of vitamin B₁₂, analogs with abolished binding to the ubiquitous protein transcobalalmin II showed a reduced uptake in non-targeted tissues. In the case of FR-targeting, it was the combined application with the antifolate pemetrexed that led to an improved tumor selectively of folic acid-based radioconjugates while undesired uptake in FR-positive kidneys could be reduced. Targeting of intracellular tumor markers such as the enzymes hexokinase or human thymidine kinase 1 turned out to be a more problematic strategy for SPECT tracers, particularly those that are based on radiometals, compared to the targeting of cell surface-exposed tumor markers. The necessity of the targeting agent to permeate cancer cell membranes via carrier systems or passive diffusion to reach intracellular targets could be a hindrance for a proper function of the targeting system in particular if the radioconjugate is composed of a bulky radiometal complex.

Finally, it has to be critically acknowledged that only a small selection of examples for tracer designs could be included in this chapter. The immense opportunities for the design of radiopharmaceuticals and the enormous potential it provides for future development of new and improved SPECT tracers holds great promise for early clinical application of novel imaging agents in oncology.