Exploring metabolism in vivo using endogenous ¹¹C metabolic tracers

Abstract

Cancer and other diseases are increasingly understood in terms of their metabolic disturbances. This thinking has revolutionized the field of ex vivo metabolomics and motivated new approaches to detect metabolites in living systems including proton magnetic resonance spectroscopy (¹H-MRS), hyperpolarized ¹³C MRS, and positron emission tomography (PET). For PET, imaging abnormal metabolism in vivo is hardly new. Positron-labeled small-molecule metabolites have been used for decades in humans, including ¹⁸F-FDG, which is used frequently to detect upregulated glycolysis in tumors. Many current ¹⁸F metabolic tracers including ¹⁸F-FDG have evolved from their ¹¹C counterparts, chemically identical to endogenous substrates and thus approximating intrinsic biochemical pathways. This mimicry has stimulated the development of new radiochemical methods to incorporate ¹¹C and inspired the synthesis of a large number of ¹¹C endogenous radiotracers. This is in spite of the 20-minute half-life of ¹¹C, which generally limits its use in patients to centers with an on-site cyclotron. Innovation in ¹¹C chemistry has persisted in the face of this limitation, because (1) the radiochemists involved are inspired (2) the methods of ¹¹C incorporation are diverse and (3) ¹¹C compounds often show more predictable in vivo behavior, thus representing an important first step in the validation of new tracer concepts. In this mini-review we will discuss some of the general motivations behind PET tracers, rationales for the use of ¹¹C, and some of the special challenges encountered in the synthesis of ¹¹C endogenous compounds. Most importantly, we will try to highlight the exceptional creativity employed in early ¹¹C tracer syntheses, which used enzyme-catalyzed and other “green” methods before these concepts were commonplace.

Introduction

The study of endogenous radiotracers holds special appeal in nuclear medicine. Of the positron-emitting nuclei (¹⁵O, ¹³N, ¹¹C) corresponding to atoms commonly found in endogenous metabolites, ¹¹C is the most versatile and best-studied, with a half-life (20 minutes) allowing its incorporation into many small-molecules of biologic interest. As this review will describe, numerous endogenous ¹¹C positron emission tomography (PET) radiotracers have been reported, including sugars, amino acids, nucleosides, and antioxidants. Since these are chemically and biochemically identical to their ¹²C counterparts, in many respects they offer the purest study of human metabolism. Increasingly, numerous highly prevalent diseases are understood in terms of their metabolic derangements, including cancer, diabetes, and fatty liver disease. Recent developments in the understanding of these diseases have led to a resurgence of interest in metabolomics, studied both ex vivo and in living systems. Several imaging techniques are well-suited to this analysis, including ¹H magnetic resonance spectroscopy (MRS), hyperpolarized ¹³C MR spectroscopy, and PET.

The success of ¹⁸F-FDG in the study of human cancers highlights the power of metabolic technologies in understanding and treating human disease. In the case of ¹⁸F-FDG, the target is highly glycolytic cells, that upregulate glucose (GLUT) transporters as well as hexokinase, thus trapping the phosphorylated ¹⁸F-FDG adduct intracellularly. This simple mechanism is the basis for cancer detection, staging, and treatment evaluation in several clinical scenarios. However, as any radiochemist will appreciate ¹⁸F-functionalized metabolites frequently do not perform as intended. Since endogenous molecules do not contain fluorine, ¹⁸F tracers are not sterically, chemically, or biochemically equivalent to their unmodified counterparts. As we will highlight comparing PET versions of Vitamin C, ¹⁸F substitution can markedly alter the transport properties of the molecule. Furthermore, unintended reactions in vivo in particular ¹⁸F defluorination can significantly confound interpretation of imaging data. In the context of these obstacles, radiosynthesis of the endogenous ¹¹C species may be worth the additional expense and difficulty- providing a starting point for modification with longer half-life nuclei.

In this review, we will summarize the basic approaches used in PET to identify tumors and other abnormalities, including affinity-based, microenvironment-responsive, and metabolic strategies. The primary role of endogenous ¹¹C radiotracers is in this final category, whereby the mechanism of contrast depends on host metabolism of the positron-emitting substrate. We will survey the basic methods employed to generate ¹¹C endogenous tracers, several of which have used living systems. For example, an early synthesis of ¹¹C glucose relied on the metabolism of ¹¹C CO₂ during photosynthesis. Increasingly, the ingenuity of radiochemistry labs has allowed chemical synthesis of ¹¹C endogenous molecules from simple building-blocks, including ¹¹C methyl-iodide and ¹¹C cyanide. We will also discuss special challenges faced in the synthesis of endogenous ¹¹C radiotracers, including the development of enantomerically-pure ¹¹C amino acids, and unintended incorporation of carrier into the target molecule. Finally, we will survey the success stories in this field, whereby ¹¹C endogenous tracers have been used to interrogate basic biochemistry. As interest in metabolism expands and evolves, PET will be used in conjunction with other metabolic technologies to explore the chemistry of living systems.

PET tracers- general strategies

From a clinical perspective, PET is already a highly useful technology with the potential for application to numerous new disease targets. Whereas the most common techniques used in clinical imaging, computed tomography (CT) and magnetic resonance imaging (MRI) largely provide anatomic information, PET is employed to investigate function, based on the biodistribution of a positron-emitting nucleus in the body. Positron-emitting nuclei include ¹¹C and ¹⁸F, which are appealing based on their relatively long half-lives (20 minutes for ¹¹C and 120 minutes for ¹⁸F), and minimal perturbations of known drug or metabolite structures. For the purposes of this review, we will limit our discussion to these nuclei, although remarkable progress has been made in recent years using other nuclei (eg. ⁶⁸Ga), which can also be incorporated into targeted small molecules via chelating groups(1–4). In a typical clinical PET study, a radiotracer labeled with one of these nuclei is administered intravenously to a patient. Following a delay, images of the patient are acquired via a PET scanner, which is frequently incorporated into a dual-modality instrument, PET-CT or PET-MRI for simultaneous acquisition of functional and anatomic information.

The specific mechanisms of PET tracer uptake and retention are increasingly investigated more rigorously, in cancer and other diseases. In general, these mechanisms require both disease-specific retention of the PET tracer, and elimination of nonspecific activity or “background.” There are three basic PET tracer strategies: (1) affinity based (2) microenvironment-sensitive and (3) metabolic, whereby retention of the tracer depends on metabolism by host machinery. PET tracers using these divergent strategies can be used to study the same disease. For example, the metabolic probe ¹⁸F-FDG has been used to identify the characteristic pattern of hypometabolism in Alzheimer’s disease, to complement the application of affinity-based amyloid and tau tracers.

Affinity-based PET probes represent a very large number of published tracers. In many cases, the starting point is known pharmacology for a given receptor-ligand pair. Incorporation of the PET nucleus is often performed to minimize steric or electronic perturbations, so as to retain high-affinity binding. Recent examples of high-affinity PET tracers include those developed for neurodegenerative disease, in particular targeting amyloid plaques and tau-derived neurofibrillary tangles(5–8). The development of high-affinity PET tracers for beta-amyloid highlights a fairly typical progression for affinity-based probes. An early, major accomplishment in this area was the development of Pittsburgh compound B, ¹¹C PiB whose structure was derived from thioflavin T, a dye used for histologic staining of misfolded protein aggregates(9). The chemical structure of ¹¹C PiB as well as an early ¹¹C PiB PET scan in an Alzheimer’s patient are shown in Figure 1. Since the use of short half-life ¹¹C tracers is largely restricted to large medical centers with an on-site cyclotron, subsequent efforts focused on the development of high performance ¹⁸F tracers including ¹⁸F-florbetapir (¹⁸F-AV-45)(10). Currently, ¹⁸F-AV-45 is commercially available (Siemens), allowing use at centers without a cyclotron or radiopharmaceutical facility. The success of affinity-based methods is further highlighted by the success of recent approaches targeting prostate-specific membrane antigen (PSMA), which is highly expressed in a number of prostate cancers. Two approaches show high promise, including a phosphoramidate-derived tracer which uses substrate mimicry to yield a high affinity ¹⁸F ligand(11). Another successful ligand is ⁶⁸Ga-PSMA, which takes advantage of low-cost ⁶⁸Ga generators to yield a method that has shown early promise in the detection and management of prostate cancer. In general, affinity-based PET tracers require a very low dissociation constant (K_d) in the nM or pM range and include both small-molecules and antibodies. Careful consideration to specific activity is also essential to affinity-based strategies.

Figure 1.

Affinity-based tracer Pittsburgh compound B (PiB), an ¹¹C analog of thioflavin T. This compound is used to image β-amyloid plaques seen in neurodegenerative diseases including Alzheimers. In this case the tracer showed markedly increased uptake in the brain of an Alzheimers patient relative to a healthy control. An ¹⁸F FDG is shown for comparison showing areas of hypometabolism. Figure adapted from Klunk et al. 2004.

A second strategy uses the disease microenvironment to sequester PET nuclei. The best example is ¹⁸F-fluoromisonidazole (FMISO), which undergoes chemical transformation in tissues with low oxygen tension, with multiple reductions resulting in species capable of covalent binding to macromolecules, or conjugation to reduced glutathione (GSH)(12). These transformations are shown in Figure 2, as is a FMISO scan demonstrating the effects of radiation therapy on a brain tumor(13). A large number of additional hypoxia-sensitive tracers have been reported, several of which are also imidazole-derived, but others, including ⁶⁴Cu-ATSM, rely on tumoral reduction of Cu(II)(14). Recently, other tumor-specific analytes have been targeted, including acid (H⁺)(15), reactive oxygen species (in particular H₂O₂)(16), and formaldehyde(17). These strategies have used analyte-sensitive, metastable ¹⁸F-FDG and ¹⁸F-FLT precursors to show specific accumulation in cells and preclinical models with a pH-sensitive strategy shown in Figure 3.

Figure 2.

Putative mechanism of ¹⁸F-FMISO accumulation. In hypoxic tissues sequential reduction of ¹⁸F-FMISO results in a species capable of both covalent binding to macromolecules and conjugation to reduced glutathione (GSH). The lower images show dramatic effects of tumor reoxygenation after radiation therapy in glioblastoma multiforme. Images A through C show images from a patient before therapy (C shows an ¹⁸F-FMISO-PET image). E through G show images following therapy, demonstrating marked decrease in ¹⁸F-FMISO retention despite a similar MRI appearance. Figure adapted from Masaki et al. 2016 and Narita et al. 2012.

Figure 3.

Analyte-sensing strategy using an ¹⁸F-FDG precursor. In this case the FDG precursor is labile at acidic pH, promoting decomposition to ¹⁸F-FDG in the tumoral microenvironment. In preclinical imaging studies, high tumor uptake was seen while brain activity was suppressed. Adapted from Flavell et al. 2016.

Finally, several of the most successful PET tracers depend on biotransformation for image contrast. These tracers can either be chemically identical to an endogenous substrate (¹¹C-thymidine), or a well-tolerated mimic (thymidine analog ¹⁸F-FLT). The best known metabolic PET tracer is ¹⁸F-FDG, in which ¹⁸F is substituted for a hydroxyl group at the 2-position. This substitution does not impede ¹⁸F-FDG by GLUT transporters (primarily GLUT 1,3,4) or phosphorylation at the 6-position by hexokinase, but isomerization to fructose (glucose phosphate isomerase) is effectively blocked. Thus, trapping of ¹⁸F-FDG becomes a surrogate for GLUT transporter and hexokinase expression, which are up-regulated in most cancers. Of note, the contrast obtained in a PET-FDG scan is partially related to charge trapping; the phosphorylated ¹⁸F-FDG adduct is retained in the cell, as opposed to the behavior of readily diffusible, nonpolar molecules.

Beyond glycolysis-the power of metabolic strategies

The success of ¹⁸F-FDG in oncology is perhaps not surprising, given that cancer is increasingly understood as a metabolic disease. A large number of metabolic pathways have been exploited for cancer diagnosis, including glycolysis, DNA synthesis, glutaminolysis, membrane synthesis, and antioxidant cycling. The role of metabolic reprogramming in cancer pathogenesis is highlighted in the discovery of numerous oncometabolites including 2-hydroxyglutarate (2-HG), that is produced in cancer-associated isocitrate dehydrogenase 1 (IDH1) mutations(18–20). Strikingly, 2-HG is not a mere bystander of abnormal metabolism- but rather itself an epigenetic modifier that influences DNA and histone demethylation. The discovery of this oncometabolite was quickly followed by reports of imaging methods used to detect it, with 2-HG successfully identified by ¹H-MRS using spectral editing techniques(21,22). With its power to generate structures derived from endogenous molecules, PET is well suited to interrogate the divergent metabolism seen in cancer and other diseases.

Why ¹¹C?

The PET radionuclei incorporated into endogenous molecules include ¹¹C, ¹³N, and ¹⁵O, with respective half-lives of 20 minutes, 10 minutes, and 2 minutes respectively. Of these, ¹¹C is clearly best-suited to incorporation into complex organic structures, based on the relatively long half-life and diversity of precursor molecules available for use. Endogenous ¹¹C structures are also amenable to enzymatic and even “living” syntheses, based on their homology to natural substrates. In fact, the chemistry of ¹¹C building blocks can be simpler than that of ¹⁸F precursors, providing an important starting point for evaluation of a new tracer concept.

The main justification for using ¹¹C tracers to study metabolism is their close approximation of endogenous molecules. In many cases, ¹⁸F substituted structures fail to replicate the biochemical modification needed for tracer uptake and accumulation. The difference between endogenous ¹¹C molecules and their ¹⁸F counterparts is highlighted by studies of ascorbate-derived PET tracers. One potential application of PET versions of Vitamin C is that they undergo ascorbate-recycling in response to reactive oxygen species, generating sensors that are potentially responsive to oxidative stress. Vitamin C undergoes a two-electron oxidation to dehydroascorbic acid (DHA), which in solution exists as both hydrated and bicyclic forms. The bicyclic version of DHA is postulated to undergo rapid GLUT transport, based on its structural similarity to glucose(23,24). The first synthesis of ¹⁸F-Vitamin C was accomplished with ¹⁸F incorporated at the 6-position by Kothari et al., generating a molecule incapable of forming a bicyclic species due to the substitution of ¹⁸F for a hydroxyl group(25,26). Recently, a synthesis of ¹¹C Vitamin C was reported, for which the authors reported accumulation of ¹¹C in ROS-producing cells via ascorbate recycling(27). The presumed mechanism of ¹¹C ascorbate recycling is shown in Figure 4. Another study contrasted ¹¹C and ¹⁸F analogs more explicitly, via the comparison of L-[β-11C] DOPA and 6-fluoro-[β-11C]-L-DOPA(28). Significant differences in brain radiotracer deposition were seen in a monkey model, suggesting that ¹¹C L-DOPA and ¹⁸F L-DOPA tracers are non-identical with respect to their distribution in vivo.

Figure 4.

Oxidation of [1-¹¹C] Vitamin C. (A) Two-electron oxidation of Vitamin C to dehydroascorbic acid (DHA) results in a species that exists in solution as a bicyclic hemiketal. Cyclization requires a 6-OH and this form is believed to be responsible for GLUT-mediated uptake. (B) Mechanism of [1-¹¹C] Vitamin C reactive oxygen species (ROS) sensing. Adapted from Carroll et al. 2016.

The liveliest syntheses of endogenous ¹¹C tracers

Early development of ¹¹C tracers required considerable ingenuity, harnessing living systems, and the specificity of enzymes. For example, an early synthesis of ¹¹C glucose used photosynthesis, via exposure of Swiss chard leaves to ¹¹C-CO₂ followed by separation using liquid chromatography(29). This method is contrasted with the first direct radiochemical synthesis of ¹¹C glucose using reaction of ¹¹C-HCN with D-arabinose followed by reduction(30).

One of the main advantages of using enzymes in PET radiosynthesis is their inherent substrate specificity, particularly useful for the enantiomeric resolution of racemic mixtures. A remarkable example is the synthesis of L-aromatic amino acids, ¹¹C tyrosine and ¹¹C tryptophan(31). The starting material for their syntheses was racemic ¹¹C alanine, generated from ¹¹C HCN via a Strecker synthesis. As shown in Figure 5, alanine is converted to pyruvate by oxidation/hydrolysis (D-amino acid oxidase, glutamic-pyruvic transaminase), and to tyrosine and tryptophan by the actions of β-tyrosinase and tryptophanase, respectively. Most ingenious is the use of D-amino acid oxidase, which uses flavin adenine dinucleotide (FAD) as a cofactor to selectively oxidize the D-enantiomer, allowing separation of the ¹¹C L-amino acid from the mixture.

Figure 5.

Synthesis of ¹¹C tyrosine and ¹¹C tryptophan via a single-pot enzymatic synthesis. Purification of the L-isomer was made possible by use of D-amino acid oxidase. Adapted from Antoni et al. 2015.

¹¹C building blocks for chemical synthesis

Historically, the most important technical advances in the widespread use of endogenous ¹¹C tracers were the syntheses of reactive ¹¹C building blocks (in particular ¹¹C MeI), and the development of automated/remote-controlled synthesis equipment. The building blocks most frequently used in the syntheses of ¹¹C endogenous radiotracers are summarized in Figure 6. Their use in ¹¹C tracers falls into three main categories: methylations, cyanations, and carbonyl chemistry (CO₂, CO, COCl₂). Several excellent reviews have highlighted methods of ¹¹C incorporation into radiopharmaceuticals(32–35).

Figure 6.

Common building blocks in ¹¹C radiochemical synthesis. Adapted from Scott 2009.

One of the most significant advances in ¹¹C radiotracer synthesis was the development of ¹¹C MeI. Methylations using ¹¹C MeI are an excellent way to label a substrate containing a nucleophilic oxygen, nitrogen, or sulfur. One of the first PET radiotracers to use ¹¹C MeI was ¹¹C methionine, formed by the reaction between ¹¹C MeI and a homocysteine-derived precursor. The first radiosynthesis of ¹¹C L-methionine used an S-benzyl protected precursor in Na/NH₃, which allowed both facile reduction and deprotonation of the reactive thiol(36). Although this method allowed generation of ¹¹C L-methionine with very high enantiomeric excess (> 99%), it has been replaced by a convenient technique using the cyclic precursor L-homocystine-thiolactone(37). Despite the lower reported enantiomeric excess (approximately 70–80%), this is not considered critical for most imaging applications. These syntheses are shown in Figure 7.

Figure 7.

Syntheses of ¹¹C L-methionine.

Special synthetic challenges

A significant challenge in ¹¹C radiochemistry has been the synthesis of chiral amino acids. In some cases incorporation of ¹¹C occurs away from the key chiral center, for example in the synthesis of ¹¹C methionine or ¹¹C glutamine(38). However, in other cases ¹¹C chiral amino acid synthesis has employed methods that yield racemic mixtures, in the absence of clever chemical and biochemical strategies. Synthesis of racemic [1-¹¹C] L-alanine is most commonly performed via the Strecker synthesis, in which ¹¹C-CN is incorporated into an aldehyde analog, in this case a bisulfate adduct followed by treatment with ammonium hydroxide(39). Analogous procedures have been performed for [1-¹¹C] leucine(40) and [1-¹¹C] tyrosine(41). For synthesis/purification of the pure L-enantiomer, several methods have been employed including (1) separation using high performance liquid chromatography (HPLC) on a chiral stationary phase (2) enzymatic resolution (3) use of a chiral auxiliary, usually glycine-derived (4) application of a chiral catalyst, either for selective hydrogenation or alkylation.

Alkylation of glycine-derived structures is a frequently applied method to yield [3-¹¹C] amino acids. These have included use of [(+)-2-hydroxypinanyl-3-idene] glycine tert-butyl ester(42), the Schiff base of (S)-O-[N-benzypropyl] amino] benzophenone-glycine-Ni complex(43), and the camphor-derived Oppolzer’s synthon(44) (Figure 8). These have in many cases afforded high ee syntheses of the desired L enantiomer. In addition, several chiral catalysts have been used including chiral diphosphine-rhodium(45) and more recently phase-transfer catalysts(46), applied to stereoselective hydrogenation and alkylation respectively.

Figure 8.

Radiosynthesis of [3-¹¹C] L-alanine via an Oppolzer’s synthon-derived glycine equivalent. Adapted from Langstrom et al. 2013.

An additional ¹¹C radiosynthetic challenge is the successful removal of potential reactants. In some cases these require special engineering solutions to remove reactive species. For example, use of ¹¹C-HCN is complicated by the presence of large quantities of NH₃. In the synthesis of [1-¹¹C] lactate via an acetaldehyde precursor, if NH₃ is not removed from the reaction milieu, the product is racemic [1-¹¹C] alanine(47). The possible reaction products are summarized in Figure 9. To synthesize [1-¹¹C] lactate, a trapping method was developed for ¹¹C-HCN consisting of NaOH applied to a platinum wire. This method successfully removed NH₃ from the reaction.

Figure 9.

Synthesis of [1-¹¹C] lactate via ¹¹C-cyanide depends on the successful scavenging of NH₃.

Studying the biochemistry of living systems

Basic energy metabolism- glycolysis and ketone bodies

As highlighted by the success of ¹⁸F-FDG, the altered glycolysis seen in cancer and other diseases is a major target for imaging agent development. The majority of applications have been to human cancers. As reported in 1926 by Otto Warburg, tumors have altered metabolism, reverting to a primitive phenotype for energy. Warburg’s studies found a remarkable preference of tumors for aerobic glycolysis over oxidative phosphorylation. This was a surprising result given the massive energy benefit complete oxidation of glucose confers. The “why” remains controversial, but the glycolytic phenotype seen in cancer has been leveraged in molecular imaging strategies. PET scans using ¹⁸F-FDG are currently the standard of care for identifying tumors and detecting treatment response. Increasingly, ¹⁸F-FDG is being used for other diseases including neurodegenerative disorders, seizures, inflammation, and infection.

Several metabolites involved in glycolysis have been ¹¹C labeled for metabolic study. One interesting feature of these ¹¹C molecules is the absence in many cases of a metabolic stopping point. A critical feature of ¹⁸F-FDG is that substitution at the 2-position renders its phosphorylated adduct incapable of glucose-fructose isomerization by glucose-6-phosphate isomerase. This allows metabolic trapping of ¹⁸F, and easier interpretation of PET-FDG data: retained tracer is a consequence of upregulated glucose (GLUT) transport and/or hexokinase activity in the tissues of interest. Interestingly, a similar strategy was used in the development of [2-¹¹C]-2-deoxyglucose as the ¹¹C counterpart for ¹⁸F-FDG(48,49). In contrast, mechanistic trapping of most endogenous ¹¹C metabolites is not possible, although in several cases it is likely that the label is effectively trapped in a large steady-state pool. For example, aggressive tumors and their microenvironments have large steady-state concentrations of lactate. It is therefore likely that several ¹¹C substrates are converted to, and retained as ¹¹C lactate in these contexts. Similarly, many cell types in particular immune cells have high steady-state concentrations of Vitamin C. Therefore, if ¹¹C Vitamin C or ¹¹C dehydroascorbic acid is applied to inflamed tissues, it is likely that the retained label represents ¹¹C Vitamin C.

Numerous methods have been developed to label glucose, using ¹¹C cyanide(30), ¹¹C methylenetriphenylphosphorane(50), or nitromethane(51). Pyruvate has been prepared enzymatically from ¹¹C alanine(39), via ¹¹C cyanide or carboxylation of an acyl carbanion(52). Similarly, lactate has been generated enzymatically(39) or via ¹¹C cyanide(47). Acetate can enter glycolysis via acetyl coA and its ¹¹C version has been synthesized via carboxylation of methyl-Grignard(53). ¹¹C glutamine, was synthesized recently from a chiral precursor using ¹¹C-CN(38), while [5-¹¹C] glutamate was obtained from O-acetyl-L-homoserine by enzymatic catalysis(54). Both of these metabolites can enter the TCA cycle via alpha-ketoglutarate, and glutaminolysis is upregulated in numerous human cancers(55). Finally, the role of alternate energy sources has been explored via the radiosynthesis of ¹¹C ketone bodies. In fasting, the normal brain increasingly relies on ketone bodies (β-hydroxybutyrate, acetoacetate, and acetone) produced in the liver from fatty acid β-oxidation. Both ¹¹C acetoacetate(56) and β-hydroxybutyrate have been synthesized, the latter via enzymatic conversion using β-hydroxybutyrate dehydrogenase(57).

DNA synthesis

Nucleoside analogs have a special role in antiviral therapeutics and nuclear imaging. The ¹⁸F-thymidine analog ¹⁸F-FLT has a mechanism similar to that of ¹⁸F-FDG, whereby the probe is transported by ENT-1 and phosphorylated by thymidine kinase (TK)(58,59). [2-¹¹C] thymidine has been developed and tested as a PET tracer of thymidine incorporation into deoxyribonucleic acid (DNA), and thus as a means to detect cellular proliferation non-invasively. [2-¹¹C] thymidine is synthesized from [2-¹¹C] thymine, obtained by cyclization of ¹¹C urea and diethyl beta-methylmalate(60). After purification, [2-¹¹C] thymine and 2′-deoxyribose-1-phosphate are incubated in the presence of thymidine phosphorylase to form [2-¹¹C] thymidine. Another synthesis of ¹¹C thymidine using ¹¹C MeI has been reported(61). Although sophisticated modeling may be required, several authors have argued that the flux of [2-¹¹C] thymidine into DNA can serve as a marker of tumor proliferation(62).

Reduction and oxidation- Vitamin C and uric acid

The non-invasive detection of oxidative stress could have profound implications for the diagnosis and management of numerous diseases. Dysregulation of reactive oxygen species (ROS) provides a powerful motivation to develop non-invasive biomarkers of oxidative stress. A synthesis of ¹¹C Vitamin C was recently reported, synthesized from ¹¹C cyanide and L-xylosone based on a modification of the previously reported ¹³C and ¹⁴C enrichment techniques(27). The authors studied sensitivity to numerous inflammation-relevant ROS including H₂O₂, superoxide, and hypohlorite, and showed ROS-dependent accumulation in activated neutrophils. Another potential ROS-sensitive ¹¹C radiotracer synthesized recently is [¹¹C] uric acid. The synthesis of [¹¹C] uric acid was achieved by reacting 5,6-diaminouracil with ¹¹C phosgene(63). Uric acid has a well-known role in gout, but its known redox pairs (hypoxanthine and allantoin) have implicated it in numerous redox functions in vivo. One potential challenge in developing ROS-sensitive sensors for the noninvasive detection of oxidative stress is autoradiolysis(64). Studies of ¹¹C Vitamin C found that a high specific activity synthesis, while feasible, resulted in poor stability under physiologic conditions. It remains to be shown whether lower specific activity radiosynthesis, or other stabilizing strategies will afford sufficient sensitivity for in vivo ROS detection.

Metabolism of amino acids

Amino-acid derived PET tracers are applied frequently to study human cancers, especially brain tumors. The most important amino-acid derived radiotracer synthesized is ¹¹C L-methionine, which provided one of the first applications of ¹¹C methyl-iodide radiosynthesis(36). ¹¹C L-methionine studies in two patients with glioblastoma multiforme are shown in Figure 10(65). Since the use of ¹¹C L-methionine is restricted to PET centers with an on-site cyclotron and radiochemistry facility, promising results obtained with ¹¹C L-methionine have stimulated the development of ¹⁸F-labelled aminoacid tracers, particularly O-(2-¹⁸F-fluoeoethyl1)-L-tyrosine (FET)(66), ¹⁸F-FACBC(67), and ¹⁸F L-dopa(68). The latter has been used in oncologic applications in addition to Parkinson’s disease. ¹¹C tyrosine has been used in human soft tissue sarcomas for which a protein synthesis rate was calculated(69).

Figure 10.

Brain tumor studies using ¹¹C L-methionine. In some cases this tracer allows distinction between recurrent tumor (A, pathologically proven) and radiation necrosis (B). Adapted from Terakawa et al. 2008.

Membranes, lipids and steroids

In ¹H magnetic resonance spectroscopy (¹H-MRS), choline as a component of cell membranes is interpreted as a marker for cellular membrane turnover. Elevations of choline as detected by ¹H-MRS are seen in neoplasms of both of the brain and prostate(70,71). An increased choline resonance on ¹H-MRS, as well as elevated uptake of ¹¹C choline are associated with increased cell proliferation and choline kinase activity. The most frequent synthesis of ¹¹C choline is via ¹¹C methyl-iodide, to generate [N-methyl-¹¹C] choline(72). Recently ¹⁸F-choline has been used more frequently particularly in prostate cancer patients(73). Several fatty acids have been synthesized including ¹¹C arachidonic acid(74), an important signaling molecule and ¹¹C palmitate(75), synthesized via an ¹¹C alkyl iodide. The syntheses of endogenous steroids have been reported including [17α-¹¹C] methyltestosterone(76).

Neurotransmitters

Several neurotransmitters and their precursors have been studied including ¹¹C L-tyrosine(41), L-DOPA(77), L-tryptophan(31), 5-hydroxy-L-tryptophan(31), ¹¹C epinephrine(78) and ¹¹C norepinephrine(79). ¹¹C epinephrine has been used to study cardiac sympathetic nerve function in the context of heart transplantation, since the latter is associated with decreased sympathetic nerve integrity(80). An interesting study looking at cocaine toxicity showed that cocaine inhibited ¹¹C norepinephrine uptake in the heart, in monkeys(81).

Differential metabolism and site-specific ¹¹C labeling

One attractive feature of ¹¹C endogenous compounds is that they may be labeled at different nuclei, allowing the detection of divergent metabolism. An excellent example is ¹¹C pyruvate, which may be labeled at the 1 or 3-positions. In the heart, oxidative phosphorylation requires that pyruvate be decarboxylated to synthesize acetyl-CoA. When [1-¹¹C] pyruvate is decarboxylated, labeling of metabolically inert ¹¹C-CO₂ results in no tissue retention(35). In contrast, when [3-¹¹C] pyruvate is used the label is incorporated into acetyl-CoA for import into mitochondria. An opposite result would be expected for the interrogation of highly glycolytic tumors. Specifically, reduced mitochondrial import of acetyl-CoA is typically seen in tumors, which are more likely to use aerobic glycolysis. ¹¹C pyruvate labeled at both the 1 and 3-positions would likely be incorporated into ¹¹C lactate and either retained in tumors or exported. Another excellent example of site-specific labeling used to identify the metabolic fate of nuclei was ¹¹C L-DOPA. L-DOPA may be labeled in either the β or carboxylic positions, with the label ‘lost” as ¹¹C-CO₂ by the activity of DOPA-decarboxylase for the latter(82). Imaging studies using site-specific ¹¹C labeling to demonstrate decarboxylation are shown in Figure 11.

Figure 11.

Site-specific labeling of ¹¹C probes allows detection of differential metabolism. (A) Incorporation of ¹¹C into the myocardium is seen with [3-¹¹C] pyruvate, since the label is transferred to acetyl-CoA which enters the TCA cycle. In contrast the label is “lost” following decarboxylation for [1-¹¹C] pyruvate. (B) Similar findings with [1-¹¹C] L-dope, with loss of the label as ¹¹C CO₂. Adapted from Antoni et al. 2015.

Endogenous ¹¹C tracers- synergy with other techniques

Recently, the study of endogenous ¹³C nuclei in vivo has been enabled by a new spectroscopic imaging method, namely hyperpolarized (HP) ¹³C MRS. This technique involves increasing the NMR-observable signal beyond that seen at thermal equilibrium, and has been applied to the study of several endogenous molecules including pyruvate, fumarate, bicarbonate, fructose, and ascorbic acid(83). The major advantage of this technique is direct observation of small molecules via their ¹³C chemical shifts, allowing easy identification of metabolic products. In a typical experiment, enriched ¹³C pyruvate is irradiated with microwave energy in a high magnetic field in the presence of a paramagnetic electron source, for approximately 1 hour at a temperature near absolute zero. After “polarization,” the sample is rapidly heated, dissolved in aqueous solvent and injected intravenously. The real-time conversion of the introduced ¹³C agent to various metabolites is then visualized via differences in chemical shift.

The most commonly employed hyperpolarized ¹³C substrate is [1-¹³C] pyruvic acid, which has been studied in a small cohort of prostate cancer patients(84), and in numerous preclinical oncology models. [1-¹³C] pyruvate is converted rapidly to [1-¹³C] lactate in numerous tumors, via the activity of NADH-dependent lactate dehydrogenase (LDH). The rate of this conversion has been shown to be grade-dependent in several tumor models(85). In these models, little or no ¹³C CO₂ or ¹³C HCO₃⁻ is visualized. In contrast, when [1-¹³C] pyruvic acid is introduced into the myocardium of normal animals (for example pigs), rapid decarboxylation via PDH results in detection of a large ¹³C HCO₃⁻ resonance. Absence of this signal may be a biomarker for myocardial ischemia and/or cardiac failure(86). Detection of these divergent pathways is analogous to what might be observed in similar biologic scenarios using ¹¹C pyruvate. In many cases data gleaned from ¹¹C PET and ¹³C hyperpolarized MRI are complimentary, and elucidate the key mechanisms of image contrast. For example, development of the ¹³C ascorbic acid/dehydroascorbic acid redox pair(87) for hyperpolarized MRI motivated the use of ¹¹C ascorbic acid as an endogenous redox sensor.