Development of radiotracers for oncology – the interface with pharmacology

Abstract

There is an increasing role for positron emission tomography (PET) in oncology, particularly as a component of early phase clinical trials. As a non-invasive functional imaging modality, PET can be used to assess both pharmacokinetics and pharmacodynamics of novel therapeutics by utilizing radiolabelled compounds. These studies can provide crucial information early in the drug development process that may influence the further development of novel therapeutics. PET imaging probes can also be used as early biomarkers of clinical response and to predict clinical outcome prior to the administration of therapeutic agents. We discuss the role of PET imaging particularly as applied to phase 0 studies and discuss the regulations involved in the development and synthesis of novel radioligands. The review also discusses currently available tracers and their role in the assessment of pharmacokinetics and pharmacodynamics as applied to oncology.

LINKED ARTICLES

This article is part of a themed section on Imaging. To view the other articles in this section visit http://dx.doi.org/10.1111/bph.2011.163.issue-8BJP has previously published an Imaging in Pharmacology themed section, edited by A Davenport and C Daly. To view this section visit http://dx.doi.org/10.1111/bph.2010.159.issue-4

Introduction

Drug development is a lengthy, complex and expensive process. Currently the development of a typical investigational agent takes 10–15 years and costs approximately US $800 million. Despite the time and financial commitment, attrition rates are high with less than 10% of novel anticancer drugs entering the market from phase I studies (Dimasi, 2001). This low success rate has been ascribed to two major causes; the lack of predictive preclinical models and inadequate, complex trial designs (Bergstrom et al., 2003; Marchetti and Schellens, 2007). In parallel, the area of drug discovery and development is undergoing a rapid expansion in almost all therapeutic areas of medicine, particularly in oncology. This is in part fuelled by innovative technologies such as computer-aided drug design, combinatorial chemistry, and high-throughput methodologies for studying genomics and structural biology. The use of these tools has led to an increase in the identification of potentially druggable targets that drive pathological conditions. However, due to increasing costs and limited resources, it is not possible to develop all potential drug candidates and this has initiated discussions as to how to more effectively drive drug development.

It is generally accepted that early information on exposure of target tissue to investigational drug, modulation of the desired target, and biological effect will assist in the prompt translation of agents from the laboratory to the clinic (Workman et al., 2006). It is important to establish these measurements early in the drug development pathway, either prior to phase I studies (phase 0) or as part of phase I studies. To this end positron emission tomography (PET) can be used to obtain crucial pharmacokinetic (PK) and pharmacodynamic (PD) information non-invasively, and hence the role of PET imaging in drug development is increasing.

PET is a non-invasive imaging methodology that involves the administration of compounds labelled with positron-emitting radioisotopes (radiotracers or radioligands) that are formulated for intravenous injection (radiopharmaceuticals). Radiotracers are substrates of normal physiological pathways (activated probes) or localize to particular targets because of specific binding interactions (targeted probes). These radiolabelled probes are administered at tracer amounts so as to not perturb normal physiology but allows in vivo studies under physiological conditions. PET therefore allows the investigation of substrate-target interactions or physiological processes with the advantage of not having to administer an efficacious dose of a chemical compound as with classical pharmacological investigations. As radiotracers act as substrates for normal physiological processes they need to undergo extensive preclinical validation. The development of novel PET radiotracers is a complex process that requires a multidisciplinary research group including expertise in organic chemistry, radiochemistry, pharmacology, cell biology, physics, image analysis and oncology. In this review we initially describe the process by which novel cancer imaging tracers are developed, from conception to the clinic (Figure 1). The second part of the review will focus predominantly on PET for assessing PD end-points that illustrate the hallmarks of cancer, proliferation, apoptosis and angiogenesis studies in patients, and the utility of PET as an early biomarker of therapeutic response. We will also discuss the PET imaging in phase 0 and PK studies.

Figure 1.

The developmental process involved in designing a new radiotracer for clinical use. OLINDA, organ level internal dose assessment.

Radiosynthesis

Any biological target that is present at increased or decreased levels in cancer cells can be visualized by PET. Ideally the target should be as specific as possible for the disease process, and consideration should also be given to the clinical information that might be gleaned from imaging the target or pathway. The development of a targeted radiotracer involves the synthesis of an extensive library of potential compounds for a particular target, with the expectation that only a few imaging agents will successfully progress to clinical PET studies. This library can contain several analogues of the parent compound that have a known affinity for a target and often may be based on previously known compounds. For example, in the design of a PET imaging tracer for the detection of α_vβ₃-integrin as a marker of angiogenesis, a focused library of peptides containing the arginine-glycine-aspartic acid (RGD) sequence were screened and [¹⁸F]Galacto-RGD, was taken forward into clinical development (Haubner et al., 2001). PET tracers are manufactured by automated synthesis modules the first of which was designed for the manufacture of [¹⁸F]fluorodeoxyglucose (FDG) (Iwata et al., 1984). Within this module the radiochemical hardware was interfaced with a microcomputer that controlled the 32 steps involved in the radiosynthesis of [¹⁸F]FDG (Fowler et al., 1981). More recently, the synthesis of [¹⁸F]FDG occurs via a high yield, one-pot sysnthesis reaction, which has become the synthetic pathway of most commercial [¹⁸F]FDG automation modules (Hamacher et al., 1986) (Figure 2).

Figure 2.

Chemical structures of common tracers involved in imaging. (A) [¹⁸F]FDG (B). [¹⁸F]FLT C isatin sulfonamide, [¹⁸F]ICMT-11, (D) [¹¹C]acetate, (E) [¹¹C]choline, (F) [¹²⁴I]– 2′-fluoro-5-iodo-[beta]-D-arabinouridine (FIAU).

To have clinical utility, these potential tracers must have subnanomolar or nanomolar affinity for the physiological target so as to not interfere with normal biological function. Other desirable characteristics of radiotracers for somatic tumours include:

• high specificity and/or selectivity

• high plasma clearance and low plasma protein binding

• neutral and hydrophilic to enhance elimination and reduce effective (radiation) dose

• limited or measurable metabolism and amenable to kinetic modelling

• low toxicity

• good in vivo stability because of low metabolic clearance

Of central importance, the compound must have a suitable site for labelling with radioisotopes such as carbon-11 or fluorine-18, the most commonly used radioisotopes for PET in oncology. However, research utilizing other radioisotopes including copper-64, yttrium-86 and iodine-124 is increasing. When labelling a drug or molecule of choice, an element in the target molecule is substituted by its radioisotope. Problems arise however with the substitution of hydrogen as this does not have a positron emitting isotope. In this case hydrogen is substituted by fluorine-18, which may result in altered PK compared with the native molecule but these substitutions are conservative, and in the past have yielded useful probes such as [¹⁸F](FDG).

Radioisotopes are produced by charged particle nuclear reactions within a cyclotron where a target vessel containing a gas or fluid is bombarded by protons or deuterons. These radioisotopes are obtained from the cyclotron in simple chemical forms (Table 1). For example carbon-11 could be obtained as [¹¹C]CO₂ or [¹¹C]CH₄ and fluorine-18 as either [¹⁸F]fluoride or [¹⁸F]fluorine. Once generated, radiochemistry is then used to incorporate these radioactive substrates into the compounds of interest. This process involves a number of steps, culminating in the conversion of the radioisotope to a reactive intermediate that is then used to label the compound of interest. The radiolabelled compound is isolated, for instance by high-performance liquid chromatography (HPLC), to eliminate radiochemical and chemical by-products that may have arisen during synthesis. The product is then formulated and can be used for aseptic injection.

Table 1.

Characteristics of radionuclides commonly used in the development of PET tracers

Radionuclide	Half-life (min)	Intermediate building blocks	Typical specific activity (GBq·µmol−1)
11C	20	[11C]CH3I, [11C]HCN, [11C]CO, [11C]CO2, [11C]CH4	3.4 × 105
18F	110	[18F]F-, [18F]F2	6.3 × 104
76Br	972	[76Br]Br+,[76Br]Br-	7.2 × 103
13N	10	[13N]NH4+,[13N]NH4-,[13N]NH2-	7.0 × 105
15O	2	[15O]H2O	3.4 × 106
124I	40 (days)	[124I]I+, [124I]I-	1.15 × 103
68Ga	68	[68Ga]Ga	1.02 × 105
64Cu	760	[64Cu]CuCl2	2.2 × 108–1.85 × 109

Due to the short half-life of PET radioisotopes, the labelling of compounds involves considerable amounts of radioactivity, and therefore needs to be performed within remote-controlled lead-shielded ‘hot’ cells. The short half-life of radioisotopes also results in a time constraint for PET radiolabelling, and the introduction of the radioisotope at the latest possible stage in the synthesis is desirable. The development of novel synthetic procedures to improve this ‘late labelling’ is an active area of research.

Compounds labelled with carbon-11 or fluorine-18 are produced with high specific radioactivity, that is, these compounds have a high level of radioactivity and are associated with a very small amount of the stable compound or carrier. The amount of carrier present is usually sufficient to obtain a mass spectrum of the compound. To ascertain the position of the incorporated radioisotope, co-labelling experiments can be conducted. For example, for carbon-11 chemistry, very small amounts of carbon-13-labelled agents can be mixed with the carbon-11 equivalents and used in the radiosynthesis. The purified product is then examined by carbon-13 nuclear magnetic resonance and mass spectrometry.

The commonly used radionuclides in PET that are most appropriate for labelling of drugs are summarized in Table 1. Carbon-11 is an ideal nuclide for labelling molecules of biological interest, however, its short half-life (20.4 min), complex synthesis procedures and difficulties in obtaining high specific activities are limitations for routine manufacturing by PET centres. The metallic nuclides, gallium-68, copper-64 and yttrium-86 are useful for developing peptide- and protein-based radiotracers. Fluorine-18 is ideal for a number of reasons; it has a long half-life (110 min) allowing transportation from production to clinical PET imaging sites, it has acceptable dosimetry allowing multiple studies in a given patient, in terms of production it can be produced in large quantities in a cyclotron, has high labelling yields and finally, fluorine-18 has low positron energy compared with the other nuclides, which improves the resolution of images obtained (Vallabhajosula, 2007). Radionuclides are administered to patients, where they decay by positron emission, which results in the release of two gamma rays. Gamma emission is detected by detectors placed around the subject and images obtained showing the distribution of radionuclide in the body. Improvements in anatomical localization is achieved by combining images obtained by PET with concurrent computer tomographic (CT) images obtained prior to scanning.

For a radiotracer to be developed further, a number of characteristics need to be fulfilled during the radiosynthesis phase (Pimlott, 2005). It must be possible to:

• Obtain sufficient yields (5–40% non-decay corrected, end of synthesis yields, depending on synthesis procedure) and rapid radiosynthesis. This optimization is particularly crucial in the development of a radiotracer for early clinical studies and may affect the cost of the radiotracer and logistics of these studies.

• Obtain high radiochemical purity adequate for human imaging studies (preferably 95% or higher).

• Obtain high specific radioactivity to allow injection of sufficient radioactivity into patients for imaging. The amount of compound present in a patient dose should be well below the amount that elicits a pharmacological response.

• Obtain the radiopharmaceutical in a form that is acceptable to regulatory bodies with respect to impurity profile [European Medicines Agency (EMEA) or Food and Drug Administration (FDA)].

Once radiosynthesis has successfully been developed, a standard operating procedure (SOP) can be initiated for the synthesis of the radioligand, and to ensure that the radiopharmaceutical is sterile and endotoxin-free (Figure 1). The SOP will detail in full the synthesis and specifications for production and purity analysis of these tracers so that they are synthesized in a reproducible and reliable manner. The SOP will also indicate how scanning will be conducted. SOPs are central to both preclinical and clinical validation of a novel radiotracer before it is considered to be useful for imaging the target of interest. After synthesis, radiolabelled compounds must then undergo pharmacological characterization, initially in cell culture, and then in vivo to determine the biological action and affinity for the tumour target (Figure 1), a procedure that will be discussed in more detail later.

Production and quality control

The European Community Directive on Good Clinical Practice in clinical trials requires that the manufacture of investigational medicinal products be in compliance with Good Manufacturing Practice (GMP) standards within approved facilities, and that these facilities must be capable of meeting these standards. Within the UK, facilities involved in the production of radiopharmaceuticals must apply for a manufacturing license issued by the Medicines and Healthcare Products Regulatory Agency (MHRA), and these sites are regularly monitored. In the United States, the FDA acts as an equivalent regulatory body and is responsible for issuing guidance notes and regulations, and a summary of the regulations involved in radiotracer development is given in Table 2.

Table 2.

Regulatory guidelines pertaining to the manufacture and clinical use of radioligands

Body	Locality	Regulation name	Process	Type of regulation
Medicines and Healthcare products Regulatory Agency (MHRA)	UK	Medicines Control Agency (MCA) ‘Specials’ License (MHRA, 2007)	Ensures manufacture of investigational medicinal products are in compliance with Good Manufacturing Practice (GMP) standards	Legislative and binding
			Issue guidance notes and regulations	Non-binding
		Investigational Medicinal Product ‘IMP’ or equivalent (UK ‘Specials’) (MHRA, 2010)	Ability to undertake human studies with medicinal products	Legislative and binding
Food and Drug Administration (FDA)	USA		Issue guidance notes and regulations	Non-binding
		Radioactive Drug Research Committee (RDRC) approval mechanism: Investigational new drug (IND)	Safety evaluation for entirely new radiolabelled compounds for which there has been no prior human exposure	Legislative and binding
		Radioactive Drug Research Committee (RDRC) approval mechanism: Exploratory Investigational New Drug (eIND)	Safety evaluation for situations where there has been some prior human exposure to the radiolabelled compound	Legislative and binding
		Critical Path Initiative	Action plan to assist investigators along the approval path towards IND or eIND	Encourage investigators to submit new tracers for approval. Non-binding
Radiopharmacy Committee of European Association of Nuclear Medicine (EANM)	EU (incl UK)	‘Guidelines on Good Radiopharmacy Practice (GRPP)’ (Committee, 2007)	Production of radiopharmaceuticals, specific GMP issues such as compliance to GRPP, self-inspection, acceptance criteria and ensuring appropriate staff qualifications.	Non-binding
Administration of Radioactive Substances Advisory Committee (ARSAC)	UK	Certificate for the administration of radioactive medicinal products	Certification allows specific radiolabelled compounds of pre-defined effective dose to be administered to study subjects for a specified time interval	Legislative and binding

It is not always feasible to perform detailed quality control tests prior to administration of radionuclides, particularly as the half-life of the majority of radionuclides is short. Therefore, an ongoing quality assurance system is often established prior to initial clinical studies to ensure that quality analysis occurs. For radioligands with a high specific radioactivity, it is recommended to evaluate the stability of the ligand during test runs for up to 4 h (for carbon-11 and fluorine-18) after radiosynthesis to ensure that changes in the chemical composition of the radioligand has not occurred.

A number of in-house quality control procedures need to be assessed prior to administration of the radionuclide to the patient. These include assessment of purity, nature of any impurities, sterility of the solution, absence of pyrogens and suitable pH. With regards to impurities generated during the radiosynthesis, it should be emphasized that the parent compound is given at only tracer doses, and hence exposure to minor impurities is negligible. Minor impurities, however, are still considered and limits set; (EMEA, 2009). Consideration should also be given to levels of the other components in the formulation, such as anti-oxidants and residual solvents. These concepts should all be outlined within the SOP, which are adhered to by the manufacturing unit and must be readily available for audit purposes. At many centres a simple checklist is used to ensure that the radiopharmaceutical released adheres to strict criteria.

Chemical identity and purity

Prior to the administration of the radiopharmaceutical to patients it is essential, as part of the developmental process, to provide evidence that the radiopharmaceutical prepared via the radiosynthetic route is identical to the non-radiolabelled reference compound. This may not be the case for example with fluorinated compounds such as [¹⁸F]FDG, where there are no known biological equivalent. If the radiosynthetic route differs from that of the non-radiolabelled compound then the radiosynthetic route should be documented and be available for review by an independent body. Following decay of the radiolabelled compound, the chemical identity should be determined by mass spectrometry and nuclear magnetic resonance, if possible. The purity of the radiochemical, radionuclide and radiopharmaceutical should be determined by HPLC-UV and HPLC-mass spectrometry. It is common to perform three to five runs of the radiosynthetic process once this has been determined to ensure the robustness and reproducibility of the manufacturing process and to ensure that the compound meets all clinical criteria. During this process batch numbers should be recorded including manufacturing records of the radiosynthesis and quality control analyses.

In vitro and in vivo studies

Prior to testing in human subjects, novel radiopharmaceuticals must undergo in vitro and in vivo validation in order to ascertain target specificity, biodistribution and toxicity. Radiotracers are initially assessed in vitro to assess target specificity, uptake and the impact on cellular responses. Previously, radioligands would then be transferred into animal models, now however, a number of authors have suggested the use of multicellular tumour spheroids (MTS) as a transitional model, as the cytology and the morphology of MTS resembles that of experimental tumours in mice (Fracasso and Colombatti, 2000; Kunz-Schughart et al., 2004). Therefore, the use of MTS would allow evaluation of dose and dosing regimen with less laboratory effort. Preclinical studies have shown MTS to be an appropriate in vitro system to investigate drug response to chemotherapy (Bates et al. 2000; Dubessy et al., 2000; Mueller-Klieser, 2000). A number of MTS of differing tumour types have been evaluated and shown to be good experimental models in which to investigate novel agents and PET imaging as a biomarker of response (Monazzam et al., 2006; 2007; 2009;). However, MTS does not substitute the animal systems.

Animal studies are performed as part of the development of novel radioligands to determine biodistribution, reproductive toxicology, carcinogenic toxicity and genotoxicity. An integral part of early radiotracer development is a biodistribution study. These studies are often the first conducted and are important in determining whether binding of the radioisotope is tumour specific or if it binds to other sites outside the tumour, that is, non-specific binding. High non-specific binding of a radiotracer is a common reason for the failure of a radiotracer to reach clinical studies. Furthermore, knowing the degree of non-specific binding will assist in calculating the absorbed dose by tissues (dosimetry). Toxicology and biodistribution studies assure that the radiopharmaceutical is non-toxic at doses to be administered to patients.

A key aspect of the in vivo assessment of radioligands for use in oncology is selection of appropriate tumour models. This is of particular importance with reference to biodistribution studies; if the tumour model does not express a receptor of interest for example then any proof of concept study will be negative. In a study by Ullrich and colleagues [¹⁸F]fluorothymidine ([¹⁸]FLT) was investigated as an early response marker to erlotinib, an epidermal growth factor receptor (EGFR) inhibitor in animal models of non-small cell lung cancer (Ullrich et al., 2008). The authors report a reduction in [¹⁸]FLT by PET in erlotinib-sensitive EGFR-dependent tumours; in contrast [¹⁸]FLT-PET changes were attenuated in tumours harbouring T790M secondary EGFR mutations that confer resistance to EGFR tyrosine kinase inhibitors. Similarly, in a study investigating [¹⁸]FLT-PET as a marker of response to mTOR inhibitor, rapamycin, reductions in [¹⁸]FLT-PET were not seen in PTEN wild-type resistant tumours but occurred in sensitive cell types (with PTEN loss), again highlighting the need for the target of interest to be present for proof of concept studies (Wei et al., 2008 p. 140).

Carcinogenicity studies are not required unless there is a specific concern. This is particularly true for tracers being developed for use in oncology where, in general, patients have limited life expectancy. In these patients, long-term repeat dose toxicology, genotoxicity and carcinogenicity studies are usually omitted. Similarly, reproductive toxicology studies are not required as long as patients of reproductive age will be advised to take appropriate precautions to avoid pregnancy, and pregnant and lactating patients will be excluded from any clinical studies.

In terms of toxicology studies, single-dose studies are performed in animals using the formulation to be used in patients. Single-dose toxicology studies are deemed appropriate in animals, as in the clinical setting, patients will usually receive only a single, low dose of the radioligand, and if repeated administrations are delivered, the doses are usually administered weeks apart and, from a toxicological standpoint, are considered equivalent to a single dose. All animal toxicology studies are performed under the highest quality standards possible [equivalent to Good Laboratory Practice (GLP)].

Initial dosimetry studies are performed in animals over a number of time points depending on the biologicalal and physical half-lives of the radioisotope. The radioactivity in blood, urine and tissues, expressed as a percentage of injected radioactivity, is determined to enable calculation of the predicted human effective radiation dose (mSv) to each tissue. Human dosimetry often tends to be the first human study of new radiotracers. The dose to critical organs, such as the bladder and liver, can limit the amount of radioactivity that can be injected in human studies and therefore has the potential to limit the success of the tracer.

PK studies, mainly in rodents but where possible in larger animals, play important roles in the prediction of human radiation organ dosimetry and toxicokinetics (Figure 3). This is an important concept as metabolism can affect not only the amount of tracer that is able to bind to the target, but the generation of radiolabelled metabolites may interfere with image analysis. Furthermore, rapid metabolism of the radiotracer during the imaging time frame can limit the potential usefulness of a tracer, and thus needs to be considered.

Figure 3.

Use of [¹⁸F]FLT in positron emission tomography (PET) studies (A). Transverse [¹⁸F]FLT-PET images of mice bearing the SKMEL-28 human melanoma tumour, 10 days after inoculation with either vehicle or PD0325901 treatment. (B) [¹⁸F]FLT-PET time (min) versus radioactivity curve [as measured by normalized uptake value at 60 min postinjection (NUV₆₀)] for the vehicle or PD0325901 treatment group (courtesy of Julius Leyton).

The potential immunogenicity of the radiopharmaceuticals is also of importance, particularly if the tracer is to be administered more than once to the same patient. This issue should be addressed by inspecting the structures of compounds for any possible alerts for immunogenicity by analogy with related compounds, for example, the presence of peptides or the presence of a penicilloic acid moiety.

Once animal toxicology and biodistribution studies are complete and the tracer is deemed suitable for clinical use from a scientific perspective, the radiotracer must undergo validation in the clinical setting. There are a number of regulatory bodies that manage the clinical development of radioligands to eusure that trials are run in a safe and consistent manner. These agencies tend to have similar remits in various countries, and the approval process is outlined in Table 2.

Role of imaging in early clinical trials

Pre-phase I or phase 0 studies

In 2003, the EMEA published a concept note followed by a Position Paper on the non-clinical safety studies required to support human clinical trials with a single dose of pharmacologically active compound using microdose techniques (http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2009/09/WC500002941.pdf). These guidelines affect PET in two ways: firstly in terms of the development process of radioligands, but also use of PET as clinical trial end-points (Marchetti and Schellens, 2007). As sub-pharmacological single doses of the investigational agent are administered, ultrasensitive analytical methods such as PET are required to assess the drug distribution and PK. The paper raises the possibility of a reduced preclinical safety package for microdose studies and facilitates the development of novel radioligands (VanBrocklin, 2008).

As there has been a significant decline in the number of new molecules submitted to the FDA over the past decade, the FDA responded to this by devising an action plan called the ‘Critical Path Initiative’ to facilitate the drug discovery and development process from inception to approval (FDA, 2004; 2010;). A recent initiative implemented by the FDA is the establishment of the Exploratory Investigational New Drug (eIND) application, a characteristic of which are microdosing/phase 0 studies. These studies involve the administration of a single ‘micro-dose’ or ‘tracer-dose’ of a novel agent or radioligand, either to healthy volunteers or patients. The aim of a phase 0 study is to obtain information on PK and/or distribution properties or receptor selectivity prior to phase I. The studies do not provide information about the safety or tolerability of the radiopharmaceutical. The definition of a microdose is ‘less than 1/100 of the dose of a test substance calculated (based on animal data) to yield a pharmacological effect of the test substance with a maximum dose of ≤100 µg or, in the case of biological agents, ≤30 nmol’ (CHMP, 2004). The translational research questions that justify a pre-phase I study include whether the agent distributes to the tumour, to what extent the agent distributes to normal tissues, and whether the metabolite profile in rodents and humans are similar. These studies are used to establish whether the radiopharmaceutical behaves in human subjects as expected from preclinical studies. Theoretically, microdose trials should allow an early discrimination between those lead compounds to take forward into further development and those to be discarded. The biodistribution of a compound is also important to determine the extent of non-specific binding. Pre-phase I studies can also yield information regarding PK modulation and scheduling. N-[2-(dimethylamino)ethyl]acridine-4-carboxamide (DACA) is a DNA intercalating acridine derivative that was labelled with carbon-11 in order to determine the metabolic profile of DACA, the distribution to the tumour – as a predictive marker of response – and distribution to normal tissues – as a predictor of toxicity (Saleem et al., 2008). Twenty-four patients were enrolled in the study and [¹¹C]DACA administered at 1/1000 of the phase I dose. This PK study illustrated that [¹¹C]DACA undergoes extensive metabolism and its metabolites distributed well to the tumour. Low peak concentrations of [¹¹C]DACA-derived radioactivity were obtained in the vertebrae and brain compared with other normal tissues suggesting that myelosuppression and neurotoxicity were unlikely to be dose-limiting toxicities. There was uptake within the myocardium suggesting the possibility of myocardial toxicity in the phase I study. In a similar study LY2181308, an antisense oligonucloetide to survivin, was labelled with carbon-11 and administered concurrently with therapeutic doses of unlabelled drug (Saleem et al., 2009). [¹¹C]LY2181308 uptake in both normal tissue and tumour was assessed. The authors demonstrated the utility of microdosing experiments illustrating that active drug concentrations could be achieved within the tumour and that unlabelled LY2181308 saturated normal tissue kinetics resulting in increased tumour uptake. These studies establish the feasibility of performing PET studies with radiolabelled anticancer drugs in a phase 0 setting.

PET can also been utilized to provide proof of principle for mechanism of action of the labelled drug. The effect of eniluracil, a drug that inhibits dihydropyrimidine dehydrogenase, an enzyme responsible for the degradation of 5-fluorouracil (5-FU), was studied using PET (Ahmed et al., 1999; Saleem et al., 2000a,b;). The study illustrated that in eniluracil-naïve patients, [¹⁸F]FU-derived radioactivity localized more strongly to the normal liver than in metastases. In addition there was distinct localization of radioactivity in the gall bladder consistent with hepatobiliary clearance of [¹⁸F]FBAL that disappeared in the presence of eniluracil. In patients receiving eniluracil, there was increased plasma uracil and unmetabolized [¹⁸F]FU. There was also an increase in the half-life of [¹⁸F]FU-derived radioactivity in tumours, providing proof of principle for mechanism of action of eniluracil. More recently Rosso and collegues utilized a PET imaging approach to generate kinetics of temozolamide following administration of labelled drug to patients with gliomas. The authors illustrate that the utilization of impulse response function to predict drug exposures in both tumours and healthy tissue, assuming linear drug kinetics (Rosso et al., 2009).

Following these studies a novel radioligand will undergo formal clinical evaluation prior to licensing for clinical use. A list of PET tracers currently used in the clinical setting, either as research tools or for diagnostic purposes, is given in Table 3.

Table 3.

Common PET tracers used in oncology

Isotope	Radiotracer	Target	Measured effect
11C	[11C]choline	Phosphocholine	Choline kinase activity, membrane function
	l-Methyl-[11C]methionine	Amino acid transporter	Protein synthesis, amino acid transport, tumour cell growth
	L-1-[11C]tyrosine	Amino acid uptake	Amino acid metabolism
	2-[11C]thymidine	DNA	DNA synthesis. Tumour cell proliferation
	N-methyl-[11C]temozolamide, Carbonly-[11C]temozolamide, N-[2-(dimethylamino)ethyl]acridine-4-carboxamide, [11C]-1,3-bis-(2-chloroethyl)-1-nitrosourea		Drug pharmacokinetics
18F	[18F]fluorodeoxyglucose	Glucose metabolism	Glucose transport. Aerobic and anaerobic glycolysis, glucose consumption or metabolism
	5-[18F]fluorouracil, [18F]fluorotamoxifen		Pharmacokinetics
	[18F]fluoride	Incorporation in the hydroxyapatite crystals in bone	Bone metabolism; Bone blood flow and osteoblastic activity
	1-[11C]acetate	Tricarboxylic acid cycle via acetyl coenzyme A	Lipid synthesis
	[18F]fluoroacetate
	3-[18F]fluorothymidine	Activity of thymidine kinase-1	DNA synthesis. Tumour cells proliferation
	[18F]fluoroazomycin arabinoside, [18F]fluoromisonidazole	Hypoxic cells	Tumour hypoxia
	16α-[18F]fluoro-17β-oestradiol	Oestrogen receptors in breast cancer	Receptor binding, response to anti-oestrogen therapy
	16β-[18F]-5α-dihydrotestosterone (FDHT)	Androgen receptor (ARs)	Functional status of tumour ARs
	[18F]AH111585/fluciclatide, [18F]Galacto-RGD	Angiogenesis	Integrin receptors (αVβ3) on endothelial cells of neovasculature
	[18F]oligonucleotide	Gene expression	In vivo hybridization with mRNA
	9-1-[18F]fluoro-3-hyroxy-3-propoxy- methylguanine	Gene expression	Substrate to herpes virus thymidine kinase
68Ga	[68Ga]DOTATOC, [68Ga]DOTANOC	Specific binding to somatostatin receptor SSTR-II; SSTR-II, III, V	Presence of SSTRII; SSTRII, III, IV
124I	Na[124I]	Gene expression	Sodium iodide symporter expression
15O	[15O]H2O	Perfusion	Blood flow/perfusion
	[15O]CO	Blood volume	Antivascular response
others	124I-, 64Cu-, 86Y-Labeled antibodies	Binding to tumour antigens	Specific binding to tumour associated antigenic binding sites (such as CEA, PSMA, CD20 and CD22)
	[124I]Annexin V, [64Cu]Annexin V, [18F]-Annexin V	Apoptosis	Specific binding to Phosphatidylserine (PS) on cell membrane

Drug PKs

As discussed earlier, a successful imaging agent must be retained by the target but must not affect the target, nor can it be too toxic for use with other treatment agents. In contrast, therapeutic agents are designed to interact and affect the target of interest. Labelling drugs may therefore be useful in understanding the PK of an agent particularly as part of the drug development process (Figure 4). Parameters including drug absorption, distribution, metabolism and excretion, as well as the maximum plasma concentration and area under the plasma concentration versus time curve are major determinants of toxicity and efficacy of a drug, and can be achieved by labelling drugs of interest with radioisotopes such as carbon-11, fluorine-18 or nitrogen-13 (Osman et al., 2001; Saleem et al., 2001; 2008;). PK parameters are obtained through mathematical kinetic modelling, a process that involves comparing the concentrations of radioactivity in tissues with that in arterial plasma and calculating the plasma-tissue exchange rates to derive tissue PK parameters (Saleem et al., 2000a,b;). This allows the investigator to ascertain what happens to the tracer within the body and, in particular, within the target organ.

Figure 4.

Uptake of tracers varies between normal and tumour tissue. (A, B) Time activity curves demonstrating normal brain and tumour uptake of [¹¹Cmethyl]temozolomide. (C, D) Time activity curves demonstrating normal liver and tumour uptake of [¹⁸F]5-fluorouracil.

The interpatient variability in the PK of anticancer drugs is generally high and these drugs have a narrow therapeutic index. Steady state plasma PK often poorly reflect the levels of drug within normal or tumour tissues. PET PK studies may aid in determining intra-tumour drug exposure. For example when comparing uptake of [¹¹C]-1,3-bis-(2-chloroethyl)-1-nitrosourea (BCNU) within the tumour after both intravenous and intra-arterial administration it was shown that intra-arterial administration resulted in 50 times higher intra-tumour concentrations (Tyler et al., 1986). Examples of drugs that have been evaluated by PET include [¹¹C]temozolamide, [¹⁸F]5FU, [¹⁸F]paclitaxel, [¹³N]cisplatin and [¹¹C]DACA (Ginos et al., 1987; Kissel et al., 1997; Dimitrakopoulou-Strauss et al., 1998; Saleem et al., 2001; 2003; Kurdziel et al., 2003). PK studies may also allow for the investigation of the mechanism of action of novel therapies prior to phase I. For example, the proposed mechanism of action of temozolamide was studied in patients using [¹¹C]temozolmide tracers labelled at the methyl and carbonyl positions. The study demonstrated that temozolamide undergoes tissue-specific ring-opening in humans, although this conformational change was not tumour specific (Saleem et al., 2003).

On labelling the drug it is important to ensure that no modification has occurred to the molecule, as even a slight change can alter its physicochemical properties. Initial exploratory studies of hydrogen-3 (tritium) and carbon-14 labelled compounds in animals are often informative in determining if imaging in patients will be useful. For example, if a large number of degradation products are rapidly generated, imaging may add little value in determining response. On the other hand, imaging the distribution of a drug may provide valuable insight into further development and choice of tumours to be treated.

The PK of 5-FU has been successfully studied by labelling the drug with ¹⁸F, allowing enough time for kinetic studies to be performed with good reproducibility (Harte et al., 1998). Up to 80% of 5-FU is degraded through catabolism. A PK study in 13 patients with colorectal liver metastases showed a relationship between uptake of 5-FU and tumour response; the likelihood of disease stabilization was higher and survival was longer in patients whose tumours had a higher uptake of the radiolabelled drug (Moehler et al., 1998).

When used in a PK context, labelled macromolecules, for example, antibodies that target cell surface receptors may show specific targeting to receptors as well as significant non-selective distribution in intracellular and extracellular pools related in part to their slow systemic clearance (Verel et al., 2005). This characteristic of radiolabelled macromolecules poses a challenge when they are used in a PD context. In the case of IgG entities, this requires labelling with long-lived radioisotopes, for example, copper-64 and imaging typically at 24–48 h.

PDs

PD endpoints in PET imaging can be considered as either generic or specific end-points. Generic end-points image physiologic processes such as cellular proliferation, glucose metabolism, apoptosis, hypoxia and angiogenesis. Imaging of specific end-points provides proof of principle for a proposed mechanism of action such as the use of 16α-¹⁸F-fluroro-17β-oestradiol to assess oestrogen receptor status in breast cancer (Lee et al., 2009). We will discuss a number of these probes that are currently undergoing preclinical and clinical development or in routine clinical use.

Tumour metabolism and proliferation

[¹⁸F]FDG is the commonest clinically available tracer and assesses the generic end-point of glucose utilization, as a measure of metabolism. Despite the increasing number of potential radiotracers, PET is generally still synonymous with [¹⁸F]FDG. Metabolic imaging with [¹⁸F]FDG relies on differential utilization of glucose in cancer cells compared with normal cells (Kelloff et al., 2005). [¹⁸F]FDG enters the cell and is phosphorylated by hexokinase to [¹⁸F]FDG-6-phosphate. [¹⁸F]FDG-6-phosphate cannot undergo any further metabolism and is trapped within the cell, with the rate of accumulation of [¹⁸F]FDG-6-phosphate being proportional to the rate of glucose utilization. Therefore, the more metabolically active the cell the more glucose is utilized. The major limitation with [¹⁸F]FDG is the false positivity seen within areas of inflammation (Subhas et al. 2005; Rosenbaum et al., 2006). Furthermore, as [¹⁸F]FDG relies on increased cellular proliferation compared with surrounding tissues, it has low sensitivity in slow growing, well-differentiated and less metabolically active tumours (Shreve et al. 1999; Burton et al., 2004). Moreover, [¹⁸F]FDG distributes within the renal collecting system obscuring pelvic lesions.

[¹⁸F]FDG has been extensively studied both for the staging of tumours and response to cytotoxic therapy having an overall average sensitivity of 84% and a specificity of 88% in cancer (Gambhir et al., 2001). More recently [¹⁸F]FDG has been used to monitor responses to targeted therapies particular those that affect glucose homeostasis such as the mTOR and Akt inhibitors (Contractor and Aboagye, 2009) (Table 4). The most impressive results with [¹⁸F]FDG were seen in the early trials of imatinib mesylate, a c-kit inhibitor used in the management of gastrointestinal stromal tumours (GISTs). In these tumours, response on [¹⁸F]FDG-PET was seen as early as 24 h after therapy, with the response on PET imaging correlated with progression-free survival (Stroobants et al., 2003). This is of particular importance in early phase clinical trials as standard radiological assessment of these tumours often fails to show any change in tumour size. Currently [¹⁸F]FDG is approved for the staging and response assessment of a number of tumour types. A number of guidelines have been developed for the assessment of response including the European Organization for Research and Treatment of Cancer (EORTC) and the National Cancer Institute guidelines (Young et al. 1999; Shankar et al., 2006). More recently the PERCIST guidelines were suggested by Wahl and colleagues but remain to be adopted (Wahl et al., 2009).

Table 4.

Selected clinical studies of the use of [¹⁸F]FDG for monitoring therapy response

Tumour type	Therapy	Change in FDG uptake with therapy	Reference
Lymphoma		Reduction after 7 days correlates with progression free survival and overall survival	Romer et al., 1998
		Reduction after two cycles correlation with complete response, progression free survival and overall survival	Haioun et al., 2005
Gastrointestinal stromal tumour	Imatinib	Reduction after 7 days; correlated with response	van Oosterom et al., 2001
	Imatinib	Reduction at 1 month correlated with response	Holdsworth et al., 2007
Biliary tract cancer	Bevacizumab, gemcitabine and oxaliplatin	Reduction at 6 weeks; correlation with progression-free survival and overall survival	Zhu et al., 2010
Non-small cell lung cancer	Gefitinib	Reduction on day 2, associated with improved progression free survival	Sunaga et al., 2008
Breast cancer	Cyclophosphamide, doxorubicin, vincristine, and prednisolone or docetaxel	Reduction after first cycle predictive of response	Smith et al., 2000
	Epirubicin, cyclophosphamide or epirubicin, paclitaxel	Reduction after first cycle predictive of response	Schelling et al., 2000
	Lapatinib	Reduction after 1 month correlated with response	Kawada et al., 2007
Prostate cancer	Goserelin	Reduction 1 after−5 months of therapy	Oyama et al., 2001
Gastric cancer	Cetuximab, irinotecan, 5-fluorouracil	Reduction correlated with response, time to progression and overall survival	Di Fabio et al., 2007
Gastro-oesophageal cancer	Epirubicin, cisplatin and 5-fluorouracil	Reduction at 14 days correlates with response and overall survival	Ott et al., 2006

Proliferation

Proliferation is one of the hallmarks of malignancy (Hanahan and Weinberg, 2000) and the majority of chemotherapeutic agents act to inhibit proliferation of cancer cells. It has been shown in a number of tumour types that the rate of proliferation, as indicated by Ki-67 and PCNA expression and S-phase fraction of tumour samples, is predictive of both survival and prognosis (Stuart-Harris et al., 2008; Yerushalmi et al., 2010) There are inherent difficulties in patient acceptability of routine biopsies, particularly if performed serially. Furthermore, there may be issues of tumour accessibility and, because of tumour heterogeneity, the biopsy sample may not be representative of the proliferative activity of the whole tumour. There is a need therefore to assess tumour proliferation, non-invasively.

Thymidine is a nucleoside utilized in DNA replication in proliferating cells, and both thymidine and its analogues have been extensively studied as potential tracers of cellular proliferation. Thymidine was first labelled with tritium, which is still used for in vitro work. During the process of development, thymidine was labelled with carbon-11. However, [¹¹C]thymidine undergoes rapid metabolism resulting in significant levels of labelled metabolites within the blood (Shields et al., 1992). The fluorinated analogue, [¹⁸F]FLT, has a suitable nuclide half-life (109.8 min) for clinical use and stability against phosphorylase-mediated degradation in vivo and is now extensively used as a marker of proliferation (Shields et al., 1998).

After injection, [¹⁸F]FLT enters the cell by both Na⁺-dependent active nucleoside transporters and by passive diffusion. [¹⁸F]FLT follows the salvage pathway of DNA synthesis and like thymidine undergoes phosphorylation by thymidine kinase–1 (TK1) to [¹⁸F]FLT-monophosphate (Seitz et al., 2002). [¹⁸F]FLT is a selective substrate for TK1 whereas thymidine is also phosphorylated by TK2. TK1 is virtually absent in quiescent cells but is increased in proliferating cells (Sherley and Kelly, 1988; Munch-Petersen et al., 1995). Phosphorylated [¹⁸F]FLT is not significantly incorporated into DNA and is trapped within the cytosol. The rate limiting step for [¹⁸F]FLT accumulation is the initial phosphorylation by TK1; it is also the rate limiting step in the salvage pathway of DNA synthesis, therefore the handling of [¹⁸F]FLT reflects cellular proliferation (Shields et al., 1998).

In a paper by Barthel and colleagues the authors investigated the ability of [¹⁸F]FLT-PET to monitor response to chemotherapy in mice bearing sarcomas injected with 5-FU (Barthel et al., 2003). The authors reported reduced uptake of [¹⁸F]FLT following treatment that correlated with the PCNA labelling index within the tumour and a reduction in tumour volume. The authors then considered the role of TK1 in [¹⁸F]FLT-PET imaging by contrasting the uptake of [¹⁸F]FLT in TK1 knock-out and functional heterozygote lymphoma models (Barthel et al., 2005). They reported higher uptake in TK1 heterozygote lymphoma compared with the knock-out model, which correlated with higher TK1 protein expression within the tumours and accumulation of [¹⁸F]FLT. These studies were then reproduced in a clinical setting. In the seminal paper by Kenny and colleagues, [¹⁸F]FLT-PET was reported as being an early response marker in patients receiving chemotherapy for breast cancer when compared with RECIST assessment (Figure 5) (Kenny et al., 2005; 2007;). Furthermore, the [¹⁸F]FLT uptake parameters within the breast tumours correlated with Ki-67 expression within surgical resection specimens lending further support of [¹⁸F]FLT as a PD marker for early clinical response.

Figure 5.

[¹⁸F]FLT-positron emission tomography images to assess early response (1 week) to epirubicin, cyclophosphamide and 5-fluorouracil chemotherapy in a patient with T3N0M0 breast cancer. (A) Pretreatment uptake is seen in the right breast and (B) post-treatment response is observable as a reduction of [¹⁸F]FLT uptake (courtesy of Dr. Laura Kenny).

However, [¹⁸F]FLT has several limitations (Barwick et al., 2009). Firstly, [¹⁸F]FLT uptake in tumours depends on TK1 activity, including therapy-induced activation of the salvage pathways (Perumal et al., 2006; Pillai et al., 2008; Kenny et al., 2009). Therefore, agents that only inhibit de novo DNA synthesis such as fluoropyrimidines may increase FLT uptake. Other therapeutic agents inhibit both pathways and subsequently reduce [¹⁸F]FLT uptake. Moreover, tumours with low TK1 expression or low proliferative activity may not be readily identified (Bading and Shields, 2008). High physiological uptake of [¹⁸F]FLT in liver and bone marrow also limits the assessment of disease in these organs. However, the recent introduction of kinetic filtering has led to significant signal reduction from the liver, improving visualization (Gray et al., 2010). There is evidence to suggest that membrane localization of nucleoside transporters can be enhanced in tumours in response to a number of chemotherapy agents, which may result in a ‘flare’ response with [¹⁸F]FLT. Perumal and colleagues investigated the role of [¹⁸F]FLT-PET in monitoring intra-tumour inhibition of thymidylate synthase in vivo 1–2 h following 5-FU injection into mice bearing RIF-1 fibrosarcomas (Perumal et al., 2006). The authors reported an increase in the 60 min [¹⁸F]FLT tumour/heart radioactivity ratio in drug treated mice compared with vehicle controls. Plasma and tumour deoxyuridine levels increased significantly, but TK1 expression and ATP remained unchanged. Cellular assays for membrane uptake suggested a functional role of type-1 equilibrative nucleoside transporter (ENT-1). The authors hypothesized that [¹⁸F]FLT uptake is mediated by redistribution of ENT-1 transporters to the plasma membrane 1–2 h after treatment. This was utilized clinically by Kenny and colleagues in breast cancer patients receiving capecitabine therapy where the authors suggested that PET retention parameters obtained following capecitabine therapy could be utilized as a PD marker for TS inhibitors (Kenny et al., 2009). There have been a number of studies considering the uptake of [¹⁸F]FLT in both the preclinical (Table 5) and clinical (Table 6) setting, with differing results according to the type of therapy utilized and at differing time points, both of which need to be considered in the utilization of [¹⁸F]FLT as a response marker.

Table 5.

Preclinical studies of the use of [¹⁸F]FLT for monitoring therapy response

Table 6.

Clinical studies of the use of [¹⁸F]FLT for monitoring therapy response

Tumour type	Therapy	Change in FLT uptake with therapy	Reference
Lung	Gefitinib	Reduction after 7 days; correlated with response	Sohn et al., 2008
	Chemo/radiotherapy	Reduction at day 2, 8 and 29	Everitt et al., 2009
Gliomas	Bevacizumab and irinotecan	Reduction at 6 weeks; correlation with progression-free survival	Chen et al., 2007
	Temozolamide	Correlation with tumour response	Galldiks et al., 2010
Germ cell tumours	Combination chemotherapy	Reduction; no association with response	Pfannenberg et al., 2010
Non-Hodgkin's lymphoma	CHOP, R-CHOP, rituximab	Reduction after 2 days; correlated with response. No change with rituximab	Herrmann et al., 2007
Breast cancer	Various chemotherapy and hormone therapy	Reduction at 2 weeks; correlated with response	Pio et al., 2006
	Combination 5-FU, epirubicin, cyclophosphamide	Reduction at 1 week; correlated with response	Kenny et al., 2007
Colorectal	Combination 5-FU and radiotherapy	Reduction at 2 weeks; no correlation with response	Wieder et al., 2007
Oropharyngeal cancer	Radiotherapy	Reduction at 1 week	Troost et al., 2010
	Radiotherapy	Increase 2–3 months compared with baseline; correlation with presence of disease	Been et al., 2009
Melanoma	CTLA4 blockade	No change in tumour uptake	Ribas et al., 2010
Colon (COLO205 cells)	Checkpoint kinase 1 (Chk1)inhibitor and docetaxel	Reduction with docetaxel and combination treatment at day 2 + 16	Zhang et al., 2009

The choline pathway can also be imaged as a marker of cell growth, and clinical response biomarker in malignancy. Phosphocholine is a biosynthetic product of choline kinase and is a precursor for membrane phosphatidylcholines, a class of phospholipids and a major component of biological membranes. The expression and activity of phosphocholine is up-regulated in a number of tumour types including prostate, lung, breast and colon cancer (Ramirez de Molina et al., 2002; 2007; Glunde et al., 2005). This increase in phosphocholine levels seen in malignancy is largely attributable to the up-regulation of choline kinase that occurs with malignant transformation. However, increased phosphocholine levels may also be due to increased breakdown via phospholipase C (Glunde et al., 2004). Choline can be radiolabelled with either carbon-11 ([¹¹C]choline) or flurorine-18 ([¹⁸F]fluorocholine, or [¹⁸F]FCH). As [¹¹C]Choline has reduced excretion by the urinary collecting system it has been used clinically in the imaging of pelvic tumours.

[¹¹C]Choline has been most extensively used in the assessment of nodal disease in prostate cancer (Hara et al., 1998; de Jong et al., 2003; Reske et al., 2006; Scher et al., 2007). However, the utility of [¹¹C]choline as an early response biomarker for the management of prostate cancer is yet to be established. Promising preclinical work by Krause and colleagues with [¹¹C]choline showed utility for early therapeutic response to docetaxel in a prostate cancer xenograft model (Krause et al. 2010). Kenny and colleagues report on a subgroup of patients treated with trastuzumab who had early responses measured by [¹¹C]choline-PET (Kenny et al., 2010). These findings require further investigation. [¹⁸F]-fluoromethyl-choline (FCH) has the advantage of longer half-life compared with [¹¹C]choline, but these tracers undergo oxidation, with metabolites being detected within the plasma soon after administration. Thus the relative contribution of both the parent compound and the metabolites are difficult to ascertain when a late imaging protocol is used. Leyton and colleagues have recently reported data with a deuterated analogue,[¹⁸F]-fluoromethyl-[1,2-²H₄]-choline (D4-[¹⁸F]FCH), which exhibits greater in vivo stability, and sensitivity particularly for late tumour imaging. Moreover, the authors report that D4-[¹⁸F]FCH was a biomarker of response in mice treated with the mitogenic inhibitor of the extracellular signal-regulated kinase, PD0325901 (Leyton et al., 2009).

Apoptosis

Apoptosis, or programmed cell death, is the likely mechanism behind the tumourcidal effects of both standard chemotherapeutic agents and many novel agents (Hanahan and Weinberg, 2000). An early event in apoptosis involves phosphatidylserine, which is normally confined to the cytoplasmic leaflet of the cell translocating to the external leaflet of the cell membrane. This has been exploited for imaging by labelling annexin V, an endogenous protein that has a high affinity for membrane bound phosphatidylderine, for imaging of apoptosis in vivo (Blankenberg et al., 1998). Annexin V was initially labelled with technetium-99 m and utilized as a SPECT radionuclide, to detect cardiac thrombosis in patients with atrial fibrillation (Blankenberg et al., 1999; Hofstra et al., 2000). SPECT imaging differs from PET imaging in that it involves the administration of a gamma-emitting tracer. An improved labelling methodology resulted in the development of [^99mTc]-HYNIC-annexin V, which has been used in the assessment of lung cancer, where a patient responding to therapy had an increased uptake of the radiotracer (Blankenberg, 2008). Other annexin V-based radioisotopes with longer half-lives have been developed including iodine-124 with the specific advantage that its long half-life allows the characterization of longer-term biologicalal processes, and [¹⁸F]-annexin V (Glaser et al., 2003; Yagle et al., 2005).

There is still considerable interest in imaging apoptosis, as this is the predominant outcome from many cytotoxic and mew mechanism-based anticancer agents. [¹⁸F]-labelled 5-fluoropentyl-2-methyl-malonic acid ([¹⁸F]-ML10) is a ¹⁸F radiotracer which is a member of the ApoSense family of small-molecule apoptosis probes that utilize the γ-carboxyglutamic-acid (Gla)-domain protein as a potential critical structural moiety, enabling binding to cells with anionic phospholipid surfaces (Cohen et al., 2009). Early studies reported that induction of apoptosis is associated with cellular uptake of the probe, which is also caspase-dependent, and is blocked by caspase inhibition. Uptake of the tracer occurs in parallel with binding of Annexin-V and disruption of mitochondrial membrane potential. Uptake and accumulation of ([¹⁸F]-ML10 is lost upon membrane disruption and therefore may be utilized to differentiate apoptosis versus necrosis (Cohen et al., 2009). Whilst there are currently no published studies of the utility of [¹⁸F]-ML10 in the imaging of malignancies, a recent study has validated the use of an analogue of ML-10, ML-9. In this study [³H]-ML-9 was used to detect apoptosis induced by doxorubicin in in vivo models of malignancy (Grimberg et al., 2009). The authors report marked uptake of [³H]-ML-9 following treatment which correlated with both caspase activation and DNA fragmentation. Clinical validation of [¹⁸F]-ML10 as a marker of apoptosis is underway.

The process of cell death involves the family of caspases that execute cell death. Given the role of these enzymes in cell death there is considerable interest in imaging caspases. Isatin sulphonamides are novel agents in development that bind selectively to the active site of caspase-3 (and -7). Work by our group illustrates the utility of a specific [¹⁸F]isatin sulfonamide, [¹⁸F]ICMT-11, which binds to a range of drug-induced apoptotic cancer cells in vitro and to murine lymphoma xenografts (Smith et al., 2008; Nguyen et al., 2009). Furthermore, the increased signal intensity in tumours after drug treatment was associated with increased apoptosis. This tracer is currently undergoing further validation.

Angiogenesis

Angiogenesis is an essential process that supports tumour growth and metastasis through neovascularization that is driven by growth factors including vascular endothelial growth factor (VEGF) (Shibuya, 2008). There are two main strategies for the targeting of angiogenesis for therapeutic benefit: direct inhibition of angiogenesis growth factors and their receptors, for example, the anti-VEGF inhibitor bevacizumab, and indirect inhibition of blood flow through targeting of established vasculature, for example, the tubulin binding drug combretastatin A4-phosphate. Despite the introduction of these agents into clinical practice, there are few validated imaging biomarkers that can currently be used to predict or assess response to these therapies. Naturally there is considerable interest in this area

There have been a number of PET methodologies developed for imaging angiogenesis. Direct imaging of VEGF and its receptor has been studied by a number of investigators with IgG radiolabelled with long-lived isotopes but these are yet to be used routinely in the clinic. In a study by Collingridge and colleagues, the free and membrane bound VEGF in tumours was imaged directly in mice through utilization of an anti-VEGF antibody labelled with iodine-124 ([¹²⁴I]-SHPP-VG76e) (Collingridge et al., 2002). This study highlighted the need to evaluate antibody diagnostics following labelling to assure that they retain immunoreactivity and the use of control antibodies to assess blood pooling. Specific imaging of VEGF at 24 h after injection of the radiotracer was demonstrated in tumours expressing different levels of VEGF, as well as following presaturation of VEGF binding sites by unlabelled VG76e. In a phase I study, HuMV833, a humanized anti-VEGF monoclonal antibody, was labelled with iodine-124 and the distribution of the antibody, as well as its clearance, was assessed (Jayson et al., 2002). The results indicated heterogenous uptake both within individual tumours and between individual patients. Recently, bevacizumab was labelled with zirconium-89 and assessed in an ovarian cancer model in vivo (Nagengast et al., 2007). The study reported higher uptake of radiolabeled bevacizumab compared with control radiolabelled IgG in the human xenografts. However, compared with other studies investigating radiolabelled bevacizumab the tumour uptake was lower. This may have been as there is no clear evidence that VEGF is up-regulated in this tumour model, which again underscores the importance of appropriate tumour models in proof of concept studies. With respect to the targeting of the VEGF receptor, Cai and colleagues recently published an in vivo study of [⁶⁴Cu]-1,4,7,10-tetraazacyclododecane-N,N′,N″,N‴-tetraacetic acid (DOTA)-VEGF₁₂₁ (Cai et al., 2006). The report suggests rapid and specific uptake of [⁶⁴Cu]-DOTA-VEGF₁₂₁ in tumours that express high levels of VEGFR-2. Furthermore, the authors have developed a VEGFR-2-specific PET tracer based on mutated VEGF₁₂₁ that has lower binding affinity for VEGFR-1 without compromising VEGFR-2 binding affinity (Wang et al., 2007). Given the angiogenesis inducing properties of VEGF isoforms, it will be interesting to see how VEGFR-targeting probes are transferred into clinical use.

There is increasing interest in imaging integrin α_vβ_3. Integrins are a family of cell-adhesion receptors that allow the interaction of tumour cells and tumour endothelial cells with the extracellular matrix (ECM). The integrin α_vβ₃ is highly expressed on activated endothelial cells during angiogenesis. The tripeptide amino acid sequence, RGD, is present in a number of ECM proteins including vitronectin, fibronectin, and thrombospondin. Cyclic peptide ligands possessing the RGD sequence have high affinity for these integrins and, when labelled appropriately with positron- or γ-emitting radioisotopes, can be used for imaging α_vβ₃ receptor by PET. Haubner and colleagues synthesized [¹⁸F]Galacto-RGD for evaluation of tumour angiogenesis in a number of tumour types (Haubner et al., 2001). In these studies it was noted that tracer uptake corresponded to expression of α_vβ₃ and tracer uptake could be blocked by excess unlabelled compound. [¹⁸F]Galacto-RGD has been evaluated clinically in a number of tumour types illustrating good biodistribution, low background uptake and heterogeneous uptake within tumours that correlated with expression of α_vβ₃ within tumour samples. More recently a new pegylated RGD-based radioligand for α_vβ₃ ([¹⁸F]AH111585/[¹⁸F]fluciclatide) was developed with high tumour retention. In a recent phase I study, Kenny and colleagues reported that [¹⁸F]fluciclatide is retained in tumour tissues and detects primary and metastatic breast cancer by PET (Kenny et al., 2008 #125). A number of other clinical studies are currently underway utilizing this tracer as a response marker and results are awaited.

Visualization of blood flow offers another strategy for evaluating the vascular phenotype. Clinically, [¹⁵O]H₂O has been extensively used to measure tissue perfusion in response to anti-vascular therapy. [¹⁵O]H₂O is freely diffusible into and out of tissue water and is biologically inert, having the desired qualities for a tracer of perfusion. In a seminal study in phase I patients, [¹⁵O]H₂O-PET was performed in patients receiving the anti-vascular agent combrestatin A4. The authors reported a 30–60% reduction in tumour flow at higher dose levels. This trial was important in delineating the role of PET imaging in phase I drug development. In particular, results from the PET study confirmed the mechanism of action of combrestatin A4 and allowed for the selection of dosing schedule to be taken forward into phase II (Collins, 2003).

Conclusion

In summary, the development of a novel radiotracer is a complicated process involving an experienced multi-disciplinary research team. Described here are the initial steps in identifying and developing a potential radioligand for clinical use. As described, this involves extensive in vitro and in vivo characterization. A number of oncology drugs have been radiolabelled for PET imaging of PK and various drugs have been evaluated for target modulation and response using PET. Imaging of the tumours and normal tissues preclude extrapolation of plasma PK-PD to tissue PK-PD and the inherent assumptions associated with that as these are measured directly by the imaging method. As oncology continues to move towards the concept of personalized medicine, the role of PET imaging in achieving this aim will develop further. PET imaging allows a non-invasive, functional assessment of pathophysiological processes occurring within a patient, and is becoming an essential tool in the assessment of disease diagnosis, progression and response to therapy. Furthermore, as the cost of drug development increases, PET may be useful in demonstrating proof of concept for the proposed molecular and biological mechanisms of action for a drug, as well as being a useful biomarker. This is an area where reduced regulatory burden on investigators will boost innovation and rapid transition of research technology into patient tools.

Glossary

Abbreviations

5-FU

5-fluorouracil

ARSAC

Administration of Radioactive Substances Advisory Committee

BCNU

1,3-bis-(2-chloroethyl)-1-nitrosourea

CT

computer tomography

DACA

N-[2-(dimethylamino)ethyl]acridine-4-carboamine

DOTA

1,4,7,10-tetraazacyclododecane-N,N′,N″,N′″-tetraacetic acid

ECM

extracellular matrix

EGFR

epidermal growth factor receptor

eIND

Exploratory Investigational New Drug

EMEA

European Medicines Agency

ENT

equilibrative nucleoside transporter

EORTC

European Organisation for Research and Treatment of Cancer

FDA

Food and Drug Administration

GLP

Good Laboratory Practice

GMP

Good Manufacturing Practice

GRPP

Guidelines on Good Radiopharmacy Practice

IMP

investigational medicinal product

MCA

Medicines Control Agency

MHRA

Medicines Healthcare products Regulatory Agency

ML-10

5-fluoropentyl-2-methyl-malonic acid

MTS

multicellular tumour spheroids

PD

pharmacodynamics

PET

positron emission tomography

PK

pharmacokinetics

RDRC

Radioactive Drug Research Committee

RGD

arginine-glycine-aspartic acid

SOP

standard operating procedure

TK

thymidine kinase

VEGF

vascular endothelial growth factor

Conflicts of interest

The authors have no conflict of interest to declare.

Supporting Information

Teaching Materials; Figs 1–5 as PowerPoint slide.