Positron emission tomography molecular imaging for drug development

Abstract

Human in vivo molecular imaging with positron emission tomography (PET) enables a new kind of 'precision pharmacology', able to address questions central to drug development. Biodistribution studies with drug molecules carrying positron-emitting radioisotopes can test whether a new chemical entity reaches a target tissue compartment (such as the brain) in sufficient amounts to be pharmacologically active. Competition studies, using a radioligand that binds to the target of therapeutic interest with adequate specificity, enable direct assessment of the relationship between drug plasma concentration and target occupancy. Tailored radiotracers can be used to measure relative rates of biological processes, while radioligands specific for tissue markers expected to change with treatment can provide specific pharmacodynamic information. Integrated application of PET and magnetic resonance imaging (MRI) methods allows molecular interactions to be related directly to anatomical or physiological changes in a tissue. Applications of imaging in early drug development can suggest approaches to patient stratification for a personalized medicine able to deliver higher value from a drug after approval. Although imaging experimental medicine adds complexity to early drug development and costs per patient are high, appropriate use can increase returns on R and D investment by improving early decision making to reduce new drug attrition in later stages. We urge that the potential value of a translational molecular imaging strategy be considered routinely and at the earliest stages of new drug development.

Figure 1

Effect of increasing doses of the competitive P-gp substrate cyclosporine A (CsA) on [¹¹C]-loperamide uptake in porcine brain. Increasing doses of CsA lead to increased net uptake of the [¹¹C]-loperamide with greater competition for the transporter. Note that the prominent ‘hot spot’ in the upper two images localizes to the pituitary gland, which sits outside the blood–brain barrier. Images were acquired from the same animal scanned sequentially on the same day

Figure 2

Chemical structures of 3-N-[¹¹C-methyl]-temozolomide (A) and [4-¹¹C-carbonyl]-temozolomide (B)

Figure 3

One image set from an illustrative drug occupancy study illustrating the radioligand signal before (upper) and after (lower) administration of cold drug competing for the same binding site. Brighter regions define increased radioisotope concentration. With repetition of a similar image pair over a range of doses of cold drug (or by varying timing of radioligand injection after cold drug administration), varying plasma concentrations at the time of scanning allow estimation of an in vivo IC₅₀ for the cold drug based on measures of relative displacement of the radiotracer

Figure 4

An example of data from target occupancy studies with two molecules in candidate selection. The molecules were antagonists and previous work suggested that occupancy by approximately 80% or more would be needed for the desired pharmacological effects (horizontal broken line). However, both molecules had recognized potential toxicities at plasma concentration shown by the vertical solid line. The in vivo human target occupancy-plasma concentrations defined in separate sets of PET experiments (see panels A and B) established a range of plasma concentrations (and thus doses) over which pharmacological effects were likely to be seen. In doing so, they also estimated the therapeutic index for the molecules (grey areas). The molecule used in the study for panel A, which has the higher therapeutic index, was selected for further development

Figure 5

[¹⁸F]-FDG PET study of a patient with an abdominal ovarian tumour (arrow). A significant decrease in [¹⁸F]-FDG uptake (SUV_max) together with volumetric tumour reduction was observed at the second visit. Images courtesy of Dr A. Saleem, GSK Clinical Imaging Centre

Figure 6

(A) Group statistical maps (overlaid on structural MRI image) showing baseline fMRI BOLD response to receipt of monetary rewards. (B) Relationship between study population variability in midbrain dopamine receptor subtype 3 (D3R) availability and amygdala response to reward. Correlation between midbrain [¹¹C]-(+)-PHNO BP_ND and amygdala activation to monetary rewards for individual subjects studied, demonstrating that subjects with higher midbrain D3R availability have greater amygdala BOLD response to rewards

Figure 7

(A) Structures for alternately radiolabelled forms of the radioligand WAY100635 which show different relative radioisotope accumulation in hippocampus (hip) and medial temporal cortex (MTC) relative to the cerebellum (cerebell) in rat and human because of different routes of metabolism in the two species (B)