Although the concept of “biological validation” is potentially as broad as one can possibly imagine, a few minimal key but non-trivial experiments are required to demonstrate that the radiopharmaceutical visualizes the biological process or pathway for which it has been designed. The main goal of these biological validation experiments is to ensure that the radiopharmaceutical specifically interacts with the biological target and is not just retained in a non-specific fashion. In vivo stability, and in vivo biodistribution are also a few of the basic requirements for tracer characterization.
In vitro studies and cross-validation
Although the present white paper is focused on radiopharmaceuticals developed for oncological applications, many concepts described herein are likewise applicable to tracers developed for other diseases. In oncology, the most common targets for radiopharmaceuticals are oncoproteins that are linked to key deregulated pathways. Evidence of target-mediated uptake must be shown in vitro via incubation of the tracer with cells expressing the target. Elevated radiotracer levels in these target-expressing cells confirms specificity versus low to negligibly bound tracer values in negative or low-expressing control cells. Target-positive cells can have a natural abundance of the target, such as ad-hoc tumor cells. Alternatively, cells can be engineered to express or downregulate the target, essentially creating isogenic models with similar genetic backgrounds. Normal tissue cells or cell with low expression of the target must be included as control. Expression has to be cross-validated by a different technique, the most common being western blots. Immunohistochemistry (IHC), bioluminescence imaging (BLI), flow cytometry or any technique able to independently confirm the expression level of the target adds value to the validation.
Additionally, target-mediated uptake must be confirmed with an intervention aimed at reducing the signal. This can include a variety of techniques, some more commonly used than others. Addition of an excess of the cold reference compound — at least 10 × , and up to 100 × , more than the radiotracer, aimed to competitively saturate the target and reduce the uptake of the radiopharmaceutical, the so-called “blocking experiment”, is perhaps the most common proof of target-mediated uptake of the drug, as opposite as non-specific uptake. Alternatively as discussed above, engineered knockdown cells or cells treated with drugs to modulate the target’s expression (i.e. inhibitors) can be employed. When possible, demonstrating differential tracer uptake in a panel of cell lines that have varying levels of target — usually high versus low — provides a thorough proof of target-mediated uptake of the radiopharmaceuticals. These cell lines either possess wild-type, knockdown or knock in target expression, or modulated via treatment with inhibitors using the drug vehicle as a control. Specificity of the tracer can also be confirmed by comparing uptake of a non-specific tracer. Low cellular uptake of a scrambled peptide or a non-specific mAb isotype validates specificity of the tracer.
Along these lines, peptide and antibody-based radiopharmaceuticals should be investigated for internalization kinetics upon binding to the target antigen. Tracer internalization is usually demonstrated as a function of time with incubation periods ranging from minutes to days. Information on the rate of tracer internalization renders meaningful insights on outcomes of radioligand therapy and dose selection as determined via image-guided PET or SPECT imaging.
For radiolabeled antibodies, the immunoreactivity or immunoreactive fraction (IF} is determined to establish what fraction of the radiolabeled mAb retains its ability to bind to its target antigen. The IF of a modified mAb is typically significantly less than 100% due to chelate/linker conjugation, radiosynthesis conditions, and damage from radiation during storage [
21,
22]. During preclinical development, establishment of the IF of new tracers is a prerequisite. There is currently no consensus or defined acceptable range. Experts in the field deem acceptable immunoreactivity limits are > 70% if using live cells and > 80% if cell-free or immobilized antigens are utilized. Values below this range will likely affect targeted uptake of the agent as well as non-specific binding and clearance. There is no standard approach to determining IF. The most common protocol was developed by Lindmo et al
. wherein the IF is linearly extrapolated to conditions that are under infinite antigen excess [
23]. This method is deemed to derive the “true” IF versus the apparent IF value, determined under limited excess antigen conditions. However, the Lindmo protocol has drawbacks especially when antigen density on the cell surface is unknown or limited [
24,
25].
The binding affinity, defined as the strength of association between the drug and its target ligand, is paramount when designing a mAb- (or small molecule)-based radiopharmaceutical where the radiolabeled compound is targeting a cell surface receptor, integrin, or other type of protein. Affinity is typically expressed as the equilibrium dissociation constant, K
D, which is the ratio of how rapidly the drug dissociates (k
off) to how fast it binds (k
on). Thus, the smaller the K
D, the higher the binding affinity. It is commonly thought that the highest affinity mAbs (10
−10–10
−12 M) are usually favored in drug development. However, in the case of tumor delivery, evidence suggests the mAbs do not distribute homogeneously throughout the tumor. They preferentially localize in the tumor periphery, within close proximity to blood vessels as a consequence of slow dissociation kinetics (k
off ~ 28 h) [
26,
27]. This creates a binding site barrier, where mAb diffusion into the tumor core is hindered [
28]. In contrast, low affinity mAbs (K
d > 10
−7, k
off ~ 10 s) rapidly dissociate and move further into other regions of the tumor, but with the drawback of low tumor uptake and retention. A K
D “sweet spot” ranges between 10
−8–10
−9 M. In contrast, the tumor penetration of small molecule radiotracers is not as dependent on K
D, but still remains a critical parameter to be established in development. Collectively, binding affinity dictates a significant role in tumor association and penetration of radiopharmaceuticals for imaging and therapy.
A key parameter in the development of a radiopharmaceutical is its binding potential (BP). It is a key measure of receptor occupancy derived from in vitro radioligand binding experiments. BP is calculated as the ratio of the total density of receptors in a tissue (Bmax) to affinity (KD). BP = Bmax KD
Of note, measuring in vivo receptor occupancy preclude receptors bound to endogenous ligands. For in vivo imaging studies wherein radiotracer concentrations are very low and occupying only a small portion of available receptors, the BP is indicative of bound versus free tracer concentration in a state of equilibrium [
29].
In vivo validation
In vivo metabolism is a very useful indication of how long the tracer remains intact in vivo as it accumulates within the tumor target prior to degradation.
Studies are usually performed by extracting the tracer and/or its metabolites at specific time points from tissues of interest and by analyzing the extract via radio-HPLC. A thorough analysis involves extraction and analysis of several tissues of interest (
i.e. blood, liver, kidney, heart, tumor, even spleen). At minimum, metabolites from the blood and urine are reported. Typically the % intact compound in the organ is normalized to an “organ blank”, where the compound is added to an organ ex vivo and then homogenized, centrifuged, and the amount of compound that remains with the pellet is taken into account. An example of how this is done is described by Boswell et al., [
30] where the in vivo stability of
64Cu-chelates was determined. The % authentic intact
64Cu-ligand complex = %P x %C x [1 + (% Pellet/% Super)] x % E, where % P = purity of injectate as determined by radio-iTLC; %C =
64Cu chelator complex determined by integration of the size-exclusion HPLC chromatogram; % Pellet/% Super = the ratio of the radioactivity in the pellet and supernatant of the organ blank; and % E = the extraction efficiency of the harvested organ after injection of
64Cu-labeled chelate. A figure should report the HPLCs of the intact standard staggered with representative HPLCs of the organs analyzed. Analysis can be performed at times ranging from 5 min to 24 h post-injection or longer, depending on the PK of the tracer. In addition, % intact compound at any time point should be reported with standard deviation for a minimum of three animals [
31]. In general, an in vivo metabolism study informs on the viability of the tracer to accumulate in the tumor while remaining intact within a certain timeframe. If the tracer degrades rapidly prior to reaching the target, poor contrast is potentially obtained stemming from non-target tissue binding of radiometabolites.
Ex vivo biodistribution is the standard for characterizing the pharmacokinetic (PK) properties of the radiotracer. However, when and if possible, PET or SPECT imaging based biodistribution is equally accurate and sometimes preferable. This approach requires fewer animals and maintains the context of size and local environment which is critical for detection. Dynamic imaging scans, particularly with small molecule tracers, are also useful for in vivo PK analysis. They should be reported as Time Activity Curves (TACs) and expressed as %ID/cc.
Each radiopharmaceutical is designed to target a specific biological phenomenon. Regardless, in order to exert its role, a radiopharmaceutical requires a mechanism of retention or delivery to the target tissue. Mechanisms of delivery or retention can either be protein/ligand binding, [
32] metabolic trapping as is the case for [
18F]FDG [
33] or [
18F]FAZA [
34], an in situ secondary reaction such as phosphorylation, or, simply, a shift in polarity [
35]. These methods of radiotracer uptake should be identified, and, when possible experimentally proven.
Proof of target-mediated uptake must be demonstrated in vivo. Tumor bearing mice must be injected with the radiopharmaceutical, and imaging must be taken at different time points. With mAb-based tracers, the mAb must be cross-reactive to the target antigen. Fully human or humanized mAbs will require testing in established human xenografts or patient-derived tumors with overexpressed targets using immunocompromised hosts. When possible and if cross-reactivity of the tracer to the target is not an issue, employment of syngeneic tumors in animal models with an intact immune system will provide a better understanding of the role played by the immune system on the tumor uptake of the radiotracer [
36]. Engineered mouse models, such as transgenic mouse models or knock-out models, are a very valuable addition to the validation of a new radiopharmaceutical.
For diagnostic agents, it is paramount to identify the timepoint that provides the best signal-to-noise ratio. For short-lived isotopes such as fluorine-18, dynamic imaging must be taken for at least one hour, followed by static images. Time points vary depending on the half-life of the agent: for short-lived isotopes, static scans can be recorded up to 3 h post injection. For mAb radiotracers bearing longer lived isotopes (e.g. copper-64, zirconium-89), imaging as early as 1–4 h p.i. has been reported, followed by acquisitions at 24 through 144 h p.i. It is worth nothing that imaging of anesthetized animals for prolonged periods of time should be done with caution since certain anesthetics can alter biological processes, which can affect tracer distribution. This can be attenuated by acquiring dynamic scans for at least 10 min to evaluate the initial biodistribution of the tracer, then statically at the 1 h or at later timepoints.
For preclinical validation, quantification of PET imaging is usually reported as % injected dose per cubic centimeter (% ID/cc). Standardized uptake values (SUVs) are also used in the preclinical field and are the most commonly used metric in the clinical field. Quantification of the agent accumulation as tumor-to-muscle or, for orthotopic tumors, tumor to the corresponding healthy tissue is important metric for quantification of contrast. Similar to the in vitro experiment described above, target-mediated uptake can be proven by using tumors with different expression of the target, or the target can be modulated pharmacologically, by means of inhibitors or pharmacological doses of the corresponding non-radioactive compound.
When cancer metabolic pathways are targeted, it is important to clarify what enzymatic or signaling pathway is responsible for the tracer accumulation in the tumor cells. When the tracer is an analogue of an endogenous metabolite, often a direct readout of the enzymes responsible for the tracer’s entry into the tumor cells (transporters) is provided. Notable examples are [
18F]FDG and [
18F]fluoroglutamine ([
18F]FGln), which are transported inside the tumor cells by mainly GLUT1/3 and ASCT2 [
37,
38], respectively, and retained intracellularly shortly after. In this case, results of a metabolite analysis can strengthen our understanding of the relevant applications the tracer can be used for. [
18F]FGln for instance, can be employed to assess the overexpression of ASCT2, that might or might not translate into overactivation of glutaminolysis. However, as an example, using [
18F]FGln as direct pharmacodynamic marker of glutaminase inhibitors can be complicated, because glutaminase is downstream the “trapping” point for the PET tracer. It can, however, be correlated to glutaminase inhibition, if upregulation of the transporters is occurring in that context [
39]. Once an initial validation is performed, numerous applications of a new radiopharmaceutical can be investigated, and the list can be endless. More mouse models, such as orthotopic tumors, syngeneic mouse models, or patient-derived xenografts, can be employed, together with various types of interventions.
Biological applications span from understanding the pharmacokinetics/biodistribution of a drug via molecular imaging by using a radiolabeled version of the drug [
40], or understanding the pharmacodynamics of a drug by using an imaging agent that reports on a secondary mechanism [
19]. Imaging can also be extremely useful in understanding early response to therapy both as positive predictor or negative predictor [
34,
41‐
45]. Finally, most of the validations reported herein are applicable to the development of radiopharmaceutical for systemic radiotherapy, that however, requires further robust and controlled tests to assess efficacy.