KEY CHOICES FOR ACCURATE REPRESENTATION

By acting directly on the immune system, immunotherapies, particularly immune checkpoint inhibitors such as PD-L1, seem to show encouraging results in the fight against cancer. Thanks to molecular imaging by Positron Emission Tomography (PET), it is possible to predict for each patient the effectiveness of this type of treatment with greater precision. PET can quantify the expression of a predictive biomarker for immunotherapy such as PD-L1 and provide a so-called body-identity mapping. This requires the use of radiolabeled ligands directed against PD-L1 and the real-time visualization of its expression in all tumor lesions. ImmunoPET, coupling the sensitivity of PET with the affinity of monoclonal antibody-based ligands, has allowed many advances in this field. However, the choice of the format of the antibody-based ligands is crucial to ensure an accurate representation of PD-L1 distribution.

Recently, more and more clinical and preclinical trials involve radiolabeling of antibody fragments (Fab), as their clearance is faster and their tumor penetration higher than immunoglobulins (IgG). Despite their attractive PET imaging properties, using these small molecules is more complex and risky. Notably, the radiolabeling process may have a greater influence on in vivo behaviors compared to IgG.

The most common strategy for radiolabeling antibodies or antibody fragments relies on the ligation (conjugation) of a radiolabeled electrophilic prosthetic group to one of its surface-accessible lysine residues in a random fashion. But this approach lacks regioselectivity and the drug to antibody ratio (DAR) is not strictly controlled. Non-random ligation methods (specific bioconjugations) are however of great interest. Enzymatic bioconjugation, in particular, represents an interesting alternative with several advantages. Indeed, the labelling conditions are soft, at physiological pH and temperature, and the number of possible ligations as well as the regioselectivity can be totally controlled.

A DIFFERENT BIODISTRIBUTION DEPENDING ON THE LABELLING

In their new study published in Molecular Pharmaceutics, researchers from the BioMaps laboratory (SHFJ) and the Molecular Engineering for Health Service (SIMoS, DMTS) have developed a chemoenzymatic approach for the radiofluorination of an anti-PD-L1 Fab and compared the in vivo properties of their radiolabeled product to those of two other versions of this Fab stochastically labeled by conventional methods (with Fluorine 18 or Zirconium 89).

They achieved rapid and fully automated radiofluorination^[1] with high labeling yield on an anti-PD-L1 Fab through the action of lipoate-protein ligase. They then compared the impact of specific versus random bioconjugation on the in vivo properties of the Fab.

Their data suggest that the choice of the method used for radiolabeling (specific or random) has little impact on Fab pharmacokinetics. In contrast, biodistribution is different depending on whether Fab is labeled with ¹⁸F or ⁸⁹Zr, particularly in the liver, kidney, heart, bone and joints. These differences between these two labeling pathways could be explained by the metabolism and natural affinity of zirconium for bone and cartilage tissue.

The non-specific release of ⁸⁹Zr, even in small amounts, may lead to misinterpretation of the images by immunoPET. Therefore, ¹⁸F labeling better reflects Fab pharmacokinetics and has several advantages for better patient management.

[1] via a prosthetic group derived from a pyridine structure;([¹⁸F]FPyOctA)