Increased Imaging Ligand Hydrophilicity and Improved Pharmacokinetic Properties Provides Enhanced In Vivo Targeting of Fibroblast Activation Protein

Abstract

In this work, the impact of physicochemical modifications on pharmacokinetics and in vivo targeting of a small molecule fibroblast activation protein inhibitor (FAPI) imaging ligand in a murine model of rheumatoid arthritis was evaluated. While similar ligands have been well-reported in oncology for molecular imaging and radiotherapy, there are limited reports of FAPI derivatives in targeted applications in immunology. As inflammation may increase both specific and non-specific delivery of targeted agents in general, we sought to identify the optimal targeted molecular imaging probe characteristics for efficient cell surface engagement. A series of FAPI derivatives were synthesized and their physicochemical properties modified via conjugation of fluorescent dyes and/or an albumin-binding small molecule. The impact of these modifications on cell surface binding affinity was assessed using an overexpressing line. Additionally, a thorough mechanistic characterization of fibroblast activation protein (FAP) cell surface internalization was evaluated in both overexpressing and endogenously expressing cells. Lastly, the pharmacokinetics and in vivo uptake in inflamed arthritic paws were characterized via near-infrared (NIR) imaging. All targeted molecular imaging agents tested maintained strong nanomolar binding affinity to cell surface FAP independent of chemical modification. The murine fibroblast-like synoviocytes expressed lower absolute cell-surface FAP compared to a transfected line, and the net internalization half-life measured for the transfected cells via flow cytometry was 7.2 h. The unmodified FAPI ligand exhibited the poorest in vivo targeting, likely resulting from its large apparent volume of distribution (62.7 ml) and rapid systemic clearance (t1/2 = 0.5 h). Conjugation of a charged, hydrophilic AF647 fluorophore decreased systemic clearance (t1/2 = 2.1 h) and demonstrated a 2-fold improvement in blocking FAPI-800CW engagement of FAP in vivo when compared to blocking of FAPI-800CW with FAPI with up to 2.8-fold improvements noted for the equivalent albumin binding construct comparison.