Résumé
Objectives: The programmed death protein (PD-1) and its ligand programmed death-ligand 1 (PD-L1) play critical roles in a major checkpoint pathway in which cancer cells can evade detection by the immune system. A same-day PET imaging agent capable of measuring PD-L1 status in both primary and metastatic lesions simultaneously could be an important tool for optimizing PD-1 treatments in mono- and combination therapies. The purpose of this work is to evaluate the tumor targeting potential of two anti-PD-L1 Adnectins (derived from the 10th type III domain of human fibronectin (10Fn3); ~ 10 kDa) after 18F-fluorine labeling and characterize their properties as PET radioligands for the PD-L1 receptor. Methods: Two anti-human PD-L1 Adnectins, were labeled with 18F using a two-step radiochemical synthesis. This synthesis featured a direct fluorination of a prosthetic group, [18F](3-(2-(2-(2-(2-azidoethoxy)ethoxy)ethoxy)ethoxy)-2-fluoropyridine, followed by a copper free click conjugation with a ring constrained alkyne modified Adnectin. The tracer binding characteristics of these 18F labeled anti-human-PD-L1 Adnectins were evaluated in human non-small cell lung cancer (NSCLC) tissue samples using in vitro autoradiography. Immunohistochemistry staining was performed on these tumors to see if tracer uptake correlated to regions of high PD-L1 expression. Dynamic microPET imaging studies were carried out in mice bearing bilateral PD-L1 (+) (L2987, human lung) and PD-L1(-) (HT-29, human colon) xenografts after injection with [18F]ADX_5417_E01( n=7) or [18F]ADX_5322_A02 (n=7) Adnectins. Blocking studies were also performed in which mice received a co-injection of 3 mg/Kg of unlabeled anti-human PD-L1 Adnectin. In all cases following the 2 hr. imaging scan, the mice were euthanized and biodistribution studies were performed. Finally, in vivo biodistribution and radiation dosimetry was measured by microPET in cynomolgus monkey. Results: The anti-PD-L1 Adnectins bound to human and cynomologus monkey PD-L1 with Kd in the picomolar range (10-13 pM), as measured by surface plasmon resonance. These anti-PD-L1 Adnectins were successfully labeled with 18F. Autoradiography studies shows both [18F]Adnectins display specific and saturable binding to human NSCLC tissues. Autoradiography and IHC staining of these human tumor slices confirmed that PD-L1 expression patterns matched [18F]ADX_5322_A02 and [18F]ADX_5417_E01 binding patterns. In vivo PET imaging using these agents clearly visualized PD-L1 expression in L2987 xenografted mice. Two hours after administration, a 3.5-fold higher tumor uptake (2.41 ±0.29 %ID/g) was observed in the L2987 (PD-L1(+)) xenograft tumor compared to 0.82±0.11 %ID/g in the HT-29 (PD-L1 (-)) xenograft. A co-administration of 3 mg/kg unlabeled anti-human PD-L1 Adnectin reduced the tumor uptake at 2 h post-injection by more than 70%, while HT-29 uptake remained relatively unchanged, demonstrating PD-L1 specific binding. PET studies in cynomolgus monkeys confirmed binding to PD-L1(+) tissue (e.g. spleen) with minimal nonspecific background signal exclusive of primary clearance organs. Radiation dosimetry of [18F]ADX_5322_A02 indicates an estimated single administration dose limit of >6 mCi for an average human subject.The 18F-labeled Adnectins were rapidly cleared from the blood through the kidneys and bladder. Conclusion: Both [18F]Adnectins showed promising characteristics in preclinical evaluation. MicroPET imaging studies demonstrated the feasibility of using these labeled Adnectins to image tumor PD-L1 status. These derivatives are promising in vivo, same day imaging agents for evaluating PD-L1 protein expression on cancerous cells. Clinical studies with these same day imaging agents are underway to better understand the heterogeneity and dynamics of PD-L1 expression in human tumors.