In vitro evaluation of [3H]CPPC for imaging CSF-1R

Objectives: Carbon-11 labeled CPPC selectively targets microglia via macrophage colony stimulating factor-1 receptor (CSF-1R) in the central nervous system (CNS). This study aims to characterize [³H]CPPC as a new in vitro tool for radioligand binding methods in the CNS and peripheral tissues and screen a library of compounds derived from PLX3397 as potential PET probes for CSF-1R. Herein, autoradiography (ARG) studies were carried out to: (i) establish distribution of [³H]CPPC in a rat lipopolysaccharide (LPS) and mouse glioblastoma (GL261-Luc; GBM) model of neuroinflammation compared to TSPO shown by [³H]PBR28 ARG and IBA-1 immunofluorescence; (ii) determine the specificity of [³H]CPPC by ARG using CSF-1R targeting reference compounds in LPS rat brain and intraperitoneal (i.p.) LPS mouse spleen tissues; (iii) compare the target density given by [³H]PBR28 and [³H]CPPC binding in CNS tissues compared to peripheral tissues; (iv) explore selectivity over CNS targets and kinases and; (v) utilize [³H]CPPC as a tool radioligand to screen novel CSF-1R inhibitors.

Methods: Three models of inflammation were generated including a rat model of intrastriatal LPS injection, a mouse GBM model, and an i.p. LPS mouse model. Fresh frozen and fixed tissues were prepared from brain and spleen for radioligand binding assays and immunofluorescent staining with anti- CSF-1R and IBA-1 antibodies. ARG was carried out in 100% fetal bovine serum at room temperature over 2 hours using increasing concentrations of [³H]CPPC or [³H]PBR28. Competitive binding ARG was carried out using tissues co-incubated with [³H]CPPC and reference (PLX3397, PLX5622 or BLZ-945) or test compounds. Competitive binding with [³H]CPPC in spleen tissue was used to screen a library of 20 PLX3397 derivatives as leads for new PET radioligands. Cold CPPC underwent a kinase panel screen (Eurofins) for in vitro selectivity.

Results: Non-specific binding of [³H]CPPC was high in normal rodent brain tissue. [³H]CPPC binding did not align with expression of inflammation as shown by TSPO distribution ([³H]PBR28) using autoradiography and IBA-1 immunofluorescence. Saturation binding of [³H]CPPC was carried out in LPS rat tissues however saturation was not achieved. Due to high non-specific binding in CNS tissues, LPS-treated spleen tissues were used to calculate a B_max of 1339 fmol/mg and a K_d for CSF-1R of 4.3 nM with [³H]CPPC. To explore specificity, [³H]CPPC was co-incubated with known CSF-1R inhibitors and minimal displacement was observed with BLZ-945 (25%), PLX5622 (16%) and PLX3397 (15%) while high specific binding in LPS spleen tissues was discovered. Multiple binding sites were observed. Unlabeled CPPC lacked selectivity and showed off-target binding to a substantial number of kinase targets (204 of 403 tested) at a concentration relevant to in vitro radioligand binding assays (10 μM).

Conclusions: Herein we report the first in vitro evaluation of [³H]CPPC as a tool radioligand in several rodent models of neuroinflammation where both adequate saturability or specific binding were not observed. However [³H]CPPC may have utility for the evaluation of peripheral targets and screening of CSF-1R inhibitors.