BackgroundThe C-X-C chemokine receptor type 4 (CXCR4) is overexpressed in many cancers, e.g. multiple myeloma and acute leukemia, yet solely [68Ga]PentixaFor is used for clinical PET imaging. The aim of this study was to develop and assess a second generation Al18F-labeled D-amino acid peptide based on the viral macrophage inflammatory protein II for CXCR4 targeted molecular imaging. MethodsWe designed a library of monomer and multimer constructs and evaluated their binding affinity for human and mouse CXCR4. Based on these results, we selected the best vector molecule for development of an Al18F-labeled ligand, [18F]AlF-NOTA-2xDV1(c11sc12s), which was further evaluated in a cell-based binding assay to assess its binding properties and specificity for CXCR4. Next, pharmacokinetics and tumor uptake of [18F]AlF-NOTA-2xDV1(c11sc12s) were evaluated in naïve mice and mice with xenografts derived from U87.CXCR4 cells. Finally, we performed an imaging study in a non-human primate to assess the in vivo distribution of this novel radioligand in a species closely related to humans. ResultsThe lead ligand AlF-NOTA-2xDV1(c11sc12s) showed six-fold higher affinity for human CXCR4 compared to Ga-Pentixafor. The corresponding radiotracer was obtained in a good radiochemical yield of 40.1 ± 13.5 % (n = 4) and apparent molar activity of 20.4 ± 3.3 MBq/nmol (n = 4) after optimization. In U87.CD4.CXCR4 cell binding assays, the total bound fraction of [18F]AlF-NOTA-(2×)DV1(c11sc12s) was 32.4 ± 1.8 %. This fraction could be reduced by 82.5 % in the presence of 75 μM AMD3100. In naïve mice, [18F]AlF-NOTA-2xDV1(c11sc12s) accumulated in organs expressing mouse CXCR4, e.g. the liver (SUVmean (mean standardized uptake value) 75 min p.i. 11.7 ± 0.6), which was blockable by co-injecting AMD3100 (5 mg/kg). In U87.CXCR4 xenografted tumor mice, the tumor uptake of [18F]AlF-NOTA-2xDV1(c11sc12s) remained low (SUVmean 0.5 ± 0.1), but was reduced by co-administration of AMD3100. Surprisingly, [18F]AlF-NOTA-2xDV1(c11sc12s) exhibited a similar biodistribution in a non-human primate as in mice indicating off-target binding of [18F]AlF-NOTA-2xDV1(c11sc12s) in liver tissue. We confirmed that [18F]AlF-NOTA-2xDV1(c11sc12s) is taken up by hepatocytes using in vitro studies and that the uptake can be blocked with AMD3100 and rifampicin, a potent organic anion-transporting-polypeptide (OATP)1B1 and OATP1B3 inhibitor. ConclusionThe second generation D-peptide AlF-NOTA-2xDV1(c11sc12s) showed high affinity for human CXCR4 and the corresponding radiotracer was produced in good radiochemical yields. However, [18F]AlF-NOTA-2xDV1(c11sc12s) is not specific for CXCR4 and is also a substrate for OATP1B1 and/or OATP1B3, known to mediate hepatic uptake. Therefore, D-amino acid peptides, based on the viral macrophage inflammatory protein II, are not the prefered vector molecule for the development of CXCR4 targeting molecular imaging tools.