Abstract
1092
Objectives: High affinity CXCR4 antagonists with long receptor residence times have been developed which have been radiolabelled with a range of isotopes. Our aim was to provide a series of gallium-68 labelled molecules based on high affinity azamacrocyclic binding units. Imaging CXCR4 chemokine receptor expression levels in cancers as a prognostic marker or to inform treatment selection is of interest with the peptidic gallium-68 agent Pentixafor of increasing clinical interest. We have targeted the CXCR4 chemokine receptor using configurationally restricted tetraazamacrocyclic transition metal complexes to give nanomolar level receptor affinities and long receptor residence times, offering an alternative approach to gallium-68. Methods: The probes can be labelled with gallium-68 in a second chelating component or prosthetic group. The influence of structural variation in chelator and spacing group was investigated using in vitro assays to select compounds for in vivo imaging and optimised radiolabelling protocols. Radiolabelling with gallium-68 was carried out with analysis by radio-TLC and conditions developed for semi-preparative HPLC purification. Intracellular calcium flux in response to CXCL12 was evaluated in vitro with non-radioactive compounds U87-CXCR4 glioblastoma cells along with cell binding experiments using the selected radiotracers. In vivo 90 minute dynamic PET/CT imaging was carried out on CD1 mice bearing U87-CXCR4 xenografts. Blocking studies were carried out using a higher affinity probe. Comparison with pentixafor was carried out although this tracer does not bind to murine CXCR4 and so only tumour comparisons are valid.
Results: A series of compounds with different CXCR4 binding units (cross bridged cyclam and cyclam/ zinc(II), nickel(II) or copper(II)) and different length spacers with isothiocyanate or amide functional groups linking to the gallium-68 chelator (NOTA, DOTA, NOTAGA or DOTAGA starting chelators). Four probes were selected for imaging studies with affinities in the range 20-142 nM. Gallium-68 was used to label the chelators to form a radiometal complexes in yields of 34-96% with analysis by radio-TLC before purification. HPLC methods were developed for separation of the radiolabelled compounds along with reformulation for imaging studies. All were shown to be stable in serum up to 3 hour at 37ºC. In vitro experiments showed binding to cells overexpressing the CXCR4 receptor. In vivo imaging showed SUV max values for tumour slightly lower than pentixafor for the best tracer tested (2.1 vs. 2.9 at 90 mins) however tumour accumulation was similar at earlier timepoints. Blocking with 10 mg/kg of a higher affinity antagonist (IC50 = 4 nM) shows a 78% reduction of radioactivity in the U87.CXCR4 tumour in the best tracer from this set (NOTA with amide linker). Conclusions: The initial structures tested for gallium-68 labelling disrupted receptor binding due to the positioning of an acetate group, however, this was addressed by the inclusion of a spacer group or alternate chelator. High affinity probes and optimised methods for radiolabelling/ separation were developed. There is impact of chelator and linker group choice on probe imaging properties in vivo. Comparisons with pentixafor are favourable and with further optimisation similar properties could be attained. Research support: We gratefully acknowledge Yorkshire Cancer Research (Grant: HEND376), Dr Assem Allam and the Daisy Appeal Charity (Grant: DAhul0211) for funding. Figure: Examples from in vivo studies. Subjects implanted with CXCR4 over-expressing tumours injected with 2-8 MBq of gallium-68 tracer. Two animals shown as coronal maximum intensity projections (46-66 minutes). For SPO342 the images on the right are following injection of a blocking agent. A plot of SUV max is shown for the tumour. Analysis for the liver was carried out for macrocyclic binding unit tracers (pentixafor does not bind to murine CXCR4).
