Allocation of TSPO-PET signals to cellular and extracellular compartments in dependence of blood-brain barrier disruption in experimental glioblastoma

Abstract

Abstract Background Translocator protein (TSPO)-PET imaging facilitates monitoring of glioblastoma in preclinical models and patients. However, specificity of TSPO-PET signals remains to be investigated. In this study, we aimed to decipher exact contributions of cellular and extracellular compartments including the impact of blood-brain barrier (BBB) disruption to TSPO-PET signals in an experimental glioblastoma mouse model. Methods Mice with implanted glioblastoma (SB28; n = 36 early-stage, n = 39 late-stage; GFP(+)) were injected with the TSPO tracer [ 18 F]GE-180 (21 MBq ± 1.5 MBq). Dynamic 60 min microPET scans were performed with and without prior blocking by excessive cold radiotracer. Tumors were dissociated, followed by cell sorting of tumor cells, tumor-associated microglia/macrophages (TAMs, CD11b(+)) and non-tumor/non-TAM (i.e. remaining) cells with subsequent gamma emission measures and flow cytometry (scRadiotracing) to calculate radioactivity per single cell. PET signals and single-cell tracer uptake in tumors and non-lesional hemispheres were compared between blocked and unblocked conditions. N = 10 mice were intravenously injected with fluorescent Dextran to investigate BBB disruption by confocal microscopy in correlation with TSPO-PET signals of the same animals. Results TSPO-PET indicated strong but incomplete signal reduction after blocking in tumors (early-stage: -72%; late-stage: -59%, p < 0.001) and contralateral hemispheres (early stage: -65%, late-stage: -63%, p < 0.001) compared to unblocked animals. We found nearly complete blocking of radiotracer uptake across all analyzed cell fractions (tumor cells: -95%, p = 0.0039; TAMs: -98%, p < 0.0001; remaining cells: -99%, p = 0.0033) compared to unblocked mice, regardless of tumor stages. There was a significant correlation of PET signal reduction ( R =-0.824, p = 0.0034) as well as residual PET signal upon blocking ( R = 0.809, p = 0.0046) with Dextran intensity. Conclusion Specificity of TSPO-PET signals in experimental glioblastoma reaches 72% in early-stage tumors, but decreases to 59% in late-stage SB28 tumors, due to progressive contributions of BBB disruption. Non-specific TSPO tracer uptake is driven by BBB disruption and fully allocated to the extracellular compartment, whereas cellular TSPO tracer uptake shows strong specificity.