Developing a selective sphingosine-1-phosphate-5 (s1p5) radiotracer to image oligodendrocytes using preclinical positron emission tomography (PET)

In multiple sclerosis (MS), the myelin sheaths (the lipid sheathing around neurons, which are produced by oligodendrocytes) degenerate, leading to a loss of function, neuronal degeneration and disability. Although there have been advances in recent years, there is still no cure for MS or even a method to facilitate imaging oligodendrocytes activity in vivo, other than the established structural techniques such as Magnetic Resonance Imaging (MRI). In this thesis we aimed to develop novel Positron Emission Tomography (PET) ligands to specifically bind to oligodendrocytes in the central nervous system in vivo. The ligands target the sphingosine-1-phosphate-5 (S1P5) receptor as a potential marker of oligodendrocyte function with PET. The ligands selected (TEFM180 and TEFM78) were developed from a drug development library of agonists targeting S1P5 that had shown high affinity and selectivity for their target. An agonist radiotracer would be of particular interest as S1P5 is a G-protein coupled receptor and agonists show a bias to binding receptors in the active state. If successfully translated, these ligands could aid in improving trials of novel therapeutics, improving assessment of disease progression and importantly furthering our understanding of MS pathobiology. To deliver on this project's overarching aim (developing a new S1P5 PET radiotracer), the first priority was to evaluate S1P5 as a specific target on oligodendrocytes. For this, in situ hybridisation and immunofluorescence staining techniques were applied on adult naïve rat brain tissue sections. S1P5 was stained alongside various markers of oligodendrocyte developmental stage as well as markers for other central nervous system (CNS) cell types (astrocytes, microglia, and neurons). This enabled the staging of S1P5 expression at the protein level and its co-localisation and co-expression with specific cell type markers. Alongside characterisation of S1P5 expression at the RNA and protein level in the mammalian brain, two lead small molecules were investigated as potential selective PET radiotracers for S1P5. One of those molecules was TEFM180, a compound amenable to carbon-11 labelling, which was labelled with tritium and used to conduct receptor ligand binding assays and autoradiography experiments on naïve rat brain tissue. The other compound was TEFM78, a lead candidate for fluorine-18 radiolabelling. [18F]TEFM78 was radiolabelled and used for in vivo radiometabolite experiments, plasma free-fraction experiments, and dynamic PET scans on naïve rats. Kinetic modelling was conducted on the [18F]TEFM78 PET scans with input function data collected. Results from experiments conducted in this project showed there was co-localisation between S1P5 and Plp1 using in situ hybridisation and between S1P5 positive cells and CC1 positive cells using immunofluorescence staining techniques. Olig2 positive cells did not co-localise with S1P5 positive cells in the majority of cases however there was some co-localisation in a subset of cells. NG2 positive cells did not co-localise with S1P5 positive cells. GFAP and Iba1 did not co-localise with S1P5 positive cells and the cells were morphologically distinct. There was co-localization between NeuN positive cells and S1P5. In vitro receptor ligand binding assays showed that total and non-specific binding rose at increasing [3H]TEFM180 concentrations and high concentrations of rat brain protein were required to obtain a low degree of specific binding. Higher specific binding was measured using in vitro autoradiography techniques (37.62 to 70.96%). However the binding did not correlate with S1P5 immunofluorescence staining. TEFM78 was successfully radiolabelled with [18F] and used for in vivo PET studies, however productions did have a low yield and relatively low molar activity. Radiometabolite studies showed moderate metabolism of [18F]TEFM78 in rats (44% at 1 hour post-injection) and high plasma protein binding in both rat and human blood (>98%). In in vivo PET scans, [18F]TEFM78 cleared the blood rapidly, entered the rat brain and had higher uptake in white matter rich regions compared with grey matter regions. Kinetic modelling was completed on the scans with invasive input function and it was found that a 1-tissue model was preferred for this data. The total volume of distribution (VT) was 1.39 ± 0.06 mL/ccm in the whole brain, 1.58 ± 0.04 mL/ccm in the white matter and 1.39 ± 0.07 mL/ccm in grey matter. The immunofluorescence results confirm that S1P5 remains a target of interest to investigate oligodendrocytes in the context of MS, however the neuronal expression seen requires further investigation. [3H]TEFM180 is a sub-optimal ligand with low specific target engagement, however we demonstrated that this chemical scaffold is capable of crossing the blood-brain barrier and entering the CNS at time points favourable for a PET radiotracer, meaning future and optimised candidates from this structure could be more successful. [18F]TEFM78 is a promising ligand for PET imaging, however the radiosynthesis of [18F]TEFM78 should be improved to gain a higher molar activity to enable application of this technology in preclinical models of MS and potentially augment translational potential to clinical use.