6.1 SUMMARY

Inflammation is a dynamic and complex process in which multiple cell types and signaling pathways are involved, meant to restore homeostasis after injury or pathogen invasion. Comparable with the role of macrophages in peripheral inflammation, microglia cells play a prominent role in neuroinflammation. On the one hand, the role of microglia is beneficial by maintaining homeostasis, and being involved in brain development, neurogenesis and ageing [1,2]. On the other hand, activation of microglia has been implicated in the pathogenesis of several neurodegenerative and neurological diseases, e.g. Alzheimer’s disease (AD), Parkinson’s disease (PD), multiple sclerosis (MS), stroke, and possibly schizophrenia and depression [2,3]. In healthy conditions, microglia appear as ramified cells, scavenging the brain. In pathological conditions, microglia adopt a more amoeboid shape and move to the site of inflammation. Activation phenotypes of microglia appear as a continuum, with on one end a neuroprotective, anti-inflammatory phenotype and on the other end a neurotoxic, pro-inflammatory phenotype. Activation of microglia is dynamic, and suggested to change dependent on the stage of a disease, or even the disease itself. Upon microglial activation, receptor expression (e.g. of purinergic receptors) is altered, and differs between activation phenotypes [2].
Purinergic P2X7 and P2Y12 receptors are considered to be involved in the neuroinflammatory response. Their level of expression on activated microglia is dependent on the state of activation. While P2X7 receptor upregulation is associated with the pro-inflammatory phenotype, P2X12 receptor upregulation is associated with the anti-inflammatory phenotype. As neuroinflammation is a dynamic process, being able to study the expression of both receptors in vivo, could provide new insights in the exact role of microglial activation, in particular in the progression of neurodegenerative diseases. An excellent tool that can be used to serve this purpose is positron emission tomography (PET).

In chapter 1, a general introduction on the basic concepts of PET imaging and neuroinflammation is provided, in order to delineate the rationale behind the research described in this thesis.

Chapter 2 gives an overview of the most recent developments in the field of PET imaging of neuroinflammation. The main part of this overview is dedicated to PET tracers targeting the translocator protein 18 kDa (TSPO), which to date still is the most investigated target in imaging of microglial activation, both in preclinical and clinical studies. However, targeting TSPO comes with certain limitations, especially with respect to large differences in binding affinity of most tracers between human subjects, caused by the rs6971 single nucleotide polymorphism. Whereas focal and acute neuroinflammation in stroke and MS can be visualised fairly well using TSPO PET, imaging of neuroinflammation in AD is still hampered by low signal-to-noise ratios of the currently used tracers. Therefore, next to TSPO tracer development, also other existing and emerging molecular targets, including the P2X7 receptor, for development of tracers for imaging of neuroinflammation are discussed in chapter 2.

Over recent years, the P2X7 receptor gained increasing interest as a novel microglial target for in vivo imaging of neuroinflammation. In chapter 3, the synthesis of one of the first PET tracers targeting the P2X7 receptor is described. Desmethyl precursor (8) was synthesised via a 7-step synthesis route, and subsequently reacted with [11C]methyl iodide to obtain [11C]A-740003. Evaluation of this tracer in healthy rats revealed a moderate metabolic stability, but little to no brain uptake. Therefore, further in vivo evaluation of [11C]A-740003 was discontinued.
Nonetheless, [11C]A-740003, as well as its tritium-labelled analogue [[3H]A-740003, could be used in vitro to evaluate expression of theP2X7 receptor in animal models of neuroinflammation, which is described in chapter 4. Neuroinflammation can be induced locally in vivo in e.g. mice and rats by stereotactic injection of neurotoxins. Three different neurotoxins were used. Stereotactic injection of quinolic acid (QA) in striatum evokes a neuroinflammatory response, in which activated microglia predominantly adopt a pro-inflammatory phenotype. Autoradiography experiments on brain sections of this particular model revealed increased binding of [[3H]A-740003 in the lesioned striatum compared with the contralateral striatum. Binding of [[3H]A-740003 was shown to correlate with binding of [3H]PK-11195, a radioligand targeting TSPO, indicating a similar origin for upregulation of theP2X7 receptor and TSPO. Nevertheless, in another pro-inflammatory animal model of neuroinflammation, induced by intrastriatal injection of lipopolysaccharide (LPS), no difference in binding was observed between injected and non-injected striatum. Following intracerebroventricular injection of interleukin-4 (IL4), inducing an anti-inflammatory phenotype of activated microglia, no difference in binding of [11C]A-740003 was observed between affected and vehicle injected mouse brains. Taken together, in vitro autoradiography results in animal models of neuroinflammation are in line with the notion thatP2X7 receptor upregulation is associated with the pro-inflammatory microglial phenotype rather than the anti-inflammatory phenotype.

Adamantanylbenzamide 1, a potentP2X7 receptor antagonist developed by AstraZeneca [4], was shown enter the CNS, but also showed a poor pharmacokinetic profile. Based on this literature lead, Wilkinson et al. [5] have designed a series in which the hydrogen atoms at the bridgeheads of the adamantane moiety were substituted by fluorine atoms, to prevent in vivo hydroxylation of this position. Indeed, increased stability was achieved in rat and mouse liver microsomes [5]. In chapter 4, carbon-11 methylation of the adamantanylbenzamide series and subsequent preclinical evaluation is described. [11C]1-4 could all be obtained in high radiochemical yield and with high radiochemical purity and molar activity, and thus all four tracers were evaluated in healthy rats. Organ distribution, including sufficient brain uptake, was similar for all tracers. Metabolite analysis confirmed significantly increased metabolic stability of [11C]4 (42 ± 2% of intact tracer 45 min p.i.) compared with [11C]1-3 (25 ± 1%, 15 ± 2% and 16 ± 1% of intact tracer 45 min p.i., respectively) in rat plasma. Based on these results, PET imaging was performed using [11C]4 in a rat model with local overexpression of the humanP2X7 receptor, which was achieved via injection of an adeno-associated viral vector expressing the humanP2X7 receptor [6]. By unilateral injection of the humanP2X7 receptor expressing vector in striatum, the contralateral striatum could serve as internal control. Uptake of [11C]4 was 1.5-fold higher in the humanP2X7 receptor expressing striatum compared with the contralateral striatum (SUVs 2.85 ± 0.69 vs. 2.30 ± 0.36) from 2 min p.i. throughout the remainder of the scanning time (60 min). Uptake of [11C]4 could be blocked by pre-treatment with a non-structurally relatedP2X7 receptor antagonist (JNJ-47965567; 30 mg·kg-1, administered subcutaneously 45 min prior to tracer injection), indicating specific binding of [11C]4 to the humanP2X7 receptor in vivo.

Moving towards a more clinically relevant use of [11C]4, in vitro autoradiography was also performed on sections of post mortem brain material of human AD patients. Although immunohistochemical staining showed a slight increase ofP2X7 receptor expression in AD patients when compared with post mortem brain material of non-neurological age-matched controls, no significant difference in binding of [11C]4 was observed. This may be a matter of resolution, however, as already shown previously with TSPO PET, AD may not be the optimal disease for validation of novel tracers for neuroinflammation, as the response in AD is mild and non-focal, compared with a more acute and focal microglial response in for instance MS or stroke.

To identify novel targets for PET imaging of neuroinflammation, co-expression networks were generated from publicly available gene expression datasets, which is described in chapter 5. Analysis of expression data was guided by pre-defined criteria: i) expression preferentially in microglia compared with macrophages; ii) higher expression in anti-inflammatory microglia phenotype compared with non-stimulated or pro-inflammatory phenotype; iii) similar modulation in rodents as in humans. Apart from biological criteria, selected genes were also classified from a radiopharmaceutical perspective, e.g. preference for cell surface expression over intracellular expression and the availability of selective ligands in the literature to be used as lead compounds for PET tracer development. Keeping all criteria in mind, the P2X12 receptor came up as a promising translational biomarker. Subsequently, in in vitro experiments, P2ry12 mRNA was shown to be increased after treatment of primary human microglia cultures with IL4, but not after treatment with LPS or vehicle. In addition, a similar experiment in primary human macrophage cultures revealed no mRNA expression of Pry12 in any condition. Taken together, these data confirmed the potential of the P2X12 receptor as a target for PET imaging of the anti-inflammatory phenotype of microglia. As the P2X12 receptor is a well-known drug target for anti-coagulants, numerous selective antagonists with high affinity for this receptor have been developed. A highly potent P2X12 receptor antagonist [7] was selected as lead for development of the first PET tracer targeting the P2X12 receptor, [11C]5, which was obtained from precursors 3 and 6 via rhodium-mediated carbon-11 carbonylation. In vitro evaluation of [11C]5 indeed revealed higher tracer binding in IL4 injected mouse brains compared with vehicle injected mouse brains. Surprisingly, autoradiography on brain sections of a mouse model of stroke (MCAO) showed decreased tracer binding in the affected hemisphere compared with the contralateral hemisphere, and these results were confirmed by immunohistochemical staining for the P2X12 receptor. Decreased binding was most pronounced 3 days after MCAO, suggesting microglia switch to a more anti-inflammatory phenotype at later time points after stroke (7 and 10 days). In addition, the same phenomenon was observed in post mortem brain material of a patient who died from a stroke. These results confirm down-regulation of the P2X12 receptor in the pro-inflammatory microglial phenotype, and would allow for in vivo imaging of the dynamics of activated microglia in disease. However, in vivo evaluation in rats proved that [11C]5 is rapidly metabolised, likely due to cleavage of the ester moiety. Furthermore, no uptake of the tracer was observed in brain, and combined with the high uptake in duodenum, this points towards [11C]5 being a substrate for P-glycoprotein (P-gp), a brain efflux transporter.

References
[1] Jacobs AH, Tavitian B. Noninvasive molecular imaging of neuroinflammation. J. Cereb. Blood Flow Metab.
2012, 32, 1393-1415
[2] Colonna M, Butovsky O. Microglia Function in the Central Nervous System During Health and Neurodegeneration. Annu. Rev. Immunol. 2017, 35, 441-468
[3] Skaper SD, Facci L, Giusti P. Mast cells, glia and neuroinflammation: partners in crime? Immunology 2013, 141, 314-327
[4] Baxter A, Bent J, Bowers K, Braddock M, Brough S, Fagura M, Lawson M, McInally T, Mortimore M, Robertson M, Weaver R, Webborn P. Hit-to-Lead Studies: The Discovery of Potent Adamantane AmideP2X7 Receptor Antagonists. Bioorg. Med. Chem. Lett. 2003, 13, 4047-4050
[5] Wilkinson SM, Barron ML, O’Brien-Brown J, Janssen B, Stokes L, Werry EL, Chishty M, Skarratt KK, Ong JA, Hibbs DE, Vugts DJ, Fuller S, Windhorst AD, Kassiou M. Pharmalogical evaluation of novel bioisosteres of an adamantanyl benzamideP2X7 receptor antagonist. ACS Chem. Neurosci. 2017, doi: 10.102/acschemneuro.7b00272
[6] Ory D, Celen S, Gijsbers R, Van Den Haute C, Postnov A, Koole M, Vandeputte C, Andrés J-I, Alcazar J, De Angelis M, Langlois X, Bhattacharya A, Schmidt M, Letavic MA, Vanduffel W, Van Laere K, Verbruggen A, Debyser Z, Bormans G. Preclinical evaluation of aP2X7 receptor selective radiotracer: positron emission tomography studies in a rat model with local overexpression of the humanP2X7 receptor and in non-human primates. J. Nucl. Med. 2016, 57, 1436-1441
[7] Bach P, Antonsson T, Bylund R, Bjorkman JA, Osterlund K, Giordanetto F, van Giezen JJ, Andersen SM, Zachrisson H, Zetterberg F. Lead optimization of ethyl 6-aminonicotinate acyl sulfonamides as antagonists of the P2X12 receptor. Separation of the antithrombotic effect and bleeding for candidate drug AZD1283. J. Med. Chem. 2013, 56, 7015-7024