Radiosynthesis and initial preclinical evaluation of [¹¹C]AZD1283 as a potential P2Y12R PET radiotracer

Abstract

Intro:

Chronic neuroinflammation and microglial dysfunction are key features of many neurological diseases, including Alzheimer’s Disease and multiple sclerosis. While there is unfortunately a dearth of highly selective molecular imaging biomarkers/probes for studying microglia in vivo, P2Y12R has emerged as an attractive candidate PET biomarker being explored for this purpose. Importantly, P2Y12R is selectively expressed on microglia in the CNS and undergoes dynamic changes in expression according to inflammatory context (e.g., toxic versus beneficial/healing states), thus having the potential to reveal functional information about microglia in living subjects. Herein, we identified a high affinity, small molecule P2Y12R antagonist (AZD1283) to radiolabel and assess as a candidate radiotracer through in vitro assays and in vivo positron emission tomography (PET) imaging of both wild-type and total knockout mice and a non-human primate.

Methods:

First, we evaluated the metabolic stability and passive permeability of non-radioactive AZD1283 in vitro. Next, we radiolabeled [¹¹C]AZD1283 with radioactive precursor [¹¹C]NH₄CN and determined stability in formulation and human plasma. Finally, we investigated the in vivo stability and kinetics of [¹¹C]AZD1283 via dynamic PET imaging of naïve wild-type mice, P2Y12R knockout mouse, and a rhesus macaque.

Results:

We determined the half-life of AZD1283 in mouse and human liver microsomes to be 37 and > 160 min, respectively, and predicted passive CNS uptake with a small amount of active efflux, using a Caco-2 assay. Our radiolabeling efforts afforded [¹¹C]AZD1283 in an activity of 12.69 ± 10.64 mCi with high chemical and radiochemical purity (>99%) and molar activity of 1142.84 ± 504.73 mCi/μmol (average of n = 3). Of note, we found [¹¹C]AZD1283 to be highly stable in vitro, with >99% intact tracer present after 90 min of incubation in formulation and 60 min of incubation in human serum. PET imaging revealed negligible brain signal in healthy wild-type mice (n = 3) and a P2Y12 knockout mouse (0.55 ± 0.37%ID/g at 5 min post injection). Strikingly, high signal was detected in the liver of all mice within the first 20 min of administration (peak uptake = 58.28 ± 18.75%ID/g at 5 min post injection) and persisted for the remaining duration of the scan. Ex vivo gamma counting of mouse tissues at 60 min post-injection mirrored in vivo data with a mean %ID/g of 0.9% ± 0.40, 0.02% ± 0.01, and 106 ± 29.70% in the blood, brain, and liver, respectively (n = 4). High performance liquid chromatography (HPLC) analysis of murine blood and liver metabolite samples revealed a single radioactive peak (relative area under peak: 100%), representing intact tracer. Finally, PET imaging of a rhesus macaque also revealed negligible CNS uptake/binding in monkey brain (peak uptake = 0.37 Standard Uptake Values (SUV)).

Conclusion:

Despite our initial encouraging liver microsome and Caco-2 monolayer data, in addition to the observed high stability of [¹¹C]AZD1283 in formulation and human serum, in vivo brain uptake was negligible and rapid accumulation was observed in the liver of both naïve wildtype and P2Y12R knockout mice. Liver signal appeared to be independent of both metabolism and P2Y12R expression due to the confirmation of intact tracer in this tissue for both wildtype and P2Y12R knockout mice. In Rhesus Macaque, negligible uptake of [¹¹C] AZD1283 brain indicates a lack of potential for translation or its further investigation in vivo. P2Y12R is an extremely promising potential PET biomarker, and the data presented here suggests encouraging metabolic stability for this scaffold; however, the mechanism of liver uptake in mice should be elucidated prior to further analogue development.

1. Introduction

Microglia, the resident innate immune effector cells of the central nervous system (CNS), perform critical functions to maintain a healthy brain environment including passive surveillance of the brain, rapid response to insult or infection, and post-injury repair [1,2]. Conversely, chronic maladaptive microglial activation plays a significant role in driving pathology in neurological diseases such as Alzheimer’s Disease (AD) and multiple sclerosis. However, there are no robust methods to specifically detect these cells and their function in vivo. Positron emission tomography (PET) is a highly sensitive functional imaging modality well-suited to non-invasive, longitudinal visualization of molecular processes relevant to such diseases. Unfortunately, currently available PET biomarkers for microglial function suffer from significant drawbacks [3]. For example, translocator protein 18 kDa (TSPO), the most common PET biomarker for microglial imaging, is hampered by its unclear functional role in neuroinflammation, low dynamic range between healthy and diseased subjects, as well as variable binding affinity due to a genetic polymorphism in humans (Ala147Thr) [4]. Moreover, expression of TSPO, along with other emerging PET markers of interest (e.g., CSF1R, CB2, P2X7), is not entirely specific to microglia, but is also detected on peripheral myeloid lineage cells, astrocytes, neurons, and/or endothelial cells [3]. Cumulatively, these factors complicate image interpretation and underscore the need for more specific, dynamic, and functionally relevant PET biomarkers for microglial function.

A recent genetic study identified the adenosine diphopshate (ADP) receptor, P2Y12R, as an important gene strongly associated with the homeostatic phenotype of microglial activation. In the healthy adult brain, P2Y12R protein is constitutively and exclusively expressed on microglia, with rapid and significant reduction upon the onset of a pro- inflammatory state [5,6]. Notably, microglial P2Y12R recognizes ADP released in the context of brain injury/infection and plays a key role in driving morphological changes, activation, and migration of microglia to sites of injury [5,7,8]. Changes in P2Y12R expression and function are dynamic and complex, varying both temporally and spatially in conditions with a strong neuroinflammatory component. Numerous published studies have identified P2Y12R as a promising functional biomarker of neuroinflammation in multiple difficult to manage diseases including AD, multiple sclerosis, and neuropathic pain [5,7,9–15]. This combination of microglial-specific CNS expression, functional relevance in the context of neuroinflammation, and its established role in multiple clinical contexts makes P2Y12R a highly promising and potentially versatile PET biomarker for detecting and studying microglial function in different disease contexts.

Although P2Y12R is a well-characterized pharmacological target, most known antagonists are not suitable for use as CNS PET tracers. De Vries and colleagues labeled the first radiotracer candidate for P2Y12R, [¹¹C]5, a piperazinyl-pyridine urea initially designed as a therapeutic P2Y12R antagonist (Fig. 1) [15]. [¹¹C]5 was thoroughly studied in vivo in healthy rats and in vitro using autoradiography of mouse and human postmortem brain tissue from stroked and healthy control subjects. Although autoradiography results were extremely promising, [¹¹C]5 is not a viable tracer for in vivo analyses due to reported low metabolic stability and poor brain uptake [16]. More recently, this same team of researchers have radiolabeled and evaluated two structurally distinct high affinity thienopyrimidine-based P2Y12R antagonists [17]. Both tracers were found to be highly specific for P2Y12R (via autoradiography) and blood-brain barrier (BBB) penetrant in vivo in healthy rats, however, transporter mediated active efflux precluded further development [17].

Fig. 1.

P2Y12R Antagonists (A) AZD1283 with potential radio-labelling site highlighted in red, (B) [¹¹C]5, and (C), (D) new thienopyridine based radiotracers with labelling site highlighted in red. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Following an extensive literature search, we identified the clinical drug candidate AZD1283 as a highly promising P2Y12R PET radiotracer candidate. AZD1283 is an acyl sulfonamide based P2Y12R antagonist amenable to ¹¹C-radiolabeling related to, but structurally distinct from, [¹¹C]5. Specifically, both molecules were designed during the same drug development process: AZD1283 was discovered during efforts to optimize the metabolic stability and solubility of [¹¹C]5, which has somewhat higher affinity (IC₅₀ of 6.3 nM versus 11 nM for AZD1283) [16,18,19]. Encouraged by the improved pharmacokinetic/pharmacodynamic properties of AZD1283, we elected to radiolabel this molecule and evaluate its imaging properties. Our specific goals were to 1) characterize [¹¹C]AZD1283 both in vitro and in vivo and 2) evaluate the translational potential of [¹¹C]AZD1283 as a CNS PET radiotracer. Here we report PET imaging of [¹¹C]AZD1283 in wild-type and P2Y12R knockout mice, as well as non-human primates, in order to account for considerable inter-species differences between rodents and primates when gauging translational potential. Finally, we also report ex vivo biodistribution and metabolite analysis in mice, in order to further understand the fate of this molecule in vivo.

2. Materials and methods

2.1. Caco-2 permeability assay

Caco-2 cells were maintained in Dulbecco’s Modified Eagle Medium (DMEM) in an atmosphere of 5% CO₂. For transport experiments, cells were seeded at a concentration of 5 × 10⁵/well on polycarbonate filter inserts and monolayers allowed to grow and differentiate for 21 ± 4 days before use in experiments. Apparent permeability coefficients were determined for apical(A) to basal(B), A → B, and B → A directions with and without the presence of a transporter inhibitor (Elacridar). Test items and reference compounds were dissolved in Hank’s balanced salt solution (HBSS) at pH 7.4 to yield a final concentration of 10 μM. Assays were performed in HBSS containing 25 mM HEPES (pH 7.4) at 37 °C. Prior to the study, monolayers were washed in pre-warmed HBSS. At the start of the experiments, prewarmed HBSS containing test compound(s) was added to the donor side of the monolayer, and pure HBSS was added to the receiver side. Aliquots were taken from the receiver side over the course of the 2 h incubation period; aliquots of the donor side were taken at 0 and 2 h. Aliquots were diluted with an equal volume of methanol/water with 0.1% formic acid containing the internal standard, and the mixture was analyzed by Liquid Chromatography-Mass Spectrometry (LC-MS/MS). The apparent permeability coefficients (P_app) were calculated using the formula: P_app = (dC_rec/dt)/(A × C_{0, donor}) × 10⁶, with dC_rev/dt being the change in concentration in the receiver compartment with time, C_0,donor the concentration in the donor compartment at time 0, and A the area of the compartment with the cells.

2.2. Microsome stability assay

Human or mouse liver microsomes were separately combined at a final concentration of 11.25 mg protein/compound with K₃PO₄ (100 mM; pH 7.4), MgCl₂ (10 mM), and test compound (1 μM), and pre-incubated (10 min, 37 °C). NADPH (1 mM) was added to initiate reactions (total volume 100 μL). Reactions were quenched at sequential timepoints (0, 10, 20, and 40 min) with Clem stop solution (100 μL, 625 ng/mL) (Cyprotex) in acetonitrile. Samples were centrifuged at 4000g for 20 min, diluted (75 μL into 75 μL 0.1% formic acid in water), and analyzed by LC-MS/MS.

2.3. Radiosynthesis and characterization

2.3.1. General considerations

All chemicals were acquired from commercial sources and used without further purification unless otherwise stated: The synthetic precursor for radiolabeling, a brominated analogue of AZD1283, was synthesized to order by ACME Bioscience (Palo Alto, CA) and used without further purification (For characterization and quality control details, please see Supplementary Information).

2.3.2. Radiosynthesis and purification

The synthesis of [¹¹C]cyanide precursor was performed using a GE Process Cabinet (ProCab) as follows: 4.77 Ci of [¹¹C]CO₂ were generated with GE PETtrace cyclotron via the ¹⁴N(p,α)¹¹C reaction. [¹¹C]CO₂ was purified through a molecular sieves column (4 Å, 60–80 mesh), combined with a stream of H₂ and flowed across a bed of nickel catalyst at 400 °C to generate [¹¹C]CH₄, which was subsequently reacted with NH₃ gas in the presence of platinum catalyst (tightly coiled platinum wire) at 950 °C to yield [¹¹C]NH₄CN; this was delivered to a GE TRACERlab FX_FN synthesis module and trapped in a mixture of 0.5 M aqueous KOH (2 μL) and a solution of kryptofix 2.2.2 (K_2.2.2, 3 mg) in tetrahydrofuran (THF, 1 mL) previously placed in the module’s reactor. In order to minimize the ammonia concentration in the reaction solution, the ProCab outlet line was connected to the reactor immediately before the delivery of the [¹¹C]NH₄CN. After the trapping of the radioactive precursor (as evidenced by a radioactivity plateau in the reactor radioactivity detector), a solution of synthetic precursor (Br-1) (1 mg, 0.002 mmol) and tetrakis- triphenylphosphine Palladium(0) (Pd(PPh₃)₄, 2 mg, 0.002 mmol) in THF (500 μL) was added to the reactor. The resulting mixture was heated to 90 °C and stirred for 5 min in a pressurized reactor. The reaction mixture was then cooled to room temperature, and diluted with 70:30 MeCN:H₂O (500 μL). The crude reaction mixture was purified via semipreparative HPLC (column: Phenomenex Gemini 5 mm C18 110 Å, 250 × 10 mm; mobile phase A: H₂O with 0.1% trifluoroacetic acid (TFA) by volume; mobile phase B: MeCN with 0.1% TFA by volume; program: 50–90% B in 20 min at 5 mL/min; standard retention time: 11.5 min) to furnish radiochemically pure [¹¹C]AZD1283. Tracer was trapped on a C-18 plus lite sep pak (Waters), rinsed with water (10 mL), and then eluted to a product vial with EtOH (0.5 mL) and saline (0.9% NaCl for injection, 4.5 mL) to provide formulated, radiochemically pure [¹¹C]AZD1283 suitable for injection. Purity and identity of final product were confirmed via analytical HPLC (column: Phenomenex Gemini 5 mm C18 110 Å, 250 × 4.6 mm; mobile phase A: H₂O with 0.1% TFA by volume; mobile phase B: MeCN with 0.1% TFA by volume; program: 70–95% B in 15 min at 1 mL/min; standard retention time: 7.5 min).

2.3.3. In vitro human plasma stability

To assess stability in serum, 20 μL of formulated [¹¹C]AZD1283 (46 μCi) was added to 330 μL of human plasma, and the mixture vortexed briefly. Aliquots (70 μL) were transferred to individual Eppendorf tubes and incubated at 37 °C. Samples were quenched at regular intervals (0, 5, 15, 30, and 60 min) with 140 μL of ice-cold MeCN and centrifuged at 12,500g for 10 min. Supernatant was transferred to an HPLC vial and analyzed via HPLC (Agilent LC 1200 series) using analytical QC conditions as described above. Controls consisting of radiotracer incubated without plasma were run at 0 and 60 min.

2.4. Preclinical evaluation and ex vivo analysis in mice

2.4.1. General

Rodent experiments were conducted in accordance with the Administrative Panel on Laboratory Animal Care (APLAC) at Stanford University, which is accredited by the Association for the Assessment and Accreditation of Laboratory Animal Care (AAALAC International).

All mice were housed in groups of 5 in a temperature-controlled environment under a 12-h light/dark schedule with ad libitum access to food and water. For all imaging and ex vivo studies, female C57BL/6 naïve wild-type mice were used (3 months old) with the exception of one male P2Y12R knockout mouse (C57BL/6 background; 11 months old) for in vivo 60 min dynamic PET imaging. Based on preliminary studies with the available mice, sex as a variable was determined not to impact the results or conclusions of this study. Consequently, experiments were not repeated with sex-matched mice given the need to limit the number of mice in accordance with our APLAC guidelines. Prior to imaging, mice were anesthetized using isoflurane gas (2–3.5% for induction, 1–2.5% for maintenance) and ophthalmic ointment placed in both eyes to prevent corneal ulcers. A proper anesthetic depth, as determined by the lack of response to a toe pinch and a regular respiratory rate, was ensured for each mouse prior to placement in the rodent PET/CT scanner. All mice were maintained under anesthesia while injected intravenously with the radiotracer and for the duration of PET/CT imaging, and respiration was monitored for the duration of image acquisition.

2.4.2. Dynamic PET/CT imaging and analysis of [¹¹C]AZD1283 in mice

Four mice (3 naïve wild-type females and one P2Y12R total knockout male) were placed in a 2 × 2 mouse holder and loaded into an Inveon dPET (Siemens) for 60 min dynamic PET imaging. Data were acquired in list mode format over 60 min (4 × 15 second frames, 4 × 1 minute frames, 11 × 5 minute frames) commencing just prior to tail-vein injection of tracer (300 μCi) and reconstructed via OSEM-3D with scatter correction. A transmission scan was acquired to correct for attenuation during image reconstruction. Following PET imaging, mice were transferred to a SOFIE GNEXT PET/CT for CT acquisition. PET/CT data were co-registered and analyzed using Inveon Research Workplace software to generate images and quantify tracer uptake in regions of interest (ROIs). A 3D ROI was manually defined for the whole brain using a summed_to_minute image, then applied to the full dynamic data set to generate a time activity curve (TAC).

2.4.3. Murine ex vivo biodistribution studies

Following acquisition of 60 min dynamic PET data and CT, blood was collected via cardiac puncture (at approximately 65 min post-injection while mice were still anesthetized) and all mice immediately euthanized via transcardial perfusion and subsequent thoracotomy. Briefly, an incision was made below the diaphragm and ribcage, and the diaphragm cut open to expose the heart. Subsequently, A 25 gauge needle (Covidien Monoject) was inserted into the left ventricle. The right atrium was cut to allow flow and the animal transcardially perfused by pushing phosphate buffered saline (Gibco) through the left ventricle (either 20 mL or until the liver is cleared of blood). Following perfusion, organs (blood, brain, heart, kidney, liver, lung, spleen) were collected from mice and weighed. Radioactivity in each organ of interest was quantified via counting with a gamma counter (Hidex AMG). Radioactivity was expressed as %ID/g in each organ.

2.4.4. Ex vivo determination of radiometabolites in blood and liver

[¹¹C]AZD1283 was administrated to a separate group of naïve wild-type mice, and after thirty minutes blood was collected via cardiac puncture and livers removed following euthanization via cervical dislocation and transcardial perfusion. Blood (500 μL) was placed in a heparinized tube and centrifuged at 1800g for 4 min. 200 μL of separated plasma was transferred to a tube containing ice cold acetonitrile (300 μL) and centrifuged at 9400g for 4 min. For liver samples, a single lobe was placed in a vial containing 500 μL of ice-cold acetonitrile and homogenized via immersion blender, keeping samples on ice for this process to minimize heating during homogenization. Samples were centrifuged at 9400g for 4 min. All samples were kept on ice between processing steps. For all samples, 150 μL supernatant was transferred to an HPLC vial for analysis. 100 μL of supernatant and whole pellet were transferred to separate tubes for gamma counting (Hidex AMG) and quantification of extraction efficiency.

2.5. Preclinical evaluation in rhesus macaque

2.5.1. General considerations

Primate studies were conducted in accordance with the standards set by the University of Michigan Institutional Animal Care & Use Committee (IACUC), which is accredited by the Association for the Assessment and Accreditation of Laboratory Animal Care (AAALAC International).

The University of Michigan PET Center has maintained 2 rhesus macaques for ~19 years and the monkeys are individually housed in adjacent steel cages (83.3 cm high × 152.4 cm wide × 78.8 cm deep) equipped with foraging boxes. They are currently housed in adjacent cages as repeated attempts to socially house them in the same cage have been unsuccessful due to aggressive incompatibility. Cages are metal and contain gridded floors for radiation safety reasons (radioactive waste is contained to the gridded floor and is easier to clean). Temperature and humidity are carefully controlled, and the monkeys are kept on a 12 h light/12 h dark schedule. Monkeys are fed Lab Fiber Plus Monkey Diet (PMI Nutrition Intl. LLC, Shoreview MN, USA) that is supplemented with fresh fruit and vegetables daily. Water and enrichment toys (manipulanda and food-based treats) are available continuously in the home cage.

2.5.2. MicroPET imaging and analysis of [¹¹C]AZD1283 in rhesus macaque

Imaging studies were performed on a Concorde P4 microPET (Knoxville, TN) in an intact, mature female rhesus monkey (n = 1, animal weight 9.36 kg). The animal was anesthetized in the home cage with Telazol and transported to the PET facility, where the subject was intubated for mechanical ventilation; anesthesia was continued with isoflurane and maintained throughout the duration of the PET scan. A venous catheter was inserted into one hind limb and the monkey was placed on the PET gantry with its head secured to prevent motion artifacts. Ten minutes later, 4.90 mCi of [¹¹C]AZD1283 was administered in a bolus dose over 1 min, and the brain imaged for 60 min (5 × 1 min frames, 2 × 2.5 min frames, 2 × 5 min frames, 4 × 10 min frames). Emission data were collected beginning with the injection and continued for 60 min. Vitals (HR, SPO₂, EtCO₂ and respiratory rate) were collected for the entire duration of the scan. Data were corrected for attenuation and scatter and reconstructed using the three-dimensional–maximum a priori method (3D MAP algorithm). A ROI was defined for the whole brain on multiple planes by using a summed image. The volumetric ROI was then applied to a full dynamic data set to generate a TAC.

3. Results and discussion

3.1. Preliminary characterization of AZD1283

First, we assessed key pharmacokinetic and pharmacodynamic properties to characterize AZD1283 in our own hands. We found the half-life to be 37 and > 159 min through in vitro incubation with mouse and human liver microsomes, respectively. Notably, this half-life is compatible with our selected imaging isotope, carbon-11 (t_1/2 = 20 min). Caco-2 monolayer assays (a proxy for lipophilic membrane permeability) indicated passive CNS uptake of AZD1283 with likelihood for minor efflux (AB = 4.48, BA/AB = 5.3). To supplement these data, we used a CNS Multiple Parameter Optimization (MPO) tool to evaluate relevant computational properties (e.g. ClogP, molecular weight ‘MW,’ total polar surface area ‘tPSA’) in order to gauge the probability of CNS penetration [20]. We determined the CNS MPO score for AZD1283 to be 4.2, indicative of passive CNS uptake and alignment of key ADME properties [20]. Comparison of these data with published data on [¹¹C]5 was particularly encouraging: MPO scores for the two molecules (4.2 for AZD1283 vs. 3.6 for [¹¹C]5) demonstrated AZD1283 to have properties more conducive CNS uptake (e.g., lower tPSA and MW). While cLogP, a common metric for lipophilicity, was closer to ideal for [¹¹C]5, published data for the logD of AZD1283 (2.8, determined at pH 6.8) was reassuring for lipophilicity conducive to passive CNS uptake [19]. Additionally, in light of low metabolic stability and lack of CNS uptake for [¹¹C]5, liver microsome and caco-2 assays were promising for in vivo imaging with AZD1283 [16]. Cumulatively, these preliminary data, in conjunction with the excellent in vitro performance of the structurally similar [¹¹C]5, and improved PK/PD and physicochemical properties of AZD1283, motivated us to proceed with radiosynthesis of and in vivo imaging with [¹¹C]AZD1283.

3.2. Radiosynthesis and in vitro characterization of [¹¹C]AZD1283

We radiolabeled [¹¹C]AZD1283 with [¹¹C]HCN via palladium-catalyzed cross coupling to cyanate a brominated precursor molecule (Fig. 2). [¹¹C]AZD1283 was reliably synthesized (n = 3) in sufficient yield (12.69 ± 10.64 mCi) and molar activity (1142.84 ± 504.73 mCi/μmol), with high chemical and radiochemical purity (>99%) confirmed via analytical HPLC (Fig. 2). Purified [¹¹C]AZD1283 was trapped on a C18 plus light sep-pak and reformulated into saline (4.5 mL) containing 10% ethanol (0.5 mL), suitable for administration to animals for in vivo imaging studies. Prior to in vivo studies, we tested the stability of [¹¹C] AZD1283 in human plasma and formulation vehicle: HPLC analysis after 0, 5, 15, 30, and 60 min incubation of formulated tracer in human plasma showed [¹¹C]AZD1283 to be metabolically stable in vitro, with no loss of parent or detection of radio-metabolites after 60 min (Fig. 3). The radiotracer was also stable in formulation and resistant to radiolysis for >90 min post end-of-synthesis (supp Fig. 1). While susceptibility of the carboxylic acid ester moiety on the substituted pyridine ring of [¹¹C] AZD1283 to esterases in the blood and subsequent short metabolic half-life were concerns, the demonstrated in vitro stability in plasma suggests that tracer amounts of [¹¹C]AZD1283 may be stable in the blood on a timescale compatible with carbon-11 PET imaging studies. Resistance to radiolysis also alleviates a need for ascorbic acid in formulation, further simplifying and streamlining the radiosynthesis of this molecule. These encouraging preliminary data, in conjunction with a straightforward and robust radiolabeling method, motivated us to evaluate our new tracer in both mice and non-human primates.

Fig. 2.

(A) Radiosynthetic scheme for [¹¹C]AZD1283. (B) HPLC chromatograms confirming identity of [¹¹C]AZD1283 (gamma, red) co-injected with cold AZD1283 standard (UV 214 nm, blue). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 3.

HPLC analysis of human plasma samples after 0 (red), 5 (blue), 15 (green), 30 (pink), and 60 (brown) minutes incubation showed [¹¹C]AZD1283 to be metabolically stable in vitro, with no loss of parent tracer or appearance of radio-metabolites. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.3. Murine microPET imaging, ex vivo biodistribution, and metabolism studies

Next, we performed 60 min dynamic [¹¹C]AZD1283 microPET/CT imaging in naïve wild-type female mice (n = 3) (Fig. 4) and a P2Y12R total KO male mouse (n = 1) (supp fig). Negligible signal was detected in the brain of healthy wild-type mice over the course of the dynamic PET study, indicating a lack of passive BBB permeability (peak uptake = 3.46 ± 2.0%ID/g at 15 s post injection and decreasing to 0.55 ± 0.37%ID/g at 5 min post injection, indicative of perfusion as opposed to uptake) (Fig. 5) (Supp Fig. 4). Strikingly, high signal was detected in the liver within 5 min of administration (58.28 ± 18.75%ID/g), and persisted for the remaining duration of the scan (Fig. 5). Dynamic PET imaging in the P2Y12R knockout mouse revealed a very similar pattern of uptake and distribution, showing no CNS uptake and minimal binding outside the liver, while slightly increased signal in the kidney at the earliest time-points relative to naïve wild-type mice was observed.

Fig. 4.

Summed representative PET/CT images of a naïve wild-type female mouse at (A) 0–45 s, (B) 10–20 min, (C) 30–40 min and (D) 50–60 min after intravenous injection of [¹¹C]AZD1283. H = heart, Liv = liver, and black arrows point to carotid arteries (i.e., tracer signal in blood pool).

Fig. 5.

(A) Time-activity curves (TACs) showing PET signal in the whole brain of 3 naïve wild-type mice and 1 P2Y12 KO mouse over the duration of a 60 min dynamic scan. (B) TAC of PET signal in the liver of 3 naïve wild-type mice throughout 60 min dynamic scan. (C) Graph of radioactive signal in tissues from naïve wild-type mice (n = 4) using ex vivo gamma counting. (D) HPLC chromatograms of processed supernatant extracted from liver and blood samples from mice perfused 30 min after injection of [¹¹C]AZD1283.

We corroborated in vivo imaging results with ex vivo biodistribution for four naïve mice, which demonstrated a mean %ID/g of 0.9% ± 0.40, 0.02% ± 0.01, and 106 ± 29.70% in blood, brain, and liver, respectively (Fig. 5). In order to further elucidate the utility of this molecule in vivo and better characterize the dominant radioactive species contributing to the high liver signal, we administered [¹¹C]AZD1283 to three additional naïve mice and sacrificed 30 min after administration of the radiotracer. Blood (n = 2) and liver (n = 3) samples were collected for metabolite analysis, and analyzed via HPLC (Fig. 5). Notably, HPLC analysis of both blood and liver metabolite samples revealed a single radioactive peak consistent with that of intact [¹¹C]AZD1283. Gamma counting of supernatant and pellet prior to HPLC analysis confirmed extraction efficiency to be 64.43% and 63.15% for liver (n = 1) and blood (n = 1) samples, respectively.

Despite promising in vitro data, the negligible brain uptake and rapid accumulation of radioactivity in the liver demonstrated by PET studies show that [¹¹C]AZD1283 is not suitable for in vivo imaging of P2Y12R in the CNS in mice. Interestingly, ex vivo biodistribution confirmed minimal signal remaining in the blood, despite P2Y12R expression on platelets. This pattern of uptake and distribution, with hepatic accumulation, was also observed upon imaging of a P2Y12R total knockout mouse. Together, these data suggest that the tracer is being rapidly taken up (or metabolized) in a fashion unrelated to specific binding by or interaction with P2Y12R. While this pattern of uptake and distribution, in conjunction with the presence of functional groups including a carboxylic acid ester, is initially indicative of rapid metabolism, multiple studies (in vitro human plasma stability, in vitro liver microsome stability in both mice and humans, and ex vivo analysis of both blood and liver in mice) suggest the molecule is metabolically stable at tracer concentrations. Furthermore, published selectivity studies confirm AZD1283 is selective for P2Y12R over other purinergic receptors; that is, does not display significant binding to any other related receptors [21]. Cumulatively, these data suggest that [¹¹C]AZD1283 is rapidly taken by the liver via a P2Y12R- and metabolism-independent mechanism. In addition to rapid liver uptake, [¹¹C]AZD1283 does not cross the intact BBB [22]. This may be due to a site on the scaffold, likely the sulfonamide proton, that is susceptible to deprotonation in vivo, as suggested by the need for trifluoracetic acid in HPLC solvents; such a reaction would yield a predominantly negatively charged species unlikely to cross the BBB. Alternatively, rapid sequestration of the tracer by the liver may limit bioavailability for brain uptake. Furthermore, the relatively high molecular weight and polar surface area, while an improvement over previously tested tracers, may still be prohibitively high for CNS uptake of this small molecule.

3.4. PET imaging of rhesus macaque

While [¹¹C]AZD1283 did not enable in vivo detection of P2Y12R in the CNS of mice, the appreciable differences in physiology (e.g. metabolic rate, efflux transporter expression on BBB) between mice and higher species such as nonhuman primates motivated us to also assess this tracer in a rhesus macaque in order to more accurately predict in vivo behavior in humans. Similar to our murine studies, 60 min dynamic PET imaging of the rhesus macaque brain also demonstrated negligible CNS uptake/binding, as reflected by sagittal brain PET images (n = 1), and low peak brain SUV (peak SUV = 0.37) (Fig. 6). Notably, in contrast to images in mice where activity localized almost exclusively to the liver, radioactivity is seen throughout the head in rhesus (e.g. in the nasal cavity), but not within the brain. However, it is not possible to rule out relatively high liver uptake in primates, given that only the head was scanned due to dimensional constraints of the field of view of the scanner used for image acquisition. Collectively, these data confirm that [¹¹C]AZD1283 is not suitable for further evaluation for in vivo CNS imaging of P2Y12R due to a lack of CNS uptake, most likely due to an inability to passively penetrate the intact BBB.

Fig. 6.

Sagittal PET image of rhesus macaque summed 0–60 min showed negligible CNS uptake/binding, reflected by (A) Sagittal microPET brain images (n = 2), and (B) low peak brain SUV (peak SUV = 0.37).

4. Conclusion

We reliably synthesized [¹¹C]AZD1283, a PET radiotracer based on a known high affinity P2Y12R antagonist, in high radiochemical purity and good molar activity. [¹¹C]AZD1283 was found to rapidly accumulate in the liver of both naïve wildtype and P2Y12R total knockout mice in vivo. Ex vivo metabolite analyses confirm that this signal is likely due to intact tracer, suggesting a mechanism independent of both P2Y12R expression and metabolism governing liver uptake in mice. Furthermore, published studies confirming selectivity of AZD1283 for P2Y12R suggests that binding of AZD1283 to another target is presumably not a potential confounder [21]. In the healthy rhesus macaque, negligible CNS uptake indicated a lack of potential for in vivo imaging or clinical translation for this molecule. P2Y12R remains a target of great interest for PET imaging, and a wealth of literature concerning the AZD1283 chemical scaffold make [¹¹C]AZD1283 a promising starting point for development and optimization of a new generation of improved radiotracers. While the data presented here are encouraging for metabolic stability of this scaffold, the mechanism of liver uptake in mice should be elucidated prior to development of further analogues based on this scaffold. Future work will elucidate the factors impacting rapid liver uptake, and leverage published X-ray crystallography and quantitative structure activity relationship (QSAR) to develop novel high affinity P2Y12R antagonists amenable to radiolabeling, with properties conducive to CNS uptake (i.e., lower molecular weight, lower tPSA, and minimizing negative charge at physiologic pH). Alternatively, structurally distinct pharmacophores are being considered in parallel. Development and translation of a P2Y12R PET tracer suitable for use in vivo will be a significant step forward in PET imaging of microglia, and thus warrants further research efforts. Such tracer would have the potential to greatly enhance our ability to noninvasively study functional alterations in microglia and neuroinflammation in a host of neurological diseases.