Synthesis and Initial In Vivo Evaluation of [11C]AZ683—A Novel PET Radiotracer for Colony Stimulating Factor 1 Receptor (CSF1R)

Positron emission tomography (PET) imaging of Colony Stimulating Factor 1 Receptor (CSF1R) is a new strategy for quantifying both neuroinflammation and inflammation in the periphery since CSF1R is expressed on microglia and macrophages. AZ683 has high affinity for CSF1R (Ki = 8 nM; IC50 = 6 nM) and >250-fold selectivity over 95 other kinases. In this paper, we report the radiosynthesis of [11C]AZ683 and initial evaluation of its use in CSF1R PET. [11C]AZ683 was synthesized by 11C-methylation of the desmethyl precursor with [11C]MeOTf in 3.0% non-corrected activity yield (based upon [11C]MeOTf), >99% radiochemical purity and high molar activity. Preliminary PET imaging with [11C]AZ683 revealed low brain uptake in rodents and nonhuman primates, suggesting that imaging neuroinflammation could be challenging but that the radiopharmaceutical could still be useful for peripheral imaging of inflammation.


Introduction
Colony Stimulating Factor 1 Receptor (CSF1R, M-CSF, or cFMS) is a class III receptor tyrosine kinase [1] that regulates immune response by controlling the survival and activity of macrophages and macrophage-like cells [2]. Aberrant activity of CSF1R, or its endogenous ligands (CSF1 and IL-34), plays a role in many disorders that have an immune/inflammatory component [3]. Specifically, chronic inflammation caused by increased activity of macrophages due to increased CSF1R response is present in many autoimmune disorders such as rheumatoid arthritis, inflammatory bowel disease, and autoimmune nephritis, among others [4,5]. The contribution of CSF1R to symptomatic Alzheimer's Disease (AD) is also well known, due in part to its proliferative effects on microglia, which are associated with neuroinflammation, a hallmark clinical symptom of AD [6,7]. A mechanism for CSF1R involvement in inter-neuronal transmission of pathogenic tau protein by microglia was also recently elucidated [8]. Involvement of CSF1R in certain types of cancers, such as gliomas, also correlates with poor disease prognosis, as proliferation of CSF1R-controlled tumor-associated macrophages (TAMs) correlates with tumor angiogenesis and metastasis [4,[9][10][11].
Consequently, CSF1R inhibitors (both small molecules and biologics) have been proposed as a means of controlling inflammation in this multitude of diseases and disorders via macrophage depletion/regulation [12]. Many CSF1R inhibitors, including some that are kinase-specific, can be found in both academic and patent literature [4,5,13], and several have proceeded to clinical trials for PET imaging of CSF1R therefore remains underdeveloped and herein we attempt to address this issue through development of [ 11 C]AZ683, a new radiopharmaceutical for CSF1R. AZ683 (Figure 1) was selected because it has >250-fold selectivity for CSF1R over 95 other kinases, low plasma protein binding, a good pharmacokinetic (PK) profile, and both fluorine and N-methyl moieties which are potential sites for radiolabeling with 18 F or 11 C, respectively [22][23][24]. Moreover, AZ683 has low nanomolar affinity for CSF1R (Ki = 8 nM; IC50 = 6 nM), making it ideal for PET studies which typically utilize nanomoles-picomoles of radiotracer, and the cLogP of the neutral (uncharged) compound is 3.1 which suggests that it should cross the BBB. Since N-methylation of the desmethyl precursor with [ 11 C]MeI (or [ 11 C]MeOTf) was envisioned to be simpler than 18 F-labeling of this scaffold, the synthesis and carbon-11 radiolabeling of [ 11 C]AZ683 was undertaken for initial evaluation and is described herein. We also report preliminary evaluation of the radiotracer as a CSF1R imaging agent in rodent and non-human primate PET imaging studies.

Synthesis of Reference Standard and Precursor
AZ683 reference standard 6a and N-desmethyl precursor 7 were synthesized via modified literature procedures in five and six steps, respectively (Scheme 1) [23]. Both syntheses diverged from a common intermediate 4. This common intermediate was synthesized via condensation of 4-bromo-3-ethoxyaniline (1) with diethylethoxymethylenemalonate to yield 2. This was followed by cyclization/chlorination with POCl3 and tetrabutylammonium chloride to form chloroquinoline 3. A subsequent SNAr reaction with 2,4-difluoroaniline yielded intermediate 4. Buchwald-Hartwig crosscoupling was then used to couple either N-Boc piperazine or N-methylpiperazine with 4, yielding intermediates 5a and 5b for the reference standard and precursor, respectively. Amidation of the ethyl ester of 5 was performed using formamide/NaOEt to generate reference standard 6a and N-Boc protected precursor 6b. Final deprotection of the Boc group of 6b with trimethylsilyl chloride in methanol furnished precursor 7. Precursor 7 and reference standard 6a were readily separable on

Synthesis of Reference Standard and Precursor
AZ683 reference standard 6a and N-desmethyl precursor 7 were synthesized via modified literature procedures in five and six steps, respectively (Scheme 1) [23]. Both syntheses diverged from a common intermediate 4. This common intermediate was synthesized via condensation of 4-bromo-3-ethoxyaniline (1) with diethylethoxymethylenemalonate to yield 2. This was followed by cyclization/chlorination with POCl 3 and tetrabutylammonium chloride to form chloroquinoline 3. A subsequent S N Ar reaction with 2,4-difluoroaniline yielded intermediate 4. Buchwald-Hartwig cross-coupling was then used to couple either N-Boc piperazine or N-methylpiperazine with 4, yielding intermediates 5a and 5b for the reference standard and precursor, respectively. Amidation of the ethyl ester of 5 was performed using formamide/NaOEt to generate reference standard 6a and N-Boc protected precursor 6b. Final deprotection of the Boc group of 6b with trimethylsilyl chloride in methanol furnished precursor 7. Precursor 7 and reference standard 6a were readily separable on analytical and semipreparative Phenomenex Luna C18 columns using a 30% ethanolic eluent buffered with sodium phosphate at a pH of 6.6 (see Materials and Methods for details).

Preclinical PET Imaging
Initial evaluation of the imaging properties of [ 11 C]AZ683 was undertaken in female Sprague-Dawley rat. [ 11 C]AZ683 was administered via intravenous tail vein injection and rodent brain imaging was conducted for 60 min. To our surprise, [ 11 C]AZ683 showed very little brain uptake but did show high uptake and retention in the pituitary and thyroid glands (Figure 2, left). Although both glands are known for expression of CSF1R protein (thyroid) and CSF1R RNA (thyroid and pituitary) [26], we assume the very high uptake is more likely indicative of non-specific binding associated with the lipophilic nature of the compound (Table 1). Since inter-species differences are sometimes apparent between rodents and non-human primates due to the higher metabolic rate in rodents and differing BBB efflux systems, imaging in rhesus macaque brain was also performed. The primate imaging results largely mirrored the rat data, with fairly poor brain influx during the early

Preclinical PET Imaging
Initial evaluation of the imaging properties of [ 11 C]AZ683 was undertaken in female Sprague-Dawley rat. [ 11 C]AZ683 was administered via intravenous tail vein injection and rodent brain imaging was conducted for 60 min. To our surprise, [ 11 C]AZ683 showed very little brain uptake but did show high uptake and retention in the pituitary and thyroid glands (Figure 2, left). Although both glands are known for expression of CSF1R protein (thyroid) and CSF1R RNA (thyroid and pituitary) [26], we assume the very high uptake is more likely indicative of non-specific binding associated with the lipophilic nature of the compound (Table 1). Since inter-species differences are sometimes apparent between rodents and non-human primates due to the higher metabolic rate in rodents and differing BBB efflux systems, imaging in rhesus macaque brain was also performed. The primate imaging results largely mirrored the rat data, with fairly poor brain influx during the early frames, followed by almost complete washout and little brain retention in a normal brain (Figure 2, right). There was some retention in the central region of the brain that was likely ventricular uptake and, as before, the pituitary gland could be observed in frame and showed a much greater degree of uptake than brain. Overall though, brain uptake in monkey was higher than in rat and there was perhaps some focal uptake in the monkey cerebellum (standardized uptake value (SUV)~0.3-0.4 at late time points). Given that the cerebellum is an area of known CSF1R expression in humans [26], and CSF1R function is thought to be conserved between vertebrates [27], this signal could correspond to CSF1R, presumably associated with microglia found in the monkey cerebellum [28]. However, this will need to be confirmed in future in vitro experiments with primate brain sections. Target receptor density of CSF1R could ostensibly be low in a non-diseased control animal and would explain poor brain retention, but again normal CSF1R levels are challenging to quantify in vivo as they are transient and expected to fluctuate with the turnover of macrophages and microglia. However, low receptor density would not limit first pass brain influx and efflux which was also quite low. Overall these PET imaging data suggest imaging CSF1R associated with neuroinflammation using [ 11 C]AZ683 may be challenging, but that uptake in monkey could be sufficient to observe accumulation in a brain inflammation model. There is literature precedent for TSPO radiotracers with low brain uptake being used to successfully image inflammation in rat models [29,30]. Moreover, the present studies do not rule out labeling the scaffold with a longer-lived PET radionuclide (e.g., 18 F or 124 I) and using a prolonged infusion protocol so that sufficient radiotracer accumulates at sites of inflammation. [ 11 C]AZ683 could also possibly be used for imaging of peripheral CSF1R to evaluate its role in inflammation outside of the brain.
Given that [ 11 C]AZ683 possesses properties mostly consistent with BBB permeability (Table 1) [23,31,32], the lack of brain uptake was unexpected and the reasons for it are unclear. It is possible that [ 11 C]AZ683 is a substrate for an efflux transporter on the BBB and since, for example, P-glycoprotein transporter (P-gp) expression is higher in rodents than monkeys (and humans) [33], this could explain the 2-3-fold higher uptake of the radiotracer observed in monkey brain. Given the differences in type and expression levels of efflux transporters between species, monkeys are better for predicting the role of P-gp in limiting brain penetration of drugs in humans [33]. However, as we take a conservative view towards primate safety, methods to determine whether efflux activity is responsible for the low brain uptake of [ 11 C]AZ683 (e.g., cyclosporin A blockade of the P-gp transporter [34]) have not been pursued at this time. Alternatively, in this case, cLogP estimates (Table 1) may not be a good indicator of BBB permeability. [ 11 C]AZ683 has multiple groups containing nitrogen and oxygen atoms which are ionizable, corresponding to multiple pKa values (Figure 3 [32]). We do not expect the primary amide to limit BBB permeability since we conduct brain imaging with other primary amide-containing radiopharmaceuticals such as [ 11 C]LY2795050 [35]. Understanding the relationship between cLogP of charged species as a function of pH is complicated [31], but it is likely that AZ683 is charged at physiological pH and this could be the reason for poor brain uptake. The oxygen and nitrogen atoms in question also participate in hydrogen bonding, and cLogP-the total number of oxygen and nitrogen atoms in a drug molecule (N + O) offers information about logBBB. If cLogP-(N + O) >0, logBBB is likely to be positive and the drug has a good probability of entering the CNS [31]. In the case of AZ863, cLogP-(N + O) = −4, suggesting the number of oxygen and nitrogen atoms may be too high for good CNS penetration. All of these issues should be considered in the design of next generation CSF1R radiopharmaceuticals going forward.

General Considerations
All the chemicals were purchased from commercially available suppliers and used without purification. Automated flash chromatography was performed with Biotage Isolera Prime system. High-performance liquid chromatography (HPLC) was performed using a Shimadzu LC-2010A HT system. 1 H NMR spectra were acquired using a Varian 400 apparatus (400 MHz) in CDCl 3 or CD 3 OD. δ are reported in ppm relative to tetramethylsilane (δ = 0), J are given in Hz. Mass spectra were measured on an Agilent Technologies (Santa Clara CA, USA) Q-TOF HPLC-MS or Micromass (Manchester, UK) VG 70-250-S Magnetic sector mass spectrometer employing the electrospray ionization (ESI) method.

General Considerations
All the chemicals (except for reference standard 6a and precursor 7 noted above) were purchased from commercially available suppliers and used without purification: sodium chloride, 0.9% USP and Sterile Water for Injection, USP were purchased from Hospira; Dehydrated Alcohol for Injection, USP was obtained from Akorn Inc. (Lake Forest IL, USA) HPLC was performed using a Shimadzu (Kyoto, Japan) LC-2010A HT system equipped with a Bioscan B-FC-1000 radiation detector, and HPLC columns were acquired from Phenomenex (Torrance CA, USA). Other synthesis components were obtained as follows: sterile filters were acquired from MilliporeSigma (Burlington MA, USA); C18 Vac 1cc Sep-Paks were purchased from Waters Corporation (Milford MA, USA); Sep-Paks were flushed with 5 mL of ethanol followed by 10 mL of sterile water prior to use.

Quality Control Testing of [ 11 C]AZ683
Visual inspection Doses were visually examined and required to be clear, colorless, and free of particulate matter.

Dose pH
The pH of the doses was determined by applying a small amount of the dose to pH-indicator strips and determined by visual comparison to the scale provided. pH needs to be between 4.5 and 7.5, and the pH of each [ 11 C]AZ683 dose synthesized in this study was 5.0.

Analytical HPLC
Analytical HPLC was performed using a Shimadzu LC-2010A HT system equipped with a Bioscan B-FC-1000 radiation detector (column: Phenomenex Luna C18, 5µ, 4.6 × 150 mm; mobile phase: 27%  Figure 5 for a typical analytical HPLC trace) and coinjection with unlabeled reference standard 6a confirmed radiochemical identity (see Figure 6 for a coinjection HPLC trace).

Animal Husbandry and Housing
Husbandry and housing for rodents and primates is provided by the University Laboratory for Animal Medicine (ULAM) at UM, and animal facilities are in compliance with the regulations defined by the US Department of Agriculture (USDA).
Monkeys: The University of Michigan PET Center has maintained 2 rhesus macaques for 15 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 do 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.
Rodents: Rats are housed in Allentown #3 micro ventilated cages (27 cm wide × 49 cm deep × 27 cm high, floor area 923 Sq cm) with animal housing densities set by ULAM and the Guide for the Care and Use of Laboratory Animals. Housing is located on ventilated racks with continuous water and air supply exchange. All animals are provided with LabDiet 5LOD as well as enrichment materials, and are on a light schedule of 12 h light/12 h dark.

Primate Imaging Protocol
Primate imaging studies were done using a mature female rhesus monkey (weight = 9.4 kg, n = 1). The animal was anesthetized in the home cage with ketamine and transported to the PET imaging suite. The monkey was intubated for mechanical ventilation, and anesthesia was continued with isoflurane. Anesthesia was 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. Following a transmission scan, the animal was injected i.v. with [ 11 C]AZ683 (145.0 MBq) as a bolus over 1 min, and the brain imaged for 60 min (5 × 2 min frames-4 × 5 min frames-3 × 10 min frames).

PET Image Analysis
Emission data were corrected for attenuation and scatter, and reconstructed using the 3D maximum a priori (3D MAP) method. By using a summed image, regions of interest (ROI) were drawn on multiple planes, and the volumetric ROIs were then applied to the full dynamic data set to generate time-radioactivity curves.

Conclusions
In conclusion, we have developed a radiosynthesis of [ 11 C]AZ683 for PET imaging of CSF1R and neuroinflammation that provides doses suitable for preclinical use. However, preliminary preclinical PET imaging in rodents and nonhuman primates revealed low brain uptake of [ 11 C]AZ683. Overall, these PET imaging data suggest imaging CSF1R associated with neuroinflammation using [ 11 C]AZ683 could be challenging and emphasize that high affinity, good selectivity, and appropriate drug-like properties do not guarantee that a compound will make a good radiopharmaceutical for in vivo brain PET. Nevertheless, uptake in monkey could be sufficient to observe accumulation in a brain inflammation model. These studies also do not rule out labeling the scaffold with a longer-lived PET radionuclide (e.g., 18 F or 124 I) and using a prolonged infusion protocol to ensure that sufficient radiotracer accumulates at sites of inflammation for imaging and quantitation of CSF1R. [ 11 C]AZ683 could potentially also be used for imaging of peripheral CSF1R to evaluate its role in inflammation outside the brain. Future evaluation in animal models of inflammation appears warranted.