神经炎症是中枢神经系统面对伤害刺激的适应性反应，与多种神经退行性 疾病紧密相关。小胶质细胞是中枢神经系统中的免疫细胞，当中枢神经系统出 现炎症时，小胶质细胞会增殖、分化，大脑会招募小胶质细胞移动到受伤部位 进行免疫反应来控制炎症。集落刺激因子 1 受体(colony stimulation factor 1 receptor, CSF1R)在脑中特异性表达于小胶质细胞，是小胶质细胞成像的理想靶 点。正电子计算机断层扫描成像(positron emission tomography, PET)作为高度定 量的分子影像技术，具有成像过程组织穿透性强、能在分子水平实现体内生物 过程表征等优点，其在辅助药物开发和临床疾病诊断等领域具有重要应用。因 此，研制靶向于CSF1R的 PET显像剂能够实现小胶质细胞的特异性成像，探 索小胶质细胞活化导致的神经炎症在神经退行性疾病中的核心作用，同时辅助 CSF1R 中枢神经药物的开发，为疾病的治疗提供新的途径。目前，[11C]CPPC 是研究最深入的 CSF1R PET 显像剂，且已经进入临床探索阶段。该探针及其 衍生物都有较高的结合力和良好的脑摄取，但其激酶选择性差。因此，基于对 CPPC 分子的 N-苯基酰胺结构的修饰，我们开展了靶向脑内 CSF1R PET探针 的研发工作。本论文主要包括以下两方面的研究： 一、6-氟吡啶甲酰苯胺类CSF1R PET探针的研究 本章选取 CPPC 分子骨架的 N-苯基酰胺结构，将呋喃环改为易于氟标记 的 6-氟吡啶结构，设计合成了一系列探针分子 2.4a–f，并筛选出对 CSF1R 有 较高亲和性和高选择性的探针分子2.4b (IC50 = 21.3 ± 7.7 nM)，其激酶选择性 优于 CPPC。随后，以三甲基季铵盐作为标记前体，通过与 18F 离子发生芳基 取代反应制备了探针[18F]2.4b，放射性标记产率为 28.03% (n = 4)，比活度为 58 - 87.1 GBq/μmol。此外，[18F]2.4b 具有合理的脂溶性(logD7.4 = 2.89)，且在 生盐溶液和小鼠血清中具有较高的体外稳定性。小鼠生物分布实验表明， [18F]2.4b 具有较高的初始脑摄取(brain10min = 3.30% ID/g)，但清除速率较慢 (brain60min = 2.71% ID/g)。该探针在血液中的摄取较低(< 1% ID/g)，有利于获得 较高的血脑比；且骨摄取较低(< 2.7% ID/g)，说明没有明显的体内脱氟现象。 此外，体内代谢实验表明[18F]2.4b 在小鼠脑部稳定，注射探针 30 分钟后母体 化合物含量大于 95%。在大鼠 PET 成像中，[18F]2.4b 表现出较高的脑摄取 (SUVmax = ~1.35)和较慢的脑清除，且没有明显的骨摄取，与小鼠动物分布结果 一致。猴子 PET 成像表明，[18F]2.4b 在非灵长类脑部同样具有较高的摄取，在丘脑中的最高值为 ~1.97 SUV；相较于大鼠，该探针在猴子脑部有一定的脑 清除。最后，[18F]2.4b 在脂多糖(lipopolysaccharides，LPS)诱导的炎症模型小 鼠中的脑摄取明显高于对照组小鼠，在探针注射 30 min 后，脑摄取分别为 0.98% ID/g 和 1.79% ID/g，升高了 92.3%；且该信号能够被 CSF1R 抑制剂 BLZ945 和 CPPC 有效抑制，摄取值分别减少 20.9% (p = 0.047)和 21.4% (p = 0.064)。该实验表明，探针[18F]2.4b 能够有效地、特异地探测炎症小鼠脑部 CSF1R表达水平的变化，是非常有潜力的用于神经炎症PET成像的显像剂。 二、呋喃和吡咯甲酰苯胺类CSF1R PET探针的研究 通过第一部分工作，我们发现吡啶环的引入降低了抑制剂对 CSF1R 的亲 和力。此外，根据报道，呋喃和吡咯结构有利于增加 CPPC 类抑制剂的亲和 力。因此，我们选择了呋喃和吡咯甲酰苯胺作为母体结构，通过在其 2、4 位 引入含有氟原子的基团，比如4-氟哌啶、3-氟吡咯烷和3-氟吖啶等，设计合成 一系列抑制剂分子3.4a - o。活性测定表明，呋喃甲酰苯胺类化合物3.4d (IC50 = 2.1 ± 0.7 nM)和 3.4n (IC50 = 1.3 ± 0.1 nM)对 CSF1R的亲和力与CPPC (IC50 = 2.2 nM) 相当，且明显高于化合物2.4b (IC50 = 21.3 ± 7.7 nM)。然而，对3.4d的 放射性18F标记未能获得成功，可能的原因是标记过程中哌啶环 4-位离去基团 (对甲基苯磺酰基和甲基磺酰基)未被 18F 离子取代，而发生了消去反应得到了 四氢吡啶基团。在含有 3-氟吖啶基团的 3.4n 的标记中，我们以对硝基苯磺酰 基为离去基团，通过与 18F 离子发生亲核取代反应制备了探针[18F]3.4n，放射 性标记产率为5.5% (n = 3)，比活度为30.1 - 35.2 GBq/μmol。此外，[18F]3.4n 的 logD7.4 值为 3.87，在生盐溶液中具有较高的体外稳定性。小鼠生物分布实 验表明，与[18F]2.4b 相比，[18F]3.4n 具有更高的初始脑摄取(brain2min = 4.31% ID/g)和快速的脑清除速率(brain60min = 0.85% ID/g)，且骨摄取更低(< 2.30% ID/g)。最后，[18F]3.4n 在 LPS 炎症模型小鼠中的脑摄取明显高于对照组小 鼠，在注射探针30 min后，脑摄取分别为0.63% ID/g和 1.25% ID/g，升高了 99.7%；且该信号能够被 CSF1R抑制剂 BLZ945和自抑制有效抑制，摄取值分 别减少26.0% (p = 0.015)和15.7% (p = 0.063)。该实验表明，探针[18F]3.4n能够 有效地、特异地探测炎症小鼠脑部 CSF1R 表达水平的变化，在体外亲和力、 脑摄取和脑清除速率等性质中优于[18F]2.4b，具有较大的临床应用潜力。

Neuroinflammation is inflammation of the central nervous system (CNS). It may be initiated in response to a variety of cues, including brain injury, infection or neurodegenerative diseases. Microglia are the resident immune cells in the CNS that are activated in response to these cues. Colony stimulation factor 1 receptor (CSF1R) is specifically expressed in microglia in the brain and is an ideal target for microglia imaging. As a highly quantitative imaging technique, positron emission tomography (PET) has the advantages of tissue penetration and the ability to characterize in vivo bioprocesses at the molecular level, which is used in the fields of assisting drug discovery and clinical diagnosis. Therefore, the development of PET imaging probes targeting CSF1R can achieve microglia-specific imaging, explore the central role of neuroinflammation due to microglia activation in neurodegenerative diseases, and aid the development of central nervous system drugs for CSF1R, which can provide a new pathway for the clinical therapy. Recently, [11C]CPPC is the most developed CSF1R PET probe and is already in the clinical exploration phase. Both [11C]CPPC and its derivatives have high binding affinity and good brain uptake, but it is of poor kinase selectivity. Therefore, based on the modification of the formanilide scaffold of CPPC, we develop a series of CSF1R PET probes. This thesis consists of the following two main studies: I. Study of 6-fluoropyridine carboxanilide-based CSF1R PET probes The goal of this chapter was to design and synthesize compounds 2.4a-f by introducing the 6-fluoropyridine moiety that is easy for fluorine-labeled. Then, compound 2.4b (IC50 = 21.3 ± 7.7 nM) was screened, which showed a better kinase selectivity than that of CPPC. Subsequently, [18F]2.4b was obtained by aryl substitution reaction with 18F- using trimethyl quaternary ammonium salt as the precursor, with a radiochemical yield of 28.03% (n = 4) and the molar activity of 58 - 87.1 GBq/μmol. In addition, [18F]2.4b has a reasonable lipid solubility (logD7.4 = 2.89) and has high in vitro stability in saline and mouse serum. Biodistribution experiments showed that [18F]2.4b started with a high initial brain uptake (brain10min = 3.30% ID/g) but followed by a slow clearance rate (brain60min = 2.71% ID/g). The low uptake of [18F]2.4b in blood (< 1% ID/g) favored a high blood-to-brain ratio;and the low bone uptake (< 2.7% ID/g) indicated that there was no significant in vivo defluoridation. In addition, in vivo metabolism experiments demonstrated that [18F]2.4b was stable in mouse brain with more than 95% of the parent compound 30 min after probe injection. In SD rat PET imaging, [18F]2.4b exhibited a high brain uptake (SUVmax = ~1.35) and slow brain clearance without significant bone uptake, consistent with the results of biodistribution. PET/MRI imaging in monkeys showed that [18F]2.4b had similarly high uptake in the non-primate brain, with a maximum value of ~1.97 SUV in the thalamus; compared to SD rats, [18F]2.4b showed brain clearance in the monkey brain. Finally, lipopolysaccharides (LPS)-induced inflammation model mice significantly increased the brain uptake of [18F]2.4b, compared to control mice, with the brain uptake increasing by 92.3% from to 0.98% ID/g to 1.79% ID/g. 30 min after probe injection, the brain uptake was able to be inhibited by CSF1R inhibitors BLZ945 and CPPC effectively, which was reduced by 20.9% (p = 0.047) and 21.4% (p = 0.064), respectively. This experiment demonstrated that [18F]2.4b can effectively and specifically detect the change in CSF1R expression levels in the brains of inflammation mice and is a very promising probe for PET imaging of neuroinflammation. II. Study of furanocarbonylanilide-based and pyrrolobenzoylaniline-based CSF1R PET probes Through the first part of our work, we found that the introduction of the pyridine ring decreased the affinity of CSF1R inhibitors. In addition, according to reports, the furan and pyrrole moieties were favorable to increase the affinity of CSF1R inhibitors. Therefore, we chose furanocarbonylanilide and pyrrolobenzoylaniline as scaffold and synthesized a series of compounds 3.4a-o by introducing 4-fluoropiperidine, 3-fluoropyrrolidine, and 3-fluoroacridine at the 2 position and 4-position of formanilide. Compound 3.4d (IC50 = 2.1 ± 0.7 nM) and 3.4n (IC50 = 1.3 ± 0.1 nM) showed comparable affinity for CSF1R as CPPC (IC50 = 2.2 nM) and much higher than compound 2.4b (IC50 = 21.3 ± 7.7 nM). However, the radiosynthesis of [18F]3.4d was unsuccessful, possibly due to the fact that the leaving groups (tosyl and methanesulfonyl) were not replaced by 18F- during labeling. Instead, the elimination reaction took place to obtain the tetrahydropyridine moiety. [18F]3.4n was prepared by nucleophilic substitution reaction with 18F- using p nitrobenzenesulfonyl as the leaving group, with a radiochemical yield of 5.5% (n = 3) and the molar activity of 30.1 - 35.2 GBq/μmol. Furthermore, the logD7.4 value for [18F]3.4n is 3.87 and [18F]3.4n had a high in vitro stability in saline. Biodistribution experiments demonstrated that [18F]3.4n had a higher initial brain uptake (brain2min = 4.31% ID/g) and more rapid brain clearance rate (brain60min = 0.85% ID/g) and lower bone uptake (< 2.30% ID/g) compared to [18F]2.4b. Finally, lipopolysaccharides (LPS)-induced inflammation model mice significantly increased the brain uptake of [18F]3.4n, compared to control mice, with the brain uptake increasing by 99.7% from to 0.63% ID/g to 1.25% ID/g. 30 min after probe injection, the brain uptake was able to be inhibited by CSF1R inhibitors BLZ945 and self-blockade effectively, which was reduced by 26.0% (p = 0.015) and 15.7% (p = 0.062), respectively. This experiment demonstrated that [18F]3.4n can effectively and specifically detect the change in CSF1R expression levels in the brains of inflammation mice, and its affinity, brain uptake and brain clearance rate superior to those of [18F]2.4b, which has great potential for clinical imaging.