Neuroimaging of indoleamine 2,3-dioxygenase 1 (Ido1) in Alzheimer's disease: Development of a novel 18F-labeled PET tracer Authors: Hua Cheng, Nilay S Narayan, Yongle Wang, Yanli Wang, Yafei Huang, Hongzhi Yang, Ramya Tokala, Jinwu Yan, Joseph William Downey, Anne Siewko, Breanna Lizeth Varela, Madelyn Rose Hallisey, Tiffany A Geronimo, Mary Margaret Banick, Rudolph E Tanzi, Can Zhang, Yulong Xu, Changning Wang Published: 2026 ABSTRACT AbstractIDO1-mediated dysregulation of the kynurenine pathway is a pivotal driver of neuroinflammation in Alzheimer's disease (AD). Despite its clinical significance, PET probes specifically targeting IDO1 in the AD brain remain scarce. In this study, we first established a strong correlation between IDO1 expression and AD progression through bioinformatic analysis. Subsequently, a novel PET tracer, [18F]XW-032, was developed via ruthenium-mediated deoxyfluorination. [18F]XW-032 demonstrated superior blood-brain barrier (BBB) permeability and a favorable metabolic profile. In vitro autoradiography and in vivo dynamic PET/CT imaging in 5XFAD mice confirmed its high binding specificity and robust imaging performance, further validated by competitive blocking studies with unlabeled IDO1 inhibitors. Notably, [18F]XW-032 effectively visualized regional IDO1 heterogeneity in the AD brain. These findings position [18F]XW-032 as a potent molecular tool for early AD diagnosis and for monitoring the efficacy of IDO1-targeted therapeutic interventions.Keywords: Molecule imaging, Alzheimer's disease, IDO1, Neuroimaging, Neurodegeneration Introduction Neuroinflammation plays a critical role in the onset and progression of central nervous system (CNS) diseases [[1], [2], [3]], including AD [4,5] amyotrophic lateral sclerosis (ALS) [6,7], Parkinson's disease (PD) [8,9], cancer with brain metastases [10] etc. As a result of neuroinflammation, 95% of the essential amino acid tryptophan (Trp) is catabolized through the kynurenine pathway (KP), leading to the production of multiple neuroactive metabolites [11]. KP plays an indispensable role in the metabolism involved in several brain diseases. Its downstream metabolites, particularly quinolinic acid (QUIN), an NMDA receptor agonist, are of special interest associated with neurological diseases [12]. Several enzymes can catalyze the degradation of Trp, including indoleamine 2,3-dioxygenase 1 (IDO1), indoleamine 2,3-dioxygenase 2 (IDO2), and tryptophan 2,3-dioxygenase (TDO) [13]. Among these, IDO2 exhibits a markedly lower capacity for Trp metabolism compared to IDO1, while TDO is primarily expressed in the liver rather than in neural tissues [14,15]. IDO1 is among the most extensively studied enzymes in neurological diseases, owing to its role in catalyzing the rate-limiting step of the kynurenine pathway [16]. Due to the importance of IDO1 plays in the KP, it has been identified as one of the potential contributors to the development of several CNS diseases, especially AD [17,18]. Notably, IDO1 exhibits pronounced region-specific elevation in areas most susceptible to AD-related neuroinflammation, such as the hippocampus and prefrontal cortex. [19,20]. This makes IDO1 not only a driver of pathology but also a promising biomarker for monitoring inflammatory status in AD.In recent years, numerous studies have revealed a strong association between IDO1 and AD, as AD is characterized by pronounced neuroinflammatory responses, IDO1 inhibitors could help control the inflammatory processes in AD [11]. Besides that, IDO1 is a promising target in AD imaging. PET tracers targeting IDO1 is a powerful tool for IDO1 visualization in a non-invasive way [21]. IDO1 PET tracers have emerged as a promising immunotherapeutic target in oncology, and several radiotracers have been developed for cancer imaging [22]. As shown in Fig. 1, despite the clinical advancement of IDO1-targeted immunotherapies in oncology, including Indoximod [23], Epacadostat [24], and BMS-986205 (Phase 3) [25], as well as Navoximod, LY3381916 [26], and PF-0680003 (Phase 1), their application in CNS disorders is severely hindered by the paucity of imaging tools capable of crossing the blood-brain barrier (BBB). Non-invasive visualization of IDO1 through PET imaging is essential for early diagnosis and therapeutic longitudinal monitoring; however, current tracers are primarily optimized for peripheral tumors and often exhibit suboptimal brain kinetics. Representative Inhibitors of IDO1 entering clinical research. To find a PET/CT tracer for neurological diseases, the CNS drugability must be considered [27]. To address this unmet need, we identified XW-032—a potent IDO1 inhibitor with favorable drug-like properties and predicted BBB permeability—as a scaffold for tracer development [28]. In this study, we report the synthesis and characterization of a novel Fluorine-18 labeled PET radiotracer, [18F]XW-032, for IDO1 neuroimaging. Our findings demonstrate high binding affinity, excellent selectivity, and, for the first time, the feasibility of visualizing IDO1 fluctuations in AD mouse models, thereby providing a robust molecular tool for investigating the neuroinflammatory landscape of Alzheimer's disease. Methods In this research, all chemical reagents and solvents were obtained with a purity of ACS grade or higher, thus no further purification was necessary. Ethanol (EtOH), acetonitrile, and anhydrous dimethyl sulfoxide (DMSO) were acquired from chemical suppliers like Millipore Sigma. The 18F isotope was generated using a Siemens Eclipse HP cyclotron. All animal experiments in this study were carried out at Massachusetts General Hospital (MGH) with Public Health Service (PHS) Assurance of Compliance under Number A3596-01. The animal research protocols detailed were examined and authorized by MGH's Institutional Animal Care and Use Committee (IACUC), referred to as the Subcommittee on Research Animal Care (SRAC). PET/CT imaging was performed on mice anesthetized with isoflurane to ensure comfort and minimize distress. Throughout all procedures, a group of experienced animal care technicians carefully supervised the animals' safety. Veterinary staff members were assigned to oversee the mice's daily health and well-being. The mice were kept in appropriate cages that supported both their physical and behavioral needs, and they were given continuous access to food and water. Furthermore, any required dietary supplements were administered according to the guidance of the veterinary team. Public dataset re-analysis Transcriptomic data was reanalyzed from publicly available dataset (GSE33000) in this study. Gene expression profiles from Alzheimer's disease (AD) and control brain samples were obtained from previously published cohorts (GSE33000). Raw or normalized expression matrices were downloaded and processed using standard pipelines in R studio. Count data were normalized using DESeq2. Inflammation-related genes (including IFNG, IL1B, TNF, and STAT1) were selected based on biological relevance. For correlation analyses, pairwise relationships between IDO1 expression and indicated genes were evaluated using Pearson correlation coefficients. Differences between two groups were assessed using the Wilcoxon rank-sum test. Linear regression lines were fitted for visualization. Violin plots and scatter plots were generated using ggplot2. Chemistry Step 1 (Compound 3):To a solution of acetyl chloride (6.96 g, 88.60 mmol, 6.30 mL, 3.08 equiv) in MeOH (105 mL) at 0 °C was stirred for 30 min, followed by the addition of 4-(4-bromophenyl)butanoic acid (7.0 g, 28.80 mmol, 1.0 equiv). The reaction mixture was then stirred at 25 °C for 2 h. Upon completion, the reaction was quenched by adjusting the pH to 7 with saturated aqueous K2CO3 and diluted with H2O (100 mL). The mixture was extracted with DCM (2 × 100 mL). The combined organic layers were dried over Na2SO4, filtered, and concentrated under reduced pressure to afford the crude product. Methyl 4-(4-bromophenyl)butanoate (7.0 g, 27.22 mmol, 94.5% yield) was obtained as a yellow oil and used in the next step without further purification.1H NMR (400 MHz,CDCl3) δ ppm 7.41 (d, J = 8.0 Hz, 2H), 7.06 (d, J = 8.0 Hz, 2H), 3.67 (s, 3H), 2.61 (t, J = 7.2 Hz, 2H), 2.32 (t, J = 7.6 Hz, 2H), 1.90-1.97 (m, 2H). LC-MS [M + H]+: 257.1 Step 1 (Compound 3):To a solution of acetyl chloride (6.96 g, 88.60 mmol, 6.30 mL, 3.08 equiv) in MeOH (105 mL) at 0 °C was stirred for 30 min, followed by the addition of 4-(4-bromophenyl)butanoic acid (7.0 g, 28.80 mmol, 1.0 equiv). The reaction mixture was then stirred at 25 °C for 2 h. Upon completion, the reaction was quenched by adjusting the pH to 7 with saturated aqueous K2CO3 and diluted with H2O (100 mL). The mixture was extracted with DCM (2 × 100 mL). The combined organic layers were dried over Na2SO4, filtered, and concentrated under reduced pressure to afford the crude product. Methyl 4-(4-bromophenyl)butanoate (7.0 g, 27.22 mmol, 94.5% yield) was obtained as a yellow oil and used in the next step without further purification.1H NMR (400 MHz,CDCl3) δ ppm 7.41 (d, J = 8.0 Hz, 2H), 7.06 (d, J = 8.0 Hz, 2H), 3.67 (s, 3H), 2.61 (t, J = 7.2 Hz, 2H), 2.32 (t, J = 7.6 Hz, 2H), 1.90-1.97 (m, 2H). LC-MS [M + H]+: 257.1 Step 2 (Compound 5):A mixture of methyl 4-(4-bromophenyl)butanoate (7.0 g, 27.22 mmol, 1.0 equiv), bis(pinacolato)diboron (7.60 g, 29.95 mmol, 1.1 equiv), potassium acetate (8.02 g, 81.67 mmol, 3.0 equiv), and Pd(dppf)Cl2 (666.97 mg, 0.817 mmol, 0.03 equiv) in DMSO (70 mL) was degassed and purged with N2 (three cycles). The reaction mixture was then stirred at 85 °C for 4 h under a nitrogen atmosphere. After completion, the reaction mixture was diluted with H2O (200 mL) and extracted with EtOAc (3 × 100 mL). The combined organic layers were dried over Na2SO4, filtered, and concentrated under reduced pressure to afford a crude residue. The residue was purified by column chromatography on silica gel (hexanes/ethyl acetate = 100:1 to 3:1) to afford methyl 4-[4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)phenyl]butanoate (6.0 g, 19.72 mmol, 72.5% yield) as a yellow oil.1H NMR (400 MHz, CDCl3) δ ppm 7.74 (d, J = 8.0 Hz 2H), 7.19 (d, J = 7.6 Hz, 2H), 3.67 (s, 3H), 2.67 (t, J = 7.6 Hz, 2H), 2.32 (t, J = 7.2 Hz, 2H), 1.90-2.10 (m, 2H), 1.34 (s, 12H). LC-MS [M + H]+: 305.6 Step 2 (Compound 5):A mixture of methyl 4-(4-bromophenyl)butanoate (7.0 g, 27.22 mmol, 1.0 equiv), bis(pinacolato)diboron (7.60 g, 29.95 mmol, 1.1 equiv), potassium acetate (8.02 g, 81.67 mmol, 3.0 equiv), and Pd(dppf)Cl2 (666.97 mg, 0.817 mmol, 0.03 equiv) in DMSO (70 mL) was degassed and purged with N2 (three cycles). The reaction mixture was then stirred at 85 °C for 4 h under a nitrogen atmosphere. After completion, the reaction mixture was diluted with H2O (200 mL) and extracted with EtOAc (3 × 100 mL). The combined organic layers were dried over Na2SO4, filtered, and concentrated under reduced pressure to afford a crude residue. The residue was purified by column chromatography on silica gel (hexanes/ethyl acetate = 100:1 to 3:1) to afford methyl 4-[4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)phenyl]butanoate (6.0 g, 19.72 mmol, 72.5% yield) as a yellow oil.1H NMR (400 MHz, CDCl3) δ ppm 7.74 (d, J = 8.0 Hz 2H), 7.19 (d, J = 7.6 Hz, 2H), 3.67 (s, 3H), 2.67 (t, J = 7.6 Hz, 2H), 2.32 (t, J = 7.2 Hz, 2H), 1.90-2.10 (m, 2H), 1.34 (s, 12H). LC-MS [M + H]+: 305.6 Step 3 (Compound 7):A mixture of methyl 4-[4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)phenyl]butanoate (6.0 g, 19.72 mmol, 1.0 equiv), 2-bromo-5-fluoropyridine (4.17 g, 23.67 mmol, 1.2 equiv), Pd(dppf)Cl2 (1.44 g, 1.97 mmol, 0.10 equiv), and Cs2CO3 (19.28 g, 59.17 mmol, 3.0 equiv) in dioxane (80 mL) and H2O (20 mL) was degassed and purged with N2 (three cycles). The reaction mixture was then stirred at 100 °C for 8 h under a nitrogen atmosphere. After completion, the reaction mixture was concentrated under reduced pressure to afford a crude residue. The residue was purified by column chromatography on silica gel (hexanes/ethyl acetate = 100:1 to 3:1) to afford methyl 4-[4-(5-fluoropyridin-2-yl)phenyl]butanoate (1.0 g, 3.29 mmol, 16.7% yield) as a white solid. 1H NMR (400 MHz, CDCl3) δ ppm 8.52 (d, J = 2.8 Hz, 1H), 7.85 (d, J = 8.4 Hz, 2H), 7.65-7.75 (m, 1H), 7.45-7.50 (m, 1H), 7.25-7.35 (m, 2H), 3.67 (s, 3H), 2.70 (t, J = 7.6 Hz, 2H), 2.35 (t, J = 7.6 Hz, 2H), 1.95-2.05 (m, 2H). LC-MS [M + H]+:274.3 Step 3 (Compound 7):A mixture of methyl 4-[4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)phenyl]butanoate (6.0 g, 19.72 mmol, 1.0 equiv), 2-bromo-5-fluoropyridine (4.17 g, 23.67 mmol, 1.2 equiv), Pd(dppf)Cl2 (1.44 g, 1.97 mmol, 0.10 equiv), and Cs2CO3 (19.28 g, 59.17 mmol, 3.0 equiv) in dioxane (80 mL) and H2O (20 mL) was degassed and purged with N2 (three cycles). The reaction mixture was then stirred at 100 °C for 8 h under a nitrogen atmosphere. After completion, the reaction mixture was concentrated under reduced pressure to afford a crude residue. The residue was purified by column chromatography on silica gel (hexanes/ethyl acetate = 100:1 to 3:1) to afford methyl 4-[4-(5-fluoropyridin-2-yl)phenyl]butanoate (1.0 g, 3.29 mmol, 16.7% yield) as a white solid. 1H NMR (400 MHz, CDCl3) δ ppm 8.52 (d, J = 2.8 Hz, 1H), 7.85 (d, J = 8.4 Hz, 2H), 7.65-7.75 (m, 1H), 7.45-7.50 (m, 1H), 7.25-7.35 (m, 2H), 3.67 (s, 3H), 2.70 (t, J = 7.6 Hz, 2H), 2.35 (t, J = 7.6 Hz, 2H), 1.95-2.05 (m, 2H). LC-MS [M + H]+:274.3 Step 4 (Compound 8):To a solution of methyl 4-[4-(5-fluoropyridin-2-yl)phenyl]butanoate (1.0 g, 3.66 mmol, 1.0 equiv) in THF (5 mL) was added LiOH·H2O (307 mg, 7.32 mmol, 2.0 equiv) and H2O (5 mL). The reaction mixture was stirred at 25 °C for 2 h.Upon completion, the mixture was diluted with H2O (20 mL) and washed with EtOAc (2 × 10 mL). The aqueous phase was acidified to pH 7 with 1 M HCl, then extracted with EtOAc (3 × 20 mL). The combined organic layers were dried over Na2SO4, filtered, and concentrated under reduced pressure to afford 4-[4-(5-fluoropyridin-2-yl)phenyl]butanoic acid (600 mg, 2.26 mmol, 61.7% yield, 97.5% purity) as a white solid. 1H NMR (400 MHz, DMSO-d6) δ ppm 12.06 (s, 1H), 8.63 (d, J = 2.8 Hz, 1H), 7.88-8.12 (m, 3H), 7.75-7.85 (m, 1H), 7.30 (d, J = 8.4 Hz, 2H), 2.64 (t, J = 7.6 Hz, 2H), 2.24 (t, J = 7.6 Hz, 2H), 1.78-1.87 (m, 2H). 13C NMR (126 MHz, DMSO-D6) δ 174.85, 153.26, 142.60, 137.99, 135.46, 129.38, 127.06, 124.11, 121.33, 34.64, 33.59, 26.19. LC-MS [M + H]+: 260.1 Step 4 (Compound 8):To a solution of methyl 4-[4-(5-fluoropyridin-2-yl)phenyl]butanoate (1.0 g, 3.66 mmol, 1.0 equiv) in THF (5 mL) was added LiOH·H2O (307 mg, 7.32 mmol, 2.0 equiv) and H2O (5 mL). The reaction mixture was stirred at 25 °C for 2 h.Upon completion, the mixture was diluted with H2O (20 mL) and washed with EtOAc (2 × 10 mL). The aqueous phase was acidified to pH 7 with 1 M HCl, then extracted with EtOAc (3 × 20 mL). The combined organic layers were dried over Na2SO4, filtered, and concentrated under reduced pressure to afford 4-[4-(5-fluoropyridin-2-yl)phenyl]butanoic acid (600 mg, 2.26 mmol, 61.7% yield, 97.5% purity) as a white solid. 1H NMR (400 MHz, DMSO-d6) δ ppm 12.06 (s, 1H), 8.63 (d, J = 2.8 Hz, 1H), 7.88-8.12 (m, 3H), 7.75-7.85 (m, 1H), 7.30 (d, J = 8.4 Hz, 2H), 2.64 (t, J = 7.6 Hz, 2H), 2.24 (t, J = 7.6 Hz, 2H), 1.78-1.87 (m, 2H). 13C NMR (126 MHz, DMSO-D6) δ 174.85, 153.26, 142.60, 137.99, 135.46, 129.38, 127.06, 124.11, 121.33, 34.64, 33.59, 26.19. LC-MS [M + H]+: 260.1 Standard Compound (XW-032): Compound 8 (25 mg) was dissolved in DMF (1.5 mL), then 4-fluoroaniline (11.4 μL), HATU (58.4 mg), and DIPEA (34.84 μL) were combined and then stirred for 1 h at room temperature. The reaction was quenched with water, then washed 3 times with Ethyl Acetate (EA). The collected organic layers were then washed with brine once and then dried over Na2SO4. The reaction mixture was then filtered using vacuum filtration and then isolated using a rotary evaporator. The isolated compound was then purified using reverse phase column chromatography (EA:Hexane = 1:1). After purification, the compound was isolated again using a rotary evaporator. 1H NMR (500 MHz, Chloroform-d) δ 8.51 (d, J = 4.3 Hz, 1H), 7.83 (s, 2H), 7.72 – 7.64 (m, 1H), 7.44 (q, J = 4.0 Hz, 3H), 7.27 (t, J = 6.8 Hz, 3H), 6.98 (d, J = 7.2 Hz, 2H), 4.11 (t, J = 6.9 Hz, 1H), 2.74 (d, J = 6.8 Hz, 2H), 2.29 (d, J = 6.9 Hz, 2H), 2.07 (d, J = 7.2 Hz, 2H). 13C NMR (126 MHz, DMSO-D6) δ 197.99, 195.29, 170.77, 158.87, 156.65, 154.50, 145.70, 145.14, 137.78, 135.99, 129.39, 126.65, 124.57, 122.52, 120.06, 114.27, 36.15, 33.01, 25.23. LC-MS [M + H]+:353.1. Precursor: Compound 8 (100 mg) was dissolved in DMF (6 mL), then 4-aminophenol (44.7 μL), HATU (220 mg), and DIPEA (134 μL) were combined and then stirred for 1 h at room temperature. The reaction was quenched with water, then washed 3 times with Ethyl Acetate (EA). The collected organic layers were then washed with brine once and then dried over Na2SO4. The reaction mixture was then filtered using vacuum filtration and then isolated using a rotary evaporator. The isolated compound was then purified using reverse phase column chromatography (EA:Hexane = 1:1). After purification, the compound was isolated again using a rotary evaporator. 1H NMR (500 MHz, Chloroform-d) δ 8.52 (d, J = 2.9 Hz, 1H), 8.00 (s, 1H), 7.85 (d, J = 8.2 Hz, 2H), 7.69 (dd, J = 8.8, 4.3 Hz, 1H), 7.45 (d, J = 2.8 Hz, 1H), 7.33 – 7.28 (m, 4H), 6.97 (s, 1H), 6.76 (d, J = 8.8 Hz, 2H), 5.08 (s, 1H), 2.75 (d, J = 7.4 Hz, 2H), 2.32 (t, J = 7.4 Hz, 2H), 2.09 (q, J = 7.4 Hz, 2H). 13C NMR (126 MHz, DMSO-D6) δ 171.09, 153.39, 141.81, 137.20, 134.27, 131.81, 130.22, 127.76, 124.74, 122.13, 121.41, 117.05, 36.10, 33.80, 28.72. LC-MS [M + H]+: 351.2. Molecular modeling The protein crystal structure of IDO1 (PDB number, 6AZW) was obtained from the Protein Data Bank [29]. Molecular docking studies were performed using the Schrödinger suite. Protein structures were prepared using the Protein Preparation Wizard, including the addition of hydrogens, assignment of bond orders, removal of redundant chains, and energy minimization. The receptor grid was generated based on the defined binding site. Ligands were prepared using LigPrep and minimized to obtain low-energy conformations. Docking was carried out using Glide, and the resulting poses were exported in PDB format. Docking results were analyzed and visualized using PyMOL. Radiosynthesis of [18F]XW-032 Precursor (5 mg) and CpRu(COD)Cl (12.76 mg, 41.46 μmol) were dissolved in EtOH (50 μL) in a 0.5 ml vial and heated at 85 °C for 30 min. 2-chloro-1,3-bis(2,6-diisopropylphenyl)-1H-imidazolium chloride (ClIm) (19.05 mg, 41.46 μmol) and MeCN (150 μL) were then added, and the mixture was used directly for [18F]fluoride elution. Cyclotron-produced [18F]fluoride was trapped on a preconditioned PS-HCO3 QMA cartridge, rinsed with MeCN (1.0 mL), and eluted with the reaction solution into a borosilicate vial. The cartridge was further washed with DMSO (150 μL) and DMSO/MeCN (1:1, 50 μL), and the combined eluates were collected. The sealed vial was heated at 125 °C for 30 min. After cooling, the reaction was quenched with 0.1% TFA in H2O and purified by semi-preparative HPLC (C18, 5 μm, 9.4 × 250 mm; 0.1% TFA in H2O/MeCN = 46:54; 5.0 mL/min). The HPLC chromatogram [18F]XW-032 and co-injection HPLC chromatogram of [18F]XW-032 and XW-032 were display in support information (Fig. S1 and S2). PET/CT acquisition The Rodent PET/CT acquisition has been described in published paper [30]. Male C57BL/6 mice (n = 4∼8, 6 months old) were placed in a Triumph PET/CT scanner (Gamma Medica, Northridge, CA) and anesthetized with 2% isoflurane (Patterson Vet Supply, Inc., Greeley, CO, USA) via inhalation. At the beginning of the PET scanning process, each mouse received an intravenous injection of 0.2 mL saline solution containing 5% [18F]XW-032 via a catheter placed in the lateral tail vein. To evaluate the blocking effect, a separate group of mice was pre-administered with unlabeled XW-032 at a dose of 0.3 mg/kg prior to the radiotracer injection. Each PET scan lasted for 60 min and was followed by a CT scan. The collected PET data were reconstructed using a three-dimensional maximum likelihood expectation maximization (3D-MLEM) algorithm, achieving a spatial resolution of 1 mm full width at half maximum (FWHM). PET/CT image analysis The PET and CT images that underwent reconstruction were processed with PMOD software 4.0 (PMOD Technologies Ltd. Zurich, Switzerland). Manual generation of anatomical volumes of interest (VOIs) was performed within spheres, guided by high-resolution CT structural images. Time-activity curves (TACs) were exported as decay-corrected activity per unit volume. The TACs were quantified as percent injected dose per unit volume (% ID/g) for subsequent analysis. All data in this paper performed a statistical significance test via t-tests (and nonparametric tests). Autoradiography in vitro Mouse brain slices (20 μm) were pre-incubated in 50 mM Tris-HCl buffer for 30 min, followed by incubation with [18F]XW-032 (1 mCi/L) in fresh buffer. For blocking studies, standard compound XW-032 (10 μM) was included. After incubation, brain slices were washed with cold Tris-HCl buffer and distilled water, air-dried, and exposed to an imaging plate (BAS-MS2025, GE Healthcare). Autoradiograms were acquired, and regions of interest (ROIs) were defined visually. Signal intensity was quantified using ImageJ. Results Bioinformatic analysis of the correlation between Ido1 and AD To validate the correlation between IDO1 and AD, we reanalyzed the public dataset using the GSE33000 dataset, containing transcriptomic data from human AD brains. As IDO1 is known to be upregulated in response to pro-inflammatory stimuli, we first assessed the expression profiles of key inflammatory mediators, including STAT1, IL1B, TNF, and IFNG. As shown in Fig. 2A, Differences between two groups were assessed using the Wilcoxon rank-sum test, several inflammatory markers exhibited significant upregulation in AD patients compared to healthy controls. Specifically, the expression levels of IFNG, IL1B, and TNF were significantly upregulated in the AD group (P < 0.001), suggesting a robust inflammatory response within the AD brain microenvironment. Interestingly, STAT1, a downstream transcription factor of interferon signaling, also showed a significant increase in expression, indicating the activation of the JAK/STAT signaling axis during AD progression. Bioinformatic analysis of IDO1 and AD. (A) Identification of Pro-inflammatory Cytokine (IFNG, IL1B, STAT1, TNF)Dysregulation in Brains of AD Patients. (B) Correlation Analysis Between Pro-inflammatory Cytokines and IDO1 Expression. (C) Proposed Mechanism of IDO1-Mediated Neuroscience in AD. IDO1 is a rate-limiting enzyme in the kynurenine pathway, often induced by inflammatory stimuli. We performed Pearson correlation analyses to evaluate the relationship between the identified pro-inflammatory cytokines and IDO1 expression (Fig. 2B). IFNG showed a remarkably strong positive correlation with IDO1 (r = 0.792, P < 0.001), suggesting it might be the primary driver of IDO1 induction. Moderate positive correlations were observed for TNF (r = 0.477, P < 0.001) and STAT1 (r = 0.229, P < 0.001). In contrast, IL1B showed no significant correlation with IDO1 expression (r = 0.086, P = 0.132). Based on the expression data and correlation results, we constructed a mechanistic model to illustrate how the AD brain environment promotes neurodegeneration through the IDO1/Kynurenine pathway (Fig. 2C). The pathological hallmarks of AD—including Aβ plaques, neurofibrillary tangles, and activated microglia/astrocytes—trigger the release of pro-inflammatory cytokines (IFNG, IL1B, STAT1, TNF) [31] [32]. Specifically, IFNG activates the JAK/STAT1 signaling pathway [33], which directly promotes the transcription of IDO1 [34]. The subsequent upregulation of IDO1 drives overactivation of the kynurenine pathway, resulting in two major pathological consequences: neurotoxicity and immunosuppression. Collectively, these findings identify the IFNG–STAT1–IDO1 axis as a critical link between chronic neuroinflammation and neuronal loss in AD, and provide a molecular framework for the elevated signal observed in our IDO1-targeted PET imaging. PET Radioligand design and properties Given the pivotal role of IDO1 in tumor immune evasion [35] and its emerging significance in CNS disorders, the development of an IDO1-targeted PET tracer requires a stringent balance between binding potency, selectivity, and BBB permeability. To develop a PET tracer targeting IDO1 in AD, we strategically modified a potent IDO1 inhibitor, XW-032, as the lead scaffold. The discovery of XW-032 began with high-throughput screening (HTS) to identify the initial hit compound XW-01, followed by iterative optimization to yield a structurally novel scaffold with potent IDO1 inhibitory activity (IC50 = 21 ± 5 nM) [28]. The radiolabeling site was selected by replacing a fluorine atom on the phenyl ring with fluorine-18, resulting in the novel tracer [18F]XW-032 (Fig. 3A). The chemical design and radiolabeling strategy. (A) Fluorine-18 radiolabeling strategy based on IDO1 inhibitor. (B) The Pharmaceutical and chemical properties of XW-032. (C) Predicted binding pattern of XW-032 and IDO1 (PDB: 6AZW). Physicochemical evaluation confirmed the CNS-drug likeness of XW-032 (Fig. 3B). The molecule possesses an optimized molecular weight (MW = 354.1) and a topological polar surface area (tPSA) of 41.46 Å2, well below the 60 Å2 threshold for BBB penetration. Its lipophilicity (cLogP = 4.6) and a Schrodinger software predicted CNS activity score (scope: −2∼2) of highest +2, further supporting robust brain uptake and minimal active efflux. Molecular docking with crystal structure (PDB: 6AZW) was performed to explain the binding modality (Fig. 3C). The docking results revealed that the tracer occupies the active site with high complementarity. As shown in the zoomed-in view (Fig. 3D), the molecule establishes key interactions with several conserved residues, including hydrophobic interactions and potential hydrogen bonding with residues such as S167, F226, and L331. The fluoro-pyridine and phenyl moieties are well-accommodated in the hydrophobic pockets, stabilizing the ligand-protein complex. These theoretical findings provide a robust structural basis for the high binding affinity observed in subsequent in vitro and in vivo evaluations. Chemical Synthesis of compound XW-032 and precursor As shown in Scheme 1, the synthesis procedure of XW-032 involves a multi-step procedure designed to construct a molecule with good biological activity and drug-like properties. Initially, compound 3 is obtained via an esterification reaction, followed by the preparation of the boronate intermediate 5 through borylation. A subsequent Suzuki coupling yields compound 7. This is then subjected to alkaline hydrolysis to afford the key intermediate 8. Finally, an acid–base condensation furnishes the reference standard compound XW-032 and precursor of the tracer. In total, 6 chemical steps are required, yielding good results, thus providing a solid foundation for subsequent radiolabeling. Chemical Synthesis of XW-032a and Radiology Chemical Synthesis of [18F]XW-032b aReagents and conditions of Chemical Synthesis: (a) MeOH,0-25 °C, 2.5 h; (b) Pd(dppf)Cl2 DCM, KOAc, DMSO, 85 °C, 4 h; (c) Pd(dppf)Cl2, Cs2CO3, Dioxane, H2O, 100 °C, 8 h; (d) LiOH.H2O, THF, H2O, 25 °C, 2 h; (e) HATU, DIPEA, DMF; (f) HATU, DIPEA, DMF. bReagents and conditions of Radiology Chemical Synthesis: (a) Ethanol, 85 °C,30 min; b. CH3CN, DMSO, 130 °C, 30 min. Radiology Chemical Synthesis procedure The radiosynthesis of [18F]XW-032 was accomplished via a two-step continuous process (Scheme 1). Initially, the phenolic precursor was reacted with the ruthenium complex CpRu(COD)Cl in ethanol at 85 °C to generate the η6 π-coordinated Ru–phenol intermediate without the need for isolation. Subsequently, a CH3CN solution of ClIm was introduced, producing a stable complex suitable for eluting [18F]fluoride from a Chromafix 30-PS-HCO3 anion exchange cartridge. The eluted activity was combined with a DMSO/CH3CN (1:1, v/v) mixture and heated at 125 °C for 30 min to achieve 18F-deoxyfluorination. Following the reaction, the crude mixture was purified by semipreparative HPLC to get [18F]XW-032. The product fraction was collected and reformulated. [18F]XW-032 was obtained in 25–31% radiochemical yield (non-decay-corrected, EOS), with radiochemical purity >95% (Fig. S1 and S2) and molar activity of 107.5–111.2 GBq/μmol (EOB). PET imaging studies of [18F]XW-032 in mice To evaluate the brain-imaging potential of [18F]XW-032, dynamic PET/CT scans were performed in healthy mice. The tracer exhibited rapid blood-brain barrier (BBB) penetration and sustained cerebral retention, characterized by a favorable maximum standardized uptake value (SUVmax around 0.7) and slow clearance kinetics (Fig. 4A–B). To verify target specificity, a self-blocking study was conducted using unlabeled XW-032 (0.3 mg/kg). A transient increase in total brain uptake was initially observed in the blocking group due to peripheral target saturation. A significant reduction of TACs was observed after blood normalization in the blocking group (Fig. 4C), normalized the brain uptake curves with the maximum radioactivity in the blood at each time point (Fig. S3). AUC (Area under curve) of baseline is significantly higher than that of the blocking group (Fig. 4D). These results confirm that the brain signal of [18F]XW-032 is primarily driven by specific IDO1 binding. PET/CT imaging of baseline and blocking group. (A) PET/CT obtained procedure and images of mouse brains (20−60 min) of [18F]XW-032. (B) SUV of whole-brain time-activity curves (TACs) under self-blocking (0.3 mg/kg) and baseline (n = 3). (C) Normalized TACs of [18F]XW-032. (normalized the brain uptake curves with the maximum radioactivity in the blood at each time point). (D) AUC (Area under curve) of SUV (Standardized Uptake Value). (E) Representative PET/CT regional images of [18F]XW-032. (F-G) Baseline and self-blocking regional-brain TACs of under self-blocking (0.3 mg/kg) and baseline (n = 3). (H-G) Normalized TACs of [18F]XW-032 in the Cortex Hippocampus and striatum. (K-M) AUC of the Cortex, Hippocampus, and striatum. Detailed voxel-based analysis across eight functional brain regions revealed a heterogeneous distribution of [18F]XW-032 (Fig. 4E). In healthy mice, all examined regions, including the cortex, hippocampus, and striatum, showed stable tracer retention with distinct spatial profiles (Fig. 4F–G). Quantitative comparisons of plasma-normalized SUV showed that the cortex and hippocampus exhibited significantly higher baseline uptake compared to the blocking group after the initial distribution phase (>6 min), as further evidenced by AUC analysis (Fig. 4H–L). In contrast, although the striatum showed a trend toward specific binding, the difference did not reach statistical significance (Fig. 4M), potentially reflecting the low basal IDO1 expression in the absence of neuroinflammation. Collectively, these findings demonstrate that [18F]XW-032 can effectively map the regional expression of IDO1 in vivo. To evaluate the binding affinity and specificity of the novel probe [18F]XW-032 in brain tissues, in vitro radio-autophotography was performed on brain sections from AD mice, wild-type (WT) (Fig. 5A). The quantitative analysis for gray values (Fig. 5B) demonstrated that [18F]XW-032 exhibited the highest uptake in the brain sections of AD mice (253.5 ± 12.4). This uptake was significantly higher than that observed in WT mice (224.2 ± 10.8, P < 0.01), correlating with the upregulated IDO1 expression in the AD pathological environment. To further confirm the target specificity, a blocking group was established by co-incubating AD sections with unlabeled XW-032. The signal intensity in the blocking group was significantly reduced (134.8 ± 14.2, P < 0.001), indicating that the binding of [18F]XW-032 is highly specific to the IDO1 protein. Furthermore, beyond traditional P-value testing, we employed estimation graphics to visualize the magnitude and precision of the biological effects (Fig.s 5C and 5D). The Gardner-Altman/Cumming plots revealed a substantial mean difference between the AD and Blocking groups, with the 95% confidence interval (CI) clearly excluded from zero (Fig. 5C), confirming a robust displacement effect. Similarly, the estimation plot in Fig. 5D highlights the precision of the difference between AD and WT groups, emphasizing the probe's sensitivity in distinguishing AD-related IDO1 elevation from physiological levels. Autoradiographic test. (A) Autoradiographic images of AD, WT and Blocking. (B) The relative radioactive uptake. (C) Estimation Plot showing the difference in means between AD and blocking groups. (D) Estimation plot showing the difference in means between AD and WT. Dynamic PET/CT imaging was performed in 5XFAD and age-matched WT mice to evaluate the in vivo brain uptake of [18F]XW-032. There's a published paper showed increased IDO1 activity/expression in in the 5×FAD mouse model [36]. Representative images (20–60 min post-injection) showed higher brain uptake in 5XFAD mice compared with WT controls (Fig. 6A). Quantitative analysis of SUV and %ID/g demonstrated distinct kinetic profiles between the two groups. As shown in Fig. 6B and D, TACs indicated rapid brain uptake of [18F]XW-032 in both groups, reaching uptake peak levels within minutes after injection. However, in 5XFAD mice, tracer retention was prolonged, whereas faster washout was observed in WT mice, resulting in higher brain-to-blood contrast in the AD model. AUC analysis further showed a significant increase in the 5XFAD group compared with WT (P < 0.05 in SUV analysis and P < 0.01 in %ID/g). These results are consistent with the increased IDO1-associated signal observed in autoradiography, supporting the ability of [18F]XW-032 to differentiate AD-related pathology from normal brain tissue in vivo. Whole-body biodistribution of [18F]XW-032 was evaluated in mice brain uptake remained relatively stable, consistent with the preceding analysis (Fig. S4). PET/CT imaging in 5XFAD and WT mice of [18F]XW-032. (A) PET/CT images of mouse brains (20−60 min). (B) %ID/g TACs of whole-brain. (C) AUC (Area under curve) of %ID/g. (D) SUV whole-brain TACs of AD and WT mice. (F) AUC (Area under curve) of whole-brain SUV. (F) Representative PET/CT regional images of 5XAD and WT mice. (G-H) Regional-brain SUV TACs of 5XAD and WT mice. (I-K) SUV TACs of [18F]XW-032 in the Cortex Hippocampus and striatum. (L-N) AUC of the Cortex, Hippocampus, and striatum. To further characterize the neuroanatomical binding profile of [18F]XW-032, we performed a detailed VOI (Volume of Interest) analysis across eight functional brain regions (Fig. 6F). The SUV in AD (5XFAD) mice was globally increased across all examined regions compared to WT controls (Fig.s 6G-6H). Notably, in regions heavily associated with neuroinflammation, like the Cortex, Hippocampus, and Striatum, the TACs showed a distinct and sustained separation between the two groups (Fig.s 6I-6K). Following the initial distribution phase (>20 min), the probe retention in these inflammatory hotspots remained significantly higher in the AD group, while the WT group exhibited typical clearance kinetics. Quantitative comparison of the AUC for these three key regions revealed robust statistical differences: Cortex (P = 0.0196), Hippocampus (P = 0.0237), and Striatum (P = 0.0169) (Fig.s 6L-6N). The higher AUC values in these regions validate that [18F]XW-032 can map the spatial heterogeneity of IDO1 expression in vivo, consistent with the localized neuroinflammatory activity characteristic of the 5XFAD model. Overall, compared to the WT group, [18F]XW-032 demonstrated statistically significant differences in both whole-brain uptake and uptake in key brain regions within the AD model. Stability studies in vitro and in vivo of [18F]XW-032 in mice Additional in vitro stability studies of [18F]XW-032 were performed in PBS and saline. In addition, the in vivo metabolic stability of [18F]XW-032 was evaluated using urine and plasma samples. As shown in Fig. S5, The results demonstrated that [18F]XW-032 remained stable during the investigated period, with radiochemical purity of >95% after incubation in PBS and saline for 60 min. Analysis of plasma and urine samples collected at 60 min after intravenous administration showed that approximately 95% of the detected radioactivity corresponded to intact [18F]XW-032, indicating satisfactory in vivo metabolic stability. Discussion IDO1 is overexpressed in the inflammatory milieu of CNS diseases (Fig. 7); by depleting tryptophan, it mediates associated immune responses, thereby serving as a potential target for both immunotherapy and molecular imaging [37]. In 2024, Minhas et al. published a paper reporting that Aβ and tau oligomers activate the IDO1 enzyme within astrocytes, thereby rendering them unable to supply neurons with their metabolic "fuel"—lactate. However, pharmacological inhibition of IDO1 can miraculously reactivate this metabolic support pathway, restoring cerebral glucose metabolism across various mouse models of AD and human cell models, while also rescuing long-term memory and synaptic plasticity [18]. This highlights the critical significance of IDO1 upregulation in both the treatment and diagnosis of Alzheimer's disease. This study aims to investigate the relationship between IDO1 and AD, and to develop PET probes capable of specifically mapping IDO1 expression in the brains of AD patients. With the ultimate goal of guiding the diagnosis and treatment of AD. The successful development of the IDO1 efficiency tracer will be a significant milestone in the molecular imaging of the KP in AD. By integrating rational design, advanced ruthenium-mediated radiochemistry, bioinformatic validation, and multi-scale imaging, this study supports IDO1 as a PET target associated with neuroinflammatory processes. Developing IDO1 PET probes based on the high expression of IDO1 in the brains of AD patients to aid in AD diagnosis, treatment, and drug development. A cornerstone of our findings is the mechanistic link between AD-driven neuroinflammation and IDO1 expression. Analysis of the GSE33000 dataset suggests that IDO1 upregulation in AD is associated with a pro-inflammatory environment, including increased IFNG, IL1B, STAT1, and TNF signaling, which has been further shown to synergistically induce IDO1 expression via the IFNG, TNF relevant signaling pathways, consistent with previous reports [38,39]. As illustrated in our proposed mechanism (Fig. 7), the AD pathological microenvironment triggers a shift in macrophage/microglial polarization toward the pro-inflammatory M1 phenotype. Consistent with previous studies, the results show an elevated release of cytokines such as IFNG, IL-1, and TNF, which act as potent exogenous stimuli to promote IDO1 expression [40]. We found that IFNG exhibits the strongest correlation with IDO1 upregulation. This "AD-cytokine-IDO1" axis not only validates IDO1 as a surrogate biomarker for neuroinflammation in AD but also suggests that IDO1 has high potential as a PET imaging target in AD models. Based on the conclusion that IDO1 expression is upregulated in the brains of AD patients, we proceeded to develop an IDO1 brain probe. We chose XW-032 as a potential IDO1 PET tracer due to its optimized molecular architecture [28]. The probe's low tPSA and balanced lipophilicity (cLog P) facilitate efficient BBB penetration, while molecular docking confirms its stable binding within the IDO1 catalytic pocket. XW-032 exhibits the most favorable combined profile even head-to-head comparison with existing IDO1 PET tracers, both oncology-focused and substrate-based tracers (Table S1) [[41], [42], [43]]. Building upon a high-affinity IDO1 inhibitor scaffold, we utilized a ruthenium-mediated deoxyfluorination strategy to achieve efficient fluorine-18 labeling. The ruthenium-mediated deoxyfluorination approach proved operationally straightforward, moisture-tolerant, and compatible with commercially available automated synthesizers. Optimization of temperature and reagent ratios afforded consistent yields, highlighting the method's suitability for routine radiopharmaceutical production [44]. The BBB penetration ability of [18F]XW-032 was proved by baseline, and the specificity was initially proved by the blocking group. However, the heterogeneity of probe uptake across different brain regions has not been fully elucidated, primarily because inflammatory responses in the brains of healthy mice are very mild. The specificity of [18F]XW-032 was further rigorously validated by our in vitro autoradiography, where the excess unlabeled inhibitor successfully blocked the [18F]XW-032 signal, and the bioinformatic-led discovery of IDO1 elevation in AD-vulnerable regions. In vivo PET imaging demonstrated that [18F]XW-032 effectively crosses the blood-brain barrier (BBB) and accumulates in the brains of 5XFAD mice, including the cortex and hippocampus. The significantly higher AUC values in the Cortex and Hippocampus are particularly noteworthy. These regions are known to exhibit substantial Aβ deposition and associated microglial activation, which may contribute to increased IDO1 expression. The observed differences in AUC between 5XFAD and WT mice indicate that the tracer [18F]XW-032 can map the spatial heterogeneity of the neuroinflammatory landscape and distinguish pathological from baseline conditions. Overall, we demonstrated that [18F]XW-032 is a promising probe with excellent specificity, favorable brain penetrability, and desirable metabolic stability via in vivo and in vitro studies. Notably, in experiments using AD mice, [18F]XW-032 exhibited significantly higher uptake in both the whole brain and inflammation-associated regions compared with wild-type (WT) mice. Compared with TSPO imaging, which can be affected by genetic polymorphisms (e.g., rs6971), IDO1-targeted imaging may provide complementary information related to metabolic and enzymatic changes within the KP [45]. [18F]XW-032 may have potential for clinical translation as a tool for detecting neuroinflammatory changes and monitoring therapeutic interventions targeting IDO1. As illustrated in Fig. 7, [18F]XW-032 may have several potential clinical applications, including early detection of neuroinflammatory changes by identifying elevated IDO1 expression, supporting patient stratification by reflecting neuroinflammatory status, and serving as a noninvasive tool for drug development by allowing longitudinal assessment of target engagement and treatment response [46]. However, this study still requires further data to support subsequent clinical translation. In summary, the present study demonstrates that [18F]XW-032 enables imaging of IDO1-associated signals in vivo. By linking inflammatory processes to PET imaging, this tracer provides a potential approach for investigating disease-related changes in AD and may support the development of targeted therapeutic strategies. Conclusions This study establishes [18F]XW-032 as a highly specific and brain-permeable PET radiotracer for imaging IDO1 in Alzheimer's disease. By integrating rational design, ruthenium-mediated synthesis, and multiscale validation, we demonstrated that [18F]XW-032 effectively maps IDO1 elevation driven by the TNF neuroinflammatory axis. The significant tracer retention observed in the cortex and hippocampus of 5XFAD mice underscores its sensitivity to localized pathological changes. Our findings position [18F]XW-032 as a robust molecular imaging tool for early AD diagnosis, patient stratification, and the longitudinal monitoring of IDO1-targeted therapeutic efficacy. This work provides a new dimension for non-invasive assessment of the kynurenine pathway in neurodegenerative disorders. CRediT authorship contribution statement Hua Cheng: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Nilay S. Narayan: Writing – original draft, Formal analysis, Data curation, Conceptualization. Yongle Wang: Investigation, Data curation. Yanli Wang: Investigation, Data curation. Yafei Huang: Investigation, Data curation. Hongzhi Yang: Investigation, Data curation. Ramya Tokala: Investigation, Data curation. Jinwu Yan: Investigation, Data curation. Joseph William Downey: Investigation, Data curation. Anne Siewko: Investigation, Data curation. Breanna Lizeth Varela: Investigation, Data curation. Madelyn Rose Hallisey: Investigation, Data curation. Tiffany A. Geronimo: Investigation, Data curation. Mary Margaret Banick: Investigation, Data curation. Rudolph E. Tanzi: Supervision, Formal analysis. Can Zhang: Supervision, Data curation. Yulong Xu: Writing – review & editing, Supervision, Formal analysis. Changning Wang: Writing – review & editing, Supervision, Funding acquisition, Formal analysis. Declaration of generative AI and AI-assisted technologies During the preparation of this work, the author(s) used ChatGPT (OpenAI) for language editing and polishing. The author(s) reviewed and edited the output as needed and take full responsibility for the content of the published article. Declaration of competing interest The authors declare no conflicts of interest. Acknowledgments This study is supported by the funding from the National Institutes of Health (AG081195 and AG086433), the Cure Alzheimer's Fund and NIH grant S10OD034259. The authors express their sincere appreciation to the radio pharmacy and imaging teams at the Athinoula A. Martinos Center of MGH for their indispensable contributions to executing rodent studies and facilitating radioisotope production. Footnotes Supplementary data to this article can be found online at https://doi.org/10.1016/j.mtbio.2026.103682. Contributor Information Yulong Xu, Email: yulong.xu@mgh.harvard.edu. Changning Wang, Email: cwang15@mgh.harvard.edu. Supplementary data The following is the Supplementary data to this article: Multimedia component 1 Data availability Data will be made available on request. References Associated Data Supplementary Materials Multimedia component 1 Data Availability Statement Data will be made available on request.