Novel Positron Emission Tomography Radiotracers for Imaging Mitochondrial Complex I

Abstract

Mitochondrial dysfunction has been indicated in neurodegenerative and other disorders. The mitochondrial complex I (MC-I) of the electron transport chain (ETC) on the inner membrane is the electron entry point of the ETC and is essential for the production of reactive oxygen species. Based on a recently identified β-keto-amide type MC-I modulator from our laboratory, an ¹⁸F-labeled positron emission tomography (PET) tracer, ¹⁸F-2, was prepared. PET/CT imaging studies demonstrated that ¹⁸F-2 exhibited rapid brain uptake without significant wash out during the 60 min scanning time. In addition, the binding of ¹⁸F-2 was higher in the regions of the brain stem, cerebellum, and midbrain. The uptake of ¹⁸F-2 can be significantly blocked by its parent compound. Collectively, the results strongly suggest successful development of MC-I PET tracers from this chemical scaffold that can be used in future mitochondrial dysfunction studies of the central nervous system.

Graphical Abstract:

INTRODUCTION

Alzheimer’s disease (AD) is a neurodegenerative disorder characterized by a progressive deterioration of cognitive functions. Approximately, 1 in 10 individuals over the age of 65 suffers from AD, a figure set to increase to 1 in 3 over the age of 85.¹ Although numerous studies have been dedicated to understanding the mechanisms leading to AD,^2–5 the exact etiology of AD remains elusive. Among the studied pathogenic factors, mitochondrial dysfunctions in brain cells have been suggested to play critical roles in the ultimate development of AD.^6–9 Mitochondria are dynamic organelles that control cellular energy production, redox signaling, and cell survival.¹⁰ The bioenergetic function of mitochondria is executed by the electron transport chain (ETC) via oxidative phosphorylation and aerobic adenosine triphosphate (ATP) production. The ETC is ubiquitously expressed throughout the inner mitochondrial membrane and consists of mitochondrial complexes I–V. Among the five ETC complexes, mitochondrial complex I (MC-I) is the entry point for electrons into the ETC and catalyzes electron transfer from nicotinamide adenine dinucleotide to ubiquinone.¹¹ Accrued evidence indicates that the MC-I abnormalities decrease cellular energy production and mitochondrial respiration in the brain, which have been recognized as an essential contributor to AD.^12,13 Furthermore, mitochondria are the major organelles that produce the majority of reactive oxygen species (ROS) through MC-I and MC-III.^14–16 Recent studies have indicated the pathological roles of mitochondrial ROS in the central nervous system (CNS) inflammatory responses.¹⁷ Genetic studies from Alzheimer’s Disease Neuroimaging Initiative (ADNI) also revealed the increased risk of AD development with mutations of MC-I.^18,19

While mitochondrial dysfunction has shown essential roles in AD, much is still debated and their contributions to disease development remain elusive. Therefore, a quantitative and noninvasive monitoring of MC-I in the human brain under healthy and pathological conditions would be valuable to help understand the contributions of MC-I dysfunction to disease pathology and to serve both diagnostic and therapeutic purposes. In this regard, positron emission tomography (PET) would be an ideal method for imaging the biological functions of MC-I. PET has been developed to visualize molecular targets in inaccessible tissues, particularly in the brain, therefore offering distinct advantages for AD-related research.²⁰ Given the roles of MC-I in the production of ATP, ROS, its link to neuroinflammation and involvement in the development of neurodegenerative disorders including AD, MC-I might serve as a meaningful biomarker for mitochondrial dysfunction, neuroinflammation, and oxidative stress. Consequently, PET radiotracers of MC-I will be valuable for both fundamental and translational research. Indeed, over the past few years, considerable efforts have been implemented toward developing MC-I PET radiotracers, and early attempts using rotenone and dihydrorotenol as potential PET radiotracers to image MC-I have been conducted in animal models.^21,22 Recently, several chemically distinct PET radiotracers have been developed and tested to image MC-I in preclinical studies including ¹⁸F-BMS, ¹⁸F-BCPP-EF, ¹⁸F-BCPP-BF, and ¹¹C-BCPP-EM (Figure 1).^23–35 Among them, ¹⁸F-BCPP-EF is being evaluated in clinical studies in healthy individuals and AD patients.^36–39 These radiotracers also target the rotenone binding site of MC-I and show relatively high nonspecific binding in the brain. Given that multiple sites exist within MC-I and they have different roles in the production of ROS and ATP, novel PET radiotracers of MC-I that target sites other than the rotenone site are needed to help understand its pathological roles and facilitate therapeutic development.

Figure 1.

Chemical structures of ¹⁸F-BMS, ¹⁸F-BCPP-EF, ¹⁸F-BCPP-BF, and ¹¹C-BCPP-EM.

Recently, our group has successfully developed a lead compound, ZCM-I-1 (Figure 2), showing promising neuroprotective activities both in vitro and in vivo in preclinical AD models.^40,41 Mechanistic studies revealed that this lead compound selectively modulates the function of MC-I by binding to the MC-I flavin mononucleotide (FMN) site, representing a novel mechanism of action.⁴² Inspired to further investigate the pathological roles of MC-I in AD and other neurodegenerative disorders, herein, we describe the development of a PET radiotracer of ZCM-I-1 as a potential living brain imaging tool.

Figure 2.

Chemical structures of rationally designed lead compound ZCM-I-1 and analog 2.

RESULTS AND DISCUSSION

Before we start the attempt for radiolabeling of ZCM-I-1, we further confirmed its interaction with MC-I using a cellular thermal shift assay (CESTA).^43,44 Mouse brain mitochondria and MC-I were isolated and purified following the procedures that we described recently.⁴² As shown in Figure 3A, upon incubation with ZCM-I-1, the stability of NDUFV1, the subunit of the FMN binding site and the major protein labeled by a photoaffinity labeling probe of ZCM-I-1,⁴² from the MC-I was increased from the range of 54 to 67.3 °C, as evidenced by the Western blotting results. Collectively, the CESTA results suggested the physical interaction of ZCM-I-1 with the brain MC-I.

Figure 3.

(A and B) Protein stability of the NDUFV1 subunit of complex I. Mouse brain mitochondrial complex I was incubated with (A) ZCM-I-1 or (B) compound 2 (1 μM) for 1 h. The proteins were resolved by SDS-PAGE and probed by the anti-NDUFV1 antibody. The image represents results from one of the three independent experiments. (C) MST binding of compound 2 to MC-I.

Next, radiochemistry was attempted for ZCM-I-1 to develop a suitable PET radiotracer. By analyzing the chemical structure of ZCM-I-1, the logic consideration for radiolabeling would be the 5-OCH₃ group on the indole ring with a –O¹¹CH₃ to maintain the intact structure of ZCM-I-1 and its physicochemical properties and favorable pharmacology toward MC-I.^42–44 However, synthetically, multistep radiochemistry, including –OH or –NH protection and deprotection reactions, is needed, and this challenge will hinder the future clinical applications. Our recent structure–activity relationship (SAR) studies of ZCM-I-1 demonstrated that a 4-F analog of ZCM-I-1 (compound 2, Figure 2) exhibits comparable protective potency in the phenotypic screening assay.⁴⁵ The CESTA also supported its physical interaction with the NDUFV1 subunit of MC-I (Figure 3B). We also measured the binding affinity of this compound to the purified MC-I from the bovine heart using the microscale thermophoresis (MST) assay,⁴⁶ taking advantage of the autofluorescence of MC-I.⁴⁷ As shown in Figure 3C, compound 2 binds to the MC-I with a K_D of approximately 500 nM. Therefore, taking advantage of the presence of a 4-F substituent on compound 2, we decided to make an ¹⁸F-labeled radiotracer, ¹⁸F-2 (Scheme 1). An ruthenium-mediated radiofluorination method was employed to convert ZCM-I-1 to afford ¹⁸F-2 by using air-stable ruthenium complex CpRu(COD)Cl, chloroimidazolium chloride, and ¹⁸F-fluoride as reported.⁴⁸ Specifically, ZCM-I-1 was first activated by CpRu(COD)Cl to form an η6 π-coordinated ruthenium–phenol complex, which was used to elute ¹⁸F-fluoride off an anion exchange cartridge by the subsequent addition of chloroimidazolium chloride. The elution efficiency of ¹⁸F-fluoride achieved 89 ± 6% (n = 5). After adding solvent (DMSO/acetonitrile (v/v) = 1:1), the reaction mixture was heated to 125 °C for 30 min followed by high-performance liquid chromatography (HPLC) purification and reformulation to afford ¹⁸F-2 in sterile saline containing <10% (v/v) ethanol. The overall yield for the radiolabeling chemistry was 11 ± 4% (n = 3, decay corrected) in a total synthesis time of 70 ± 5 min from end of cyclotron bombardment. The molar activity was around 2 Ci/μmol, and the radiochemical purities were consistently confirmed being greater than 95% at time of injection.

Scheme 1.

Radiochemical Synthesis of ¹⁸F-2

After the successful radiochemistry, a 60 min dynamic PET scan was attempted in C57BL/6 mice (25–30 g, male, n = 2) after administration of ¹⁸F-2 via tail vein injection to determine its biodistribution (Figure 4). The PET imaging with ¹⁸F-2 revealed a high uptake in the liver (7.16% ID/cc at 30 min post-injection) and kidney (8.17% ID/cc at 30 min post-injection). A gradual increase in the uptake with time was also observed in the kidney, suggesting that ¹⁸F-2 may be mainly excreted through urine. However, the proximal tubular epithelial cells in the kidney contain ample mitochondria, which helps maintain the homeostasis of the cells to keep the best conditions for optimum kidney functions. Therefore, the intensive interactions between ¹⁸F-2 and mitochondria, leading to elevated uptake in the kidney, cannot be ruled out. Notably, ¹⁸F-2 displayed relatively slow clearance rates in the brain (brain_5min/brain_60min = 1.02), blood (blood_5min/blood_60min = 1.24), heart (heart_5min/heart_60min = 1.29), and lung (lung_5min/lung_60min = 1.19). A relatively low uptake in muscle was observed for ¹⁸F-2 followed by a gradual increase. A similar pattern was also observed in the bone for ¹⁸F-2, implying possible defluorination of ¹⁸F-2 in the bone tissue.

Figure 4.

Biodistribution histogram of ¹⁸F-2 in C57BL/6 mice. All data are the mean. Data are expressed as the percentage of injected dose per cubic centimeter (% ID/cc).

Next, we further analyzed the properties of ¹⁸F-2 in the brain tissue. The representative PET/CT image (sagittal, summed images 20–60 min) and time–activity curves (TACs) are shown in Figure 5. ¹⁸F-2 exhibited a modest but rapid uptake to reach a constant level (~1.60% ID/cc) without significant washout during the scan period. To further verify the clearance rate of ¹⁸F-2, we performed a 120 min dynamic PET scan in C57BL/6 mice (25–30 g, male, n = 4). The results still showed no washout between 60 and 120 min (see the Supporting Information, Figure S1). Since compound 2 is a neutral and noncovalent compound, this suggests that ¹⁸F-2 has a relatively long retention time in the brain tissues. However, the roles of blood flow and tissue extraction efficiency need to be investigated to fully understand the retention properties of this compound in the future since radiotracers with long retention time have been tested for cardiac imaging.^21,22,24 Notably, ¹⁸F-2 showed higher binding signals in the brain stem, cerebellum, and midbrain, and this might indicate either higher MC-I expression or increased blood flow in these brain sections of the rodent. To test the binding specificity of ¹⁸F-2 in the brain, ideally, a different agent from a different chemical class that shares the same mechanism of action should be used to conduct the blocking studies. However, in this regard, there are no known ligands that share the same mechanism of action as ZCM-I-1 by binding to the FMN site of MC-I. Although resveratrol has been reported to bind to the FMN site,⁴⁹ given its promiscuity and PAIN structure, it may not provide valuable information as a blocking agent. Therefore, we performed self-blocking experiments in C57BL/6 mice with its parent compound 2 (Figure 6A, 25–30 g, male, n = 2). However, the brain uptake was significantly increased in self-blocking mice compared with that of baseline mice, which could be, to some extent, attributed to the increased free ¹⁸F-2 in blood after pretreatment with compound 2 (see the Supporting Information, Figure S1). Therefore, the radioactivity accumulation is presented as the whole brain radioactivity normalized by maximal blood radioactivity. Pretreatment of compound 2 (5 mg/kg) led to a significant reduction in the brain radioactivity (~34%, based on the area under the curve from 20 to 60 min). Surprisingly, under the self-blocking condition, no significant elimination of ¹⁸F-2 was observed during the PET scan. This suggests a strong binding of this radiotracer or nonreversible binding in the mouse brain. Although the properties of 2 suggest it as a favorable CNS therapeutic agent, the kinetics of ¹⁸F-2 shown here warrant further optimization of the lead structure to develop MC-I PET tracers that are suitable for quantitative imaging analysis of brain function as the imaging probe. Collectively, the characterization of ¹⁸F-2 in mice suggested the potential development of MC-I PET radiotracers by the further SAR optimization process using ZCM-I-1 as a lead compound.

Figure 5.

Brain regional radioactivity in mice. (A) Sagittal sections were superimposed on the corresponding template MRI. (B) Coronal sections were superimposed on the corresponding template MRI. (C) Transverse sections were superimposed on the corresponding template MRI. (D) TACs for brain regions. Blue, cortex; red, striatum; green, thalamus; yellow, cerebellum; brown, brain stem; purple, midbrain; and dark green, hippocampus. All data are the mean (n = 6). Data are expressed as the percentage of injected dose per cubic centimeter (% ID/cc).

Figure 6.

(A) Specificity of ¹⁸F-2 toward MC-I in the mouse brain. TACs of whole brain uptake of ¹⁸F-2 in baseline and self-blocking conditions. All data are the mean. “*” means whole brain radioactivity normalized by maximal blood radioactivity. (B) Percentages of parent ¹⁸F-2 in the plasma and brain homogenate of C57BL/6 mice (n = 4) at 30 min post-injection.

To evaluate the stability of ¹⁸F-2, radiometabolites in the plasma and brain homogenates of C57BL/6 mice were studied at 30 min post-injection. The percentages of parent ¹⁸F-2 and the corresponding radiometabolites, which were determined by radio-HPLC (see the Supporting Information), are depicted in Figure 6B. The radioactivity in plasma corresponding to ¹⁸F-2 was 81% at 30 min post-injection. Four radiometabolites, three of which are more polar and one is less polar than ¹⁸F-2, were observed in the plasma with 19% radioactivity in total. The possible metabolic sites within ¹⁸F-2 were predicted by SMARTCyp (see the Supporting Information, Table S1).⁵⁰ The fraction corresponding to ¹⁸F-2 in mouse brain homogenates was 85% at 30 min post-injection. However, a notable difference in metabolism in the brain homogenates compared with that in the plasma is that only two radiometabolites, one of which is more polar, while the other one is less polar than ¹⁸F-2, were observed with 15% radioactivity, and they are identical to the ones found in the plasma. These results indicate that some of the ¹⁸F-labeled metabolites might be able to pass through the blood brain barrier. Thus, further structure optimization of ¹⁸F-2 is still needed.

Based on the results of in vivo PET studies and radiometabolite analysis, we further preliminarily evaluated the involvement of MC-I in AD by conducting ¹⁸F-2 PET imaging studies in wild-type and AD transgenic mice, given that the literature suggested the roles of MC-1 dysfunction in AD.^12,13 To this end, 5× familial Alzheimer’s disease (FAD) transgenic mice (25–30 g, female, n = 4) and wild-type mice (25–30 g, female, n = 4) were employed in the PET/CT imaging studies with ¹⁸F-2 (Figure 7A). As shown in Figure 7B, the results indicated a slight decrease in the brain radioactivity of ¹⁸F-2 in the 5× FAD transgenic mice compared to the wild-type controls. The brain washout profile in 5× FAD transgenic mice is similar to that of wide-type mice. These results suggest that ¹⁸F-2 may have the potential to differentiate AD from normal mice and also imply that the expression levels of MC-I are decreased in AD transgenic mice compared to the wide-type mice, which agrees with previously published studies.³²

Figure 7.

PET imaging studies of ¹⁸F-2 in wild-type (25–30 g, female, n = 4) and 5× FAD transgenic mice (25–30 g, female, n = 4). (A) Representative PET sagittal images in wild-type and 5× FAD mice (summed 20–60 min). (B) TACs of the whole brain uptake of ¹⁸F-2 in wild-type and 5× FAD mice. All data are the mean. Data are expressed as the percentage of injected dose per cubic centimeter (% ID/cc).

CONCLUSIONS

In conclusion, further characterization of a recently identified MC-I modulator, ZCM-I-1, from our laboratory, using the CESTA supported its physical interaction with the mouse brain MC-I. This also encourages development of ZCM-I-1-based PET radiotracers as imaging probes to better understand the pathological roles of MC-I dysfunction in AD and other disorders and to support clinical development of MC-I-targeted therapeutics. Based on the SAR results of ZCM-I-1, a novel PET radiotracer, ¹⁸F-2, was synthesized and characterized in C57BL/6 mice by PET/CT imaging. The results demonstrated that ¹⁸F-2 exhibits a rapid and high uptake in the liver and kidney. In addition, a gradual increase of uptake was observed in the kidney with time. Furthermore, ¹⁸F-2 shows slow clearance in the brain, blood, heart, and bone. Further analysis of ¹⁸F-2 in the brain tissue revealed a rapid uptake without significant wash out during the scanning time. Higher binding signals of ¹⁸F-2 were noted in the brain stem, cerebellum, and midbrain regions. More importantly, the radioactivity of ¹⁸F-2 in the brain was significantly reduced upon pretreatment with its parent compound, hinting its specific binding in the brain. However, the kinetics of ¹⁸F-2 in the brain tissue and metabolic studies suggested the need of further optimization to develop suitable CNS PET radiotracers for MC-I. Further characterization of ¹⁸F-2 in transgenic AD mice suggested its potential to differentiate the binding in AD mouse brains from wild-type mouse brains. Taken together, the structure of ZCM-I-1 represents a novel chemical scaffold to develop suitable CNS MC-I PET radiotracers that will help image the biological processes of MC-I under pathological conditions.

MATERIALS AND METHODS

General Information.

All commercially available reagents were used without further purification unless otherwise stated. Analytical thin layer chromatography was performed using Silica Gel GF254 plates (Merck Millipore co. ltd, 0.2 mm thick). Compounds were purified using CombiFlash Rf 150 (Teledyne ISCO co. ltd). ¹H and ¹³C spectra were recorded on a Bruker 500 MHz. Chemical shifts in ¹H NMR spectra were reported in parts per million (ppm) on the δ scale from an internal standard of CDCl₃ (7.26 ppm). Data were reported as follows: chemical shift (δ ppm), multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet, and br = Broad), coupling constant in hertz (Hz), and integration. Chemical shifts of ¹³C NMR spectra were reported in ppm from the central peak of CDCl₃ (77.0 ppm) on the δ scale. MS data were recorded on an Agilent Technologies 6310 quadrupole mass spectrometer.

All animal studies were carried out at Massachusetts General Hospital (PHS Assurance of Compliance No. A3596-01). The Subcommittee on Research Animal Care (SRAC) serves as the Institutional Animal Care and Use Committee (IACUC) for the Massachusetts General Hospital. SRAC reviewed and approved all procedures detailed in this paper.

PET/CT imaging was performed in anesthetized (isoflurane) mice to minimize discomfort. Highly trained animal technicians monitored animal safety throughout all procedures, and veterinary staff were responsible for daily care. All mice were socially housed in cages appropriate for the physical and behavioral health of the individual animal and were given unlimited access to food and water, with additional nutritional supplements provided as prescribed by the attending veterinary staff.

Preparation of Compound 2.

Compound 2 was synthesized following our reported procedures.⁴⁰

5-(4-Fluorophenyl)-N-(2-(5-methoxy-1H-indol-3-yl)ethyl)-3-oxo-pentanamide (Compound 2): ¹H NMR (400 MHz, methanol-d₄) δ 7.21 (d, J = 8.78 Hz, 1H), 7.10–7.15 (m, 2H), 7.06 (d, J = 2.26 Hz, 1H), 7.04 (s, 1H), 6.91–6.96 (m, 2H), 6.76 (dd, J = 2.38, 8.66 Hz, 1H), 3.81 (s, 3H), 3.49 (t, J = 7.15 Hz, 2H), 2.91 (t, J = 7.15 Hz, 2H), 2.74–2.79 (m, 4H); ¹³C NMR (100 MHz, methanol-d₄) δ 205.8, 169.1, 164.1, 161.7, 155.2, 138.3, 133.6, 131.1, 131.1, 129.2, 124.5, 116.2, 115.9, 113.1, 113.1, 112.8, 101.6, 56.6, 45.3, 41.5, 29.6, 26.3. HRMS (m/z) (M + Na)⁻: calcd. For C₂₂H₂₃FN₂NaO₃ 405.1585, found 405.1588.

Cellular Thermal Shift Assay (CESTA).

Mitochondrial complex I was purified from C57BL/6 mice as described.⁴² Purified complex I (700 ng) was incubated with ZCM-I-1 or compound 2 for 1 h on ice in no BSA buffer (215 mM mannitol, 75 mM sucrose, 1 mM EGTA, and 20 mM HEPS in purified water, pH 7.4). The samples were then heated to the designated temperatures on a Bio-Rad T100 thermal cycler for 3 min, cooled to 25 °C for 3 min, and further cooled to 4 °C for 30 s. The samples were then centrifuged at 20,000 g at 4 °C for 20 min, resolved by SDS-PAGE, and probed with NDUFV1 antibody (AbCam, ab210558).

Microscale Thermophoresis Assay.

A Monolith NT. Automated instrument was used to measure the K_D value. A range of concentrations of compound 2 were incubated with purified MC-I (400 nM) from the bovine heart for 40 min in PBS-T assay buffer. The samples were loaded into NanoTemper glass capillaries, and a series of ligand-induced changes in thermophoresis curves was gained using an excitation power of 100% and a high MST power. K_D values were calculated using the mass action equation with NanoTemper software from duplicate reads of an experiment.

Radiosynthesis of ¹⁸F-2.

ZCM-I-1 (3.0 mg, 7.9 μmol) and CpRu(COD)Cl (7.3 mg, 23.7 μmol) were added to EtOH (0.05 mL) in a 2 mL borosilicate glass vial. The vial was capped, and the reaction mixture was stirred at 85 °C for 30 min. The vial was removed from the heating block and allowed to stand for 1 min at room temperature. To the vial, N,N-bis(2,6-diisopropyl)phenyl-2-chloroimidazolium chloride (10.9 mg, 23.7 μmol) and 0.15 mL of MeCN were added, and the resulting solution was drawn into a 1.0 mL polypropylene syringe.

Target water from the cyclotron containing ¹⁸F-fluoride was loaded with a syringe onto a solid-phase exchange (SPE) Chromafix 30-PS-HCO₃ cartridge. The cartridge was washed with MeCN (1.0 mL). The cartridge was inverted and fitted with a female × female Luer adapter. With the syringe, which contained the corresponding solution of ZCM-I-1, ruthenium complex, and N,N-bis(2,6-diisopropyl)phenyl-2-chloroimidazolium chloride, ¹⁸F-fluoride was eluted into an 8 mL borosilicate glass vial. The cartridge was washed with DMSO (0.15 mL) followed by DMSO/MeCN (0.05 mL, 1:1 (v/v)). The reaction vial, which contained 0.4 mL of the reaction mixture, was sealed with a Teflon-lined cap and was heated at 125 °C for 30 min. Then, the reaction mixture was allowed to stand for 1 min at room temperature.

The radioactive mixture containing ¹⁸F-2 was then applied to a reverse-phase semipreparative HPLC (Phenomenex Gemini-NX 5u C18 110A, 250 × 10 mm, 5.0 mL/min, a gradient of 10–90% CH₃CN in H₂O with 0.1% formic acid; 0–1 min, 10% CH₃CN; 1–10 min, 10 to 90% CH₃CN; and 10–15 min, 90% CH₃CN). A radioactive fraction having a retention time of 10.5 min was collected in a flask and diluted in water (20 mL). The final product was reformulated by loading onto an SPE C-18 cartridge (Waters WAT020515 Sep-Pak Plus Short C18), rinsing with water (4 × 5 mL), eluting with EtOH (0.3 mL), and diluting with saline (2.7 mL). The chemical and radiochemical purity of ¹⁸F-2 was tested by analytical HPLC (VARIAN Puruit XRs 5 C18, 150 × 4.6 mm) and elution with a gradient of 10–90% solvent B in solvent A (solvent A was 0.1% TFA in H₂O, and solvent B was CH₃CN), at a flow rate of 2 mL/min. Confirmation of the identity of ¹⁸F-2 was achieved by coinjection with compound 2 as the reference standard. For the determination of specific activity, mass (μmol) of ¹⁸F-2 with a known radioactivity was determined by HPLC comparison of UV absorbance at 254 nm with those of known concentrations of nonradioactive compound 2.

PET/CT Imaging Studies.

Mice were used for the baseline and blocking studies. All animals were anesthetized with inhalational isoflurane at 3% in a carrier of 2 L/min medical oxygen and maintained at 2% isoflurane for the duration of the scan. The mice were fixed on the bed of a Triumph Trimodality PET/CT scanner (Gamma Medica, Northridge, CA) in the prone position and injected with ¹⁸F-2 (150–200 μL; ~200 μCi) via a lateral tail vein catheterization at the start of PET acquisition. For blocking studies, compound 2 (5 mg/kg, iv) was injected at 10 min prior to ¹⁸F-2 injection. Dynamic PET acquisition lasted for 60 min and was followed by CT for anatomic coregistration. PET data were reconstructed using a 3D-MLEM method resulting in a full width at half-maximum resolution of 1 mm. Reconstructed images were exported from the scanner in the DICOM format along with an anatomic CT. These files were imported to PMOD software (PMOD Technologies LLC, version 4.0). Volumes of interests, including the whole brain, striatum, cortex, thalamus, midbrain, cerebellum, brain stem, and hippocampus, were placed referencing the MRI template software. TACs were exported in terms of decay-corrected activity at specified time points with gradually increasing intervals. TACs were expressed as the percentage of injected dose per cubic centimeter (% ID/cc).

Radiometabolism Studies.

After the intravenous injection of ¹⁸F-2, C57BL/6 mice (n = 4) were sacrificed at 30 min post-injection. Blood and whole brain samples were quickly removed. The blood samples were centrifuged at 4000 rpm for 4 min at 4 °C to separate the plasma. Then, the supernatant was collected in a small vial containing acetonitrile. The mixture was vortexed for 10 s and centrifuged at 4000 rpm for 4 min at 4 °C for deproteinization. The mouse brains were homogenized in a cold acetonitrile/water (1/1, v/v) solution. The homogenate was centrifuged at 4000 rpm for 4 min at 4 °C. Then, the supernatant was collected. An aliquot of the supernatant obtained from the plasma or brain homogenate was injected into the radio-HPLC system and analyzed using an Agilent Eclipse Plus 3.5 um C18 column (4.6 mm × 100 mm) at a flow rate of 2 mL/min in a gradient of 10–95% CH₃CN with 0.1% formic acid in H₂O with 0.1% formic acid (0–3 min, 10% CH₃CN with 0.1% formic acid; 3–10 min, 10 to 95% CH₃CN with 0.1% formic acid; 10–12 min, 95% CH₃CN with 0.1% formic acid). The retention time of ¹⁸F-2 is 8.8 min. The percentages of parent ¹⁸F-2 and the corresponding radiometabolites were calculated as [peak area for ¹⁸F-2 or corresponding radiometabolites]/[total peak area] × 100.

ABBREVIATIONS

AD

Alzheimer’s disease

ADNI

Alzheimer’s Disease Neuroimaging Initiative

ATP

adenosine triphosphate

AUC

area under the curve

CESTA

cellular thermal shift assay

CNS

central nervous system

ETC

electron transport chain

FMN

flavin mononucleotide

HPLC

high-performance liquid chromatography

MC-I

mitochondrial complex

MST

microscale thermophoresis

NADH

nicotinamide adenine dinucleotide

OXPHOS

oxidative phosphorylation

PET

positron emission tomography

ROS

reactive oxygen species

SAR

structure–activity relationship

TACs

time–activity curves

TLC

thin layer chromatography