|Home | About | Journals | Submit | Contact Us | Français|
L.W., R. C. and M. F. contributed equally to this work.
A suitable TSPO PET ligand may visualize and quantify neuroinflammation in living brain. Herein we report a 18F-ligand, [18F]2 ([18F]FDPA) is radiolabeled in high yield and high specific activity based on our spirocyclic iodonium ylide (SCIDY) strategy. [18F]2 demonstrated saturable specific binding to TSPO, substantially-elevated brain uptake and slow washout of bound PET signal in the preclinical models of brain neuroinflammation (cerebral ischemia and Alzheimer’s disease).
The translocator protein 18 kDa (TSPO) is a five trans-membrane domain protein and located with outer mitochondrial membrane. TSPO is expressed predominantly in steroid-synthesizing endocrine organs including heart, kidneys, adrenal cortex, testis or ovary.1 One of the most well-characterized functions of TSPO is its involvement in the translocation of cholesterol from the outer to the inner mitochondrial membrane,2 which is critical to the synthesis of steroids. In the central nervous system (CNS), TSPO is dramatically upregulated in neuroinflammatory conditions (activated microglial cells).3, 4 Abnormal overexpression of TSPO has been found as a pivotal pathological signal in the progression of many other neurological disorders. In particular, neuroinflammation occurs at the pre-plaque stage5 of Alzheimer’s disease (AD) and serves as a major driving force.6 Recent studies have also revealed that the TSPO expression was closely associated with infarction induced by ischemia.7 Therefore, TSPO is a potential biomarker for detecting inflammation associated with brain disorders, and for monitoring treatment response of anti-inflammatory therapies.
Positron emission tomography (PET) is a non-invasive imaging technology that is capable of visualizing and quantifying biochemical and pharmacological processes in vivo. A TSPO-specific PET ligand is utilized as a translational imaging tool to evaluate the status of neuroinflammation (activated microglia) in living brain.8, 9 [11C]PK11195 was the first extensively evaluated PET ligand for TSPO; however, low signal-to-noise ratio and high non-specific binding hampered its further advance.10 Intensive efforts therefore have been made to improve PET imaging quality and quantitative process with the development of a rich library of TSPO PET ligands.11–13 Particularly a PET ligand [18F]1 ([18F]DPA-714; Figure 1), developed by James and Kassiou et al, has shown improved bioavailability, lower nonspecific binding, and higher non-displaceable binding potential (BPND).14, 15 Initial evaluation of this radiopharmaceutical in healthy human volunteers validated its biodistribution profile towards TSPO.16 Clinical PET studies using [18F]1 were further conducted on patients with amyotrophic lateral sclerosis,17 post-stroke18 and AD,19 demonstrating this ligand may be useful in assessing the extent of neuroinflammation.
N,N-diethyl-2-(2-(4-[18F]fluorophenyl)-5,7-dimethylpyrazolo[1,5-a]pyrimidine-3-yl)-acetamide [18F]2 ([18F]FDPA), a fluoroaryl analog of [18F]1, in which the fluorine atom is directly attached to terminal aromatic ring, is another promising PET ligand (Figure 1). Its intrinsic structure (aryl-18F bond) offers an improved metabolic profile via substantially-reduced formation of radiolabeled metabolites in brain and plasma compared with [18F]1 as observed by Solin and Dollé et al.20, 21 Several radiofluorination reactions were reported to prepare [18F]2 via [18F]fluoride ion under various labeling conditions, including nitroarene, aryltrimethylammonium salt and diaryliodomium salt as precursors, that resulted in limited radiochemical conversions (< 3%).22 Electrophilic fluorination starting with carrier-added [18F]F2 to prepare [18F]Selectfluor as a labeling reagent was recently performed by Solin and co-workers, to prepare [18F]2, and resulted in 15 ± 3% decay corrected yields with a low specific activity of 7.8 ± 0.5 GBq/μmol (0.21 ± 0.01 Ci/μmol).23 However, because a low-cold-mass PET ligand injection is critical to detect inflammatory events,24, 25 there is an unmet and urgent need for [18F]2 with high specific activity to evaluate the potential of this radioligand in PET imaging of neuroinflammation.
On the basis of our previous work, precise radiofluorination of non-activated aromatics can be achieved with spirocyclic iodonium ylide (SCIDY) precursors.26–29 These bench-stable precursors provide efficient, regiospecific and direct radiolabeling via [18F]fluoride, demonstrating a broad substrate scope for non-activated and/or sterically-hindered (hetero)arenes. The operation is simple, metal-free and suitable for routine and automated production. Herein, we report an efficient radiosynthesis of [18F]2 based on SCIDY precursors from no-carrier added [18F]fluoride and preliminary PET imaging evaluation in preclinical models of ischemia and AD.
As shown in Scheme 1, we developed an efficient route and improved the literature procedure30 to obtain non-radioactive standard 10 (FDPA), and the key intermediate 11 for our SCIDY strategy. Microwave-assisted synthesis was utilized to assemble N,N-diethylamide 6 with dramatically-reduced reaction time (40 min) while maintaining a good yield (55%). The cyclization reaction triggered by hydrazine and acetic acid, led to fluoroaminopyrazole 8 in 88% yield. Finally, condensation between 8 and acetylacetone afforded 10 as a white crystalline solid in 75% yield. The key compound 11 was also prepared in an analogous synthetic route with an overall yield of 41% in three steps from 3-(4-iodophenyl)-3-oxopropanenitrile 4. We found that oxidation of 11 by using Oxone in mixed solvents of CHCl3 and trifluoroacetic acid was critical for this type of transformation, leading to the corresponding iodo(III) intermediate 12 in ca. 60% conversion. The intermediate was used without further purification in the subsequent coupling reactions with spirocyclic auxiliaries (SPI5 and SPIAd)29 to afford SCIDY precursors 13 (22%, two steps from 11) and 14 (34%, two steps from 11), respectively.
In vitro binding affinity of 10 for TSPO was measured from competitive binding of TSPO-selective radioligand [3H]PK11195 (Figure S1). 10 showed excellent binding affinity (Ki: 2.0 ± 0.8 nM) which is consistent with prior reports31 and comparable with that of DPA-714 (Ki: 7.0 ± 0.4 nM).14 The selectivity of 10 towards central benzodiazepine receptor (CBR) binding was reported as Ki > 1 mM,22 in comparison to that of DPA-714 (Ki > 10 μM).14 These results indicated that [18F]2 could be utilized to study in vivo binding profiles for TSPO by PET.
Based on our prior reports of utilizing SCIDY precursors to enable radiofluorination on non-activated arenes from no-carrier added [18F]fluoride,26, 29 we developed an efficient radiolabeling method from SCIDY precursor 14 to produce [18F]2. Specifically, we carried out comprehensive optimization of radiofluorination conditions, including bases, solvents, reaction temperatures and precursor loadings (Figure 2). We first set out the radiofluorination with 14 using different types and loading amounts of bases, including tetraethylammonium bicarbonate (TEAB) and K2CO3/K222 in DMF. Unfortunately, under these conditions only trace amounts of [18F]2 were detected. These fruitless results promoted us to study the thermostability of SCIDY precursors. While compound 14 alone showed no decomposition when heating at 120 °C for 10 min, the addition of TEAB (1 mg) led to a rapid decomposition even without heating, which indicates alternative milder bases/additives may be necessary to achieve high radiolabeling efficiency (Figure S2). To our delight, combination of K2C2O4/K222 gave the desired product [18F]2 with >20% radiochemical conversion (RCC). Furthermore two quaternary alkylammonium salts, namely, tetrabutylammonium methanesulfonate (TBAOMs)32 and tetraethylammonium perchlorate (TEAOCl4) were identified as the most promising additives (> 30% RCCs) for further evaluation (Figure 2A & Figure S3). By optimizing reaction time and temperature we found that TBAOMs was superior to TEAOCl4, and RCCs reached to 40 ± 3% (n = 3) at 120 °C for 15 min (Figure 2B & 2C). Together with solvent effect (Figure 2D) and the amounts of precursor loading (Figure S4), we identified that the combination of 12 mg of TBAOMs, 1 mg of SCIDY precursor 14 in CH3CN at 120 °C for 15 min provided the highest RCC of 63 ± 5% (n = 6). In addition, we also tested SPI5-based SCIDY precursor 13 under the optimal conditions (Figure S5), which only provided 16 ± 3% RCC (n = 3). These results were consistent with our prior observation that SPIAd precursors were more efficient than SPI5 precursors under 18F-labeling reactions.29 As a result, the radiosynthesis of [18F]2 was achieved in 45 ± 8% decay-corrected yield within 80 min with >99% radiochemical and chemical purity and specific activity of 96 ± 22 GBq/μmol (2.6 ± 0.6 Ci/μmol; see experimental section). The identity of [18F]2 was confirmed by co-injection with 10 (Figure S6). No radiolysis was observed up to 120 min after formulation. [18F]2 was stable in vitro in mouse serum at 37 °C for 120 min. The efficient radiosynthesis with high specific activity, radiochemical purity and stability enabled the subsequent ex vivo biodistribution and preclinical in vivo PET imaging studies.
Lipophilicity can be used, in part, as a predictive value for assessing blood brain barrier (BBB) permeability with an optimal range of 1.0–3.5.33 Using liquid-liquid partition (“the shake flask method”), LogD7.4 value of [18F]2 was determined to be 2.34 ± 0.05 (n = 5), which is comparable with several reported TSPO radioligands such as [18F]DPA-714 ([18F]1; LogD7.4 2.44)14 and [18F]GE-180 (LogD7.4 2.95).31
The uptake, distribution and clearance of [18F]2 were studied in mice at three time points (2, 25 and 45 min) post injection (Figure 3, Table S1 and S2). At 2 min p.i., high uptake (>5 %ID/g) was observed in the lungs, heart, liver, kidneys, pancreas and adrenal glands. After the initial phase the radioactivity levels in most tissues decreased rapidly, while the signals in the kidney, spleen and adrenal glands continued to increase during the whole period. The radioligand was efficiently cleared from blood (2 min/45 min ratio >10), and high uptake of [18F]2 in the liver, kidney and small intestine indicated urinary and hepatobiliary excretion, as well as possible intestinal re-uptake. The uptake of [18F]2 in the present study is consistent with the distribution of TSPO in mice, as high expression in the lungs, heart, kidneys and adrenal glands has been previously observed.1 In particular, since TSPO expression in microglia cells of normal brain has been found to be in a low level,34 brain uptake of [18F]2 was moderate with 3.69 %ID/g at 2 min p.i., and the radioactivity washout was reasonable with 1.15 %ID/g at 45 min time point (2 min/45 min ratio 3.2). In addition, negligible bone uptake was detected (<1 %ID/g), showing little or no defluorination occurred in vivo, which was also confirmed by PET imaging studies (Figure S7). Based on these results, we further evaluated if [18F]2 is a suitable PET ligand targeting TSPO in preclinical models of neuroinflammatory brain.
Due to the low TSPO expression in normal brain,34 we utilized a focal cerebral ischemic rat model (see section 10 in SI) in which TSPO expression was up-regulated with activated microglia35 while the blood-brain barrier was intact or not seriously disrupted.36 As shown in Figure 4, PET images and time-activity curves of [18F]2 showed maximum uptake in the ipsilateral side (ischemic site) reached a peak of 1.20 SUV at 10 min. Blockade with PK11195 (3 mg/kg) decreased PET signal by ca. 80%, indicating high in vivo specificity and low non-specific binding. The SUV(2/40min) ratio of 2.9 showed a rapid washout rate of non-specific binding under blocking conditions. In the contralateral side, [18F]2 showed limited peak brain uptake (0.54 SUV) and marginal changes between baseline and blocking conditions. We also observed a rapid clearance of non-specific binding (SUV2/40min = 5.9) probably attributed to low expression of TSPO in normal brain.34 High ipsilateral to contralateral ratios at 10 min and 90 min p.i. were determined as 5.41 ± 0.12 and 4.26 ± 0.34, respectively. Pretreatment with PK11195 also substantially diminished heterogeneous radioactive signals between ipsilateral and contralateral sides.
We also carried out a simplified reference-tissue model (SRTM) analysis and determined BPND using the contralateral side as the reference region.36 The BPND values in the ipsilateral sides were determined to be 4.02 ± 1.32 (p < 0.05) for [18F]2, which was higher than [11C]PK11195 (1.59 ± 0.33, p < 0.05).36
Liu et al. demonstrated that TSPO was mainly expressed in activated microglia in APP/PS1 transgenic mouse brains by immunofluorescence studies between TSPO and Aβ as well as the association of Aβ deposition and TSPO-positive microglia in human AD brain sections.37 Thus as proof of concept, we explored in vivo binding profile of [18F]2 in a transgenic mouse model (APP/PS1) of AD. [18F]2 rapidly crossed the BBB and increased to 1.50 ± 0.13 SUV at 3 min p.i., demonstrating 1.6-fold higher peak brain uptake and slow washout (possibly attributed to increased specific and non-specific binding) in APP/PS1 brain compared with age-matched control (Figure 5).
The binding specificity of [18F]2 to TSPO was further confirmed by autoradiography studies in ischemia and AD brains. The ipsilateral side of ischemic rat brain showed substantially increased signal levels than that of contralateral side (ratio 7:1; Figure 6). The difference in radioactivity levels between the ipsilateral and contralateral sides was diminished by co-incubation with PK11195 (Figure 6B), leaing to a dramatic reduction (>98%) of ipsilateral binding (Figure 6C). We also noted increased whole brain uptake in APP/PS1 mouse brain slices compared with age-matched controls (Figure S8). These in vitro findings are consistent with the in vivo binding specificity of [18F]2 by PET studies.
The SCIDY-based methodology provides a high-yield and high-specific-activity radiosynthesis of [18F]2 that enabled preliminary PET studies in preclinical models of focal cerebral ischemia and Alzheimer’s disease. These proof-of-concept results not only warrant [18F]2 as a promising imaging biomarker to elucidate the involvement of microglia activity in preclinical models of ischemia and AD, but also provide the basis for designing new pyrazolo[1,5-a]pyrimidine based TSPO PET ligands. Further validation including immunofluorescent staining, radiometabolite analysis, PET imaging studies in higher species among different polymorphism are outlined to evaluate the suitability of [18F]2 as a potential imaging tool for clinical translation.
To a solution of 11 (50 mg, 0.11 mmol) in trifluoroacetic acid (0.39 mL) and chloroform (0.13 mL) was added Oxone (100mg, 0.165 mmol) under stirring. The mixture was stirred at room temperature for 60 min, then concentrated in vacuo. To the crude was added EtOH (0.8 mL), then a solution of SPIAd (25.3 mg, 0.11 mmol) in 10% Na2CO3 (0.5 mL), followed by additional 10% Na2CO3 (0.3 mL) to adjust to pH 9. The mixture was stirred at room temperature for 70 min, then diluted with H2O (5 mL), extracted with CH2Cl2 (5 mL × 3). The combined organic layers were dried over MgSO4 and concentrated in vacuo. The residue was purified by column chromatography on silica gel (0–10% MeOH in EtOAc) to afford ylide precursor 14 as a white solid (26.1 mg, 34% yield).
[18F]Fluoride solution (0.74–1.48 GBq; 20–40 mCi) was delivered and trapped on a Waters Sep-Pak light QMA cartridge (pretreated with 1 mL of 8.3% aqueous NaHCO3 and 20 mL H2O) by nitrogen gas. A solution of TBAOMs (12.0 mg) in CH3CN/H2O (v/v, 1/1, 1.0 mL) was used to elute [18F]fluoride into a V-shaped vial sealed with a Teflon-lined septum. The vial was heated to 110 °C while nitrogen gas was passed through a P2O5-Drierite™ column into the vented vial. When no liquid was visible, it was removed from heat, anhydrous CH3CN (1 mL) was added, then the heating was resumed until dryness. This step was repeated three times. The vial was then cooled to room temperature under nitrogen flow. A solution of precursor 14 (1.0 mg) in anhydrous CH3CN (1 mL) was added into the reaction vial and heated at 120 °C for 12 min. CH3CN/H2O (v/v, 1/1, 2.0 mL) was added to the mixture, which was then injected to a semi-preparative HPLC system. HPLC purification was completed on a semi-preparative Phenomenex Luna C18 column (250 × 10 mm, 5 μm) eluting with CH3CN/0.1 M NH4•HCO2(aq) (v/v, 1/1) at a flowrate of 5 mL/min. The retention time for [18F]2 was 17.5 min. The radioactive fraction corresponding to the desired product was collected and diluted with 23 mL H2O. The mixture was then loaded onto a C18 SPE cartridge (pre-activated with 5 mL EtOH followed by 10 mL of H2O). The cartridge was washed with 20 mL of H2O and dried by nitrogen gas for 5 min. The product [18F]2 was then eluted with 1.5 mL of CH3OH, and collected in a sterile flask, evaporated to dryness by nitrogen gas at 60 °C for 20 min, and reformulated in a saline solution (3 mL) containing 100 μL of 25% ascorbic acid in H2O and 100 μL of 20% Tween® 80 in ethanol. The synthesis time was ca. 80 min from end-of-bombardment. Radiochemical and chemical purity were measured by analytical HPLC (Phenomenex Luna C18, 250 × 4.6 mm, 5 μm, UV at 254 nm; CH3CN/0.1 M NH4•HCO2(aq) (v/v, 7/3) at a flowrate of 1.0 mL/min). The identity of [18F]2 was confirmed by the co-injection with unlabeled 10. Radiochemical yield was 45 ± 8% (decay-corrected) with >99% radiochemical purity and 96 ± 22 GBq/μmol (2.6 ± 0.6 Ci/μmol) specific activity.
We thank the National Institute of Mental Health’s Psychoactive Drug Screening Program for the compound screening. R.C. is supported by China Scholarship Council (201506250036). S.H.L. is a recipient of an NIH career development award (DA038000).