Radiosynthesis and Preclinical Evaluation of an 18F-Labeled Triazolopyridopyrazine-Based Inhibitor for Neuroimaging of the Phosphodiesterase 2A (PDE2A)

The cyclic nucleotide phosphodiesterase 2A is an intracellular enzyme which hydrolyzes the secondary messengers cAMP and cGMP and therefore plays an important role in signaling cascades. A high expression in distinct brain areas as well as in cancer cells makes PDE2A an interesting therapeutic and diagnostic target for neurodegenerative and neuropsychiatric diseases as well as for cancer. Aiming at specific imaging of this enzyme in the brain with positron emission tomography (PET), a new triazolopyridopyrazine-based derivative (11) was identified as a potent PDE2A inhibitor (IC50, PDE2A = 1.99 nM; IC50, PDE10A ~2000 nM) and has been radiofluorinated for biological evaluation. In vitro autoradiographic studies revealed that [18F]11 binds with high affinity and excellent specificity towards PDE2A in the rat brain. For the PDE2A-rich region nucleus caudate and putamen an apparent KD value of 0.24 nM and an apparent Bmax value of 16 pmol/mg protein were estimated. In vivo PET-MR studies in rats showed a moderate brain uptake of [18F]11 with a highest standardized uptake value (SUV) of 0.97. However, no considerable enrichment in PDE2A-specific regions in comparison to a reference region was detectable (SUVcaudate putamen = 0.51 vs. SUVcerebellum = 0.40 at 15 min p.i.). Furthermore, metabolism studies revealed a considerable uptake of radiometabolites of [18F]11 in the brain (66% parent fraction at 30 min p.i.). Altogether, despite the low specificity and the blood–brain barrier crossing of radiometabolites observed in vivo, [18F]11 is a valuable imaging probe for the in vitro investigation of PDE2A in the brain and has potential as a lead compound for further development of a PDE2A-specific PET ligand for neuroimaging.


Introduction
The 3 ,5 -cyclic nucleotide phosphodiesterase 2A (PDE2A) belongs to a superfamily of highly conserved enzymes (PDE1-11). These omnipresent PDEs are the only enzymes degrading the secondary messengers cAMP and cGMP, thereby terminating the intracellular signalling pathways addressed by G-protein coupled receptors (GPCRs) as well as by natriuretic peptides and nitric oxide, respectively. The high relevance of cyclic nucleotide signalling for physiological processes makes PDEs in general a choice for exploration as therapeutic targets for the treatment of numerous pathophysiological conditions. Furthermore, because of isoform specific expression profiles, PDEs are also considered biomarkers for certain diseases of, e.g., the nervous, cardiovascular, and immune systems as well as of cancer [1].

Organic Syntheses of the New Inhibitor 11 and the Precursor Compounds 13 and 19
Starting from the esters 1 and 2, the hydroxybutyl sidechain was introduced by metalation with isopropylmagnesium chloride lithium chloride complex (iPrMgCl · LiCl), followed by a Grignard reaction with butanal using a modified procedure from Giovannini et al. [20] (Scheme 1A). Subsequently, the hydroxy group was protected as 2tetrahydropyranyl ether (2-THP) and the resulting esters 5 and 6 were converted into the benzohydrazides 7 and 8. Cyclization with 3-chloro-6-cyclopropyl-2-methylpyrido [2,3b]pyrazine [20] was accomplished by the use of 1,8-diazabicyclo [5.4.0]undec-7-en (DBU) as a base for hydrazide activation [25] followed by the addition of PPh 3 /C 2 Cl 6 as a reagent for dehydration and ring closure to the triazole [26] in a one-pot reaction at room temperature. Deprotection of the resulting fluoro-derivative 9 with pyridinium p-toluenesulfonate (PPTS) yielded the new inhibitor 11. Oxidation of the secondary alcohol with Dess-Martin periodinane (DMP) produced compound 12 which was needed as an intermediate reference for the radiosynthesis. Miyaura borylation [27] of the bromo derivative 10 yielded the boronic acid pinacol ester 13, one of the two different radiosynthesis precursor compounds. The synthesis of the other precursor started with the oxidation of 14 [28] to the corresponding nitrosubstituted compound 15 using a modified procedure from Hahm et al. [29] (Scheme 1B). Cyclization to the triazolo compound 18 was accomplished after the protection of the ketone as the ketal 16 and conversion of the ester to the hydrazide 17. The final deprotection with p-toluenesulfonic acid (TsOH) yielded the nitro precursor 19.

Radiosyntheses
As depicted in Scheme 2, two strategies have been drafted for the generation of [ 18 F]11. While with strategy A a copper-mediated radiofluorination (CMRF) was envisaged, strategy B is based on a common nucleophilic aromatic substitution reaction using a carbonyl function in para-position for activation of the labeling position. With the deprotection (strategy A) and reduction (strategy B), respectively, a second reaction step was necessary for both procedures.

Copper-Mediated Radiofluorination Experiments-Strategy A
The attempts to synthesize the intermediate [ 18 F]9 by CMRF have been performed with a boronic acid pinacol ester (13) as a precursor (Scheme 2) under different reaction conditions adopted from the literature [30][31][32][33][34][35] (see Table 1). In most of the reported procedures rather high amounts of precursor (30-60 µmol) are used which is however sometimes not practicable as for example in our study due to an elaborated organic precursor synthesis. Generally, for the here described experiments only 2 mg (3.4 µmol) of the protected precursor 13 were used and the labeling reactions were conducted at temperatures ranging from 110-140 • C with [Cu(OTf) 2 (py) 4 ] as copper catalyst. For monitoring of the labeling progress, aliquots were taken every 5 min (total reaction time 20 min) and analysed by radio-TLC. In the first labeling experiments, [ 18 F]tetra-n-butylammonium fluoride ([ 18 F]TBAF, generated from [ 18 F]fluoride and (n-Bu) 4 NHCO 3 under azeotropic drying) in N,N-dimethylacetamide (DMA)/tert-butanol (tert-BuOH) with a precursor-to-copper catalyst ratio of 1:3.5 (entry 1 in Table 1) was chosen as these conditions have been shown to be efficient for the labeling of several radiotracers in our lab [36,37]. However, an RCY of only 0.6% could be achieved which prompted us to test n-butanol as probably a more efficient alcohol additive [32] without achieving an improvement. Furthermore, the reduction of the content of [Cu(OTf) 2 (py) 4 ] to a ratio of precursor-to-copper catalyst of 1:1 did not increase the fraction of labeling product (entry 2 in Table 1). Moreover, the use of a higher amount of precursor (8.5 µmol, entry 4) did not give higher labeling yields. In order to exclude a methodical failure of the labeling procedures, a standard labeling system was established using 4-biphenylboronic acid pinacol ester and [Cu(OTf) 2 (py) 4 ]. As shown in the right column of Table 1, the synthesized 4-[ 18 F]fluorobiphenyl could be produced in sufficient amounts under the selected reaction conditions.
In the next step, the conventional potassium carbonate-kryptofix-222 system [ 18 F]F − / K 222 /K 2 CO 3 (generated by elution of [ 18 F]fluoride from an anion-exchange cartridge with K 2 CO 3 , addition of K 222 and azeotropic drying) was used as fluorination agent with N,Ndimethylformamide (DMF) as solvent. However, no product formation could be observed at precursor-to-copper catalyst ratios of 1:1.5 and 5:1 (entries 5 and 6).
As already discussed by the group of Gouverneur, late-stage 18 F-incorporation by CMRF can bear risks, in particular when the compound to be radiolabeled contains Nheterocycles [39,40]. Thus, the 18 F-fluorination might be inhibited due to a complexation of copper by the lewis base nitrogen atoms generating non-reactive copper species. To investigate the potential influence of the N-heterotricyclic triazolopyridopyrazine core on the labeling process, we performed a test reaction with the standard labeling system (4-biphenylboronic acid pinacol ester) under the addition of an equimolar amount of 12. The obtained RCY of 60% indicates, that the N-heterocycles do not prevent the formation of reactive copper-fluorination species. Based on this result, we assume that the unsuccessful radiofluorination of [ 18 F]9 might be caused by a sterical hindrance due to the bulky tricyclic moiety substituted in ortho-position to the also space-consuming boronic acid pinacol ester leaving group of the phenyl ring and thus preventing the attack of the Cu(II)-18 F-complex.

Aromatic Nucleophilic Substitution Reactions-Strategy B
The aromatic nucleophilic substitution reaction of the nitro precursor 19 (Scheme 2) was first investigated with the conventional [ 18 F]F − /K 222 /K 2 CO 3 system using different solvents and temperatures (entries 1-3 in Table 2). However, neither the use of acetonitrile (ACN) at 110 • C nor DMF or DMSO at 150 • C resulted in the formation of the intermediate Inspection of samples of the reaction mixtures with UV-HPLC revealed the entire decomposition of the precursor in all three solvents already after 5 min reaction time. Therefore, we assumed that the precursor might be sensitive to the basic labeling conditions and tested [ 18 F]TBAF as a milder nucleophilic 18 F-fluorination agent. Labeling reactions were performed in ACN, ACN/DMSO, tert-butanol and DMF (entries 4-7 in Table 2). While the reaction in tert-butanol did not result in the formation of [ 18 F]12, maximum RCYs of 21% could be achieved with ACN. In contrast to the [ 18 F]F − /K 222 /K 2 CO 3 system, intact precursor could be detected in all solvents even after 20 min labeling time. All reactions were performed with small precursor amounts of about 1-2 mg. To reduce the keton function of [ 18 F]12 in the second reaction step, sodium borhydride (NaBH 4 , 2M solution) with n-butanol and lithium aluminum hydride (LiAlH 4 , 1M solution) have been tested as reducing agents by direct addition to the labeling reaction mixture at 35 • C. The reduction was completed already after 5 min reaction time; however, with LiAlH 4 the formation of a slightly more hydrophilic radioactive byproduct has been observed accounting for about the same amount as the desired product.
Finally, for this two-step one-pot radiolabeling procedure on the basis of the nitro precursor 19, highest RCYs for the radiofluorination step were achieved with [ 18 F]TBAF in ACN at 110 • C after a reaction time of 15 min. For the second step, the reduction system NaBH 4 /n-butanol at 35 • C turned out to be most useful for a quantitative conversion of the intermediate [ 18

Automated Radiosynthesis of [ 18 F]11
The developed reaction conditions were translated to an automated procedure using the TRACERlab FX2 N synthesis module (GE Healthcare). The synthesizer setup is described in the experimental part. In brief, after trapping and elution of [ 18 F]fluoride by a solution of tetra-n-butylammonium hydrogencarbonate from an anion exchange cartridge, the labeling reaction of the azeotropically dried [ 18 F]TBAF with the nitro precursor 19 was performed in ACN at 110 • C for 15 min. For the subsequent reduction step, sodium borhydride in n-butanol was added and the reaction proceeded at 35 • C for 5 min. To isolate [ 18 F]11, the crude reaction mixture was diluted with a solution of aqueous 20 mM ammonium formate, acetonitrile and water and directly applied to a semi-preparative radio-HPLC system. The radiotracer fraction was collected at a retention time of 28-30 min and reformulated by solid phase extraction (SPE) using a C18-cartridge. The long retention time was needed in order to separate the radiotracer from the remaining nitro precursor, whose keto function was also reduced and therefore eluted close in front of the radiotracer ( Figure S12 in Supplementary Materials). The obtained radiotracer eluate was transferred out of the hot cell, concentrated and formulated in sterile isotonic saline. Consequently, the formulated product contains 10% of ethanol and had a radiochemical concentration of about 1 MBq/µL. In a total synthesis time of~80 min, [ 18 F]11 could be produced with a high radiochemical purity of ≥99%, a radiochemical yield of 2.2 ± 0.7% (n = 8) and molar activities in the range of 11-20 GBq/µmol (n = 4, end of synthesis EOS) with starting activities of 6-8 GBq. The identity of [ 18 F]11 has been confirmed by analytical radio-HPLC of the final product spiked with the reference compound 11 ( Figure S13 in Supplementary Materials).
In order to estimate the lipophilicity of [ 18 F]11, the partition coefficient was determined by the shake flask method using n-octanol and phosphate-buffered saline as partition system. With this method a logD value of 1.79 ± 0.66 (n = 4) could be determined. The chemical stability of the radiotracer was investigated by incubation in phosphate-buffered saline (PBS) and n-octanol at 30 • C. [ 18 F]11 proved to be stable in these media and inspection by radio-HPLC revealed no defluorination or degradation within 60 min of incubation time.  11 in horizontal brain sections obtained from female SPRD and male Wistar rats. High levels of binding are seen in distinct layers of the somatosensory cortex, in different hippocampal regions and in the basal ganglia (Figure 1), the only brain structures with high PDE2A immunoreactivity in rat [41]. For the other brain regions and in particular the cerebellum, all characterized by much lower PDE2A expression [41], much weaker binding of [ 18 11 in horizontal brain sections obtained from female SPRD and male Wistar rats. High levels of binding are seen in distinct layers of the somatosensory cortex, in different hippocampal regions and in the basal ganglia (Figure 1), the only brain structures with high PDE2A immunoreactivity in rat [41]. For the other brain regions and in particular the cerebellum, all characterized by much lower PDE2A expression [41], much weaker binding of [ 18 F]11 was detectable.    11 in rat brain in vitro. Representative autoradiogram of a 12 µm cryosection of one female SPRD rat incubated with 9.5 nM [ 18 F]11. The regions Co = somatosensory cortex (layers denoted by Roman numerals), CPu = caudate putamen, and Hip = hippocampus (Cornu Ammonis subfields indicated by CA1-3) are presented separately at a higher magnification. Further abbreviations: Cb = cerebellum.  As illustrated in Figure 2, the saturability of the bindings sites of [ 18 F]11 has been proven by co-incubation with an excess of non-labeled 11. Furthermore, the binding of [ 18 F]11 was completely blocked by the PDE2A-specific inhibitors BAY60-7550 [15] and PF-05180999 [6]. Altogether, these in vitro findings indicate virtually exclusive and high specific binding of [ 18 F]11 towards rat PDE2A.

Biological Evaluation
To further characterize the binding site of [ 18 F]11 in PDE2A expressing regions of the rat brain, we performed quantitative in vitro autoradiography with co-incubation of rat brain cryosections with a fixed concentration of As illustrated in Figure 2, the saturability of the bindings sites of [ 18 F]11 has been proven by co-incubation with an excess of non-labeled 11. Furthermore, the binding of [ 18 F]11 was completely blocked by the PDE2A-specific inhibitors BAY60-7550 [15] and PF-05180999 [6]. Altogether, these in vitro findings indicate virtually exclusive and high specific binding of [ 18 F]11 towards rat PDE2A.
To further characterize the binding site of [ 18 F]11 in PDE2A expressing regions of the rat brain, we performed quantitative in vitro autoradiography with co-incubation of rat brain cryosections with a fixed concentration of     11 in rat brain by homologous competition. Representative autoradiograms of 12 µm cryosections of one female SPRD rat incubated with 6.8 nM [ 18 F]11 alone (total binding) or in the presence of different concentrations of 11. Non-specific binding of the radioligand has been determined by co-incubation with 1 µM BAY60-7550. The grey scale corresponds to values of 1100-11,327 quantum level/pixel and is valid for all autoradiograms. Abbreviations: Cb = cerebellum, Co = cortex, CPu = caudate putamen, and Hip = hippocampus. Altogether, the autoradiographic studies indicated that [ 18 F]11 binds with high affinity and specificity towards PDE2A in the brain. A binding pattern almost identical to the Altogether, the autoradiographic studies indicated that [ 18 F]11 binds with high affinity and specificity towards PDE2A in the brain. A binding pattern almost identical to the one reported for the tritiated version of the PDE2A-specific PF-05270430 in rat brain [42] as well as to the immunohistochemically determined distribution of PDE2A in rat [41] along with an affinity and an inhibitory potency towards PDE2A in the low-nanomolar range justified the investigation of the potential of [ 18 F]11 for the imaging of PDE2A in the brain in vivo.

Metabolism of [ 18 F]11 in CD-1 Mice and SPRD Rats
To support the interpretation of the following pharmacokinetic studies, the metabolism of [ 18 F]11 was evaluated in both mice and rats. Non-anesthetized female CD-1 mice (n = 2) and one anesthetized female SPRD rat received a single i.v. injection of [ 18 F]11 (~30 MBq) via the tail vein and a blood sample was obtained at 30 min p.i. Immediately afterwards, the animals were sacrificed for sampling of the brains.
After deproteinization by treating the plasma and brain samples with a mixture of ACN/H 2 O and extraction of the activity (recoveries ≥ 91%), aliquots were analyzed by radio-HPLC. In rat, the fraction of non-metabolized [ 18 F]11 in the plasma and brain accounted for about 50% and 66%, respectively. In the corresponding radio-chromatograms of the rat plasma samples ( Figure 5A) several radiometabolites were detected eluting in front of the radiotracer and thus assumed to be more polar than [ 18 F]11. One single radiometabolite demonstrates an apparently lipophilic character because it elutes after [ 18 F]11. For the brain ( Figure 5B), a comparable elution pattern is observed in principle, however, with lower intensities of the more polar radiometabolites. Regarding the metabolism of [ 18 F]11 in mouse, no considerable differences could be found in the peak pattern except in the intensities ( Figure 5C,D). Thus, in mice the radiotracer is faster degraded resulting in fractions of non-metabolized [ 18 F]11 in plasma and brain accounting for about 7/11% and 42/50% (n = 2) of the total activity at 30 min p.i. ometabolite demonstrates an apparently lipophilic character because it elutes after [ 18 F]11. For the brain ( Figure 5B), a comparable elution pattern is observed in principle, however, with lower intensities of the more polar radiometabolites. Regarding the metabolism of [ 18 F]11 in mouse, no considerable differences could be found in the peak pattern except in the intensities ( Figure 5C,D). Thus, in mice the radiotracer is faster degraded resulting in fractions of non-metabolized [ 18 F]11 in plasma and brain accounting for about 7/11% and 42/50% (n = 2) of the total activity at 30 min p.i.

Biodistribution and Central Nervous System Penetration Study in CD-1 Mice and SPRD Rats
The biodistribution of [ 18 F]11 was further investigated by a dynamic PET scan in one female CD-1 mouse. A moderate brain uptake of [ 18 F]11 is indicated by a peak value of the time-activity curve (TAC) of the brain of SUV = 0.55 at 2.8 min p.i. suggesting a limited passage of the radiotracer through the blood-brain barrier. According to the immunohistological data [41] and the autoradiographic results, the brain regions striatum, cortex and hippocampus were defined as PDE2A-rich regions. However, we observed no significant differences in the uptake of activity in these regions in comparison to the PDE2A-poor reference region cerebellum as illustrated in Figure 6A. Similar TAC peak values (0.71; 0.47; 0.70 at 2.8 min vs. 0.61 at 0.8 min) and washout kinetics indicate poor specificity of [ 18 F]11 in vivo ( Figure 6A,B). The high percentage of radiometabolites in the brain (~50% of the activity at 30 min post injection) is probably the main confounding factor hindering the PET imaging with [ 18 F]11.
Evaluation of the whole-body distribution derived from PET imaging (Figure S14 and Table S1 in Supplementary Materials) revealed a low accumulation of activity in the bladder, spleen, muscle, gallbladder, kidney, and liver (SUV values < 5 at all time points investigated), indicating low renal or hepato-biliary excretion of [ 18 F]11. An intestinal excretion instead is strongly suggested from the high and continuosly increasing accumulation of activity in the small intestine (maximum at 60 min p.i. with SUV~35). Altogether the biodistribution of activity after i.v. administration does not recapitulate the expression pattern of PDE2A [43].
To assess potential species differences in the pharmacokinetics of [ 18 F]11, the tracer was administered intravenously in one female SPRD rat for investigation by dynamic PET. Figure 7A illustrates a higher but still moderate brain uptake of [ 18 F]11 in the rat in comparison to the mouse confirmed by a peak value of 0.97 at 0.6 min ( Figure 7B). The PDE2A-rich regions caudate nucleus and putamen, cortex and hippocampus present a slightly higher uptake over time than the PDE2A-poor region cerebellum (signal-tobackground ratio between 1.1 and 1.4) insufficient for imaging purposes as observed on the PET images ( Figure 7A). The non-negligible fraction of 34% brain penetrant radiometabo-lites at 30 min p.i. probably contributes to high background noise, hindering the detection of a potential specific binding of [ 18 F]11 in vivo. Furthermore, the signal-to-background ratio in vivo under dynamic conditions might per se be lower than anticipated from the autoradiographic experiments performed under equilibrium conditions in vitro (caudate putamen-to-cerebellum ratio of about 2) [44].
The biodistribution of [ 18 F]11 was further investigated by a dynamic PET scan in one female CD-1 mouse. A moderate brain uptake of [ 18 F]11 is indicated by a peak value of the time-activity curve (TAC) of the brain of SUV = 0.55 at 2.8 min p.i. suggesting a limited passage of the radiotracer through the blood-brain barrier. According to the immunohistological data [41] and the autoradiographic results, the brain regions striatum, cortex and hippocampus were defined as PDE2A-rich regions. However, we observed no significant differences in the uptake of activity in these regions in comparison to the PDE2Apoor reference region cerebellum as illustrated in Figure 6A. Similar TAC peak values (0.71; 0.47; 0.70 at 2.8 min vs. 0.61 at 0.8 min) and washout kinetics indicate poor specificity of [ 18 F]11 in vivo ( Figure 6A,B). The high percentage of radiometabolites in the brain (~50% of the activity at 30 min post injection) is probably the main confounding factor hindering the PET imaging with [ 18 F]11. Evaluation of the whole-body distribution derived from PET imaging ( Figure S14 and Table S1 in Supplementary Materials) revealed a low accumulation of activity in the bladder, spleen, muscle, gallbladder, kidney, and liver (SUV values < 5 at all time points investigated), indicating low renal or hepato-biliary excretion of [ 18 F]11. An intestinal excretion instead is strongly suggested from the high and continuosly increasing accumulation of activity in the small intestine (maximum at 60 min p.i. with SUV~35). Altogether the biodistribution of activity after i.v. administration does not recapitulate the expression pattern of PDE2A [43].
To assess potential species differences in the pharmacokinetics of [ 18 F]11, the tracer was administered intravenously in one female SPRD rat for investigation by dynamic PET. Figure 7A illustrates a higher but still moderate brain uptake of [ 18 F]11 in the rat in comparison to the mouse confirmed by a peak value of 0.97 at 0.6 min ( Figure 7B). The These results demonstrated (i) an average brain uptake of [ 18 F]11 in mice and rats with (ii) no significant uptake in PDE2A-expressing regions and (iii) a contribution of radiometabolites to the measured PET signals due to the passage of a non-negligible fraction of radiometabolites through the blood-brain barrier in both species despite a slightly slower metabolization in rat in comparison to mouse.
Altogether, regardless of the limited success to confirm the excellent in vitro potential of [ 18 F]11 to detect PDE2A protein in the brain in vivo, we conclude that [ 18 F]11 has potential as a starting point for further structural optimization to develop a PDE2A-specific PET ligand for neuroimaging of this enzyme in vivo.
PET images ( Figure 7A). The non-negligible fraction of 34% brain penetrant radiometabolites at 30 min p.i. probably contributes to high background noise, hindering the detection of a potential specific binding of [ 18 F]11 in vivo. Furthermore, the signal-to-background ratio in vivo under dynamic conditions might per se be lower than anticipated from the autoradiographic experiments performed under equilibrium conditions in vitro (caudate putamen-to-cerebellum ratio of about 2) [44]. These results demonstrated (i) an average brain uptake of [ 18 F]11 in mice and rats with (ii) no significant uptake in PDE2A-expressing regions and (iii) a contribution of radiometabolites to the measured PET signals due to the passage of a non-negligible fraction of radiometabolites through the blood-brain barrier in both species despite a slightly slower metabolization in rat in comparison to mouse.
Altogether, regardless of the limited success to confirm the excellent in vitro potential of [ 18 F]11 to detect PDE2A protein in the brain in vivo, we conclude that [ 18 F]11 has potential as a starting point for further structural optimization to develop a PDE2A-specific PET ligand for neuroimaging of this enzyme in vivo.

Syntheses
General Procedure 1 (GP1) An 0.33 M solution of ethyl 2-halo-5-iodobenzoate in dry THF was cooled to −40 to −50 • C and the temperature was maintained during the whole procedure. 1.1 eq of 1.3 M iPrMgCl · LiCl in THF were added during 15 to 20 min. After 1 h stirring, 1.2 eq of freshly distilled butanal was added over 10 to 15 min and the mixture was stirred for an additional hour. An aqueous saturated NH 4 Cl solution (20 mL) was added and the cooling was removed. After the solution was warmed to room temperature, 100 mL DCM was added and the aqueous phase was extracted three times with 20 mL DCM. The combined organic phases were dried over MgSO 4 , filtered and the solvent was removed under reduced pressure. The crude product was purified by flash chromatography. General Procedure 2 (GP2) A 0.33 M solution of the benzyl alcohol in dry DCM was cooled in an ice bath. Subsequently 1 mol% TsOH · H 2 O and 1.2 eq 3,4-dihydro-2H-pyran were added. After stirring for 16 h at room temperature, 10 mL of an aqueous saturated NaHCO 3 solution was added and the mixture was stirred for 1 h and then extracted three times with 20 mL DCM. The combined organic phases were dried over Na 2 SO 4 , filtered and the solvent was removed at reduced pressure. The crude product was purified by flash chromatography.

General Procedure 3 (GP3)
To a solution of the ester in methanol, hydrazine hydrate was added and the mixture was stirred at reflux or at room temperature for the given time. Water was added and the aqueous phase was extracted with DCM. The combined organic phases were dried over Na 2 SO 4 , filtered and the solvent was removed at reduced pressure.

2-Fluoro-5-(1-[(tetrahydro-2H-pyran-2-yl)oxy]butyl)benzohydrazide 7
Compound 7 was prepared according to GP3 from 526 mg (1.0 eq, 1.62 mmol) ester 5, 7.0 mL methanol and 787 µL (10.0 eq, 16.2 mmol) hydrazine hydrate. The solution was heated to reflux for 4 h, cooled to room temperature and 20 mL water and 20 mL DCM were added. The mixture was extracted three times with 20 mL DCM. The crude product was purified by flash chromatography using a gradient from 0 to 5% methanol in DCM yielding 309 mg (61%) of the diastereomeric mixture of the hydrazide as a viscous colorless oil.  Compound 8 was prepared according to GP3 from 2.21 g (1.0 eq, 5.73 mmol) ester 6, 10 mL methanol and 2.8 mL (10.0 eq, 57.3 mmol) hydrazine hydrate. The solution was heated to reflux for 5.5 h, cooled to room temperature and 20 mL water and 20 mL DCM were added. The mixture was extracted three times with 20 mL DCM. The crude product was purified by flash chromatography using a gradient from 0 to 5% methanol in DCM yielding 1.88 g (89%) of the diastereomeric mixture of the hydrazide as a viscous colorless oil.  To a 0.1 M solution of 1.0 eq hydrazide in ACN were added 6.0 eq DBU and 3-chloro-6-cyclopropyl-2-methylpyrido[2,3-b]pyrazine. After stirring at room temperature for 4 h, 2.0 eq PPh 3 was added. 15 min later, the solution was placed in an ice bath and 2.0 eq C 2 Cl 6 was added portionwise. The mixture was stirred for an additional hour at room temperature, silica gel was added and evaporated to dryness at reduced pressure. The residue was purified by flash chromatography.
TLC were added. After a further 22 h at 80 • C, 10 mL water was added and the mixture was extracted three times with 20 mL DCM. The combined organic phases were dried over Na 2 SO 4 , filtered and the solvent was removed at reduced pressure. The crude product was purified by flash chromatography using a gradient from 0 to 100% ethyl acetate in cyclohexane. Further purification with preparative HPLC at isocratic conditions using 80% ACN in water yielded 67 mg (23%) of the diastereomeric mixture of the boronic acid pinacol ester as a beige solid.
TLC  A mixture of 1.7 mL (6.0 eq, 16.7 mmol) 30% H 2 O 2 and 10.5 mL DCM was cooled to 0 • C and 2.71 mL (7.0 eq, 19.5 mmol) trifluoroacetic anhydride were added slowly. At room temperature, a solution of 656 mg (1.0 eq, 2.79 mmol) ethyl 2-amino-5-butyrylbenzoate (14) in 3.5 mL DCM was added slowly. After stirring for 20 h, the solution was cooled to 0 • C, 10 mL saturated Na 2 SO 3 solution was added and the mixture was extracted three times with 20 mL DCM. The combined organic phases were dried over Na 2 SO 4 , filtered and the solvent was removed at reduced pressure. The crude product was purified by flash chromatography using a gradient from 0 to 10% ethyl acetate in cyclohexane yielding 294 mg (40%) of the nitro compound as a pale yellow oil. To a solution of 265 mg (1.0 eq, 1.00 mmol) 15 and 1.66 mL (10.0 eq, 9.99 mmol) triethyl orthoformate in 5.0 mL ethanol were added 71 µL (1.0 eq, 1.00 mmol) acetyl chloride. After stirring for 24 h at reflux, the solution was cooled to room temperature, 5 mL saturated NaHCO 3 solution was added and the mixture was extracted three times with 20 mL DCM. The combined organic phases were dried over Na 2 SO 4 , filtered and the solvent was removed at reduced pressure. The crude product was purified by flash chromatography using a gradient from 0 to 50% ethyl acetate in cyclohexane yielding 201 mg (58%) of the ketal as a pale yellow oil. Compound 17 was prepared according to GP3 from 188 mg (1.0 eq, 0.554 mmol) 16, 5.0 mL methanol and 538 µL (20.0 eq, 11.1 mmol) hydrazine hydrate. The solution was stirred for three days at room temperature. Subsequently, 30 mL water was added and the mixture was extracted five times with 20 mL DCM. The crude product yielded 177 mg (98%) of the hydrazide as a pale yellow oil, which was used without further purification.  Radio thin layer chromatography (radio-TLC) was performed on silica gel (Polygram ® SIL G/UV 254 ) pre-coated plates with a mixture of dichloromethane/methanol 10/1 (v/v) as eluent. The plates were exposed to storage phosphor screens (BAS-IP MS 2025, FUJI-FILM Co., Tokyo, Japan) and recorded using the Amersham Typhoon RGB Biomolecular Imager (GE Healthcare Life Sciences, Freiburg, Germany). Images were quantified with the ImageQuant TL8.1 software (GE Healthcare Life Sciences).
Analytical chromatographic separations (radio-HPLC) were performed on a JASCO (JASCO Deutschland GmbH, Pfungstadt, Germany) LC-2000 system, incorporating a PU-2080Plus pump, AS-2055Plus auto injector (100 µL sample loop), and a UV-2070Plus detector coupled with a gamma radioactivity HPLC detector (Gabi Star, raytest Isotopenmessgeräte GmbH, Straubenhardt, Germany). Data analysis was performed with the Galaxie chromatography software (Agilent Technologies) using the chromatograms ob- − cartridge (ABX, Radeberg, Germany) which was preconditioned by rinsing with 5 mL of an aqueous 0.5 M solution of NaHCO 3 and 10 mL water. After loading, the cartridge was washed with 2.0 mL of anhydrous methanol and dried with a stream of nitrogen for 3 min. The activity was then eluted with 20 mg DMAPOTf (synthesized according to the procedure described by Zhang et al. [35]) dissolved in 500 µL of anhydrous methanol achieving activity recoveries of 85 %. The methanol was evaporated under a stream of nitrogen at 60 • C. The dry [ 18 F]DMAPF was then dissolved in 800 µL of the respective solvent (DMA or DMI) and divided into two portions. One portion was used for labeling the boronic acid pinacol ester precursor 13 (2 mg, 3.4 µmol in 300 µL solvent) and the other portion for the reference reaction with 4-biphenylboronic acid pinacol ester (2 mg, 7 µmol in 300 µL solvent; Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany). Before addition of the precursor solutions, the [ 18 F]DMAPF was treated with the copper catalyst [Cu(OTf) 2 (py) 4 ] (in a molar ratio of 1.5 to precursor) and the mixtures were stirred for 2 min at room temperature. The 18 F-labelings were performed at 115 • C. To determine radiochemical yields, samples were taken for radio-TLC and radio-HPLC at different time points (5, 10, 15, 20 min F]fluoride was dried by azeotropic distillation as described above and the respective final fluorination agent was dissolved in 1.0 mL of labeling solvent and divided in two portions. Thereafter, a solution of 1-2 mg of precursor 19 dissolved in 300 µL of the appropriate solvent was added, and 18 F-labeling was performed at different temperatures in dependence of the solvent used. To analyze the reaction mixture and to determine radiochemical yields, samples were taken for radio-TLC and radio-HPLC at different time points (5, 10, 15, 20 min).
[ 18 F]Fluoride (6-8 GBq) was trapped on a Sep-Pak ® Accell QMA light cartridge ( Figure 8, entry 1) and eluted into the reactor with 1150 µL of an aqueous ACN solution of TBAHCO 3 (150 µL 0.075 M TBAHCO 3 , 700 µL ACN, 300 µL water, entry 2). After azeotropic distillation for 4 min at 50 • C, 2.5 mL of ACN was added and the mixture was further azeotropically dried for 7 min at 70 • C. Thereafter, 1.0 mg (2.4 µmol) of the nitro precursor 19 dissolved in 800 µL of ACN (entry 4) was added, and the reaction mixture was stirred at 110 0 C for 12 min. After cooling to 30 • C, 24 µL (48 µmol) of a 2M sodium borhydride solution (2M NaBH 4 in triethylene glycol dimethyl ether, Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany) in 400 µL n-butanol was added (entry 5) and reduction of the keton function was performed within 5 min at 30 • C. After cooling, the reaction mixture was diluted with 2.0 mL of 20 mM aqu. ammonium formate, 0.5 mL water and 0.5 mL ACN (entry 6) and transferred into the injection vial (entry 7). Semi-preparative radio-HPLC was performed using a Reprosil-Pur C18-AQ column (250 × 10 mm; 10 µm; Dr. Maisch HPLC GmbH, Germany) with a solvent composition of 50% ACN/20 mM NH 4 OAc aq. at a flow rate of 6.5 mL/min (entry 8). [ 18 F]11 was collected in the dilution vessel (entry 9) previously loaded with 40 mL of H 2 O. Final purification was performed by passing the solution through a Sep-Pak ® C18 light cartridge (entry 10), followed by washing with 2 mL of water (entry 11) and elution of [ 18 F]11 with 1.2 mL of EtOH (entry 12) into the product vial (entry 13). The ethanolic solution was transferred out of the hot cell and the solvent was reduced under a gentle argon stream at 75 0 C to a final volume of 10-50 µL. Afterwards the radiotracer was diluted in isotonic saline to obtain a final product containing 10% of EtOH (v/v).
parative radio-HPLC was performed using a Reprosil-Pur C18-AQ column (250 × 10 mm; 10 µ m; Dr. Maisch HPLC GmbH, Germany) with a solvent composition of 50% ACN/20 mM NH4OAcaq. at a flow rate of 6.5 mL/min (entry 8). [ 18 F]11 was collected in the dilution vessel (entry 9) previously loaded with 40 mL of H2O. Final purification was performed by passing the solution through a Sep-Pak  C18 light cartridge (entry 10), followed by washing with 2 mL of water (entry 11) and elution of [ 18 F]11 with 1.2 mL of EtOH (entry 12) into the product vial (entry 13). The ethanolic solution was transferred out of the hot cell and the solvent was reduced under a gentle argon stream at 75 ⁰C to a final volume of 10-50 µ L. Afterwards the radiotracer was diluted in isotonic saline to obtain a final product containing 10% of EtOH (v/v).

Determination of Stability and logD Value
The determination of stability and logD value has been performed according to the procedures described by Sadeghzadeh and Wenzel et al. [45].

In Vitro Autoradiography
Frozen sagittal 12 µm brain sections obtained from female SPRD and male Wistar-Han rats (10-12 weeks old) were thawed, dried in a stream of cold air, and pre-incubated for 15 min with incubation buffer (50 mM TRIS-HCl, pH 7.4; supplemented with 1 mM MgCl 2 , 1 mM CaCl 2 , 2 mM KCl, and 2 U/mL Adenosine deaminase) at ambient temperature. Afterwards, the buffer was decanted and the cryosections dried again in a stream of cold air before covering with the radioligand in incubation buffer (~0.1 MBq [ 18 F]11/mL; A m at the time of the incubation~10 GBq/µmol) for incubation at ambient temperature for 90 min. For the determination of the non-specific binding the radioligand solution was supplemented with 1 µM of the PDE2-specific ligands BAY-60-7550 (IC50 for human PDE2A = 4.7 nM, for bovine PDE2A = 2.0 nM [46]) or PF-0518099 (IC50 for PDE2A = 1.6 nM [6]).
For the confirmation of the saturability of the binding sites as well as for the homologous competition experiment performed to estimate the K D of 11, the radioligand solution was supplemented with different concentrations of non-labeled 11 (0.1 nM-1 µM). The incubation was terminated by pouring off the incubation solution and washing the slides twice with ice-cold buffer (50 mM TRIS-HCl, pH 7.4/4 • C) for 2 min and rinsing by dipping in ice-cold demineralised water. Immediately afterwards, the sections were dried in a stream of cold air and exposed along with appropriate radioactive standards (1 µL of serial dilutions of the radioligand in incubation buffer, pipetted on a microscopic glass slide and air dried) to an 18 F-sensitive image plate for 120 min. Finally, the imaging plate was scanned at the highest possible resolution (pixel size 12.6 µm × 12.6 µm) in a phosphor imager (HD-CR 35; Duerr NDT GmbH, Bietigheim Bissingen, Germany) and the digital image analysed with an image analysis program (Aida Image Analyzer v. 4.26; Elysia-raytest GmbH, Straubenhardt, Germany).
For the determination of the protein concentration, four dried 12 µm cryosections were scraped from the microscopic glass slides, transferred in an Eppendorf tube and solubilised with 200 µL of demineralised water. The protein concentration of this solution was determined by a commercial assay (Pierce™ BCA Protein Assay Kit, Thermo Fisher Scientific, Schwerte, Germany). The area of the cryosections was determined during the analysis of the digitized autoradiographic images, and the volume was calculated by multiplying the area with the thickness of the slices. Finally, the protein concentration of the tissue was calculated (37 mg protein/g wet weight).

Metabolite Studies
Two non-anesthetized female CD-1 mice (30 and 33 g, 90 days) were injected intravenously with [ 18 F]11 (28 MBq = 94 nmol/kg and 25 MBq = 78 nmol/kg). The animals were anesthetized by isoflurane inhalation at 28 min p.i. and blood was collected from the retro-orbital plexus, followed by cervical dislocation at 30 min p.i. Immediately afterwards, the brains were isolated, washed twice with isotonic saline and stored on ice. Plasma was obtained from the whole blood sample by centrifugation (8.000 rpm, 2 min, ambient temperature). All samples were weighed and the radioactivity was measured in a dose calibrator (ISOMED 2010 Dose Calibrator, NUVIA Instruments GmbH, Dresden, Germany). Afterwards, the brains were homogenized in 1 mL demineralized water on ice (Potter S, B. Braun Biotech, Melsungen, Germany).
Metabolism in rats has been investigated in the animals applied in PET studies (protocol reported below). In brief, blood was taken from two anesthetized female SPRD rats 30 min after intravenous injection of [ 18 F]11 followed by decapitation and isolation of the brains. The blood samples and the brains were processed as described above.
Protein precipitation was performed by the addition of ice-cold ACN/H 2 O (9:1, v/v) in a ratio of 4:1 (v/v) of organic solvent to plasma or brain homogenate, respectively. The samples were vortexed for 3 min, equilibrated on ice for 3 min, and centrifuged for 10 min at 10,000 rpm and the supernatant collected. The precipitates were washed with 100 µL of the solvent mixture and subjected to the same procedure. The combined supernatants were concentrated at 70 • C under argon flow to a final volume of approximately 100 µL and analysed by analytical radio-HPLC with a gradient system (see radiochemistry general section). The peak corresponding to the radiotracer in the radio-chromatogram was identified by co-injecting the sample with the reference compound 11. To determine the percentage of activity in the supernatants compared to total activity, aliquots of each step as well as the precipitates were quantified by an automated gamma counter (1480 WIZARD, Perkin Elmer, Turku, Finland). For plasma and brain samples recoveries of ≥91% of total activity were obtained.

In Vivo PET Studies in CD-1 Mice and SPRD Rats
For the time of the experiments, one female CD-1 mouse (10 weeks; 30.5 g) and one female SPRD rat (9 weeks; 256 g) were kept in a dedicated climatic chamber with free access to water and food under a 12:12 h dark:light cycle at a constant temperature (24 • C). The animals were anesthetized (Anaesthesia Unit U-410, agntho's, Lidingö, Sweden) with isoflurane (2.0%, 300 mL/min) delivered in a 60% oxygen/40% air mixture (Gas Blender 100 Series, MCQ instruments, Rome, Italy) and their body temperature maintained at 37 • C with a thermal bed system for the duration of the PET studies. A dynamic PET scan (PET/MR 1T, Mediso nanoScan ® , Hungary) started 20 s before intravenous injection into the tail vein of [ 18 F]11 and was performed in mouse for 60 min (4.9 MBq in 150 µL isotonic saline; 20.5 nmol/kg; A m : 8.9 GBq/µmol, EOS) in rat for 30 min (29.7 MBq in 400 µL isotonic saline; 12 nmol/kg; A m : 11 GBq/µmol, EOS). Each PET image was corrected for random coincidences, dead time, scatter and attenuation (AC), based on a whole body (WB) MR scan. The reconstruction parameters for the list mode data were 3Dordered subset expectation maximization (OSEM), 4 iterations, 6 subsets, energy window: 400-600 keV, coincidence mode: 1-5, ring difference: 81. Preceding the PET scan a T1 weighted WB gradient echo sequence (TR/TE: 20/6.4 ms, NEX: 1, FA: 25, FOV: 64 × 64 mm, Matrix: 128 × 128, STh: 0.5 mm) was performed for AC and anatomical orientation. PET Images were analyzed with PMOD (PMOD Technologies LLC, v. 4.1, Zurich, Switzerland). The respective brain regions were identified using the mouse brain atlas template Ma-Benveniste-Mirrione-T2 or the rat brain atlas template Schwarz-T2. The regional activity data (radioactivity concentration, kBq/mL) were normalized to the injected dose (MBq) and corrected for the animal weight (g) to provide standardized uptake values (SUV, g/mL). Finally, the results of the PETscans are expressed as mean SUV of the respective region of interest ROI.

Conclusions
With the aim to develop a specific radiotracer for neuroimaging of the cyclic nucleotide phosphodiesterase PDE2A by PET, a new 18 F-labeled triazolopyridopyrazine-based PDE2A inhibitor was developed. [ 18 F]11 was obtained in a two-step one-pot radiolabeling procedure, and the establishment of an automated radiosynthesis in the TRACERlab synthesis module enabled a reproducible production with sufficient radiochemical yields and high radiochemical purity. In vitro autoradiography indicated high affinity and specific binding of [ 18 F]11 towards PDE2A in the rat brain with apparent K D and B max values of about 0.2 nM and 20 pmol/mg protein in the PDE2A-rich brain region nucleus caudate and putamen. In vivo PET/MR studies in mice and rat revealed a moderate brain uptake, however, no considerable accumulation of activity in PDE2A-specific regions was observed. Furthermore, in vivo metabolism studies revealed a significant fraction of brain-penetrant radiometabolites of [ 18 F]11. Altogether, [ 18 F]11 shows excellent in vitro properties for the investigation of the enzyme PDE2A in the brain and is a suitable starting point for further developments towards a successful imaging agent for PET neuroimaging of PDE2A.