Synthesis and in vitro evaluation of [¹⁸F]FECIMBI-36: A potential agonist PET ligand for 5-HT_2A/2C receptors

Abstract

Radiosynthesis and in vitro evaluation of [¹⁸F]-2-(4-bromo-2,5-dimethoxyphenyl)-N-(2-(2-fluoroethoxy)benzyl)ethanamine, ([¹⁸F]FECIMBI-36) or ([¹⁸F]1), a potential agonist PET imaging agent for 5-HT_2A/2C receptors is described. Syntheses of reference standard 1 and the corresponding des-fluoroethyl radiolabeling precursor (2) were achieved with 75% and 65% yields, respectively. In vitro pharmacology assay of FECIMBI-36 by [³H]-ketanserin competition binding assay obtained from NIMH-PDSP showed high affinities to 5-HT_2AR (K_i = 1 nM) and 5-HT_2CR (K_i = 1.7 nM). Radiolabeling of FECIMBI-36 was achieved from the boc-protected precursor 2 using [¹⁸F]-fluoroethyltosylate in presence of Cs₂CO₃ in DMSO followed by removal of the protective group. [¹⁸F]1 was isolated using RP-HPLC in 25 ± 5% yield, purity ≥95% and specific activity 1–2 Ci/μmol (N = 6). In vitro autoradiography studies demonstrate that [¹⁸F]1 selectively label 5-HT_2A and 5-HT_2C receptors in slide-mounted sections of postmortem human brain using phosphor imaging. Our results indicate the potential of [¹⁸F]1 for imaging 5-HT_2A/2C receptors in the high affinity state in vivo using PET imaging.

5-HT₂Rs belong to the class of G-protein coupled receptors (GPCRs) and have three 3 subtypes, namely 5-HT_2AR, 5-HT_2BR and 5-HT_2CR with 70–80% sequence homology.^1–4 5-HT_2AR is the most abundant excitatory 5-HT receptor in human brain and plays an essential role in a number of physiological processes and psychiatric disorders including dementia, schizophrenia, suicidal behavior and major depression.^1–8 5-HT₂Rs are also implicated in the therapeutic effects of some antidepressants as well as antipsychotics.^9–11 5-HT_2BR is less abundant in the central nervous system (CNS), whereas 5-HT_2CR is abundant in the choroid plexus, and at a low level in the striatum, hippocampus, neocortex and amygdala.^1,10 5-HT₂R can be present in a high or low agonist affinity binding state.¹ In the CNS, 5-HT_2AR are found in large amounts in the forebrain and neocortex and low levels in brainstem nuclei and hippocampus.^12,13

Although there are several radioligands available for in vivo imaging of 5-HT_2AR, none can detect changes in the high affinity state in pathological conditions. Whether any alteration of 5-HT_2AR in major CNS disorders involves the high-affinity conformation is not known.^14–18 An agonist positron emission tomography (PET) radiotracer would have many advantages for 5-HT_2AR imaging because agonists bind the functionally relevant high-affinity sub-population of GPCRs.^14–18 The availability of a well-characterized agonist PET tracer would facilitate the measurement of minimally effective dose and in addition, a combination of antagonist and agonist PET data would allow separate quantification of the high- and low-affinity conformational states of the 5-HT₂R.

[¹¹C]CIMBI-5 (Fig. 1); 5-HT_2AR K_i = 0.044 nM, E_max = 85%, 5-HT_2BR K_i = 73 nM and 5-HT_2CR K_i = 2 nM,¹² is the first 5-HT_2AR agonist PET tracer studied by others and us in vivo in pigs and nonhuman primates.^12–16 More recently, [¹¹C]CIMBI-36, a bromoanalog of CIMBI-5, was developed by Ettrup et al. and evaluated in pig and nonhuman primates.^17,18 [¹¹C]CIMBI-36 is currently being examined in human subjects for its ability to quantify 5HT_2AR by PET imaging.^17–19 PDSP data for CIMBI-36 shows that it has a 5HT_2AR K_i = 0.5 nM and E_max = 87%, 5HT_2BR K_i = 0.5 nM, and 5-HT_2CR K_i = 1.7 nM.¹⁷ PET studies in nonhuman primates suggest that cortical binding of [¹¹C]CIMBI-36 is mostly due to 5-HT_2AR, whereas, tracer uptake in choroid plexus can be attributed to the binding to 5-HT_2CR population. PET blocking studies of [¹¹C]CIMBI-36 with the 5HT_2AR antagonist ketanserin in human subjects reveal a cortex to cerebellum ratio <2.^17,18 Furthermore, the radioligand has slow washout, particularly in cortex, a brain region of high receptor density.¹⁹ The slow kinetics make accurate kinetic modeling difficult, and this would be a serious limitation for research use of [¹¹C]CIMBI-36. To overcome this disadvantage, PET imaging of 5-HT_2AR using [¹⁸F] radioligands (t_1/2 = 110 min) would allow longer duration scanning and enable valid modeling.^15,16 Additionally, [F-18] tracers due to their 110 min half-life, permit deliveries to PET Centers that lack an onsite cyclotron and thereby allow cost effective multicenter clinical studies. We identified the fluoroethyl version of CIMBI-36 (FECIMBI-36, 1) as a potential agonist ligand for 5-HT_2AR [¹⁸F]-tracer development. Herein we describe the synthesis, pharmacology assessment, radiolabeling and in vitro evaluation of [¹⁸F]FECIMBI-36 in slide-mounted sections of postmortem human brain.

Figure 1.

Chemical structures of CIMBI-5, CIMBI-36 and FECIMBI-36.

The synthesis of desfluoroethyl CIMBI-36 (2) was achieved according to minor modifications of a published procedure.^14,20 The synthesis of reference standard FECIMBI-36 (1) was achieved in 75% yield by reacting the corresponding boc-protected precursor (2) with 1-bromo-2-fluoroethane and NaH, followed by removal of the protective group with trifluoroacetic acid at room temperature (Scheme 1).²¹ The affinities of FECIMBI-36 for 5-HT_2AR and various other biogenic amine receptors and transporters and functional assays were determined through the National Institute of Mental Health Psychoactive Drug Screening Program (NIMH-PDSP) (Table 1).

Scheme 1.

Synthesis of FECIMBI-36 and radiosynthesis of [¹⁸F]FECIMBI-36. Reagents and conditions: (i) Fluoroethyl bromide, NaH, DMF, 110 °C, 12 h; (ii)TFA/DCM, 15 min, 75%, (iii) [¹⁸F]FEOTs, Cs₂CO₃, DMSO, 120°C, 20min; (iv) TFA/CAN (1:1), 80°C, 10min.

Table 1.

Affinity and selectivity of FECIMBI-36

Targets	Ki (nM)	Targets	Ki (nM)
5-HT1A	1433	D1	>10,000
5-HT1B	>10,000	D2	1394
5-HT1D	>10,000	D3	1076
5-HT1E	>10,000	KAR	>10,000
5-HT2A	1.0	D4	>1000
5-HT2B	2.8	D5	>10,000
5-HT2C	1.7	DAT	>10,000
5-HT3	>10,000	DOR	>10,000
5-HT7	>10,000	EP	>10,000
Adenosine	>10,000	GABA	>10,000
α1A–1C	>10,000	H4	2126
α2A	551	HERG	>10,000
α2B	1256	KOR	558
α2C	656	M	>10,000
β1	>10,000	H3, H4	>10,000
β2	3220	MDR1	>10,000
β3	>10,000	H2	1983
BZP	>10,000	H1	150
Ca++	>10,000	MOR	149
AMPA	>10,000	mGluR	>10,000
NET	208	NMDA	>10,000
NK	>10,000	SERT	284
Sigma1	499	Sigma2	51
Na+	>10,000	NT1	>10,000
CB1, CB2	>10,000	I	>10,000
mGluRs	>10,000	VMAT	>10,000
NOP	>10,000	NMDA	>10,000
PBR	876	NT	>10,000
Oxytocin	>10,000	PKC	>10,000
Smoothened	>10,000	V1–2	>10,000
Agonist assays	5-HT2AR	5-HT2BR	5-HT2CR
	Emax: 63.4%	Emax: 18.4%	Emax: 100%

K_i values measured against [³H]ketanserin show that FECIMBI-36 has high affinity for 5-HT₂Rs (Table 1). FECIMBI-36 showed no significant affinities for other tested brain targets, pGP and MDRs and selected affinities are shown in Table 1. Agonist potency assays revealed that E_max values of FECIMBI-36 for 5-HT_2AR, 5-HT_2BR and 5-HT_2CR are 63.4%, 18.4% and 100%, respectively, with no antagonistic properties (Table 1).

Radiosynthesis of [¹⁸F]FECIMBI-36 was achieved by a one pot, two step reaction with the phenolate of 2 (Scheme 1). In summary, reaction of [¹⁸F]FEOTs with precursor 2 in the presence of Cs₂CO₃ afforded [¹⁸F]1 in 25 ± 5% yield and >95% chemical and radiochemical purities (N = 6).²² The [¹⁸F]FEOTs in turn was synthesized by a procedure previously reported by us.²³ Total synthesis and purification time required for the radiosynthesis was 50 min at EOS (N = 6). The specific activity of the radioproduct was 1–2 Ci/μmol at EOS. Partition coefficient (logP) of [¹⁸F]1 was 4, indicating that the radioligand should have adequate lipophilicity for passive brain entry.²⁴ The radioproduct was stable in saline–ethanol (9:1) for up to 6 h based on analytical HPLC analyses and no detectable [¹⁸F]radiofluorination was found over 6 h in the formulated solution. We determined the potential of [¹⁸F]CIMBI-36 to label binding sites in sections of human brain using modifications of a method described by Finnema et al.^18,25 Sections (20 μm) of prefrontal cortex Brodmann area (BA9), hippocampus, temporal cortex and choroid plexus from a coronal slab of hemisphere, and for comparison, sections of cerebellum from the same 37 year old male case, were incubated with [¹⁸F]1 for 60 min at 37 °C The 5-HT_2AR antagonist ketanserin (10 μM) was used to define nonspecific binding. Following incubation, sections were rapidly dried and exposed to phosphorimaging screens for 60 min.

As is evident from Figure 2, there is binding of [¹⁸F]1 to 5HT₂R in prefrontal cortex, temporal cortex, hippocampus and choroid plexus. The nonspecific binding was reduced when the washing time was increased from 6 min to 20 min (data not shown). Binding was observed throughout the gray matter of the prefrontal cortex. Specific binding (total binding − nonspecific binding × 100) is approximately 40% of total binding in prefrontal cortex (Fig. 3). The distribution of specific binding corresponded to the distribution of the 5-HT_2A antagonist ligand [³H]ketanserin,^26,27 suggesting that the [¹⁸F]FECIMBI-36 specific binding is largely to the 5-HT_2AR in gray matter of the prefrontal cortex and other regions including the temporal cortex and hippocampus. As shown in Figure 2, specific binding is also found in the choroid plexus, a region where the highest concentration of 5HT_2CR is present. No appreciable binding was found in cerebellum, which is consistent with the known distribution of 5-HT₂R in human brain. The cerebellar binding of [¹⁸F]1 is lower than [¹¹C]CIMBI-36 and has better target to non-target ratio in vitro.

Figure 2.

Phosphor images of [¹⁸F]1 in postmortem human brain sections. Total (left upper and lower images) and nonspecific (right upper and lower images) binding of [¹⁸F]FECIMBI-36 in human brain sections. Nonspecific binding was determined by ligand displacement with 10 μM ketanserin. Note there is more binding in gray matter than in white matter and there is binding and displacement in the choroid plexus.

Figure 3.

Specific binding of [¹⁸F]1 in postmortem human brain sections. Percent specific binding of [¹⁸F]FECIMBI-36. Orange bars are the mean of measurements from two IDs and blue bars are measurements from one ID. All regions were assayed in triplicate within an experiment.

In summary, we synthesized FECIMBI-36 and determined its binding affinity and selectivity to the 5HT₂R and a variety of other receptors and transporters. The radiosynthesis of [¹⁸F]1 was performed successfully in 25 ± 5% yield and the total time required for radiosynthesis was 50 min from EOS with excellent chemical and radiochemical purities and high specific activity. Autoradiography studies show high binding of [¹⁸F]1 to 5-HT_2A/2C receptors in various brain regions of postmortem human brain sections from nonpsychiatric controls. Our results indicate [¹⁸F]FECIMBI-36 as a potential agonist radiotracer with 110 min half-life for sensitive and accurate in vivo detection of 5HT_2A/2CR in neuropsychiatric disorders, and to measure in vivo occupancies of novel drugs under development.