Development of [¹⁸F]ASEM, a specific radiotracer for quantification of the α7-nAChR with positron-emission tomography

Abstract

The alpha-7 subtype of the nicotinic acetylcholine receptor (α7-nAChR) is fundamental to physiology; it mediates various brain functions and represents an important target for drug discovery. Exploration of the brain nicotinic acetylcholine receptors (nAChRs) using positron-emission tomography (PET) will make it possible to better understand the important role of this receptor and to study its involvement in schizophrenia, bipolar disorder, Alzheimer's and Parkinson's diseases, drug dependence, inflammation and many other disorders and simplify the development of nicotinic drugs for treatment of these disorders.

Until recently, PET imaging of α7-nAChRs has been impeded by the absence of good radiotracers. This review describes various endeavors to develop α7-nAChR PET tracers by several research groups including the author’s group. Most initial PET tracers for imaging α7-nAChRs did not exhibit suitable imaging properties due to their low specific binding. Recently discovered [¹⁸F]ASEM is the first highly specific α7-nAChR radioligand and it was recently translated to human PET imaging.

1. Introduction

Nicotinic cholinergic receptors (nAChRs) are neurotransmitter-gated cationic channels that are present in the central nervous system (CNS), autonomic and sensory ganglia, and various non-neuronal cells. Two nAChR subtypes, α4β2- and α7-nAChR, are the most abundant nAChRs in the CNS[1].

The α7-nAChR subtype is highly expressed in the human brain and this subtype has been implicated in the pathophysiology of a variety of brain disorders and conditions including schizophrenia, Alzheimer's disease, bipolar disorder, traumatic brain injury, anxiety, depression, multiple sclerosis, inflammation, and drug addiction[1-10].

As was demonstrated in post-mortem studies, the density of α7-nAChRs in human brain tissue is significantly altered in many disorders:

Schizophrenia

In autoradiography and immunochemistry studies, Freedman et al[11] and others [12-17] have demonstrated a significant post-mortem reduction (25-54%) of α7-nAChR binding or expression in the hippocampus and cortex of subjects with schizophrenia vs. controls.

Alzheimer’s disease

A characteristic of Alzheimer’s disease is degeneration of cholinergic neurons[18]. A number of reports have described a significant loss of α7-nAChRs in the cortex and hippocampus of patients with Alzheimer’s disease [15, 19] (see also review[20]).

Bipolar disorder

An autoradiography study using an α7-nAChR radiotracer demonstrated an increased binding in the hippocampus and perirhinal cortex in the brain slices of the subjects suffering from bipolar disorder[21].

Traumatic brain injury

Traumatic brain injury is a significant public health problem with almost 2 million documented cases per year in the USA, with a mortality of 20%[22, 23]. Several reports found a significant reduction (30-70%, ex vivo or in vitro) of α7-nAChRs in animal models of traumatic brain injury [8, 24-26], suggesting that alteration of the α7-nAChR is a crucial component of the biochemical perturbation caused by traumatic brain injury.

The difference in the denstity of α7-nAChRs in the brain between healthy subjects and patients suffering from various disorders was quantified in post-mortem studies, but it never was observed in the living human brain. Non-invasive quantification of α7-nAChRs in humans would provide a better understanding of their role in various CNS disorders and could also simplify the development of nicotinic drugs for treatment of these disorders [27-32].

PET provides the best opportunity for quantification of receptors in the human brain – better than any other clinical imaging modality [33, 34]. However, since the invention of the PET technique in 1975 fewer than 40 of the existing receptors in the human brain have been imaged due to the lack of available PET radiotracers (see http://www.nimh.nih.gov/research-priorities/therapeutics/cns-radiotracer-table.shtml). Until recently, one of the major cerebral receptors lacking an appropriate PET radioligand for human imaging was α7-nAChR. The recently developed PET radioligand [¹⁸F]ASEM has opened new avenues in noninvasive imaging of this receptor system in human subjects.

2. Initial PET radioligands for α7-nAChRs

In principle, a quality α7-nAChR radioligand for PET should exhibit the same set of characteristics as PET tracers for most other brain receptors: 1) a high specific and low non-specific binding in vivo; 2) high selectivity versus non-target binding sites; 3) reversible brain kinetics with good blood-brain barrier permeability; 4) radiochemistry that is suitable for short-lived isotopes; and 5) low radiation burden and toxicity. These general requirements for PET radiotracers have been summarized in many reviews [35-37].

While all general requirements must be met, the high specific binding is the most demanding property in the development of α7-nAChR radiotracers. Specific PET tracers for brain receptors are expected to obey the Eckelman’s criterion that B_max/K_D ≥ 10 (B_max = binding site density; K_D = binding affinity constant of the radiotracer)[38, 39]. The concentration of the α7-nAChR binding sites in the primate brain is low (B_max = 5 – 15 fmol/mg protein or 1.5 – 12 fmol/mg tissue)[12, 40, 41]. Consequently, the expected binding affinity for a quality α7-nAChR PET radioligand must be in a sub-nanomolar range. This binding affinity requirement challenged the development of suitable α7-nAChR radioligands (see reviews[35, 42, 43]).

Investigators have been attempting to develop α7-nAChR radioligands for in vivo imaging since the pioneering work of the Dolle[44] (Orsay, France) and Pomper[45] (Baltimore, US) in 2001-2005. Both groups radiolabeled quinuclidine derivatives (Fig. 1) that were previously reported by AstraZeneca as potential α7-nAChR drugs. Unfortunately, these radiotracers did not exhibit a sufficient signal-to-noise ratio in lab animals and were not translated to humans.

Fig. 1.

First radioligands for emission tomography imaging of α7-nAChR [44, 45].

In 2005-2010 many researchers, including our own group, worked on the development of a clinically viable α7-nAChR PET radioligand, and about two dozen α7-nAChR compounds were radiolabeled with [¹⁸F] or [¹¹C]. As summarized in the recent reviews[35, 46-48], those efforts did not lead to an α7-nAChR PET radioligand with sufficient in vivo specificity.

[¹¹C]CHIBA-1001 was the only α7-nAChR radioligand translated to human PET in the past, but it showed low target-to-non-target ratios in the brain (<1.3) in a single human PET scan[49] (Fig. 2). Further blocking experiments using PET in humans demonstrated some specificity of the [¹¹C]CHIBA-1001 binding that, however, was not sufficiently high for reliable receptor quantification. This result led [¹¹C]CHIBA-1001 inventors to a conclusion that a better PET radioligand is necessary[50]. The low specific binding of [¹¹C]CHIBA-1001 in the human brain is in agreement with its relatively low in vitro binding affinity ([³H]CHIBA-1001, rat or human K_D = 120 - 193 nM) and its inadequate α7-nAChR regional distribution in the rodent brain[51, 52].

Fig. 2.

Left: Structure of [¹¹C]CHIBA-1001. Right: PET images of human brain with [¹¹C]CHIBA-1001. Panel a Magnetic resonance images (MRI) of the corresponding slices. Panel b Static images acquired from 0 to 90 min after injection of [¹¹C]CHIBA-1001 expressed as SUV. Panel c A parametric image for the total distribution volume of [¹¹C]CHIBA-1001 generated using Logan graphical analysis. The data from 30 to 90 min were applied to the Logan plot analysis. Reprinted from[49] with permission of Copyright Clearance Center's RightsLink service.

The most recent α7-nAChR PET radioligands, [¹⁸F]AZ11637326[53, 54], [¹¹C]NS14492[55] and [¹⁸F]NS10743[56, 57] (Fig. 3), exhibited some specific binding in the brains of lab animals, but their specificities were also insufficient for human PET. The main reason for the deficient PET properties of most of the initial radioligands was due to the low binding affinity of these compounds (see for review[48]). Another recent PET tracer [¹¹C]A752274 (Fig. 3) that was developed by collaboration of Abbott Laboratories and Johns Hopkins University exhibited very high binding affinity (K_i = 0.092 nM). Unfortunately, [¹¹C]A752274 is a polar compound with low lipophilicity (logD_7.4 = -2.7) that shows low brain uptake in animals, which makes it inappropriate for brain PET[58].

Fig. 3.

Recent α7-nAChR PET radioligands, [¹⁸F]AZ11637326[53, 54], [¹¹C]NS14492[55], [¹⁸F]NS10743[56, 57] and [¹¹C]A-752274[58].

3. Development of [¹⁸F]ASEM[48, 59]

3.1 Synthesis of ASEM

In 2012 Abbott Laboratories disclosed a number of α7-nAChR ligands that were synthesized as potential drug candidates[60]. One of the compounds of the series was 3-(1,4-diazabicyclo[3.2.2]nonan-4-yl)dibenzo[b,d]thiophene 5,5-dioxide (Fig. 4), an α7-nAChR selective ligand with exceptionally high binding affinity, K_i = 0.023 nM, and the ability to penetrate the blood-brain barrier in lab animals[60]. These properties made this dibenzothiophene compound an attractive lead for development of quality α7-nAChR PET tracers.

Fig. 4.

Structure of 3-(1,4-diazabicyclo[3.2.2]nonan-4-yl)dibenzo[b,d]thiophene 5,5-dioxide, an α7-nAChR selective ligand with high binding affinity that was developed by Abbott[60].

Even though the Abbott lead (Fig. 4) is difficult to radiolabel with the PET radionuclide ¹¹C, the presence of an electron-withdrawing sulfonyl group opened an opportunity for making [¹⁸F]fluoro-derivatives of this compound via the nucleophilc aromatic substitution with [¹⁸F]fluoride. PET chemists from JHU synthesized a series of fluoro-derivatives of 3-(1,4-diazabicyclo[3.2.2]nonan-4-yl)dibenzo[b,d]thiophene 5,5-dioxide. Within the series, 4-(6- fluorodibenzo[b,d]thiophen-3-yl)-1,4-diazabicyclo[3.2.2]nonane 5,5-dioxide (JHU82132, ASEM) and its para-isomer (JHU82108, para-ASEM) exhibited the best α7-nAChR in vitro binding affinity and high selectivity vs. heteromeric nAChR subtypes or 5-HT₃[48, 59] (see Table 1). The abbreviation “ASEM” stands for ALPHA-SEVEN (Α-СЕМЬ) in Greek-Russian as was suggested by Dr. R.F. Dannals from Johns Hopkins University.

Table 1.

Inhibition in vitro binding affinities (K_i, nM) of ASEM, para-ASEM and the Abbott lead toward α7-nAChR, heteromeric nAChR subtypes and 5-HT₃ [48].

Compound	α7-nAChR a	Heteromeric nAChR subtypes b						5-HT3c	Selectivity
		α2β2	α2β4	α3β2	α3β4	α4β2	α4β4		α7/α4β2	α7/5HT3
	0.3, 0.5	-	-	-	-	-	-	660d	-	-
	0.37, 0.45	>10000	4000	1000	709	562	1000	230	1370	561
	1.32, 1.35	1000	8000	2000	5000	885	3000	505	663	378

^aRat cortical membranes, radiotracer [¹²⁵I]α-bungarotoxin (0.1 nM), K_D = 0.7 nM

^bInhibition in vitro binding assay of all heteromeric nAChR subtypes was performed with stably transfected HEK293 cells and [³H]epibatidine (0.5 nM), K_D = 0.021 nM (α2β2-nAChR), K_D = 0.084 nM (α2β4-nAChR), K_D = 0.034 nM (α3β2-nAChR), K_D = 0.29 nM (α3β4-nAChR), K_D = 0.046 nM (α4β2-nAChR), K_D = 0.094 nM (α4β4-nAChR).[80]

^cHuman 5-HT₃ recombinant/HEK293 cells, radiotracer [³H]GR65630 (0.35 nM), K_D = 0.5 nM

^dthe K_i value is taken from[60]

Prior radiolabeling of [¹⁸F]ASEM and further animal experiments the JHU group forecasted the PET imaging value of this compound by comparison of its in vitro binding affinity versus the previous best α7-nAChR PET tracers [¹⁸F]AZ11637326, [¹¹C]NS14492 and [¹⁸F]NS10743(Fig. 3). The binding assay of all four compounds was performed under the same assay conditions (Table 2). This head-to-head comparison demonstrated that the α7-nAChR affinity of ASEM is 1-2 orders of magnitude superior to the previous radiotracers[48] and in vivo specific binding of [¹⁸F]ASEM was expected to be proportionally greater.

Table 2.

Comparison of in vitro α7-nAChR inhibition binding affinities of ASEM versus previous PET radioligands NS14492, NS10743, AZ11637326.

Compound	Ki, nM a[48]
NS14492	20.4
NS10743	38.0
AZ11637326	3.3
ASEM	0.37, 0.45

^aThe binding assay performed under the same conditions: rat cortical membranes, radiotracer [¹²⁵I]α-bungarotoxin (0.1 nM), K_D = 0.7 nM (commercial assay, CEREP, www.cerep.fr).

In addition to the promising binding affinity, other molecular determinants of ASEM that are important for the blood-brain barrier permeability (molecular weight = 358; lipophilicity logD_7.4 = 2.0; polar surface area = 49) are within the optimal range for most brain PET tracers[61, 62].

The appropriate in vitro properties of ASEM for PET and its potential suitability for [¹⁸F]-radiolabeling were the driving forces behind the radiosynthesis of [¹⁸F]ASEM and further in vivo experiments with this radiotracer. The isomer para-ASEM was selected as a back-up compound since it manifested fairly comparable in vitro binding affinity properties as ASEM (Table 1).

3.2 Radiosynthesis of [¹⁸F]ASEM and para-[¹⁸F]ASEM

Among the main requirements for [¹⁸F]-labeled (half-life t_1/2 = 109.77 min) PET tracers is an efficient radiolabeling and its suitability for automation and preparation of the final product with high specific radioactivity, purity and radiochemical yield. The high specific radioactivity of PET radiotracers for neuroreceptors is a general requirement because of the concerns about the binding competition between the [¹⁸F] radiotracer and its natural non-radioactive [¹⁹F] isotopomer (carrier), which is always present in the radiolabeled product [63]. For reliable quantification analysis and also due to the safety concerns, a dose of PET radiotracer should occupy less than 5% of the available binding sites. This general requirement makes it necessary for a radiotracer to contain a low mass of carrier or, in other words, exhibit high specific radioactivity. Due to the low density of α7-nAChRs in the brain (see above) the radiotracers for this receptor should be used for PET studies only at a specific radioactivity greater than 5000 – 10000 mCi/µmol [35].

[¹⁸F]ASEM and para-[¹⁸F]ASEM were prepared by a Kryptofix-222^® – assisted reaction of the corresponding nitro-precursors, PRE-ASEM and PRE-para-ASEM, with [¹⁸F]fluoride (Fig. 5). The radiosynthesis was performed remotely in a PC-controlled radiochemistry synthesis module (Microlab, GE). The final radiolabeled products were purified by preparative high-performance liquid chromatography (HPLC) and solid-phase extraction and formulated as sterile apyrogenic solutions in 7% ethanolic saline. Both radiotracers, [¹⁸F]ASEM and para- [¹⁸F]ASEM, were prepared with comparable radiochemical yields of 16 ± 6% (n=14) (non-decay-corrected), specific radioactivities in the range of 330 - 1260 GBq/μmol (9-34 Ci/μmol), and a radiochemical purity greater than 99%[48].

Fig. 5.

Radiosynthesis of [¹⁸F]ASEM and para-[¹⁸F]ASEM using a PC-controlled radiochemistry synthesis module (Microlab, GE) [48].

With increased demand for the [¹⁸F]ASEM for human and animal PET studies, a fast microwave-assisted synthesis with improved radiochemical yield (25-50%) was further developed[64]. The microwave method allows a routine preparation of about 500 mCi [¹⁸F]ASEM per batch with high specific radioactivity (>10000 mCi/µmol) and radiochemical purity (>98%), which means that it can be used for several PET scans within the same day.

4. Pre-clinical studies with [¹⁸F]ASEM in mice

4.1 Biodistribution in CD1 mice

After the successful radiosynthesis of [¹⁸F]ASEM, the radiotracer biodistribution, in vivo binding specificity, selectivity and brain kinetics were evaluated in CD1 mice [48].

In the ex vivo biodistribution experiments each mouse received an injection of [¹⁸F]ASEM into a lateral tail vein [48]. The animals were sacrificed at certain time points (5 – 120 min) and the brain regions were quickly dissected and assayed in a gamma-counter and time-radioactivity curves (Fig. 6) were generated. The study demonstrated that [¹⁸F]ASEM readily entered the mouse brain. The peak brain uptake (7.5% injected dose/g tissue) was seen at 5 min post injection, followed by a gradual decline. The time-radioactivity curves showed that the highest accumulation of [¹⁸F]ASEM radioactivity occurred in the colliculus, hippocampus and frontal cortex, intermediate radioactivity was observed in the striatum and the rest of the brain and the lowest radioactivity was seen in the cerebellum. This distribution of [¹⁸F]ASEM radioactivity is comparable to the in vitro distribution of α7-nAChRs in rodent brain tissue [65, 66].

Fig. 6.

Regional distribution of [¹⁸F]ASEM in CD-1 mice. Data: mean %injected dose/g tissue ± SD (n = 3). Abbreviations: Coll = superior and inferior colliculus; Hipp = hippocampus; FrCtx = frontal cortex; Rest = rest of brain; Str = striatum; CB = cerebellum. Reprinted from [48] with permission of Copyright Clearance Center's RightsLink service. The study demonstrated that [¹⁸F]ASEM exhibits high brain uptake with regional distribution that matches the distribution of the α7-nAChR in the rodent brain and reversible brain kinetics.

The high α7-nAChR specificity of [¹⁸F]ASEM binding was demonstrated by the blocking experiments in CD1 mice with the selective α7-nAChR partial agonist SSR180711 (Fig. 7). For this study each animal was injected with a mixture of [¹⁸F]ASEM and SSR180711. As expected, the [¹⁸F]ASEM accumulation in the α7-nAChR-rich brain regions was dose-dependently blocked by SSR180711. This result proved that [¹⁸F]ASEM uptake in the mouse brain is specific and mediated by α7-nAChRs. The dose escalation blockade also demonstrated that [¹⁸F]ASEM is a suitable tool for in vivo evaluation of potential new α7-nAChR drugs.

Fig. 7.

Dose dependent blockade of [¹⁸F]ASEM (0.07 mCi, specific radioactivity = 7900 mCi/μmol, i.v.) accumulation by intravenous co-injection of SSR180711, a selective α7-nAChR partial agonist (doses 0.02 mg/kg, 0.2 mg/kg, 1 mg/kg, 3 mg/kg) in the CD-1 mouse brain regions 90 min after the injection. *P ≤ 0.01, significantly different from controls (ANOVA). Data: mean %injected dose/g tissue ± SD (n=3). Abbreviations: Coll = superior and inferior colliculus; Hipp = hippocampus; Ctx = cortex; Str = striatum; Th = thalamus; Rest = rest of brain; CB = cerebellum. Reprinted from [48] with permission of Copyright Clearance Center's RightsLink service

One of the main characteristics of PET radiotracers is binding potential (BP_ND), which is the ratio of specific-to-nonspecific binding [67]. Good PET tracers are expected to demonstrate BP_ND>1. The binding potential values of [¹⁸F]ASEM in the mouse α7-nAChR – rich regions cortex (5.3), hippocampus, (5.5) and colliculus (8.0) (Table 3)[48] were high and sufficient for the receptor quantification.

Table 3.

Approximate binding potential (BP_ND) values (unitless) of [¹⁸F]ASEM and para-[¹⁸F]ASEM in the mouse brain regions. Data: mean ± SD (n = 6) [48].

Compound	Region	Superior & inferior colliculus	Hippocampus	Cortex
[18F]ASEM		8.0 ± 1.6	5.5 ± 1.7	5.3 ± 1.2
para-[18F]ASEM		2.0 ± 0.5	3.1 ± 0.7	2.0 ± 0.3

The α7-nAChR in vivo selectivity of [¹⁸F]ASEM binding in the mouse brain was tested by the blocking experiments with several non-α7-nAChR CNS drugs (Fig. 8). The drugs were ondansetron (selective 5-HT₃ antagonist), SCH23390 (D₁- and D₅-antagonist and 5-HT_1C/2C agonist), ketanserin (5-HT₂/5-HT_2C antagonist), naltrindole (selective δ-opioid antagonist), cytisine (α4β2-nAChR-selective partial agonist). None of these drugs except the positive control SSR180711 reduced accumulation of [¹⁸F]ASEM radioactivity when compared to the baseline controls[48]. The absence of blockade with the α4β2-nAChR-selective cytisine and 5-HT₃-selective ondansetron was especially remarkable because α7-nAChR ligands often are not selective and bind at these receptors.

Fig. 8.

Effect of various CNS drugs (2 mg/kg, s.c.) on accumulation of [¹⁸F]ASEM in CD-1 mouse brain regions 90 min after injection of tracer expressed as %ID/g tissue. Abbreviations: Coll = superior and inferior colliculus; Hipp = hippocampus; Ctx = cortex; CB = cerebellum; REST = rest of brain. Data are mean ± SD (n=3). *P < 0.01, significantly different from controls. Columns that do not include the asterisk are insignificantly different from controls (P > 0.01) (ANOVA, single-factor analysis). The graph demonstrates that unlike the positive control (SSR180711) all non-α7-nAChR CNS drugs do not have an effect on the cerebral uptake of [¹⁸F]ASEM that is α7-nAChR selective in vivo. Drugs: SSR180711 – selective a7-nAChR partial agonist; Ondansetron - selective 5-HT₃ antagonist; SCH23390 -D₁- and D₅-antagonist and 5- HT_1C/2C agonist; Ketanserin - 5-HT₂/5-HT_2C antagonist; Naltrindole - Selective δ-opioid antagonist; Cytisine – α4β2-nAChR agonist [48].

These studies in CD1 mice showed that [¹⁸F]ASEM labels α7-nAChR receptors in the mouse brain with high degree of specificity and selectivity [48] (Figs. 6-8). In the opinion of the [¹⁸F]ASEM developers the high specific binding of this radiotracer is chiefly attributed to its superior binding affinity versus the previous α7-nAChR radiotracers. This is supported by the correlation of the binding affinity and binding potential of [¹⁸F]ASEM and other radiotracers (Fig. 9).

Fig. 9.

Correlation of the BP_ND^cortex (unitless) vs. 1/K_i (nM⁻¹) of α7-nAChR PET radioligands [¹⁸F]AZ11637326, [¹¹C]NS14492, [¹⁸F]NS10743, [¹⁸F]para-ASEM and [¹⁸F]ASEM (y = 1.91x + 0.52, R² = 0.98). All K_i values were obtained under the same binding assay conditions. Reprinted from [48] with permission of Copyright Clearance Center's RightsLink service

In a parallel set of experiments the para-[¹⁸F]ASEM was also tested in CD1 mice. The pattern of regional distribution of para-[¹⁸F]ASEM was comparable to that of [¹⁸F]ASEM. However, in agreement with its lower binding affinity para-[¹⁸F]ASEM exhibited lower specific binding in mice than [¹⁸F]ASEM (Tables 1 and 3).

4.2 Radiometabolite analysis of [¹⁸F]ASEM in blood and brain

Most PET radiotracers undergo metabolism and generate various radiometabolites that, ideally, do not accumulate in the brain and spoil the quality of the PET image and thus reduce the accuracy of the quantification of the targeted receptor. Conventionally, the brain radiometabolites are considered to be insignificant for accurate quantification if the fraction of the parent radiotracer in the brain is greater than 95%.

The HPLC analysis of blood samples from CD-1 mice, and, also, from baboons[59] and human subjects[68] demonstrated that the parent compound [¹⁸F]ASEM was gradually metabolized to the same hydrophilic radiometabolites in all three species. Luckily, only a small fraction of [¹⁸F]ASEM radiometabolites penetrates the blood-brain barrier and the main radioactive compound in the brain is the parent [¹⁸F]ASEM (>95%), as it is shown by the HPLC analysis of the mouse brain tissue. Because the radiometabolite fraction in the brain tissue is insignificant, mathematical PET modeling of the radiometabolites is not necessary for quantification of α7-nAChRs with [¹⁸F]ASEM [59].

4.3 Clinical α7-nAChR drugs block the [¹⁸F]ASEM binding in the mouse brain.

Several drugs that target α7-nAChRs are now in the clinical phases of development for treatment of cognitive deficit in various pathologies [27-29]. DMXB-A was the first selective α7-nAChR agonist to demonstrate cognitive enhancement and improvement in negative symptoms in patients with schizophrenia[69, 70]. The more recent drug EVP-6124 (Encenicline) is an α7-nAChR selective partial agonist that is now in a number of clinical trials for treatment of Alzheimer’s disease and schizophrenia (see https://clinicaltrials.gov). It was of interest to test the blocking effect of the clinical doses of the α7-nAChR drugs in the PET experiments with [¹⁸F]ASEM.

DMXB-A dose-dependently blocked the [¹⁸F]ASEM binding in the α7-nAChR - rich brain regions in mice (Fig. 10a)[68]. The blocking effect was significant when a clinical equivalent dose was used. This result demonstrates the potential feasibility for evaluating the effect of the clinical drug DMXB-A in human subjects with [¹⁸F]ASEM, and opens new ways to study the biochemical mechanism of drugs for treatment of cognitive performance in patients with schizophrenia. In addition to DMXB-A, similar [¹⁸F]ASEM blocking studies were performed with two other nicotinic drugs [68] in clinical trials that bind at α7-nAChR, EVP-6124[71] and Varenicline (binds at α4β2- and α7-nAChRs [72]) (Figs. 10b,c).

Fig. 10.

Baseline versus blockade studies of [¹⁸F]ASEM with mouse-equivalent doses of clinical α7-nAChR drugs in CD-1 mice. Data: %ID/g tissue ± SD (n = 4). The control mice were treated with vehicle saline. Abbreviations: CB = cerebellum, Hipp = hippocampus; Ctx = cortex. Statistics in all three graphs: *P < 0.01, blockade is significantly different from controls (ANOVA). a: DMXB-A (GTS-21), dose-escalation. Note: a mouse-equivalent dose = 25 mg/kg[81] of the clinical dose (150 mg). 90 min post [¹⁸F]ASEM injection. b: EVP-6124, a mouse-equivalent dose (0.18 mg/kg) of the clinical dose (1 mg). 60 min post [¹⁸F]ASEM injection. c: Varenicline, a mouse-equivalent dose (0.18 mg/kg) of the clinical dose (1 mg). 60 min post [¹⁸F]ASEM injection. The graph demonstrates that in vivo binding of [¹⁸F]ASEM in the mouse brain regions enriched with α7-nAChR is significantly blocked by the α7-nAChR drugs DMXB-A, EVP-6124 and varenicline. Reprinted from[68] with permission of Copyright Clearance Center's RightsLink service.

It is noteworthy, that clinical dose equivalents of α7-nAChR drugs EVP-6124 and DMXB-A only partially blocked the binding of [¹⁸F]ASEM in the mouse brain (23-28% and 45-55%, respectively) (Fig. 10a,b). This degree of blockade is comparable to the currently accepted degree of α7-nAChR occupancy required to achieve clinical or behavioral efficacy. The α7-nAChR receptors have five binding sites distributed between five α7 subunits. When an agonist is applied to a population of α7-nAChRs, the maximum α7-nAChR receptor activation may occur when two of the five possible binding sites are occupied[73, 74]. At higher concentrations, agonists desensitize the α7-nAChR and may bind to all five binding sites. In agreement with this mechanism, pre-clinical studies and clinical trials suggest that efficacious concentrations of EVP-6124 (Encenicline) is low and sufficient to occupy only one binding site on the α7-nAChRs [71, 75]. It was hypothesized that endogenous acetylcholine (ACh) is required to bind to another site on the α7-nAChR in order to activate the receptor channel opening. Because ACh exhibits low α7-nAChRs binding affinity the complex ACh*α7-nAChR quickly dissociates after the receptor channel opening and, thus, α7-nAChRs are activated by low clinical doses of EVP-6124, but are not desensitized.

This mechanism of action agrees with the observed degree of [¹⁸F]ASEM blockade with EVP-6124 and DMXB-A and suggests that in the future human PET studies with clinically efficacious doses of α7-nAChR agonists the binding of [¹⁸F]ASEM will not be blocked more than 20% - 40%.

4.4 Distribution of [¹⁸F]ASEM in DISC1 mice, a rodent model of schizophrenia.

Post-mortem research demonstrated significantly lower density of α7-nAChRs in the brains of schizophrenia subjects vs. controls (see Introduction for references). Mutant DISC1 mice provide a model for brain and behavioral phenotypes seen in schizophrenia[76]. Based on the favorable imaging properties identified in control CD1 mice, the [¹⁸F]ASEM binding was investigated in DISC1 mice. In agreement with the reduced density of α7-nAChRs in the brain tissue of schizophrenic subjects, the [¹⁸F]ASEM binding in the α7-nAChR – rich brain regions was significantly lower in the DISC1 mice when compared to control animals (data not shown, see paper[59]). This result emphasizes the potential utility of [¹⁸F]ASEM for imaging of α7-nAChRs in schizophrenia.

5. PET-[¹⁸F]ASEM imaging of α7-nAChRs in baboons.

The successful brain distribution studies of [¹⁸F]ASEM in rodents (see above) provided a foundation for further pre-clinical evaluation in baboons [59]. The main purpose of these studies was to determine if (1) [¹⁸F]ASEM exhibits the brain regional distribution that matches the distribution of α7-nAChRs in non-human primates; (2) the binding is α7-nAChR specific; (3) the brain pharmacokinetics of [¹⁸F]ASEM are suitable for quantification of α7-nAChRs and (4) to evaluate the PET imaging characteristics of [¹⁸F]ASEM by mathematical modeling methods.

In PET baboon experiments [¹⁸F]ASEM exhibited high and reversible brain uptake that peaked (500% standardized uptake value (%SUV)) at 20 min post injection of the radiotracer (Fig. 11)[59]. The distribution pattern of the [¹⁸F]ASEM radioactivity was heterogeneous (thalamus > insula > anterior cingulate cortex > putamen > hippocampus > cortical regions > pons ~ cerebellum ~ corpus callosum) and consistent with previously published in vitro distribution of α7-nAChRs in non-human primates [40, 41, 77].

Fig. 11.

Baseline cerebral time-activity curves (TACs) after bolus administration of [¹⁸F]ASEM in three baboons. The graph demonstrates a substantial heterogeneous brain uptake of [¹⁸F]ASEM that matches the distribution of α7-nAChR in non-human primates[40, 41, 77] and reversible brain kinetics. Data: mean Standardized Uptake Values (%SUV) ± SD (n = 3). Reprinted from[59] with permission of SNMMI.

The α7-nAChR specificity of the [¹⁸F]ASEM binding in the baboon brain was established in the blocking PET experiments. The dose-dependent blockade of [¹⁸F]ASEM uptake with selective α7-nAChR partial agonist SSR180711 demonstrated that the binding of [¹⁸F]ASEM is mediated by α7-nAChR (Fig. 12).

Fig. 12.

Sagittal (Top row), and trans-axial (Middle and Bottom rows) views of V_T images of [¹⁸F]ASEM in the same baboon for a baseline PET scan (B), and after administration of 0.5 mg/kg (C) and 5 mg/kg (D) of SSR180711, a selective α7-nAChR partial agonist. MR images (A) indicate locations of selected brain structures including the cingulate cortex (Cg), thalamus (Th), and caudate nucleus (CN), which are indicated by + in the V_T images (D). The V_T images are displayed using the same minimum and maximum values for all scanning conditions. These data demonstrate the dose dependent blockade of [¹⁸F]ASEM in baboon brain and provide evidence that [¹⁸F]ASEM is specific and mediated by α7-nAChR. The images also suggest that there is no reference region devoid of α7-nAChRs. Reprinted from[59] with permission of SNMMI.

The baseline and blocking PET studies allowed calculation of one of the main PET imaging characteristics of [¹⁸F]ASEM in baboon, the binding potential (BP_ND = 3.9 – 6.6)[59]. The value of BP_ND is quite high for PET and suitable for reliable quantification of α7-nAChRs. For comparison, all previous α7-nAChR radioligands exhibited BP_ND < 1[35, 46-48].

6. First-in-human PET - [¹⁸F]ASEM imaging of α7-nAChRs[68].

Previously, the lack of a specific PET radioligand has impeded the accurate mapping and quantification of α7-nAChRs in the living human brain. In the first-in-human PET studies in five healthy male subjects under an IND, [¹⁸F]ASEM radioactivity peaked in the brain at 20 min after bolus injection with a robust uptake value of 400 %SUV (Fig. 13). The brain pharmacokinetics were reversible and demonstrated a decline of radioactivity concentration after the peak[68]. The highest regional brain uptake (total volume of distribution V_T, calculated with plasma reference graphic analysis) was seen in the parietal cortex (V_T=22±1.8) ~ putamen (21.8±2.9) > thalamus (20.9±3.0) > cingulate (19.6±2.2) > temporal lobes (19.7±1.8) ~ frontal lobes (19.3±1.9) > hippocampus (V_T=17.9±1.9); moderate uptake was found in the cerebellum (15.1±1.6) and brainstem (14.8±1.7); and the lowest uptake was in the corpus callosum (9.9±2.0). The regional distribution of [¹⁸F]ASEM in the human brain matches the post-mortem human and non-human primate data [59, 78, 79]. The test-retest variability (TRV) was 10.8±5.1%, which charcterises [¹⁸F]ASEM as a reproducible PET radiotracer.

Fig. 13.

Baseline PET/[¹⁸F]ASEM TAC’s (%SUV ± SD (n = 5)) in healthy human males. Abbreviations: Pu = putamen; Pr = precuneous; Pa = parietal lobe; Th = thalamus; Fr = frontal lobe; Cg = cingulate; Oc = occipital; Tp = temporal lobe; Hp = hippocampus; CN = caudate nucleus; Cb = cerebellum; CC = corpus callosum. The distribution of [¹⁸F]ASEM in the human brain regions is comparable with non-human primate (see for review [34]) and human post-mortem distribution of α7-nAChR [47-48]. The brain kinetics of [¹⁸F]ASEM are reversible. Reprinted from[68] with permission of Copyright Clearance Center's RightsLink service.

7. Conclusion

In summary, during the past decade we have witnessed substantial efforts to develop a PET radioligand for quantification of α7-nAChRs in the human brain. Several research groups have radiolabeled a number of α7-nAChR compounds with [¹⁸F] and [¹¹C] for PET. Unfortunately, these radioligands did not show suitable specific binding in vivo due to insufficient α7-nAChR binding affinity. The recently developed [¹⁸F]ASEM, a highly α7-nAChR specific and selective radiotracer for brain PET, demonstrated excellent in vivo imaging properties in the rodents and baboons and was successfully translated to human subjects. [¹⁸F]ASEM opens new horizons for studying α7-nAChRs in the living human brain.

Fig. 14.

Averaged (n = 5) transaxial images of a spatially normalized V_T map of [¹⁸F]ASEM and matching MRI in healthy control subjects. Cerebellum (Cb) and medial temporal cortex (mdT; panel a) show relatively low V_T values and hippocampus (Hp; panel b) show medium V_T values. The insula (In), putamen (Pu), and thalamus (Th) are shown in panel c, and middle frontal (mFC), parietal (PC), and occipital (OC) cortices (panel d) exhibit high V_T values in the human brain. Red dots on MRI images indicate outlines of cortical and subcortical structures. Reprinted from[68] with permission of Copyright Clearance Center's RightsLink service.