5-HT Radioligands for Human Brain Imaging With PET and SPECT

Abstract

The serotonergic system plays a key modulatory role in the brain and is the target for many drug treatments for brain disorders either through reuptake blockade or via interactions at the 14 subtypes of 5-HT receptors. This review provides the history and current status of radioligands used for positron emission tomography (PET) and single photon emission computerized tomography (SPECT) imaging of human brain serotonin (5-HT) receptors, the 5-HT transporter (SERT), and 5-HT synthesis rate. Currently available radioligands for in vivo brain imaging of the 5-HT system in humans include antagonists for the 5-HT_1A, 5-HT_1B, 5-HT_2A, and 5-HT₄ receptors, and for SERT. Here we describe the evolution of these radioligands, along with the attempts made to develop radioligands for additional serotonergic targets. We describe the properties needed for a radioligand to become successful and the main caveats. The success of a PET or SPECT radioligand can ultimately be assessed by its frequency of use, its utility in humans, and the number of research sites using it relative to its invention date, and so these aspects are also covered. In conclusion, the development of PET and SPECT radioligands to image serotonergic targets is of high interest, and successful evaluation in humans is leading to invaluable insight into normal and abnormal brain function, emphasizing the need for continued development of both SPECT and PET radioligands for human brain imaging.

1. INTRODUCTION

The serotonergic (5-HT) system plays a key modulatory role in many central nervous system (CNS) functions. Multiple serotonergic receptor subtypes and region-specific innervations from the dorsal and medial raphe nuclei within the brainstem result in a complex pattern of modulatory control over various physiological, emotional, and cognitive processes. These processes include control of mood and sleep, regulation of cognitive performance, learning and memory, modulation of ingestive behavior, and influence on reward circuits that mediate, for example, motivation.^1–4 Serotonergic dysfunction has been implicated in the etiology of many psychiatric disorders including depression,⁵ anxiety,⁶ schizophrenia,⁷ and other neurological disorders such as Alzheimer’s disease,^8,9 and epilepsy.¹⁰ Although our basic understandings of the serotonergic system are derived from animal models, the development of noninvasive brain imaging techniques, such as positron emission tomography (PET) and single positron emission computerized tomography (SPECT), increasingly allows the study of the serotonergic system in the human brain.¹¹

This review covers the history and current status of the development of PET and SPECT radioligands for imaging serotonergic targets within the brain. First, we briefly introduce the serotonergic system and the uses of PET and SPECT imaging in general. A more detailed description of the different radioligands follows.

A. 5-HT Targets for PET and SPECT

The 5-HT receptors are among the most diverse group of neurotransmitter receptors in the human genome and the 5-HT system is also one of the phylogenetically oldest systems. Currently, 14 structurally and pharmacologically distinct mammalian 5-HT receptor subtypes have been described. Based on their structure, affinity for different ligands, and second messenger pathway, they are assigned to one of the seven families, 5-HT_1–7.¹² All 5-HT receptors, except the 5-HT₃ receptor, are G-protein coupled seven transmembrane spanning receptors (GPCRs). The 5-HT₃ receptor is a ligand-gated sodium ion channel. In addition, the 5-HT transporter (SERT) responsible for 5-HT reuptake and 5-HT synthetic enzymes, especially tryptophan hydroxylase, are also targets for tracer development. Even though much still needs to be elucidated, it is known that each of these targets has its own distinct pattern of distribution and function in the brain.^1,13–15 The 5-HT targets and their characteristics are summarized in Table I.

Table I.

5-HT Receptor Properties and Available Selective Radioligands

Target	Nomenclature (transduction)		Highest Bmax fmol/mg protein (region)	Selective ligands	Selective radioligand	PET radioligand
GPCRs	5-HT1A	(Gi/o)	860–3,140* (hipp)63 300–1,9103 (ctx)	✔	✔	✔
	5-HT1B	(Gi/o)	400–500* (g pall)141 3963 (occ ctx)141	✔	✔	✔
	5-HT1D	(Gi/o)	226 (g pall)144	✔	(✔)	–
	5-ht1e	(Gi/o)	224 (put)144	–	–	–
	5-HT1F	(Gi/o)	79 (rat wb)145	✔	✔	–
	5-HT2A	(Gq/11)	570 (ctx)149	✔	✔	✔
	5-HT2B	(Gq/11)	ND	✔	–	–
	5-HT2C	(Gq/11)	688 (pig ch plx)264	✔	✔	–
	5-HT4	(Gs)	223 (stri)287	✔	✔	✔
	5-ht5a	(?)	ND	✔	–	–
	5-ht5b	(?)	ND	–	–	–
	5-HT6	(Gs)	215 (stri)303	✔	✔	✔
	5-HT7	(Gs)	68 (thal)310	✔	✔	–
Ion channel	5-HT3	–	465 (pig ctx)271	✔	✔	–
Transporter	SERT	–	587 (ctx)322 682 (thal)320	✔	✔	✔

Receptor nomenclature is agreed by IUPHAR (International Union of Basic and Clinical Pharmacology). The term “receptor” is only applied to entities for which operational, structural, and signal transduction information is available; thus 5-ht_1E, 5-ht_5A, and 5-ht_5B have lower case letters indicating that no function has yet been attributed to them. B_max values are taken from the literature (references provided as numerical superscript) and are from human brain unless otherwise stated. Units are equivalent fmol/mg protein; * denotes a simple conversion from fmol/mg tissue assuming that10% of tissue weight is protein. N.B. B_max values vary between tracers, species, and laboratories but only one such example is provided here. Hipp, hippocampus; ctx, cortex; occ, occipital; g pall, globus pallidus; stri, striatum; put, putamen; wb, whole brain; ch plx, choroid plexus; thal, thalamus; ND, not determined. ✔ indicates that a ligand is available. -- indicates that no ligand is available. (✔) Ligands also have 5-HT_1B affinity. ✔ in PET probe column indicates a PET tracer is available that shows promise in human subjects.

2. IN VIVO BRAIN IMAGING

In vivo brain imaging is an important and widely used tool for the study of the living brain. For in vivo pharmacology, PET and SPECT are most often used for quantification of brain receptor concentrations, and since the first PET study was published in 1983,¹⁶ this neuro-receptor mapping technique has been used to study various neurotransmitter systems in health and disease.^17–20 In brain, PET has also been widely used to study glucose metabolism,^21–24 blood–brain barrier transport,^25–28 and neurotransmitter release.^29,30 PET and SPECT imaging are also useful for tracking pharmacokinetic and pharmacodynamic properties of drugs, and can be used to determine the occupancy of therapeutic drugs, which ultimately can be used to estimate the optimal doses to be used in Phase II studies. For example, it has been demonstrated that between 60 and 80% occupancy of D₂ receptors is required for anti-psychotic medication efficacy and that beyond this level, side effects are likely to occur.³¹ Further, treatment with clinically effective doses of the selective serotonin reuptake inhibitors (SSRIs) paroxetine or citalopram is associated with approximately 80% occupancy of SERT.³² For that reason, the pharmaceutical industry has increasingly taken advantage of these techniques to determine doses of new agents. SPECT is a more widely accessible functional imaging tool than PET, due to its lower costs and use of radioligands with longer half-lives. In spite of its lower sensitivity and resolution as compared to PET, SPECT is widely used for brain imaging, and radioligands suitable for this technique are in great demand.

Significant discoveries within the 5-HT system in human brain have been made following the development of selective PET and SPECT radioligands, for example, that the 5-HT_1A receptor is reduced in social anxiety and panic disorder^33,34 whereas the 5-HT_2A receptor may be increased in depressed individuals following treatment.³⁵ However, the full potential of 5-HT-related treatments cannot be realized before appropriate radioligands selective for all target receptors and enzymes have been developed and validated, including those that will allow endogenous 5-HT release and synaptic levels to be measured.³⁰

A. The Principles of PET and SPECT

Radioisotopes that are typically used include ¹¹C, ¹³N, ¹⁵O, ¹⁸F, and ⁷⁶Br in PET and ^99mTc and ¹²³I in SPECT. Permissive of the molecular structure, these can be incorporated into potential tracers of interest. Radioligands for imaging brain neurotransmitter receptors and transporters are based structurally on receptor antagonists or agonists. Since the radioligands are targeted to work when administered in tracer doses (typically below 5 μg per injection and with receptor occupancies of less than 5%), they should rarely elicit pharmacological effects.

The vast majority of PET and SPECT radioligands available today are antagonists; there are many more molecules and chemical series of antagonists available from pharmaceutical drug discovery programs and they tend to be easier to develop. For GPCRs, antagonist radiotracers label the whole population of receptors with the same affinity, thus providing a good indication of total receptor number and higher B_max, whereas agonist radioligands preferentially label the high-affinity states of the receptor that are capable of eliciting signalling events. Compared with the total number of receptors, the distribution of receptors in the high-affinity state is predicted to be a more precise outcome in studies with a functional aim³⁶ or studies attempting to image the release or depletion of endogenous agonist neu-rotransmitter.^37,38 Agonist radioligands should also provide a better measure of occupancy for new agonist pharmacological agents or any agent at the “functional” proportion of the receptor. Because of these considerations, the interest in agonist radioligands is increasing.

B. Overview of Successful Radioligand Properties

In order to generate quantifiable images of cerebral binding, radioligands must possess certain qualities. Radioligands suitable for in vitro quantification by competition, saturation, or autoradiographical methodologies cannot necessarily be used in vivo. This is mainly because of the need to penetrate the blood–brain barrier and not be excluded by pumps such as P-glycoprotein (P-gp), but also due to problems arising from high non-specific binding or issues with radiosynthesis, metabolism, or pharmacokinetics. The main properties necessary for a radioligand to be useful for in vivo imaging are as follows.

Radiosynthesis should be as simple and fast as possible, especially with short-lived ¹¹C, giving a high specific radioactivity to avoid any significant pharmacological blockade of the receptor in question with “cold” compound from the synthesis.

The radioligand must possess high affinity toward the target receptor, normally in the nanomolar range, to maximize the target to background ratio. On the other hand, to achieve suitable reversible receptor kinetics, the affinity must not be too high. The association rate should be high, and the dissociation rate should be long enough to be measured, but short enough for the washout phase to be appreciable within the time frame of scanning.

The radioligand must bind selectively to the target of interest. To obtain meaningful data for target distribution studies, the binding potential measured by PET or SPECT using the radioligand must be obtained from binding to the desired target only. Since the binding potential is the product of the number of receptors available for binding (B_avail) and the affinity (1/K_D), these two factors must be considered when determining the selectivity of a radioligand. Often the affinity of a radioligand, as 1/K_i, toward the desired target is determined in a competition assay against a well-known ligand. This is then compared with the affinities of the ligand toward a wide spectrum of other receptors and binding sites. K_i values toward other targets that are more than 20–100 times higher than for the desired target are normally acceptable. Acceptability, however, depends on B_avail for the binding sites in question. If B_avail for the desired target is much higher than B_avail for non-targets, then the binding potential will still selectively reflect the wanted target, and vice versa if a non-target binding site is very abundant, since then even low affinity to this non-target will influence the total signal. In order to address the selectivity of a radioligand one can perform an in vivo blocking study, where the radioligand is co-administered with a known selective antagonist that binds only to the target.

Generally, the radioligand should be moderately lipophilic to achieve adequate penetration of the blood–brain barrier without incurring excessive non-specific binding to brain tissue or very slow brain clearance. Moderate lipophilicity does not, however, guarantee brain entry. For example, some moderately lipophilic compounds are substrates for efflux transporters at the blood–brain barrier, such as P-gp, and these compounds are effectively excluded from entry into brain.³⁹

Plasma clearance rate of radioligand should also be considered. Rapid clearance results in difficulties with accurate determination of the input curve, particularly at late time points and this complicates the subsequent mathematical modelling with an arterial input function.

The metabolites produced should either be polar, so excluded from the brain, or unlabelled, so as not to interfere with the signal obtained in brain.³⁹ Interference from lipophilic radiometabolites can, however, be taken into account if it is possible to establish a bolus–infusion design (e.g.,⁴⁰) or if the radiometabolites do not have specific binding in the brain.

One thing that must be considered when designing and synthesizing potential PET radioligands is the likelihood of being able to label them successfully without interfering with their pharmacology. ¹¹C-labelling is usually feasible because of the presence of carbon atoms in all organic compounds. The introduction of the small electronegative ¹⁸F atom in place of hydrogen or hydroxyl can be unpredictably beneficial or detrimental to the pharmacological properties of the ligand in terms of its affinity and selectivity. Introduction of a bulky¹²³I atom in place of hydrogen generally affects pharmacology and will increase ligand lipophilicity. The half-life of the radionuclide is an important consideration. The relatively short half-life of ¹¹C (t_1/2 = 20.4 min) is suitable for following quite rapid pharmacokinetics, and may permit more than one study session in the same subject in a single day. However, the necessity to produce ¹¹C-labelled radioligands on-site is a logistical disadvantage. Longer-lived ¹⁸F (t_1/2 = 109.8 min) for PET and ¹²³I (t_1/2 = 13 hr) for SPECT do not require on-site production and are suitable for following slower pharmacokinetics. Alternative labels are being investigated, such as the SPECT label ^99mTc (t_1/2 = 6 hr), because of its relative safety and availability.

3. CURRENT RADIOLIGANDS FOR IMAGING THE 5-HT SYSTEM

A multitude of radioligands exist for in vitro studies of serotonergic targets, and over the last decade, we have seen an impressive increase in the number of useful PET and SPECT radioligands. With this review, we describe the PET and SPECT radioligands published to date for imaging the serotonergic system. Table I summarizes the 5-HT targets, their brain distribution, and density and availability of selective ligands. Table II summarizes the development history of all published PET and SPECT ligands for those targets where tracer development has been undertaken. In addition, since the relative success of a radioligand can be estimated through its frequency of use and publication rate, these data are given in Table III.

Table II.

Development of PET and SPECT Radioligands for the 5-HT System Designed for Use in Human Subjects

Target	Type	Radioligand	Rodent	Nonhuman primate	Human	Ongoing studies	Reason for failure	Known problems
5-HT1A	SPECT	[123I]p-MPPI	✔	✔	×	–	No specific 5-HT1A binding Rapid metabolism Possible brain exclusion via efflux transporter	–
	PET	[11C]WAY-100635	✔	✔	✔	✔	–	Fast systemic metabolism; difficult kinetic modelling with arterial input function, reference tissue models complicated by radiometabolites in plasma
		[11C]CPC-222	✔	ND	✔	ND		Lower signal-to-background ratio than [11C]WAY-100635
		[11C](R)-RWAY	×	✔	×	–	Possible influx of lipophilic radiometabolite	P-gp substrate in rodent
		[11C]DWAY	✔	✔	✔	ND		Unreliable radiolabelling technique; low radioactive yield
		[18F]6FPWAY	ND	✔	×	–	Moderate uptake	–
		[18F]MPPF	✔	✔	✔	✔	–	P-gp substrate; fast brain clearance and low uptake
		[18F]FCWAY	✔	✔	✔	✔	–	Defluorination of parent compound; sub-optimal imaging
		[18F]MefWAY	ND	✔	ND	–
		[11C]NAD-299	ND	✔	ND	–
		[11C]CUMI-101	✔	✔	✔		–	Partial, not full agonist
5-HT1B	PET	[11C]AZ10419369	ND	✔	✔		–
		[11C]P943	ND	✔	✔	✔	–
5-HT2A	SPECT	[123I]DOI	×	×	–	–	Low target-to-background ratio	Non-selective (5-HT2C)
		[123I]MSP	✔	ND	ND	ND
		[123I]-R91150	✔	✔	✔	✔	–	Lower target-to-background ratio than equivalent PET tracers
		[123I]-3-I-CO	×	ND	–	–	–	Low target-to-background ratio, possible P-gp substrate
	PET	[11C]Ketanserin	ND	ND	×	–	Low target-to-background ratio, fast metabolism	Non-selective (α1, H1, 5-HT2C)
		[11C]NMSP	ND	ND	✔	–	–	Non-selective (D2)
		[11C]MBL	ND	✔	✔	–		Non-selective (some D2, α1, 5-HT1 and 5-HT2C)
		[18F]setoperone	✔	✔	✔	✔	–	Non-selective (D2)
		[18F]altanserin	✔	✔	✔	✔	–	Need for bolus-infusion due to radiometabolites requiring complex modelling using other methods. Possible mixed pharmacology
		[18F]deuteroaltanserin	ND	✔	✔	✔	–
		[18F]RP62203	✔	ND	–	–		Multi-step radiosynthesis
		[11C]MDL100907	✔	✔	✔	✔	–	Arterial input function needed
		[18F]MH.MZ	×	ND	ND	–	Extensive first-pass metabolism, slow washout	-
		(R)-[18F]MH.MZ	✔	ND	ND
		[11C]Cimbi-5 and 36	✔	✔ (pig)	ND	ND
5-HT3	PET	[11C]MDL 72222	×	×	–	–	Lack of specific binding, high lipophilicity	Low 5-HT3 brain density
		[11C]YM060, [11C]Y-25130	×	–	–	–	Low brain uptake, low lipophilicity, ionisation of tertiary amide
		[11C]KF17643	×	–	–	–	No specific binding
		[11C]S21007	×	×	–	–	No specific binding
		[18F]MR18445	×	×	–	–	No specific binding
		[11C]NMQ	×	×	–	–	Rapid kinetics, high non-specific binding
5-HT4	SPECT	[123I]SB207710	×	✔	ND	–	–
	PET	[11C]SB207145	✔	✔ (pig)	✔	✔	–	Slow kinetics
5-HT6	PET	[18F]12ST05	×	–	–	–	No specific binding	–
		[11C]GSK215083	ND	✔ (pig)	✔
		[11C]GSK224558	ND	(pig only)	ND	–		Inferior ligand compared with [11C]GSK215083
5-HT7	PET	[11C]DR4446	ND	×	–	–	Minimal specific binding	–
SERT	SPECT	β-[123I]CIT, nor-β-[123I]CIT	✔	✔	✔	✔	–	Non-selective (DAT, NET)
		[123I]ADAM	✔	✔	✔	✔	–	Lower resolution than equivalent PET tracers. Ratio method of quantification over-estimates specific binding
	PET	[11C]McN5652	✔	✔	✔	✔	–	Slow brain uptake, irreversible kinetics complicate quantification
		[11C]DASB	✔	✔	✔	✔	–
		[11C]MADAM	ND	✔	✔	✔	–
		[18F]ADAM	✔	✔	ND	–	–
5-HT synthesis		[11C]-AMT	✔	✔	✔	✔	–	Low target-to-background ratio, analysis complicated by possible blood-brain barrier exchange and additional metabolic pathways
		[11C]-HTP	✔	✔	✔	✔	–	Lack of correlation with [11C]-AMT in direct comparison suggests further validation is necessary

✔ indicates that ex vivo, PET, or SPECT studies were performed and the ligand was considered successful in the species identified. In certain instances, radioligands were tested in pig brain either in addition to, or instead of nonhuman primate. × indicates that ex vivo, PET, or SPECT studies were performed but the ligand was not considered successful in the species indicated. “Ongoing studies” indicate that the ligand is still being used for research or clinical purposes in humans. ND indicates not yet determined, i.e. experiments have not been published in the species indicated. Missing data are either not applicable (−) or not available/not known (blank cells).

Table III.

Successful and Promising PET and SPECT Radioligands for Serotonergic Targets in Humans

Target	Radioligand	First in man	Animal studies	Human studies	Research institutions
5-HT1A	[11C]WAY-100635	1995	10	80	12
	[18F]MPPF	2000	27	21	6
	[18F]FCWAY	2000	7	11	1
	[11C]CUMI-101	2008a	2	1a	1
5-HT1B	[11C]AZ10419369	2008	3	1	1
	[11C]P943	2009	1	3	1
5-HT2A	[123I]-5-I-R91150	1997	9	19	6
	[18F]setoperoneb	1990	3	36	7
	[18F]altanserin	1994	5	51	10
	[18F]deuteroaltanserin	1998	2a	5a	1
	[11C]MDL100907	1998	5	6	5
5-HT4	[11C]SB207145	2008	3a	3a	2
5-HT6	[11C]GSK215083	2008	1	2	1
SERT	Beta-[123I]CITb	1993	20	87	18
	[123I]ADAM	2005	11	28	12
	[11C]DASB	2000	17	48	16
	[11C]MADAM	2005	1	7	2
5-HT synthesis	[11C]-AMT	1997	2c	17d	2
	[11C]-HTP	1991	4	10	3

“First in man” indicates the year when the tracer was first tested inhuman subjects as determined from publication of full papers or conference proceedings. “Animal studies” indicate the number of published manuscripts that include data from any animal species other than humans; these include ex vivo, SPECT, and/or PET studies in rat, cat, pig, dog, or nonhuman primate. “Human studies” indicate the number of PET or SPECT published manuscripts.

^aPublications may also include conference proceedings. “Research institutions” indicates the number of centers using the tracer in human subjects. The number of publications and research institutions indicated are correct at the time of writing (July 2010).

^bTracer has multiple binding sites but only those reporting on the serotonergic target of interest are included.

^cMany animal studies have been performed with the carbon-14 equivalent of this tracer but these are not listed here.

^dMajority of studies are measuring kynurenine metabolism in tumors rather than 5-HTsynthesis per se.

A. 5-HT_1A Receptor Radioligands

The 5-HT_1A receptor is one of the best characterized receptors in the serotonergic family. This is especially due to its role as an inhibitory autoreceptor in the raphe nuclei and the possible implications of this role for the treatment of depression and anxiety with serotonin reuptake inhibitors and the potential for the treatment of learning and memory deficits.^4,41,42

A number of reviews have detailed PET radioligands available for the study of 5-HT_1A receptors,^43,44 and many have been summarized quite recently.⁴⁵ Therefore, not all radioligands will be described in detail here. Many of the 5-HT_1A radioligands were based on WAY-100635 (N-[2-[4-(2-methoxyphenyl)piperazin-1-yl]ethyl]-N-pyridin-2-ylcyclohexane-carboxamide), which in its carbonyl-¹¹C-labelled form is widely used for 5-HT_1A receptor imaging. In total, 39 potential PET 5-HT_1A radioligands are described in Kumar and Mann,⁴⁵ but to date, only two of these ligands are in frequent use for determining 5-HT_1A receptor densities; [carbonyl-¹¹C]WAY-100635 and [¹⁸F]MPPF (4-[¹⁸F]fluoranyl-N-[2-[4-(2-methoxyphenyl)piper-azin-1-yl]ethyl]-N-pyridin-2-ylbenzamide). [¹⁸F]MefWAY (4-([¹⁸F]fluoranylmethyl)-N-[2-[4-(2-methoxyphenyl)piperazin-1-yl]ethyl]-N-pyridin-2-ylcyclohexane-1-carboxamide) and agonist ligands [¹¹C]CUMI-101 ([O-methyl-¹¹C]2-(4-(4-(2-methoxyphenyl)-6-piperazin-1-yl)butyl)-4-methyl-1,2,4-triazine-3,5(2H,4H)dione) and [¹⁸F]F15599 (3-chloro-4-[¹⁸F]fluorophenyl-(4-fluoro-4{[(5-methyl-pyrimidin-2-ylmethyl)-amino]-methyl}-piperidin-1-yl)-methanone) seem to be promising future radioligands.^46–48 The agonists are of growing interest for this target because of the functional differences between pre- and post-synaptic 5-HT_1A receptors, which may translate into differences in binding. Here we have chosen to describe these radioligands and a few others that have been important for 5-HT_1A receptor radioligand development. A PET image of 5-HT_1A brain receptors obtained in healthy volunteers using [¹¹C]CUMI-101 is shown in Figure 1.

Figure 1.

Six PET radioligand brain images in healthy control individuals, taken with different serotonergic markers. Transverse, sagittal, and coronal sections are shown. (A) 5-HTsynthesis. Normalized [¹¹C]AMT Trapping (K*), average image from 60 healthy controls. Image courtesy of P. Gravel, M. Leyton, M. Diksic and C. Benkelfat, McGill University Health Center (MUHC), Montreal, Canada. (B) 5-HT_1A receptor. HRRTPET image of [¹¹C]CUMI-101brain distribution, from Center for Integrated Molecular Brain Imaging (Cimbi), Copenhagen, Denmark. (C) 5-HT_1B receptor. Fused MR and PET images of [¹¹C]AZ10419369 brain distribution. Average images from 3 to 93 min after injection. Image courtesy of K. Varnäs, C. Halldin and L. Farde, Karolinska Institutet, Dept of Clinical Neuroscience, Stockholm, Sweden. (D) 5-HT transporter (SERT). HRRT PET image of [¹¹C]DASB brain distribution, from Center for Integrated Molecular Brain Imaging (Cimbi), Copenhagen, Denmark. (E) 5-HT_2A receptor. HRRT PET image of [¹⁸F]altanserin brain distribution, from Center for Integrated Molecular Brain Imaging (Cimbi),Copenhagen, Denmark. (F) 5-HT₄ receptor. HRRTPET image of [¹¹C]SB207145 brain distribution, from Center for Integrated Molecular Brain Imaging (Cimbi),Copenhagen, Denmark.

B. 5-HT_1A SPECT Radioligands

1. [¹²³I]p-MPPI

[¹²³I]p-MPPI (4-[¹²³I]iodo-N-[2-[4-(2-methoxyphenyl)piperazin-1-yl]ethyl]-N-pyridin-2-ylbenzamide) was initially a promising 5-HT_1A SPECT radioligand. In the rat, brain uptake was rapid and high, the maximum ratio of hippocampus to cerebellum was 3.3 and specific binding was blocked by the 5-HT_1A receptor ligands, 8-OH-DPAT and WAY-100635.⁴⁹ A similar profile was seen in the nonhuman primate brain, suggesting [¹²³I]p-MPPI could prove useful as a SPECT radioligand. Specific binding could be blocked by pretreatment with 8-OH-DPAT or WAY-100635, as is seen both in ex vivo and in vitro autoradiographic studies.⁵⁰ By contrast, [¹²³I]p-MPPI did not display specific localization to 5-HT_1A receptor-rich brain areas in humans. This may be due to rapid in vivo metabolism causing breakdown of the amide bond, precluding its use in humans.⁵¹ Another perhaps more likely possibility is that this radioligand was excluded from human brain by an efflux transporter, such as P-gp, since close structural analogues, such as the fluoro-analogue MPPF, are known to be P-gp substrates.⁵² This experience illustrates the difficulties imposed by translating findings in other species to humans.

2. ^99mTc-Labeled Radioligands

Several attempts to develop a ^99mTc-labelled 5-HT_1A SPECT radioligand have also been reported. Progress was initially slow because of low brain uptake for many early candidates, but more recent publications suggest that further development could result in a successful radioligand for use in humans.^53–58

C. 5-HT_1A Antagonist PET Radioligands

1. [¹¹C]WAY-100635

WAY-100635 (N-[2-[4-(2-methoxyphenyl)-1-piperazinyl]ethyl]-N-2-pyridinylcyclohexane-carboxamide) is a selective and high-affinity 5-HT_1A receptor antagonist (K_D = 0.2–0.4 nM). It was originally found to have very low affinity for all other tested binding sites,⁵⁹ but recent evidence has suggested that it is also a potent D₄ agonist (K_D = 2.4 nM⁶⁰). The latter property is unproblematic for PET and SPECT imaging because D₄ receptors exist in brain at very low density relative to 5-HT_1A receptors. The binding of tritiated WAY-100635 to postmortem rodent and human brain has been mapped autoradiographically^61–63 and the binding of ¹¹C-labelled WAY-100635 to living nonhuman primate and human brain with PET.^64,65 Regional binding densities obtained from [¹¹C]WAY-100635 PET studies have been compared with agonist and antagonist binding in human postmortem tissues⁶² and good correlations between the 5-HT_1A receptor distribution were found for autoradiography versus in vivo imaging.

The first radioligand tested with PET in human subjects was [O-methyl-¹¹C]WAY-100635⁶⁵ which, unfortunately, after systemic injection produced a lipophilic radiometabolite, [O-methyl-¹¹C]WAY-100634, that crossed the blood–brain barrier.⁶⁶ To avoid this problem, the position of the ¹¹C- label was placed in the carbonyl position of WAY-100635 instead of the methyl position, to give a new radioligand, [carbonyl-¹¹C]WAY-100635. The new label position improved the target to background radioactivity ratio in both nonhuman primate and human PET studies,⁶⁷ so this radiotracer is currently the most utilized for in vivo imaging of cerebral 5-HT_1A receptors (Table I). A database of 5-HT_1A binding with PET in normal volunteers was published in 2002.⁶⁸ As reviewed by Kumar and Mann,⁴⁵ [¹¹C]WAY-100635 binding has been investigated in numerous studies of patients with psychiatric or neurological disorders. Since 2007 more than 25 further PET studies using [¹¹C]WAY-100635 have been published to date, extending the list of human disorders susceptible to changes in 5-HT_1A receptor binding to panic disorder,³⁴ bipolar depression,⁶⁹ and anorexia nervosa⁷⁰ among others.

One drawback with [¹¹C]WAY-100635 is that it undergoes a fast systemic metabolism, which makes kinetic modelling difficult, because the arterial input function is less well determined at later time points.⁷¹ To circumvent this problem, reference tissue methods have been used for quantification of [¹¹C]WAY-100635 data. Receptor-poor cerebellar white matter may serve as a reference region.^72,73 However, this approach may also be associated with some inaccuracy, since the rapid appearance of radiometabolites in plasma violates the fundamental assumption that both the reference region and the pool of free tracer in the binding regions continue to exchange significant quantities of tracer with the blood pool.⁷⁴

2. [¹¹C]CPC-222

CPC-222 (N-[2-[4-(2-methoxyphenyl)-1-piperazinyl]ethyl]-N-(2-pyridinyl)-2-bicyclo-[2,2,2] octanecarboxamide) is a WAY-100635 derivative developed in an attempt to improve its metabolic profile, and in particular its resistance to amide bond hydrolysis. Initial PET studies were promising, demonstrating high hippocampal to cerebellar radioactivity ratios in rats and humans.^75,76 As predicted, the metabolism of [¹¹C]CPC-222 is slower than that of [¹¹C]WAY-100635. However, [¹¹C]CPC-222 gives lower signal than [¹¹C]WAY-100635 and there are no further studies published with this radioligand.

3. [¹¹C](R)-RWAY

The R-enantiomer of RWAY ((2R)-1-(azepan-1-yl)-4-[4-(2-methoxyphenyl)piperazin-1-yl]-2-phenylbutan-1-one) was designed in another attempt to provide a radioligand more resistant than WAY-100635 to amide bond hydrolysis. (R)-RWAY is a high-affinity 5-HT_1A receptor antagonist. This ligand is structurally quite similar to WAY-100635, but the direction of its amide group is reversed. Because of this feature, RWAY is less susceptible to amide hydrolysis in vivo. In PET studies of nonhuman primate studies, [¹¹C](R)-RWAY cerebral binding in 5-HT_1A receptor-rich regions was high and could be blocked by pretreatment with WAY-100635.^77,78 In humans, although target to background ratios were acceptable (up to 3), determinations of distribution volumes were unstable, possibly due to the greater presence of a lipophilic radiometabolite in human plasma compared with that in primate.⁷⁹ This limits the usefulness of [¹¹C](R)-RWAY for human brain imaging.

4. [¹¹C]DWAY

[¹¹C]Desmethyl-WAY-100635 ([¹¹C]DWAY; [carbonyl-¹¹C]N-[2-[4-(2-hydroxyphenyl)-1-piperazinyl]ethyl]-N-2-pyridinylcyclohexanecarboxamide) is a putative low level radio-metabolite of [¹¹C]WAY-100635.⁶⁷ It has been shown that DWAY has very similar pharmacology to WAY, and that [carbonyl-¹¹C]DWAY exhibits favorable PET characteristics. These include, higher brain uptake than [¹¹C]WAY-100635 at equipotent doses in human brain, suitable pharmacokinetics, and a radioactive signal comparable to that of [¹¹C]WAY-100635 in rat and monkey.^80,81 However, further human studies utilizing this ligand have not yet been forthcoming, perhaps due to the difficulty of setting up the radiolabelling process. Despite attempts to improve the production methodology, radiochemical yield has continued to be low.⁸²

5. [¹⁸F]6FPWAY

In an attempt to design an alternative 5-HT_1A receptor PET ligand with a better metabolic profile and a longer radioactive half-life, bromo- and fluoro-analogues of WAY-100635 were produced: 6FPWAY (N-(2-(1-(4-(2-methoxyphenyl)-piperazinyl)ethyl))-N-(2-(6-fluoropyridinyl))-cyclohexanecarboxamide) and 6BPWAY (N-(2-(1-(4-(2-methoxyphenyl)-piperazinyl)ethyl))-N-(2-(6-bromopyridinyl))cyclohexanecarboxamide).⁸³ The introduction of the halogen groups was predicted to reduce the rate of metabolism and to provide opportunities for labelling with ¹⁸F or ⁷⁶Br. Both halogenated ligands were initially labelled with ¹¹C in their carbonyl positions for preliminary PET studies in nonhuman primates and were found promising. Satisfactory radioactive yields were obtained. Both radioligands and especially [¹¹C]6FPWAY accumulated in brain areas according to 5-HT_1A receptor density and gave high target to background ratios. Metabolism of [¹¹C]6BPWAY was reduced, as intended. While the metabolism of [¹¹C]6FPWAY was comparable with that of [¹¹C]WAY-100635, the radiometabolites were quite polar; therefore, 6FPWAY was subsequently labelled with ¹⁸F⁸⁴ and tested in monkey PET studies. [¹⁸F]6FPWAY demonstrated resistance to defluorination, but only moderately high uptake in 5-HT_1A receptor-rich regions.⁸⁵ Therefore, this radio-ligand has not seen use in human subjects.

6. [¹⁸F]MPPF

[¹⁸F]MPPF is another successful 5-HT_1A PET ligand.⁸⁶ Its use in animal and human PET studies, specifically in relation to its potential for measuring changes in endogenous 5-HT, has recently been reviewed by Aznavour and Zimmer.⁸⁷ Similar to WAY-100635, MPPF acts as a reversible, competitive antagonist at 5-HT_1A receptors, and it has high 5-HT_1A receptor affinity (K_D = 0.3 nM) and selectivity.^86,88 Optimization of production methods has resulted in a simple one-step procedure giving high yield.⁸⁹ [¹¹C]MPPF displays low nonspecific binding in the human brain and its distribution matches that of postmortem human brain 5-HT_1A receptor distribution.^90,91 [¹⁸F]MPPF is a P-gp substrate in rats,^52,92 and although it is still unknown whether this applies in humans, it has not detracted from its clinical application. A number of [¹⁸F]MPPF studies in patients have now been published, including in epilepsy,⁹³ cognitive impairment, and Alzheimer’s disease,⁹⁴ migraine,^95,96 and depression.⁹⁷ A database of PET normative data with age- and gender-related binding variables is also available,⁹⁸ as well as test–retest data.⁹⁹

Although [¹⁸F]MPPF binding to 5-HT_1A autoreceptors in the raphe nuclei seems sensitive to endogenous 5-HT levels at least in some studies,¹⁰⁰ it is still questionable if such sensitivity applies elsewhere in the brain. Studies in healthy volunteers and depressive patients in remission have concluded that [¹⁸F]MPPF binding is not sensitive to reductions in 5-HT after tryptophan depletion.^101,102 One study, however, reported an increase in several brain regions in [¹⁸F]MPPF binding potential following sleep, which is presumed to reduce 5-HT release.¹⁰³ In a recent review, it was concluded that currently available 5-HT_1A receptor radioligands do not appear to be sensitive to endogenous 5-HT.³⁰

In order to improve brain uptake, the desmethylated analogue [¹⁸F]DMPPF was synthesized through a two-step radiochemical procedure. Brain uptake was indeed found to be higher, and the compound showed better signal-to-noise ratio and a slower clearance in rats compared to [¹⁸F]MPPF.¹⁰⁴ This radioligand has to our knowledge not yet been tested in vivo in humans.

7. [¹⁸F]FCWAY

A series of radioligands based on fluorocyclohexyl analogues of WAY-100635 were developed by Lang et al.¹⁰⁵ Following evidence of high target to background ratios, several of these radioligands were further developed, and in particular two proved to be potentially useful as PET ligands. [¹⁸F]FCWAY ([¹⁸F]trans-4-fluoro-N-2-[4-(2-methoxyphenyl)piperazin-1-yl]ethyl]-N-(2-pyridyl)cyclohexanecarboxamide) was demonstrated to possess high 5-HT_1A affinity (K_i = 0.25 nM) and a high hippocampal to cerebellar binding ratio and thus had potential for imaging 5-HT_1A receptor density. [3-cis-¹⁸F]FCWAY, which has a lower affinity (K_i = 1.2 nM) and consequently faster pharmacokinetics, was proposed to be more useful for measuring dynamic changes in receptors, e.g., competition with endogenous 5-HT.^106,107

[¹⁸F]FCWAY has been used to image 5-HT_1A receptors in epilepsy,^108–111 panic disorder,¹¹² and post-traumatic stress disorder.¹¹³ However, defluorination of the parent compound leads to high bone uptake of radioactivity, which interferes with optimal imaging of superficial brain areas. Radiodefluorination in human subjects can be abolished by pre-administration of disulfiram, resulting in enhanced receptor visualization.^114,115 However, the major defluorination issue with this radioligand may be the reason for its use not expanding beyond a single PET centre.

So far, despite its slower defluorination, [3-cis-¹⁸F]FCWAY has not yet been utilized in further PET studies, possibly due to its lower affinity.

8. [¹⁸F]MefWAY

[¹⁸F]MefWAY was developed in an attempt to produce an ¹⁸F-labelled ligand that would be stable to defluorination in vivo.¹¹⁶ As an analogue of WAY-100635, [¹⁸F]MefWAY has very comparable affinity and gives a higher target to background radioactivity ratio than many other 5-HT_1A radioligands, including [¹⁸F]MPPF. In nonhuman primates, the signal is comparable with that obtained with [carbonyl-¹¹C]WAY-100635 and superior to [¹⁸F]MPPF.⁴⁶ [¹⁸F]MefWAY appeared stable to radiodefluorination in monkey. Although this radioligand seems very promising for further evaluation in human subjects, none have yet been published.

9. [¹¹C]NAD-299

NAD-299 ([R]-3-N,N-dicyclobutylamino-8-fluoro-3,4-dihydro-2H-1-benzopyran-5-carbox-amide) has several promising characteristics: subnanomolar affinity (K_D = 0.17 nM), favorable lipophilicity, high selectivity in rat brain,¹¹⁷ and high specific binding in postmortem human brain.¹¹⁸ In the nonhuman primate, [¹¹C]NAD-299 displayed good PET characteristics. These include rapid accumulation in brain and high uptake in 5-HT_1A-rich brain areas like frontal cortex (ratio of 3 to cerebellum) and raphe nuclei that was displaceable by WAY-100635. The radiometabolites were more polar than the parent radioligand, and slower to accumulate than those of WAY-100635.¹¹⁹ However, the study of this radioligand in humans has not been reported.

D. 5-HT_1A Agonist PET Radioligands

The development of a successful 5-HT_1A agonist PET radioligand has proved difficult. There are several examples based on the agonist 8-OH-DPAT,⁴⁵ but only a couple have been reasonably successful in vivo. [¹¹C]MPT ([O-methyl-¹¹C]2-(4-[4-(7-methoxynaphthalen-1-yl)-piperazin-1-yl]butyl)-4-methyl-2H-[1,2,4]triazine-3,5-dione) gave a good signal in the baboon brain, but slow washout and immeasurable plasma-free fraction limit its utility,¹²⁰ and [¹¹C]MMT ([O-methyl-¹¹C]2-(4-[4-(3-methoxyphenyl)piperazin-1-yl]-butyl)-4-methyl-2H-[1,2,4]-triazine-3,5-dione) had impressive in vitro characteristics, but PET studies in baboon failed because of low specific binding and fast clearance.¹²¹

1. [¹¹C]CUMI-101

Further optimization of structure–activity relationship in MPT analogues produced [¹¹C]MMP (now known as [¹¹C]CUMI-101) ([O-methyl-¹¹C]2-(4-(4-(2-methoxyphenyl)-piperazin-1-yl)butyl)-4-methyl-1,2,4-triazine-3,5(2H,4H)dione), a high affinity partial agonist compound (K_i = 0.15 nM) with satisfactory radiochemical yield, and which appeared superior to both [¹¹C]MPT and [¹¹C]MMT in terms of binding ratios and wash out times.¹²² PET studies in baboons confirmed good 5-HT_1A selectivity with high specific binding that was displaceable by WAY-100635 and 8-OH-DPAT. [¹¹C]CUMI-101 has polar radio-metabolites that do not cross the blood–brain barrier, and in vivo modelling data in baboons have been published.⁴⁸ Test–retest studies in the baboon were satisfactory (8–13%) and it was suggested that this radioligand may be sensitive to endogenous changes in 5-HT,¹²³ although studies in rodents do not support this.¹²⁴ In humans, [¹¹C]CUMI-101 has promising features and may preferentially label the high affinity site (Fig. 1¹²⁵), but it remains to be seen whether it can be used to image endogenous 5-HT release in humans or whether it will make a good in vivo radioligand in clinical populations.

2. Other 5-HT_1A Agonist Radioligands

The development of other 5-HT_1A agonists is ongoing and several new leads have been investigated. F15599 is an agonist with a K_i of 2.2 nM, and over 1,000-fold selectivity with respect to a wide range of other receptors, transporters, ion channels, and enzymes.¹²⁶ It seems that F15599 preferentially activates postsynaptic 5-HT_1A receptors in rat frontal cortex.¹²⁷ It contains fluorine in a way that is compatible with ¹⁸F labelling, and the radiosynthesis and validation of [¹⁸F]F15599 as a potential PET radioligand were recently reported.⁴⁷ Using autoradiography, this study reported similar receptor distributions using [¹⁸F]F15599 and [¹⁸F]MPPF in rat and cat brain. In vivo microPET showed a rapid accumulation of the radioligand in rat brain but with a cortex to cerebellum ratio of only 1.6. Similar results were obtained in the cat brain. The low target to background ratio augurs poorly for the usefulness of [¹⁸F]F15599 as a radioligand for human 5-HT_1A receptor PET studies.

S14506 (1-[2-(4-fluorobenzoylamino)ethyl]-4-(7-methoxy-naphthyl)piperazine) is another high-affinity ligand (K_d = 0.79 nM) with reasonable selectivity demonstrated in in vitro and ex vivo studies. It was recently labelled to produce [¹¹C]S14506 and [¹⁸F]S14506 and tested for its suitability as an in vivo tracer in rat and monkey PET studies.¹²⁸ Unfortunately, due to low uptake and low signal-to-background ratio neither tracer was deemed suitable for in vivo imaging.

E. 5-HT_1A Radioligands: Conclusions

In conclusion, there are three radioligands in current use for PET studies of the 5-HT_1A receptor in human subjects: [carbonyl-¹¹C]WAY-100635, [¹⁸F]MPPF, and [¹⁸F]FCWAY. A further two promising radioligands are emerging: [¹⁸F]MefWAY and [¹¹C]CUMI-101. [carbonyl-¹¹C]WAY-100635 is the most widely used 5-HT_1A receptor radioligand. It has the advantage of a high target-to-background ratio and its distribution matches that of the 5-HT_1A receptors. Its only disadvantage is its fast metabolism in plasma and consequently, there are difficulties with its accurate quantification. [¹⁸F]MPPF has the advantage of the longer lived ¹⁸F-label, and it also selectively labels the 5-HT_1A receptors with a low non-specific binding. Its major disadvantage is its low brain uptake. [¹⁸F]FCWAY also benefits from being ¹⁸F-labelled, and displays affinity and selectivity comparable to its analogue WAY-100635 and also to [¹⁸F]MPPF. [¹⁸F]FCWAY is, however, prone to defluorination in vivo, which severely compromises its use for some brain regions, especially cortex. [¹⁸F]MefWAY is also analogous to WAY-100635, and its resistance to defluorination makes it a promising radioligand. [¹¹C]CUMI-101 is a high affinity 5-HT_1A (partial) agonist radio-ligand that displays high specific binding in the baboon and also appears suitable for imaging the high affinity site within human brain.

F. 5-HT_1B Receptor Radioligands

The evolution of the study of 5-HT_1B receptors has been complicated because 5-HT_1B pharmacology was long confused with that of 5-HT_1D receptors found in the rodent brain. Initially, 5-HT_1B and 5-HT_1D were considered to be species homologues of the same receptor, but it was later revealed that the human 5-HT_1D receptor encompassed two receptor sub-types, 5-HT_1Dα and 5-HT_1Dβ, encoded by separate genes; 5-HT_1Dα became what is now known as the 5-HT_1D receptor and 5-HT_1Dβ became the 5-HT_1B receptor.¹⁵ Most studies suggest that in humans, the 5-HT_1B receptor makes up the vast proportion of 5-HT_1B/D receptors in the brain.^129,130

Interest in the receptor was enhanced by the discovery of sumatriptan and related 5-HT_1B/D agonist drugs for the treatment of migraine.¹³¹ The majority of 5-HT_1B ligands have mixed 5-HT_1B/1D pharmacology, but selective 5-HT_1B ligands do exist and are now of interest due to their emerging utility in the study of depression, aggression, and drug reinforcement.^132–134 Radioligands have been synthesized to facilitate in vitro studies of the 5-HT_1B receptor and development of PET radioligands for the 5-HT_1B receptor is ongoing. Recent interest has been sparked by the suggestion that even 5-HT_1B antagonist ligands may be sensitive to displacement by endogenous 5-HT.¹³⁵

G. 5-HT_1B PET Radioligands

1. [¹¹C]AZ10419369

[¹¹C]AZ10419369 (5-methyl-8-(4-[¹¹C]methyl-piperazin-1-yl)-4-oxo-4H-chromene-2-carboxylic acid(4-morpholin-4-yl-phenyl)-amide) has recently been reported to be a selective 5-HT_1B antagonist suitable for PET studies in nonhuman primates¹³⁶ and humans.¹³⁷ [³H]AZ10419369 has high affinity (K_D = 0.4 nM) and high specific binding to the human 5-HT_1B receptor,¹³⁸ although its selectivity remains to be documented. Initial [¹¹C]AZ10419369 PET studies in nonhuman primates demonstrated a high brain uptake and a regional brain distribution in line with that previously reported for the 5-HT_1B receptor, also correlating with autoradiographic images with [³H]AZ10419369. Other promising characteristics included a relatively high in vitro target-to-background ratio (cortex/cerebellum =2.3 in human tissue), displaceable binding by a 5-HT_1B specific antagonist AR-A000002, a slow metabolism, and no lipophilic radiometabolites. This ligand may also demonstrate sensitivity to displacement by endogenous 5-HT in nonhuman primates.¹³⁵ The first human PET study with [¹¹C]AZ10419369 suggested similarly good in vivo characteristics.¹³⁷ This radioligand displayed rapid uptake, a good target-to-background ratio, no evidence of the formation of radiometabolites, and a binding distribution that correlated with receptor autoradiographical studies and known 5-HT_1B densities. In addition, the binding potentials obtained with reference tissue modelling correlated well with those arising from arterial input kinetic modelling. A PET image of 5-HT_1B brain receptors obtained in healthy volunteers using [¹¹C] AZ10419369 is shown in Figure 1.

2. [¹¹C]P943

The development and validation of the 5-HT_1B radioligand, [¹¹C]P943 (R-1-[4-(2-methoxy-isopropyl)-phenyl]-3-[2-(4-methyl-piperazin-1-yl)benzyl]-pyrrolidin-2-one), has very recently been reported, including modelling approaches for its quantification in the human brain.¹³⁹ P943 is a high affinity 5-HT_1B antagonist (K_i = 1.2 nM), with affinities for most other serotonergic receptors being more than 100-fold lower. It has some affinity toward the 5-HT_1D receptor (K_i = 12 nM), but only a very small fraction of [¹¹C]P943 was expected to bind to the 5-HT_1D receptors due to their low abundance. This has not yet been confirmed, since the only in vivo blocking study so far was done using the mixed 5-HT_1B/1D receptor antagonist, GR127935.¹⁴⁰ A comparison between autoradiography data and binding potentials measured in vivo by human PET, however, shows a fairly good correlation and suggests that the signal is indeed specific to the 5-HT_1B receptor. Radiosynthesis of [¹¹C]P943 is simple, and the substance is taken up slowly into the brain with a peak after 20 min. On the contrary, the washout appears to be relatively fast. Another promising characteristic is the slow metabolism of the radioligand with the production of polar radiometabolites only. Cerebellar gray matter can be used as a reference region, as this is devoid of 5-HT_1B receptors,¹⁴¹ and reference tissue models could potentially be used.¹³⁹ Even though pharmacological displacement of [¹¹C]P943 has not yet been performed in humans to confirm the specificity of the radioligand, [¹¹C]P943 appears to be a promising radioligand for the quantification of 5-HT_1B receptor densities in the human brain in vivo. The first patient studies are now emerging; the first in alcohol dependence¹⁴² and another in depression.¹⁴³

H. 5-HT_1D, 5-ht_1e, and 5-HT_1F Radioligands

The distribution of 5-ht_1e, 5-HT_1D, and 5-HT_1F receptors has been mapped in brain tissue.^144,145 However, a lack of selective ligands for these receptors has hampered research into their function and potential for drug discovery. There are no reported selective ligands of any kind for the 5-ht_1e receptor. 5-HT_1D and 5-HT_1F receptors are of interest in migraine treatment.^146,147 There now exists a few selective ligands for the 5-HT_1D receptor,¹⁴⁷ and a selective radioligand for in vitro characterization is available for the 5-HT_1F receptor.¹⁴⁸ To date, no selective PET or SPECT radioligands for in vivo brain imaging of the 5-HT_1D, or 5-HT_1F receptor have been developed.

I. 5-HT_2A Receptor Radioligands

5-HT_2A receptors and their mRNAs have been extensively mapped in the brain by auto-radiography, in situ hybridization, and immunocytochemical techniques.^149–152 There are several selective 5-HT_2A ligands available, and a review of certain 5-HT₂ ligands, their affinities, efficacies, and indications was published several years ago.¹⁵³ 5-HT_2A receptors are of interest for many reasons: they are a primary target of psychedelic compounds, contribute to the efficacy of many antipsychotic medications, and are involved in the etiology or treatment of various psychiatric disorders.^17,153,154

The selective radioligands available for studying 5-HT_2A receptors include [³H]RP62203 (fananserin, (N-[3-[4-(4-fluorophenyl)piperazin-1-y1]propyl]-1,8-naphthalenesultam)), [³H]ke-tanserin (3-[2-[4-(4-fluorobenzoyl)piperidin-1-yl]ethyl]-1H-quinazoline-2,4-dione), [³H]MDL 100,907 ((R)-4-(l-hydroxy-1-(2,3-dimethoxyphenyl)methyl)-N-2-(4-fluorophenylethyl)piperidine), and [³H]altanserin (3-[2-[4-(4-fluorobenzoyl)piperidin-1-yl]ethyl]-2-sulfanylidene-1H-quinazolin-4-one). Attempts have been made to label these ligands for PET and/or SPECT imaging with mixed results, as described below. Only one successful SPECT ligand for imaging of 5-HT_2A receptors has so far been produced: [¹²³I]-R91150 (4-amino-N-[1-[3-(4-fluorophenox-y)propyl]-4-methyl-4-piperidinyl]-5-iodo-2-methoxybenzamide), also termed [¹²³I]-5-I-R91150. [¹¹C]MDL 100,907 and especially [¹⁸F]altanserin have both become successful PET ligands, among others that show promise. A PET image of 5-HT_2A brain receptors obtained in healthy volunteers using [¹⁸F]altanserin is shown in Figure 1.

J. 5-HT_2A SPECT Ligands

1. [¹²³I]DOI

Scintigraphic images of 5-HT₂ receptors in the human brain were obtained in the 1970s, well before the first PET or SPECT studies emerged. Sargent and colleagues demonstrated the preferential accumulation of radioactivity in brain after administration of radio-labelled DOI (1-(2,5-dimethoxy-4-iodophenyl)-2-aminopropane) and DOB (1-(2,5-di-methoxy-4-bromophenyl)-2-aminopropane).^155,156 As these two drugs are routinely used to test 5-HT₂ receptor function in vivo, there is a plethora of data on their pharmacology and kinetics.

[¹²⁵I]DOI is a radiolabelled non-selective 5-HT₂ agonist, often used to detect 5-HT_2A/2C receptors in the brain with autoradiography.^157,158 Very early imaging studies utilizing racemic DOI with a brominated label were not particularly successful.¹⁵⁵ However, it was later revealed that the R-enantiomer displayed the highest affinity and selectivity,¹⁵⁹ and therefore [¹²³I]R-DOI was synthesized for in vivo testing as a potential SPECT ligand in rodents and baboon.¹⁶⁰ [¹²³I]R-DOI had good brain penetration, with high accumulation in 5-HT_2A receptor-rich brain areas, but its low target-to-background ratio and its inability to be displaced by ketanserin or cold DOI suggested a high nonspecific uptake and made the ligand unsuitable for in vivo imaging by SPECT.

2. [¹²³I]MSP

After the butyrophenones emerged as the foundation for 5-HT_2A selective antagonist ligands, F[¹¹C]MSP (8-[3-(4-fluorobenzoyl)propyl]-1-methyl-1,3,8-triazaspiro[4,5]decan-4-one) and [¹²³I]MSP (8-[3-(4-iodobenzoyl)propyl]-1-methyl-1,3,8-triazaspiro[4,5]decan-4-one) were developed as potential PET and SPECT ligands, respectively.¹⁶¹ F[¹¹C]MSP was quickly abandoned, due to problems with low specific radioactivity. Nevertheless, in vivo pharma-cokinetic and brain binding characterization of [¹²³I]MSP in mice suggested sufficient uptake and retention in the brain, with regional accumulation corresponding to known 5-HT_2A receptor distribution. Binding in frontal cortex (target-to-background ratio of 3.5) could be partly blocked by pre-treatment with ritanserin and IMSP but not by ketanserin. No further studies with [¹²³I]MSP have been published.

3. [¹²³I]-R91150

The SPECT radioligand [¹²³I]-R91150 has several different names in the literature. Originally, it was called either [¹²³I]-5-I-R91150¹⁶² or [¹²³I]-R93274,¹⁶³ but it is now often named [¹²³I]-R91150, even though R91150 originally was used to designate the noniodinated analogue of 5-I-R91150. For simplicity, we will refer to it as [¹²³I]-R91150.

¹²⁵I-labelled R91150 was originally synthesized by Mertens et al.¹⁶⁴ and further characterized by Terrière et al.¹⁶⁵ It has high affinity (K_D = 0.11 nM) and selectivity for 5-HT_2A receptors in vitro, and preferential retention in rat frontal cortex in vivo with a target-to-background ratio of 10. Preliminary SPECT studies in baboons using [¹²³I]-R91150 were also promising.¹⁶³ Target-to-background ratios were lower than in rats (1.3–1.5 in cortical areas), and kinetics and metabolism were faster. Only polar radiometabolites were seen and binding could be blocked by administration of ketanserin (5-HT₂ antagonist) but not by raclopride (D₂ antagonist). In humans, frontal cortex-to-cerebellum ratios reached values of 1.4, and remained stable, probably due to a slow dissociation rate of [¹²³I]-R91150.¹⁶²

In early human studies, a single bolus injection of [¹²³I]-R91150 was used to calculate the specific uptake ratio using the simple tissue ratio method at pseudoequilibrium. A later study validated this method against full kinetic compartmental analysis and found a good correlation between compartmental modelling with arterial plasma input and the tissue ratio method.¹⁶⁶ Catafau et al.¹⁶⁷ further evaluated the in vivo properties of [¹²³I]-R91150 by conduction of a dose-dependent displacement of [¹²³I]-R91150 binding with ketanserin.

Despite the lower signal-to-noise ratio of [¹²³I]-R91150 compared with the available PET radioligands, the widespread availability of SPECT facilities made [¹²³I]-R91150 quite popular. A string of publications appeared, based on [¹²³I]-R91150-SPECT determined in vivo 5-HT_2A receptor occupancy by various ligands^168–170 and changes in 5-HT_2A density involved in disease states including cognitive decline,¹⁷¹ suicidal behavior,^172,173 and anorexia nervosa.^174,175 The influence of age and gender on binding potential in healthy subjects has also been determined.¹⁷⁶ Due to its success, R91150 was also labelled with ¹⁸F and evaluated as a potential PET radioligand. Ex vivo data from mice were promising, but the six-step radiosynthesis prevents a more widespread use of this radioligand.¹⁷⁷

4. [¹²³I]-3-I-CO

Fu et al.¹⁷⁸ developed a series of halogenated novel compounds based on common structural features of altanserin and MDL 100,907 to provide a new series of 4′-substituted phenyl-4-piperidinylmethanol and benzoyl-4-piperidine derivatives. Affinity for the 5-HT_2A receptor was determined and the compounds were further evaluated for selectivity for 5-HT_2A versus 5-HT_2C, 5-HT₆, and 5-HT₇, as well as dopamine D₂ and adrenergic α₁ and α₂ receptors. Several promising compounds were reported, of which three have been labelled with ¹²³I and evaluated as potential SPECT ligands^179–181: [¹²³I]-(4-fluorophenyl)(1-[2-(2-iodophenyl) ethyl]piperidin-4-yl)methanone, [¹²³I]-(4-fluorophenyl)(1-[2-(4-iodophenyl)ethyl]piperidin-4-yl) methanone, and [¹²³I]-(4-fluorophenyl)[1-(3-iodophenethyl)piperidin-4-yl]methanone ([¹²³I]-3-I-CO). Of these [¹²³I]-3-I-CO was the most promising ligand with its high affinity (K_i = 0.51 nM) and selectivity toward 5-HT_2A receptors. In vivo, it readily entered the rat brain and its binding was displaceable by ketanserin. In addition, in rats no radiometabolites entered the brain. However, target-to-background ratio was low, and it is possible that [¹²³I]-3-I-CO is a substrate for P-gp.¹⁸²

K. 5-HT_2A PET Radioligands

1. [¹¹C]Ketanserin and [¹⁸F]FEK

The first mention of a PET study specifically aiming at imaging 5-HT₂ receptors in the brain utilized [¹¹C]ketanserin¹⁸³ based on its known favorable in vitro and in vivo characteristics, although it also displays high histamine sub-type-1 (H₁) receptor affinity and lower affinities for the α₁-adrenoceptor and the 5-HT_2C receptor. Here, preferential accumulation in frontal cortex relative to cerebellum was described, and the binding could be blocked by pretreatment with unlabeled chlorpromazine. However, the target-to-background ratio was low in the human brain, and the radiotracer metabolized fast.¹⁸³ An analogue to ketanserin, [¹⁸F]fluoroethylketanserin ([¹⁸F]FEK) was subsequently developed. It had better PET properties¹⁸⁴ at least in the baboon, where the frontal cortex-to-cerebellar ratio was 2.5, and pretreatment with ketanserin blocked the specific binding. Despite these initial promising observations, further [¹⁸F]FEK human studies have not been published.

2. [¹¹C]NMSP

[¹¹C]NMSP (N-methylspiperone, 8-[4-(4-fluorophenyl)-4-oxobutyl]-2-methyl-4-phenyl-2,4,8-triazaspiro[4.5]decan-1-one) is a dual D₂/5-HT₂ ligand. Although NMSP has high affinity for both receptors, the majority of the specific binding in neocortex is due to 5-HT₂ receptor binding, since D₂ receptors are only present at low density in cortex, thus justifying its use as a 5-HT₂ radioligand in this area.^185,186

[¹¹C]NMSP has been used primarily as an imaging tool to visualize D₂ receptor binding in striatum, but was also used in early PET studies to estimate changes in cortical 5-HT₂ receptor binding in, for example, aging,¹⁸⁷ schizophrenia,¹⁸⁸ and 5-HT₂ receptor occupancy by antipsychotic medications such as risperidone,¹⁸⁹ clozapine,^190,191 MDL 100,907,¹⁹² and flupentixol.¹⁹³ At that time, the more selective PET radioligands, [¹⁸F]setoperone and [¹⁸F]altanserin, had yet to be fully characterized.

3. [¹¹C]MBL

The radiolabelled LSD derivative, N¹-([¹¹C]-methyl)-2-Br-LSD ([¹¹C]-MBL), has also been tested as a PET radioligand.^194,195 MBL has high 5-HT_2A receptor affinity in vitro (K_i = 0.5 nM), eight-fold lower affinity for D₂ receptors, high affinity for 5-HT_2C receptors,¹⁹⁶ and weak α₁-adrenoceptor and 5-HT₁ receptor interactions.¹⁹⁵ Nonetheless, initial studies in baboons suggested that [¹¹C]-MBL selectively labelled cortical regions and was blocked by ketanserin, and human studies suggested the same; highest labelling in frontal, temporal, and parietal cortex (cortical-to-cerebellar ratios: 1.7–2.7) with lower levels observed in caudate and putamen and lowest in cerebellum. The lack of further PET imaging papers could be indicative of its nonselective pharmacological profile combined with the emergence of better radiotracers for imaging the 5-HT_2A receptors.

4. [¹⁸F]Setoperone

The structure of setoperone is related to that of ketanserin. [¹⁸F]setoperone (6-[2-[4-(4-[¹⁸F]fluorobenzoyl)piperidin-1-yl]ethyl]-7-methyl-2,3-dihydro-[1,3]thiazolo[3,2-a]pyrimidin-5-one) in vivo showed high binding in baboon brain in areas known to be rich in 5-HT_2A receptors such as cerebral cortex, but also high binding in striatum.¹⁹⁷ The target-to-background ratio was 3 in cortex and binding could be blocked by pretreatment with spiperone and ketanserin. In striatum, radioligand binding was fully prevented by spiperone but only partly by ketanserin suggestive of a significant contribution from dopamine D₂ receptors. In humans, pretreatment with ketanserin, sulpiride, and prazosin confirmed that the [¹⁸F]setoperone signal was due to 5-HT₂ receptors in frontal cortex, to D₂ receptors in striatum and that no significant α₁-adrenoceptor binding was evident.¹⁹⁸ In vivo, [¹⁸F]setoperone produced radiometabolites, but these were less lipophilic than the parent compound. Due to the differential localization of the 5-HT_2A relative to D₂ receptors, [¹⁸F]setoperone became a relatively successful 5-HT_2A PET ligand despite its affinity for D₂ receptors. It has been used to estimate the occupancy of 5-HT_2A receptors by antipsychotics^199–201 and to assess possible changes in 5-HT_2A density in diseases such as Alzheimer’s disease, migraine, stroke, and depression, and in response to electro-convulsive shock therapy.^202–207

5. [¹⁸F]Altanserin

Altanserin, like setoperone, is a fluorobenzoyl derivative structurally related to ketanserin. Its K_i for 5-HT_2A receptors is 0.13 nM, so given its α₁-adrenoceptor affinity of 5 nM and D₂ affinity of 62 nM, the majority of the signal can be attributed to 5-HT_2A receptor binding. The first [¹⁸F]altanserin study was published in 1991,²⁰⁸ and today, [¹⁸F]altanserin is the most frequently used 5-HT_2A receptor PET radioligand (Table II and Fig. 1).

The potential of [¹⁸F]altanserin as a PET radioligand was first demonstrated in the rat brain where the frontal cortex-to-cerebellum ratio is 11, and its binding can be blocked by pretreatment with ritanserin, ketanserin, and setoperone, and only to a minor degree by D₂ ligands.²⁰⁸ In a follow-up study in humans, the synthesis methodology was improved,²⁰⁹ ketanserin pretreatment blocked [¹⁸F]altanserin binding in the human brain,²¹⁰ and displaced [¹⁸F]altanserin binding.⁴⁰ Although [¹⁸F]altanserin produces radiometabolites, these were initially considered not to interfere with binding characteristics.²¹¹ Later, it was realized that the radiometabolites of [¹⁸F]altanserin contributed to the nonspecific binding, and thus quantification required complex kinetic modelling.^212,213 To simplify the quantification and to take into account the lipophilic radiometabolite, a bolus–infusion approach was developed. Initially, 6 hr of infusion were required to obtain equilibrium measurements,²¹⁴ but a later study showed that this time could be reduced to 2 hr, making infusion studies more feasible.⁴⁰ The bolus–infusion paradigm with [¹⁸F]altanserin was later shown to have excellent test–retest reliability, particularly in large brain regions with high binding.²¹⁵

Following the positive validation studies, [¹⁸F]altanserin was used to determine changes in 5-HT_2A receptor density in relation to aging,^216–219 depression,^220,221 anorexia nervosa/bulimia,^222,223 obsessive–compulsive disorder,²²⁴ cognitive decline,^225,226 Tourette’s syndrome,²²⁷ risk or onset of schizophrenia,^228–231 and in relation to the personality trait neuroticism.^232,233 A database of 5-HT_2A binding in healthy volunteers has been published,²¹⁸ and it has been reported that binding in healthy subjects correlates with body mass index²³⁴ but does not vary with gender.²³⁵ Furthermore, twin studies have shown that [¹⁸F]altanserin binding has a strong genetic component.²³⁶

6. [¹⁸F]Deuteroaltanserin

The significant advantages of [¹⁸F]altanserin include its specific brain uptake, kinetics that allow for bolus–infusion schedules, a high target-to-background ratio, and a high reproducibility. Prior to the development of suitable bolus–infusion regimes, formation of lipophilic radiometabolites from [¹⁸F]altanserin was considered a major disadvantage because of the resulting complex quantification methods in data analysis. Development of a PET radioligand based on [¹⁸F]altanserin, which did not give rise to radiometabolites crossing the blood–brain barrier, led to the synthesis of [¹⁸F]deuteroaltanserin.²³⁷ The two deuterium atoms, which are present in place of hydrogen atoms, were hypothesized to retard metabolism. This was indeed the case, and [¹⁸F]deuteroaltanserin was reported to have better brain uptake in baboon and humans compared with [¹⁸F]altanserin, using a bolus–infusion paradigm. In humans, cortical-to-cerebellar ratio was increased by 26% above that observed previously for [¹⁸F]altanserin, suggesting it might be a superior PET radioligand.²³⁷ In a subsequent study by the same group, the test–retest reliability was also quite good.²³⁸

Staley et al.²³⁹ directly compared [¹⁸F]altanserin and [¹⁸F]deuteroaltanserin in baboons under equilibrium receptor binding conditions. They demonstrated 5-HT_2A specificity of both tracers by injecting the 5-HT_2A antagonists ketanserin and SB46349B, and they concluded that [¹⁸F]deuteroaltanserin was essentially equivalent to [¹⁸F]altanserin for 5-HT_2A receptor imaging in the baboon.

Since then, only two further studies have been published using [¹⁸F]deuteroaltanserin in humans; one to demonstrate that oestrogen replacement therapy increases prefrontal 5-HT_2A receptor density²⁴⁰ and the other suggesting that 5-HT_2A receptors are reduced in cortical regions in Alzheimer’s disease.²⁴¹ It remains to be seen whether this ligand will become as successful as its predecessor.

7. [¹⁸F]RP62203

Although [¹⁸F]setoperone and especially [¹⁸F]altanserin are both successful PET radio-ligands, they both display somewhat mixed pharmacology, as described above, which could limit their utility. [³H]RP62203 was shown to be a 5-HT_2A antagonist with very high affinity (K_i = 50.0 pM) and a good selectivity profile in rat cerebral cortex,²⁴² without any significant D₂ and α₁ receptor affinity. Subsequent in vivo binding characteristics of ¹⁸F-labelled RP62203 ([¹⁸F]RP62203) revealed specific binding that correlated with the known 5-HT_2A receptor distribution in the rat brain. Binding was four times higher in cortical regions than cerebellar regions and was abolished by prior dosing with ritanserin. This suggested that [¹⁸F]RP62203 might make a good PET radioligand.²⁴³ Shortly after, this was confirmed in rat studies by a different group²⁴⁴ who found that [³H]RP62203 gave a cortex-to-cerebellum activity ratio in rats in vivo of 9. However, the radiosynthesis of [¹⁸F]RP62203 is multi-step and therefore probably too demanding for regular implementation. For this reason and perhaps also because of the subsequent emergence of [¹¹C]MDL 100,907, no further reports on [¹⁸F]RP62203 have been seen.

8. [¹¹C]MDL 100,907

MDL 100,907 is a reversible, highly selective 5-HT_2A ligand with subnanomolar affinity (K_D = 0.14–0.19 nM).²⁴⁵ Despite differences in their in vitro selectivity profiles, the binding of [¹⁸F]altanserin and [³H]MDL 100,907 to the 5-HT_2A receptor is quite comparable.²⁴⁶ [¹¹C]MDL 100,907 showed promise in nonhuman primates where it accumulated in 5-HT_2A receptor-rich regions, with a neocortex-to-cerebellum ratio of 4.5 that was abolished after injection of ketanserin.²⁴⁷ Later, radioligand binding and autoradiography studies using [³H]MDL 100,907 and [¹¹C]MDL 100,907 confirmed its selectivity and high specific binding in rat, nonhuman primate, and human brain,^{152,245,248,249} making it the first truly selective 5-HT_2A receptor ligand. In parallel, favorable PET characteristics for [¹¹C]MDL 100,907 were reported in humans.²⁵⁰ MDL 100,907 has moderate lipophilicity and the binding potential is 4–6 times higher in neocortex than cerebellum, where binding is low.

Further studies have validated the methodology for modelling [¹¹C]MDL 100,907 binding in PET studies, and find that a two-tissue compartment model using arterial input is superior to reference tissue models.^251,252 This somewhat complicates the use of [¹¹C]MDL 100,907, since compartment modelling requires the metabolite-corrected arterial plasma radioactivity concentration to be determined during each scan. A recent study suggests that noninvasive graphical analysis (NIGA) is comparable with the two-tissue compartment model,²⁵³ which greatly enhances the applicability of [¹¹C]MDL 100,907. [¹¹C]MDL 100,907 has so far only been used in a limited number of clinical studies including determination of 5-HT_2A receptor binding in patients recovered from depression³⁵ and in patients with obsessive–compulsive disorder.²⁵⁴

9. [¹⁸F]MH.MZ and (R)-[¹⁸F]MH.MZ

The advantages of [¹¹C]MDL 100,907 over [¹⁸F]altanserin include its lack of radio-metabolites, and its higher 5-HT_2A receptor selectivity. Its disadvantages include the short half-life of the ¹¹C label. To combine the favorable properties of each ligand, Herth et al.²⁵⁵ developed MH.MZ (FE1-MDL 100,907) ((3-fluoro-ethoxy-2-methoxyphenyl)-1-[2-(4-fluoro-phenyl)ethyl-4-piperidine-methanol), by replacing an O-methyl group within MDL 100,907 with a labelled fluoroethyl moiety. MH.MZ has a lower affinity for the 5-HT_2A receptor than MDL 100,907 and altanserin (K_i = 3 nM), but its visualization by autoradiography was good, nonspecific binding was low, and competition studies suggested relatively high 5-HT_2A selectivity. The high specificity was later confirmed, as it was shown that [¹⁸F]MH.MZ has very high K_i values for a whole range of other receptors including all other serotonergic receptors (the lowest being for 5-HT_2C where K_i = 71 nM).²⁵⁶ Following this [¹⁸F]MH.MZ was evaluated in vivo and it was found that the radioligand readily entered rat brain and gave a cortex-to-cerebellum ratio of 2.7. It was observed that [¹⁸F]MH.MZ undergoes extensive first-pass metabolism, which significantly reduced its bioavailability. Further, its time–activity curves showed a very slow washout from the rat brain. One polar radiometabolite was found in ex vivo homogenates from rat brain, but this was considered to be contamination from blood still located in vessels in the brain rather than from brain gray matter tissue.²⁵⁶ To further improve selectivity and increase affinity, a new series of similar compounds was developed including the enantioselective derivative, (R)-[¹⁸F]MH.MZ.²⁵⁷ This compound demonstrated higher affinity (K_i = 0.72 nM), lower nonspecific binding, and higher target-to-background ratio (ratios of between 3.3 and 3.9 in cortex relative to cerebellum), as demonstrated by PET in rats.²⁵⁸ Whether [¹⁸F]MH.MZ or (R)-[¹⁸F]MH.MZ has better characteristics than the 5-HT_2A radioligands already used in human PET still remains to be settled.

10. [¹¹C]CIMBI Compounds

As mentioned above, agonist tracers could potentially enable imaging of the active, high-affinity state of receptors, which may provide a more meaningful assessment of membrane-bound receptors. The first radiolabelled high-affinity 5-HT_2A receptor agonist was 2-(4-iodo-2,5-dimethoxyphenyl)-N-(2-[¹¹C-O-methyl]methoxybenzyl)ethanamine ([¹¹C]Cim-bi-5). In the pig brain, [¹¹C]Cimbi-5 showed a cortex-to-cerebellum binding ratio in the same order of magnitude as [¹⁸F]altanserin, and it was displaceable by ketanserin in both rats and pigs.²⁵⁹ Subsequently, the in vivo validation of an additional nine novel 5-HT_2A receptor agonist PET tracers in the pig brain has been presented,²⁶⁰ and of these, [¹¹C]Cimbi-36 had the most favorable kinetics and the highest target-to-background ratio. This series of compounds seem to be the first promising radioligands for the investigation of 5-HT_2A agonist binding in the living human brain.

L. 5-HT_2A Radioligands: Conclusions

Five of the above described radioligands for the 5-HT_2A receptor have successfully been used in human studies; the SPECT radioligand [¹²³I]-R91150, and the PET radioligands [¹⁸F]setoperone, [¹⁸F]altanserin, [¹⁸F]deuteroaltanserin, and [¹¹C]MDL 100,907.

[¹²³I]-R91150 displays a lower signal-to-noise ratio compared to the available PET radioligands, but the widespread availability of SPECT facilities makes it a practical imaging tool for, e.g., drug occupancy studies, in spite of the lower resolution provided by SPECT.

Of the PET radioligands [¹⁸F]altanserin has continued to be most widely used, despite its lipophilic radiometabolite. This use is especially due to its longer lived ¹⁸F-label, which enables the application of a bolus/infusion paradigm. Imaging data obtained from [¹⁸F]altanserin binding in the human brain are highly reproducible and the large number of publications based on this radioligand provides a convenient reference for new findings. The radioligand [¹⁸F]setoperone is less selective than [¹⁸F]altanserin and is being used less and less. [¹¹C]MDL 100,907 is a more selective 5-HT_2A ligand than [¹⁸F]altanserin in vitro, but is much less widely used as a 5-HT_2A radioligand for in vivo studies than [¹⁸F]altanserin (Table III). The reason for this could be that the modelling issues make [¹¹C]MDL 100,907 less attractive and only time will tell whether the quantification of [¹¹C]MDL 100,907 with NIGA will change this. Similarly, it will be interesting to see whether (and how) imaging 5-HT_2A receptors with agonist radioligands differs from that of antagonist radioligands.

M. 5-HT_2B Radioligands

The 5-HT_2B receptor is predominantly expressed in peripheral tissues such as cardiac and intestinal tissue. 5-HT_2B mRNA and protein has been detected in human brain, but only in a few discrete nuclei.²⁶¹ Although only low levels of 5-HT_2B receptors are present in the brain,²⁶² certain selective compounds have started to emerge that may permit further investigation of their function.²⁶³ However, no selective radioligands have been developed so far.

N. 5-HT_2C Radioligands

Very high levels of the 5-HT_2C receptor are found in the epithelial cells of the choroid plexus, but they are also present in gray matter.^264–266 Due to the lack of selective radioligands, this receptor is less well studied than its pharmacologically related 5-HT_2A receptor. Selective nonlabelled ligands for the 5-HT_2C receptor exist, but currently all the labelled 5-HT_2C radioligands have shared pharmacology with other receptors (e.g. Hamedah et al.²⁶⁷), and none have been developed for SPECT or PET imaging. This is disappointing given the potential role of this receptor in feeding behavior and mood. Many atypical antipsychotics are potent 5-HT_2C antagonists,²⁶⁸ which has been suggested to relate to their propensity to cause weight gain and the new antidepressant agomelatine is thought to act in part by blocking 5HT_2C receptors so elevating noradrenaline and dopamine levels in cortex.²⁶⁹

O. 5-HT₃ Receptor Radioligands

The 5-HT₃ receptor is unique among 5-HT receptors in being a ligand-gated ion channel. Its structure consists of a pentamer of subunits that come together to form a cation channel. Five different subunits have been identified: 3A–E. The 3A subunits have been shown to form functional homomers, the others form hetero-oligomeric pentamers, with the inclusion of a 3A subunit being necessary for functional integrity.²⁷⁰ Located on both central and peripheral neurons, the highest brain levels are found in the dorsal vagal complex of the brain stem, with low levels found in the forebrain.^271–274

Growth in the availability of selective ligands for the 5-HT₃ receptor came about once the anti-emetic properties of 5-HT₃ receptor antagonists were established and is still of interest as an emerging targets in irritable bowel syndrome and alcohol abuse.^275,276 Several selective agonists and antagonists are available, of which some have been labelled with ³H or ¹²⁵I for in vitro studies (e.g. Hewlett et al.²⁷⁷). In all, eight potential PET ligands have been tested in the attempt to create a successful radioligands for 5-HT₃ receptors. Six were unsuccessful due to a lack of specific binding and two because of poor brain uptake.

P. 5-HT₃ SPECT Radioligands

1. Zacopride Derivatives

Ebert and co-workers developed a series of zacopride derivatives as potential high-affinity 5-HT₃ antagonists radiolabelled with ¹²⁵I.²⁷⁷ Most of these failed due to various reasons such as low 5-HT₃ receptor affinity or a mixed pharmacology. One promising analogue [¹²⁵I]DAIZAC ((S)-5-chloro-3-iodo-2-methoxy-N-(1-azobicyclo-[2.2.2]oct-3-yl)benzamide) displayed high affinity and very high selectivity for the 5-HT₃ receptor over a wide range of other CNS receptors tested. Zacopride is a well-known mixed 5-HT₃ receptor antagonist and 5-HT₄ agonist, but DAIZAC showed a selectivity for 5-HT₃ over 5-HT₄ receptors >120 times that of (S)-zacopride. Despite these promising features, the ligand has not yet been further developed for SPECT imaging.

Q. 5-HT₃ PET Radioligands

1. [¹¹C]MDL 72222

The first candidate PET ligand for imaging the 5-HT₃ receptor was [¹¹C]MDL 72222 ((8-methyl-8-azabicyclo[3.2.1]octan-3-yl)3,5-dichlorobenzoate), a selective 5-HT₃ receptor antagonist. In the rat brain and baboon brain an initial study showed that [¹¹C]MDL 72222 rapidly crossed the blood–brain barrier and was distributed throughout the brain. Prior treatment with unlabelled MDL 72222 did not displace [¹¹C]MDL 72222 in brain areas with 5-HT₃ receptors, suggesting a lack of specific binding, probably due to a combination of its high lipophilicity and the relatively low density of 5-HT₃ receptors.²⁷⁸

2. [¹¹C]YM060, [¹¹C]Y-25130, and [¹¹C]KF17643

Ishiwata et al.^279,280 radiolabelled several specific 5-HT₃ receptor antagonists: Y-25130 (azasetron): N-(1-azabicyclo[2.2.2]oct-3-yl)-6-chloro-4-methyl-3-oxo-3,4-dihydro-2H-1, 4-benzoxazine-8-carboxamide), YM060: (R)-5-[1-methyl-3-indolyl)carbonyl]-4,5,6,7-tetra-hydro-1H-benzimidazole, and KF17643: endo-8-methyl-8-azabicyclo[3.2.1]oct-3-yl 2-(n-propyloxy)-4-quinolinecarboxylate for PET imaging. [¹¹C]Y-25130 and [¹¹C]YM060 brain uptake in mice was low for both compounds, possibly due to low lipophilicity of [¹¹C]Y-25130, or ionization of the tertiary amine or both. In contrast to [¹¹C]Y-25130 and [¹¹C]YM060, [¹¹C]KF17643 exhibited high uptake in mice and appeared metabolically stable. However, no specific binding to the 5-HT₃ receptor was detected in mice.

3. [¹¹C]S21007

[¹¹C]S21007 (5-(4-benzylpiperazin-1-yl)4Hpyrrolo[1,2-a]thieno[3,2-e]pyrazine), a 5-HT₃ partial agonist with nanomolar affinity, has also been evaluated as a potential PET ligand in rat and baboon.^281,282 [¹¹C]S21007 failed to give a specific binding signal from the 5-HT₃ receptor in the baboon since blocking with unlabelled S21007 was not associated with any changes in [¹¹C]S21007 binding. Despite favorable kinetics, brain penetration, lack of contamination by labelled metabolites, high affinity, and selectivity [¹¹C]S21007 failed in detecting the 5-HT₃ receptors, and this points to the difficulty in measuring the rather low levels of 5-HT₃ receptors in the brain.

4. [¹⁸F]MR18445

After the failure of the first arylpiperazine derivative, [¹¹C]S21007, MR18445 (4-[4-(4-fluor-obenzyl)piperazino]-7-methoxypyrrolo[1,2-α]quinoxaline) was synthesized and examined as a PET radioligand for 5-HT₃ receptor imaging, this time with an ¹⁸F label. [¹⁸F]MR18445 uptake in rat brain was rapid and high, and there was no evidence of radiometabolites. However, neither autoradiography in rats nor PET studies in baboons revealed any specific binding.²⁸³

5. [¹¹C]NMQ

N-[methyl-¹¹C]Methylquipazine ([¹¹C]NMQ, (2-[1-(4-methyl)-piperazinyl]quinoline)), a derivative of quipazine, was examined for suitability as a PET radioligand in rat and nonhuman primate.²⁸⁴ In rat brain uptake and wash-out were very rapid. In the nonhuman primate brain kinetics were similar; initial rapid accumulation was followed by a slower decrease with all regions approaching that of nonspecific binding. Binding was high in 5-HT₃ receptor-rich regions, and although prior treatment with cold quipazine decreased binding in the medulla compared to the cerebellum, nonspecific binding was also high. Overall, this suggests that if the specific binding signal could be further improved, an arylpiperazine derivative might become a better radioligand for PET studies of the 5-HT₃ receptor than those previously reported.

R. 5-HT₃ Radioligands: Conclusions

Despite a number of research centres undertaking a concerted effort to develop 5-HT₃ selective PET and SPECT tracers, it seems that the very discrete localization and relatively low levels of 5-HT₃ receptors in the brain makes it a very difficult target to image in vivo. Ligands with even higher affinity and/or specific activity and lower nonspecific binding are required if this is to happen. More recently, several new benzoxazole derivatives have been synthesized with a view to PET radioligand development,²⁸⁵ but these have yet to be tested in vivo.

S. 5-HT₄ Receptor Tracers

5-HT₄ receptors are involved in learning and memory and are potential targets for the treatment of Alzheimer’s disease.²⁸⁶ Positively coupled to adenylate cyclase, downstream functions include regulation of GABA-A currents in cortical neurones, and enhancement of neurotransmitter release (e.g. Ach, DA, 5-HT, GABA). Various radioligands have been used to map them in brain tissue, showing that they are predominantly found in the mesolimbic and nigrostriatal system with highest densities in the caudate nucleus, putamen, nucleus accumbens, globus pallidus, and substantia nigra.^287–289 Potential PET ligands are beginning to emerge. So far, only one PET ligand, [¹¹C]SB207145 (8-amino-7-chloro-(N-[¹¹C]methyl-4-piperidylmethyl)-1,4-benzodioxan-5-carboxylate), has been successfully evaluated in humans.^290,291 A PET image of 5-HT₄ brain receptors obtained in healthy volunteers using [¹¹C]SB207145 is shown in Figure 1.

T. 5-HT₄ Radioligands

1. [¹²³I]SB207710

In parallel with the development of the very high affinity 5-HT₄ antagonist [³H]GR113808, Brown et al.²⁹² reported the binding characteristics of another potent, selective 5-HT₄ radioligand [¹²⁵I]SB207710 ([¹²⁵I](1-butylpiperidin-4-yl)methyl-8-amino-7-iodo-2,3-dihy-drobenzo[b][1,4]dioxine-5-carboxylate), and soon after its distribution was mapped by autoradiography in rat brain,²⁹³ confirming its potency, selectivity, and correlation with [³H]GR113808 labelled sites. Pike et al.²⁹⁴ characterized its efficacy as a radioligand in vivo, with the perspective of creating a SPECT ligand. In rats, [¹²⁵I]SB207710 entered the brain but had a rapid clearance rate. Radioactive accumulation correlated with the known in vitro 5-HT₄ receptor distribution, with a striatum-to-cerebellum ratio of 3.4. [¹²³I]SB207710 was then examined by SPECT in nonhuman primates. The tracer readily accumulated in 5-HT₄ receptor-rich areas, i.e., striatum with a target-to-background ratio of 4. The binding was displaceable by the antagonist SB204070, and cerebellum showed no specific binding. These results indicated that [¹²³I]SB207710 might become a suitable candidate for SPECT studies in humans, but further studies have not yet been published.

2. [¹¹C]SB207145

The 5-HT₄ receptor antagonist SB207145 was radiolabelled with ¹¹C by Gee et al.²⁹⁵ and evaluated for its potential as a PET radioligand for 5-HT₄ imaging. The first preliminary in vivo data showed that the radioligand readily entered the brain and was distributed according to known 5-HT₄ receptor distribution, and it was further shown that binding was sensitive to prior treatment with cold SB207145 or with the structurally dissimilar 5-HT₄ antagonist SB207040. A later study confirmed the promising characteristics of the radioligand in the pig brain, and further explored radioligand metabolism and binding kinetics. [¹¹C]SB207145 proved to have slow but reversible kinetics in high-binding brain regions, and using cerebellum as a reference region, the simplified reference tissue model was found to be the model of choice in the pig brain.²⁹⁶ [¹¹C]SB207145 was subsequently successfully evaluated in the human brain²⁹⁰ (Fig. 1). Here a binding potential of 3.1 was reported in caudate nucleus and binding was displaceable by the selective 5-HT₄ receptor inverse agonist piboserod (SB207266). In this study, a comprehensive quantification of the binding of [¹¹C]SB207145 to cerebral 5-HT₄ receptors in the human brain in vivo was further provided. Estimation of distribution volumes and binding potentials of [¹¹C]SB207145 showed good test–retest reproducibility and time-stability. The blocking study with piboserod confirmed that the cerebellum is a suitable reference region devoid of specific binding, and that non-specific binding is constant across brain regions. Subsequently, it was shown that [¹¹C]SB207145 is not displaceable by increased levels of endogenous 5-HT.²⁹¹ Together this shows that [¹¹C]SB207145 can be used for quantitative PET measurements of 5-HT₄ receptors in the human brain.

U. 5-ht₅ Radioligands

Two genes for the 5-ht₅ receptor have been described: A and B. The 5-ht_5a receptor is expressed in human brain, e.g. in the suprachiasmatic nucleus (SCN), but low levels of expression have made characterization very difficult. There are no selective agonists, but two selective antagonists have recently become available,²⁹⁷ one of which has been shown to alter circadian function in vivo.²⁹⁸ Only mRNA for the 5-ht_5b receptor has so far been found in humans and it is thought that no functional protein is produced due to stop codons.^299,300 There are no available radioligands for either of the 5-ht₅ receptors.

V. 5-HT₆ Radioligands

5-HT₆ receptors are found exclusively in the CNS and are predominantly expressed in the striatum, limbic system, and cortex.^301,302 They are of particular interest because of their involvement in learning and memory.⁴ Selective radioligands for in vitro studies are available, such as [¹²⁵I]SB258585, but its development as an in vivo imaging tool was hampered due to poor brain penetration.³⁰³ More recently, radioligands for PET imaging of the 5-HT₆ receptor have been developed by GSK, of which [¹¹C]GSK215083 3-[(3-fluorophenyl)-sulfonyl]-8-(4-[¹¹C]methyl-1-piperazinyl)quinoline is the most promising.

1. [¹⁸F]12ST05

Based on the selective 5-HT₆ antagonist SB271046 a novel series of compounds was developed. The specificities of these compounds toward the 5-HT₆ receptor were tested.³⁰⁴ This lead to the evaluation of [¹⁸F]12ST05 ([¹⁸F]-N-[2-(1-[(4-fluorophenyl)sulfonyl]-1H-indol-4-yloxy)ethyl]-N,N-dimethylamine) as the first potential 5-HT₆ PET radioligand.³⁰⁵ Synthesis technique, specific activity, and uptake kinetics were all found to be satisfactory, with the radioligand showing excellent brain penetration in rat and cat brain. In vitro, it had high affinity, and ex vivo autoradiography demonstrated high labelling in regions known to contain 5-HT₆ receptors, which to some extent was displaceable by unlabelled 12ST05 and SB258585. However, the initial evaluation in vivo in cat did not reveal any specific binding to the cerebral 5-HT₆ receptors, and does not seem to have been further developed for PET.³⁰⁵

2. [¹¹C]GSK215083

Recent conference proceedings identified two novel compounds that might be suitable for PET studies of the 5-HT₆ receptor. Based on the 3-benzenesulfonyl-8-piperazine-1-yl-quinoline scaffold, [¹¹C]GSK215083 and [¹¹C]GSK224558 (8-(4-[¹¹C]methyl-1-piperazinyl)-3-[(2-(trifluoromethoxy)phenyl)sulfonyl]quinoline) were both investigated in pig brain^306,307 and [¹¹C]GSK215083, the more superior ligand, subsequently showed promise in a human PET study.³⁰⁸ It has high affinity for 5-HT₆ (in vitro K_i, 0.16 nM), but also has high 5-HT_2A affinity (in vitro K_i, 0.79 nM). However, the differential localization of 5-HT_2A and 5-HT₆ receptors (predominantly cortical and striatal respectively) combined with the ~5-fold difference in affinity meant that it was still possible to discriminate between these two receptor types in vivo. A recent study in nonhuman primates demonstrated this; [¹¹C]GSK215083 binding was preferentially blocked by unlabelled GSK215083 within striatum relative to cortex, suggesting an ~8-fold difference in affinity in vivo. Concerns of a possible mass dosing effect were also addressed. Although BP values are relatively low, [¹¹C]GSK215083 represents the only available 5-HT₆ PET radioligand to date.³⁰⁹

W. 5-HT₇ Radioligands

5-HT₇ receptors are found in several brain regions such as hippocampus, thalamus, and hypothalamus, especially the SCN, where they are thought to be involved in circadian rhythm, sleep, and mood regulation.^310,311 Three different splice variants of the human 5-HT₇ receptor have so far been found, but they do not seem to differ markedly in their pharmacological profile.^312,313 Several potent and selective ligands for the 5-HT₇ receptor have been developed, but so far only [¹¹C]DR4446 has been prepared and evaluated in vivo as a potential 5-HT₇ PET radioligand, which was unsuccessful. A recent attempt to produce a fluorinated radioligand also failed; the lead compound demonstrated very low specific binding in ex vivo autoradiography.³¹⁴

1. [¹¹C]DR4446

DR4446 (1-methyl-2a-[4-(4,5,6,7-tetrahydrothieno[3,2-c]pyridin-5-yl)butyl]-2a,3,4,5-tetrahydro-1H-benz[cd]indole-2-one) was shown to possess moderate affinity (K_i = 9.7 nM) and high selectivity for 5-HT₇ receptors over other 5-HT receptors.³¹⁵ [¹¹C]DR4446 was therefore synthesized and tested in the nonhuman primate brain,³¹⁶ revealing good brain penetrability, metabolic stability, but a minimal specific binding component. No reports on [¹¹C]DR4446 in humans have been published.

X. SERT Tracers

SERT is the serotonin reuptake transporter and is one of three monoamine reuptake transporters in the brain, along with transporters for dopamine (DAT) and noradrenaline (NET).^317,318

Because of the interest in SERT generated by the success of its inhibitors in the treatment of depression and anxiety disorders, successful imaging of SERT in living human brain was for a long time the topic of intensive investigation, and several PET and SPECT ligands have been developed for this purpose. A recent, very detailed review of SERT imaging by PET and SPECT can be found in Huang et al.³¹⁹ so the whole spectrum of radioligands available will not be discussed at great length here.

Many of the ligands initially tested as PET and SPECT radioligands were labelled derivatives of successful antidepressant drugs like the SSRI paroxetine,³²⁰ the SNRI venla-faxine,³²¹ and the tricyclics imipramine³²² and cyanoimipramine.³²³ These radiotracers were, however, unsuccessful in vivo, mainly due to insufficient specific to nonspecific binding in the human brain.

Initially, better images of SERT in the human brain came from the cocaine derivative SPECT ligand [¹²³I]β-CIT, and later with the selective but kinetically irreversible PET ligand [¹¹C](+)McN5652. Today, the most successful line of SERT radioligands are those developed from diarylsulfides such as [¹²³I]ADAM, and especially [¹¹C]MADAM, and [¹¹C]DASB. A PET image of 5-HT transporter binding obtained in healthy volunteers using [¹¹C]DASB is shown in Figure 1.

Y. SPECT Radioligands for SERT

1. β-[¹²³I]CIT and Nor-β-[¹²³I]CIT

β-[¹²³I]CIT (2-β-carbomethoxy-3-β-(4-iodophenyl)tropane) is a nonselective monoamine transporter SPECT ligand that has approximately equal affinity for all three transporters. It was developed to improve the metabolic instability of the PET ligand [¹¹C]cocaine, and was evaluated as a SPECT ligand in both nonhuman primates³²⁴ and in humans.³²⁵ Despite its nonselectivity, β-[¹²³I]CIT has been used for imaging both DAT and SERT, by taking advantage of the differential localization of these transporters (striatum and midbrain, respectively) and the different tracer kinetics in these regions. Various analogues of β-[¹²³I]CIT have been made through modest modifications of the structure, such as labelling at different positions and creation of fluoroalkyl analogues to decrease the time to reach pseudoequilibrium,³²⁶ but these efforts were mainly directed at improving its effectiveness for imaging the dopamine reuptake transporter (DAT). The development of nor-β-[¹²³I]CIT (2-β-carbomethoxy-3-β-(4-iodophenyl)nortropane)^327,328 produced a radioligand with supposedly increased specific binding for SERT over DAT in humans, but it remains controversial whether nor-β-[¹²³I]CIT is a better radioligand for SERT than β-[¹²³I]CIT.³²⁹

Today, several hundred publications (Table II shows publications only related to SERT) using β-[¹²³I]CIT demonstrate the level of success this radioligand has enjoyed especially for measuring the dopamine transporter. β-[¹²³I]CIT and nor-β-[¹²³I]CIT have been used to image SERT density in human midbrain in a number of conditions such as depression,^330,331 generalized anxiety disorder,³³² obsessive–compulsive disorder,³³³ panic disorder,³³⁴ and Parkinson’s disease,³³⁵ among others. The disadvantage of these SPECT ligands is their inability to selectively image SERT.

2. Diarylsulfides

The diarylsulfides are a class of potent SSRIs that so far have given rise to three potential SPECT ligands: [¹²³I]IDAM,^336–338 [¹²³I]ODAM,³³⁹ and [¹²³I]ADAM.^340,341

3. [¹²³I]ADAM

The most successful of the diarylsulfide SPECT radioligands listed above is [¹²³I]ADAM (2-((2-((dimethylamino)methyl)phenyl)thio)-5-iodophenylamine), which has been shown to be potent, selective, and have a high target-to-background ratio in human studies.^340–343 It has been used for estimating SERT occupancy of SSRIs^344–348 and to demonstrate changes in SERT binding density in depression,³⁴⁹ migraine,³⁵⁰ borderline personality disorder,³⁵¹ and night eating syndrome.³⁵² Quantification of the SERT binding with [¹²³I]ADAM SPECT is most often done with a ratio method, based on data acquired from 200 to 240 min. This has, however, been shown to overestimate the specific binding by about 10%. The most reliable outcome is based on a 0–120 min SPECT acquisition followed by Logan noninvasive modelling.³⁴³

Z. PET Radioligands for SERT

1. [¹¹C]McN5652

The first selective PET radioligand for imaging SERT in the human brain was [¹¹C](+) McN5652 (trans-1,2,3,5,6,10-β-hexahydro-6-[4-(methylthio)phenyl]pyrrolo-[2,1-a]isoquino-line).³⁵³ Its use has been limited by a relatively low target-to-background ratio in vivo, as well as a slow brain uptake and irreversible kinetics that complicate quantification in high binding regions.^354,355 Despite this, [¹¹C](+)McN5652 PET has been used to investigate SERT binding in ecstasy use,³⁵⁶ impulsive aggression,³⁵⁷ and depression.³⁵⁸ In a series of experiments, the fluoromethyl analogue of [¹¹C](+)McN5652 was developed and evaluated as a ¹⁸F-labelled PET radioligand³⁵⁹ and it was concluded that [¹⁸F](+)-FMe-McN5652 has better features than [¹¹C](+)McN5652 for SERT imaging with PET. No further studies using [¹⁸F](+)-FMe-McN5625 have been forthcoming, perhaps because of the subsequent success of the diarylsulfides.

2. Diarylsulfides

Many PET radioligands for measuring SERT have been developed from the diarylsulfide series including [¹¹C]ADAM,³⁶⁰ [¹⁸F]ADAM,^361,362 [¹¹C]S-Me-ADAM,³⁶³ [¹¹C]DAPA,³⁶⁴ [¹¹C]AFM,³⁶⁰ [¹¹C]AFA,³⁶⁵ [¹¹C]EADAM,³⁶⁶ [¹¹C]DAPP,³⁶⁷ [¹¹C]AFE,³⁶⁸ [¹¹C]MADAM,³⁶⁹ and [¹¹C]DASB.³⁷⁰ All these ligands demonstrated selectivity for SERT, with varying target-to-background ratios and brain kinetics. Following a comparison between [¹¹C]ADAM, [¹¹C]DASB, [¹¹C]DAPA, and [¹¹C]AFM in reference to [¹¹C]McN5652, [¹¹C]DASB emerged as a favourite due to its ability to measure SERT binding within a short scan time.³⁶⁰ Today [¹¹C]DASB and [¹¹C]MADAM have become the most popular SERT PET radioligands.

3. [¹¹C]DASB

[¹¹C]DASB (3-amino-4-(2-dimethylamino-methyl-phenylsulfanyl)-benzonitrile) is one of a series of ¹¹C-labelled arylthiobenzylamines developed by Wilson and Houle.³⁷¹ [¹¹C]DASB displayed good selectivity (negligible affinity for more than 40 other receptors) and high affinity. Ex vivo studies further demonstrated saturable binding in rats.³⁷⁰ PET studies of [¹¹C]DASB in humans showed that the uptake distribution matched that expected for SERT and binding kinetics could be quantified with reference tissue models.^367,372

It was quickly realized that [¹¹C]DASB was a superior PET ligand to [¹¹C](+)McN5652, both in baboons³⁷³ and in humans³⁵⁴ but also superior to several other promising new PET ligands.³⁶⁰ Furthermore, test–retest data showed high reproducibility and reliability.³⁷⁴ [¹¹C]DASB PET has since been used to measure SERT occupancy at clinical doses of fluoxetine, citalopram, sertraline, duloxetine, and venlafaxine.^32,375–377 These were followed by patient studies, investigating a multitude of conditions including MDMA use,³⁵⁶ depression,^378–380 schizophrenia,³⁸¹ obsessive–compulsive disorder,³⁸² alcoholism,³⁸³ and bipolar disorder.³⁸⁴ [¹¹C]DASB binding has further been investigated in healthy individuals in relation to personality traits,³⁸⁵ seasonal changes,³⁸⁶ and familial risk for mood disorders.³⁸⁷

4. [¹¹C]MADAM

After demonstrating promising characteristics in nonhuman primate PET studies,³⁶⁹ [¹¹C]MADAM (N,N-dimethyl-2-(2-amino-4-methylphenylthio)benzylamine) was tested in humans.³⁸⁸ The brain accumulation pattern was consistent with human postmortem SERT binding with [³H]MADAM³⁸⁹ as well as [³H]imipramine, [³H]paroxetine, and [³H]citalopram.^390,391 In a test–retest study, [¹¹C]MADAM had excellent reproducibility when done several weeks apart.³⁹² Since then, [¹¹C]MADAM has been used to estimate relative SERT occupancy of citalopram and escitalopram³⁹³ and to investigate the relationship between 5-HT_1A and SERT binding in healthy young men and women.^394–397

5. [¹⁸F]ADAM

The ¹⁸F-labelled form of ADAM, 4-[¹⁸F]ADAM (N,N-dimethyl-2-(2-amino-4-[¹⁸F]fluor-ophenylthio)benzylamine) has been synthesized and a synthesis methodology suitable for PET imaging has been reported.^398,399 With microPET and autoradiography, [¹⁸F]ADAM has a rapid brain uptake and high target-to-background ratios and the specific signal was displaceable by paroxetine.⁴⁰⁰ Further studies in rat and monkey PET confirmed that bio-distribution, dosimetry, and toxicity data support its investigation in human subjects.⁴⁰¹

AA. SERT Radioligands: Conclusions

Several SERT radioligands are available and among these the most successful are [¹²³I]ADAM, [¹¹C]DASB, and [¹¹C]MADAM. The most widely used PET SERT radioligand is [¹¹C]DASB because of its good selectivity, high reproducibility, and simple quantification. The primary reason why new radioligands are still being developed is that the low density of SERT binding sites in neocortex is challenging for accurate measurement. [¹¹C]AFM with its high target-to-background ratio might show improved detection of cortical SERT density, and thus be a valuable addition to the existing SERT radioligands. It would also be valuable to have an ¹⁸F-labelled SERT radioligand, and [¹⁸F]ADAM shows promise in this respect.

BB. 5-HT Synthesis Radiotracers

Some effort has been devoted toward developing PET tracers for probing presynaptic 5-HT synthesis. 5-HT in the brain is synthesized from dietary tryptophan, and tracer candidates should thus be analogues of tryptophan and follow the 5-HT synthesizing pathway. α-[¹¹C]-methyl-L-tryptophan ([¹¹C]AMT or α[¹¹C]MTrp) and 5-hydroxy-L-[β-¹¹C]tryptophan ([¹¹C]HTP) are substrates in the first and second enzymatic steps in the biosynthesis of 5-HT, and have both been tested for their ability to determine 5-HT synthesis rate in the human brain. A PET image of sites of 5-HT synthesis within brain using [¹¹C]AMT in healthy volunteers is shown in Figure 1.

1. [¹¹C]AMT

[¹¹C]AMT is a ¹¹C-labelled methyl analogue of tryptophan and a substrate for tryptophan hydroxylase. [¹¹C]AMT is not used for protein synthesis and is not catabolized by the monoamine oxidases, and is thus trapped by the 5-HT synthesis pathway.⁴⁰² The use of [¹¹C]AMT to measure 5-HT synthesis rate was evaluated ex vivo⁴⁰³ and also in vivo in animals.^404,405 The first study in humans showed that uptake values were stable within an individual, which was taken as an indication of the potential of [¹¹C]AMT to serve as an index of individual 5-HT synthesis capacity.⁴⁰⁶ Furthermore, changes in the uptake with age and gender were found to be consistent with previously reported biochemical measurements of 5-HT in brain tissue.⁴⁰⁷ There are, however, several caveats with the use of [¹¹C]AMT. The signal-to-noise ratio is rather low, there is a time delay between uptake and conversion of the radiotracer,^405,408 and in some elegant nonhuman primate studies it has been found that [¹¹C]AMT binding seems to be primarily driven by blood–brain barrier exchange, rather than 5-HT synthesis rate.⁴⁰⁹ Further, under pathologic conditions, [¹¹C]AMT is metabolized by means of the kynurenine pathway and the measurements thus reflect both pathways.⁴¹⁰

Clinically, [¹¹C]AMT has been used to determine changes in the 5-HT synthesis rate during acute changes in mood⁴¹¹ and during antidepressant treatment,⁴¹² but its main use is now in the study of pathological conditions such as epilepsy and endocrine tumors where uptake can be marked so it assists in diagnoses.

2. [¹¹C]HTP

Conversion of tryptophan to 5-hydroxytryptophan (5-HTP) is the rate-limiting step in 5-HT synthesis, but neither tryptophan hydroxylase nor amino acid decarboxylase is saturated in the process of 5-HT synthesis. Therefore, in principle, measurement of either enzymatic step may be used to determine the rate of 5-HT synthesis. Further, since the endogenous level of 5-HTP in the brain is normally very low, the rate of decarboxylation to 5-HT is most probably equal to its formation. This has lead to the use of [¹¹C]HTP in PET as a measure of 5-HT synthesis rate.⁴¹³ In vivo nonhuman primate brain studies showed that following injection of [¹¹C]HTP, the accumulation of its radiometabolite decreases following co-injection with unlabelled 5-HTP⁴¹⁴ and increases after co-administration of vitamin B6,⁴¹⁵ a cofactor for the decarboxylase, indicating that the radiotracer is sensitive to pharmacological challenges. In humans, [¹¹C]HTP utilization rates were shown to be region-dependent and could be altered by serotonergic pharmacological challenges,⁴¹⁶ suggesting that data from PET studies using [¹¹C]HTP might be interpreted as a measure of brain 5-HT turnover. A later study showed that human PET data from this tracer could be reliably modelled when using a model that included irreversible trapping of the tracer.⁴¹⁷ As for [¹¹C]AMT, the majority of studies using [¹¹C]HTP are in diagnostic investigations in pathological conditions rather than investigating 5-HT synthesis per se.

[¹¹C]AMT and [¹¹C]HTP binding were found to correlate in some postmortem studies,^417,418 but not in a head-to head comparison study in rhesus monkeys.⁴¹⁹ More in vivo validation studies are required if firm conclusions are to be drawn about the suitability of these tracers for estimating 5-HT synthesis.

4. CONCLUSIONS

Currently, well-validated radiotracers for imaging the 5-HT system exist as antagonist radioligands for the 5-HT_1A, 5-HT_1B, 5-HT_2A, and 5-HT₄ receptors, and for SERT.

Three radioligands are currently used for PET studies of the 5-HT_1A receptor: [¹¹C]WAY100635, [¹⁸F]MPPF, and [¹⁸F]FCWAY. There are four suitable tracers for the 5-HT_2A receptor: the SPECT tracer [¹²³I]5-I-R91150, the nonselective PET tracer [¹⁸F]setoperone, and the selective tracers [¹⁸F]altanserin and [¹¹C]MDL 100,907. Several SERT radioligands exist and among these the most successful are [¹²³I]ADAM, [¹¹C]DASB, and [¹¹C]MADAM.

Several promising radioligands are currently emerging. Of these, the two 5-HT_1B ligands [¹¹C]AZ10419369 and [¹¹C]P943, the 5-HT₄ receptor radioligand [¹¹C]SB207145, and the 5-HT₆ receptor radioligand [¹¹C]GSK215083 are the only ones where successful evaluations in humans have been published so far.

Unfortunately, despite significant investment and many failed attempts, there are currently no PET or SPECT radioligands for the study of 5-HT₃ receptor in humans. In the past, a lack of selective ligands appears to be the reason for the lack of radioligand development programs for 5-HT_1D/e/F, 5-HT_{2B/C, 5-HT5}, and 5-HT₇ receptors. However, there are now a number of selective ligands available for many of these targets, which should facilitate renewed effort to tracer development at these targets. Emerging functions for these receptors make them interesting targets for future study with in vivo brain imaging.

From the present review it is clear that development of radioligands for in vivo brain imaging in humans is a long process of trial and error. Despite substantial effort to predict the usefulness of a potential radioligand for in vivo brain studies, most radiotracers fail during tracer development. Among the main reasons for this are insufficient blood–brain barrier passage, too high nonspecific binding, inability to displace the binding of the radioligand with an unlabelled target-specific ligand, production of lipophilic radiometabolites, or kinetic properties that complicate quantification. Preclinical evaluations in both rats and larger species are useful for selecting the best candidates, but often species differences make it difficult to directly translate results from animals to the human situation.²⁸ Therefore, one cannot firmly conclude on the suitability of a radioligand for human studies before a thorough evaluation in humans has been performed, including blocking, modelling, and test–retest studies.

It is also evident from this review that once a radioligand for imaging of a serotonergic target has been successfully evaluated, the interest in using the radioligand for studies of a whole range of human conditions and for establishing receptor occupancy from clinical doses of drug compounds is huge. In addition, the potential use of agonist tracers for imaging “functional” receptors in vivo may provide us with additional information that cannot be obtained using antagonist tracers. They may also be more susceptible to changes in endogenous neurotransmitter, a property that may provide new avenues to understanding disease. With this in mind we strongly encourage the continuing efforts to develop radioligands for imaging of serotonergic targets.

Biographies

Dr. Louise M. Paterson trained in Pharmacology at the University of Bristol, UK, where she obtained her B.Sc. (2000). She continued her studies at Bristol and was awarded a Ph.D. in Pharmacology (2004), characterizing the imidazoline-2 binding site and its relationship with monoamine oxidase. This was followed by a series of post-doctoral positions in translational pharmacology at the University of Bristol between 2004 and 2009. These projects combined pre-clinical laboratory work with healthy volunteer and patient studies, developing her expertise in sleep and depression and the involvement of the 5-HT system. She recently moved to Imperial College London, where she is pursuing her interest in combining translational medicine with human brain imaging. She has recently published a critical review of the use of PET and SPECT to measure the release of endogenous 5-HT.

Dr. Birgitte R. Kornum gained a B.Sc. degree in Chemical Engineering from the Danish Technical University (2000) and continued her training at the University of Copenhagen, where she gained an M.Sc. degree in Human Biology (2003) and a Ph.D. degree from the School of Neuroscience (2009). Her Ph.D. project focussed on using the pig as a model in brain research, and using this model she, among other things, performed in vitro and in vivo evaluations of new radiotracers for PET imaging. She now holds a post doc position at Stanford Center for Sleep Sciences.

Professor David J. Nutt (DM, FRCP, FRCPsych, FMedSci) received his undergraduate training in medicine at Cambridge and Guy’s Hospital, and continued training in neurology to MRCP. After completing psychiatric training in Oxford, he continued as a lecturer and later as a Wellcome Senior Fellow in psychiatry. He then spent two years as Chief of the Section of Clinical Science in the National Institute of Alcohol Abuse and Alcoholism in NIH, Bethesda. In 1988, he returned to the UK to set up the Psychopharmacology Unit at the University of Bristol, an interdisciplinary research group spanning Psychiatry and Pharmacology before moving to Imperial College London in 2008 where he is currently the Edmund J. Safra Professor of Neuropsychopharmacology and Director of the Neuropsychopharmacology Unit. His main research interests are in merging pharmacological approaches with brain imaging techniques, in particular PET imaging of brain disorders such as addiction and affective disorders.

Dr. Victor W. Pike trained as a chemist at the University of Birmingham (UK), where he gained his B.Sc. (1972) and Ph.D. degrees (1976). After a postdoctoral position, he entered a Staff Scientist position at the MRC Cyclotron Unit (Hammersmith Hospital/Imperial College, UK) in 1978 where he later became Head of its Chemistry and Engineering Section. In 2001 he moved to the National Institutes of Health (Bethesda, Maryland) to become Chief of the PET Radiopharmaceutical Sciences Section in the newly established Molecular Imaging Branch of the National Institute of Mental Health, where he has continued his primary interests in all aspects of PET radiotracer development, and especially radiotracers for brain imaging.

Professor Gitte M. Knudsen received her training as a neurologist at the University hospitals in Copenhagen, and obtained her D.M.Sc. degree in 1994 (Application of the double-indicator technique for measurement of blood-brain barrier permeability in humans). She subsequently conducted research within cerebrovascular regulation and since 1998 her main research interest has been within molecular brain imaging, in particular of the 5-HT system. She is the Director of the Neurobiology Research Unit (NRU) and of the Center for Integrated Molecular Brain Imaging (Cimbi) and is employed at Rigshospitalet and Professor at the University of Copenhagen.