The role of the serotonergic system and the effects of antidepressants during brain development examined using in vivo PET imaging and in vitro receptor binding

Abstract

Serotonin (5-HT) and the serotonergic system, which includes the serotonin transporter (SERT) and the two G protein-coupled 5-HT1A and 5-HT1B receptors, are implicated in the pathophysiology and treatment of several neuropsychiatric disorders including major depressive disorder (MDD) and anxiety. Two classes of antidepressants—selective serotonin reuptake inhibitors (SSRIs), which block SERT, and tricyclic antidepressants (TCAs), which block several monoamine transporters—alter 5-HT levels and modulate the serotonergic system. However, the transient suicidal ideation associated with SSRIs led the U.S. Food and Drug Administration to issue a black-box warning in 2004 on their use in juveniles. In addition, both SSRIs and TCAs exhibit high response variability; roughly only one-third of patients treated with these agents undergo remission. This variability in response is believed to arise from gene-environment interaction; this includes events that unfold during early-life development, and may subsequently influence behavior, treatment outcome, and serotonergic function. My PhD thesis explored several facets of the serotonergic system and its involvement in MDD and anxiety, including: 1. the long-term effects of fluoxetine (an SSRI) administered during brain development in rhesus monkeys imaged with positron emission tomography (PET) for two serotonergic markers: SERT and 5-HT1A receptors (Paper I); 2. the antidepressant effects of escitalopram (an SSRI) or nortriptyline (a TCA) on 5-HT1A/1B receptors in the rat brain, explored using a gene-environment model (Paper II); 3. the functionality and selectivity of [11C]CUMI–101, a 5-HT1A receptor PET radioligand, in rodent and primate brain (Paper III) ; and 4. the disparate findings regarding 5-HT1A receptor PET imaging studies in MDD, and their implications for research studies (Paper IV).

Paper I (Am J Psychiatry, 2014) examined the long-term effects of fluoxetine administered to juvenile rhesus monkeys who, as young adults, were imaged with two selective PET radioligands: [11C]DASB for SERT, and [11C]RWAY for 5-HT1A receptors. An equal number of monkeys separated from their mothers at birth—an animal model of human childhood stress—were also studied. At birth, 32 male rhesus monkeys were randomly assigned to either maternal separation or normal rearing conditions. At age two, half (N = 8) of each group was randomly assigned to fluoxetine or placebo for one year. To eliminate the confounding effects of residual drug, monkeys were scanned at least 1.5 years after drug discontinuation. Social interactions were assessed both during and after drug administration. Results indicated that fluoxetine persistently upregulated SERT, but not 5-HT1A receptors, in both neocortex and hippocampus. Whole-brain, voxel-wise analysis localized SERT upregulation in lateral temporal and cingulate cortices. Neither maternal separation by itself nor the rearing-by-drug interaction was significant for either radioligand. Fluoxetine did not significantly affect behavioral measures.

In order to investigate the issue of response variability, Paper II (Neurosci Lett., 2014) examined the antidepressant effects of two agents—escitalopram (an SSRI) and nortriptyline (a TCA)—on 5-HT1A and 5-HT1B receptors in the rat brain using a gene-environment model. The effects of genetic vulnerability were modeled using the Flinders Sensitive Line, a rat model of depression and its control counterpart, the Flinders Resistant Line. The effects of environmental vulnerability were modeled using the maternal separation paradigm. Rats (n=105) were drawn from four groups reflecting experimental crossing of strain and early-life stress to assess the effects of escitalopram or nortriptyline compared to vehicle. Quantitative in vitro autoradiography was performed using [125I]MPPI (5-HT1A) and [125I]CYP (5-HT1B) in prefrontal cortex (PFC) and hippocampus. Stringent, Bonferroni-corrected statistical analyses showed significant strain-by-rearing-by-treatment interactions in PFC 5-HT1A and hippocampal 5-HT1B receptors. Either genetic or environmental vulnerability reduced serotonergic binding. However, the effects of vulnerable genotype and environment were not additive. Antidepressants, in general, increased serotonergic binding.

Paper III (J Nucl Med., 2014) assessed the functionality and selectivity of CUMI-101 using in vivo PET and in vitro receptor binding. Initially, [11C]CUMI–101 was reported in 2007 as a putative agonist, selective for the 5-HT1A receptor. Based on this, we also scanned 32 monkeys with [11C]CUMI–101 to compare the results between an agonist vs. an antagonist PET radioligand for the 5-HT1A receptor. Intriguingly, a 2011 study reported that CUMI-101 exhibited potent antagonistic behaviors at 5-HT1A receptors in the rat brain. Using [35S]GTPγS functional assay, we replicated this finding in rats, and extended the study to also include primates. In primates, CUMI-101, unlike 8-OH-DPAT, which is a 5-HT1A receptor agonist, did not stimulate [35S]GTPγS binding even up to a concentration of 10 μM. In addition, both in vivo PET and in vitro receptor binding demonstrated that CUMI-101 exhibited cross-reactivity with α1 adrenoceptors. In vitro binding to α1 adrenoceptors was highest in thalamus (>45%) and lowest in neocortex and cerebellum (<10%) in rat, monkey, and human brain. In vivo uptake of [11C]CUMI-101 was completely blocked only when both WAY-100635 (a 5-HT1A receptor antagonist) and prazosin (an α1 adrenoceptor antagonist) were co-administered.

Paper IV (NeuroImage, 2012) critically reviewed factors that may have led to the disparate findings surrounding 5-HT1A receptor imaging in MDD using the PET radioligand [carbonyl- 11C]WAY-100635. Two key confounding ‘technical’ factors were highlighted: the use of the cerebellum as a reference region, and the imprecision of measuring the concentration of parent radioligand in arterial plasma—the method considered to be the ‘gold standard’. The notion that these technical factors may have confounded results is underscored by the finding that different results were obtained from the same study cohort depending on whether the outcome measure was normalized to cerebellar gray matter or plasma free fraction. The fact that not every study obtained plasma free fraction precluded a meta-analysis. Our study recommends that, in the future, individual centers acquire data using the ‘gold standard’ arterial input to address methodological concerns. This would also allow researchers to meaningfully pool the data, allowing them to reach a consensus regarding putative alterations in 5-HT1A receptor function in MDD.

Taken together, the findings of these four papers lead to some important conclusions regarding the role of the serotonergic system in the pathophysiology and treatment of MDD and anxiety. Our finding that fluoxetine leads to persistent SERT upregulation (Paper I) underscores the need for practitioners to exercise caution in prescribing SSRIs to juveniles, though no concrete implications should be drawn regarding whether the persistence of these effects is ultimately ‘good’ or ‘bad’. Paper II sheds light on the manner in which complex gene/environment interactions shape how antidepressants affect 5-HT1A/1B receptors in the brain, an issue that contributes directly to research regarding the etiology and pathophysiology of MDD, and the complex interaction between the disorder and the antidepressants used to treat it. Paper III provides important evidence that [11C]CUMI-101 is neither an agonist nor selective for the 5-HT1A receptor.

Finally, the evidence presented in Paper IV makes key recommendations regarding the pooling of data necessary for future human PET studies to enter a more collaborative phase. This would allow researchers to clarify discrepancies and advance psychiatric research, particularly given the planning and labor-intensive costs of achieving these goals.