Novel Pharmacological Treatments for Generalized Anxiety Disorder: Pediatric Considerations

Abstract

Pediatric anxiety disorders such as generalized anxiety disorder (GAD) are common, impairing, and often undertreated. Moreover, many youth do not respond to standard, evidence-based psychosocial or psychopharmacologic treatment. An increased understanding of the gamma-aminobutyric acid (GABA) and glutamate neurotransmitter systems has created opportunities for novel intervention development for pediatric GAD. This review examines potential candidates for pediatric GAD: eszopiclone, riluzole, Eglumegad (LY354740), pimavanserin, agomelatine. The pharmacology, preclinical data, clinical trial findings and known side effects of these compounds are reviewed, particularly with regard to their potential therapeutic relevance to pediatric GAD. Notwithstanding numerous challenges, some of these agents represent potential candidate drugs for pediatric GAD. Further treatment development studies of agomelatine, eszopiclone, pimavanserin and riluzole for pediatric GAD also have the prospect of informing the understanding of GABAergic and glutamatergic function across development.

Introduction

Pediatric anxiety disorders affect up to 20% of youth and contribute to social impairment, poor academic performance, and family stress (Essau et al., 2000). Anxiety disorders increase the risk of developing depressive disorders (Pine, 1999), substance use disorders, and suicidality (Asselmann et al., 2018; Husky et al., 2012). Generalized anxiety disorder (GAD) is one of the more common anxiety disorders in childhood and is characterized by diffuse, difficult-to-control anxiety that typically emerges during adolescence (Beesdo-Baum & Knappe, 2012).

Putative variations in the serotonin transporter gene (5-HTT) may predispose children and adolescents to developing GAD (Tsuang et al., 2004). However, the neurochemical basis of GAD in children and adolescents is likely more wide-ranging and complex with contributions from serotonin, norepinephrine, glutamate, and GABA systems. This complexity creates clinical challenges in precision medicine approaches to tailoring treatment in youth with GAD but also presents immense research opportunities for developing new pharmacological treatments. Neurodevelopmental considerations are important in considering the pathophysiology of anxiety as the glutamatergic and GABAergic neurotransmitter systems also undergo substantial changes, potentially facilitating more heterogeneity in the clinical picture, pathophysiology, and pharmacological response in childhood GAD (Pine, 1999).

Current, evidence-based pharmacologic treatments for GAD as main diagnosis in children and adolescents are limited to selective serotonin reuptake inhibitors (SSRIs) and serotonin norepinephrine reuptake inhibitors (SNRIs) (Dobson et al., 2019). Duloxetine is the only antidepressant approved by the FDA for pediatric GAD (Jeffrey R Strawn et al., 2015).

Despite the promise of novel therapeutics, there are significant challenges to developing more effective treatments. These challenges relate not only to the fluctuating course of anxiety disorders but to developmental factors and inherent limitations of clinical trials design. In youth, anxiety symptoms—like depressive symptoms—tend to fluctuate over time (Costello et al., 2005). Thus, a group of anxious youth may have symptoms and variance of symptom measures as well as the autocorrelation of symptoms that are determined by non-treatment. From a clinical trial design standpoint, even with current statistical approaches, some degree of stationarity is necessary to detect treatment-related improvement. Treatments are also largely borrowed by those that have demonstrated efficacy in adults with anxiety disorders (J. R. Strawn et al., 2018).This approach assumes that the pathophysiology and by extension the neuropharmacology of these disorders is the same in pediatric and adult populations. However, differences in neurocircuitry have been observed in anxious adults relative to anxious youth. Measures of anxiety symptoms that have been used in pediatric (and adult studies) are encumbered by issues related to unidimensionality and over representation of some symptoms (e.g., somatic symptoms) relative to psychic symptoms or functional impairment (e.g., relationship dysfunction and avoidant behavior). Importantly, problems introduced by the unidimensionality of these measures are amplified in the pediatric population that often has greater co-morbidity (Kendall et al., 2010).

This review examined potential investigational pharmacologic treatments for pediatric GAD with the objective of informing related, clinical research, and drug development. Embase and Medline from 2008 to 2019 were searched with the assistance of a research librarian using keywords listed for this review. Studies and manuscripts that were published in English were included. The reference list was finalized in March 2019 through review of initial literature search and discussion among all authors. The reference list of each manuscript was reviewed for additional potential references. Additional manuscripts were reviewed and included by PEC and JS as deemed necessary for a comprehensive manuscript.

(1). Rationale for GABA and glutamate neurotransmitter systems for pediatric GAD

In pre-clinical models, stress enhances excitatory neurotransmission, mediated by elements of the glutamatergic neurotransmission system and chronic stress increases vulnerability to developing anxiety disorders (Simon & Gorman, 2006). Glutamatergic tone in the anterior cingulate cortex correlates with the severity of anxiety symptoms in adolescents with GAD (Jeffrey R Strawn et al., 2014). Thus, glutamate-modulating agents may have a role in treating severe anxiety disorders. Acute administration of N-Methyl-D-Aspartic acid (NMDA) receptor antagonists has anxiolytic effects in preclinical and clinical models (Javitt, 2004; Plaznik et al., 1994). Conventional antidepressants and NMDA antagonists decrease glutamatergic neurotransmission without histologic changes in the neuron which may be associated with remission from depression or anxiety disorders via restoration of a basal rate of neurogenesis (Duman, 2002; Jacobs et al., 2000; Manji et al., 2003; Santarelli et al., 2003). Notably, prior work suggests that glutamatergic and GABAergic balances may regulate the expression of BDNF (Marmigère et al., 2003).

(2). Promising Drugs

Eszopiclone

Eszopiclone, a γ-aminobutyric acid (GABA_A) receptor agonist, is rapidly absorbed after oral administration and extensively metabolized by CYP3A4 and CYP2E1. N-desmethylzopiclone, the pharmacologically active primary metabolite, is distributed in brain and among other tissues and then eliminated via urine and saliva. In pediatric studies of eszopiclone, body weight significantly affects clearance and volume of distribution (Maier et al., 2011; Sunovian Pharmaceuticals Inc., 2012). With a 7.5 mg dose, the time to peak concentration (T_MAX) is 1 hour and the t_½ is 6 hours. As eszopiclone is weakly protein bound, drug interactions related to protein binding are not concerning. CYP3A4 and CYP2E1 play a role in the metabolism of eszopiclone. Eszopiclone does not inhibit CYP 1A2, 2A6, 2C9, 2C19, 2D6, 2E1, or 3A4 (Sunovion Pharmaceuticals Inc., 2014).

In pre-clinical rodent models of insomnia, eszopiclone decreases acetylcholine release in sleep promoting regions of pontine reticular formation (Hambrecht-Wiedbusch et al., 2010) and in the amygdala suggesting its potential in regulation of fear and anxiety as well as learning, memory, sleep and autonomic control (Hambrecht-Wiedbusch et al., 2014). In terms of anxiety, low dose eszopiclone (1–10 mg/kg) had promising effects controlling anxiety and anxiety related insomnia on rats (Huang et al., 2010). Eszopiclone may also facilitate neurogenesis. Co-administration of eszopiclone with fluoxetine significantly increased survival of marked newborn cells in the adult rat dorsal hippocampus (Su et al., 2009). In another trial twice daily eszopiclone treatment strongly enhanced the survival of proliferating cells in hippocampal dentate gyrus cells in rats (Methippara et al., 2010). This effect may be due to responsiveness of adult born hippocampal cells GABAergic stimulation (Nakamichi et al., 2009). Neurogenesis involves proliferation of neural progenitors, survival of newly proliferating cells, and differentiation to neural phenotypes, all of which may be involved in the pathophysiology and treatment of anxiety (Revest et al., 2009).

Eszopiclone is currently indicated for insomnia in adults and 6 randomized controlled trials supporting this indication (Sunovion Pharmaceuticals Inc., 2014). In one of these trials add-on eszopiclone to escitalopram resulted in greater improvement in HAM-A scores at each week, even when the insomnia item was removed, and also enhanced the possibility of response and remission in a 10 week long insomnia trial. Despite discontinuation of eszopiclone, treatment differences in anxiety measures were maintained (Pollack et al., 2008).

In a more recent trial comparing cognitive behavioral therapy (CBT) and eszopiclone combined therapy versus eszopiclone alone in adults with sleep disorder (N=29), greater improvement in self-reported anxiety (self-rating anxiety scale, SAS) was observed in the combined group (Zhang et al., 2018). Another study of perimenopausal women with impaired daytime functioning due to insomnia as well as some depressive and/or anxiety symptoms demonstrated improvement in insomnia as well as depression and anxiety symptoms in eszopiclone group compared to placebo (Joffe et al., 2010).

There are no pediatric trials of eszopiclone in GAD. One study of youth with attention deficit hyperactivity disorder (ADHD) and insomnia (age 6–17 years, N=486) failed to detect differences in efficacy for insomnia and safety between eszopiclone and placebo (Sangal et al., 2014).

In general, eszopiclone is well tolerated (Joffe et al., 2010; Pollack et al., 2008; Zhang et al., 2018). Most frequently reported adverse event was metallic taste (Joffe et al., 2010; Pollack et al., 2008); other adverse events were reported as somnolence, headache, dry mouth, dizziness, jitteriness and heart palpitations (Greenblatt & Zammit, 2012). There may be an impairment of driving in the initial waking hours(Greenblatt & Zammit, 2012). Among youth with ADHD, the most frequent adverse events were headache, dysgeusia and dizziness (Sangal et al., 2014).

In terms of safety concerns for pediatric GAD studies, eszopiclone is a schedule IV controlled substance. However eszopiclone may still be a promising candidate for study and development in pediatric GAD, particularly given concerns about tolerability limitations associated with benzodiazepines in pediatric GAD and other anxiety disorders (Dobson et al., 2019).

Riluzole

The exact mechanism of action of riluozole is unknown. Theories postulate that riluzole inhibits glutamate release considering its role in amyotrophic lateral sclerosis (ALS). But there might be other processes involved. Riluzole inhibits glutamic acid release in cultured neurons, from brain slices, and from corticostriatal neurons in vivo (Doble, 1996) and decreases release of glutamate from synaptic terminals throughout the CNS (Prakriya & Mennerick, 2000). Riluzole also blocks some of the postsynaptic effects of glutamic acid by noncompetitive blockade of N-methyl-D-aspartate (NMDA) receptors(Doble, 1996), increases glutamate reuptake (Frizzo et al., 2004), include inhibition of voltage-dependent sodium channels in CNS neurons (Urbani & Belluzzi, 2000), enhances of hippocampal alpha-amino-3-hydrosy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor subunit (GluR1 and GluR2) expression (Du et al., 2007). Riluzole stimulates growth factor synthesis, including BDNF (Katoh-Semba et al., 2002; Mizuta et al., 2001); promotes neurogenesis, neurite branching, and neurite outgrowth (Shortland et al., 2006).

Riluzole is well-absorbed and its bioavailability is approximately 60%. Riluzole is highly protein-bound (90%) and has a t_½ 12 hours. Metabolism is mostly hepatic and consists of hydroxylation mainly and glucuronidation. CYP1A2 is the principal isoenzyme responsible for N-hydroxylation. Primary excretion pathways are urine (90%) and feces (5%)(ITF Pharma Inc., 2018; Sanofi-Aventis U.S. LLC., 2008).

In preclinical models of anxiety (elevated plus maze, light/dark and open field tests), the anxiolytic effects of riluzole are comparable to diazepam (Sugiyama et al., 2012). Riluzole infusion into the basolateral amygdala in rats improved recognition memory and resulted in anxiolytic effect in the elevated plus-maze test and decreased freezing time in contextual fear conditioning test with foot shocks (Sugiyama et al., 2017; Sugiyama et al., 2018). Riluzole attenuated anxiety-like behaviors in mice that were pre-treated with veratrine, a sodium channel activator and anxiety inducer, in the open field test (Ohashi et al., 2015).

Other preclinical studies are available in various models including traumatic brain injury, spinal cord injury, Huntington’s disease, ischemia and ALS showing neuroprotective effects of riluzole(Wahl & Stutzmann, 1999). In another trial, riluzole stimulated human platelets to release BDNF, which is a contributor molecule in brain development, neuronal plasticity and neuronal survival (Turck & Frizzo, 2015).

One open-label study suggested that riluzole (100 mg/day) in adults with GAD may reduce anxiety and is associated with increased hippocampal N-acetylaspartate increase (as a marker of neuronal integrity) (Mathew et al., 2005). Eighteen GAD patients (seven of the participants with comorbid panic disorder) received riluzole monotherapy and 8 patients (44%) remitted based on HAM-A score. Most of comorbid panic disorder patients (5 out of 7) completed the trial and all had significant improvement in PDSS (Abdallah et al., 2013; Mathew et al., 2008).

Riluzole’s side effect profile was favorable and preliminary results regarding riluzole for the treatment of severe mood, anxiety and impulsive disorders were encouraging (Zarate & Manji, 2008). In trials of adults with GAD, the most common adverse events during the trial were insomnia/sleep disturbance, nausea, abdominal discomfort, sedation, and dry mouth. No serious adverse events were noted. One patient who completed the trial required a dose reduction of riluzole due to sedation. Transient increase in alanine aminotransferase ranging from 1.1 to 1.8 times the upper limit of normal was observed; for some patients, it normalized within 2 weeks while receiving riluzole and remained normal at endpoint; and for the remaining patients, aminotransferase values normalized after discontinuation of riluzole. No patient exhibited symptoms of hepatic toxicity (Mathew et al., 2005; Mathew et al., 2008).

The safety of riluzole in pediatric patients has been examined in two OCD trials. Six treatment resistant OCD patients received riluzole for 12 weeks in an open label trial. All subjects reached a target dose of 100 mg daily without limitation of adverse events aside from one patient who experienced drowsiness. Increases in lactate dehydrogenase and liver transaminases were observed with riluzole and both resolved on their own despite continued riluzole treatment (Grant et al., 2007). An RCT with 60 pediatric OCD patients (mean age 14 years) did not demonstrate a significant difference between riluzole and placebo with respect to adverse events(Grant et al., 2014). One serious adverse event (pancreatitis) was recorded in the riluzole group which resolved. Five other patients with asymptomatic elevations of transaminases were recorded. There were no differences in primary and secondary outcome measures among the riluzole and placebo groups. This patient sample was treatment refractory and taking a variety of other medications concurrently (Grant et al., 2014).

In considering safety, riluzole may be promising agent to study in pediatric GAD. The safety data from the prior pediatric OCD clinical trial are promising. Although efficacy was not demonstrated for OCD, this sample was highly treatment refractory and taking variable concomitant medications (Grant et al., 2014). Further study as monotherapy in pediatric GAD may be worth considering.

Eglumegad (LY354740)

Eglumegad (LY354740) is a selective agonist of Group II mGluRs (Schoepp et al., 2003). The mGlu2/3 belongs to Group II mGluRs that are located mainly presynaptically. Their major role is to inhibit neurotransmitter release (Nicoletti et al., 2011). They have an established role in synaptic plasticity regulation and in drug addiction (Lindsley et al., 2016).

The clinical use of Eglumegad (LY354740) is limited by its poor gastrointestinal absorption and poor bioavailability (3–5%) (Rorick-Kehn et al., 2006). To improve Eglumegad (LY354740) absorption and further assess the therapeutic effects of mGlu2/3 agonists, a pharmacologically inactive peptidyl prodrug (L-alanine) form of the active compound Eglumegad (LY354740), named LY544344, was developed. LY544344 is an Eglumegad (LY354740) prodrug that increases Eglumegad (LY354740) bioavailability. In rodents, oral administration of LY544344 improves its bioavailability by 10-fold compared to Eglumegad (LY354740) and produces comparable behavioral effects observed with parenteral administration of Eglumegad (LY354740) (Rorick-Kehn et al., 2006). Pharmacokinetic data are currently available for Eglumegad (LY354740) in animals (Johnson et al., 2002). In rats and dogs, oral Eglumegad (LY354740) has a linear pharmacokinetic profile from 30 to 1000 mg/kg. The bioavailability is 10% and 45 % respectively. The primary elimination route was urine in both animals. Metabolism of Eglumegad (LY354740) was evaluated in vitro using rat and dog liver microsomes and rat liver slices. Neither rats nor dogs metabolized Eglumegad (LY354740). In summary, Eglumegad (LY354740) is poorly absorbed in rats, moderately absorbed in dogs and rapidly excreted as unchanged drug in the urine (Johnson et al., 2002; Pilc, 2003).

Rat models of anxiety have supported a mechanistic role for an imbalance between glutamatergic and GABAergic tone in panic attacks(Johnson & Shekhar, 2012). Group II metabotropic glutamate (mGlu) receptors are located in areas that are thought to be implicated in panic networks such as dorsomedial hypothalamic regions and dorsal periaqueductal gray areas (Johnson et al., 2013). Eglumegad (LY354740) has anxiolytic activity in many rodent stress and anxiety models, including fear-potentiated startle, elevated plus maze, lactate-induced panic, and stress-induced hyperthermia (Helton et al., 1998; Kłodzińska et al., 1999; Shekhar & Keim, 2000; Spooren et al., 2002). Eglumegad (LY354740) had similar efficacy to benzodiazepines in animal models, but had no evidence for central nervous system (CNS) depression of motor function, learning, and memory (Helton et al., 1998; Tizzano et al., 2002). Pretreatment with Eglumegad (LY354740) inhibits panic-like response without producing sedation in rats (Shekhar & Keim, 2000). Neuroprotective characteristics of mGlu2/3 receptor agonists were also shown in animal models of cerebral ischemia(Bond et al., 2000; Bond et al., 1999) and mGlu2/3 knock out mice(Corti et al., 2007).

A double-blind trial in patients with panic disorder revealed no statistically significant differences between Eglumegad (LY354740) and placebo in reducing Panic Disorder Severity Scale (PDSS) scores and the clinical global impression (CGI) (Bergink & Westenberg, 2005). However, in infusion studies with cholecystokinin tetrapeptide (CCK-4)—a reliable panicogen—suggested benefit for the pro-drug LY544344 (which has higher bioavailability than Eglumegad [LY354740])(Kellner et al., 2005).

One double-blind trial for adults with GAD (N=738)(Michelson et al., 2005) that randomized patients to Eglumegad (LY354740) (100 or 200 mg BID), placebo, or lorazepam (4–5 mg daily) demonstrated that Eglumegad (LY354740) had greater efficacy compared to placebo and comparable reductions to lorazepam over 5 weeks. The tolerability profile of Eglumegad (LY354740) was superior to lorazepam (Michelson et al., 2005). Reported side effects can be found in Table 1. A subsequent study of LY544344 in adults with GAD (Dunayevich et al., 2008) randomized patients to LY544344 (8 or 16 mg BID) or placebo; however, secondary to a preclinical concern for treatment-emergent seizures (dog at 60 mg/kg, and in rat and mouse studies at doses 70 mg/kg) the study was terminated and active patients were discontinued after only 29 patients (26.1%) completed the double-blind treatment phase and entered the single-blind placebo, drug discontinuation assessment phase. Primary efficacy analysis showed a positive treatment effect for patients in the LY544344 treatment groups relative to placebo treatment (p=0.027). Patients who received LY544344 16 mg b.i.d., but not 8 mg b.i.d., had greater mean improvement from baseline on the HAM-A total score than those who received placebo (Dunayevich et al., 2008).

Table 1.

Summary of the mechanism of action, pharmacokinetics, and pharmacodynamics of the potential drugs for GAD

	Mechanism of action	Bioavailability	Half-life	Metabolism	Excretion
Agomelatine	MT1 and MT2 receptor agonist. 5-HT2C receptor antagonist.	Rapidly absorbed after oral administration (>75%), but absolute bioavailability after first-pass metabolism is (<5%).	2.3 hours.	Metabolized by hepatic CYP450, CYP1A2, and CYP2C9 to 3-hydroxy-agomelatine (inactive), and 7-desmethyl-agomelatine (inactive).	Eliminated by urinary excretion of its metabolites.
Eglumegad	Selective mGlu2/3 agonist.	Oral bioavailability in humans is only 3–5% due to limited gastrointestinal absorption. The oral bioavailability of the drug was ~ 10% in rats and 45% in dogs.	No human studies for half-life.	No human studies for metabolism.	Primarily eliminated via renal excretion.
Eszopiclone	Unknown, likely interactions with GABA-receptor complexes binding domains near or linked allosterically to benzodiazepine receptors.	Rapidly absorbed after oral administration.	~6 hours.	Extensively metabolized by CYP3A4 isoforms and CYP2E1 into its active metabolite, N-desmethylzopiclone, which is then distributed in brain and other tissues.	Eliminated via excretion in urine (up to 75% as metabolite), and saliva (<10% as parent drug).
Pimavanserin	5-HT2A antagonist/inverse agonist. Some affinity for 5-HT2C and sigma-1 receptors (σ1R).	Pimavanserin has dose-proportional pharmaco-kinetics after single oral doses from 17 to 255 mg (0.5–7.5 times the dosage recommendation). The bioavailability of pimavanserin oral tablets and oral solution is identical.	Pimavanserin: ~57 hours. N-desmethylated metabolite: ~200 hours.	Primarily metabolized via CYP3A4 and CYP3A5 to form the active N-desmethylated metabolite.	Feces (<2% as parent drug); urine (<1% as parent drug; <1% as metabolites).
Riluzole	Exact mechanism is unknown. Likely glutamate inhibitor and noncompetitive NMDA receptor blocker.	Approximately 60% by oral administration.	~12 hours.	Almost 88% of the drug is metabolized by either oxidation (via CYP1A2) or glucoronidation (via UGT-HP4). Some metabolites are pharmacologically active.	Feces (5%); urine (90%, 2% as unchanged).

Note: NMDA N-Methyl-D-aspartic acid, 5-HT 5-hydroxytryptamine (serotonin) receptor, MT melatonergic receptor, CYP cytochrome p enzyme, UGT-HP4 UDP-glucuronosyltransferase enzyme.

There are no trials of LY544344 in pediatric patients. Further study in children and adolescents could prove challenging as the adult clinical trial was halted for preclinical concerns related to seizure risk.

(3). Other Potential Drugs

Agomelatine

Agomelatine, a selective melatonergic MT1 and MT2 receptor agonist may have hypnotic effects (Fornaro et al., 2010). Agomelatine is also an antagonist at serotonergic 5-hydroxytryptamine 2_C (5-HT_2C) receptors that are implicated in underlying neurobiology of anxiety related behaviors(De Berardis et al., 2013). Additionally, the anxiolytic activity of agomelatine may be related to its modulation of GABAergic pathways and 5-HT_2C antagonism (De Berardis et al., 2013; Fornaro et al., 2010; Loiseau et al., 2006).

It is rapidly absorbed (>75%) after oral administration, although, extensive first-pass metabolism lowers its absolute bioavailability to <5% (Zupancic & Guilleminault, 2006). Thus, intranasal and transdermal formulations that have been developed may circumvent its extensive first-pass metabolism (Fatouh et al., 2017; Said et al., 2017). Agomelatine is mainly metabolized by hepatic CYP450 CYP1A2 and CYP2C9 to its main metabolites, 3-hydroxy-agomelatine and 7-desmethyl-agomelatine. Agomelatine is primarily eliminated by urinary excretion of metabolites and has a mean half-life (t_½) of approximately 2 h.

Preclinical studies demonstrated anxiolytic effects of agomelatine under basal conditions and in fear conditioning models (Guardiola‐Lemaitre et al., 2014). Agomelatine has demonstrated anxiolytic effects in a wide variety of preclinical models of anxiety including prenatal restraint stress (Morley-Fletcher et al., 2011), hypercortisolemic mice (Rainer et al., 2012), social defeat (Tuma et al., 2005), and social isolation rearing [SIR] model (Regenass et al., 2018). Additionally, agomelatine increased hippocampal brain-derived neurotrophic factor (BDNF) levels in another rat model (Lu et al., 2018; Marmigère et al., 2003). Chronic (3 weeks) agomelatine administration increased cell proliferation and neurogenesis in the ventral dentate gyrus, a region implicated in fear-related behaviors. Chronic agomelatine treatment facilitated neuronal survival throughout the dentate gyrus, although this effect appears to be restricted to developing neurons (Banasr et al., 2006). Another trial with light stress-exposed rat brains revealed that agomelatine stimulated neurogenesis and prevented apoptosis (Yucel et al., 2016). Agomelatine appeared to protect the rat brain from cerebral ischemia and reperfusion injury by suppressing apoptosis (Chumboatong et al., 2017). Also, chronic agomelatine treatment prevented transcription of the interleukin–1β and metabotropic glutamatergic receptor (mGluR) genes (Rossetti et al., 2018).

In three randomized, placebo-controlled, clinical trials, agomelatine was superior to placebo in adults with GAD (Stein et al., 2017; Stein et al., 2014; Stein et al., 2008) and, in another RCT, was more tolerable, but equally effective, compared to escitalopram in adults with GAD (Stein et al., 2018). High and low-dose (10 and 25 mg/day) agomelatine groups outperformed the placebo group, although only the high dose group showed a robust difference in remission rate compared to placebo (40.2 versus 12.5%) (Stein et al., 2017). When escitalopram and agomelatine compared over the course of 12 weeks, both significantly decreased Hamilton Anxiety Rating Scale (HAM-A) scores, however non-inferiority of agomelatine versus escitalopram was not established (Stein et al., 2018). Agomelatine (25–50 mg/kg) has been reported to be clinically effective and well tolerated in adults with GAD (Buoli et al., 2017; Stein et al., 2018) but is not approved by the FDA. At present there are no studies of agomelatine in pediatric patients.

In regards to safety, agomelatine, escitalopram, and placebo have similar tolerability in adults with GAD (Stein et al., 2017; Stein et al., 2014; Stein et al., 2008; Stein et al., 2018). The most frequently reported adverse events included dizziness, nausea, headache, nasopharyngitis, diarrhea and somnolence. Adverse events leading to discontinuation were less frequent in the agomelatine-treated patients compared to those who received placebo and escitalopram. The most recent agomelatine study (Stein et al., 2018) reported that agomelatine had superior tolerability compared to escitalopram with respect to headache, nausea, insomnia, dizziness, anxiety, hyperhidrosis, and diarrhea. There were no differences between agomelatine and escitalopram in terms of other adverse events. No suicidal behavior was reported during the study.

Rare cases of severe and life-threatening liver injury have been reported (Friedrich et al., 2016; Štuhec, 2013). Agomelatine is contraindicated in patients with impaired liver function (Freiesleben & Furczyk, 2015) and pre-treatment transaminases should be evaluated with subsequent hepatic profiles being obtained after 3, 6, 12, and 24 weeks(European Medicines Agency, 2016). If an elevation of transaminases is detected, these exams should be repeated within 48 h and agomelatine should be discontinued if it increases more than three times the higher limit of normal value (European Medicines Agency, 2016).

There are safety concerns to consider for future study in children and adolescents. Liver toxicity is a challenge to consider in further development of agomelatine for pediatric GAD. Agomelatine-induced hepatotoxicity appears to be rare and unpredictable (Gahr et al., 2013). Analysis of 9,234 patients who were treated with agomelatine showed that serum transaminases increased three fold higher than the upper normal limit in 1.3 and 2.5 % of patients treated with 25 and 50 mg of agomelatine, respectively, compared with 0.5 % for placebo. The incidence of abnormal transaminases was low and dose dependent. No patient-specific factors were identified regarding potential risk factors and withdrawal of agomelatine led to rapid recovery (Perlemuter et al., 2016).

Pimavanserin

Pimavanserin, is a potent selective serotonin 5-HT_2A inverse agonist with a combination of inverse agonist and antagonist activity to a lesser extent for 5-HT_2C and sigma-1 receptors (σ1R) (ACADIA Pharmaceuticals Inc, 2016). Pimavanserin has no affinity for the dopaminergic, adrenergic, histaminergic, or muscarinic receptors which separates it from other antipsychotic drugs (Vanover et al., 2006).

The mean t_½ for pimavanserin and its active metabolite (N-desmethylated metabolite) are approximately 60 hours and 200 hours, respectively. Pimavanserin demonstrates dose-proportional pharmacokinetics and is predominantly metabolized by CYP3A4 and CYP3A5 and to a lesser extent by CYP2J2, CYP2D6, and various other CYP and FMO enzymes. CYP3A4 is responsible for the formation of its major active metabolite and in vitro data suggest that pimavanserin does not irreversibly inhibit major hepatic and intestinal human CYP enzymes. Elimination is both via urine and feces (ACADIA Pharmaceuticals Inc, 2016).

In rodents, activation of 5-HT_2C receptors in the basolateral amygdala complex induces anxiogenic behaviors (Campbell & Merchant, 2003; Hackler et al., 2007)and 5-HT_2C receptor antagonists have shown to have anxiolytic properties (Aloyo et al., 2009). The 5-HT_2C and 5-HT_2A receptors have been implicated in the etiology and treatment of various psychiatric disorders. Activation of 5-HT_2C receptors, with agonists results in feelings of anxiety and panic in humans (Gatch, 2003). Pimavanserin’s activity on 5-HT_2A and 5-HT_2C receptors may suggest its potential on anxiety and fear conditions. Serotonin 2A receptors are concentrated in the limbic system and are important for mediating fear. There is evidence showing that the processing of emotionally loaded information is modulated by the 5HT_2A receptor and that the expression of these receptors related to anxiety traits (Frokjaer et al., 2008). Individuals with higher serotonin activity tend to have more fearful personalities, and interestingly, animals with a deficiency of 5-HT_2A receptors lack normal fear reactions (Moresco et al., 2002).

Pimavanserin has not yet been studied in clinical trials for GAD. In adults, pimavanserin and other 5-HT_2A receptor inverse agonists attenuated insomnia in few trials through increases in slow wave sleep (Ancoli-Israel et al., 2011; Monti, 2010; Rosenberg et al., 2008; Teegarden et al., 2008). There are no pediatric trials of pimavanserin in pediatric patients with GAD, or any other disorder.

Overall, clinical trials for pimavanserin in Parkinson’s demonstrated an acceptable safety profile (Bozymski et al., 2017). Prescription information reported most frequent side effects in Parkinson patients in the pimavanserin arms in Parkinson’s disease trials and in healthy volunteers as drowsiness, nausea, peripheral edema, confusion, hallucinations, constipation, and gait disturbance (Vanover et al., 2007). Pimavanserin may prolong QT intervals and concomitant administration of medications that prolong the QT interval should be avoided (ACADIA Pharmaceuticals Inc, 2016).

The primary safety concern for development of pimavanserin is the knowledge gaps related to its safety profile in children and adolescents. Initial and ongoing studies would need to examine this in a systematic fashion (ACADIA Pharmaceuticals Inc, 2016). A secondary challenge is the cost of pimavanserin—approximately $3,000/month.

Conclusion

Many children and adolescents with GAD do not respond completely to standard interventions such as CBT, SSRIs, and SNRIs (Bushnell et al., 2019). There is an urgent need to expand pharmacologic treatment options for pediatric GAD. Moreover, GABAergic and glutamatergic dysfunction are implicated in the pathophysiology of GAD in children, adolescents and adults, yet in youth, relatively few treatment studies have targeted these systems (Figure 1). These neurotransmitter systems undergo substantial shifts in childhood and adolescence (Benarroch, 2012; Chugani et al., 2001). This trajectory is poorly understood in GAD but could present strategic opportunities for pharmacological intervention, thereby sparing the need for lifelong treatment. In this regard, agomelatine, eszopiclone, and riluzole are promising agents to consider in for clinical development in pediatric GAD. Given that an adult clinical trial of Eglumegad (LY354740) was discontinued in adults due to safety concerns it is unlikely that this drug will have promise for clinical development in children (Dunayevich et al., 2008).

Figure 1.

GABAergic and Glutamatergic Modulators. Agomelatine modulates (GABAergic receptors and pathways broadly. Eglumegad (LY354740) is a selective agonist of presynaptic Group II MGluRs, Eszopiclone is GABA A receptor agonist. Riluzole inhibits glutamate release and blocks the postsynaptic effects by noncompetitive antagonism of N-methyl-D-aspartate (NMDA) receptors. Further, riluzole increases glutamate uptake and enhances hippocampal alpha-amion-3–3hydrosy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor subunit (GluR1 and GluR2) expression.

Previous anxiolytic drug discovery has largely focused on single targets (highly selective agents acting on specific molecular targets). However, the notion of better and safety drugs for GAD has not been realized. Multimodal drug candidates in development, like agomelatine, with the intention of re-equilibrating perturbed circuits, may potentially lead to more effective treatments (Figure 2).

Figure 2.

Pharmacodynamic Targets of Novel Agents for Anxiety. Pharmacodynamic targets of novel agents for anxiety. Agomelatine is a selective melatonin 1 and 2 receptor agonist. Pimavanserin is a selective serotonin 5-HT2A inverse agonist. Pimavanserin also has inverse agonist and antagonist activity for 5-HT2C and sigma-1 receptors. Riluzole inhibits voltage-dependent sodium channels.

Each of these candidates has unique strengths and barriers to consider for future study. Broadly, eszopiclone and riluzole may have the most promise in the short-term for development in pediatric GAD as there are existing safety data in children. Preclinical work with eszopiclone and riluzole also are encouraging. Eszopiclone and riluzole also have a favorable cost profile compared to other agents. Agomelatine may be another promising agent to consider for pediatric GAD with respect to prior preclinical work and adult clinical trials. Unfortunately at present there are substantial barriers to developing LY354740 and pimavanserin for pediatric GAD related to safety, preclinical seizure evidence and cost. Further preclinical studies of agomelatine, eszopiclone, pimavanserin and riluzole in the context of neurodevelopment are critical to advance mechanistic understanding of anxiety in children and adolescents and to facilitate drug discovery.

Abbreviations

ACh

acetylcholine

ADHD

attention deficit hyperactivity disorder

AMPA

alpha-amino-3-hydrosy-5-methyl-4-isoxazolepropionic acid

ALS

amyotrophic lateral sclerosis

BDNF

brain-derived neurotrophic factor

CBT

cognitive behavioral therapy

CCK-4

cholecystokinin tetrapeptide

CNS

central nervous system

CYP

cytochrome P450

EPS

extrapyramidal symptoms

h

hours

HAM-A

Hamilton anxiety rating scale

HDRS-17

Hamilton depression rating scale

5-HT

5-hydroxytryptamine (serotonin) receptor

5-HTT

serotonin transporter gene

5-HT2C

5-hydroxytryptamine c receptor

GABA

gamma-aminobutyric acid

GAD

generalized anxiety disorder

mGluR

metabotropic glutamatergic receptor

M1 and M2

melatonin receptors

NMDA

N-Methyl-D-aspartic acid

OCD

obsessive compulsive disorder

σ1R

sigma-1 receptors

SIR

social isolation rearing

SNRIs

serotonin norepinephrine reuptake inhibitors

SSRIs

selective serotonin reuptake inhibitors

t_½,

half-life