Pathophysiology of TD

Data from the early 1990s supported the concept that TD was a manifestation of D₂ receptor upregulation and supersensitivity related to chronic reduction in dopaminergic neurotransmission, primarily from postsynaptic receptor blockade.Reference Teo, Edwards and Bhatia ⁹ Dopamine D₂ receptors are expressed on striatal medium spiny neurons and function in an inhibitory manner to reduce the velocity and amplitude of movements through activation of the so-called the indirect basal ganglia pathway.Reference Aquino and Lang ¹¹ The development of increased D₂ receptor sensitivity would thus be expected to induce hyperkinesia. This model was supported by in vivo animal data demonstrating the development of D₂ receptor upregulation and supersensitivity after exposure to D₂ antagonists, and by the clinical observation that withdrawal of D₂ antagonists in humans resulted in TD exacerbationReference Teo, Edwards and Bhatia ⁹ ; however, animal data also demonstrate that these phenomena occur very quickly after drug exposure, contrary to the clinical course of TD, and are rapidly reversible after withdrawal of the dopamine antagonist, implying that the persistent forms of TD in humans may have differing mechanisms than that associated with withdrawal dyskinesia.Reference Casey ^{12 ,} Reference Mahmoudi, Levesque and Blanchet ¹³ Moreover, while increased striatal D₂ receptor binding can be seen in patients exposed to chronic D₂ blockade, this effect is not necessarily correlated with the presence of dyskinesia in imaging or postmortem studies.Reference Teo, Edwards and Bhatia ⁹ Although the D₂ upregulation/supersensitivity hypothesis for TD appears lacking in humans, primates exposed to clozapine or haloperidol experienced significant D₃ receptor upregulation in the haloperidol cohort, with the extent of D₃ binding in the nigrostriatal regions correlating with TD intensity.Reference Mahmoudi, Levesque and Blanchet ¹³ There is also support for the D₃ hypothesis from genetic studies associating certain D₃ receptor polymorphisms with increased TD risk.Reference Segman, Goltser and Heresco-Levy ¹⁴ Given the large overlap in sequence homology and ligand affinity between D₂ and D₃ receptors, selective D₃ agents have only recently been developed for in vivo human neuroimaging, so future imaging studies may shed light on the viability of this concept.Reference Le Foll, Wilson, Graff, Boileau and Di Ciano ¹⁵

Genetic markers have also implicated numerous pathways involved in striatal dopaminergic signaling, including serotonin and dopamine receptor variations,Reference Segman, Goltser and Heresco-Levy ¹⁴ and polymorphisms in the gamma-aminobutyric acid (GABA) transporterReference Son, Lee and Yoon ¹⁶ and GABA_A receptor.Reference Inada, Koga and Ishiguro ¹⁷ As will be discussed below, modulation of vesicular monoamine transporter type 2 (VMAT2) is a promising treatment modality, and several polymorphisms in the VMAT2 gene have been associated with increased TD risk.Reference Zai, Tiwari and Mazzoco ¹⁸

Other hypotheses have been advanced over the years related to in vivo animal data, and clinical and genetic human studies. The involvement of free radicals and other oxidative mechanisms was suggested 2 decades ago on the basis of animal and a small number of human studies.Reference Lohr and Browning ¹⁹ These hypotheses fell into disfavor due to the inconsistent results from vitamin E trials,Reference Lohr, Kuczenski and Niculescu ²⁰ but they have not been completely abandoned, as some genetic studies point to increased risk among those with polymorphisms in the free radical scavenger enzyme superoxide dismutase and related anti-oxidative enzymesReference Teo, Edwards and Bhatia ^{9 ,} Reference Cho and Lee ²¹ and to markers related to systemic inflammation.Reference An, Tan and Shi ²² Recent animal studies have indicated that peroxisome proliferator-activated receptor (PPAR) agonists exhibit neuroprotective properties, leading to exploratory studies of the PPAR-gamma agonist pioglitazone and PPAR-alpha agonist fenofibrate in rat models of TD, with positive results.Reference Grover, Kumar, Singh, Vikram and Budhiraja ²³ The observations that phenylketonuria was associated with TD risk and that administration of large phenylalanine doses worsened dyskinetic symptoms in patients with TD led to the hypothesis that failure to clear central nervous system phenylalanine might underlie TD pathophysiology. Branched chain amino acids compete with phenylalanine for transport across the blood–brain barrier, and their use was associated with TD improvement in a number of studies; however, most of these were open label and were performed by investigators at 1 institution.Reference Richardson, Small, Read, Chao and Clelland ²⁴ No new studies have appeared in the past decade.

Among the more novel findings from genetic association studies are loci associated with development, cellular signaling, and neuroplasticity.Reference Teo, Edwards and Bhatia ^{9 ,} Reference Aquino and Lang ¹¹ That TD might be best viewed as a disorder of synaptic plasticity has emerged as a leading unifying hypothesis that brings together basic science and genetic findings with the clinical observation that TD shows limited reversibility after withdrawal of offending agents.Reference Aquino and Lang ¹¹ While striatal D₂ receptor hypersensitivity might be the initial manifestation of D₂ antagonist exposure, ongoing D₂ blockade creates secondary effects on the plasticity of glutamatergic synapses of striatal interneurons. Aberrant glutamatergic signals to cortical structures that also have impaired plasticity results in a situation wherein withdrawal of dopamine antagonists fails to generate the expected symptomatic reversal.Reference Teo, Edwards and Bhatia ⁹ Not only does this model suggest that certain glutamate-based strategies might be effective for TD,Reference Teo, Edwards and Bhatia ⁹ but it also explains why signal interruption via surgical pallidotomy or bilateral deep brain stimulation (DBS) of the internal part of the globus pallidus (GPi) have been reported as beneficial in those with intractable TD.Reference Pouclet-Courtemanche, Rouaud and Thobois ²⁵

Imaging studies and peripheral markers in schizophrenia patients with TD reflect this underlying cellular dysfunction. Reduced basal ganglia and thalamic volume is seen among those with TD, with the greatest reductions found in the caudate nucleus.Reference Sarró, Pomarol-Clotet and Canales-Rodríguez ²⁶ S100B is a calcium binding protein expressed by astrocytes and involved in numerous cell regulatory processes. Peripheral S100B levels are increased after central nervous system cellular insults, and data in schizophrenia patients not only reveal higher S100B levels among those with TD, but also that serum S100B levels positively correlate with abnormal movement rating scale scores.Reference Zhang, Xiu and Chen da ²⁷

Treatment

After an extensive review of the literature, the American Academy of Neurology (AAN) found few evidence-based therapies for TD in 2013 and concluded that the following agents are either not recommended, or have insufficient data to support (or refute) their use: acetazolamide, bromocriptine, baclofen, buspirone, diltiazem, galantamine, eicosapentaenoic acid, levetiracetam, vitamin E, vitamin B6, thiamine, selegiline, melatonin, nifedipine, yi-gan san, biperiden discontinuation, botulinum toxin type A, electroconvulsive therapy, α-methyldopa, reserpine, and pallidal DBS.Reference Bhidayasiri, Fahn and Weiner ²⁸ The AAN review also noted that data are insufficient to support or refute TD improvement by withdrawing causative agents or switching from typical to atypical antipsychotics.Reference Bhidayasiri, Fahn and Weiner ²⁸ Vitamin E in particular had proponents based on early small case series, but these findings failed to replicate in larger controlled studies.Reference Vijayakumar and Jankovic ^{2 ,} Reference Soares and McGrath ²⁹ Evidence for branched chain amino acid preparations also remains inconclusive due to the paucity of controlled studies.Reference Vijayakumar and Jankovic ²

Among the small number of evidence-based options are clonazepam, ginkgo biloba, amantadine, and tetrabenazine. In a double-blind, crossover, 12-week, placebo-controlled study, clonazepam (mean dose 3.5 mg/d) was associated with 35% improvement in dyskinesia symptoms (n=19), although tolerance developed after 5–8 months of use in the 5 subjects whose treatment continued up to 9 months.Reference Thaker, Nguyen, Strauss, Jacobson, Kaup and Tamminga ³⁰ The investigators noted that a 2-week washout was sufficient to restore the antidyskinetic effect of clonazepam. Of the 4 published studies of amantadine use, 2 employed double-blind, placebo-controlled, crossover designs. The first, an 18-week trial, reported 15% improvement in dyskinesia ratings on an amantadine dose of 300 mg/d,Reference Angus, Sugars, Boltezar, Koskewich and Schneider ³¹ while the second study randomized patients to amantadine 100 mg/d or placebo for 2 weeks each (with a 4-day washout between treatment arms) and found a 22% reduction in abnormal involuntary movement scale (AIMS) scores.Reference Pappa, Tsouli, Apostolou, Mavreas and Konitsiotis ³² Extract of ginkgo biloba (EGb) is an antioxidant that possesses free radical scavenging properties. A standardized extract (EGb-761) was examined in a 12-week, double-blind, placebo-controlled trial in which inpatients with schizophrenia and TD were randomly assigned to EGb-761 240 mg/d (n=78) or placebo (n=79). EGb-761 was well tolerated, with 96.8% of subjects completing the study. There was a significantly greater decrease in endpoint AIMS total score in the patients treated with EGb-761 compared to the placebo cohort (p<0.0001), with ≥30% reduction in AIMS noted in 51.3% of EGb-761 but only 5.1% of the placebo group.Reference Zhang, Tan, Zhang, Chan, Wu and Zhou ³³

By the 1960s, it was known that TD was the result of increased dopamine signaling, and this led to the search for agents that could modulate dopamine neurotransmission without directly antagonizing postsynaptic receptors. Reserpine’s effect on presynaptic vesicle monoamine content would not be elucidated until the 1960s; however, by the mid-1950s it was known that reserpine had antipsychotic properties and was useful for movement disorders such as Huntington’s disease, but with significant tolerability issues related to orthostasis.Reference Chandler ^{34 ,} Reference Lazarte, Petersen and Baars ³⁵ Tetrabenazine (TBZ) was developed in the 1950s as an antipsychotic based on in vivo models that predicted reserpine-like effects, but with markedly reduced orthostasis risk.Reference Quinn, Shore and Brodie ³⁶ The first TD study with TBZ was published in 1972 with the authors rationalizing the choice of TBZ due to its lower risk for hypotension than reserpine.Reference Kazamatsuri, Chien and Cole ³⁷ Though TBZ has been available in Canada, Great Britain, and Europe for decades, it was not approved in the US until August 15, 2008, with an indication for the management of chorea in patients with Huntington’s disease.

While the mechanism that differentiated reserpine’s and TBZ’s clinical properties was not understood, by the mid-1980s it became clear that integral membrane transporters were necessary to package neurotransmitters into synaptic vesicles of presynaptic neurons (Figure 1).Reference Scherman and Weber ³⁸ This discovery led to the characterization of multiple vesicular transporters, including those with specificity for acetylcholine,Reference Roghani, Feldman and Kohan ³⁹ or for monoamines such as dopamine, serotonin, norepinephrine, epinephrine, and histamine.Reference Erickson, Eiden, Schafer and Weihe ⁴⁰ The vesicular monoamine transporter (VMAT) was found to exist in 2 isoforms (VMAT1 and VMAT2) that vary in distribution: VMAT1 is expressed mainly in the peripheral nervous system, while VMAT2 is expressed mainly in monoaminergic cells of the CNS.Reference Erickson, Schafer, Bonner, Eiden and Weihe ⁴¹ TBZ’s improved tolerability profile was related to the fact that it was a specific and reversible VMAT2 inhibitor, while reserpine was an irreversible and nonselective antagonist of both VMAT isoforms. TBZ and reserpine also have different binding sites on VMAT2 (Figure 2).

Figure 1 Location and function of VMAT. ADP: adenosine biphosphate, ATP: adenosine triphosphate, VMAT: vesicular monoamine transporter.

Figure 2 VMAT2 structure.Reference Jankovic and Clarence-Smith ⁴⁶ Adapted from: Jankovic J, Clarence-Smith K. Tetrabenazine for the treatment of chorea and other hyperkinetic movement disorders. Expert Review of Neurotherapeutics 2011; 11(11):1509-2, reprinted by permission of the publisher Taylor & Francis Ltd. (http://www.tandfonline.com).

Investigation of TBZ’s metabolism revealed that it is rapidly and extensively converted into 2 isomers, α-dihydrotetrabenazine (DH-TBZ) and β-DH-TBZ, which have high affinity for VMAT2 and are the pharmacologically active agents.Reference Kilbourn, Lee, Vander Borght, Jewett and Frey ^{42 ,} Reference Kilbourn, Lee, Heeg and Jewett ⁴³ The α-DH-TBZ isomer is metabolized via cytochrome P450 (CYP) 2D6 and 3A4 into inactive metabolites, while β-DH-TBZ is metabolized solely via 2D6.Reference Mehanna, Hunter, Davidson, Jimenez-Shahed and Jankovic ^{44 ,} Reference Muller ⁴⁵ Due to the short half-life of DH-TBZ and the existence of 2D6 polymorphisms, use of TBZ for Huntington’s disease carries recommendations for thrice daily (TID) dosing, and for CYP 2D6 genotyping to screen for poor metabolizer status when exceeding 50 mg/d.Reference Jankovic and Clarence-Smith ⁴⁶ To obviate these issues, 2 different pharmacological strategies were explored to moderate TBZ’s metabolism, to permit once-daily dosing, and also to improve tolerability.

Deutetrabenazine

The use of the stable isotope deuterium to replace selected hydrogen atoms in a molecule can result in a compound with similar pharmacodynamic properties but different kinetics, as the carbon-deuterium covalent bond requires 8 times more energy to break than a carbon-hydrogen bond. ⁴⁷ A deuterated form of TBZ deutetrabenazine (Deut-TBZ) was synthesized (Figure 3) with such purpose in mind. While the active metabolites of Deut-TBZ retain the VMAT2 affinity of the nondeuterated DTBZ forms, the substitution of deuterium for hydrogen at specific positions markedly slows the breakdown of metabolites, resulting in a pharmacokinetic profile with longer metabolite duration of action, greater active drug exposure (Figure 4), and less impact of 2D6 genotype on drug exposure, eliminating the need for genotyping. ^{47 ,} Reference Frank, Testa and Stamler ⁴⁸ Deut-TBZ was first studied in Huntington’s chorea in a 13-week, double-blind, placebo-controlled, parallel-group study in which 90 patients were randomized 1:1 to receive Deut-TBZ or placebo twice daily.Reference Frank, Testa and Stamler ⁴⁸ The study involved an 8-week titration period and 4-week maintenance period followed by a 1-week washout. The maximum daily Deut-TBZ dose was 48 mg, but was reduced to 36 mg in those receiving a strong CYP 2D6 inhibitor (bupropion, fluoxetine, or paroxetine). No dose modification was needed based on 2D6 genotype. There was a 36.4% reduction in total maximal chorea score for Deut-TBZ compared to 14.4% for placebo.Reference Frank, Testa and Stamler ⁴⁸ Importantly, adverse effects were comparable between both groups, with 1 drop-out in the Deut-TBZ arm vs 2 in the placebo arm. The only adverse event occurring in ≥5% of Deut-TBZ subjects and at a rate ≥2 times that of placebo was somnolence: 11.1% for Deut-TBZ vs. 4.4% for placebo.

Figure 3 Structures of tetrabenazine and deutetrabenazine.

Figure 4 Kinetics of tetrabenazine and deutetrabenazine. ⁴⁷

A subsequent TD study was performed in a similar design with 117 subjects randomized in 1:1 manner to Deut-TBZ or placebo.Reference Anderson, Factor and Hauser ⁴⁹ The population demographics were as follows: mean age 54.9±9.8 years, 59% female, 79% Caucasian, 80.5% of whom were receiving ongoing dopamine antagonists, mean TD duration of 75.0±81.9 months. The mean baseline AIMS score for items 1–7 was 9.6±3.9, with 85.8% of subjects having AIMS≥6. Study treatment retention was high, with 6 drop-outs in the Deut-TBZ arm vs 7 in the placebo arm. There was a mean 3.0 point decrease in AIMS for Deut-TBZ compared to 1.4 for placebo (p=0.019). Among those with baseline AIMS≥6, there was a 3.4 point decrease in AIMS for Deut-TBZ compared to 1.9 for placebo (p=0.027). There were no adverse effects that occurred in ≥5% of Deut-TBZ subjects and at a rate ≥2 times the rate in placebo.

Valbenazine

Each of TBZ’s active metabolites α-and -DH-TBZ possess multiple chiral centers, yielding a total of 8 possible isomers (Figure 5), each of which has different VMAT2 activity.Reference Muller ^{45 ,} Reference Yao, Wei and Wu ⁵⁰ Characterization of isomers with greatest VMAT2 affinity (Table 1) led to the development of valbenazine, a prodrug that is metabolized into the most active DH-TBZ isomers.Reference Muller ⁴⁵ Importantly, valbenazine was designed to be metabolized slowly, and thereby minimize high peak plasma concentrations, decrease peak-to-trough ratios, and reduce inter-subject variability. The T_Max for the active metabolites is 4–10 hours with a half-life of approximately 20 hours, allowing once-daily dosing.Reference O’Brien, Jimenez and Hauser ⁵¹ Due to the limited range of metabolites (2), it was also designed to limit off-target receptor binding that was theoretically possible with certain DH-TBZ isomers.Reference Muller ⁴⁵ Data from a randomized, 6-week, double-blind, placebo-controlled, dose-titration study in subjects with TD was published in 2015 (n=100).Reference O’Brien, Jimenez and Hauser ⁵¹ The population demographics were as follows: mean age 56.2±10.3 years, 57% male, 79% Caucasian, 73% of whom were receiving ongoing antipsychotic treatment, mean TD duration of 7 years. The mean baseline AIMS score for items 1–7 was 8.0±4.0. The valbenazine starting dose was 25 mg, and this could be escalated in 25 mg increments every 2 weeks to a maximum of 75 mg. At study endpoint, 75% of subjects were on the maximum dose, and there were no study drop-outs due to adverse events in the valbenazine cohort. There was a mean 3.6 point decrease in AIMS for valbenazine compared to 1.1 for placebo (p=0.0005), with lower baseline severity and mood diagnosis (vs schizophrenia spectrum) moderating factors that improved treatment response.Reference Josiassen, Remington and Burke ⁵² The following adverse events occurred in ≥5% of valbenazine subjects and at a rate ≥2 times that of placebo: fatigue: 9.8% for valbenazine vs 4.1% for placebo; headache: 9.8% for valbenazine vs 4.1% for placebo; decreased appetite: 7.8% for valbenazine vs 0% for placebo.

Figure 5 Dihydrotetrabenazine.

Table 1 Dihydrotetrabenazine (DH-TBZ) isomer affinity for VMAT2 (Ki±SEM nM)Reference Yao, Wei and Wu ⁵⁰

A subsequent 6-week study using similar design was presented in 2016 using dosing data derived from the prior trial.Reference Marder, Knesevich and Hauser ⁵³ In this design 234 subjects with TD were randomized in a 1:1:1 manner to placebo, valbenazine 40 mg once daily or valbenazine 80 mg once daily. Completion rates were high (87.6%), with only 2 drop-outs due to adverse events in each of the placebo and 40 mg arms, and 3 in the 80 mg group. There were no adverse events in either valbenazine arm that exceed 5% in frequency and were ≥2 times that of placebo. Subject demographics were similar to the prior trial but with higher baseline AIMS scores: mean 10.4±3.6.Reference Marder, Knesevich and Hauser ⁵³ The prior trial demonstrated a least squares mean 2.5 point difference in AIMS score between valbenazine and placebo, while for this trial the least squares mean change from baseline to week 6 was 3.1 points between valbenazine 80 mg (–3.2) and placebo (–0.1) (p<0.001). The effect size (Cohen’s d) was large, at 0.90.

Disclosures

Dr. Meyer reports having received speaking or advising fees from Acadia, Alkermes, Forum, Merck, Otsuka-USA, Sunovion, and Teva.