Movement Disorders Moment: Personalized Medicine in Parkinson Disease

Parkinson disease (PD) is a neurodegenerative condition caused by the loss of dopaminergic neurons in the substantia nigra. This neuronal loss results in decreased dopamine production and dopamine levels, precipitating unintended or uncontrollable movements, including tremors, muscle stiffness, slow movement, and impaired balance and coordination. PD affects ~1 million adults in the United States. The increasing prevalence and economic burden of PD, expected to affect 1.6 million adults and exceed $79 billion in annual healthcare costs by 2037, signals the need for increased research efforts and improved treatment approaches.¹ There are no curative therapies; the therapeutic strategy focuses on symptom management to improve quality of life.²

Medications that increase dopamine levels in the brain comprise the cornerstone of PD therapy. These medications increase dopamine through different mechanisms (Figure 1), including dopamine replacement (levodopa [L-DOPA]), stimulating dopamine production (dopamine agonists), or inhibiting dopamine degradation (catechol-O-methyltransferase [COMT] and monoamine oxidase inhibitors).³ Response to these medications, such as symptom improvement (efficacy) and adverse effects (safety), is a highly variable and complex phenotype, influenced by patient-specific factors (eg, age, stage in therapy). In addition, genetic factors may contribute to variability in response. Thus, pharmacogenetics, which integrates pharmacology and genomics to identify genetic predictors of medication response, may provide valuable information to help tailor PD therapy.⁴

The Pharmacogenomics Knowledgebase (PharmGKB) curates comprehensive pharmacogenetic information, including clinical gene–drug guidelines, pharmacogenetic information in drug labeling, prescribing information, and clinical gene–drug annotations.⁴ Clinical annotations summarize the collective evidence to define the relationship between a specific genetic variant and medication, with level 1 or 2 denoting high to moderate evidence supporting the association, level 3 denoting low evidence, and level 4 denoting no association. The currently curated clinical annotations related to antiparkinsonian medications are limited to low-evidence associations (level 3).⁵ A recent systematic review examined the evidence for tailoring PD treatment on the basis of the 5 most studied genes in PD research.⁶ Herein, we review antiparkinsonian medications with PharmGKB clinical annotations and discuss their relevance. 

Figure 1. Mechanism of action of select antiparkinsonian drugs. 

Abbreviations: 3-MT, 3-methoxytyramine; 3-O-MD, 3-O-methyldopa; AADC, aromatic L-amino acid decarboxylase; COMT, catechol-O-methyltransferase; DAT, dopamine transporter; DOPAC, 3,4-dihydroxyphenylacetic acid; L-DOPA, levodopa; MAO-B, monoamine oxidase B. Created in BioRender. Davis, B. (2025) https://BioRender.com/2pmv0bg

Antiparkinsonian Medications

Levodopa

L-DOPA, a dopamine precursor, is prescribed to ~86% of people with PD, and forms the cornerstone of therapy for motor symptoms.⁷ L-DOPA demonstrates subjective symptomatic improvement in the majority of users (94%) and clinical improvement in 43% of individuals.⁸ L-DOPA is primarily metabolized by aromatic L-amino acid decarboxylase to form dopamine, with secondary metabolism by COMT to form 3-O-methyldopa.⁹ Prolonged L-DOPA use causes side effects due to disease progression, altered brain chemistry, and drug-induced changes in dopamine signaling and other neurotransmitter systems.¹⁰ After a decade of L-DOPA therapy, most individuals who use L-DOPA encounter adverse effects, such as excessive daytime somnolence, dyskinesias, gastrointestinal (GI) issues, motor fluctuations, hallucinations, and impulse control disorders.^13,14 Among L-DOPA users, 25% to 45% experience GI symptoms, such as nausea, vomiting, constipation, dysphagia, and loss of appetite.^10,11

Dopamine Agonists

Dopamine agonists (eg, ropinirole, pramipexole) are prescribed to ~39.4% of people with PD, improving symptoms in ~43% of those treated^.7 Pramipexole is minimally metabolized and is largely excreted unchanged. Ropinirole is primarily metabolized by CYP1A2. The most common side effects of pramipexole and ropinirole are nausea (28% to 60%), somnolence (22% to 40%), and dizziness (25% to 40%).^12,13 More than one-third of individuals treated with dopamine agonists report adverse reactions resulting in therapy cessation.¹⁴

Monoamine Oxidase B Inhibitors

Monoamine oxidase B (MAO-B) inhibitors (eg, selegiline, rasagiline) are used by 30% of people with PD, producing clinically significant improvement in ~66% of treated individuals.^7,15 Selegiline and rasagiline undergo metabolism primarily by CYP2B6 and CYP1A2, respectively. Common adverse effects vary by medication, and include nausea (20%), dizziness (14%), and abdominal pain (8%) for selegiline, and headaches (14%), arthralgia (7%), and dyspepsia (7%) for rasagiline.^16,17

Catechol-O-Methyltransferase Inhibitors

COMT inhibitors (eg, entacapone, tolcapone) are used by 11.5% of people with PD.⁷ When used as an adjunct therapy to L-DOPA, they extend L-DOPA half-life and improve symptoms in ~76% of individuals.¹⁸ Entacapone undergoes primary hepatic metabolism by isomerization followed by glucuronidation and works only in the peripheral circulation.¹⁹ In contrast, tolcapone crosses the blood–brain barrier to act both peripherally and centrally at presynaptic neurons. Tolcapone is primarily metabolized through glucuronidation, with minor contributions from COMT, CYP3A4, CYP2A6, and N-acetylation.²⁰ Individuals treated with entacapone or tolcapone commonly experience GI adverse effects (10% to 14% with entacapone, 30% with tolcapone) and dyskinesia (25% with entacapone, 42% with tolcapone). In addition, tolcapone causes sleep disorders in ~24% of individuals.^19,20

Pharmacogenetics 

Among antiparkinsonian medications clinically annotated in PharmGKB, genes encoding enzymes involved in dopamine degradation (COMT), dopamine receptors (DRD2, DRD3), dopamine transport (SLC6A3, SLC22A1), and glutamate receptor regulation (HOMER1) have been found to influence response.⁵ Notably, the majority of associations are in pharmacodynamic, rather than pharmacokinetic (eg, metabolic), pathways. At the drug level, L-DOPA has the most clinical annotations related to both efficacy (SLC6A3) and toxicity (DRD2, DRD3, HOMER1), whereas entacapone (COMT), rasagiline (DRD2), and pramipexole (DRD3) annotations are related to efficacy.^21-26 Ropinirole and tolcapone have no annotated associations. 

Pharmacogenetic studies investigating efficacy have used the Unified Parkinson’s Disease Rating Scale (UPDRS) and the Stand-Walk-Sit (SWS) test to assess clinical response. UPDRS is a 4-part standardized tool used to evaluate PD severity and progression. Scores are commonly evaluated before and after treatment, with lower scores indicating clinical improvement.²⁷ SWS is a functional assessment that evaluates changes in completion time, gait stability, and postural transitions to monitor PD progression and treatment response. The Table summarizes antiparkinsonian pharmacogenetic studies included in PharmGKB clinical annotations. Variant population frequencies based on the Ensembl genome browser (GRCh38 release 114; European Bioinformatics Institute/Wellcome Trust Sanger Institute, Cambridge, UK) are presented in Figure 2. 

Figure 2. Genotype frequencies of genes related to antiparkinsonian medications were assessed across diverse populations, including those of African ancestry (AFR), African ancestry in the southwestern United States (ASW), American ancestry (AMR), Mexican ancestry in Los Angeles (MXL), East Asian ancestry (EAS), European ancestry (EUR), Utah residents of Northern and Western European ancestry (CEU), and South Asian ancestry (SAS), using data from the 1000 Genomes Project phase 3 allele frequencies.

Catechol-O-Methyltransferase

COMT encodes an enzyme involved in the degradation of catecholamines, including dopamine, epinephrine, and norepinephrine.²⁸ In addition, COMT degrades L-DOPA, reducing the amount available for conversion to dopamine. As PD progresses, COMT inhibitors may be added to help manage L-DOPA effects.²⁹ COMT activity varies substantially between individuals, with COMT rs4680G>A (Val158Met) contributing to this variability. The Val (G) allele results in higher basal COMT activity, whereas the Met (A) allele results in lower activity.²⁶

The influence of COMT rs4680 (Val158Met) on entacapone response to L-DOPA was evaluated in individuals with COMT Val/Val (high activity; n=17) or Met/Met (low activity; n=16) genotypes randomized to L-DOPA combined with either entacapone or placebo.²⁶ Among Val/Val carriers, entacapone cotreatment significantly increased “On” time (39±10 minutes) compared with placebo (9±9 minutes; P=.003), whereas no difference was observed between entacapone and placebo when limited to Met/Met carriers (P=.28). When limited to entacapone-treated individuals, Val/Val carriers had more pronounced COMT inhibition in red blood cells and also exhibited a greater increase in L-DOPA area under the plasma drug concentration–time curve (AUC) compared with Met/Met carriers (62±6% vs 34±8%; P=.01). Significant genotype-by-treatment interactions were identified for “On” time (P=.04), L-DOPA AUC (P=.04), and COMT inhibition (P=.02). Although this study did not include individuals with the Val/Met genotype, a recent pharmacokinetic study in 54 Japanese individuals found that entacapone increased L-DOPA AUC to a greater extent in Val/Val carriers (1.59±.26-fold) compared with Val/Met (1.41±.36-fold) and Met/Met (1.28±.21-fold) carriers (P<.05).³⁰

Frequencies of the COMT high-activity Val/Val genotype are highest among African populations (~50% to 55%); Figure 2; COMT), and therefore, these individuals may be more likely to benefit from add-on entacapone therapy.³¹ However, individuals of African ancestry have been underrepresented in COMT inhibitor trials and pharmacogenomic studies, limiting the generalizability of these findings. Larger and more diverse studies are needed to establish the clinical utility of COMT genotype-guided therapy.

Dopamine Receptors (DRD2 and DRD3)

There are 5 types of dopamine receptors (D1 through D5), with varying function and expression patterns. D2 and D3, encoded by DRD2 and DRD3, respectively, are involved in locomotion, memory, and learning (D2), cognition and impulse control (D3), and attention and sleep (D2/D3).³² Variants in DRD2 and DRD3 have been linked to altered safety and efficacy of rasagiline, pramipexole, and L-DOPA.

Among 692 individuals with early PD in the pharmacogenetic substudy of the ADAGIO double-blind, placebo-controlled, multicenter trial, individuals with the DRD2 rs2283265 C/C genotype exhibited a 2-point reduction in peak UPDRS scores (mean change −2.02±.40) after 12 weeks of rasagiline therapy.²¹ Conversely, individuals with A/C or A/A genotypes experienced clinical worsening (mean change in UPDRS .74±.78). Similar results were observed for DRD2 rs1076560 C/C, in high linkage disequilibrium with rs2283265 (data not shown). However, only rs2283265 remained significant after Bonferroni correction (P=.047). African and European populations have the highest frequencies of the DRD2 rs2283265 C/C genotype (Figure 2; DRD2 rs2283265), whereas Asian and Hispanic populations have much lower frequencies. Thus, efficacy may vary among populations and requires further evaluation.

DRD3 rs6280 (Ser9Gly) is associated with pramipexole response and L-DOPA toxicity.³³ Among 30 pramipexole-treated individuals, 60% of Ser/Ser genotype carriers experienced at least a 20% improvement in UPDRS scores at 2 months, compared with 13% of Ser/Gly or Gly/Gly carriers (P=.024).³⁵ DRD3 rs6280 has also been linked to L-DOPA–associated toxicities.^22,25Ser/Gly and Ser/Ser genotype carriers exhibited a 4.4- and 9.7-fold increase in visual hallucinations, respectively (P<.001).²⁵ In addition, DRD3 rs6280 Ser/Ser and DRD2 rs1799732 Ins/Ins genotypes were both independently associated with an ~2-fold increase in the prevalence ratio of L-DOPA–associated GI adverse effects (P<.05). Moreover, individuals with both DRD2 rs1799732 Ins/Ins and DRD3 rs6280 Ser/Ser risk genotypes had a nearly 5-fold increase in GI toxicity (prevalence ratio 4.63; 95% CI, 1.52–14.09; P=.0007), suggesting additive effects.²² African populations have lower frequencies of DRD2 rs1799732 Ins/Ins and DRD3 rs6280 Ser/Ser genotypes (Figure 2). Therefore, these associations may not be generalizable across populations. 

Solute Carrier Transporters (SLC6A3 and SLC22A1)

Solute carrier (SLC) transporters are membrane-bound proteins (>400 across 65 families) that transport a wide range of endogenous and exogenous substances (eg, ions, neurotransmitters, xenobiotics). They exhibit varying substrate specificity and play essential roles in physiologic and pharmacologic processes.³⁴

SLC6A3 encodes the dopamine transporter, a crucial regulator of dopamine neurotransmission.²⁸SLC6A3 rs3836790 is a functional variant represented by alleles containing 5 or 6 repeats. The influence of rs3836790 on L-DOPA efficacy was evaluated among 61 individuals as part of a multicenter, parallel-group, double-blind, placebo-controlled trial, randomized to receive add-on methylphenidate or placebo.²⁴ Individuals with the 6/6 genotype had an average UPDRS motor score improvement of 17.2±4.0, compared with 12.3±2.5 for individuals with the 5/5 and 5/6 genotypes (P<.0001). This remained significant after adjusting for age, sex, weight, disease duration, and L-DOPA dose (P=.0004). In addition, the 6/6 genotype was associated with improvement in the number of steps (P=.02), completion time (P=.016), and freezing gait episodes (P=.004) on the SWS test. When limited to individuals who received methylphenidate (n=28), SLC6A3 rs3836790 6/6 was associated with motor UPDRS score (P=.0002), number of steps (P=.0003), completion time (P=.0009), and number of freezing gait episodes (P=.017) in the “On” condition, but only SWS completion time (P=.027) in the “Off” condition.

SLC22A1, encoding organic cation transporter 1 (OCT1), transports endogenous monoamines (eg, dopamine, serotonin, epinephrine), cationic drugs, and xenobiotics.²⁸ The relationship between SLC22A1 rs622342 and cumulative defined daily doses of L-DOPA plus coprescribed antiparkinsonian medications (eg, dopamine agonists, COMT inhibitors, anticholinergics, amantadine) and time from L-DOPA initiation to death from any cause was evaluated in 99 L-DOPA–treated individuals.³⁵ Each rs622342 C allele was associated with a .34 increase in defined daily doses (95% CI, .064–.62) of any antiparkinsonian medication. Average survival time after L-DOPA initiation was highest among A/A genotype carriers (6.9 years), and lower for A/C (5.2 years) and C/C (4.4 years) genotype carriers (hazard ratio, 1.47; 95% CI, 1.01–2.13). However, given the different combinations of medications and doses, it is difficult to determine how these findings would translate clinically. European and Hispanic populations have higher frequencies of the C allele compared with East Asian and South Asian populations (Figure 2; SLC22A1). Similarly, people with PD in the United States and Europe have higher average L-DOPA dose requirements compared with individuals in Japan.³⁶SLC22A1 variation may offer an explanation for this observed variability.

Homer Scaffold Protein 1 (HOMER1)

HOMER1 encodes a neuronal protein that regulates glutamate receptor function.²⁸ The influence of HOMER1 rs4704559 on L-DOPA–associated adverse effects was evaluated in 205 individuals (83.7% European; 16.3% African).²³ Individuals with A/G or G/G genotypes had a lower prevalence of dyskinesia (31.8%) compared with A/A carriers (68.2%; P=.043). After multiple analyses using Poisson regression, researchers found that individuals with the G allele had a 38.5% lower prevalence of dyskinesia (P=.0009) and a 48.5% lower prevalence of visual hallucinations (P=.020). African and Hispanic populations have the highest frequencies of the G allele (Figure 2; HOMER1). However, the incidence of L-DOPA–associated dyskinesia (~30% to 35% after several years of treatment) does not vary among populations.³⁷ Based on current evidence, clinical factors, including L-DOPA dose, younger age at onset, and disease duration, are more predictive of dyskinesia than HOMER1 genetic variation.

Future Directions and Challenges 

Pharmacogenetic testing provides a promising path to personalizing PD treatment.⁶ Although current evidence is limited by sample size and diversity, a number of the PD pharmacogenetic associations discussed have been identified in substudies or as a component of PD randomized trials.^21,24,26 This contrasts with many other pharmacogenetic studies that have used cohort or observational studies.³⁸ Whereas this provides a stronger approach to assess pharmacogenetic predictors of response, evaluation is complicated by complexity and variability in PD symptoms, subjective assessments, varying definitions of therapeutic response across studies, and numerous drug combinations that cannot be adequately assessed. In evaluating the impact of genetic variation, differences in population frequencies of variants found to influence the efficacy or safety of antiparkinsonian medications are important considerations.

In addition, the majority of PD pharmacogenetic studies have used a selective candidate gene approach, prioritizing the evaluation of genes in dopamine-related pathways (eg, DRD2, DRD3, SLC6A3). Few of the studies discussed assessed genes in pharmacokinetic pathways, such as phase I (eg, cytochrome P450 enzymes) or phase II (eg, glucuronidation) metabolism.²¹ Given that many pharmacogenetic associations with established clinical utility involve pharmacokinetic pathways, these may also influence the efficacy and safety of antiparkinsonian medications and warrant further investigation. Whereas we focused on PharmGKB clinical annotations for this evaluation, these annotations are manually indexed and curated and therefore may not represent the totality of PD pharmacogenetics evidence. To address this limitation, PharmGKB has recently added the PGxMine pipeline to automatically extract variant-drug sentences from PubMed, in addition to reference-list chaining.^39,40 This method identified 78% of pharmacogenetic-based articles, an increase from only 33% using manual indexing.³⁹ However, machine-generated entries are not curator-reviewed, so unpublished or new PD pharmacogenetic findings may not be considered in assignment of current PharmGKB evidence levels.

Conclusion

Variants in key genes, including COMT, DRD2, DRD3, HOMER1, SLC6A3, and SLC22A1, have been found to affect the efficacy or safety of commonly used antiparkinsonian medications, including L-DOPA, dopamine agonists, MAO-B inhibitors, and COMT inhibitors.⁵ Given the sizable proportion of people with PD who experience inadequate motor response or adverse effects, genetic factors may help explain variable response. However, evidence for these associations is limited. Further research is needed to validate existing findings, identify additional gene–drug interactions, and assess the utility of genotype-guided therapy in personalizing treatment for PD.