P.5.d.009 - Predictive genetic model for levodopa-induced dyskinesia in patients with Parkinson’s disease

Abstract

Predictive genetic model for levodopa-induced dyskinesia in patients with Parkinson's disease S.A. Ivanova(1), V.M. Alifirova(2), M.B. Freidin(3), I.V. Pozhidaev(4), O.Y. Fedorenko(4), N.A. Bokhan(4), I.A. Zhukova(5), N.G. Zhukova(5), B. Wilffert(6), A.J.M. Loonen(6) (1)Mental Health Research Institute, Molecular Genetics and Biochemistry Laboratory, Tomsk, Russia (2)Siberian State Medical University, Department of Neurology and Neurosurgery, Tomk, Russia (3)Tomsk National Research Medical Center of the Russian Academy of Sciences, Research Institute of Medical Genetics, Tomsk, Russia (4)Tomsk National Research Medical Center of the Russian Academy of Sciences, Mental Health Research Institute, Tomsk, Russia (5)Siberian State Medical University, Department of Neurology and Neurosurgery, Tomsk, Russia (6)University of Groningen, Groningen Research Institute of Pharmacy, Groningen, The Netherlands Parkinson's disease (PD), a common neurodegenerative disorder caused by the loss of the dopaminergic input to the basal ganglia, is commonly treated with levodopa (L-DOPA). The use of this drug, however, is severely limited by adverse effects. Levodopa-induced dyskinesia (LID) is one of these and characterized by involuntary muscle movements that occur as a consequence of chronic levodopa treatment. LID is a substantial barrier to effective symptomatic management of PD as up to 45% of L-DOPA users develop LID within 5 years [1]. Clinical heterogeneity of LID suggests a significant role of endogenous factors in determining their prevalence. Some evidences suggest a relationship between LID and specific genetic variants, such as polymorphisms in the genes controlling enzymes responsible for drug and monoamine metabolism, neurotransmitter receptors and proteins involved in oxidative stress or antioxidant function [2–4]. Objective: To investigate a contribution of polymorphic variants of neurotransmitter receptors and cytochrome genes in the development of LID in PD patients. Methods: A total of 212 PD patients who received L-DOPA therapy were studied. Dyskinesia was assessed by using the Abnormal Involuntary Movement Scale (AIMS). DNA extraction and genotyping were conducted according to standard protocols and blind to the clinical status of the subjects. Genotyping was carried out for 72 SNPs of DRD1, DRD2, DRD2/ANKK1, DRD3, DRD4, HTR2C, HTR3A, HTR3B, HTR6, HTR2A, HTR1A, HTR1B, CYP1A2*1F, CYP2D6*3, CYP2D6*4, CYP2C19*3, CYP2C19*17, CYP2C19*2, and GSTP1 genes using MassARRAY® Analyzer 4 (Agena Bioscience™) and the set SEQUENOM Consumables iPLEX Gold 384. Discriminant analysis and receiver operating curve (ROC)-analysis were carried out to build a genetic predictive model for dyskinesia. Results: Group of PD patients consists of 149 females and 83 males (age ranging from 40 to 86 years, average age 68.7 ± 7.6 years). The mean age of onset is 60.04 ± 9.46 years, average disease duration is 9.79 ± 5.57 years. Dyskinesia was reported in 57 (26.9%) patients. The best discriminant model was obtained with the following predictors: rs11721264, rs165774, rs3758653, rs4245147, rs6313, rs1364043, rs2734849, rs324035, rs6311, rs11246226 and rs4244285. These polymorphisms are localized in the following genes: DRD3 (rs11721264, rs324035), DRD4 (rs3758653, rs11246226), DRD2 (rs4245147, rs2734849), HTR2A (rs6313, rs6311), HTR1A (rs1364043). The discriminant model using this set of SNPs gives the error of classification about 13% and the AUC 0.795. Depending on the anticipated frequency of LID, positive and negative predictor values varied between 0.745–0.834 and 0.864–0.916, respectively. We hypothesized in our previous studies that the pathological basis of LID might be degeneration of indirect pathway medium spiny neurons [5]. These indirect pathway medium spiny neurons carry type 2 family dopamine receptors (DRD2, DRD3, DRD4), and HTR2A receptors. Moreover, dopamine release may be promoted by inhibiting serotonergic neurotransmission. Hence, the current findings are well in line with this hypothesis. Conclusion: The resulting panel of 11 SNPs provides a sufficiently high accuracy of LID prediction. The use of this panel in a prospective study will clarify the prospects for its application in clinical practice for predicting risk of LID in patients with PD.