<?xml version="1.0"?>
<metadata xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:dc="http://purl.org/dc/elements/1.1/"><dc:title>Pharmacogenetic markers of adverse events of dopaminergic treatment in Parkinson's disease</dc:title><dc:creator>Redenšek,	Sara	(Avtor)
	</dc:creator><dc:creator>Dolžan,	Vita	(Mentor)
	</dc:creator><dc:creator>Trošt,	Maja	(Komentor)
	</dc:creator><dc:subject>Parkinson's disease</dc:subject><dc:subject>pharmacogenetics</dc:subject><dc:subject>dopaminergic treatment</dc:subject><dc:subject>adverse events</dc:subject><dc:subject>genetic polymorphisms</dc:subject><dc:subject>personalized medicine</dc:subject><dc:subject>clinical-pharmacogenetic models</dc:subject><dc:description>Introduction
Parkinson's disease (PD) is the second most prevalent neurodegenerative disease. The main pathological hallmark is the death of dopaminergic neurons in the nigrostriatal pathway, which mainly affects motor function. Therefore dopamine replacement strategy has a central role in PD management. The most prescribed dopaminergic drugs in PD are levodopa and dopamine agonists. Symptomatic dopaminergic treatment is very effective, but poses risk for occurrence of motor and non-motor adverse events. Motor fluctuations and dyskinesia are the most frequent motor adverse events. Motor fluctuations present as oscillations between good and bad motor symptom control, while dyskinesia usually manifest as involuntary movements. They mostly occur after long-term levodopa administration, however they are sometimes observed shortly after treatment initiation. Non-motor adverse events, such as excessive daytime sleepiness and sleep attacks, visual hallucinations, nausea/vomiting, orthostatic hypotension, peripheral oedema, and impulse control disorders, are associated with both levodopa and dopamine agonists. Although some risk factors for adverse events are known, their occurrence and time to occurrence in individual patients cannot be predicted yet. Genetic variability in the dopaminergic and other pathways affecting the development and progression of PD could serve as foundation for research and identification of pharmacogenetic markers of adverse events. 
Dopaminergic pathway is the most heavily disrupted neurotransmitter pathway in the pathogenesis of PD. Polymorphic genes encode several enzymes (DDC, COMT, MAOB), transporters (SLC6A3, SLC18A2, SLC7A5, SLC22A1), receptors (DRD1-5) and other molecules (SV2C) leading to inter-individual variability in the capacity of dopamine synthesis, transport, degradation, and signalling. Several pharmacogenetic studies searching for associations between single nucleotide polymorphisms (SNPs) of above mentioned genes and response to dopaminergic treatment have already been performed. 
Dopaminergic treatment outcome may also be influenced by pathways related to molecular pathogenesis of the disease. Neuroinflammation and oxidative stress related pathways contribute to PD pathogenesis. Several SNPs in the IL1&amp;#946;, TNF&amp;#945;, and IL6 have already been associated with PD susceptibility. Genetic variability and/or activities of the inflammasome NALP3, glutathione peroxidase (GPX1), catalase (CAT), superoxide dismutase (SOD2), and nitric oxide synthase (NOS1) have been associated with PD susceptibility. Additionally, genetic variability of the neurodevelopment and apoptosis pathways have also been associated with PD susceptibility.
Genetic variability of dopaminergic pathway has been extensively studied, but only individual genes were analysed. Furthermore, pathways affecting PD development and progression may play an important role in the treatment response, but no data on this has been reported so far. 
Aim of the study and hypotheses
The aetiology of adverse events is not completely understood and their occurrence cannot be clearly predicted. Some clinical and pharmacogenetic factors were already identified as potential markers of the adverse events. However, a comprehensive evaluation of the genetic variability in (i) dopaminergic pathway and pathways of (ii) neuroinflammation, (iii) oxidative stress, (iv) neuron development, proliferation, and differentiation, and (v) apoptosis has not been conducted. The aim of this study was to use pathway-based approach to evaluate SNPs from the above listed pathways as potential pharmacogenetic markers of adverse events of dopaminergic treatment. Furthermore, we wanted to construct clinical and clinical-pharmacogenetic models for prediction of time to occurrence of motor complications after initiation of levodopa treatment. 
In this study we tested the following hypotheses:
• Genetic variability in the pathways of transport, metabolism, and mechanism of action of dopamine, levodopa and other dopaminergic substances influences the occurrence and/or time to occurrence of adverse events of dopaminergic treatment in PD.
• Genetic variability in pathways affecting the development and progress of PD has an influence on occurrence and/or time to occurrence of adverse events of dopaminergic treatment in PD.
• Clinical-pharmacogenetic models can help us to evaluate the simultaneous influence of certain genetic polymorphisms and clinical factors on the occurrence of adverse events of dopaminergic treatment in PD.
Participants and methods 
A total of 231 unrelated PD patients were enrolled in this retrospective study at the Department of Neurology, University Medical Centre Ljubljana, Slovenia between October 2016 and April 2018. Inclusion criteria were (i) diagnosis of PD according to the UK Parkinson Disease Society Brain Bank criteria by an experienced movement disorders specialist, (ii) available clinical data, (iii) at least three months of an ongoing levodopa and/or dopamine agonists treatment duration. The study protocol was approved by the Slovenian Ethics Committee for Research in Medicine (KME 42/05/16). All of the patients signed the informed consent.
Each patient donated a sample of their peripheral blood for DNA isolation within withdrawal for other routine laboratory tests. All patients were genotyped for 37 SNPs in the genes involved in the processes of (i) dopaminergic pathway, (ii) neuroinflammation, (iii) oxidative stress, (iv) neuron development, proliferation, and differentiation, and (v) apoptosis.
Median and 25th to 75th percentile range were used to describe central tendency and variability of continuous variables, while frequencies were used to describe the distribution of categorical variables. The agreement of genotype frequencies with Hardy-Weinberg equilibrium was examined by Chi-squared test. Logistic regression was used to calculate odds ratios (ORs), 95% confidence intervals (95%CIs), and p-values to examine the associations of selected SNPs and clinical data with the risk for adverse events. Allele with the highest frequency was used for reference. Cox proportional hazards models were used to estimate the association of SNPs and clinical covariates with the time to occurrence of motor complications after levodopa treatment initiation. Results were reported as hazard ratios (HR), 95%CI, and p-values. Models were built using Cox analysis with LASSO (Least Absolute Shrinkage and Selection Operator) penalization due to a large number of explanatory variables in comparison to the number of events. For both models the time-dependent receiver operating characteristic (ROC) curves were constructed, where sensitivity, specificity, and tAUC were assessed. The predictive scoring system was estimated by selecting the threshold that provided the maximized sum of the cross-validated true positive rate and true negative rate. Cross-validation was applied on all the predictive accuracy estimates to avoid biased and overoptimistic results. Statistical analyses were carried out by IBM SPSS Statistics, version 21.0 (IBM Corporation, Armonk, NY, USA) and the R software.
Results
Within the first part of the study we studied associations between genetic variability of the dopaminergic pathway and the occurrence of adverse events of dopaminergic treatment. We observed some important associations: visual hallucinations were associated with COMT rs165815 C allele (OR=0.34; 95%CI=0.16-0.72; p=0.004), DRD3 rs6280 C allele (OR=1.88; 95%CI=1.00-3.54; p=0.049), and DRD3 rs6280 CC genotype (OR=3.31; 95%CI=1.37-8.03; p=0.008); orthostatic hypotension was associated with DDC rs921451 C allele (OR=1.86; 95%CI=1.07-3.23; p=0.028), DDC rs921451 CT genotype (OR=2.30; 95%CI=1.26-4.20; p=0.007), DDC rs3837091 AGAG- genotype (OR=1.94; 95%CI=1.07-3.51; p=0.028), and SLC22A1 rs628031 AA genotype (OR=2.57; 95%CI=1.11-5.95; p=0.028); peripheral oedema were associated with SLC22A1 rs628031 AA genotype (OR=4.00; 95%CI=1.62-9.88; p=0.003); impulse control disorders were associated with SLC22A1 rs628031 AA genotype (OR=3.16; 95%CI=1.03-9.72; p=0.045); and dyskinesia were associated with SLC22A1 rs628031 GA genotype (OR=0.48; 95%CI=0.24-0.98; p=0.043), and with SLC22A1 rs628031 A allele (OR=0.48; 95%CI=0.25-0.92; p=0.027).
In the analysis of the genetic variability of the neuroinflammation pathway only one association showing a nonsignificant trend was observed. The IL1&amp;#946; rs1143623 C allele was associated with lower odds for developing orthostatic hypotension (OR=0.57; 95%CI=0.32-1.00, p=0.050), while heterozygotes had nominally significant lower odds for developing orthostatic hypotension (OR=0.51; 95%CI=0.28-0.93, p=0.028). 
Genetic variability of the oxidative stress pathway also showed some important associations with the occurrence of adverse events of dopaminergic treatment: peripheral oedema were associated with CAT rs1001179 A allele (OR=0.32; 95%CI=0.15-0.68, p=0.003) and CAT rs1001179 GA genotype (OR=0.32; 95%CI=0.15-0.69, p=0.004); excessive daytime sleepiness and sleep attacks were associated with NOS1 rs2682826 A allele (OR=1.75; 95%CI=1.00-3.06, p=0.048) and NOS1 rs2682826 AA genotype (OR=3.75; 95%CI=1.23-11.45, p=0.020); nausea/vomiting was associated with SOD2 rs4880 T allele (OR=0.49; 95%CI=0.25-0.94, p=0.031) and SOD2 rs4880 CT genotype (OR=0.48, 95%CI=0.24-0.98, p=0.045); impulse control disorders were associated with NOS1 rs2682826 A allele (OR=2.59; 95%CI=1.09-6.19, p=0.032) and NOS1 rs2682826 GA genotype (OR=2.74, 95%CI=1.13-6.64, p=0.025).
Variables associated with time to occurrence of motor fluctuations after initiation of levodopa treatment were: age at diagnosis (HR=0.97; 95%CI=0.96-0.99; p&lt;0.001), time from diagnosis to initiation of levodopa treatment (HR=1.36; 95%CI=1.25-1.49; p&lt;0.001), NOS1 rs2293054 (GG: Ref.; AA: HR=0.36; 95%CI=0.13-1.00; p=0.051), DRD2 rs1799732 (CC: Ref.; --: HR=8.89; 95%CI=1.19-66.18; p=0.033), and DRD3 rs6280 (TT: Ref.; CC: HR=2.04; 95%CI=1.16-3.60; p=0.014).
Variables associated with time to occurrence of dyskinesia after initiation of levodopa treatment were: age at diagnosis (HR=0.96; 95%CI=0.95-0.98; p&lt;0.001), beta-blockers (HR=0.60; 95%CI=0.36-1.00; p=0.051), time from diagnosis to initiation of levodopa treatment (HR=1.23; 95%CI=1.11-1.37, p&lt;0.001), CAT rs1001179 (GG: Ref.; AA: HR=2.60; 95%CI=1.17-5.79; p=0.019), SOD2 rs4880 (CC: Ref.; TT: HR=0.54; 95%CI=0.30-0.98; p=0.043), SLC22A1 rs628031 (GG: Ref.; GA: HR=0.63; 95%CI=0.40-1.00; p=0.048; AA: HR=0.53; 95%CI=0.29-0.99; p=0.047), DRD2 rs1799732 (CC: Ref.; --: HR=8.66; 95%CI=1.16-64.86; p=0.036), and NRG1 rs3735781 (AA: Ref.; GA: HR=0.65; 95%CI=0.41-1.99; p=0.051).
The clinical-pharmacogenetic model for motor fluctuations included age at diagnosis (HR=0.99), time from diagnosis to initiation of levodopa treatment (HR=1.24), COMT rs165815 (HR=0.90), DRD3 rs6280 (HR=1.03), and BIRC5 rs9904341 (HR=0.95). With the ROC curve after five years of treatment (cross-validated AUC=0.70, sensitivity=52.2%, specificity=82.3%). We constructed the ROC curve for data after five years of treatment for the clinical model for comparison (AUC=0.68; sensitivity=48.4%, specificity=81.9%).
The clinical-pharmacogenetic model for dyskinesia included sex (HR=1.07), age at diagnosis (HR=0.97), tremor-predominant PD (HR=0.88), beta-blockers (HR=0.95), alcohol consumption (HR=0.99), time from diagnosis to initiation of levodopa treatment (HR=1.15), CAT rs1001179 (HR=1.27), SOD2 rs4880 (HR=0.95), NOS1 rs2293054 (HR=0.99), COMT rs165815 (HR=0.92), and SLC22A1 rs628031 (HR=0.80). With the ROC curve of five year treatment (cross-validated AUC=0.68, sensitivity=54.1%, specificity=66.1%). We constructed the ROC curve for data after five years of treatment for the clinical model for comparison (AUC=0.71; sensitivity=79.8%, specificity=48.4%).
Discussion
In this study we evaluated the effect of 34 SNPs from five different biological pathways and several clinical parameters on the occurrence and also time of occurrence of adverse events of dopaminergic treatment in PD patients. 
We observed a protective effect of the COMT rs165815 C allele against visual hallucinations This SNP has never been studied in association with this phenotype. However, COMT rs4680 already showed association with visual hallucinations. On the other hand, DDC genotypes were nominally significantly associated with the elevated odds for occurrence of orthostatic hypotension. It was suggested that the studied SNPs reduce DDC expression or activity, which could lead to lowered concentrations of noradrenalin and consequential vasodilatation. This mechanism may support the occurrence of orthostatic hypotension due to dopaminergic treatment. Furthermore, the SLC22A1 rs628031 appeared to be a promising genetic marker of several adverse events of dopaminergic treatment. It was associated with higher odds for development of orthostatic hypotension, peripheral oedema, and impulse control disorders. It was also associated with lowered odds for development of dyskinesia. Since this SNP appeared to be important in the occurrence of several adverse events, it may play a role in the overall drug action and may thus modify the overall susceptibility to adverse events. DRD3 rs6280 CC genotype was observed as a risk factor for visual hallucinations. This SNP may lead to modifications in intracellular signalling via higher binding affinity
Our study showed one, however nonsignificant, association of the orthostatic hypotension with IL1&amp;#946; rs1143623. Some studies already explored and observed an association between inflammation and orthostatic hypotension, but our study was the first to point out an association with a genetic component. This SNP has already showed an association with PD susceptibility.
CAT rs1001179 was significantly associated with the occurrence of peripheral edema. The mechanism of this adverse event is not fully understood. Our results suggest that oxidative pathway might be involved in its occurrence. We observed that the NOS1 rs2682826 is associated with the excessive daytime sleepiness and sleep attacks and impulse control disorders. The NOS1 genotype might affect sleep cycle according to some papers. Additionally, NOS1 rs2682826 has already been associated with several psychiatric diseases, such as depression and anxiety. SOD2 rs4880 was associated with lowered odds for nausea/vomiting occurrence. Some studies have already showed association of oxidative stress in the enteric nervous system with development of gastrointestinal complications.
We constructed clinical and clinical-pharmacogenetic models for prediction of time to occurrence of motor complications and compared them. These models could serve for construction of algorithms for identification of patients with elevated risk for earlier occurrence of motor complications. The difference in predictive capacities of the clinical and clinical-pharmacogenetic models was not clinically important. The cross-validated AUCs, specificities, and sensitivities for clinical and clinical-pharmacogenetic models for motor fluctuations did not differ significantly. In the case of dyskinesia the clinical-pharmacogenetic model had higher specificity in comparison to the clinical model. This means that the probability of an actual adverse event in the case of a positive result is higher under the clinical-pharmacogenetic model. All of the models correctly predicted the occurrence of the adverse event in the first five years of levodopa treatment in the majority of cases.
Conclusions 
With this study we have shown that genetic variability influences the occurrence and time to occurrence of adverse events of dopaminergic treatment in PD.
We have confirmed or at least partially confirmed the following hypotheses:
• SNPs of the dopaminergic pathway influence the occurrence of several adverse events of dopaminergic treatment in PD.
• IL1&amp;#946; rs1143623 from the neuroinflammation pathway shows a trend towards an association with the occurrence of orthostatic hypotension as an adverse event of dopaminergic treatment in PD.
• SNPs from genes from the oxidative stress pathway influence the occurrence of several non-motor adverse events of dopaminergic treatment in PD.
• Clinical-pharmacogenetic models may enable a simultaneous evaluation of the influence of clinical and genetic factors on the time of occurrence of motor complications due to levodopa treatment in PD.</dc:description><dc:date>2019</dc:date><dc:date>2020-08-12 09:55:54</dc:date><dc:type>Doktorsko delo/naloga</dc:type><dc:identifier>118016</dc:identifier><dc:identifier>VisID: 13240</dc:identifier><dc:identifier>COBISS_ID: 302939136</dc:identifier><dc:language>sl</dc:language></metadata>
