Hypothesis: Do miRNAs Targeting the Leucine-Rich Repeat Kinase 2 Gene ( LRRK2 ) Influence Parkinson's Disease Susceptibility?

Abstract

Parkinson's disease (PD) is a frequently occurring neurodegenerative motor disorder adversely impacting global health. There is a paucity of biomarkers and diagnostics that can forecast susceptibility to PD. A new research frontier for PD pathophysiology is the study of variations in microRNA (miRNA) expression whereby miRNAs serve as “upstream regulators” of gene expression in relation to functioning of the dopamine neuronal pathways. Leucine-Rich Repeat Kinase 2 (LRRK2) is a frequently studied gene in PD. Little is known about the ways in which expression of miRNAs targeting LRKK2 impact PD susceptibility. In a sample of 204 unrelated subjects (102 persons with PD and 102 healthy controls), we report here candidate miRNA expression in whole blood samples as measured by real-time PCR (hsa-miR-4671-3p, hsa-miR-335-3p, hsa-miR-561-3p, hsa-miR-579-3p, and hsa-miR-3143) that target LRRK2. Using step-wise logistic regression, and controlling for covariates such as age, gender, PD disease severity, concomitant medications, and co-morbidity, we found that the combination of has-miR-335-3p, has-miR-561-3p, and has-miR-579-3p account for 50% of the variation in regards to PD susceptibility (p < 0.0001). Notably, the hsa-miR-561-3p expression was the most robust predictor of PD in both univariate and multivariate analyses (p < 0.001). Moreover, the biological direction (polarity) of the association was plausible in that the candidate miRNAs displayed a diminished expression in patients. This is consistent with the hypothesis that decreased levels of miRNAs targeting LRRK2 might result in a gain of function for LRRK2, and by extension, loss of neuronal viability. To the best of our knowledge, this is the first clinical association study of the above candidate miRNAs' expression in PD using peripheral samples. These observations may guide future clinical diagnostics research on PD.

Introduction

Parkinson's Disease (PD) is the second most common and progressive neurodegenerative complex disease after Alzheimer's disease. The progressive degeneration of neurons results from an elusive interaction between genetic and environmental factors (Pringsheim et al., 2014). The incidence studies in elderly populations have been reported to vary between 410 and 529/100000 (de Lau et al., 2004). The variations of PD incidence among populations might offer clues about the role of genetic and environmental factors. For instance, the incidence among persons over the age of 55 was found in the range of 410–529/100000 in Europe and 1.5–17/100000 in Asia (Van Den Eeden et al., 2003). PD is often sporadic with about 10% of patients suggesting a positive family history.

Several specific chromosomal loci, including the leucine-rich repeat kinase 2 (LRRK2) gene located on chromosome 12q12, have been identified in relation to PD (Klein et al., 2012). LRRK2 gene mutations are the most common known cause of late-onset autosomal dominant and sporadic PD (Brice et al., 2005, Ozelius et al., 2006). On the other hand, relatively little is known about the ways in which expression of miRNAs targeting LRKK2 impact PD susceptibility. miRNAs are small endogenous noncoding RNA molecules about 19–21 nucleotids (nt) in length. In general, miRNAs play an important role in post-transcriptional regulation of gene expression via, for example, translational repression or mRNA degradation. The translational repression occurs by base pairing between the mature miRNAs and 3′-UTR or 5′-UTR of target mRNA (Ardekani and Naeini, 2010).

In a sample of 204 unrelated subjects, 102 persons with PD (65 male and 37 female) and 102 healthy (44 male and 58 female) controls, we report here candidate miRNA expression in whole blood samples as measured by real-time PCR (hsa-miR-4671-3p, hsa-miR-335-3p, hsa-miR-561-3p, hsa-miR-579-3p, and hsa-miR-3143) that target LRRK2. This is the first report, to the best of our knowledge, of these miRNAs with pertinence for LRRK2 gene function, and by extension, for PD susceptibility. The findings might pave the way for greater research emphasis on “upstream” regulation of LRRK2 gene function, and stimulate future clinical diagnostics development using miRNA pathways as a new focus in PD precision medicine research.

Materials and Methods

Subjects

The participants were recruited from the Neurology Department of the Medical Faculty of Gaziantep University in southeast Turkey. The study group was comprised of 102 unrelated Turkish individuals with PD and 102 unrelated Turkish healthy volunteer controls. The patients with PD were diagnosed according to the National Institute of Neurological Disorders and Stroke (NINDS) criteria (Bhidayasiri et al., 2013). The control subjects had no PD history. Additionally, we obtained detailed clinical and demographic data concerning age, gender, PD disease staging, co-morbidity related medical history (e.g., hypertension, diabetes mellitus), family history of PD (noted as positive if a patient had at least one family member with PD), prescription drug use history for PD (grouped as users of either L-DOPA, DOPA-Agonist, Rasajilin, L-DOPA + DOPA-Agonist, DOPA-Agonist + Rasajilin, or L-DOPA + DOPA-Agonist + Rasajilin) and drugs for other diseases if applicable. Staging of PD was made with the Hoehn-Yahr scale (classification stage: 1, 1.5, 2, 2.5, 3, 4, 5) (Martín et al., 2015).

The study was reviewed and approved by the institutional ethics oversight committee of the Faculty of Medicine at Gaziantep University. All subjects provided written informed consent. A 10 mL blood sample from each subject was drawn into a coded heparinized tube for molecular genetics analyses.

Rationale for the selection of miRNA genes for expression analyses

Eleven distinct readily accessible online web-based databases (MicroCosm Targets, TargetScan, miRNAMap, microInspector, PicTar, miRWalk, mirbase, mirdb, Patrocles, Diana lab, and Human MiRNAs and Diseases databases) were used to estimate and identify the miRNAs targeting the LRRK2 gene (Griffiths et al., 2006; John et al., 2006; Vlachos et al., 2012). Subsequently, their expressions in the blood were checked using the Ferrolab (Ferrolab et al., 2010) and the Miracle (Zotos et al., 2012) databases. Finally, the microRNA database was used (Wang, 2008) in order to demonstrate the efficacy values (matching scores) of the match with their target genes of selected miRNAs. The expression of five candidate miRNAs (hsa-miR-4671-3p, hsa-miR-335-3p, hsa-miR-561-3p, hsa-miR-579-3p, and hsa-miR-3143) in whole blood samples were measured by real-time PCR.

Reverse transcriptase PCR reactions (RT-PCR)

RNA was extracted from whole blood using the modified method of Trizol® (Avison, 2005). Reverse transcriptase reactions were carried out with an RT kit (miScript II RT Kit, 218161, Qiagene) by the manufacturer's protocol. The 25 μL PCR included 10 ng total RNA, 5X miScript HiSpec Buffer, 10X miScript Nucleics Mix, miScript Reverse Transcriptase Mix, and RNAse free water. The reaction was performed on an automated Thermal Cycler (G-Storm, UK) with condition for 30 min at 16°C, 30 min at 42°C, 5 min at 85°C, and then cDNA samples were held at −20°C. The expression levels for cDNAs belonging to the patient or control subjects were determined with hsa-miR-26b endogenous control gene by Real-Time PCR and data were calculated by quantitative-comparative analysis.

Quantitative-Comparative CT (ΔΔCT) real-time PCR

Quantitative-Comparative CT (ΔΔCT) Real-time PCR was performed in a Stratagene MX 300P Real-Time PCR System (Agilent Technologies In.) using the Plexor™ Analysis Software. The hsa-miR-26b-5p (Qiagene) was used as endogenous control microRNA for other miRNAs. We used as a reference RNA sample (Life Technologies, AM7155). The 25 μL PCR included 5 μL RT-PCR products, 2X SYBRGreen Universal PCR Master Mix (Qiagen), and 900 nM of each primer (hsa-miR-4671-3p, hsa-miR-335-3p, hsa-miR-561-3p, hsa-miR-579-3p, and hsa-miR-3143) (Qiagen). The reactions were incubated in a 96-well plate of preincubation at 50°C for 2 min and at 95°C for 10 min, followed by 40 cycles at 95°C for 15 seconds, 55°C for 30 sec, and at 70°C for 30 sec. All reactions were run in triplicate. ΔΔCT values were used and calculated of 2^-ΔΔCT values (Livak et al., 2001).

Statistical analyses

The data were first normalized to be appropriate for analysis. We initially carried out univariate analysis using various methods for comparison purposes. The models were subsequently refined by systematically dropping variables that did not explain a significant proportion of the differences between cases and controls. Models were compared using likelihood ratio tests Bayesian information criteria. Analyses were performed using Stata version 11 (StataCorp, Texas, USA). The chi-squared test was used to determine the gender distribution between PD patients and controls.

Because the study key aim was to identify the miRNAs that might predict subjects' status as being a PD patient versus a healthy control, we used step-wise logistic regression while controlling for potential confounding covariates including age, gender, PD disease stage, concomitant medications, and co-morbidity. The association analyses of each miRNA expression data were entered as a univariate predictor initially while accounting for the above co-variates. Subsequently, we performed more nuanced bivariate analyses where two miRNA expression data were entered as a predictor variable in the logistic regression analyses. The findings were expressed with two-tailed p-values (declared significant if less than 0.05), 95% confidence intervals (CIs), odds ratios and the contribution of each variable measured by pseudo R² in regards to the strength of the prediction of the PD versus control status with a given miRNA variable.

Results

We report here on a study sample of 204 individuals, 102 PD patients and 102 healthy controls. The patient group had 65 males and 37 females, whereas the control group had 44 males and 58 females. The age distribution of the patients and controls were 67.08 ± 1.12 and 62.70 ± 0.85, respectively. While these differences in demographics in terms of age and gender were significant (p < 0.05), we controlled for demographic variation in multivariate analyses. In all, gender and age each accounted for 3% of the variation for PD susceptibility.

The data concerning the expression of each of the five candidate miRNAs in whole blood samples are shown in Table 1. Statistical analyses of the miRNA expression data, odds ratios, and 95% confidence intervals (CIs) are presented in Table 2. Finally, step-wise multivariate analyses and the candidate biomarkers with greatest impact for PD susceptibility are shown in Table 3. The latter Table 3 illustrates that the step-wise multivariate logistic regression was carried out by removing individual variables at each stage, while controlling for the potential confounding variables noted under the statistical analyses (age, gender, co-morbidity, concomitant medications, etc.), so as to identify the minimal predictive model. Of note, in general the expression of each of the five miRNAs was lower in the PD patients compared to the healthy controls (Table 1).

Table 1.

miRNA Expression Values in Parkinson's Disease Patients and Healthy Controls

miRNA name	Patients/Controls (N)	Median Expression Value (Patients/Controls)	25%	75%
hsa-miR-4671-3p	102/102	0.4266	0.2797	0.5312
		0.4434	0.3161	0.6606
hsa-miR-335-3p	102/102	0.5329	0.4230	0.6939
		0.7918	0.7453	0.8562
hsa-miR-561-3p	102/102	0.4422	0.3544	0.5241
		0.7806	0.7192	0.8562
hsa-miR-579-3p	102/102	0.3232	0.1909	0.4193
		0.7019	0.6034	0.8231
hsa-miR-3143	102/102	0.4466	0.2849	0.5777
		0.5832	0.4319	0.6886

Median (50^th percentile), first quartile (25^th percentile), third quartile (75^th percentile), are reported here. N, Numbers of individuals.

Table 2.

Univariate Analyses of the Five miRNA Expression Data (Predictor Variables) for Association with Parkinson's Disease

Univariate Predictor Variable	Odds Ratio	p-value	95% CI
hsa-miR-4671-3p	2.332	0.44	[0.268–20.251]
hsa-miR-335-3p	0.051	0.039	[0.003–0.861]
hsa-miR-561-3p	0.001	0.001	[0.000–0.027]
hsa-miR-579-3p	0.048	0.038	[0.003–0.846]
hsa-miR-3143	0.863	0.900	[0.087–8.575]

An odds ratio of less than 1.0 indicates lesser expression in patients than healthy controls. P-values <0.05 were considered significant without correction for multiple testing. However, with 5 miRNA expression data serving as predictor variables, we note that the hsamiR-561-3p significance level (p = 0.001) would still survive correction for multiple testing (where significant = p < 0.01). As this is the first study of these 5 miRNAs in relation to the Parkinson's disease, future studies in larger samples are required for potential diagnostics' development. Boldface indicates statistically significant.

Table 3.

Multivariate Step-wise Logistic Regression Analyses of miRNA and Demographic Predictor Variables for Association with Parkinson's Disease

Statistical model variables	Likelihood Ratio Chi ²	Statistical significance P-value	Variance explained (%) pseudo R ²
Full model
hsa-miR-4671-3p; hsa-miR-335-3p; hsa-miR-561-3p; hsa-miR-579-3p; hsa-miR-3143; Age; Gender	145.85	<0.0001	52
Full model minus age
hsa-miR-4671-3p; hsa-miR-335-3p; hsa-miR-561-3p; hsa-miR-579-3p; hsa-miR-3143; Gender	144.61	0.0000	51
Full model minus gender
hsa-miR-4671-3p; hsa-miR-335-3p; hsa-miR-561-3p; hsa-miR-579-3p; hsa-miR-314-3; Age	143.82	<0.0001	51
Full model minus demographic factors (age and gender)
hsa-miR-4671-3p; hsa-miR-335-3p; hsa-miR-561-3p; hsa-miR-579-3p; hsa-miR-3143	142.55	<0.0001	50
Minimal model from step-wise logistic regression
hsa-miR-335-3p; hsa-miR-561-3p and hsa-miR-579-3p	142.26	<0.0001	50
Minimal model from step-wise logistic regression including the demographic data
hsa-miR-335-3p; hsa-miR-561-3p; hsa-miR-579-3p; Age; Gender	145.26	<0.0001	51
Bivariate analysis of the miRNAs from the minimal model
hsa-miR-335-3p and hsa-miR-561-3p	138.24	<0.0001	49
hsa-miR-561-3p and hsa-miR-579-3p	136.49	0.0000	48
hsa-miR-335-3p and hsa-miR-579-3p	118.83	0.0001	42
Demographic data
Age only	9.49	0.0021	3
Gender only	8.75	0.0031	3
Age and gender	17.07	0.0002	6

Discussion

The present study makes the following salient observations. First, we found in univariate analyses that hsa-miR-335-3p; hsa-miR-561-3p, and hsa-miR-579-3p are significantly associated with PD susceptibility (p < 0.05). While the results are not corrected for multiple testing, we suggest that hsamiR-561-3p expression was highly significant for association with PD at a p = 0.001 level, thus suggesting that this biomarker offers formidable potential for study in future studies of PD.

Second, using step-wise logistic regression, and controlling for covariates such as age, gender, PD disease severity, concomitant medications, and co-morbidity, we found that the combination of hsa-miR-335-3p, hsa-miR-561-3p, and hsa-miR-579-3p account for 50% of the variation in regards to PD susceptibility (p < 0.0001) (Table 3). Notably, again, the hsa-miR-561-3p expression was the most robust in both univariate and multivariate analyses (p < 0.001).

Third, the biological direction (polarity) of the observed significant associations was plausible in that the candidate miRNAs displayed a diminished expression in patients. This is consistent with the idea that the decreased levels of miRNAs that target LRRK2 might result in a gain of function for LRRK2, and by extension, loss of neuronal viability (Esteves et al., 2014; Plowey et al., 2014).

We note, however, that the study is in part limited for generalizing the results because we analyzed samples in whole blood and not in the brain tissue. On the other hand, studies in peripheral tissues may also offer advantages by virtue of accessibility in future clinical diagnostics discovery research on PD.

miRNAs are functional in transcriptional and post-transcriptional regulation of gene expression by regulating the 5′-UTR or 3′-UTR of the target mRNAs. On this basis, human LRRK2 gene expression levels are in part regulated by the action of miRNAs. While LRRK2 has been of intensive research focus in the past, its upstream regulators such as miRNAs and their expression have not been studied. The present study thus makes an initial attempt to fill this knowledge gap in the literature concerning genetic basis of PD susceptibility. The pathogenesis of LRRK2 is mostly due to gain of function as suggested above. Indeed, some studies have observed that LRRK2 may be associated with loss of neuronal viability (Li et al., 2014).

Additionally, animal models are used often in research on the LRRK2 gene. Overexpression of LRRK2 leads to decrease in locomotor activity and loss of dopamin neurons in Drosophila (Liu et al., 2008). Similarly, overexpression of LRRK2 gene results in degeneration of dopamine neurons in C. elegans (Saha et al., 2009). These data collectively suggest that the downregulation of the miRNAs observed in the present study may potentially contribute to an ostensible increase in PD susceptibility.

At present, it is very difficult to establish diagnosis at an early or preclinical stage of PD. Therefore, there is a real-life need to develop biomarkers targeting pathways hitherto understudied and yet promising potential for future clinical diagnostics development. miRNA expression can therefore serve as a viable avenue in support of PD biomarker and diagnostics research. Such studies can be conducted not only in peripheral tissues but also in the brain tissue and/or the cerebrospinal fluid (CSF). Genetic studies have been providing useful leads in other neurological disorders such as stroke (Black et al., 2015) which also deserve attention in studies of persons with PD.

Conclusion

The expression of hsa-miR-561-3p in particular showed a significant reduction in whole blood samples from persons with PD. To the best of our knowledge, this is the first clinical expression association study of the above candidate miRNAs in PD using peripheral samples. These observations may guide future clinical biomarker and diagnostics research on PD in studies with a larger sample size and in the brain tissue and/or the CSF. Additionally, the tandem study of miRNA genetic variation and the LRRK2 gene expression and protein abundance is also warranted for robust and innovative diagnostics development to identify the individuals at risk for PD in the future.

Footnotes

Acknowledgments

This study was supported by intramural infrastructure research facilities of Gaziantep University Faculty of Medicine and the University of Cape Town.

Author Disclosure Statement

The authors have no financial conflict of interest.

References

Alieva

, Filatova

, Karabanov

, Illarioshkin

, Limborska

, Shadrina

, and Slominsky

. (2015). miRNA expression is highly sensitive to a drug therapy in Parkinson's disease. Parkinsonism Relat Dis, 21, 72–74.

Ardekani

, and Naeini

. (2010). The role of microRNAs in human diseases. Avicenna J Med Biotechnol, 2, 161–179.

Avison

. (2005). Measuring Gene Expression. Garland Science.

Berg

, Schweitzer

, Leitner

, Zimprich

, Lichtner

, Belcredi

, and Wszolek

. (2005). Type and frequency of mutations in the LRRK2 gene in familial and sporadic Parkinson's disease. Brain, 128, 3000–3011.

Bhidayasiri

, and Reichmann

. (2013). Different diagnostic criteria for Parkinson disease: What are the pitfalls?. J Neural Transmission, 120, 619–625.

Black

, Wang

, and Wang

. (2015). Ischemic stroke: From next generation sequencing and GWAS to community genomics?. OMICS, 19, 451–460.

Botta‐Orfila

, Morató

, Compta

, Lozano

, Falgàs

, Valldeoriola

, and Molinuevo

. (2014). Identification of blood serum micro‐RNAs associated with idiopathic and LRRK2 Parkinson's disease. J Neurosci Res, 92, 1071–1077.

Esteves

, Swerdlow

, and Cardoso

. (2014). LRRK2, a puzzling protein: Insights into Parkinson's disease pathogenesis. Exper Neurol, 261, 206–216.

Feng

, Jankovic

, and Wu

. (2015). Epigenetic mechanisms in Parkinson's disease. J Neurol Sci, 349, 3–9.

10.

Ferrolab

, and Group

. (2010). miRiam: A reliable unsupervised method for detecting miRNA binding sites on mRNAs: April 2010 Release. J RNAi Gene Silencing, 6, 1.

11.

Griffiths-Jones

, Grocock

, Van Dongen

, Bateman

, and Enright

. (2006). miRBase: microRNA sequences, targets and gene nomenclature. Nucleic Acids Res, 34, D140–D144.

12.

Heman-Ackah

, Hallegger

, Rao

, and Wood

. (2014). RISC in PD: The impact of microRNAs in Parkinson's disease cellular and molecular pathogenesis. Front Mol Neurosci, 6, 40.

13.

John

, Sander

, and Marks

. (2006). Prediction of human microRNA targets. Methods Mol Biol, 342, 101–113.

14.

Khan

, Jain

, Lynch

, Pavese

, Abou-Sleiman

, Holton

, and Gibbons

. (2005). Mutations in the gene LRRK2 encoding dardarin (PARK8) cause familial Parkinson's disease: Clinical, pathological, olfactory and functional imaging and genetic data. Brain, 128, 2786–2796.

15.

Klein

, and Westenberger

. (2012). Genetics of Parkinson's disease. Cold Spring Harbor Persp Med, 2, a008888.

16.

Lesage

, and Brice

. (2009). Parkinson's disease: From monogenic forms to genetic susceptibility factors. Human Mol Genetics, 18, R48–R59.

17.

, Tan

, and Yu

. (2014). The role of the LRRK2 gene in Parkinsonism. Mol Neurodegener, 9, 47.

18.

Livak

, and Schmittgen

. (2001). Analysis of relative gene expression data using real-time quantitative PCR and the 2−ΔΔCT method. Methods, 25, 402–408.

19.

Liu

, Wang

, Yu

, Li

, Wang

, Jiang

, and Ross

. (2008). A Drosophila model for LRRK2-linked parkinsonism. Proc Natl Acad Sci, 105, 2693–2698.

20.

, Wei

, Wu

, Hu

, Liu

, and Yuan

. (2013). Advances with microRNAs in Parkinson's disease research. Drug Design Develop Therapy, 7, 1103.

21.

Martínez-Martín

, Rodríguez-Blázquez

, Alvarez

, Arakaki

, Arillo

, Chaná

, and Serrano-Duenas

. (2015). Parkinson's disease severity levels and MDS-Unified Parkinson's Disease Rating Scale. Parkinsonism Relat Dis, 21, 50–54.

22.

Mouradian

. (2012). MicroRNAs in Parkinson's disease. Neurobiol Dis, 46, 279–284.

23.

Ozelius

, Senthil

, Saunders-Pullman

, et al. (2006). LRRK2 G2019S as a cause of Parkinson's disease in Ashkenazi Jews. N Engl J Med, 354, 424–425.

24.

Plowey

, Johnson

, Steer

, Zhu

, Eisenberg

, Valentino

, and Chu

. (2014). Mutant LRRK2 enhances glutamatergic synapse activity and evokes excitotoxic dendrite degeneration. Biochim Biophys Acta Mol Basis Dis, 1842, 1596–1603.

25.

Pringsheim

, Jette

, Frolkis

, and Steeves

. (2014). The prevalence of Parkinson's disease: A systematic review and meta‐analysis. Movement Dis, 29, 1583–1590.

26.

Saha

, Guillily

, Ferree

, Lanceta

, Chan

, Ghosh

, and Hisamoto

. (2009). LRRK2 modulates vulnerability to mitochondrial dysfunction in Caenorhabditis elegans. J Neurosci, 29, 9210–9218.

27.

Serafin

, Foco

, Zanigni

, Blankenburg

, Picard

, Zanon

, and Pramstaller

. (2015). Overexpression of blood microRNAs 103a, 30b, and 29a in l-dopa–treated patients with PD. Neurology, 84, 645–653.

28.

Thomson

, Bracken

, and Goodall

. (2011). Experimental strategies for microRNA target identification. Nucleic Acids Res, 39, 6845–6853.

29.

Van Den Eeden

, Tanner

, Bernstein

, et al. (2003). Incidence of Parkinson's disease: Variation by age, gender, and race/ethnicity. Am J Epidemiol, 157, 1015–1022.

30.

Vlachos

, Kostoulas

, Vergoulis

, Georgakilas

, Reczko

, Maragkakis

, and Hatzigeorgiou

. (2012). DIANA miRPath v. 2.0: Investigating the combinatorial effect of microRNAs in pathways. Nucleic Acids Res, 40, W498–W504.

31.

Wang

. (2008). miRDB: A microRNA target prediction and functional annotation database with a wiki interface. RNA, 14, 1012–1017.

32.

Wright Willis

, Evanoff

, Lian

, Criswell

, and Racette

. (2010). Geographic and ethnic variation in Parkinson disease: A population-based study of US Medicare beneficiaries. Neuroepidemiology, 34, 143–151.

33.

Zotos

, Roubelakis

, Anagnou

, and Kossida

. (2012). Overview of microRNA target analysis tools. Curr Bioinformat, 7, 310–323.