MicroRNAs in Parkinson's disease and emerging therapeutic targets

Abstract

Parkinson's disease (PD) is the second most common age-related neurodegenerative disorder, with the clinical main symptoms caused by a loss of dopaminergic neurons in the substantia nigra, corpus striatum and brain cortex. Over 90% of patients with PD have sporadic PD and occur in people with no known family history of the disorder. Currently there is no cure for PD. Treatment with medications to increase dopamine relieves the symptoms but does not slow down or reverse the damage to neurons in the brain. Increasing evidence points to inflammation as a chief mediator of PD with inflammatory response mechanisms, involving microglia and leukocytes, activated following loss of dopaminergic neurons. Oxidative stress is also recognized as one of the main causes of PD, and excessive reactive oxygen species (ROS) and reactive nitrogen species can lead to dopaminergic neuron vulnerability and eventual death. MicroRNAs control a range of physiological and pathological functions, and may serve as potential targets for intervention against PD to mitigate damage to the brain. Several studies have demonstrated that microRNAs can regulate oxidative stress and prevent ROS-mediated damage to dopaminergic neurons, suggesting that specific microRNAs may be putative targets for novel therapeutic strategies in PD. Recent human and animal studies have identified a large number of dysregulated microRNAs in PD brain tissue samples, many of which were downregulated. The dysregulated microRNAs affect downstream targets such as SNCA, PARK2, LRRK2, TNFSF13B, LTA, SLC5A3, PSMB2, GSR, GBA, LAMP-2A, HSC. Apart from one study, none of the studies reviewed had used agomirs or antagomirs to reverse the levels of downregulated or upregulated microRNAs, respectively, in mouse models of PD or with isolated human or mouse dopaminergic cells. Further large-scale studies of brain tissue samples collected with short postmortem interval from human PD patients are warranted to provide more information on the microRNA profiles in different brain regions and to test for gender differences.

Introduction

Parkinson's disease (PD) is the second most common age-related neurodegenerative disorder, affecting an estimated 7–10 million people worldwide (Valente et al., 2012). While less than 0.1% are affected among persons < 60 years of age, prevalence increases to 1–2% in those aged > 60 years (Shulman et al., 2011) and 2–3% in those aged > 80 years (de Lau and Breteler, 2006). The clinical main symptoms are caused by a loss of dopaminergic neurons in the substantia nigra, corpus striatum and brain cortex (Braak et al., 2004; Shulman et al., 2011). Patients exhibit a range of clinical symptoms, with the most common affecting motor function including resting tremor, rigidity, akinesia, bradykinesia and postural instability (Winklhofer and Haass, 2010). Non-motor symptoms are often an integral part of the disease and some of them, such as depression, anxiety and hyposmia, can precede the onset of Parkinsonism (Ceravolo et al., 2010). Over 90% of patients with PD have sporadic PD, also known as idiopathic PD (Thomas and Beal, 2007; Valente et al., 2012), and occur in people with no known family history of the disorder. Widespread aggregates of α-synuclein protein in the substantia nigra, together with the presence of cytoplasmic α-synuclein aggregates called Lewy bodies and α-synuclein filaments called Lewy neurites in degenerating neurons, are a pathological hallmark of sporadic PD (Winklhofer and Haass, 2010). Elevated levels of α-synuclein mRNA in substantia nigra dopamine neurons have been observed in sporadic PD (Shulman et al., 2011). Although the causes of these cases remain unclear, sporadic PD likely results from a complex interaction of environmental/acquired and genetic/inherited factors (Nuytemans et al., 2010). A small proportion of cases can be attributed to genetic factors with an autosomal or recessive pattern of inheritance and are sometimes referred to as familial Parkinson's disease. Mutations in SNCA, PARKIN, UCHL-1, PINK1, DJ-1 and LRRK2 are the origin of familial cases of Parkinson's disease, although they account for only 5–10% of patients.

MicroRNAs are abundant, endogenous, short, noncoding RNAs that act as important post-transcriptional regulators of gene expression by binding to the 3'-untranslated region (UTR) of their target mRNAs, thereby interfering with translation or causing destabilization or preferential cleavage of target RNAs (Baek et al., 2008; Ha and Kim, 2014). During the last decade, substantial knowledge has accumulated regarding the biogenesis of microRNAs, their molecular mechanisms and functional roles in a variety of cellular contexts. Altered expression of certain microRNA molecules suggests that they could have a crucial regulatory role in disorders. Increasing evidence points to inflammation as a chief mediator of PD with inflammatory response mechanisms, involving microglia and leukocytes, activated following loss of dopaminergic neurons (Rocha et al., 2015). The free radical nitric oxide (NO) plays a key role in the pathogenesis of inflammation. Under normal physiological conditions, NO has an anti-inflammatory effect, but is considered a pro-inflammatory mediator due to overproduction in abnormal situations (Sharma et al., 2007). The NO synthases (NOS) family synthesizes NO in a two-step reaction involving oxygen and many cofactors. Among the NOS isoforms (neuronal, endothelial, and inducible: nNOS, eNOs, iNOS, respectively), the nNOS is the most implicated in a wide range of functions and pathologies in the CNS. In the CNS, nNOS is located inside the postsynaptic membrane and is physically bound to N-methyl-D-aspartate (NMDA)-type glutamate receptors. Under physiological conditions, mild activation of synaptic NMDARs allows influx of Ca2+, which leads to nNOS catalytic activation (Maccallini and Amoroso, 2016). By contrast, hyperactivation of extrasynaptic NMDARs can lead to an abnormal Ca2+ influx into the postsynaptic neuron, with a subsequent overstimulation of nNOS and excessive NO production. This leads to generation of reactive oxygen and nitrogen species that cause DNA and lipid damage (Heinrich et al., 2013; Maccallini et al., 2016). Consequently, neurotransmission is impaired due to mitochondrial dysfunction and synaptic damage. NO also induces apoptosis (Cao et al., 2005). Several microRNAs (miR-939, miR-26a) have been identified to bind with the human iNOS 3'-UTR and exert a translational blockade of human iNOS synthesis (Guo and Geller, 2014). Also, overexpression of microRNA-155 decreased, whereas inhibition of microRNA-155 increased, eNOS expression and NO production in human umbilical vein endothelial cells (Sun et al., 2012).

Oxidative stress is recognized as one of the main causes of PD, and excessive reactive oxygen species (ROS) can lead to dopaminergic neuron vulnerability and eventual death. Several studies have demonstrated that microRNAs can regulate oxidative stress in in vitro and in vivo animal models of PD. Relevant microRNAs involved in regulating oxidative stress can prevent ROS-mediated damage to dopaminergic neurons, suggesting that specific microRNAs may be putative targets for novel therapeutic strategies in PD (Xie and Chen, 2016). Impairment of mitochondrial function resulting in cellular damage is also linked to aging and neurodegeneration and evidence suggests it plays a central role in the pathogenesis of PD (Winklhofer and Haass, 2010). Glutamatergic transmission and inflammatory response mechanisms are altered in striatal neurons following dopaminergic denervation (Gardoni and Bellone, 2015; Kim et al., 2015). Despite extensive research, the molecular mechanisms mediating the changes in striatal neurons following dopaminergic denervation are still unclear. Understanding the mechanisms underlying this process is important for gaining new insights into the pathogenesis of PD. A recent study suggests that the age-related decline of Dicer enzyme combined with increased cellular stress in dopaminergic neurons may compromise microRNA biosynthesis thus contributing to neurodegeneration in PD (Chmielarz et al., 2017).

Circulating microRNAs have been proposed as diagnostic biomarkers for PD and would enable detection at the earliest stages of the disease for therapy to be implemented to delay the onset or minimize the changes in the later stages of the disease. However, other organs may contribute to microRNAs in the blood so that the circulating levels may not accurately reflect the levels of specific microRNAs in the diseased brain itself (Sierzega et al., 2017). We have searched the PubMed database for studies on microRNA expression in brain tissue of patients with PD and animal models of PD and their involvement in the pathophysiology of the disease, and which might serve as therapeutic targets using microRNA mimics or antagomirs. The studies retrieved in the literature search covered the period 2007–2017.

Neuropathology/Braak Staging and Animal Models of PD

The diagnosis of PD is still largely made on clinical grounds by four cardinal signs (tremor, bradykinesia, rigidity, and postural instability) as there is no definitive laboratory test to confirm the diagnosis during life, apart from gene testing in a reduced number of cases. Non-motor symptoms may predate diagnosis by several years and a schematic has been proposed depicting normal aging and PD-related nigral cell loss over time including the time at which diagnosis typically occurs (Noyce et al., 2016). Pre-symptomatic markers of PD may include olfactory loss, depression, rapid eye movement (REM) sleep disorder, and constipation (Schapira et al., 2017). Most reviews of PD indicate that motor signs first appear when approximately 50% of substantia nigra dopaminergic neurons are lost (Marsden, 1990; Ross et al., 2004). A regression analysis of neuron counts versus duration of PD indicated that the number of neurons lost at the time of symptom onset was 31%, adjusted for age (Fearnley and Lees, 1991). At the time of 1 year post diagnosis, patients with PD may retain up to 90% of their substantia nigra dopamine neurons and 50% of their striatal dopaminergic innervation (Kordower et al., 2013).

Based on autopsy findings in patients with PD, Braak et al. (2003) reported that the intraneuronal formation of Lewy bodies and Lewy neurites has a topographically predictable progression. Accordingly Braak staging was created based on the presence of Lewy bodies and Lewy neurites. The pre-symptomatic phase usually falls within Stages 1, 2 and 3, while the symptomatic phase falls into Stages 3, 4, 5 and 6. Stage 1 (medulla oblongata): lesions initially occur in the dorsal glossopharyngeal/vagal motor nucleus and frequently in the anterior olfactory nucleus. There may also be involvement of intermediate reticular zone. Stage 2 (medulla oblongata and pontine tegmentum): this includes the pathology of stage 1 together with lesions in the caudal raphe nuclei, gigantocellular reticular nucleus, and coeruleus-subcoeruleus complex. Stage 3 (midbrain): pathology of stage 2 plus midbrain lesions, particularly in the pars compacta of the substantia nigra. Stage 4 (basal prosencephalon and mesocortex): pathology of stage 3 with lesion at prosencephalon; cortical involvement is confined to the temporal mesocortex (transentorhinal region) and allcortex (CA2-plexus)-the neocortex is unaffected. Stage 5 (neocortex): pathology of stage 4 plus lesions in higher order sensory association areas of the neocortex and prefrontal neocortex. Stage 6 (neocortex): pathology of stage 5 plus lesions in first order sensory association areas of the neocortex and premotor areas, occasionally mild changes in primary sensory areas and the primary motor field.

Neurotoxic and genetic animal models have been used to produce PD-related pathology and symptomatology (Blesa et al., 2012; Jackson-Lewis et al., 2012). Neurotoxin-based models produced by 6-hydroxydopamine (6-OHDA) and 1-methyl-1,2,3,6-tetrahydropyridine (MPTP) administration are the most widely used toxic models. Mice, rats, cats, dogs, and monkeys are all sensitive to 6-OHDA. Although similar in structure to dopamine, the presence of an additional hydroxyl group makes it toxic to dopaminergic neurons. This compound does not cross the blood-brain barrier (BBB), which necessitates its direct injection into the substantia nigra, medial forebrain bundle, or striatum. The most common use of 6-OHDA is via unilateral injection into the medial forebrain bundle or striatum. Injection of 6-OHDA into the substantia nigra kills approximately 60% of the tyrosine hydrolase (TH)-containing neurons in this area of the rodent brain with subsequent loss of TH-positive terminals in the striatum. The extent of the lesion depends on the amount of 6-OHDA injected, the site of injection, and the species used. This model does not mimic all the clinical features of PD. Dopamine depletion, nigral dopamine cell loss, and neurobehavioral deficits have been successfully achieved using this model, but it does not seem to affect other brain regions such as olfactory structures, lower brain stem areas, or locus coeruleus. Although 6-OHDA does not produce or induce proteinaceous aggregates or Lewy-like inclusions like those seen in PD, it has been reported that 6-OHDA does interact with α-synuclein (Blandini et al., 2008). 6-OHDA is frequently used as a unilateral injection because bilateral injection of this compound into the striatum produces severe adipsia, aphagia, and death (Ungerstedt, 1971).

MPTP represents the most important and most frequently used parkinsonian toxin applied in animal studies. It was shown to replicate almost all the hallmarks of PD including oxidative stress, ROS, energy failure, and inflammation. MPTP is highly lipophilic and rapidly crosses the BBB after systemic administration. Upon entering the brain, MPTP enters astrocytes and is metabolized into 1-methyl-4-phenylpyridinium (MPP+), its active metabolite which is a positively charged molecule, by monoamine oxidase-B. Once released from the astrocytes into the extracellular space via the OCT-3 transporter, MPP+ is taken up into the neuron by the dopamine transporter (DAT) and can be stored in vesicles. Inside the neuron, MPP+ is able to inhibit complex 1 of the mitochondrial electron transport chain, resulting in the release of ROS as well as decreased ATP production. MPP+ stored in vesicles is thought to expel dopamine into the extracellular space where it can be metabolized and subjected to superoxide and hydroxyl radical attack. MPTP is used mainly in nonhuman primates and mice, but has also been used in many other species such as dogs and cats. The MPTP mouse model is employed to study pathological effects of PD, while the MPTP monkey model is used mainly to study behavioral and symptomatic components of PD. The data generated by mouse models have led to a better understanding of molecular mechanisms involved in PD. .

In recent years a new generation of animal models of PD based on ectopic expression, overexpression, or intracerebral injection of α-synuclein have emerged (Visanji et al., 2016). Viral vector-mediated α-synuclein overexpression has been employed in rodents and nonhuman primates. Adeno-associated virus (AAV) vectors demonstrate high, maintained delivery of α-synuclein, with Lewy-like pathology, overt dopaminergic degeneration, and a parkinsonian behavioral phenotype in rodents (Koprich et al., 2010, 2011). These models develop inclusions of aggregated α-synuclein and/or α-synuclein-mediated neuronal cell loss replicating pathological hallmarks of PD and contributing to advances in the understanding of pathogenic mechanisms underpinning PD. Ip et al. (2017) showed that human mutated AAV1/2-A53T α-synuclein injected wild type-mice had widespread nigral and striatal expression of vector-delivered A53T α-synuclein. At 10 weeks, in AAV1/2-A53T α-synuclein mice there was a 33% reduction in TH+ dopaminergic nigral neurons, 29% deficit in striatal DAT binding, and 38% reduction in dopamine level. The mouse model has certain advantages, especially it being amenable to genetic manipulation. Transgenic mice expressing α-synuclein have been generated to try to model PD to study α-synuclein pathobiology and investigate novel therapeutics (Magen and Chesselet, 2010; Koprich et al., 2017).

Accumulating evidence indicates that L-type calcium channels are involved in brain diseases such as PD (Ortner and Striessnig, 2016) and contribute to basal metabolic stress in substantia nigra dopaminergic neurons (Sulzer and Surmeier, 2013). Cav1.2 and Cav1.3 L-type calcium channels are expressed in the substantia nigra neurons (Ortner and Striessnig, 2016). They contribute to somatodendritic Ca2+ oscillations during autonomous pacemaking or bursting in these cells (Guzman et al., 2009). It is considered that this constant Ca2+ load contributes to the vulnerability of substantia nigra neurons to degeneration in PD by enhancing mitochondrial oxidative stress (Guzman et al., 2010). Epidemiological studies show that dihydropyridines, which are antagonists of these channels, reduce the observed risk of PD (Ritz et al., 2010). This finding is surprising given the relatively low affinity of dihydropyridines for the subtype of L-type calcium channel responsible for most of the calcium entry in striatonigral dopaminergic neurons, which is one with a Cav1.3 pore-forming subunit (Sinnegger-Brauns et al., 2009; Surmeier et al., 2011). Mice in which the gene for the Cav1.3 pore-forming subunit was deleted (Cav 1.3 knockout mice) and with impaired voltage-gated Ca2+ channel activity have been used to study PD.

Human studies

Thirteen studies were found and mostly comprised both male and female patients with mean ages ranging from 68 to 80 years [Table 1]. Most of the studies indicated that the patients had sporadic PD. The three largest studies were performed on postmortem brain tissue samples from 22 patients (10 male/12 female, mean age 74 years; Tatura et al., 2016), 23 patients (5 male/2 female, Braak stages 1–3, mean age 68 years; 10 male/6 female, Braak stages 4–5, mean age 75 years; Miñones-Moyano et al., 2011) and 25 patients (4 male/2 female, Braak stages 1–2, mean age 79 years; 7 male/6 female, Braak stages 3–4, mean age 72 years; 4 male/2 female, Braak stage 5, mean age 80 years; Villar-Menendez et al., 2014). In the other studies, brain tissues samples were examined from PD patients varying in number from 3 to 20, and although the Braak stages had not been included in most, neuropathological diagnoses of PD had been made according to recognized criteria such as the presence of Lewy bodies and neuronal loss in the substantia nigra. The mean postmortem interval (PMI) for collecting brain tissue samples ranged from 5 to 46 hours.{Table 1}

MicroRNAs upregulated in PD anterior cingulate gyri samples analyzed by RT-PCR were miR-199b, -544a, -488, -221, -144 and those downregulated were miR-7, -145, -543 (Tatura et al., 2016). An analysis of PD putamen samples by NanoString nCounter microRNA assay revealed 6 microRNAs were upregulated, miR-3195, -204-5p, -485-3p, -221-3p, -95, 425-5p, and 7 were downregulated, miR-155-5p, -219-2-3p, -3200-3p, -423-5p, -4421, -421, -382-5p (Nair and Ge, 2016). Using RT-PCR, miR-34b was downregulated in putamen samples of PD patients at Braak stages 1–2 and stages 4–5 (Villar-Menendez et al., 2014). Microarray analysis of PD substantia nigra samples revealed that one microRNA was upregulated, miR-548-d, while 10 were downregulated, miR-198, -485-5p, -339-5p, -208b, -135b, -299-5p, -330-5p, -542-3p, -379, -337-5p. Expression levels of miR-198, -548d, -135b were validated with individual TaqMan assays (Cardo et al., 2014). Also in PD substantia nigra samples, 6 microRNAs were upregulated, miR-21*, -224, -373*, -26b, -106a, -301b, and with similar but milder changes in amygdala samples with a significant upregulation of miR-224 and miR-373* (Alvarez-Erviti et al., 2013). By microarray analysis, 4 microRNAs were upregulated, miR-200b*, -200a*, -195*, -424*, and -7 microRNAs were downregulated, miR-200a, -199a-3p, -148a, -451, -144, -429, -190, in PD frontal cortex samples (Thomas et al., 2012). A significant downregulation of miR-34b and miR-34c occurred in PD amygdala, substantia nigra, and frontal cortex samples (Miñones-Moyano et al., 2011). Interestingly, a significant downregulation was also found in both miR-34b and miR-34c in the amygdala, but not in the frontal cortex, of PD pre-motor cases (Miñones-Moyano et al., 2011). By RT-PCR, miR-133b was downregulated in PD midbrain samples (Kim et al., 2007).

Two of the studies had included PD patients with dementia (PDD) (Cho et al., 2013; Hoss et al., 2016). By RT-PCR, miR-205 was downregulated in PD frontal cortex, and there was no significant difference in the expression level between PD patients and PDD patients. Downregulation of miR-205 also occurred in the PD striatum (Cho et al., 2013). By microRNA sequence analysis, the levels of 64 microRNAs were downregulated whereas the levels of 61 microRNAs were upregulated in PD prefrontal cortex compared to controls, and a set of 29 microRNAs classified PD from control brain (93.9% specificity, 96.6% sensitivity). 36 microRNAs classified PDD from PD without dementia (PDN) (88.9% specificity, 81.2% sensitivity). For the majority of differentially expressed microRNAs in PD, PDD samples exhibited larger differences than PDN for the same microRNAs (Hoss et al., 2016). Among the downregulated microRNAs in PD brains were let-7i-3p/5p, miR-184, -1224, -127-5p, and among the upregulated microRNAs was miR-16-5p (Hoss et al., 2016).

In addition, three studies were made on dopaminergic neurons in PD brain tissue samples (Choi et al., 2014; Schlaudraff et al., 2014; Briggs et al., 2015). A widespread expression of miR-7 was shown in dopaminergic neurons in PD substantia nigra (Choi et al., 2014). Also there was no difference in miR-133 level in dopaminergic neurons in PD substantia nigra compared to controls (Schlaudraff et al., 2014). In dopaminergic neurons collected from PD brain tissue, 8 of the 12 upregulated microRNAs were also upregulated in males, miR-106a, -135a, -148a, -223, -26a, -28-5p, -335, -92a, while 3 were upregulated in females, let-7b, miR-106a, -95 (Briggs et al., 2015).

Animal studies

Twelve studies in mice were found and males had been used where the gender was specified. Of these studies, 8 had used the MPTP model of PD, 2 the 6-OHDA model, 1 the α-synuclein overexpression model, and 1 the L-type calcium channel Cav1.3 knockout mouse model. The ages of the mice ranged from 6 weeks to 6 months [Table 2].{Table 2}

MPTP model studies

Mice received injection(s) of MPTP 20 mg/kg or 30 mg/kg over a period ranging from 1 to 21 days (for mice aged 6–12 weeks) or 5 weeks (for mice aged 4–5 months), and sacrificed at chosen time points after the last MPTP injection (if several injections were given). Many of the studies identified downregulation of specific microRNAs in brain tissue collected from these animals. Expression of miR-7 in the midbrain of MPTP-treated mice was downregulated compared to controls (Junn et al., 2009; Zhou et al., 2016). Similarly, following MPTP administration, miR-124 was downregulated in the ventral midbrain (Wang et al., 2016) and substantia nigra (Kanagaraj et al., 2014). Also, downregulation of miR-135a-5p was found in the brain tissue of mice after administering MPTP (Liu et al., 2016). By contrast, Su et al. (Su et al., 2016) found that miR-21 was upregulated in midbrain of MPTP-treated mice compared to controls. Xiong et al. (Xiong et al., 2014) showed that miR-494 was expressed highly in the substantia nigra of MPTP-treated mice and that overexpression of miR-494, induced by injecting lentivirus containing miR-494 into the substantia nigra, exacerbated MPTP-induced neurodegeneration, with a loss of dopaminergic neurons.

Expression of miR-7116-5p was downregulated while that of miR-125-5p was upregulated in microglia from the ventral midbrain of MPTP-treated mice (He et al., 2017). Also, miR-124 expression was downregulated in substantia nigra dopaminergic neurons following MPTP administration (Wang et al., 2016).

6-OHDA model studies

Mice received a unilateral injection of 6-OHDA 3.75 µg in the medial forebrain bundle (Rivetti di Val Cervo et al., 2017) or 10 µg in the striatum (Saraiva et al., 2016). Treatment of 6-OHDA-treated mice with three transcription factors, NEUROD1, ASCL1 and LMX1A, and the microRNA miR-218, collectively designated NeAL218, reprogrammed astrocytes in vivo into induced dopamine neurons (Rivetti di Val Cervo et al., 2017). MiR-124 nanoparticles (NPs) in mice, receiving a double injection to deliver 6-OHDA into the right striatum and miR-124 NPs into the right lateral ventricle, promoted an increase in migrating neuroblasts and enhanced brain repair (Saraiva et al., 2016).

α-Synuclein overexpression model study

Expression of miR-155 in the substantia nigra was upregulated in mice receiving a unilateral injection of AAV containing α-synuclein into the substantia nigra (Thome et al., 2016).

L-type calcium channel Cav1.3 knockout mouse model study

Expression of miR-204-5p and miR-143-3p was upregulated in the hippocampus of Cav 1.3–/– knockout mice compared to controls (Gstir et al., 2014).

MicroRNAs as Therapeutic Targets for PD

Dysregulated microRNAs in PD brain tissue samples

The human studies have identified a large number of microRNAs whose levels were dysregulated in PD brain tissue samples [Table 1]. Included among those having downregulated expression were: miR-7, -145, -543 (cingulate gyri); miR-155-5p, -219-2-3p, -3200-3p, -423-5p, -4421, -421, -382-5p, -34b (putamen); miR-198, -485-5p, -339-5p, -208b, -135b, -299-5p, -330-5p, -542-3p, -379, -337-5p, -34b, -34c (substantia nigra); miR- 200a, -199a-3p, -148a, -451, -144, -429, -190, -34b, -34c -205 (frontal cortex); miR-34b, -34c (amygdala); miR-133b (midbrain); miR-205 (striatum); let-7i-3p/5p, miR-184, -1224, -127-5p (prefrontal cortex). Those with upregulated expression were: miR-199b, -544a, -488, -221, -144 (cingulate gyri); miR-3195, -204-5p, -485-3p, -221-3p, -95, 425-5p (putamen); miR-548-d, -21*, -224, -373*, -26b, -106a, -301b (substantia nigra); miR-224, -373* (amygdala); miR-200b*, -200a*, -195*, 424* (frontal cortex); miR-16-5p (prefrontal cortex). It would seem that different regions of the PD brain exhibit differently altered microRNA profiles [Figure 1], and this may reflect differences in the numbers and functional states of specific cell types present.{Figure 1}

Several dysregulated microRNAs may be potential therapeutic targets, but none of the studies had examined the effect of modifying the expression levels of chosen microRNAs in the PD brain. Choi et al. (2014) had suggested that the overexpression of miR-205 may provide an applicable therapeutic strategy to suppress the abnormal upregulation of LRRK2 protein in PD brains. Overexpression could be achieved using a miR-205 agomir or mimic. Also miR-7 protected cells from MPP(+)-induced toxicity in human dopaminergic SH-SY5Y cells (Choi et al., 2014) and the use of a miR-7 agomir may improve PD brain pathophysiology. In vitro testing using agomirs to downregulated microRNAs or antagomirs to upregulated microRNAs may provide a way of identifying possible microRNAs targets to protect dopaminergic cells exposed to neurotoxins (MPP+ or 6-OHDA).

The animal studies also identified several microRNAs whose levels were dysregulated in brain tissues of PD models [Table 2]. Included among microRNAs with downregulated expression were: miR-7 (midbrain, ventral midbrain); miR-124 (midbrain, substantia nigra); miR-135a-5p. Those with upregulated expression were: miR-21 (midbrain); miR-494, -155 (substantia nigra); miR-204-5p, -143-5p (hippocampus). He et al. (2017) found that the level of miR-7116-5p was downregulated while that of miR-125b-5p was upregulated in microglia from the ventral midbrain of MPTP mice. In MPTP model, overexpression of miR-494 by injection of lenti-494 into the substantia nigra exacerbated MPTP-induced neurodegeneration. Reprogramming of striatal astrocytes into induced dopaminergic neurons occurred on injecting lenti-NeAl218 into the dorsal striatum of 6-OHDA model (Rivetti di Val Cervo et al., 2017). Injection of miR-124 nanoparticles into the lateral ventricle of 6-OHDA model enhanced brain repair (Saraiva et al., 2016). None of the studies reviewed had used agomirs or antagomirs to reverse the levels of downregulated or upregulated microRNAs, respectively, in in vivo mouse models of PD or in in vitro with isolated dopaminergic cells.

Downstream targets of dysregulated microRNAs in brain tissue samples

Downstream targets of several important microRNAs have been indicated in the PD studies reviewed. For example, downregulated TNFSF13B (TNF superfamily member 13b) is a predicted target of upregulated miR-425-5p and LTA (lymphotoxin alpha) and SLC5A3 (soluble carrier family 5 sodium/myo-inositol cotransporter member 3) are predicted targets of upregulated miR-485-3p. The upregulated PSMB2 (proteasome subunit, beta type 2) and GSR (glutathione reductase) are predicted targets of downregulated miR-423-5p and miR-219-3p, respectively (Nair and Ge, 2016). LRRK2 (leucine-rich repeat kinase 2) is an experimentally validated target of downregulated miR-1224, and GBA (glucocerebrosidase) is a target of downregulated miR-127-5p and upregulated miR-16-5p (Hoss et al., 2016). Downregulated LAMP-2A (lysosome-associated membrane protein 2) and HSC (hematopoietic stem cell) are the predicted targets of upregulated miR-26b, -106a, -301b (Alvarez-Erviti et al., 2013). Also the 3'-UTR of transcription factor Ptx3 was identified as a potential target of downregulated miR-133b (Kim et al., 2007). Upregulated miR-21 in PD directly targeted the 3'-UTR of LAMP2A (lysosome-associated membrane protein 2A) (Su et al., 2016), while downregulated miR-135a-5p targeted the 3'-UTR of ROCK2 (rho-associated protein kinase 2) (Liu et al., 2016). Upregulated miR-204-5p and miR-143-3p were predicted to target the 3'-UTR of several ion channel mRNAs (Gstir et al., 2014). Also, the 3'-UTR of adenosine A2A receptor (A2AR) is a predicted target for downregulated miR-34b in PD (Villar-Menendez et al., 2014). Interestingly, dysregulated microRNA and target gene network related to PD may be gender-specific (Briggs et al., 2015).

Possible biological implications of some important dysregulated microRNAs in brain tissue samples

A single microRNA can regulate the expression of hundreds of target genes, so alterations in a panel of microRNAs could greatly affect the pathophysiology and outcome of PD. Use of in vivo animal models of PD together with in vitro studies using dopaminergic cells, neural progenitor cells or primary neurons exposed to MPP+ provide a means of testing specific microRNAs for protective or adverse effects. For instance, miR-7 was shown to protect dopaminergic neurons from MPP(+)-induced toxicity (Choi et al., 2014), and suppress NLRP3 inflammasone-mediated neuroinflammation in MPTP mouse model (Zhou et al., 2016). Upregulation of miR-205 may provide an applicable therapeutic strategy to suppress the abnormal upregulation of LRRK2 protein in PD brains (Cho et al., 2013). This could be achieved using lenti- or adeno-associated virus containing miR-205 or by a miR-205 agomir. It was proposed that early downregulation of miR-34b/c in PD triggers downstream transcriptome alterations underlying mitochondrial dysfunction and oxidative stress, which ultimately compromise cell viability (Miñones-Moyano et al., 2011). Upregulation of miR-34b/c may be an applicable therapeutic strategy. Also downregulation of miR-7116-5p in microglia initially potentiated the production of TNF-α in MPTP mouse model, which then amplified the production of other proinflammatory factors to contribute to dopaminergic neuron damage (He et al., 2017). Upregulation of miR-7116-5p could be tested in this model to see if it ameliorated neuron damage. The upregulation of pro-apoptotic Bim mRNA level and protein level induced by MPTP was reduced by miR-124 agomir (Wang et al., 2016). Downregulation of miR-135a-5p in MPTP mouse model was associated with an upregulation in ROCK2 that leads to phagocytosis of dopaminergic neurons (Liu et al., 2016). A miR-135a-5p agomir could possibly protect dopaminergic neurons from phagocytosis. Furthermore, the decreased level of miR-7 (Zhou et al., 2016) possibly contributes to increased α-synuclein accumulation and inflammatory response in MPTP mouse model (Junn et al., 2009). In the animal models of PD, downregulated miR-7 and miR-124 (Junn et al., 2009; Kanagaraj et al., 2014; Wang et al., 2016; Zhou et al., 2016) together with upregulated miR-21 and miR-494 (Xiong et al., 2014; Su et al., 2016) are involved in oxidative stress (Xie and Chen, 2016). In PD patients, downregulated miR-7, -34b/c and -205 (Miñones-Moyano et al., 2011; Cho et al., 2013; Villar-Menendez et al., 2014; Tatura et al., 2016) together with upregulated miR-224 (Alvarez-Erviti et al., 2013) are associated with oxidative stress (Xie and Chen, 2016).

Future Perspectives

Currently the diagnosis of patients with PD is mainly made on clinical manifestations of the disease. Molecular imaging allows a window into the pathophysiology of PD, as well as measuring the severity and progression of the disease. The dopamine terminal dysfunction can be demonstrated using positron emission tomography (PET) or single photon emission computed tomography (SPECT) with different tracers, which contribute to early and accurate diagnosis leading to appropriate medications. PET/SPECT imaging, combined with other individual information such as genetic testing, would assist in providing personalized treatment to improve clinical outcomes and minimize adverse effects (Bu et al., 2016). Neuroinflammation ligand imaging the microglial activation might guide the individualized application of non-steroidal anti-inflammatory therapy in PD patients in the future (Stoessl et al., 2011).

Further large-scale studies of brain tissue samples collected with short PMI from human PD patients are warranted to confirm the changes in microRNA expression that have been reported and to test for gender differences. Where gender was specified, all of the animal studies had used adult male mice at 6 weeks to 6 months of age. Future studies should be performed with aged animals 22 to 24 months of age. Also both male and female animals should be used, as dysregulated microRNA and target gene network related to PD may be gender-specific (Briggs et al., 2015). It has been reported that 50% to 80% of patients with PD have abnormal glucose tolerance that may be further exacerbated by L-dopa therapy (Sandyk, 1993). An observational study concluded that diabetes prevalence was closely similar between patients with PD and subjects without the disease (Becker et al., 2008). In the human studies, it is likely that many of the PD patients would have been taking medication. Animal models of PD should also incorporate possible medications that could have been used such as L-dopa, antidiabetic, antihypertensive, and antihyperlipidemic drugs. A recent clinical trial has shown that exenatide, a medication used for patients with diabetes mellitus type 2, has the potential to modify PD. Patients with sporadic PD aged 25 to 75 years received subcutaneous injections of exenatide 2 mg or placebo once weekly for 48 weeks in addition to their regular medication, followed by a 12-week washout period. Movement disorder was assessed on a rating scale at 60 weeks. Exenatide had positive effects on motor scores in PD that were sustained beyond the period of exposure. Whether exenatide affected the underlying disease pathology is uncertain (Athauda et al., 2017).

Conclusion

This review has shown the expression of a large number of microRNAs to be altered in brain tissue samples of human PD patients and experimental animal models of PD. Some of these altered microRNAs could serve as potential therapeutic targets since modifying the levels of specific microRNAs was found to have beneficial effects in animal models of PD, with improved functional outcomes. For example, miR-124 agomir delivered to the right lateral ventricle in MPTP mouse model increased the density of TH+ neurons and reduced the upregulation of Bim mRNA level and protein level induced by MPTP, leading to reduced apoptosis (Wang et al., 2016). The predicted downstream targets of many of the dysregulated microRNAs have also been identified, and these included LRRK2, LAMP2A, ROCK2, and several ion channel mRNAs. Inflammation and oxidative stress are considered to be chief mediators of PD, with NO playing a key role in the pathogenesis of this neurological disorder. The biological actions of some of the important altered microRNAs may be in regard to these mechanisms[83].

Author contributions: Bath authors contributed to the study equally.

Conflicts of interest: None declared.

Plagiarism check: Checked twice by iThenticate.

Peer review: Externally peer reviewed.

Open peer reviewer: Cristina Maccallini, University G. d'Annunzio, Italy.