Parkinson’s Disease and MicroRNAs - Lessons from Model Organisms and Human Studies

Abstract

Parkinson’s disease (PD) is a progressive, age-associated neurodegenerative disorder that affects an estimated 10 million people worldwide. PD is characterized by proteinaceous, cytoplasmic inclusions containing α-synuclein, called Lewy Bodies, which form in dopaminergic neurons in an age-dependent manner, and are associated with the emergence of characteristic PD symptoms such as resting tremor, rigidity, slow movements and postural instability. Although considerable progress has been made in recent years in identifying genetic and environmental factors that are associated with PD, early diagnosis and therapeutic options remain severely lacking. Recently, microRNAs (miRNAs) have emerged as novel therapeutic targets in various diseases, such as cancer and neurodegenerative diseases. MiRNAs have been shown to play roles in various aging and neurodegenerative disease models across phyla. More recently, studies have identified specific roles for miRNAs and their targets in the pathogenesis and progression of PD in several model organisms. Here, we discuss the evolving field of miRNAs, their association with PD, and the outlook for the future.

Introduction

Parkinson’s Disease overview

Parkinson’s disease (PD) is an age-related neurodegenerative disorder with an estimated prevalence in the industrialized world of roughly 1% of the population aged 60 and above (Lau and Breteler 2006). Recent meta-analysis has indicated that this prevalence rate nearly doubles in the population aged 80 and above (Pringsheim et al. 2014). This makes PD the second most common type of dementia among the aged population. This fact, together with an increasingly aging-population in most of the Western world, establishes PD as a key challenge in current and near-future biomedical health. PD is categorized in two main disease classes, familial and sporadic PD, which have similar symptoms but are distinguished by their cause. Familial PD is associated with mutations in genetic loci that can be traced in families, whereas sporadic PD is largely characterized as idiopathic. PD is characterized by hallmark symptoms including resting tremor, rigidity, slowness or absence of voluntary movement, postural instability, and freezing (Olanow and Tatton 1999; Dauer and Przedborski 2003; Fahn 2003). The symptoms of PD begin slowly and gradually worsen over the course of the disease (Olanow and Tatton 1999; Dauer and Przedborski 2003; Fahn 2003). In addition to these prototypical motor symptoms, patients often develop non-motor manifestations of the disease as well, including cognitive impairment, depression, fatigue, sensory impairment, sleep disorders, weight loss, and behavioral problems (Fahn 2003; Jankovic 2008). PD diagnosis can often be difficult to establish conclusively due to the fact that brain autopsy remains as the most well-established and conclusive PD diagnostic test (Jankovic 2008). Therefore, careful understanding of the hallmark features and manifestations of the disease are needed to differentiate true PD from other related disorders.

Although familial PD is less common (25% prevalence) than sporadic PD, several genetic loci have been associated with the disease. Some specific PD associated loci that have been identified as monogenetic causes of PD include the following five genes: α-synuclein (SNCA), LRRK2, PINK-1, PARKIN, and DJ-1 (Klein and Westenberger 2012; Mullin and Schapira 2015). SNCA has been found to be duplicated or triplicated in families with autosomal-dominant familial PD (Mullin and Schapira 2015). Leucine-rich repeat kinase 2 (LRRK2) was first identified in 2002 in a Japanese family with autosomal dominant Parkinsonism, but the function of LRRK2 and its role in PD are still uncertain (Klein and Westenberger 2012; Mullin and Schapira 2015). The LRRK2 protein has been implicated in autophagy, mitophagy, vesicle trafficking, cytoskeleton maintenance, and neurite growth (Mullin and Schapira 2015; Singh et al. 2021). Due to these various possible functions and the large size of the LRRK2 protein, it is hypothesized that a gain of function mutation may lead to cell death via protein-protein interactions and activation of autophagy (Klein and Westenberger 2012; Mullin and Schapira 2015). PINK-1, PARKIN, and DJ-1 have been found to cause autosomal recessive forms of PD (Klein and Westenberger 2012; Mullin and Schapira 2015). All three mutations result in the inability of the neuronal cell to clear damaged mitochondria, which results in its accumulation inside the cell and eventual cell death (Klein and Westenberger 2012; Mullin and Schapira 2015).

α-Synuclein and the Etiology of PD

The pathogenesis of PD, and other neurodegenerative diseases such as Huntington’s disease and Alzheimer’s disease, is associated with the aggregation of proteotoxic granules that accumulate inside neuronal cells and appear to lead to damage and death of these cells (Förstl and Kurz 1999). The accumulation of α-synuclein inside dopaminergic neurons is a hallmark of PD and results in the impairment of detoxification and removal of aggregated proteins, leading to cell death (Dauer and Przedborski 2003; Dawson and Dawson 2003; Galvin et al. 1999; Schapira and Jenner 2011). These intraneuronal proteinaceous cytoplasmic inclusions, containing mostly α-synuclein, are called Lewy Bodies and are a pathological hallmark of PD (Dauer and Przedborski 2003). The definitive cause of the protein aggregation is currently unknown, but many hypotheses have been reported in the literature including the dysfunction of protein degradation pathways, neuroinflammation, and mitochondrial damage (Watanabe et al. 2020; Troncoso-Escudero et al. 2018; Chen et al. 2019).

The protein α-synuclein is primarily associated with presynaptic terminals (Vargas et al. 2017) where it is hypothesized to regulate neurotransmission by promoting the assembly of the SNARE complex (Burré et al. 2010). In-vitro studies have shown that increasing the local concentration of α-synuclein promotes its fibrillization. In addition, the effect of increased concentration on fibrillization is enhanced in α-synuclein mutants such as SNCA^A53T (Conway et al. 1998). Several PD related mutations also increase the accumulation of α-synuclein within the cell, and further support these in-vitro findings. For example, several pathological mutations involve genomic multiplications of the SNCA gene that directly increase the levels of α-synuclein expression (Singleton et al. 2003). Other PD mutations of interest that promote α-synuclein accumulation are those that impair the related autophagosome/lysosome degradation machinery, such as point mutations found in Lrrk2 and GBA (Kluss et al. 2019; Mazzulli et al. 2011). The significance of these mutations is highlighted by studies in animal models of PD, which are detailed below. The differential expression profiles of α-synuclein between cell types and brain regions may therefore underlie why certain populations with higher endogenous levels of α-synuclein, such as dopamine neurons of the substantia nigra, are more vulnerable than others and why gene pathways that regulate α-synuclein expression, such as those controlled by microRNAs, may be important in the etiology of PD. This hypothesis of selective susceptibility has been previously suggested by several groups through analysis of the endogenous levels of α-synuclein in both regional and specific neuronal subtypes (Erskine et al. 2018; Taguchi and Wang 2018).

PD pathology is largely characterized by the dysfunction of autophagy and degradation pathways necessary for clearing α-synuclein aggregation (Arotcarena et al. 2019). Wild-type α-synuclein is recognized by heat shock cognate protein 70 (HSC70) and selectively cleared by the chaperone-mediated autophagy (CMA) pathway (Cuervo et al. 2004). Cuervo and colleagues found that α-syn mutants SNCA^A30P and SNCA^A53T can be recognized by CMA machinery, but these mutants are unable to be taken up into the lysosomal lumen for degradation. Overexpression of WT α-synuclein has also been shown to impair autophagy in cellular based models and mice through the inhibition of vesicular trafficking proteins such as Rab1a (Winslow et al. 2010). Mutations in the VSP35, a protein important in regulating autophagic turnover (Tsika et al. 2014) and CMA machinery (Tang et al. 2015), has been associated with increased susceptibility to PD. In addition, lysosomal enzymes such as GBA also appear to have reduced activity in PD patients, regardless of the presence of a GBA mutation (Parnetti et al. 2014). α-synuclein has also been implicated in the inhibition of the unfolded protein response (UPS), the main pathway for degrading misfolded proteins (Cook and Petrucelli 2009). Recent work by McKinnon and colleagues found that in mice, A53T α-synuclein impairs the catalytic activity of proteasome 26S, the major protease of the UPS pathway. In addition, this impairment leads to the selective accumulation of aggregation prone Ser129 phosphorylated α-synuclein (McKinnon et al. 2020). Altogether, PD pathology is characterized by dysfunction of proteostasis, autophagy, and lysosomal pathways needed to inhibit synucleinopathy.

Mitochondria dysfunction has also been heavily implicated in PD pathology (Chen et al. 2019). The presence of damaged mitochondria is a common observation in human post-mortem analysis of Lewy Body inclusions (Shahmoradian et al. 2019). Early research on post-mortem analysis of PD patients found evidence of dysfunction of mitochondrial complex-1 similar to the effects of environmental MPTP poisoning (Schapira et al. 1990). Later findings confirmed that inhibition of Complex I resulted in free radical formation, reduced amounts of adenosine triphosphate (ATP), and increased oxidative stress (T. M. Dawson and V. L. Dawson, 2003). These processes result in a cycle of increasing oxidative stress, neuronal cell injury, and neuronal cell death. The most common causes of familial PD are in fact mutations of Parkin and PINK1, which are genes associated with mitochondrial maintenance (Narendra et al. 2012). PINK1 and Parkin were both found to specifically support morphological maintenance of mitochondria (Palacino et al. 2004; Gautier et al. 2008) and collaborate to promote mitophagy of damaged mitochondria (Bertolin et al. 2013). The susceptibility of dopamine neurons in PD might be explained by the fact that the substantia nigra has been shown to be more susceptible to oxidative stress compared to other brain regions (Floor and Wetzel 1998). Recent work has explored ways in which α-synuclein pathology is causative and or a consequence of mitochondrial dysfunction. Aggregated α-synuclein has been shown to localize to the mitochondria and impair complex-1, thereby promoting oxidative stress and cell death (Ganjam et al. 2019). These mitochondrial and α-synuclein interactions were also found to precede the maturation of Lewy Bodies in cell based models (Mahul-Mellier et al. 2020). This is likely a consequence of ROS formation from damaged mitochondria, which has been linked to Lewy Body formation through promoting α-synuclein Ser129 phosphorylation (Perfeito et al. 2014). Lipids with oxidative modifications generated by mitochondrial damage have also been shown to promote aggregation of α-synuclein (Shamoto-Nagai et al. 2018). The pathogenic G2019S LRRK2 mutation has been shown to increase the kinase activity of LRRK2 while correspondingly decreasing mitophagy. A pharmacological intervention study showed that inhibitors of LRRK2 rescue the mitophagy defects seen in G2019S mutants in cell-culture and in mice (Singh et al. 2021). Altogether, mitochondrial stress seems to be exacerbated by α-synuclein aggregation, and also helps to propagate it.

The literature indicates that PD is the result of a combination of factors including mitochondrial damage, inhibition of the UPS, and the dysfunction of autophagic and lysosomal degradation pathways. These factors converge on the selective cell death of dopaminergic neurons through several known pathways. The neurotoxin 1-Methyl-4-Phenyl-Pyridinium (MPP+), which inhibits mitochondrial subunit 1, has been shown to mediate necrotic cell death and mimics the mitochondrial dysfunction seen in PD pathogenesis (Callizot et al. 2019). The mitochondrial stress leads to the activation of necroptosis as well as both caspase dependent and independent apoptosis. Apoptosis can also be triggered through α-synuclein mediated dysfunction of UPS via the direct and indirect inhibition of ATF6 signaling (Credle et al. 2015). Suppression of autophagy has additionally been shown to promote the susceptibility of cell death via apoptosis in mammalian cells (Boya et al. 2005). Pharmacologically promoting autophagy via enhancement of TFEB activation alleviates cell death from oxidative stress (Zhuang et al. 2020). PD related cell death was also recently found to be attenuated by ablation of RIP3, which inhibited neuroinflammation and apoptotic pathways (Dionísio et al. 2019). While apoptosis is evidently the main driver of dopaminergic loss in PD, other pathways such as neuroinflammation should also be considered in PD pathology. This is supported by numerous reports indicating that aggregated α-synuclein promotes inflammatory cell death by both microglia and astrocyte dysfunction (Kam et al. 2020). Understanding interventions for α-synuclein aggregation along with targeting its downstream effects on mitochondria, autophagy, proteostasis, and inflammation will likely yield effective therapeutic options for PD patients.

Models of PD

A great deal of work has gone into developing model systems for studying the etiology of PD. Animal models of PD include invertebrates such as C. elegans and D. melanogaster, and vertebrates such as rodents and non-human primates. Cellular models can also be used to better understand the pathways that underlie PD. These include primary neuronal cell culture derived from mice, induced pluripotent stem cells from mice and PD patients, and immortalized cell lines. Compared to animal models, cellular models are more conducive to high throughput screening and they also develop PD pathology more quickly (Falkenburger et al. 2016). Cell based modelling is much faster and less costly, but they are reduced systems that need to be followed up with animal testing.

Rodents

Rodents have been broadly used in neurodegenerative research because of their ease of maintenance, as well as their anatomical and genetic similarity to humans. PD pathology can be induced genetically in mice through the introduction of disease causing mutation models such as α-synuclein and Lrrk2 (Rothman et al. 2013; Xiong et al. 2018). Neurotoxins such as MPTP and 6-OHDA have been historically used to induce selective dopaminergic loss in both rats and mice (Zhang et al. 2017a; Rentsch et al. 2020). Viral vectors are also available to overexpress α-synuclein, in addition to methods to directly seed pre-formed α-synuclein fibrils into the rodent brain (Østergaard et al. 2020; Zhang et al. 2019a). To expedite the pathology of PD in rats, viral introduction of both α-synuclein and pre-formed fibrils has been shown to speed up loss of dopamine neurons and elicit long lasting neuroinflammation that more closely resembles the pathology seen in humans (Thakur et al. 2017). Rodent PD models exhibit dopaminergic loss in the substantia nigra which is accompanied by motor deficits. These motor deficits present an advantage of using rodent models, as a plethora of behavioral tests can be utilized to study the movement and strength of rodents (Taylor et al. 2010). The rotarod test, for example, can assess several pertinent behaviors including grip strength, balance, and motor coordination (Deacon 2013).

Non-Human Primates

Non-human primate models provide many advantages over other models due to their close anatomical and genetic similarities to humans. These similarities make non-human primate modeling very valuable to understanding the complex pathology of PD and the effectiveness of novel treatments. Similar to rodent models, nigrostriatal dopaminergic degeneration can be induced by neurotoxins and viral vectors (Vermilyea and Emborg 2015). Non-human primates also have the advantage of presenting non-motor systems such as cognitive impairment and sleep disturbances that better reflect symptoms (Choudhury and Daadi 2018) found in humans. While non-human primate models yield behaviors that are very similar to humans, their maintenance is demanding and expensive. For this reason, the use of non-human primates is often reserved for use in clinical trials.

C. elegans

As a model system for studies of mechanisms relevant to genetic and environmental causes of PD, C. elegans offer many advantages including, a short reproductive cycle, low maintenance costs, and robust options for genetic manipulations. In addition, since the nematode is a transparent organism, it is relatively simple to visualize the dopaminergic neurons by expressing a fluorescent protein such as GFP (Chalfie et al. 1994; Corsi et al. 2015; Berkowitz et al. 2008; Gaeta et al. 2019). C. elegans also do not have a homolog for α-synuclein, so researchers are able to study the effect of heterologous expression of human α-synuclein in nematode dopaminergic neurons. It has been found that overexpression of α-synuclein and PD-associated neurotoxicant 6-OHDA causes significant dopaminergic neuronal cell death (Harrington et al. 2010; Rudgalvyte and Wong 2015; Vistbakka et al. 2012; Martinez et al. 2017). The studies on α-synuclein indicate that human forms of the protein cause dopaminergic neuropathology in the C. elegans and suggest that the model can be useful in identifying novel therapeutic targets. Aside from α-synuclein, it has been documented that the PD genes found to be associated with familial PD have a C. elegans ortholog (Harrington et al. 2010; Rudgalvyte and Wong 2015; Vistbakka et al. 2012; Martinez et al. 2017). In particular, the genes PARKIN, PINK-1, LRRK2, and DJ-1 have orthologs that can be studied extensively in C. elegans (Harrington et al. 2010). Overall, C. elegans offers an experimental model to study the neuropathology involved in preventing dopaminergic neuronal loss, the interaction with genes associated with PD, the effectiveness of slowing symptom progression, and, in the end, changes in lifespan.

D. melanogaster

D. melanogaster is another invertebrate model that is useful for studying the pathobiology of PD. Similar to C. elegans, D. melanogaster also lacks a homolog for α-synuclein. Several α-synuclein lines have been generated in D. melanogaster which induce dopaminergic degeneration, locomotor deficits, and aggregation of α-synuclein (Mizuno et al. 2010). Transgenic lines for other genetic risk factors exist for Lrrk2 (Liu et al. 2008), GBA1 (Kinghorn et al. 2016), VPS35 (Inoshita et al. 2017), Parkin (Sang et al. 2007), DJ-1 (Meulener et al. 2005), and Pink1 (Park et al. 2006). Several of these transgenic lines also exhibit non-motor symptoms of PD (Julienne et al. 2017). While D. melanogaster lacks the complex brain anatomy found in mammalian models, they have provided major contributions to our understanding of PD. The function of Lrrk2 in vesicle tracking and the molecular consequences of α-synuclein aggregation and phosphorylation were first discovered in D. melanogaster (Imai et al. 2008; Lee et al. 2010; Takahashi et al. 2003). For these reasons, invertebrates such as D. melanogaster provide a powerful and relatively fast model with the means to understanding the disease mechanisms and convergent pathways that underlie PD. Discoveries from D. melanogaster can then be validated and further characterized in higher organisms such as rodents and human postmortem tissues.

MicroRNAs

MicroRNAs (miRNAs) are small, single-stranded RNAs, which are produced from hairpin shaped precursors. In animals, genes that encode miRNAs are transcribed into primary miRNA (pri-miRNA), which is then processed within the nucleus by Drosha (a class 2 RNaseIII enzyme) into precursor miRNA (pre-miRNA). This pre-miRNA is then transported into the cytoplasm via exportin-5 where it is further processed by Dicer (RNase III type protein) into the mature miRNA. One strand of the miRNA duplex binds to argonaute protein to form the RNA-induced silencing complex (RISC). This complex can then bind to a target messenger RNA (mRNA) through complementary paring of mRNA and miRNA. This pairing results in mRNA cleavage or translational repression. It has been shown that complementarity in the 5’ region of the miRNA, known as the seed region, is essential for target recognition specificity. Mismatches in this region of the miRNA can lead to compromised binding to the mRNA target and loss of function. This loss of function can, therefore, result in upregulation of mRNA translation leading to increased levels of the target protein due to reduced functionality of miRNA cleavage and/or translational repression (Bartel 2004; Engels and Hutvagner 2006; Macfarlane and Murphy 2009; Wahid et al. 2010; McGeary et al. 2019; Bartel 2018; Briskin et al. 2020). Since their discovery, there have been about 940 miRNAs discovered and over 1000 miRNAs predicted in humans (Macfarlane and Murphy 2009; Kozomara and Griffiths-Jones 2011; Kozomara et al. 2018). MicroRNAs have been shown to be involved in many different biological roles including cell death and proliferation, regulation of development, antiviral defense, development of the immune system, regulation of hematopoietic stem cells, the insulin pathway, brain morphogenesis, oxidative stress, neuronal cell fate and development, and a vast array of human disease processes (Bushati and Cohen 2007; Wahid et al. 2010; Gebert and MacRae 2019).

MicroRNAs and aging

Aging is one of the most significant risk factors in many chronic diseases such as type 2 diabetes, cancer, cardiovascular disease, and neurodegenerative disease (Niccoli and Partridge 2012). Nevertheless, aging was largely considered to be a passive process until the discovery that mutation to age-1 and daf-2, genes of the insulin/insulin-like growth factor-1 signaling pathway (IIS pathway), were shown to have dramatic effects on C. elegans lifespan (Friedman and Johnson 1988; Kenyon et al. 1993; Kimura et al. 1997). The IIS pathway in C. elegans begins with activation of the insulin or IGF-1 receptor, daf-2, which inhibits the activity of the forkhead box FOXO transcription factor, daf-16, through a phosphorylation cascade that included age-1/P13K. When daf-16 is phosphorylated, it stays in the cytoplasm and cannot relocate to the nucleus, where it is responsible for controlling genes associated with stress response, pathogen resistance, and metabolism. In combination, overexpression of daf-16 results in increased longevity and suppression of daf-16 results in decreased longevity (Kenyon 2010; Smith-Vikos and Slack 2012; Uno and Nishida 2016). Since then, multiple lifespan-regulated genes and pathways have been found, including TOR signaling, sirtuins, AMP-activated protein kinases, mitochondrial homeostasis, dietary/caloric restriction, and epigenetic mechanisms (Kenyon 2010; Macfarlane and Murphy 2009; Smith-Vikos and Slack 2012; Uno and Nishida 2016). In 2005, Boehm and Slack found that the miRNA lin-4 and its target lin-14, regulate lifespan in C. elegans (Boehm et al. 2005). These researchers found that the miRNA lin-4 might act to suppress translation of lin-14, which interacts with the daf-2 IIS pathway to affect aging. The study showed that reducing the activity of lin-4 resulted in shortened lifespan and accelerated aging while overexpressing lin-4 resulted in reduced lin-14 activity and extended lifespan (Boehm et al. 2005). This finding has led to the discovery of several other aging-associated miRNAs identified in C. elegans, Drosophila, and mice.

The identification of lin-4’s critical role in normal lifespan of C. elegans led to genetic surveys of other miRNAs with possible functions on aging. Several miRNAs, including miR-71, miR-238, miR-246, miR-239, and miR-34 have since been shown to affect normal C. elegans longevity (Lencastre et al. 2010; Pincus et al. 2011; Yang et al. 2013a; Boulias and Horvitz 2012). As with lin-4, it has been shown that miR-71 and miR-239 function via the IIS pathway (Lencastre et al. 2010; Pincus et al. 2011). Overexpression of miR-71 was found to increase longevity and deletion was found to reduce longevity, whereas overexpression of miR-239 was found to decrease longevity and deletion was found to increase longevity (Lencastre et al. 2010). In addition, miR-71 was also found to interact with, and be involved in, the DNA damage response pathway, which may have implications in stress response in C. elegans (Lencastre et al. 2010). The miRNAs miR-238 and miR-246 were found to be elevated in aged animals suggesting that they may be involved in increasing longevity. However, only overexpression of miR-246 was found to actually increase longevity, whereas overexpression of miR-238 had no effect on aging (Lencastre et al. 2010). Most recently, the role of miR-34 has been studied using C. elegans. One study, done in 2013 by Yang et al., showed that reducing the activity of miR-34 increased longevity in C. elegans models (Yang et al. 2013b). Another study, done in 2016 by Isik et al., showed the opposite, that overexpression of miR-34 increased longevity in C. elegans models (Isik et al. 2016a). Further research into the roles of miRNAs and aging is needed to clarify the specific interactions they may have on aging.

In Drosophila, two specific miRNAs, miR-34 and miR-14, have been found that alter lifespan. It has been shown that loss-of-function of miR-34 results in decreased longevity of Drosophila and upregulation extends normal lifespan (Liu et al. 2012). MiR-14 has also been shown to increase longevity when upregulated and decrease longevity when a deletion mutation is introduced (Xu et al. 2003).

In mouse models, it has been shown that four miRNAs are upregulated in mouse brains: let-7, miR-17, miR-26, and miR-10 (Inukai et al. 2012). It has also been found that some miRNAs act via the IIS pathway to affect longevity. Specifically, it has been shown that miR-1, miR-26, and miR-27 interact with this pathway and affect aging Mari (Mariño et al. 2010). In particular, it was found that miR-1 was upregulated in progeroid mice models suggesting that increased levels of miR-1 would result in decreased longevity (Mariño et al. 2010).

MicroRNAs and neurodegeneration

Neurodegenerative diseases, as a whole, are characterized as late-onset, progressive disorders that lead to movement or cognitive disorders. These diseases include Alzheimer’s disease (AD), amyotrophic lateral sclerosis (ALS), Huntington’s disease (HD), and Parkinson’s disease (PD). All of these neurodegenerative diseases have a common link, which is the abnormal accumulation of protein that leads to cell death. In PD, α-synuclein protein aggregation leads to the formation of Lewy bodies whereas in AD, the aggregation of β-amyloid is associated with the formation of plaques and tangles in neuronal cells. In ALS there is a characteristic aggregation of TDP-43 in upper and lower motor neurons and in HD there is accumulation of aggregated and toxic huntingtin protein in neurons of the basal ganglia (Abe and Bonini 2013; Tan et al. 2014). Underlying each disorder, there is a complex pathogenesis that involves a multitude of basic cellular pathways including genetic components, protein folding abnormalities, protein clearance problems, mitochondrial dysfunction, and apoptosis/autophagy signaling.

Recently, scientific interest in miRNAs has grown extensively, as documented above, and many researchers are looking into the role that they may have on neurodegeneration. Recent studies have shown that miRNAs work through many different pathways that are involved in neuronal survival, protein accumulation, and interactions with abnormal, aggregated proteins (Abe and Bonini 2013; Tan et al. 2014; Eacker et al. 2009). First, it has been shown that disruptions in the miRNA biogenesis pathway causes neurodegeneration (Abe and Bonini 2013; Tan et al. 2014; Eacker et al. 2009). For example, one study showed that loss of Dicer from mouse cerebellar Purkinje cells resulted in progressive degeneration of these cells and the older mice developed age-related tremors and ataxia (Schaefer et al. 2007). Similarly, another study, in Drosophila, showed that inhibition of Dicer1 resulted in dopaminergic neuronal loss and climbing deficits (Gehrke et al. 2010). These studies, and other like it, suggest that miRNAs are essential for neuronal integrity and survival and that disruption in these systems leads to neuronal cell death and symptoms associated with this loss. Second, miRNAs have been found to interact in a variety of ways in order to alter the accumulation of toxic proteins (Abe and Bonini 2013; Tan et al. 2014; Eacker et al. 2009). For example, one study demonstrated that miR-29 is intimately involved in regulating the accumulation of β-amyloid peptide. They found that miR-29 acts to regulate the β-site APP cleaving enzyme 1 (BACE1) which plays a key function in producing the toxic protein in AD. Specifically, lower expression of miR-29 resulted in higher expression of BACE1 and increased formation of β-amyloid. Conversely, overexpression of miR-29, via transfection of HEK293 cells, showed significant reductions in the amount of β-amyloid (Hébert et al. 2008). In PD, it has been demonstrated that alterations in the binding site for miR-433 on fibroblast growth factor 20 (FGF20) resulted in increased levels of α-synuclein because miR-433 was not able to downregulate translation (Wang et al. 2008). These studies show that miRNAs are also actively involved in regulating protein accumulation in neurodegeneration by altering transcription and translation of various genes associated with AD, PD, HD, and ALS. Finally, it has been documented that the toxic protein accumulations can alter the biogenesis and function of miRNAs (Abe and Bonini 2013; Tan et al. 2014; Eacker et al. 2009). In HD, it was found that accumulated Huntingtin protein causes upregulation of Repressor Element 1 Silencing factor (REST) which suppresses hundreds of genes and is located near many miRNA genes including miR-9 and miR-124 (Packer et al. 2008; Zuccato et al. 2007). In ALS, it has been shown that TDP-43, which is found in the inclusions of patients with ALS, interacts with Dicer and Drosha and can alter miRNA levels in cells (Buratti et al. 2010). These studies demonstrate that, not only do miRNAs affect accumulation of toxic protein, but also the toxic proteins can affect the synthesis and function of miRNAs. This interaction leading to altered miRNA function can drastically affect the functionality and integrity of the cell and may lead to cell death.

MicroRNAs: links between aging, stress response, and chronic disease

Aside from their role in aging, as documented above, miRNAs have been found to be associated with an organism’s response to stress as well (Lencastre et al. 2010; Leung and Sharp 2010; Prahlad and Morimoto 2009). In particular, miRNAs have the ability to interact with and regulate multiple targets and have also been shown to be involved in an organism’s response to multiple stressors including DNA damage (Wan et al. 2011; Wang and Taniguchi 2013; Wouters et al. 2011), hypoxia (Kagias and Pocock 2015), thermal regulation (Fukuoka et al. 2014; Nehammer et al. 2015), and nutrient deprivation in plants (Paul et al. 2015). These findings suggest a link between miRNAs and the pathways involved in stress response that may affect an organism’s longevity and lifespan. Previous research has shown that studying mutants with aberrant longevity or stress response can be useful in studying neurodegenerative disease models. For example, Cooper et al. showed that delaying aging is neuroprotective in PD models using C. elegans (Cooper et al. 2015). When C. elegans PD animals expressing α-synuclein were crossed with long-lived daf-2 mutants, the increase in longevity from the daf-2 mutation was found to ameliorate the deficits seen in dopamine-dependent behaviors and also lowered the sensitivity to stress that was observed in the PD animals (Cooper et al. 2015). Interestingly, the effects of some miRNAs on longevity have been shown to function, at least partially, through the daf-2 pathway (Lencastre et al. 2010). Together, these results, as well as the high degree of conservation of miRNAs, have led researchers to look at the application of miRNAs associated with aging and stress in slowing the progression of PD and other age-dependent diseases.

Association of microRNAs with Parkinson’s Disease: findings from studies in model organisms and clinical human research

C. elegans

Caenorhabditis elegans (C. elegans) has been used extensively since the 1970s as an animal model to study development and neurobiology (Kaletta and Hengartner 2006). The first C. elegans model for PD was examined in 2003 by Lakso et al., which showed that transgenic worms that overexpressed α-synuclein resulted in dopaminergic neuronal loss, loss of dendrites, increase in neuronal process breaks, and motor deficits (Lakso et al. 2003). A global expression profile of C. elegans miRNA in PD showed that twelve specific miRNAs are differentially regulated in models overexpressing α-synuclein. These include miR-50, miR-1, miR-236, miR-77, miR-64, miR-65, miR-80, miR-238, miR-58, miR-83, miR-51, miR-48, miR-241, miR-230, miR-84, and let-7 (Asikainen et al. 2010). Researchers have now started looking into possible specific roles that miRNAs may have on PD pathogenesis. Currently, the roles of let-7, miR-34 and other aging-associated miRNAs in C. elegans are under scrutiny. Studies have shown that let-7 is involved in regulating the E2F1 transcription factor and that low levels of let-7, which is observed in PD C. elegans models, results in overexpression of this factor leading to dopaminergic neuronal loss (Shamsuzzama et al. 2015, 2017). Interestingly, in two deep-sequencing surveys of C. elegans miRNA expression during aging, let-7 was observed to be one of the most highly down-regulated miRNA during aging (Lencastre et al. 2010; Kato et al. 2011). Although C. elegans let-7 was first identified because of its essential role during development (Reinhart et al. 2000), numerous studies have since demonstrated that let-7 has myriad adult specific functions across phyla, including roles in cell differentiation and as a tumor suppressor (Esquela-Kerscher and Slack 2006; Johnson et al. 2005). Together, these studies indicate adult-specific roles of let-7 and that down-regulation of let-7 during aging and in PD might be detrimental to animal health. Conversely, the miRNA miR-34 was shown to be one of the most highly up-regulated miRNAs during aging in C. elegans (Lencastre et al. 2010; Kato et al. 2011). Studies of miR-34 mutants in C. elegans demonstrated roles in regulating the IIS pathway and in stress modulation and longevity. Interestingly, both deletion and overexpression of miR-34 results in impaired stress response via alteration in daf-16 (Isik et al. 2016b; Yang et al. 2013a). These findings suggest that C. elegans stress response and aging are sensitive to normal miR-34 expression levels. Although the function of miR-34 on C. elegans PD models has not yet been tested, miR-34 has been implicated in neuronal function in Drosophila, mouse and human neurodegenerative disease (see below).

Drosophila

Drosophila has dopaminergic neurons in the central nervous system and, as in vertebrates, dopamine plays a key role in Drosophila’s learning, feeding, locomotion, and sleep cycle (West et al. 2015). These characteristics have promoted the use of Drosophila as a model to understand the pathogenesis of PD and to find potential therapeutic targets. One microRNA, miR-34, was found to have a direct link between aging and neurodegeneration. Liu et al. showed that 29 microRNAs are expressed in the adult brain, but only one, miR-34, increases with age (Liu et al. 2012). In addition, loss of miR-34 was shown to result in accelerated brain aging, brain degeneration, and decline in longevity and survival, whereas overexpression of miR-34 results in extended lifespan and reduced neurodegeneration (Liu et al. 2012). This function of miR-34 in Drosophila neuronal health was found to depend on Eip74EF, a target of miR-34 which functions in steroid hormone signaling (Liu et al. 2012). Given the already identified role of miR-34 on C. elegans stress resistance and aging (Isik et al. 2016b; Yang et al. 2013a), as well as a potential role in other forms of dementia (Dickson et al. 2013; Zovoilis et al. 2011). These studies further suggest that miR-34 might be a key, conserved miRNA with roles on aging and pathogenesis of PD. Flies expressing pathogenic variants of LRRK2 have also been shown to have impaired miRNA processing through the dysregulation of RISC, which led to decreased levels of let-7 and miR-184 (Gehrke et al. 2010).

Rodents

Recently, the role of microRNAs on the pathogenesis of PD in mouse models has been studied. In adult mice, dopaminergic-specific ablation of the miRNA processing RNAse Dicer was found to lead to dopaminergic neurodegeneration and locomotor deficits (Chmielarz et al. 2017). These results highlight the cell-autonomous importance of microRNAs in promoting cell survival of dopaminergic neurons. Two groups of researchers have also found that one microRNA, miR-7, seems to act as a regulator of α-synuclein expression and build-up in mice (Junn et al. 2009; Doxakis 2010). First, Junn et al. showed that miR-7 acts via the 3’-untranslated region (UTR) to repress α-synuclein formation in mice. This down-regulation of α-synuclein by high levels of miR-7 resulted in protection against oxidative stress and promoted survival of dopaminergic neurons. Then, Doxakis et al., independently studied miR-7, as well as miR-153, and reported similar results. These authors reported that increased levels of miR-7 and miR-153 in mouse models were associated with decreased levels of α-synuclein. Specifically, they reported that increased levels of both miR-7 and miR-153 resulted in a 30–40% decrease in endogenous α-synuclein (Doxakis 2010; Fragkouli and Doxakis 2014). Furthermore, measurements of circulating miRNAs levels in GFAP.HMOX1 mouse model of PD, which expresses the human heme oxygenase-1 (HO-1) gene in astrocytes, showed significantly lower levels of miR-153 and correspondingly increased levels of α-synuclein (Cressatti et al. 2019). These results suggest that delivery and increase in both miR-7 and miR-153 may be neuroprotective in patients with a known α-synuclein gene mutation.

Besides miR-7 and miR-153, another microRNA that has been studied and implicated as a possible link to the pathogenesis of PD is miR-155. MiR-155 has been shown to be involved in the inflammatory response to α-synuclein and the eventual neuronal cell death (Thome et al. 2016). This study showed that the loss of miR-155 resulted in a reduced inflammatory response to α-synuclein and reduction in α-synuclein-induced neurodegeneration. Also, they found that restoration of miR-155 restored the inflammatory response and promoted neuronal cell death (Thome et al. 2016). This provides a direct target that can be used therapeutically to alter the inflammatory response in dopaminergic neurons of patients with PD.

Work in rat models of PD have also highlighted important miRNAs that can be possibly be used as therapeutic targets. In a 6-hydroxydopamine rat model, treatment with a miR-200a inhibitor was shown to repress the cAMP/PKA signaling pathway and promote cell survival in the striatum (Wu et al. 2018). miR-375 mimics have also been shown to reduce cell death, neuroinflammation, and neurobehavioral defects in the 6-hydroxydopamine rat model via the inhibition of SP1 (Cai et al. 2020). 6-hydroxydopamine rat models have exhibited reduced miR-218–5p which has shown to target LASP1. Overexpression of miR-218–5p with AAVs in rats exposed to 6-hydroxydopamine was shown to be protect dopaminergic neurons and limit oxidative stress (Ma et al. 2021). Other rat models such as the rotenone induced model have been used to determine miRNAs associated with causing neurodegeneration. Jauhari and colleagues recently found that NF-κB promotes miR-146a upregulation in rats treated with rotenone. In the context of rotenone exposure, miR-146a is pro-inflammatory and targets the downregulation of Parkin, resulting in mitochondrial dysfunction and the production of reactive oxygen species (Jauhari et al. 2020). Other miRNAs have been found to be differentially expressed in the striatum of rotenone exposed rats, such as miR-26a, miR-34a, let7a, and miR-7. Horst and colleagues found that miR-26a and miR-34a were upregulated, while let7a and miR-7 were downregulated in just 10 days of daily exposure to rotenone (Horst et al. 2018).

The Role of MicroRNAs in Parkinson’s Disease – Human Studies

MicroRNA profiling in Parkinson’s disease patients

In a study of human brain tissue of PD patients, it was found that 125 miRNAs were altered at genome-wide levels in the prefrontal cortex of PD patients examined with next-generation sequencing (Hoss et al. 2016). Of these 125 miRNAs, it was shown that the levels of 64 miRNAs were downregulated and the levels of 61 miRNAs were upregulated as compared to normal tissue. This finding suggest possible roles for these miRNAs on pathogenesis of PD or as biomarkers of PD and may suggest possible therapeutic targets going forward. Specifically, one study showed that levels of miR-34 were decreased in several brain areas with varying degree in brains of PD patients compared to controls (Minones-Moyano et al. 2011). A second study showed that expression of miR-133 was specifically lowered in PD patient brain samples compared to controls, which was confirmed by RNase protection assays and Northern blotting (Kim et al. 2007). Finally, beyond brain tissue surveys, another study assessed microRNA levels as possible biomarkers of PD in patients’ serum (Ma et al. 2016). The authors found that levels of four miRNAs were significantly decreased in PD patients compared to healthy controls and these included miR-29, miR-146, miR-214, and miR-221. Specifically, miR-221 was found to be most correlated with disease and it was documented that downregulated serum miR-221 may be a potential biomarker for the evaluation of PD (Ma et al. 2016). Consistent with this, several more recent studies in mouse models and in human cell lines demonstrate that miR-221 is downregulated in PD models and is neuroprotective (see also further discussion below) (Oh et al. 2018a; Lang et al. 2020a). Two other more recent expression surveys have identified differentially expressed miRNAs associated with PD pathology. In a study where dopaminergic neurons (DA) were generated from induced pluripotent stem cells (iPSCs) generated from PD patients, 5 miRNAs were upregulated (miR-9–5p, miR-135a-5p, miR-135b-5p, miR-449a, and miR-449b-5p) while 5 others were downregulated (miR-141–3p, miR-199a-5p, miR-299–5p, miR-518e-3p, and miR-519a-3p) in the DA cells from PD patients (Tolosa et al. 2018). The upregulation of miR-9–5p and miR-135b-5p was associated with the corresponding downregulation of the transcription factors FOXA1 and NR3C1 (Tolosa et al. 2018). A more recent survey profiled miRNA serum expression in a cohort of 139 patients of various neurodegenerative diseases, including PD, and found that let-7d, miR-15b, miR-24, miR-142–3p, miR-181c and miR-222 showed altered expression in serum of PD patients. (Barbagallo et al. 2020)

The changes in miRNA expression identified in the surveys described above suggest possible functional roles of some of these miRNAs during PD. A summary of major findings that have identified miRNAs that are differentially and/or may play functional roles in PD pathology are summarized in Table 1 and in Figure 1. Below, we further discuss the research that has been done using human cell lines on several key miRNAs which have been shown to alter the expression of α-synuclein or affect other aspects of PD pathology (Figure 1B):

Table 1:

Key microRNAs implicated in the pathogenesis of Parkinson’s disease.

MicroRNA	Organism	Findings	Reference
Expression Profile	C. elegans	Global expression profile of PD models twelve differentially regulated miRNAs: let-7, miR-50, miR-1, miR-236, miR-77, miR-64, miR-65, miR-80, miR-238, miR-58, miR-83, miR-51, miR-48, miR-241, miR-230, and miR-84	(Asikainen et al. 2010)
	Human	Levels of miR-26a, miR-28–5p, miR-29b/c, miR-30b/c, miR-126, miR-147, miR-151–5p, miR-199, miR-301a, miR-335, and miR-374a/b were lower in the peripheral blood cells of PD patients compared to healthy controls	(Martins et al. 2011)
	Human	Examining blood plasma, 13 differentially expressed miRNA were found in PD patients: miR-9, miR-192, miR-222, miR-505, miR-506, miR-572, miR-626, miR-671–5p, miR-647, miR-1225–5p, miR-1307, miR-1826, miR-5488–3p	(Khoo et al. 2012b)
	Human	Expression of 4 miRNAs were significantly reduced in PD patients’ serum: miR-29, miR-146, miR-214, and miR-221	(Ma et al. 2016)
	Human	Expression of 125 miRNAs were found to be altered in the prefrontal cortex of PD patients	(Hoss et al. 2016)
	Human	Upregulation of miR-9–5p, miR-135a-5p, miR-135b-5p, miR-449a, and miR-449b-5p and downregulation of miR-141–3p, miR-199a-5p, miR-299–5p, miR-518e-3p, and miR-519a-3p in dopaminergic neurons generated from iPSC of PD patients. Upregulation of miR-9–5p and miR-135b-5p associated with downregulation of FOXA1 and NR3C1.	(Tolosa et al. 2018)
	Human	let-7d, miR-15b, miR-24, miR-142–3p, miR-181c and miR-222 showed altered expression in serum of PD patients.	(Barbagallo et al. 2020)
let-7	C. elegans	Low levels of let-7 resulted in dopaminergic neuronal loss	(Shamsuzzama et al. 2015, 2017)
	Drosophila	Found to be down-regulated in Lrrk2 pathogenic mutants and is implicated in PD pathology	(Gehrke et al. 2010)
	Mouse	Let-7a suppresses microglia inflammation caused by α-synuclein expression by targeting STAT3	(Zhang et al. 2019b)
	Rat	Let-7a was found to be downregulated in a rotenone model.	(Horst et al. 2018)
	Human	Let-7d down-regulated in 6-OHDA-induced cellular model of PD and targets caspase-3	(Li et al. 2017)
miR-7	Mouse	Increased miR-7 levels repressed α-synuclein formation and resulted in increased survival of dopaminergic neurons	(Junn et al. 2009)
	Mouse	Increased miR-7 levels were associated with decreased α-synuclein levels	(Doxakis 2010; Fragkouli and Doxakis 2014)
	Rat	miR-7a was found to be downregulated in a rotenone model.	(Horst et al. 2018)
	Human	Overexpression of mir-7 resulted in upregulation of the mTOR pathway in MPP+-induced cells acting to protect the cell against death caused by mitochondrial complex I inhibition	(Fragkouli and Doxakis 2014)
	Human	miR-7 targets RelA, a subunit of NF-kB, that allows NF-kB to be derepressed and active which then acts to protect the cell against death due to mitochondrial complex I inhibition	(Choi et al. 2014)
	Human	Ectopic expression of miR-7 reduced the expression of an α-synuclein reporter in MPP+-induced cells. Regulation of α-synuclein was affected by mitochondrial ROS levels.	(Je and Kim 2017)
	Human	FTY720, a neuroprotective drug increases the expression of miR-7a-5p in dopaminergic cells and reduces α-synuclein expression.	(Vargas-Medrano et al. 2019)
	Human	MMP+ upregulates SOX21-AS1 which regulates miR-7 expression. IRS2 is targeted by miR-7.	(Xie et al. 2021)
miR-9	Human	Upregulation of miR-9 in MPP+-induced cells	(Delavar et al. 2018)
	Human	FTY720, a PD therapeutic drug increases the expression of miR-9–5p in dopaminergic cells and reduced α-synuclein expression.	(Vargas-Medrano et al. 2019)
miR-26a	Rat	miR-26a was found to be upregulated in a rotenone model.	(Horst et al. 2018)
miR-34	C. elegans	Both overexpression and deletion of mir-34 resulted in altered stress response	(Yang et al. 2013a; Isik et al. 2016b)
	Drosophila	Loss of mir-34 results in increased brain degeneration and a decline in longevity. Overexpression of mir-34 resulted in extended lifespan and reduced neurodegeneration	(Liu et al. 2012)
	Mouse	miR-34a overexpression inhibited neuroprotection by Schisandrin B in cellular model and in PD mouse model.	(Ba et al. 2015)
	Rat	miR-34a was found to be upregulated in a rotenone model.	(Horst et al. 2018)
	Human	Reduced expression of miR-34b/c in brains of PD patients resulted in reduction of cell viability due to mitochondrial dysfunction, oxidative stress, and a reduction in ATP production.	(Minones-Moyano et al. 2011)
	Human	Inhibition of mir-34 resulted in significantly increased α-synuclein mRNA and protein levels and that overexpression of mir-34 resulted in decreased α-synuclein mRNA and protein levels	(Kabaria et al. 2015)
	Human	Neuroprotective effect of Schisandrin B in 6-OHDA-induced cellular model showed a reduction of miR-34a. miR-34a overexpression inhibited neuroprotection by Schisandrin B in cellular model by negative regulation of Nrf2.	(Ba et al. 2015)
	Human	Upregulation of miR-34a in MPP+-induced cells	(Delavar et al. 2018)
miR-133	Drosophila	miR-133 is upregulated in early stage PD flies and acts to regulate midbrain dopaminergic neurons via Pitx3	(Kong et al. 2015)
	Human	Reduced levels of miR-133b in midbrain tissue of PD patients but no difference in miR-133b expression in dopaminergic neurons.	(Kim et al. 2007; Schlaudraff et al. 2014)
	Human	Overexpression of mir-133b led to significant reductions in the number of dopaminergic neurons and overall dopamine release.	(Kim et al. 2007)
	Human	No association was found between miR-133b and a risk for PD	(Mena et al. 2010)
	Human	Reduced circulating levels of miR-133a in plasma of PD patients	(Zhang et al. 2017b)
	Human	miR-133a is downregulated in a MPP+ cell model of PD. miR-133a overexpression increases cell proliferation and inhibits both apoptosis and autophagy by inhibiting RAC1 expression.	(Lu et al. 2020)
miR-146a	Rat	miR-146 is upregulated in a rotenone model and inhibits parkin, leading to mitochondrial dysfunction and cell death.	(Jauhari et al. 2020)
miR-153	Mouse	Increased miR-153 levels were associated with decreased α-synuclein levels	(Doxakis 2010; Fragkouli and Doxakis 2014)
	Mouse	Circulating levels of miR-153 are significantly lower in the GFAP.HMOX1 mouse model of PD, which expresses the human heme oxygenase-1 (HO-1) gene in astrocytes and which has increased levels of α-synuclein.	(Cressatti et al. 2019)
	Human	Overexpression of miR-153 resulted in upregulation of the mTOR pathway and acts to protect the cell against death due to mitochondrial complex I inhibition	(Fragkouli and Doxakis 2014)
	Human	Overexpression of miR-153 resulted in downregulation of α-synuclein. Also regulates functional proteasome.	(Patil et al. 2015)
	Human	Ectopic expression of miR-153 reduced the expression of an α-synuclein reporter in MPP+-induced cells. Regulation of α-synuclein expression was affected by mitochondrial ROS levels.	(Je and Kim 2017)
	Human	Increased miR-153 levels in serum of PD patients and in MPP+-induced cells. Overexpression of miR-153 promotes oxidative stress by negative regulation of Nrf-2.	(Zhu et al. 2018)
	Human	Decreased levels of salivary miR-153 in idiopathic PD patients.	(Cressatti et al. 2020)
	Human	Long non-coding RNA SNHG1 reduces miR-153 expression by acting as a sponge, leading to activation of PTEN/AKT/mTOR pathway in MPP+-activated cells an in PD mice	(Zhao et al. 2020)
miR-155	Mouse	Low levels of miR-155 resulted in reduced inflammatory response to α-synuclein and lower dopaminergic death and that restoration of miR-155 restored this inflammatory response	(Thome et al. 2016)
miR-184	Drosophila	miR-184 is down-regulated in Lrrk2 mutants and is implicated in PD pathology	(Gehrke et al. 2010)
miR-200a	Rat	Inhibition of miR-200a was shown to limit cell death after 6-hydroxydopamine exposure via repression of the cAMP/PKA pathway.	(Wu et al. 2018)
miR-218–5p	Rat	Overexpression of miR-218–5p inhibited LASP and promoted neuroprotection in 6-hydroxydopamine model.	(Ma et al. 2021)
miR-221	Mouse	miR-221 expression is reduced in DJ-1−/− mouse brain.	(Oh et al. 2018b)
	Mouse	PD mouse model induced by MPTP and in MPP+-induced cells show increased expression of HOTAIR. HOTAIR binds to miR-221–3p and reduces NPTX2 expression leading to increased autophagy of dopaminergic neurons and decreased cell viability.	(Lang et al. 2020b)
	Human	Iron accumulation in dopaminergic neurons was shown to be involved in the pathogenesis of PD and miR-221 acts to regulate the amount of intracellular iron accumulation	(Asci et al. 2013)
	Human	Downregulated levels of serum miR-221were correlated with PD risk and may act as a biomarker for disease	(Ma et al. 2016)
	Human	miR-221 is cytoprotective in MPP+-activated cells and down-regulates pro-apoptotic factors such as BIM. DJ-1 regulates miR-221 expression through MAPK/ERK1/2 pathway.	(Oh et al. 2018b)
	Human	Long non-coding RNA SNHG1 competitively binds miR-221 in MPP+-induced cells and indirectly regulates p27/mTor pathway.	(Qian et al. 2019)
miR-375	Rat	Overexpression via miR-375 mimics were shown to be protect against cell death in a model of 6-hydroxydopamine exposure via SP1 inhibition.	(Cai et al. 2020)
miR-425	Mouse	Ablation of miR-425 exacerbates dopaminergic loss after MTPP administration. miR-425 was also specifically found to regulate the expression of necroptosis initiation factor, RIPK1	(Hu et al. 2019)
	Human	miR-245 is downregulated in dopaminergic neurons of PD patients	(Hu et al. 2019)
	Human	Long non-coding RNA SNHG7 was upregulated and miR-425 was downregulated in PD patients.	(Zhang et al. 2021)

Figure 1.

Schematic representation of interactions between key miRNAs, putative gene targets and associated consequences in PD pathology, neurodegeneration and stress response.

A. Summary of studies on animal models of PD that implicate the role of key miRNAs in functional roles that regulate genes pathways that control neuronal survival (green moves), neurodegeneration and neuronal cell death (red boxes) or stress response (grey box). Arrows indicate positive regulation while arrows terminating in T-bar indicate negative regulatory effects.

B. Summary of studies that identify key differentially regulated miRNAs in PD, their associated gene targets and predicted effects on PD-associated pathologies. Symbols as in panel A.

let-7

As discussed before, let-7 was the second identified miRNA (Reinhart et al. 2000). It is highly conserved across metazoans and has a multitude of functions in biology – both during development as well as during animal adulthood. It is also associated with human disease, including roles as a tumor suppressor (Esquela-Kerscher and Slack 2006; Johnson et al. 2005). Consistent with its ubiquity in the animal world, let-7 has been shown to have PD-associated roles across most model systems: low levels of let-7 result in dopaminergic neuronal cell loss in C. elegans, it is downregulated in LRRK2 pathogenic mutants in Drosophila, it suppresses microglia inflammation caused by α-synuclein expression in mice by targeting STAT3 and it is down-regulated in a 6-OHDA-induced cell model of PD and targets caspase-3 (Shamsuzzama et al. 2015, 2017; Zhang et al. 2019b; Li et al. 2017). Intriguingly, let-7 was observed to be one of the most highly down-regulated miRNAs during aging in C. elegans (Lencastre et al. 2010; Kato et al. 2011). This observation, together with the association of low let-7 levels with pathology in PD models, suggest that let-7 might have neuroprotective roles that are lost as the organism ages and let-7 levels decrease.

miR-34

The miR-34 family consists of three processed miRNAs that are encoded by two separate genes. MiR-34a is encoded on chromosome 1 from its own transcript while miR-34b and miR-34c are encoded on chromosome 11 and share a common transcript. Analysis of the tissue distribution for the miR-34 family has shown that the highest levels of miR-34a are in the brain, whereas, expression of miR-34b and miR-34c are highest in the lungs (Agostini and Knight 2014). Consistent with this expression profile, MPP+ expression in human cell lines significantly increased levels of miR-34a (and other miRNAs) with concomitant reduction in SIRT1, BCL2 and BDNF (Delavar et al. 2018). In a mouse model, miR-34a overexpression inhibited neuroprotection by Schisandrin B in cellular model and in PD mouse model (Ba et al. 2015). In human cell line studies, depletion of miR-34b and miR-34c resulted in reduction in cell viability due to dysregulation of mitochondrial function, oxidative stress response, and a reduction in ATP production (Minones-Moyano et al. 2011). These studies also demonstrated that a reduction in miR-34b/c was associated with a significant reduction in DJ1 and Parkin protein levels, which induces mitochondrial dysfunction, oxidative damage, and decreased cell survival. Another study showed that inhibition of miR-34b/c resulted in increases in α-synuclein expression in dopaminergic SH-SY5Y cells (Kabaria et al. 2015). Specifically, the study showed that inhibition of miR-34b and miR-34c using anti-miRs resulted in a significant increase in both α-synuclein mRNA levels as well as α-synuclein protein levels, by a magnitude of 2.2-fold and 1.7-fold respectively (Kabaria et al. 2015). This increase in protein levels was also correlated with the presence of protein aggregates inside cells. Also, they showed that overexpression of miR-34b and miR-34c resulted in significant decrease in α-synuclein mRNA levels and α-synuclein protein levels. These findings suggest that miR-34b/c targets the 3’-UTR of α-synuclein mRNA to decrease the formation of α-synuclein protein.

miR-133

The mir-133 family consists of three mature miRNAs, miR-133a-1, miR-133a-2, and miR-133b, which are encoded from chromosomes 18, 20, and 6 respectively. In Drosophila, miR-133b is upregulated in early stage PD flies and interacts with Pitx3 to regulate the maturation and function of midbrain dopaminergic neurons (Kong et al. 2015). This suggests that miR-133b may potentially play a role in the pathogenesis of PD. Studies in human cell lines have examined the role of miR-133b on PD model cells. One study showed that overexpression of miR-133b in primary midbrain cultures led to a significant reduction in the number of dopaminergic neurons and overall dopamine release (Kim et al. 2007). However, a second study documented no association between miR-133b and Pitx3 gene variants to the risk for PD (Mena et al. 2010). On the other hand, the circulating levels of miR-133a are significantly reduced in the plasma of PD patients (Zhang et al. 2017b). Similarly, in a MPP+-induced cell model of PD, miR-133a was also found to be significantly downregulated (Lu et al. 2020). Furthermore, miR-133a overexpression was found to increase cell proliferation and inhibit both apoptosis and autophagy by negative regulation of RAC1 expression (Lu et al. 2020). These results highlight that further research is warranted to further dissect the possible roles of the miR-133 family on the pathogenesis and risk of PD.

miR-221

MiR-221 is encoded in tandem with miR-222 from a gene cluster on chromosome X and has been implicated in a number of biological processes including apoptosis and cancer (Brognara et al. 2016). As explained above, miR-221 has been shown to be downregulated in PD patients and the finding of this downregulation in a patient’s serum suggests that miR-222 might be a biomarker for the disease. Also, it has been shown that iron accumulation in dopaminergic neurons and glial cells in the substantia nigra of PD patients may contribute to the development of oxidative stress, protein aggregation, and neuronal cell death (Asci et al. 2013). Specifically, these authors demonstrated that miR-221 interacts with the transferrin receptor 2 (TFR2) gene, which is responsible for encoding a protein that is expressed on the cell membrane and is involved in the uptake of iron. The study further showed that overexpression of miR-221 resulted in decreased expression of the TFR2 protein on the surface of cells suggesting that miR-221 may play a key role in regulating the amount of intracellular iron and, as a result, the amount of iron-induced oxidative stress in PD associated cells. Further connections between miR-221 and pathways of relevance in PD pathogenesis are indicated by recent studies utilizing mouse models and MPP+-induced cells. The first study showed that miR-221 expression is reduced in DJ-1^−/− mouse brain while, MPP+-induced human cells showed that miR-221 is cytoprotective in and down-regulates pro-apoptotic factors such as BIM (Oh et al. 2018b). Other studies identify possible regulators of miR-221: in a PD mouse model induced by MPTP and in MPP+-induced cells show increased expression of HOTAIR. HOTAIR binds to miR-221–3p and reduces NPTX2 expression leading to increased autophagy of dopaminergic neurons and decreased cell viability (Lang et al. 2020b). In another human cell PD model, the long non-coding RNA SNHG1 was found to competitively bind miR-221 in MPP+-induced cells and indirectly regulate p27/mTor pathway (Qian et al. 2019). These studies highlight that miR-221 is likely to have important roles as both a marker for disease and in the pathogenesis of neuronal cell death in PD and begin to identify pathways that offer mechanistic insight as to its function.

miR-7 and miR-153

As described above, both miR-7 and miR-153 have been shown to decrease levels of α-synuclein in mouse PD models (Doxakis 2010; Fragkouli and Doxakis 2014), suggesting that they may be neuroprotective in PD patients. In humans, both miR-7 and miR-153 have been shown to play important, neuroprotective roles in PD cells. Indeed, and in agreement with the mouse models, it has been shown that overexpression of miR-153 on human HEK293 cells resulted in downregulation of SNCA gene expression and a reduction in α-synuclein protein (Patil et al. 2015). The study showed that miR-153 plays a key role in regulating multiple neuronal processes including efficient proteasome activity, which is necessary to degrade misfolded and aggregated protein that is seen in PD (Patil et al. 2015). Specifically, overexpression of miR-153 acts to increase two proteins, proteasome subunit α type-1 isoform 2 (PSMA1) and Prefolding subunit 2 (PFDN2). PSMA1 is involved in proper proteasome activity and PFDN2 is responsible for transferring misfolded proteins to chaperonins ensuring proper folding. In addition, two independent studies found that miR-7 acts to protect cells against death due to mitochondrial complex I dysfunction, which is thought to play a key role in the pathogenesis of PD (Choi et al. 2014; Fragkouli and Doxakis 2014). Both researchers used 1-Methyl-4-Phenyl-Pyridinium (MPP⁺), which acts as a neurotoxin that inhibits mitochondrial Complex I, to mimic the mitochondrial dysfunction seen in PD pathogenesis. First, Fragkouli et al. showed that overexpression of both miR-7 and miR-153 resulted in upregulation of mTOR signaling, which plays key roles in cell metabolism, growth, and survival, and acted to be neuroprotective in MPP⁺ treated cells). Second, Choi et al. demonstrated that, in addition to inhibiting α-synuclein mediated cell death, miR-7 acts to protect against MPP⁺-induced toxicity (Choi et al. 2014). They showed that miR-7 protects against this toxicity by targeting RelA, a subunit of the nuclear factor-κB (NF-κB) transcription factor complex. RelA is responsible for mediating MPP+ induced cell death by suppressing NF-κB activity. MiR-7 acts to reduce RelA expression and allows NF-κB to be derepressed and active to protect against MPP+-induced toxicity. Furthermore, ectopic expression of miR-7 has been shown to reduce the expression of an α-synuclein reporter in MPP+-induced cells (Je and Kim 2017). Consistent with the previous studies, this let-7 mediated regulation of α-synuclein was affected by mitochondrial ROS levels (Je and Kim 2017). Two other studies have showed that increased expression of miR-7 is associated with neuroprotection: FTY720, a neuroprotective drug, increases the expression of miR-7a-5p in dopaminergic cells and reduces α-synuclein expression (Vargas-Medrano et al. 2019). Cells treated with MMP+ were shown to upregulate SOX21-AS1 which, in turn, binds miR-7–5p. Over-expression of miR-7–5p was shown to rescue MMP+-induced cell damage by targeting insulin receptor substrate 2 (IRS2) (Xie et al. 2021). These studies show that miR-7 and miR-153 play a key role in the pathogenesis of PD to decrease the amount of α-synuclein protein as well as protect against mitochondrial dysfunction-induced cell death.

Current and Future Research Directions

This literature review highlights the important roles that miRNAs have on the pathogenesis of PD as well as demonstrating them as possible therapeutic targets to slow the progression of PD. Functional studies of PD in human subjects are not possible, so some of the basic questions of miRNA function in PD pathology can first be addressed in models of PD. Newly emerging technologies, such as CRISPR, will allow for rapid and precise generation of knockouts and other transgenic animals that will accelerate the discovery of new therapeutic roles of miRNAs in PD. Recent work by Hu and colleagues, for example, used CRISPR to generate miR-425 deficient mice which exhibit aggravated dopaminergic degeneration and motor loss after treatment with MPTP (Hu et al. 2019). Based on their post mortem-analysis of PD brains that found miR-425 to be downregulated, Hu and colleagues hypothesized that miR-425 may have a protective role against PD pathology. This was also supported by later work in mice that indicated that miR-425 downregulates the expression of RIPK1, a protein that triggers necroptosis (Lee et al. 2010). Current advances in stem cell biology has allowed for the generation of patient derived induced pluripotent stem cells (iPSCs) which can further inform on the pathology of PD and bring precision medicine closer to reality. In 2018, Tolosa and colleagues found that the levels of 10 miRNAs were altered in dopaminergic neurons generated from iPSCs derived from PD patients compared to healthy controls. More specifically, up regulation of miR-9–5p and miR-135b-5p was found to negatively regulate PD associated transcription factors such as FOXa1 (Tolosa et al. 2018). Conditional knockout of FOXa1 in mice has been shown to downregulate tyrosine hydroxylase, leading to impairments in dopamine transmission (Pristerà et al. 2015). These experiments highlight the power of exploring new technologies and systems for delineating the role of miRNAs in the complex etiology of PD.

The current diagnosis of PD in patients traditionally relies on the presenting symptoms and the medical history of the patient. Many research groups have shown that miRNAs levels become altered in PD pathology, sparking interest in the use of miRNA levels as a biomarker for PD. One method of easily collecting data from patients and discerning levels of miRNA could be from collecting peripheral blood samples. Using microarrays, Martins and colleagues found that the levels of miR-26a, miR-28–5p, miR-29b/c, miR-30b/c, miR-126,miR-147, miR-151–5p, miR-199, miR-301a, miR-335, and miR-374a/b were lower in the peripheral blood cells of PD patients compared to healthy controls (Martins et al. 2011). Additionally, to increase the predictive power of biomarkers, miRNAs that show high sensitivity can be combined into one metric with miRNAs that exemplify high specificity for PD. Khoo and colleagues used this approach after identifying 13 differentially expressed miRNAs from the blood plasma of PD patients to increase their predictive performance (Khoo et al. 2012a). More recent advances have looked to use miRNA changes in saliva as an inexpensive and minimally invasive technique. Cressatti et al., for instance, found that the salivary levels of both miR-153 and miR-223 were lower in patients with idiopathic PD compared to the levels found in healthy controls (Cressatti et al. 2020). Another recent study by Ravanidis and colleagues found that dysregulated brain enriched miRNAs are different in patients with idiopathic PD and genetic (GBA and SNCA^A53T) PD. Using pathway enrichment analysis, the authors show that the SNCA and GBA populations comprising the genetic group are likely similar because the SNCA and GBA mutations converge on the same miRNAs that drive PD pathology (Ravanidis et al. 2020). Overall, these studies exemplify how the alterations in miRNA levels are biofluid specific and can differ between genetic PD and idiopathic PD. Future work into using miRNAs as biomarkers will benefit from future characterization on how early these miRNA changes can be observed. Earlier diagnosis may specifically help to inform patients of beneficial lifestyle changes and or treatments that may prevent the severity of their condition. In practice, collecting miRNA expression data from several different fluid sources may also help to validate a diagnosis for a patient compared to just using one source. Some limitations of using miRNAs as biomarkers should be addressed, such as the fact that medications and medical comorbidities can also affect miRNA levels and confound their use as biomarkers for PD. For example, Levodopa, a common treatment option for PD, has been shown to affect miRNA levels in the peripheral blood compared to untreated patients (Margis et al. 2011). While many questions remain, miRNAs hold great promise as effective, non-invasive biomarkers for PD.

Considering that aging is the greatest risk factor for the development of PD and the fact that many miRNAs have been discovered that alter the lifespan of organisms, miRNAs that delay aging may be attractive targets as possible new neuroprotective genes. In addition, many miRNAs of interest that are altered in PD are downregulated with aging or in PD. Therefore, a promising therapeutic intervention in PD may come in the form of delivering neuroprotective miRNAs or replacing the miRNAs that are downregulated in patients. Those miRNAs that are upregulated in PD can also be targeted as a treatment through the introduction of Anti-miRNA Oligonucleotides (AMOs) (Lima et al. 2018). Target delivery of miRNAs or AMOs to dopamine neurons can be achieved using viral vectors. Recent data from clinical trials in neurodegenerative disorders have raised concerns regarding neuroinflammation with the use of AAVs in humans (Perez et al. 2020). Future research can address these concerns by developing new AAV capsids and or modifying promoters to make viral delivery and expression more cell-type specific. Inflammation may also be circumvented with the use of other delivery methods such as nanoparticle based carriers (Wen 2016). Lipid based carriers are perhaps one of the best suited miRNA carriers for PD treatments, as they can be formulated with complex modifications that allow the LNP to overcome the multi-step barriers found in the body and also enhance cellular uptake of the carrier (Yonezawa et al. 2020). Intracortical injections of LNPs carrying siRNA have been shown to elicit widespread significant downregulation of the mRNA target in the brain (Rungta et al. 2013). More recent research has shown that incorporating a rabies virus glycoprotein enables the LNP to be delivered much less invasively through intravenous administration and still be able pass through the BBB (Conceição et al. 2016). In the future, LNP carriers can be specifically modified to promote the efficacy of PD treatment by targeting only the cell-types of interest. Chemical modifications can also be theoretically extended to the miRNAs themselves. Previous work has gone into modifying siRNAs to improve their stability and reduce their distribution to off-targeted tissues, such as conjugates of N-acetyl-D-galactosamine (GalNAc) which have been used to improve the targeting of hepatocytes in vitro and in vivo (Nair et al. 2014). To improve LNP-miRNA based PD treatments, a similar conjugation modification may be developed to increase the selectivity of miRNA for dopamine neurons. Overall, AAVs and LNPs represent promising options for PD treatment and can be specifically tailored for non-invasive miRNA and AMO delivery.

Highlights:

• Expression studies have identified many differentially regulated miRNAs and gene targets associated with Parkinson’s Disease (PD).

• In animal models of PD, mutations in specific miRNAs have been shown to affect pathogenesis relevant to PD.

• Functional studies in animal models identify conserved signatures of miRNAs and gene targets that may regulate stress response pathways associated with PD.

• Conserved miRNAs and associated gene targets represent novel targets for future therapeutic approaches in PD.