Parkinson’s disease (PD) is a common chronic progressive neurodegenerative disease in middle-aged and older people. Its main clinical manifestations are bradykinesia, resting tremor, myotonia and abnormal posture, and gait (Lee et al., 1993). About 1% of people over 65 years old suffer from PD. The main pathological features are selective and progressive degeneration of dopamine (DA) neurons in the substantia nigra compacta of the midbrain, a significant decrease of DA content in the striatum, and eosinophilic inclusion bodies–Lewy body in the residual neurons, which are mainly composed of $\alpha$-synuclein ($\alpha$-syn) (Bentwich et al., 2005). After decades of research, the diagnosis and treatment of PD are still not ideal (Kalia and Lang, 2015). An autopsy is a gold standard with the patients grouped according to pre-death clinical information. At present, the clinical diagnosis of PD is mainly through the evaluation of medical history, physical examination, imaging examination, and evaluation of the response of levodopa and other drugs, but symptoms and signs occur only when up to 80% of dopamine deficiency occur in the pars compacta of the substantia nigra (Emamzadeh and Surguchov, 2018; Lotankar et al., 2017). Only when the symptoms are fully developed can specific imaging abnormalities be observed (Ghaemi, 2002). Therefore, searching for biomarkers that can distinguish PD from other diseases with higher sensitivity and specificity is necessary, and fully tracking the progression of PD is preferable. Only in this way can we diagnose PD early.

Multiple system atrophy (MSA) is a sporadic, adult-onset, and progressive neurodegenerative disease (Fanciulli et al., 2019). The onset age is about 60 years old, but it also occurs in both young and old age (Batla et al., 2018; Wenning et al., 1997). The clinical manifestation of MSA is complex, which is the combination of autonomic nerve failure, Parkinson’s disease, and ataxia. The pathological feature of MSA is the formation of cytoplasmic inclusion bodies composed of misfolded $\alpha$-syn in glial cells (Gilman et al., 1998). MSA is currently divided into 2 clinical subtypes, including Parkinson’s syndrome as a prominent clinical subtype called the MSA-P type and cerebellar ataxia for more severe cases called the MSA-C type (Fanciulli et al., 2019; Gilman et al., 1998).

Both MSA and PD are caused by pathological aggregation of $\alpha$-syn (Yamasaki et al., 2019). Although clinical criteria for MSA have been established, prenatal diagnosis is still difficult. Previous autopsy studies showed that the diagnostic accuracy was 29%-86% (Hughes et al., 2002; Osaki et al., 2002; Wenning et al., 1997). Clinical diagnosis may also be problematic when MSA coexists with other neurodegenerative diseases with the same clinical characteristics, such as Alzheimer’s disease, Louis-related disease (Ozawa, 2004), or tauopathy (Uchikado et al., 2006). Early in the disease, the MSA-P type can be difficult to distinguish from PD (Litvan et al., 1997; Osaki et al., 2002; Seppi et al., 2005a). Data show that, clinically, approximately 10%-30% of MSA in the early stage are misdiagnosed as PD (Poewe and Wenning, 2002). The prognosis and treatment response of MSA and PD are different. Patients with MSA-P have higher morbidity and mortality than PD patients. They usually respond poorly to levodopa and develop faster dyskinesia (Muller et al., 2000; Quinn and Marsden, 1993; Seppi et al., 2005b). Early diagnosis can avoid mistaking MSA-P patients into PD pharmacological and neurosurgical tests (Lang and Lozano, 1998a, b). Therefore, it is particularly important to study the biomarkers that can accurately distinguish and diagnose MSA and PD in the early stage of the disease.

MiRNAs are a small cluster of endogenous non-coding single-stranded RNA regulatory molecules with a length of about 20-22 nucleotides, which are encoded by endogenous genes and cannot be transcribed into proteins. They inhibit the translation of target mRNAs or promote mRNAs degradation by binding to the 3’-untranslated region (3’-UTR) of target mRNAs, thus inhibiting the expression of target genes at the post-transcriptional level (Vivekanantham et al., 2015). MiRNAs are highly conservative. Their expression is sequential and tissue-specific. They participate in almost all pathological and physiological processes of mammals and play an important role in the occurrence and development of many diseases (Cao and Zhen, 2018; Hu et al., 2019; Junn et al., 2009; Vishnoi and Rani, 2017). They also play a crucial role in regulating cell proliferation, differentiation, growth, metabolism, stress response, apoptosis, and heterochromatin formation (Kim et al., 2007; Santosh et al., 2009; Tutar, 2015). Post-transcriptional regulation is an important process in the pathogenesis of PD (Dorval et al., 2012; Filipowicz et al., 2008; Hombach and Kretz, 2016; Patop et al., 2019). The role of miRNAs in the pathogenesis of PD has attracted increasing attention from researchers (Chang et al., 2017; Kim et al., 2019; Leggio et al., 2017).

MiRNA expression was monitored by qPCR in PD patients and healthy controls. Eight of 224 pre-miRNAs are highly expressed in the midbrain (Kim et al., 2007). Among them, pre-miR-133b was the most downregulated in the PD group. However, these miRNAs found in human brain tissue cannot be directly used as biomarkers for clinical use, and using brain tissue samples from PD patients to diagnose PD is impossible. Early diagnosis requires not only the identification of disease specificity but also minimally invasive biomarkers.

The low stability of RNA molecules limits their ability to be used as biomarkers. However, serum miRNAs are very stable (Jin et al., 2013). In addition to their high sensitivity and specificity, serum miRNAs are also convenient and inexpensive and have other advantages as biomarkers (Zhou et al., 2012). Geekiyanage et al. (2012) found that some serum miRNAs have similar changes in the brain. Cogswell et al. (2008) found the presence of AD-specific miRNAs in cerebrospinal fluid (CSF), suggesting that some miRNAs produced by the diseased tissue and cells can enter the CSF and then enter the peripheral blood through the circulatory system. The results of these studies provide a basis for serum miRNAs as PD biological markers.

We previously identified 7 differentially expressed miRNAs in PD patient and control sera (miR-30c, miR-31, miR-141, miR-146b-5p, miR-181c, miR-214 and miR-193a-3p), and the levels of these 7 miRNAs in serum were significantly lower in the PD patient group than in the control group (Dong et al., 2016). In this paper, we compare the difference in the expression of these serum miRNAs between 23 de novo PD cases and 23 MSA patients in order to analyze the clinical application value of the serum miRNAs in the diagnosis of PD and MSA.

Neural miRNAs may be stably packaged in microvesicles and conveyed to blood and other peripheral biofluids (urine, saliva, breast milk) (Alexander et al., 2015; Cai et al., 2018; Haqqani et al., 2013; Moldovan et al., 2013; Valadi et al., 2007). Some miRNAs have similar changes in serum and brain tissue, suggesting that changes in serum miRNAs can be used to predict changes in miRNAs in the central nervous system (CNS). For example, miR-153 and miR-223 are significantly downregulated in both nerves and circulation (Cressatti et al., 2019). MiR-9 has also been reported to be decreased in both blood and brain tissue (Jin et al., 2013). Distinguishing PD from MSA by detecting the content and types of miRNAs in serum is of great significance. Previous studies have almost always used external parameters, and increasing evidence now exists that internal references are unstable in testing serum miRNAs. In contrast, we used a short-chain RNA that was not found in humans and animals and did not affect human serum as an external reference.

Our previous work confirmed that 7 miRNAs levels (miR-30c, miR-31, miR-141, miR-146b-5p, miR-214, miR-181c, and miR-193a-3p) in the serum of patients with PD were lower compared with the controls (Dong et al., 2016). Based on the achievements of the previous work, we designed a way to further test the expression levels of the 7 miRNAs in early PD, MSA, and control groups. The results showed that the levels of miR-214 in the serum of PD patients were distinctly higher than those in the serum of the healthy control group (Fig. 1A). These results are different from our previous results that the levels of miR-214 in the serum of PD patients were reduced compared with the normal controls (Dong et al., 2016). The reason may be that the patients in the previous study had a few years of onset (the average was 6 years), but the de novo patients in this study had less than 1 year of onset. The serum miRNA levels may change dynamically in patients with early-onset and late-onset. The miR-214 levels were upregulated at the beginning of PD. This may be due to the existence of compensation in the body. We speculate that the levels decreased again as patients became older and more ill (Fig. 1D). This result provides a clinical basis for miR-214 to become an early biomarker of PD.

MiR-214 was downregulated in the mesencephalon of PD mice and MPP${}^{+}$-treated SH-SY5Y neuroblastoma cells (Wang et al., 2015). This result is in line with our previous results. In the sera of patients with PD who had PD for several years, miR-214 was significantly downregulated (Dong et al., 2016). In summary, miR-214 may be upregulated in the early PD stage (incidence within 1 year) (Fig. 1A). As the disease progresses, the level of miR-214 may decrease again (Fig. 1D). These speculations need further verification. The study demonstrated that miR-214 could restrain autophagy and the apoptosis of hippocampal neurons as a functional miRNA (Zhang et al., 2016). MiR-214 is related to the nervous system and may affect the pathogenesis of PD.

Other miRNAs in 7 miRNAs were also reported to have altered levels in autopsy brains of PD patients compared with controls. The decreased expression levels of miRNA-30 and miRNA-193 were verified in autopsy samples of PD patients (Briggs et al., 2015). MiR-181 was identified to be associated with PD from the brain and blood samples (Chatterjee and Roy, 2017). However, how these miRNAs change in MSA is still unknown. We further compared the expression differences of 7 serum miRNAs between MSA patients and healthy controls. The relative expression of 3 (miR-141, miR-193a-3p and miR-30c) of the 7 miRNAs were significantly different from the control group (Fig. 1B). By comparison, miR-141 was most important for identifying normal controls and MSA (Fig. 2A). The difference in the expression levels of 7 serum miRNAs between PD and MSA patients indicated that miR-31, miR-141, miR-181c, miR-193a-3p, and miR-214 were significantly lower in MSA than in PD (Fig. 1C). The specificity and sensitivity of these 5 miRNAs showed that they had diagnostic significance (Fig. 2B).

Our experiments further analyzed the correlations between these miRNAs expression levels and PD clinical scales. The results showed that the expression of miR-214 was negatively correlated with anxiety and PDNMS. The more severe the anxiety and nonmotor symptoms were, the lower the level of miR-214. In our study, we selected patients who did not take anti-PD drugs within 1 year of onset. Their H-Y stage and intelligence damage were relatively low. The increased level of miR-214 may be a compensatory response in the body at the beginning of the disease. With the increase of age, the miR-214 levels decreased gradually. The correlations between miRNA expression and clinical scales need to be further studied in an amplified sample size.

To date, 18 PD-related chromosomal sites have been reported, PARK1-PARK18. Some new genes have been proposed, such as GIGYF2(Park11), VPS35, and Pitx3 (Chen et al., 2013; Lautier et al., 2008; Liu et al., 2012). The genes related to MSA include SQSTM1 and SLC1A4. SNCA is a gene related to both MSA and PD.

Other groups have reported that miR-214 can target SNCA, a key protein of PD, and reduce its protein expression (Wang et al., 2015). Altered levels of miR-214 may affect SNCA protein levels and thus affect disease progression. For the MSA-related genes, SNCA was predicted to be the target gene of miR-141 (Table 2). In our work, miR-141 was the most important for identifying normal controls and MSA. These predictions have yet to be validated by more in vivo and in vitro experiments.

We have proved that miR-214 can be used as a molecular marker for PD diagnosis and its early diagnosis. Five microRNAs (miR-31, miR-141, miR-214, miR-181c, and miR-193a-3p) can be used for the antidiastole of PD and MSA.

Study conception, design, and supervision: WGL, HD, YH, and HC; history collection, assessment scales: PH and YC; sample collection: LTL, CYY, and LY; implementation of the experiment: JHY and PH; image drafting and revision: JHY; data analysis, statistical analysis: JHY and HD; manuscript writing: JHY; manuscript revising: HD, LWG, HZ, ZJZ, HZ, and QHY. All authors consented to the final version of this manuscript.

The ethics committee approved the study of the Affiliated Brain Hospital of Nanjing Medical University, and it was completed based on the ethical standards established by the 1964 Declaration of Helsinki and its later amendments. All participants provided informed consent.

Funding support from the National Key R&D Program of China [grant number 2017YFC1310300 and 2017YFC1310302]; the National Natural Science Fund [grant number 81571348]; the Jiangsu Natural Science Foundation [grant number BK20151077]; the National Key Research and Development Plan [grant number 2016YFC1306600]; the National Natural Science Fund [grant number 81701671]; and the Henan Province Key Research and Development and Promotion Special [grant number 182102310163] is greatly appreciated.

The authors declare no conflict of interest.