Effect of PLA2G6 and SMPD1 Variants on the Lipid Metabolism in the Cerebrospinal Fluid of Patients with Parkinson's Disease: A Non-targeted Lipidomics Study.

Key Summary Points

Introduction

Parkinson's disease (PD) is a common degenerative disease of the central nervous system (CNS) in elderly people [ 1 ]. Genes linked to PD encompass both autosomal dominant mutations (α-synuclein, LRRK2, VPS35, and EIF4G1) and recessively inherited mutations (PARK2, PINK1, DJ-1, SYNJ1, SMPD1, and PLA2G6) [ 2 ]. Recently, lipids have been implicated in many aspects of PD pathology [ 3 , 4 ], where an imbalance in the number of lipid molecules in the brain of patients with PD exists, leading to lipid accumulation in nerve cells and consequent inflammation. Moreover, several lipidomic studies have reported PD-specific lipid alterations in the brain and plasma of patients, including alterations in the lipid composition of lipid rafts in the frontal cortex [ 3 , 5 , 6 ]. Lipid accumulation within the dopamine neurons of patients with PD was accompanied with a lack of lipids in astrocytes from the same patients [ 4 ]. Nerve cells have a limited ability to use lipids for energy, and excess fat is transported to neighboring astrocytes to avoid the buildup of toxic byproducts. Excessive lipids were also found in microglia from patients with PD compared to healthy brain tissues [ 4 ]. High levels of glycoprotein nonmetastatic melanoma protein B were found in astrocytes in response to inflammation caused by lipid accumulation in patients with PD [ 4 ]. A recent metabolomic study [ 7 ] used liquid chromatography-mass spectrometry (LC–MS) to demonstrate the predictive metabolites of PD phenotypes. Pathway enrichment analysis revealed altered lipid metabolism associated with carnitine shuttling, sphingolipid metabolism, arachidonic acid metabolism, and fatty acid biosynthesis in patients with PD [ 7 ]. Shao et al. [ 6 ] conducted a comprehensive metabolic analysis of PD plasma using LC–MS, and found significant reductions in fatty acid and caffeine metabolites, elevations in bile acids and microbiota-derived deleterious metabolites, and alterations in steroid hormones in drug-naïve patients with PD compared to healthy individuals. Brekk et al. [ 5 ] found that cell-type-specific lipid changes may play a key role in inflammation and PD progression. Another aspect of lipid dysregulation that contributes to PD pathogenesis is oxidative stress and inflammation, with pro-inflammatory lipid mediators such as platelet-activating factor mediating the progressive neuronal progression seen in PD associated with intracellular trafficking of alpha-synuclein (α-Syn) [ 3 ]. Biophysical and lipidomic studies have shown that α-Syn preferentially binds to specific lipid families, and that these lipid–protein complexes enhance its interaction with synaptic membranes, affect its oligomerization and aggregation, and interfere with the catalytic activity of cytoplasmic lipase and lysosomal lipase, thereby affecting lipid metabolism [ 3 ]. Genetic risk factors for PD are associated with lipid metabolism and function. Genes directly or indirectly involved in glycerophospholipid and sphingolipid metabolism, such as PLA2G6 and SMPD1, were associated with PD risk [ 2 , 8 ]. However, it remains unclear whether hereditary causes affect PD sleep disorders.

Exosomes were originally considered a way for cells to excrete unwanted substances, but recent studies indicate that they play a role in intercellular communication, as well as multiple physiological and pathological processes [ 9 ]. The bilayer lipid structure of exosomes is mainly composed of cytoplasmic proteins, lipids, sphingomyelin, phosphatidylcholine, phosphatidylethanolamine, phosphatidylserine, and phosphatidylinositol, which are regulated by sphingosine [ 9 ]. Phosphatidylserine can integrate proteins on the surface of exosomes and participate in the fusion process of exosomes and cell membranes [ 9 ]. Cholesterol and sphingomyelin (SM) in exosomes also form ordered lipid raft membrane microdomains, fine-tuning membrane fluidity, and altering protein localization. Exosome membrane lipid composition may alter gamma-secretase localization and activity, thereby modulating downstream pathways associated with PD pathology [ 10 ]. In addition, exosomes are related to the secretion of α-Syn in patients with PD [ 11 ]. The cerebrospinal fluid (CSF) is a biofluid of particular interest for studying biomarkers because it is closest to the brain structures and therefore could serve as an ideal source to reflect ongoing pathologic processes. Therefore, analyzing the CSF lipidomics can provide biochemical clues to CNS diseases such as PD. In this study, we analyzed the lipid profile of CSF exosomes in four patients with PD carrying PLA2G6 mutations using ultra-performance liquid chromatography–tandem mass spectrometry. We found increased exosomal content of phospholipids in glycerophospholipids and of lysophosphatidylcholine in lysophospholipids. Surface presence of melatomin receptor MTNR1A was detectable only in patients with PLA2G6 mutations. Furthermore, we evaluated the sleep pattern by polysomnography (PSG) and found longer sleep latency and reduced sleep efficiency in patients carrying PLA2G6 mutations.

Methods

Patient and CSF Acquisition

CSF specimen collection:lumbar puncture was performed as previously described [ 12 ], and 10–15 ml of CSF was collected. All participants provided informed consent for the study. Cohort 1 consisted of patients with PD carrying PLA2G6 mutations, as previously reported by our group [ 13 ]. Patient #1 had a c.1959 T > A synonymous mutation, patient #2 had a c.1966C > G Leu656Val mutation, patient #3 had a c.2077C > G Leu693Val mutation, and patient #4 had a c.1791delC His597fx69 mutation. Cohort 2 consisted of 36 patients with PD carrying the SMPD1 gene mutation c.1522G > A, p.Gly508Arg (G508R) genotype (12 GG, 12 GA, and 12AA). In this study, we collected CSF samples from patients with GG, GA, and AA genotypes, and we included 14 other patients with PD carrying WT (WT) PLA2G6 and SMPD1 genes as controls. There were 7 male and 7 female patients 65.3 ± 5.6 years old and with a disease course of 5–10 years. The patients were inpatients of the neurology department at Sir Run Run Shaw Hospital School of Medicine at Zhejiang University from 2012 to 2020. We also included 10 healthy controls, consisting of 5 male patients and 5 female patients 61.8 ± 4.9 years old. The CSF samples from the control group were obtained from surgical patients who required spinal anesthesia for non-neurological diseases. A workflow for our study is included in Supplemental Fig. 1. Our work received approval (#112,023) from the institutional ethics committees of the Shanghai General Hospital Shanghai Jiao Tong University School of Medicine. Written consent was obtained from all the subjects, and the study was conducted following the Declaration of Helsinki.

Polysomnography (PSG) Evaluation

Sleep patterns at night were recorded using a polysomnography monitor (Sandman Elite PSG; MFI Medical, San Diego, CA, USA), and the recorded sleep parameters were as previously reported [ 14 ]: (1) total sleep time (TST); (2) sleep latency (SL); (3) sleep efficiency (SE); (4) rapid eye movement sleep (REMS); (5) arousal index (AI), which refers to the number of awakenings (duration > 15 s) per hour of sleep within the time period. Analysis Manager sleep analysis software was employed to analyze these parameters.

Sleep Quality Assessment

PD sleep scale (PDSS) was used to evaluate the overall sleep quality of patients. The scale contains 15 questions related to PD sleep disorders, including the overall quality of night sleep (PDSS-1), insomnia (PDSS-2,3), periodic limb movement disorder and restless legs syndrome (PDSS-4,5), nocturnal mental conditions (PDSS-6,7), abnormal nocturia (PDSS-8,9), nocturnal movement (PDSS-10–13), post-sleep energy recovery (PDSS-14), and daytime sleepiness (PDSS-15). The Epworth sleepiness scale (ESS) was used to evaluate the patient's daytime sleepiness. The Hamilton Depression Scale (HAMD) was used to rate the depression scale, and PD quality of life questionnaire 39 (PDQ-39) was used to assess the quality of life.

Exosome Isolation by Ultracentrifugation

Exosomes were isolated and purified using a multistage ultracentrifugation method. One milliliter of CSF was mixed with an equal volume of PBS (pH 7.4) and centrifuged at 300 g for 10 min at 4 °C. The resulting supernatant was further centrifuged at 10,000 g for 30 min at 4 °C. Subsequently, the supernatant was subjected to ultracentrifugation at 100,000 g for 70 min at 4 °C using a Sorvall WX ultracentrifuge (Thermo Fisher) to pellet the exosomes. The exosome pellets were then resuspended and washed with PBS. To ensure purity, the exosomes were pipetted with PBS and filtered through a 0.44-μm filter. The structure of the exosomes was characterized using transmission electron microscopy (TEM), following previously described methods [ 12 ]. Additionally, nanoparticle tracking analysis (NTA) was conducted as previously outlined [ 15 ].

Exosome Immunogold electron micrography (EM)

In this procedure, a falvar carbon-coated nickel grid was placed with the coated side in contact with exosomes. The samples were blocked using 0.1% bovine serum albumin (BSA; Sigma) in PBS for 30 min. Immunostaining was performed using a 1:10 dilution of anti-MTNR1A antibody (ab203038; Abcam) for 30 min. A 10-nM gold-labeled secondary antibody was used for subsequent incubation. To fix the samples, they were treated with 2.5% glutaraldehyde for 10 min. Subsequently, embedding was carried out using 2% uranyl acetate, 0.13% methylcellulose, and 0.4% uranyl acetate. After removing excess liquid, the samples were air-dried for 5 min. Finally, electron microscopy was employed to visualize the EM micrographs.

Western Blot

To extract the total protein, exosomes were lysed on ice for 20 min, followed by centrifugation at 10,000 rpm for 30 min. The concentration of total protein was determined using a BCA kit (Biyuntian Biotechnology). For Western blot analysis, each well was loaded with 20 μg of protein. The SDS-PAGE separation program consisted of 80 V for 30 min and 120 V for 60 min for efficient separation. Protein transfer to a PVDF membrane was performed using a current of 200 mA for 90 min. Subsequently, the PVDF membrane was blocked with 5% BSA blocking solution for 1 h at room temperature on a shaker. After three washes with TBST, the PVDF membrane was incubated overnight with specific antibodies, including TSG101 (ab30871; Abcam), anti-melatonin receptor 1A (MTNR1A) (ab203038; Abcam), and anti-melatonin receptor 1B (MTNR1B) (ab203346; Abcam). Following this, the PVDF membrane was incubated with horseradish peroxidase-labeled goat anti-rabbit/mouse IgG (Biyuntian) for 1 h. Finally, the membranes were imaged using the BioRad imaging system.

Extraction of Lipid Metabolites

Exosomes were disrupted by ultrasonication at 100 Hz for 5 min. Then, 100 µL of acetonitrile (chromatographic grade; Thermo Fisher) was added, and the solution was centrifuged at 12,000 rpm for 10 min at 4 °C. Next, 60 µL of the supernatant was loaded into glass autosampler vials, and 10 µL of the remaining supernatant of each group was used as a quality control group.

Chromatography

The chromatographic column used was the Kinetex XB-C18 column (100 mm × 2.1 mm, 2.6 µm; Phenomenex). The column temperature was 30 °C, and the mobile phase consisted of (A) 0.1% (v/v) formic acid aqueous solution and (B) 0.1% (v/v) acetonitrile solution of formic acid. The flow rate was 0.3 mL/min, and the sample chamber temperature was 8 °C. The gradient elution program was as follows: 0–1.0 min, 10% B; 1.0–8.0 min, 10% B to 80% B; 8.0–12.0 min, 80% B; 12.0–15.0 min, 80% B to 10% B. The injection volume was 60 µL. All operations were carried out using a high-performance liquid chromatography system (UltiMate 3000; Thermo Fisher).

Mass Spectrometry

A TripleTOF5600LC-MS system (SCIEX, Framingham, MA, USA) was used. The mass scanning range was m/z: 100–1000. The sheath gas pressure was 380 kPa, the auxiliary gas pressure was 380 kPa, and the curtain gas pressure was 170 kPa. The atomization temperature was 600 °C. Time-of-flight mass spectrometry TOF–MS scan-IDA-production scan composite mode was used, and the TOF–MS primary pre-scan and triggered secondary scan IDA ion accumulation time were 250 ms and 100 ms, respectively. The secondary trigger conditions were multiple mass deficit and dynamic background subtraction. The solution cluster voltage was ± 80 V, and the collision energy superposition was (35 ± 15) eV.

Data analysis

After converting the obtained raw data through MarkView software, peak identification, peak alignment, subtraction of solvent peaks, impurity peaks, and noise filtering were performed. The Shortcut to PeakView technology was used to promote fragment ions through the chemical intelligent peak matching algorithm. During the analysis process, the MMCD ( http://mmcd.nmrfam.wisc.edu/ ) and HMDB ( http://www.hmdb.ca ) databases were used to search, and the 3-D matrix was obtained from the exported data table. The preprocessed data matrix was imported into SIMCA-P14.1 for multivariate statistical analysis. Differential metabolites were screened according to the variable projection importance value (VIP value > 1) and independent sample t test ( p < 0.05) in the orthogonal partial least squares discriminant analysis (OPLS-DA) model.

Statistical analysis

Prism GraphPad was used for data analysis. The K–S normality test confirmed that all data were normally distributed. The data were presented as (X ± SEM). Independent t tests were used to compare data between groups. Count data were compared using the χ 2 test. A p value of less than 0.05 indicated statistical significance.

Results

PLA2G6 Mutations Cause Phenotypic Deregulation of Exosomes in the CSF of patients with PD

To determine the characteristics of exosomes isolated from CSF, TEM analysis revealed rounded structures with a size of approximately 60–150 nm from all of the samples, similar to exosomes as previously described [ 12 ]. However, when compared with exosomes from healthy controls and patients with PD carrying the WT PLA2G6 genotype, we observed larger-diameter membrane vesicles that were isolated from patients with PD carrying c.1966C > G, Leu656Val, c.2077C > G, Leu693Val, and c.1791delC, His597fx69 mutations (Fig. 1 A). Furthermore, NTA was employed to verify these results and demonstrated that the number of exosomes was reduced in the CSF of patients with PD carrying PLA2G6 mutations compared to control and patients with PD carrying the WT PLA2G6 genotype (healthy control: 5.44 ± 1.22 × 10 10 particles/ml; PD with WT PLA2G6: 4.26 ± 1.02 × 10 10 particles/ml; patient #1: 1.33 ± 0.39 × 10 10 particles/ml; patient #2: 2.03 ± 0.61 × 10 10 particles/ml; patient #3: 2.49 ± 0.47 × 10 10 particles/ml; patient #4: 2.39 ± 0.43 × 10 10 particles/ml). Additionally, the average diameter of exosomes derived from the CSF of patients with PD carrying PLA2G6 mutations was increased compared to control and patients with PD carrying the WT PLA2G6 genotype (healthy control: 85 ± 31 nm; PD with WT PLA2G6: 87 ± 37 nm; patient #1: 103 ± 26 nm; patient #2: 155 ± 30 nm; patient #3: 176 ± 39 nm; patient #4: 255 ± 46 nm) (Fig. 1 B). In summary, these results indicate that functional mutations in PLA2G6 lead to exosome structural alteration and decreased secretion, which may be implicated in PD development.

PLA2G6 Mutations Significantly Affect the Lipid Composition of CSF-Derived Exosomes in Patients with PD

PLA2G6 variants are associated with phospholipid metabolism [ 16 ]. We analyzed the effect of PLA2G6 mutations on the lipid composition of exosomes isolated from the CSF of patients with PD. The extracted exosome lipids were subjected to high-resolution mass spectrometry, and the results showed no significant difference in the constituents of glycerophospholipids, sphingomyelins, glycosphingolipids, sterol lipids, lysophospholipids, and glycolipids of exosomes derived from the CSF of patients with PD carrying PLA2G6 mutations compared to those from healthy controls and patients with PD carrying WT (WT) PLA2G6 (Fig. 2 A). However, the PC components from glycerophospholipids were higher in patient#4 (52%), patient#3 (39%), patient#1 (36%); patient#2 (37%) compared to those in healthy control (33%) and PD with WT PLA2G6 (34%) (Fig. 2 B). In addition, the PE component was lower in patient#4 (23%), patient#3 (28%), patient#1 (30%); patient#2 (29%) compared to those in healthy control (32%) and PD with WT PLA2G6 (31%) (Fig. 2 B). There were no obvious changes in PA, PG, PI, and PS among the patients (Fig. 2 B). Moreover, the LPC abundance of lysophospholipids of CSF exosomes was lower in patient#4 (8%), patient#3 (14%), patient#1 (18%); patient#2 (15%) compared to those in healthy control (19%) and PD with WT PLA2G6 (20%) (Fig. 2 C). The LPE and LPI components were significantly higher in patient#4 (LPE: 46% and LPI: 46%), patient#3 (LPE: 43% and LPI: 43%), patient#1 (LPE: 41% and LPI: 41%); patient#2 (LPE: 42% and LPI: 43%) compared to those in healthy control (LPE: 41% and LPI: 39%) and PD with WT PLA2G6 (LPE: 41% and LPI: 39%) (Fig. 2 C). These results indicate that lipid metabolism, especially PC, PE, and LPC in exosomes, was significantly affected in patients with PD carrying PLA2G6 mutations.

Melatonin Receptor was Enriched on CSF Exosome Surface in Patients with PD Carrying PLA2G6 Mutations

Melatonin (MT) is a hormone that regulates the circadian rhythm and promotes sleep via its receptors, MTNR1A and MTNR1B [ 17 ]. We analyzed the surface protein markers of exosomes isolated from the CSF of patients with PD carrying PLA2G6 mutations. Exosomes were subjected to immunogold EM with an antibody against the N-terminals of MTNR1A and MTNR1B. The results showed the surface presence of MTNR1A (Fig. 3 A) in CSF exosomes isolated from patients with PD carrying PLA2G6 mutations. However, MTNR1B was not detectable in the same samples. Additionally, there was no detectable level of MTNR1A and MTNR1B expression in exosomes isolated from healthy controls and patients with PD carrying WT PLA2G6. Western blot analysis confirmed significantly higher expression of MTNR1A protein (Fig. 3 B) in CSF-derived exosomes from patient#2, patient#3, and patient#4 compared with healthy controls and patients with PD carrying WT PLA2G6. No MTNR1B was detected in CSF exosomes from any of the patients (Fig. 3 B). These data indicate that PLA2G6 deregulated exosomes shown enrichment of melatonin receptor on the surface, which may further influence sleep disorders during PD progression.

SMPD1 Mutations Cause Deregulation of Exosome Secretion in the CSF of Patients with PD

To determine the effect of sphingomyelin phosphodiesterase 1 (SMPD1) genotypes on the characteristics of exosomes in the CSF, TEM was performed. The results demonstrated rounded structures with a size of approximately 60–150 nm (Fig. 4 A) that were isolated from all the CSF samples of patients with PD carrying SMPD1 genotypes c.1522GG, c.1522GA, and c.1522AA. Furthermore, NTA was used to analyze the exosome size and concentration. There was no significant difference in exosome size between patients with PD carrying different SMPD1 c.1522 genotypes (c.1522GG: 106 ± 27 nm; c.1522GA: 102 ± 19 nm; and c.1522AA: 104 ± 23 nm). However, the concentration of exosomes was significantly decreased in the CSF of patients with PD carrying SMPD1 c.1522GA and c.1522AA genotypes compared with those with SMPD1 c.1522GG (c.1522GG: 4.76 ± 0.69 × 10 10 particles/ml; c.1522GA: 3.51 ± 0.49 × 10 10 particles/ml; and c.1522AA: 2.14 ± 0.47 × 10 10 particles/ml) (Fig. 4 B). These results indicate that SMPD1 variation affected the extracellular vesicular formation and secretion, suggesting the potential importance in PD pathogenesis.

SMPD1 Genotypes Significantly Affect Lipid Compositions of CSF-Derived Exosomes in patients with PD

SMPD1 is an acid sphingomyelinase that can convert sphingomyelin into ceramide [ 18 ]. We analyzed the impact of SMPD1 genotypes on the lipid composition of CSF exosomes in patients with PD. High-resolution mass spectrometry showed no significant differences in glycosphingolipids, sterol lipids, lysophospholipids, and glycolipids between different SMPD1 genotypes. However, the constituents of glycerophospholipids were substantially reduced in patients with PD carrying c.1522AA (53%) compared to c.1522GA (60%) and c.1522GG (63%) (Fig. 5 A). There was no significant change in the identified sub-compositions (PC, PE, PA, PG, PI, and PS) of glycerophospholipids (Fig. 5 B). However, the abundance of sphingomyelins was significantly higher in patients with PD carrying c.1522AA (25%) compared to c.1522GA (16%) and c.1522GG (9%), indicating that SMPD1 mutation (G508R) significantly affected sphingomyelin metabolism. There was no differential abundance of lysophospholipid composition (LPC, LPE, and LPI) of CSF exosomes among different SMPD1 genotypes, c.1522GG, c.1522GA, and c.1522AA (Fig. 5 C).

Ceramide Levels were Significantly Reduced in Patients with PD Carrying Mutant SMPD1

To determine the impact of different SMPD1 genotypes on ceramide levels, exosome-derived total lipids were analyzed from patients with PD carrying varying genotypes. The total ceramide level was lower in CSF exosomes of patients with SMPD1 c.1522AA than in those with c.1522GG and c.1522GA (c.1522GG vs. c.1522GA, p = 0.0011; c.1522GG vs. c.1522AA, p = 0.0004; Fig. 6 ). These results suggest that SMPD1 G508R mutation leads to suppressed ceramide metabolism, which may involve in the sleep disorder during PD progression.

Different Types of Sleep Disorders Between Patients with PD Carrying PLA2G6 Mutations and SMPD1 Mutations

All patients underwent PSG evaluation to compare sleep patterns between groups. The results showed no significant differences in TST, REMS, and AI per hour between patients with PD carrying PLA2G6 mutations and those carrying WT PLA2G6 (Table 1 ). However, sleep latency was significantly longer in patient #1 (25.33 ± 4.56 min), patient #2 (63.21 ± 13.24 min), patient #3 (77.32 ± 12.44 min), and patient #4 (75.41 ± 23.41 min) compared to patients with PD carrying WT PLA2G6 (26.89 ± 5.21 min). Sleep efficiency was also significantly reduced in patient #1 (64.21 ± 11.74%), patient #2 (52.23 ± 15.64%), patient #3 (51.87 ± 14.82%), and patient #4 (53.74 ± 15.21%) compared to patients with PD carrying WT PLA2G6 (63.56 ± 12.36%). HAMD, ESS, PDSS, and PDQ-39 scores were also evaluated between groups. The results showed that HAMD, ESS, and PDQ-39 scores were worse, while PDSS score did not change in patients with PD carrying PLA2G6 mutations compared to those carrying WT PLA2G6 (Table 2 ).

Patients with PD carrying different SMPD1 genotypes were evaluated using PSG to compare sleep disorders between groups. There were no significant differences in TST, SL, SE, and AI between patients with PD carrying different SMPD1 genotypes (Table 3 ). However, REMS was significantly shortened in patients with PD carrying SMPD1 c.1522AA (2.51 ± 2.14 min) compared to SMPD1 c.1522GA (4.37 ± 2.74 min) and SMPD1 c.1522GG (7.32 ± 3.57 min) (Table 3 ). HAMD, ESS, PDSS, and PDQ-39 scores were not significantly different between patients with PD carrying different SMPD1 genotypes (Table 4 ).

Discussion

Several phospholipase A2 (PLA2) family members are known to regulate the metabolism of glycerophospholipids, and they can decompose the second ester bond (sn-2) of glycerophospholipids to produce fatty acids and lysophospholipids [ 19 ]. Mutations in the PLA2G6 gene are associated not only with dystonia-parkinsonism but also with typical, nondystonic parkinsonism [ 20 ]. In addition, PLA2G6 gene mutations are associated with the onset of early-onset parkinsonism [ 21 – 23 ]. PD carrying PLA2G6 mutations has a high degree of genetic and clinical heterogeneity, and the clinical phenotype of PD varies with different mutation sites, mutation types, and different races [ 24 – 26 ]. The PLA2G6 gene encodes a calcium-independent phospholipase A2-β protein, which contains different functional sites, such as the ankyrin repeat region, catalytic region, and other regions. Changes in cytoplasmic composition affect membrane homeostasis and cause neuronal damage [ 27 ]. The changes in the amino acid sequence and protein structure domains caused by mutations at different sites of PLA2G6 are different. For example, Engel et al. [ 16 ] found that the pathogenic mutations of the PLA2G6 gene (p.A341T, p.G517C) reported in patients with INAD and NBIA could significantly reduce enzyme activity. Our group previously reported that the PLA2G6 enzyme activity was almost completely lost after the (His597fx69) frameshift mutation [ 13 ]. Therefore, when there is a PLA2G6 gene mutation, the decrease in iPLA2-β enzyme activity leads to the apoptosis of dopaminergic neurons, and then the clinical manifestations of PD [ 28 ]. The results of our current study demonstrated abnormal lipid metabolism from CSF-derived exosomes from patients with PD carrying PLA2G6 mutations. PC components from glycerophospholipids were higher in patient#4: c.1791delC, His597fx69 and patient#3: c.2077C > G, Leu693Val, patient#1: c.1959 T > A, synonymous; patient#2: c.1966C > G, Leu656Val, compared to those in healthy control and PD with WT PLA2G6 (Fig. 2 B). Moreover, the LPC abundance of lysophospholipids of CSF exosomes was lower in patients with PD carrying PLA2G6 mutations compared to those in healthy control and PD with WT PLA2G6 (Fig. 2 C). Therefore, the changes in lipidome composition and function of CSF exosomes during the pathological stage of PD would influence lipid membrane fluidity of exosomes will endow different PD pathogenic genes with different biological properties of exosomes.

MT is a hormone that regulates the circadian rhythm and promotes sleep [ 17 ]. Its secretion is regulated by the SCN and follows the circadian rhythm. The secretion increases at night and decreases during the day. In addition, MT has multiple functions such as scavenging free radicals, anti-oxidation, anti-inflammatory response, and neuroprotection [ 17 ]. Results from our study demonstrate that the sleep latency was significantly longer in patient#1 (25.33 ± 4.56 min), patient#2 (63.21 ± 13.24 min), patient#3 (77.32 ± 12.44 min), and patient#4 (75.41 ± 23.41 min) compared to those in patients with PD carrying WT PLA2G6 (26.89 ± 5.21 min). The sleep efficiency was significantly reduced in patient#1 (64.21 ± 11.74%), patient#2 (52.23 ± 15.64%), patient#3 (51.87 ± 14.82%), and patient#4 (53.74 ± 15.21%) compared to those in patients with PD carrying WT PLA2G6 (63.56 ± 12.36%). Furthermore, there was surface presence of MTNR1A (Fig. 3 A) on the exosome surface in patient#2, patient#3, and patient#4. However, there was no presence of MTNR1A and MTNR1B expression in exosomes isolated from healthy control and PD with WT PLA2G6. We did not detect MT receptor expression on CSF exosomes from patients with PD carrying SMPD1 G508R (Supplemental Fig. 2). Benleulmi-Chaachoua et al. [ 29 ] found that MT receptors inhibit DA reuptake, suggesting that MT has an effect on the DA energy system. Therefore, the abnormality of MT rhythmic secretion may aggravate the progression of PD. MT acts mainly through high-affinity G protein receptors MT1 and MT2, which can act as nuclear receptors to exert anti-oxidant effects [ 30 ]. There is a high density of MT1 and MT2 receptors in the suprachiasmatic nucleus, and their expression is synchronized with the circadian rhythm [ 31 ]. The activation of MT1 and MT2 receptors can exert anti-oxidant and cytoprotective effects by controlling glutathione peroxidase, glutathione reductase, and superoxide dismutase [32; 33]. The secretion pattern of MT did not change, but compared with the normal group, the phase of the MT secretion pattern was advanced [ 34 ]. It was further found that the phase advancement of MT in patients with PD may be due to levodopa treatment producing dopaminergic effects in the central nervous system, not the disease itself [ 35 ]. Another study demonstrated that MT production and MTergic receptors MT1 and MT2 were present in patients with PD. The expression of MT was significantly lower than that of the normal group, and the secretion of MT lost the normal rhythm [34; 36]. Therefore, our results indicate that functional mutations in PLA2G6 lead to structural alteration and secretion of exosomes. Further, exosome lipid deregulation caused by PLA2G6 mutations may affect its cargo, which may further influence circadian rhythm dysfunction in PD. Future study will focus on the linkage between MTNR1A in the exosome and sleep disturbance.

About 30–50% of patients with PD are accompanied by sleep behavior disorder (RBD) [ 37 ]. RBD symptoms often appear in the early stage of PD dyskinesia symptoms [ 37 ]. PD-RBD patients have a disordered sleep structure, light sleep, and an increased frequency of awakenings, resulting in a fragmented sleep structure, reduced sleep efficiency, and reduced total sleep time, leading to excessive daytime sleepiness. The results from our study demonstrated that there were no significant differences in TST, SL, SE, and AI between patients with PD carrying different SMPD1 genotypes c.1522GG, c.1522GA, and c.1522AA (Table 3 ). However, the REMS was significantly shortened in patients with PD carrying SMPD1 c.1522AA (2.51 ± 2.14 min) compared with SMPD1 c.1522GA (4.37 ± 2.74 min) and SMPD1 c.1522GG (7.32 ± 3.57 min) (Table 3 ). HAMD, ESS, PDSS, and PDQ-39 scores were not significantly different between patients with PD carrying different SMPD1 genotypes (Table 4 ). ASMase, encoded by the SMPD1 gene, and its downstream metabolite ceramide, can trigger signal transduction events [ 18 ], change the membrane curvature by changing the lipid composition, cause drastic changes in cell membrane properties, and participate in the secretion and release of EVs. Loss of the SMPD1 gene causes Niemann–Pick disease, a lysosomal storage disorder [ 38 ]. Mutations in SMPD1 in patients with PD carrying reduced ASMase levels predate the onset of PD by 3.5–5.8 years [ 39 ]. Some studies have shown that ceramide synthesis is involved in the formation of extracellular vesicles in the lumen of multivesicular bodies independent of the ESCRT biogenesis pathway, and that the vesicles are released in the form of exosomes after the fusion of multivesicular bodies with the plasma membrane. In a study [ 40 ] of the sorting mechanism of exosomes from an oligodendrocyte cell line (Oli-Neu), exosome release was significantly reduced after the treatment of Oli-Neu cells with NSMase inhibitors. The biogenesis of exosomes occurs in sphingomyelin-rich membrane domains, and they coalesced into larger domains by forming ceramide microdomains to promote vesicle budding. The tapered structure of ceramides induces a spontaneous negative curvature of the endosomal membrane, favoring the formation of endosomes inside multivesicular bodies. Subsequent studies showed that the administration of exogenous ceramide dose-dependently increased the number of exosomes released from multiple myeloma cells [ 41 ]. The sphingomyelin synthase inhibitor D609 inhibits the conversion of ceramide to sphingomyelin leading to increased ceramide and enhanced exosome secretion [ 42 ]. Therefore, ceramide is important for exosome release. The results of our study demonstrate that, although there was no significant difference in the size of exosomes between patients with PD carrying different SMPD1 c.1522 genotypes, the concentration of exosomes was significantly decreased in the CSF of patients with PD carrying SMPD1 c.1522GA and c.1522AA genotypes compared to those with SMPD1 c.1522GG (c.1522GG: 4.76 ± 0.69 × 10 10 particles/ml; c.1522GA: 3.51 ± 0.49 × 10 10 particles/ml; and c.1522AA: 2.14 ± 0.47 × 10 10 particles/ml) (Fig. 4 B). These results indicate that SMPD1 variations affect exosome formation and secretion capability, suggesting an important role in Parkinson's disease. Furthermore, the abundance of glycerophospholipids was substantially reduced in patients with PD carrying c.1522AA (c.1522AA, 53% vs. c.1522GA, 60% vs. c.1522GG, 63%) (Fig. 5 A). The abundance of sphingomyelins was significantly higher in patients with PD carrying c.1522AA (25%) compared to c.1522GA (16%) and c.1522GG (9%), indicating that sphingomyelin metabolism was significantly inhibited by the SMPD1 mutation (G508R). More importantly, we found that the total ceramide level was lower in CSF exosomes of SMPD1 c.1522AA than in SMPD1 c.1522GG and c.1522GA (c.1522GG vs. c.1522GA, p = 0.0011; c.1522GG vs. c.1522AAm p = 0.0004, Fig. 6 A, B). These results indicate that the SMPD1 G508R mutation leads to suppressed ceramide metabolism. Ceramides are a type of sphingolipid composed of long-chain bases of sphingosine and fatty acids. Ceramide is a type of phospholipid with ceramide as the backbone, mainly including ceramide phosphorylcholine and ceramide phosphoethanolamine. Phospholipids are the main component of cell membranes, and ceramide is the main component of the intercellular matrix. Liu et al. [ 43 ] reported that microglia can regulate sleep/wake state switching through ceramide signaling to regulate thalamic reticular nucleus neurons. They found that the expression level of ceramide in the brain tissue of mouse cortex was significantly lower during the waking period than during the sleeping period [ 43 ], showing that ceramides increased with enhanced sleep drive. Ceramides are positively associated with sleep drive, which may affect waking stability by promoting sleep [ 43 ]. In addition, microglia could regulate wakefulness by transmitting ceramide signals to aTRN neurons. Sphingomyelinase promotes the formation of ceramide from sphingomyelin, and then the conical structure of ceramide disturbs the membrane curvature and fluidity, which is beneficial to microvesicle germination and multivesicular body formation to promote EV release [ 43 ]. Therefore, targeting the SMPD1 gene is expected to provide new ideas and strategies for the prevention and treatment of PD by regulating the release of exosomes and ceramide metabolism.

There are several limitations in the current study. Recruitment for patients with PD carrying PLA2G6 mutation remain challenging, results should be further validated with independent investigation. Lipidomic study lacks a unified sample processing. Other types of extracellular vesicles need lipid composition analysis.

Conclusion

We have provided for the first time independent findings of changes in lipid deregulation of exosomes in genetic risk cases of PD (PLA2G6 and SMPD1 variant carriers) and the presence of sleep disorders in these patients with PD. Although our sample size was limited, our study claims that lipid deregulation of exosomes may affect PD-related sleep disorders, which will offer new insights and potential therapeutic targets for various types of sleep disorders in Parkinson's patients.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary file1 (PDF 160 KB)

Acknowledgements

We thank the participants of the study.

Author Contributions

Yongang Li and GuiKai Ji designed the research. Yongang Li, GuiKai Ji, Mengjia Lian, Xuan Liu performed research. Yongang Li, GuiKai Ji, and Ying Xu collected subjects and analyzed data. GuiKai Ji supervised the study. Yongang Li, GuiKai Ji, and Yaxing Gui contributed to the manuscript writing.

Funding

This research was supported by Taizhou Science and Technology Plan Project (22ywb131, 22ywb137), Zhejiang Provincial Health Commission (2023RC307), Shanghai Science and Technology Program/Natural Science Foundation of Shanghai (22ZR1449800), National Natural Science Foundation of China (81401038), and Zhejiang Province Medicine Health General Research Program (2020KY602). The rapid service fee was funded by the authors.

Data Availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Conflict of Interest

Yongang Li, GuiKai Ji, Mengjia Lian, Xuan Liu, Ying Xu, and Yaxing Gui declare that they have no conflicts of interest.

Ethical Approval

The current study had approval from the Institutional Ethics Committee of Shanghai General Hospital Shanghai Jiao Tong University School of Medicine. Each patient provided written informed consent before study. The study was performed in accordance with the 1964 Declaration of Helsinki and its later amendments.

References

1. Kobylecki C. Update on the diagnosis and management of Parkinson's disease. Clin Med. 2020;20:393–398. doi: 10.7861/clinmed.2020-0220.

2. Fais M, Dore A, Galioto M, Galleri G, Crosio C, Iaccarino C. Parkinson's disease-related genes and lipid alteration. Int J Mol Sci. 2021;22:7630. doi: 10.3390/ijms22147630.

3. Alecu I, Bennett SAL. Dysregulated Lipid Metabolism and Its Role in alpha-Synucleinopathy in Parkinson's Disease. Front Neurosci. 2019;13:328. doi: 10.3389/fnins.2019.00328.

4. Hallett PJ, Engelender S, Isacson O. Lipid and immune abnormalities causing age-dependent neurodegeneration and Parkinson's disease. J Neuroinflammation. 2019;16:153. doi: 10.1186/s12974-019-1532-2.

5. Brekk OR, Honey JR, Lee S, Hallett PJ, Isacson O. Cell type-specific lipid storage changes in Parkinson's disease patient brains are recapitulated by experimental glycolipid disturbance. Proc Natl Acad Sci USA. 2020;117:27646–27654. doi: 10.1073/pnas.2003021117.

6. Shao Y, Li T, Liu Z, Wang X, Xu X, Li S, Xu G, Le W. Comprehensive metabolic profiling of Parkinson's disease by liquid chromatography-mass spectrometry. Mol Neurodegener. 2021;16:4. doi: 10.1186/s13024-021-00425-8.

7. Sinclair E, Trivedi DK, Sarkar D, Walton-Doyle C, Milne J, Kunath T, Rijs AM, de Bie RMA, Goodacre R, Silverdale M, Barran P. Metabolomics of sebum reveals lipid dysregulation in Parkinson's disease. Nat Commun. 2021;12:1592. doi: 10.1038/s41467-021-21669-4.

8. Kinghorn KJ, Castillo-Quan JI, Bartolome F, Angelova PR, Li L, Pope S, Cocheme HM, Khan S, Asghari S, Bhatia KP, Hardy J, Abramov AY, Partridge L. Loss of PLA2G6 leads to elevated mitochondrial lipid peroxidation and mitochondrial dysfunction. Brain J Neurol. 2015;138:1801–1816. doi: 10.1093/brain/awv132.

9. Mager ELASI, Breakefield XO, Wood MJ. Extracellular vesicles: biology and emerging therapeutic opportunities. Nature reviews. Drug Discov. 2013;12:347–357. doi: 10.1038/nrd3978.

10. Roncarati R, Sestan N, Scheinfeld MH, Berechid BE, Lopez PA, Meucci O, McGlade JC, Rakic P, D'Adamio L. The gamma-secretase-generated intracellular domain of beta-amyloid precursor protein binds Numb and inhibits Notch signaling. Proc Natl Acad Sci USA. 2002;99:7102–7107. doi: 10.1073/pnas.102192599.

11. Yuyama K, Sun H, Mitsutake S, Igarashi Y. Sphingolipid-modulated exosome secretion promotes clearance of amyloid-beta by microglia. J Biol Chem. 2012;287:10977–10989. doi: 10.1074/jbc.M111.324616.

12. Gui Y, Liu H, Zhang L, Lv W, Hu X. Altered microRNA profiles in cerebrospinal fluid exosome in Parkinson disease and Alzheimer disease. Oncotarget. 2015;6:37043–37053. doi: 10.18632/oncotarget.6158.

13. Gui YX, Xu ZP, Wen L, Liu HM, Zhao JJ, Hu XY. Four novel rare mutations of PLA2G6 in Chinese population with Parkinson's disease. Parkinsonism Relat Disord. 2013;19:21–26. doi: 10.1016/j.parkreldis.2012.07.016.

14. Alatriste-Booth V, Rodriguez-Violante M, Camacho-Ordonez A, Cervantes-Arriaga A. Prevalence and correlates of sleep disorders in Parkinson's disease: a polysomnographic study. Arq Neuropsiquiatr. 2015;73:241–245. doi: 10.1590/0004-282X20140228.

15. Li Y, Gu J, Mao Y, Wang X, Li Z, Xu X, Chen H, Gui Y. Cerebrospinal fluid extracellular vesicles with distinct properties in autoimmune encephalitis and herpes simplex encephalitis. Mol Neurobiol. 2022;59:2441–2455. doi: 10.1007/s12035-021-02705-2.

16. Engel LA, Jing Z, O'Brien DE, Sun M, Kotzbauer PT. Catalytic function of PLA2G6 is impaired by mutations associated with infantile neuroaxonal dystrophy but not dystonia-parkinsonism. PLoS ONE. 2010;5:e12897. doi: 10.1371/journal.pone.0012897.

17. Gunata M, Parlakpinar H, Acet HA. Melatonin: A review of its potential functions and effects on neurological diseases. Revue neurologique. 2020;176:148–165. doi: 10.1016/j.neurol.2019.07.025.

18. Deshpande D, Gupta SK, Sarma AS, Ranganath P, Jain SJ, Sheth J, Mistri M, Gupta N, Kabra M, Phadke SR, Girisha KM, Dua Puri R, Aggarwal S, Datar C, Mandal K, Tilak P, Muranjan M, Bijarnia-Mahay S, Rama Devi AR, Tayade NB, Ranjan A, Dalal AB. Functional characterization of novel variants in SMPD1 in Indian patients with acid sphingomyelinase deficiency. Human Mutat. 2021;42:1336–1350. doi: 10.1002/humu.24263.

19. Taketo MM, Sonoshita M. Phospolipase A2 and apoptosis. Biochem Biophys Acta. 2002;1585:72–76. doi: 10.1016/s1388-1981(02)00326-8.

20. Paisan-Ruiz C, Bhatia KP, Li A, Hernandez D, Davis M, Wood NW, Hardy J, Houlden H, Singleton A, Schneider SA. Characterization of PLA2G6 as a locus for dystonia-parkinsonism. Ann Neurol. 2009;65:19–23. doi: 10.1002/ana.21415.

21. Lu CS, Lai SC, Wu RM, Weng YH, Huang CL, Chen RS, Chang HC, Wu-Chou YH, Yeh TH. PLA2G6 mutations in PARK14-linked young-onset parkinsonism and sporadic Parkinson's disease. Am J Med GeneT Part B NEUROPSYCHIATRIC genet. 2012;159B(183):91. doi: 10.1002/ajmg.b.32012.

22. Tian JY, Tang BS, Shi CH, Lv ZY, Li K, Yu RL, Shen L, Yan XX, Guo JF. Analysis of PLA2G6 gene mutation in sporadic early-onset parkinsonism patients from Chinese population. Neurosci Lett. 2012;514:156–158. doi: 10.1016/j.neulet.2012.02.078.

23. Yoshino H, Tomiyama H, Tachibana N, Ogaki K, Li Y, Funayama M, Hashimoto T, Takashima S, Hattori N. Phenotypic spectrum of patients with PLA2G6 mutation and PARK14-linked parkinsonism. Neurology. 2010;75:1356–1361. doi: 10.1212/WNL.0b013e3181f73649.

24. Morgan NV, Westaway SK, Morton JE, Gregory A, Gissen P, Sonek S, Cangul H, Coryell J, Canham N, Nardocci N, Zorzi G, Pasha S, Rodriguez D, Desguerre I, Mubaidin A, Bertini E, Trembath RC, Simonati A, Schanen C, Johnson CA, Levinson B, Woods CG, Wilmot B, Kramer P, Gitschier J, Maher ER, Hayflick SJ. PLA2G6, encoding a phospholipase A2, is mutated in neurodegenerative disorders with high brain iron. Nat Genet. 2006;38:752–754. doi: 10.1038/ng1826.

25. Ramanadham S, Song H, Hsu FF, Zhang S, Crankshaw M, Grant GA, Newgard CB, Bao S, Ma Z, Turk J. Pancreatic islets and insulinoma cells express a novel isoform of group VIA phospholipase A2 (iPLA2 beta) that participates in glucose-stimulated insulin secretion and is not produced by alternate splicing of the iPLA2 beta transcript. Biochemistry. 2003;42:13929–13940. doi: 10.1021/bi034843p.

26. Smalheiser NR, Swanson DR. Calcium-independent phospholipase A2 and schizophrenia. Arch Gen Psychiatry. 1998;55:752–753. doi: 10.1001/archpsyc.55.8.752.

27. Malley KR, Koroleva O, Miller I, Sanishvili R, Jenkins CM, Gross RW, Korolev S. The structure of iPLA(2)beta reveals dimeric active sites and suggests mechanisms of regulation and localization. Nat Commun. 2018;9:765. doi: 10.1038/s41467-018-03193-0.

28. Lee LY, Ong WY, Farooqui AA, Burgunder JM. Role of calcium-independent phospholipase A2 in cortex striatum thalamus cortex circuitry-enzyme inhibition causes vacuous chewing movements in rats. Psychopharmacology. 2007;195:387–395. doi: 10.1007/s00213-007-0912-y.

29. Benleulmi-Chaachoua A, Hegron A, Le Boulch M, Karamitri A, Wierzbicka M, Wong V, Stagljar I, Delagrange P, Ahmad R, Jockers R. Melatonin receptors limit dopamine reuptake by regulating dopamine transporter cell-surface exposure. Cell Mol Life Sci CMLS. 2018;75:4357–4370. doi: 10.1007/s00018-018-2876-y.

30. Jockers R, Delagrange P, Dubocovich ML, Markus RP, Renault N, Tosini G, Cecon E, Zlotos DP. Update on melatonin receptors: IUPHAR review 20. Br J Pharmacol. 2016;173:2702–2725. doi: 10.1111/bph.13536.

31. Reiter RJ. Pineal melatonin: cell biology of its synthesis and of its physiological interactions. Endocr Rev. 1991;12:151–180. doi: 10.1210/edrv-12-2-151.

32. Choi SI, Dadakhujaev S, Ryu H, Im Kim T, Kim EK. Melatonin protects against oxidative stress in granular corneal dystrophy type 2 corneal fibroblasts by mechanisms that involve membrane melatonin receptors. J Pineal Res. 2011;51:94–103. doi: 10.1111/j.1600-079X.2011.00866.x.

33. Franco DG, Markus RP. The cellular state determines the effect of melatonin on the survival of mixed cerebellar cell culture. PLoS ONE. 2014;9:e106332. doi: 10.1371/journal.pone.0106332.

34. Videnovic A, Noble C, Reid KJ, Peng J, Turek FW, Marconi A, Rademaker AW, Simuni T, Zadikoff C, Zee PC. Circadian melatonin rhythm and excessive daytime sleepiness in Parkinson disease. JAMA Neurol. 2014;71:463–469. doi: 10.1001/jamaneurol.2013.6239.

35. Fertl E, Auff E, Doppelbauer A, Waldhauser F. Circadian secretion pattern of melatonin in de novo parkinsonian patients: evidence for phase-shifting properties of l-dopa. J Neural Transm Parkinson's Dis Dementia Sect. 1993;5:227–234. doi: 10.1007/BF02257677.

36. Adi N, Mash DC, Ali Y, Singer C, Shehadeh L, Papapetropoulos S. Melatonin MT1 and MT2 receptor expression in Parkinson's disease. Med Sci Monit Int Med J Exp Clin Res. 2010;16:BR61–BR67.

37. Romenets SR, Gagnon JF, Latreille V, Panniset M, Chouinard S, Montplaisir J, Postuma RB. Rapid eye movement sleep behavior disorder and subtypes of Parkinson's disease. Mov Disord. 2012;27:996–1003. doi: 10.1002/mds.25086.

38. Simonaro CM, Park JH, Eliyahu E, Shtraizent N, McGovern MM, Schuchman EH. Imprinting at the SMPD1 locus: implications for acid sphingomyelinase-deficient Niemann-Pick disease. Am J Hum Genet. 2006;78:865–870. doi: 10.1086/503750.

39. Alcalay RN, Mallett V, Vanderperre B, Tavassoly O, Dauvilliers Y, Wu RYJ, Ruskey JA, Leblond CS, Ambalavanan A, Laurent SB, Spiegelman D, Dionne-Laporte A, Liong C, Levy OA, Fahn S, Waters C, Kuo SH, Chung WK, Ford B, Marder KS, Kang UJ, Hassin-Baer S, Greenbaum L, Trempe JF, Wolf P, Oliva P, Zhang XK, Clark LN, Langlois M, Dion PA, Fon EA, Dupre N, Rouleau GA, Gan-Or Z. SMPD1 mutations, activity, and alpha-synuclein accumulation in Parkinson's disease. Mov Disord. 2019;34:526–535. doi: 10.1002/mds.27642.

40. Trajkovic K, Hsu C, Chiantia S, Rajendran L, Wenzel D, Wieland F, Schwille P, Brugger B, Simons M. Ceramide triggers budding of exosome vesicles into multivesicular endosomes. Science. 2008;319:1244–1247. doi: 10.1126/science.1153124.

41. Cheng Q, Li X, Wang Y, Dong M, Zhan FH, Liu J. The ceramide pathway is involved in the survival, apoptosis and exosome functions of human multiple myeloma cells in vitro. Acta Pharmacol Sin. 2018;39:561–568. doi: 10.1038/aps.2017.118.

42. Dinkins MB, Dasgupta S, Wang G, Zhu G, Bieberich E. Exosome reduction in vivo is associated with lower amyloid plaque load in the 5XFAD mouse model of Alzheimer's disease. Neurobiol Aging. 2014;35:1792–1800. doi: 10.1016/j.neurobiolaging.2014.02.012.

43. Liu H, Wang X, Chen L, Chen L, Tsirka SE, Ge S, Xiong Q. Microglia modulate stable wakefulness via the thalamic reticular nucleus in mice. Nat Commun. 2021;12:4646. doi: 10.1038/s41467-021-24915-x.

Associated Data

Supplementary Materials

Supplementary file1 (PDF 160 KB)

Data Availability Statement

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.