This article is mentioned in:

Introduction

Neonatal hyperbilirubinemia primarily occurs in term newborns and in >84% of late preterm newborns (1). Extreme neonatal hyperbilirubinemia may lead to the accumulation of bilirubin in the brain stem nuclei, cerebellum and basal ganglia, and subsequently cause acute bilirubin encephalopathy (ABE) (2). ABE is a serious disease that can cause death or a lifelong neurodevelopmental disorder, known as chronic bilirubin encephalopathy or kernicterus (3). Although the incidence of ABE has declined, the reported incidence rate in developed countries in 2009–2018 ranged from 0.4–2.7 cases per 100,000 births (4,5). Moreover, the reported incidence rates of ABE in Asia, the Middle East and Africa have been reported to be notably higher than in Western developed countries (6–8).

ABE is a clinical bilirubin-induced neurological dysfunction (BIND) characterized by poor feeding, lethargy, hypotonia and opisthotonus (9,10). The neurotoxicity is due to the diffusion of unconjugated bilirubin (UCB) into brain cells after traversing the blood-brain barrier (10). High concentrations of UCB exert neurotoxic effects by inducing a number of cellular and molecular events (e.g., neuronal excitotoxicity, mitochondrial energy failure, activation of apoptotic or necrosis pathways) (11). The in vitro exposure of neurons to UCB has been identified to induce cellular macroscopic changes, such as decreased dendritic and axonal arborization, neurite extension and ramification, and cell proliferation (12). Non-neuronal cells, such as astrocytes, microglia and oligodendrocytes, have also demonstrated sensitivity to UCB, which triggered the secretion of inflammatory mediators, including tumor necrosis factor (TNF) and interleukin (IL)-1 (13,14), upregulated the activity of matrix-metalloproteinases 2 and 9 (15) and increased caspase 3-mediated apoptosis (16). In addition, cultured vascular endothelial cells of the blood-brain barrier and the blood-cerebrospinal fluid (CSF) barrier were observed to exhibit upregulated P-glycoprotein/ABCB1 and downregulated multidrug resistance-associated protein 1/ABC, C1 expression levels, respectively, upon the exposure to high concentrations of bilirubin (17). Moreover, bilirubin-induced endoplasmic reticulum stress was previously demonstrated to contribute to inflammatory processes and apoptosis in neuronal cells, which indicated the involvement of endoplasmic reticulum stress in UCB-induced neurotoxicity (18).

However, despite numerous cell and animal experiments, the basic mechanisms of ABE are not fully understood. Over previous years, there has been an increasing interest in extracellular microvesicles/exosomes (MV/E), which are released from a number of cells, including glial cells, neurons and astrocytes in the brain (19). MV/E are membrane-bound nanovesicles that serve as important mediators of intercellular communication by transporting molecular signals, including proteins, mRNA and microRNAs, between cells (20,21). High-throughput proteomics technology has been used to identify underlying mechanisms associated with the initiation and development of numerous types of cerebral disease (22–24). However, to the best of our knowledge, few studies have investigated the molecular mechanism of bilirubin-induced neurotoxicity based on proteomics techniques or extracellular MV/E (12,25,26). Therefore, focusing on the proteomics of MV/E may elucidate the mechanisms of bilirubin-induced neurological neurotoxicity. The aim of the present study was to obtain protein profiles of extracellular vesicles from the CSF of patients with ABE. These results may help determine the pathogenesis of ABE and provide novel perspectives for the early diagnosis and treatment of ABE.

Materials and methods

Participant characteristics and CSF collection

The study protocol was approved by the Ethics Committee of Affiliated First People's Hospital of Chenzhou, Southern Medical University, University of South China (Chenzhou, China). Written informed consent was obtained from parents/caregivers of all subjects prior to the start of the study. All experimental procedures were performed in accordance with the Declaration of Helsinki (27).

All subjects in the present study were full-term newborns(n=30) recruited between January 2017 and November 2019 from the Department of Neonatology of the Affiliated First People's Hospital of Chenzhou, Southern Medical University. Patient characteristics are presented in Table I. Patients with ABE were diagnosed according to the clinical BIND score; BIND scores of 1–3, 4–6 and 7–9 are indicative of mild, moderate and severe ABE, respectively (28). The inclusion criteria for patients with ABE in the present study were: Newborns, aged ≤14 days who required exchange transfusion and had clinical BIND scores of >4 (3,28). Patients with ABE were divided into moderate (n=10) and severe groups (n=10). The control group (n=10) comprised full-term newborns who required lumbar puncture to exclude sepsis and meningitis, but who were not diagnosed with sepsis or meningitis due to normal CSF and peripheral blood test results. A total of four patients in the control group were diagnosed with pneumonia; six were diagnosed with enteritis. The demographic and clinical characteristics of the subjects are presented in Table I. Newborns with congenital cerebral malformation or secondary brain injury due to other causes (such as hypoglycemic encephalopathy, hypoxic-ischemic encephalopathy, intracranial hemorrhage, epilepsy, sepsis, meningitis or inherited metabolic diseases) were excluded from the present study.

Table I. Characteristics of the study population.

Table I.

Characteristics of the study population.

	Patients with acute bilirubin encephalopathy
Variables	Moderate group	Severe group	Control group	P-value
Number	10	10	10
Female/male	4/6	3/7	5/5
Gestational age, weeks	38.60±1.40	38.20±1.70	38.40±1.34	0.930
Birth weight, g	3,513.50±188.18	3,498.30±135.13	3,505.40±240.86	0.985
Age at admission, days	5.70±0.95	5.50±1.08	6.40±1.57	0.248
Peak serum TB levels, µmol/l	485.40±20.30	587.50±57.50	77.21±19.70	<0.001
Bilirubin-induced neurological dysfunction score at admission	4.80±0.91	8.50±0.71		<0.001
Routine test of cerebrospinal fluid
Number of cells, 106/l	7.60±2.01	7.50±1.58	6.90±1.96	0.666
Protein concentration, mg/l	810.50±94.96	790.26±107.76	745.90±131.77	0.289
Peripheral blood test
White blood cells, 109/l	8.99±1.03	9.11±1.16	8.54±1.44	0.553
C-reactive protein, mg/l	2.41±1.040	2.46±1.15	2.52±0.78	0.970
Procalcitonin, ng/ml	0.18±0.09	0.23±0.12	0.19±0.15	0.617
Cause of hyperbilirubinemia
Hemolysis, yes/no	3/7	4/6
Idiopathic, yes/no	1/9	1/9
Glucose-6-phosphate 1-dehyrodgenase deficiency, yes/no	6/4	5/5
Treatment	ET + PT	ET + PT + intravenous albumin

[i] Bilirubin-induced neurological dysfunction scores of 1–3, 4–6 and 7–9 are indicative of mild, moderate and severe acute bilirubin encephalopathy, respectively (28). Data are presented as the mean ± SD. TB, total bilirubin; PT, phototherapy; ET, exchange transfusion.

CSF samples were obtained at the time of diagnosis of ABE, while samples from controls were obtained prior to the administration of antibiotics at admission. CSF samples (2.5 ml) were collected via lumbar puncture and centrifuged at 2,000 × g for 20 min at 4°C and stored at −80°C until analysis.

Isolation of MV/E via co-precipitation

A volume of 1 ml supernatant was pipetted into an Eppendorf tube and mixed with 330 µl Ribo™ exosome isolation reagent (cat. no. C10120; Guangzhou Ribobio Co., Ltd.). The tube was subsequently placed in a refrigerator at 4°C for 30 min, then centrifuged at 15,000 × g for 2 min at 4°C. The resultant precipitate was the crude MV/E-enriched fraction, which was resuspended and further diluted with1 ml 1X PBS to achieve the detectable concentration (106 particles/ml) for further analysis.

Nanoparticle tracking analysis (NTA)

In order to characterize MV/E, NTA was performed using the NanoSight NS300 instrument (Malvern Panalytical). The crude MV/E-enriched fraction were diluted with 1 ml 1X PBS to achieve a suitable concentration (106 particles/ml), and then injected into the detection chamber (flow rate, 35 µl/second; temperature 23.1–23.7°C) equipped with a 405 nm (purple) laser. The parameters were adjusted to the best condition according to the manufacturer's instructions, and samples were measured three times. Data were analyzed using NTA 3.2 Dev Build 3.2.16 software (Malvern Panalytical) and SOP Standard Measurement. The NTA technology combines a video imaging system and laser light scattering microscopy to obtain the nanoparticle size distribution of samples in liquid suspension (29). Brownian motion of each particle was measured and analyzed individually. A light beam is used to illuminate the particles in the sample. As the particles scatter light and undergo Brownian motion, a scientific CMOS camera recorded the path of each particle to determine the mean velocity and diffusivity. The particle concentration and size distribution were calculated according to a modified Stokes-Einstein relationship (30).

Protein extraction, reduction, alkylation, digestion and isobaric tagging for relative and absolute quantification (iTRAQ)-4 plex labeling

For protein extraction, all pooled MV/E samples were removed from ice. Each sample was added to ×10 the volume of protein lysis buffer (8 M urea, 1% SDS, 1X protease inhibitor cocktail) with an ultrasonic cell disruptor (25 kHz) for 2 min in an ice bath and then allowed to react for 30 min on ice. Subsequently, precooled 100% acetone (acetone:sample ratio, 5:1) was added and incubated at −20°C overnight to precipitate the protein. The sample was centrifuged for 30 min at 12,000 × g at 4°C, and the precipitate was cleaned using 90% acetone and centrifuged again for 30 min at 12,000 × g at 4°C. Then, the supernatant was discarded, and the precipitate was fully dissolved in protein lysis buffer and centrifuged for 30 min at 12,000 × g at 4°C.

Proteins were quantified via BCA assay. For protein reduction and alkylation, protein lysis buffer was added to 50 µg protein samples to a volume of 150 µl. Then, 10 mM Bond-Breaker™ Tris (2-carboxyethyl) phosphine solution (Thermo Fisher Scientific, Inc.) was added to the sample and incubated for 60 min at 37°C. Following the incubation, 10 μl solution (40 mM iodoacetamide in 40 mM NH4HCO3) was added to the sample and incubated for 40 min at 37°C in the dark. Precooled 100% acetone (acetone: sample volume ratio, 6:1) was added to each tube for precipitation at −20°C for 4 h. Then, the sample was centrifuged at 10,000 × g for 20 min at 4°C and the precipitate was completely dissolved in 100 µl triethylammonium bicarbonate (TEAB; 100 mM).

For subsequent protein tryptic digestion, 2 µl trypsin (0.5 µg/ml) was added to each sample. Each sample was vortexed briefly, covered with parafilm to prevent evaporation and incubated overnight at 37°C. Subsequently, samples were desalted using a Sep-Pak C18 cartridge (particle size, 55–105 µm) (Waters Corporation) according to the manufacturer's instructions, and then vacuum dried. For iTRAQ-4 plex labeling, 100 μg digested peptides were dissolved in 0.5 M TEAB and labeled with the iTRAQ reagents 4-plex kit (cat. no. 43374321; Shanghai AB SCIEX Analytical Instrument Trading Co.), according to the manufacturer's protocol. iTRAQ reagents 114, 115 and 116 from the iTRAQ reagents 4-plex kit were used to label digested peptides of the control, moderate and severe groups, respectively. After reacting for 2 h at room temperature, 40 µg labeled peptides in each group were mixed together in a tube and vacuum concentrated for high-performance liquid chromatography analysis (HPLC).

Ultimate HPLC (UPLC)

In order to decrease sample complexity, first-dimensional separation of the iTRAQ-labeled mixed peptides was performed using the ACQUITY UPLC system (Waters Corporation). The peptide mixture (120 µg) was reconstituted withsolvent A (20 mM ammonium formate; 2% acetonitrile; pH 10.0) to 25 µl and loaded onto the ACQUITY UPLC BEH C18 column (grain size, 1.7 µm; 2.1×150.0 mm; temperature, 37°C; Waters Corporation) through an autosampler (temperature, 10°C). The peptides were separated at 200 µl/min for 37 min via a linear gradient, which was established by mixing solvents A and B (20 mM ammonium formate; 80% acetonitrile; pH 10.0) as follows: 0–5 min, 0.0–3.8% solvent B in solvent A; 6–16 min, 4–43% solvent B in solvent A; 17–37 min, 100% solvent B in solvent A. The ultraviolet detection wavelength was set at 214 nm. A total of 5 fractions were collected based on peak area, then vacuum dried and stored at −80°C for further liquid chromatography with tandem mass spectrometry (LC-MS/MS) analysis.

Reverse-phase LC-MS/MS

Second-dimensional reverse phase LC-MS/MS was performed at room temperature to separate and identify the digested samples (1 µg/fraction). The chromatographic separation was performed using the EASY-nLC 1200 system on a C18 column (75 µm × 25 cm; Thermo Fisher Scientific, Inc.). The mobile phase contained solvents A (2% acetonitrile with 0.1% formic acid) and B (80% acetonitrile with 0.1% formic acid). The samples were resuspended in 40 µl solvent A to a concentration of 0.5 µg/μl and were fractionated sequentially. The total fractionation time was 120 min at a flow rate of 300 nl/min via a linear gradient profile, which was established by mixing solvent A and solvent B as follows: 0–63 min, 0–23% solvent B in solvent A; 64–88 min, 23–48% solvent B in solvent A; 89–120 min, 100% solvent B in solvent A.

A Q-Exactive hybrid quadrupole-Orbitrap mass spectrometer (Thermo Fisher Scientific, Inc.) equipped with an EASY-nLC™ 1200 nanoflow LC system (Thermo Fisher Scientific, Inc.) was used for qualitative analysis using electrospray ionization in the positive mode. The conditions for MS were: Source temperature, 350°C; spray voltage, −3 kv; nebulizer gas pressure, 35 psi; and nebulizer nitrogen flow rate, 10 l/min. The data-dependent MS acquisition mode was used to select the top 20 most abundant ions for subsequent analysis. The mass spectrum scanning range (m/z) was set at 350–1,300 m/z and the mass spectrum resolution was 70,000. For MS/MS, the top 20 intense precursor ions were sequentially subjected to higher energy collision dissociation. The fragment ions were detected at a mass resolution of 17,500. Dynamic exclusion was set for 18 sec. Thermo Xcalibur™ 4.0 (Thermo Fisher Scientific, Inc.) software was used for data acquisition. A total of three biological repeats was performed to minimize the experimental variation.

Data analysis and bioinformatics

The raw data was sifted using Proteome Discoverer™ 2.1 software (Thermo Fisher Scientific, Inc.) and was searched against the UniProt Homo sapiens database (uniprot-proteome-UP000005640-Homo sapiens-20180626-73045s.fasta) using Mascot software (version 2.3.0; Matrix Science, Inc.). The search parameters were as follows: i) Dynamic modification: Oxidation (M), acetyl (Protein N-Terminus), iTRAQ4plex (Y); ii) static modification: iTRAQ4plex (K), iTRAQ4plex (N-Terminus), carbamidomethyl (C); iii) enzyme, trypsin; iv) maximum missed cleavage sites, 2; v) precursor mass tolerance, 20 ppm; vi) fragment mass tolerance, 0.05 Da; and vii) requirements for protein identification: Two peptides or ≥1 unique peptide to match. The cut-off values for differentially expressed proteins (DEPs) were set at fold change >1.2 or <0.83 and P<0.05. All P-values were corrected for multiple hypotheses testing by controlling the false discovery rate (FDR). P<0.05 and FDR <0.05 were considered to indicate a statistically significant difference.

Gene Ontology (GO) consortium (geneontology.org/) was used for functional term annotation and enrichment analysis using the UniProt accession numbers of the identified DEPs. The Kyoto Encyclopedia of Genes and Genomes (KEGG) database (kegg.jp/) was used for signaling pathway enrichment analysis. In addition, a protein-protein interaction (PPI) network of the DEPs was constructed using the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) software (version 10.5; string-db.org/) (confidence score ≥0.4). Fisher's exact values and Q-values (corrected using FDR) <0.05 were considered significant for the GO functional term and KEGG signaling pathway enrichment analysis.

Western blotting validation

A total of four DEPs [S100A9, S100A7, lactoferrin (LTF) and α defensin 1 (DEFA1)] were selected for further validation. The selection criteria for validation were: i) Fold change >1.5 or <0.67 and P<0.05; ii) DEPs were upregulated or downregulated in both the ABE and control groups; iii) results of the bioinformatics analysis (GO functional annotation and KEGG signaling pathways analysis of DEPs); iv) potential association with ABE pathogenesis, based on existing knowledge; and v) not previously studied in patients with ABE.

Proteins from harvested MV/E of CSF were extracted using protein extraction buffer [50 Tris-HCl (pH, 6.8), 100 mM DTT, 2% SDS, 0.1% bromophenol blue, 10% glycerol] and quantified via BCA assay. Proteins samples were separated by 12% SDS-PAGE at room temperature using 30 mA constant current electrophoresis and 50 μg protein was loaded per lane. The separated proteins were subsequently transferred onto PVDF membranes and blocked with 5% non-fat milk in TBST (0.1% Tween-20) for 60 min at room temperature. The membranes were washed three times with TBST for 5 min and incubated with primary antibodies at 4°C overnight. Following the primary antibody incubation, the membranes were washed three times with TBST for 5 min and incubated for 60 min at room temperature with a horseradish peroxidase-conjugated secondary antibody (1:1,000) as follows: Goat anti-mouse (cat. no. SC-3697; Santa Cruz Biotechnology, Inc.), goat anti-rabbit (cat. no. G-21234; Thermo Fisher Scientific, Inc.), rabbit anti-goat (cat. no. 81-1620; Thermo Fisher Scientific, Inc. The chromogenic reaction was developed using a chemiluminescent substrates kit (Amresco, LLC). Images were visualized using a KodakImage Station 4000R digital imaging system (Kodak). The Quantity One software (version 4.6.7; Bio-Rad Laboratories, Inc) was used for densitometric analysis. The primary antibodies (all 1:1,000) were as follows: Monoclonal anti-CD9 and anti-LTF (cat. nos. sc-59140 and sc-53498, respectively; both Santa Cruz Biotechnology, Inc.), polyclonal anti-DEFA1 (cat. no. PA5-19228; Thermo Fisher Scientific, Inc.) and polyclonal anti-S100A7 and anti-S100A9 (cat. nos. PA5-75689 and PA1-46489, respectively; both Thermo Fisher Scientific, Inc.).

Statistical analysis

Statistical analysis was performed using the SPSS statistical package (version 20.0; IBM Corp.). Continuous variables are presented as the mean ± SD of three experimental repeats. One-way ANOVA and Bonferroni's post hoc test were performed to evaluate the statistical differences between multiple group comparisons. P<0.05 was considered to indicate a statistically significant difference.

Results

NTA of MV/E from patients with ABE and controls

Representative particle concentration distribution and size scattering intensity analysis of MV/E samples is presented in Fig. 1. The particle size of MV/E from patients with ABE and controls ranged from 75–200 nm in diameter and the peak concentration of the particle of MV/E from patients with ABE and controls was at 103 nm in diameter. The three curves overlapped, which indicated that the three groups of samples were homogeneous and that there were few impurities, which suggested the successful isolation of MV/E from the CSF.

Figure 1. MV/E particle size distribution and scattering intensity analysis using nanoparticle tracking analysis. Concentration particle size curve: Representative particle size distribution of the MV/E samples obtained from patients with acute bilirubin encephalopathy and controls. MV/E, microvesicle/exosome.

Qualitative results of dysregulated proteins

A total of 182,834 MS/MS spectra were obtained via iTRAQ-based proteomics and 25,245 peptide spectrum matches (PSMs) were matched to 5,913 peptides. A total of 634 proteins were identified and 291 dysregulated proteins were selected for further study (Fig. 2; Table SI). Of the 291 proteins, 186 DEPs (75 upregulated and 111 downregulated proteins) were identified between the severe and moderate groups (Table SII). A total of 119 DEPs (37 upregulated and 82 downregulated proteins) were identified between the severe and control groups (Table SIII) and 194 DEPs (92 up- and 102 downregulated proteins) were identified between the moderate and control groups (Table SIV). Certain DEPs [serum amyloid A-1 protein (SAA1), amyloid-β precursor protein (APP), immunoglobulin-like domains (immunoglobulin heavy constant γ4 and immunoglobulin heavy variable 2–5), complement components (C4B and C5), S100A9, S100A7, DEFA1 and LTF] were identified between the ABE and control groups (Table SI).

Figure 2. The numbers of spectrums, peptides and proteins identified in the microvesicles/exosomes of controls and patients with acute bilirubin encephalopathy using tandem mass spectrometry.

GO annotation and functional term enrichment analysis of DEPs

GO annotation of all DEPs was categorized according to biological process (BP), molecular function (MF) and cellular components (CC). The majority of DEPs were associated with ‘cellular process’, ‘immune system process’ and ‘metabolic process’ (Fig. 3A) and the identified MFs indicated that the DEPs were primarily involved in ‘binding’ and ‘catalytic activity’ (Fig. 3B). In addition, the DEPs were revealed to be primarily located in the ‘cell’ and ‘extracellular region’ (Fig. 3C).

Figure 3. Gene Ontology annotation analysis of differentially expressed proteins in the microvesicles/exosomes of patients with acute bilirubin encephalopathy. (A) Biological process. (B) Molecular function. (C) Cellular component.

GO enrichment analysis revealed that the majority of the DEPs were primarily enriched in ‘immune system process’, ‘response to stimulus’ and ‘biological regulation’ in the BPs. In addition, the MFs of the DEPs were primarily associated with ‘molecular transducer activity’, ‘molecular function regulator’ and ‘transcription regulator activity’, whilst the DEPs were identified to be primarily located in ‘extracellular region part’ and ‘cell part’ (Fig. 4A).

Figure 4. Enrichment analysis of DEPs. (A) Gene Ontology functional term enrichment analysis of DEPs in patients with ABE (B) KEGG signaling pathway enrichment analysis of DEPs in patients with ABE. DEP, differentially expressed protein; ABE, acute bilirubin encephalopathy; KEGG, Kyoto Encyclopedia of Genes and Genomes CC cellular component; MF, molecular function; BP, biological process; CP, cellular processes; HD, human diseases; EIP, environmental information processing; OS, organismal systems.

KEGG signaling pathway enrichment analysis

KEGG signaling pathway enrichment analysis was performed to profile the functional differences of the proteomes of MV/E isolated from the CSF of patients with ABE and the controls. Significantly enriched signaling pathways of patients with ABE included ‘Ras signaling pathway’, ‘PI3K-AKT signaling pathway’, ‘MAPK signaling pathway’, ‘cytokine-cytokine receptor interaction’, ‘Rap1 signaling pathway’, ‘mTOR signaling pathway’ and ‘NF-κB signaling pathway’ (Fig. 4B). Other enriched signaling pathways were identified, including ‘HIF-1 signaling pathway’, ‘mTOR signaling pathway’, ‘NF-κB signaling pathway’, ‘Neuroactive ligand-receptor interaction’, and ‘ECM-receptor interaction’.

PPI of DEPs

STRING software was used to construct a potential PPI of the identified DEPs. The robust and cross-talking interactions are mapped in Fig. 5. The DEPs were discovered to be associated with the immune system and inflammation. In particular, the functions of serum amyloid A-1 protein (SAA1), amyloid-β precursor protein (APP), complement C4-B (C4B), complement C5 (C5), DEFA1, LTF, C-reactive protein (CRP), S100A7, S100A9 and lipopolysaccharide-binding protein (LBP) were demonstrated to be associated with each other.

Figure 5. Protein-protein interaction analysis of differentially expressed proteins in patients with acute bilirubin encephalopathy. Differently colored nodes represent different proteins; lines represent interactions between proteins; line width represents comprehensive score (0.4–1.0) of interaction between proteins. The wider the line, the higher the score. The default minimum score was selected as medium score (0.4). Red ellipse represents the immune and complement cascades; blue ellipse represents the immune and inflammation processes.

Verification of DEPs using western blotting

Based on the comprehensive analysis of the pathogenesis and bioinformatics of ABE, four candidate DEPs (S100A9, S100A7, LTF and DEFA1) were further verified by western blotting and an additional 12 participants were enrolled for the validation. These participants included 4 patients with moderate ABE, 4 with severe ABE and 4 controls. The expression levels of LTF and DEFA1 were significantly upregulated (Fig. 6A-C), while the expression levels of S100A7 and S100A9 were significantly downregulated (Fig. 6A, D and E) in both the moderate and severe patients with ABE compared with the control groups; these findings were consistent with the LC-MS/MS data.

Figure 6. Verification of differentially expressed proteins using western blotting. (A) Representative western blots demonstrate the expression levels of LTF, DEFA1, S100A7 and S100A9 in the MV/E isolated from the CSF. Histogram data shows the densitometric semi-quantification (mean ± SD) of (B) LTF, (C) DEFA1, (D) S100A7 and (E) S100A9 expression levels. Protein densitometric semi-quantification was normalized to the density of CD9 protein. CD9 is an exosome surface marker. *P<0.05. LTF, lactoferrin; DEFA1, α defensin 1; MV/E, microvesicle/exosome.

Discussion

The incidence of clinical ABE is high, particularly in developing countries, and subclinical BE (hypobilirubin encephalopathy) is also seen in clinical practice (6–8). The pathogenesis of BE is still unclear. In the present study, iTRAQ-based quantitative proteomic technology combined with LC/MS/MS was used to identify the proteomic profile of MV/E obtained from the CSF of patients with ABE. Certain DEPs [SAA1, APP, LBP, CRP, immunoglobulin, complement components (C4B and C5), S100A9, S100A7, DEFA1 and LTF] associated with the immune-inflammatory response were identified in the extracellular vesicles, which indicated that these extracellular vesicles may serve an important role in the pathophysiology of bilirubin-induced neurological neurotoxicity by regulating and transmitting complex communication among neurocytes.

Since the discovery of circulating extracellular vesicles in 1979 (31), there has been a rapid increase in scientific research on vesicles. Previous studies have suggested that extracellular vesicles may serve a crucial role in the pathophysiology of neurologic disorders by regulating the communication among neurocytes via complex mechanisms (32,21). Exosomes have been demonstrated to promote neurite growth and enhance axonal regeneration in injured neurons (33). Moreover, the cargo of extracellular vesicles has been discovered to be involved in the mechanisms underlying traumatic brain injury (32). However, to the best of our knowledge, the role of extracellular vesicles in bilirubin-induced neurological injury is not well characterized. In the present study, MV/E were isolated from the CSF of patients with ABE and changes in the proteomic profiles in these extracellular vesicles were identified. Certain DEPs were associated with immune-inflammation; these included SAA1, APP, LBP, CRP, immunoglobulin, complement components (C4B and C5), S100A9, S100A7, DEFA1 and LTF. In addition, GO annotation and enrichment analysis of the DEPs revealed that the majority of the DEPs were primarily associated with ‘immune system process’, ‘biological regulation’ or ‘molecular transducer activity’. KEGG signaling pathway enrichment analysis also demonstrated that these DEPs were primarily enriched in immune and inflammatory signaling pathways (‘PI3K-AKT signaling pathway’, ‘NF-κB signaling pathway’ and ‘MAPK signaling pathway’). In particular, the NF-κB signaling pathway and MAPK signaling pathway are important immune-inflammatory pathways, which have been reported to be activated by UCB at an early stage of brain development (15). These findings indicated that the DEPs identified in the extracellular vesicles of patients with CSF may serve a critical role in the pathogenesis of bilirubin-induced cerebral injury.

Previous studies have demonstrated that hazardous levels of UCB may induce the acute and chronic activation of microglia and astrocytes via the upregulation of proinflammatory gene expression levels, subsequently triggering the release of inflammatory biomarkers, such as IL-1β, TNF-α and IL-6 (14,34–37). Hazardous levels of UCB were discovered to induce the unfolded protein response and endoplasmic reticulum stress, which contributed to neuronal inflammation by inducing IL-8 expression levels via the regulation of the upstream NF-кB/PKR-like endoplasmic reticulum kinase axis (18,38,39). These results are consistent with the findings of the present bioinformatics analysis, which revealed the enrichment of the ‘NF-kappa B signaling pathway’. These bilirubin-induced immune-inflammatory signaling pathways and biomarkers have been identified previously to contribute to neuronal apoptosis, necrosis or pyroptosis (40). Altogether, these results suggested that bilirubin-induced neurological neurotoxicity may be associated with certain immune-inflammatory processes.

The present study demonstrated differential expression levels of certain proteins in the CSF of patients with ABE associated with immune-inflammation, including APP, LBP and CRP. In a previous study, APP protein was demonstrated to induce a predominantly proinflammatory phenotype of microglia in Alzheimer's disease (41). Moreover, LBP and CRP, which are non-specific markers of inflammation, were found to be significantly increased in the serum of patients with traumatic brain injury (42,43). Although these non-specific immune-inflammatory markers have been studied in the context of traumatic brain injury and neurodegenerative diseases, to the best of our knowledge, their role in the context of UCB-induced neurological dysfunction is not well characterized. The significant increase in APP, LBP and CRP expression levels in extracellular vesicles isolated from the CSF in the present study may represent the physiological and compensatory immune-inflammatory response to UCB. However, further investigations are required to verify the association between these proteins and bilirubin-induced neurological neurotoxicity.

The preset study also demonstrated the significant upregulation of complement-associated proteins, such as C4B and C5, in patients with ABE. Previous studies have reported that both neurons and glial cells are capable of synthesizing complement receptors (44,45). The complement system is hypothesized to serve an important role in brain homeostasis, neural development, central nervous system (CNS) repair, regeneration, neuroprotection, neurogenesis and the progression of pathology in neurodegenerative disorders, stroke and traumatic brain injury (46–48). These studies showed that the levels of complement components were upregulated. The present results are consistent with these studies. However, to the best of our knowledge, no previous studies have investigated the association between complement and bilirubin-induced brain injury. The present study demonstrated the significant upregulation of numerous immunoglobulin-like domains, such as immunoglobulin heavy constant γ4 and immunoglobulin heavy variable 2–5 in MV/E isolated from the CSF of patients with ABE. Immunoglobulins are an essential component of the immune response (49). These immunoglobulin-like domains have been suggested to serve as neural cell adhesion molecules that mediate interactions among nerve cells in the brain (50). Thus, extreme hyperbilirubinemia may trigger the complement system and upregulate certain immunoglobulin-like domains. However, further investigations are required to assess whether the upregulation of complement-associated proteins and immunoglobulin-like domains observed in CSF vesicles serve an important role in ABE.

S100A9, is a Ca2+-binding protein belonging to the S100 family, which is primarily derived from immunocytes (neutrophils, macrophages and myeloid-derived dendritic cells), but also expressed in astrocytes and neurons (51). S100A9 has been reported to activate multiple immune-inflammatory regulation signaling pathways, including Ras/ERK/NF-κB and PI3K/Akt in astrocytes (52,53). Moreover, the overexpression of S100A9 in neurons was identified to be involved in the infiltration of microglia and the growth of primary astrocytes in vitro via the activation of ‘Ras signaling pathways’ (54). Thus, these studies suggested that S100A9 may be an important immune inflammatory biomarker in the nervous system, which participates in the inflammatory process.

Although the upregulation of S100A9 may be involved in the immune-inflammatory response in numerous types of neuronal disorder, the present results demonstrated a significant downregulation of S100A9 expression levels in MV/E isolated from the CSF of patients with ABE. These results indicated that the balance of neurological immune-inflammatory regulation functions of S100A9 may be disrupted by hyperbilirubinemia; alternatively, the downregulation of S100A9 expression levels in the MV/E may be an intrinsic protective mechanism to trigger tissue proliferation or repair activity following the exposure to hyperbilirubinemia. S100A9 serves multiple biological roles depending on the concentration, as well as the proximal microenvironment (55). Numerous studies have demonstrated that S100A9 is involved in the modulation of cellular proliferation, differentiation and apoptosis, as well as proinflammatory and anti-inflammatory functions (56,57). At high concentrations, extracellular S100A9 exhibited proinflammatory functions and growth-inhibitory activity, induced cell apoptosis, and exerted deleterious effects by stimulating leukocyte migration and recruitment, enhancing leukocyte-endothelial contact and increasing vascular permeability (58–61). However, at low levels, extracellular S100A9 was reported to exert either anti-inflammatory properties to avoid tissue damage, or tissue proliferation/repair activity at local inflammatory sites (55). Thus, the downregulated expression levels of S100A9 may be associated with its intrinsic protective mechanism to trigger tissue proliferation or repair activity following exposure to hyperbilirubinemia. However, the association between the downregulation of S100A9 and its anti-inflammatory or intrinsic protective effects in bilirubin-induced brain injury requires further investigation. Additionally, since the S100A9 protein in the present study was detected in extracellular vesicles, it may differ from studies in which the protein was detected in total CSF (62,63). The S100A9 in extracellular vesicles represents the storage of extracellular S100A9 in the brain (64). Physiologically, there is sufficient storage of S100A9 in astrocytes or neurons; however, trauma, heat, stress, infection and inflammation have all been identified to trigger its upregulation and active release (55). Therefore, further investigations are required to determine whether the downregulated expression levels of S100A9 in extracellular vesicles are due to the release of S100A9 from extracellular vesicles into the CSF. Whether the downregulated expression levels of S100A9 and increased severity of bilirubin-induced brain damage are due to impaired neuronal function or gene suppression also remains to be elucidated.

S100A7 is also an important member of the S100 family, which contains the EF-hand calcium-binding protein (65). S100A7 is a proinflammatory protein expressed in immune cells and in neurons, microglia and astrocytes in the brain (66). S100A7 was also observed to be expressed in glial and meningeal cells, which supports the hypothesis that S100A7 may be associated with immune-inflammatory processes in the CNS (66). A previous study documented increased levels of S100A7 in the CSF of patients with Alzheimer's disease (67). However, the present study demonstrated a significant downregulation in the expression levels of S100A7 in the MV/E isolated from the CSF of patients with ABE. A study by Sharma et al (68) also reported a significant decrease in the serum levels of S100A7 in patients with acute ischemic stroke. Similar to S100A9, S100A7 expression differs between tissues and organs or in different disease states; this indicates the multifaceted nature of S100A7 function (66). The downregulated expression levels of S100A7 in the present study indicated that S100A7 may participate in BIND; however, the exact mechanism of action and function of S100A7 in this disease remains unclear. Further investigations are required to confirm the results of the present study.

In addition to the downregulation of S100A9 and S100A7 expression levels, the present study also identified the upregulation of the expression levels of certain bioactive proteins, including defensins and LTF, which regulate immune-inflammatory responses, as well as antioxidant and neuroprotective processes (69,70). The present study identified the significant upregulation of DEFA1 expression levels in the MV/E isolated from the CSF of patients with ABE. Defensins are antimicrobial peptides that serve multifaceted roles and exhibit immunomodulatory and anti-inflammatory properties (71,72). Variable expression levels of defensins have been identified in cerebral microglia and astrocytes in both the mouse and human brain, where they have been observed to serve complex roles in immunomodulatory processes (73,74). Moreover, following the injury to the CNS, microglia and astrocytes provided immune defense in a stimulus-dependent manner via the production and release of defensins (75). As ABE is a form of secondary brain injury caused by hyperbilirubinemia, the upregulated expression levels of DEFA1 indicated that DEFA1 may serve an important immunomodulatory role in the pathogenesis of bilirubin-induced brain injury. Additionally, as defensins are anti-inflammatory peptides (76), the upregulated expression levels of DEFA1 may have had an anti-inflammatory neuroprotective function by preventing the excessive inflammation in brain lesions. Thus, the findings of the present study suggested that DEFA1 may be important in the immunomodulatory and anti-inflammatory processes of bilirubin-induced neurological neurotoxicity; however, further investigations are required to elucidate the underlying mechanism.

The present study also demonstrated that the expression levels of LTF were significantly upregulated in the MV/E of patients with ABE. LTF is an iron-binding glycoprotein that belongs to the transferrin family and serves numerous beneficial biological functions, such as immunomodulatory, antioxidant and neuroprotective effects (77). Previously, LTF was observed to modulate the migration, maturation and function of immune cells (78,79). In addition, the expression levels of LTF in biological fluids were significantly upregulated in patients with inflammatory diseases (80). Moreover, in addition to the immune-inflammatory mechanisms, oxidative stress is also hypothesized to be an important pathogenetic mechanism of bilirubin encephalopathy (81). The upregulated expression levels of LTF indicated that it may be involved in maintaining hemostasis between oxidation and anti-oxidation. Previous studies have shown that ABE is partly caused by oxidative stress and brain cell damage induced by high bilirubin levels (11,81). LTF has been demonstrated to possess antioxidant properties (77). Therefore, the upregulated expression levels of LTF in the present study may be involved in maintaining homeostasis between oxidation and anti-oxidation.

LTF has been previously demonstrated to exhibit antioxidant properties, in addition to decreasing ROS generation and removing ROS from the brain; LTF was also reported to help maintain the levels of ascorbate and glutathione, which are important endogenous antioxidants (70,82). These previous findings indicated that the upregulated expression levels of LTF may improve the antioxidant capacity of the brain. Furthermore, LTF has been demonstrated to exert neuroprotective effects on the immature brain in rodent models of intrauterine growth restriction and cerebral hypoxia/ischemia (83,84). A number of studies have also revealed that upregulated expression levels of LTF served as an important neuroprotective factor, potentially due to its capacity to stimulate cell cycle progression and induce erythropoietin synthesis (77,85). Overall, the upregulated expression levels of LTF in ABE indicated that it may either modulate immune-inflammatory processes or exert antioxidant or neuroprotective effects in the pathological process of BIND. However, the mechanism linking LTF with BIND still needs to be evaluated.

In conclusion, the present study determined the proteomic profile of extracellular vesicles (MV/E) isolated from the CSF of patients with ABE. The findings of the present study provided an improved understanding of the pathophysiology of BIND. Certain proteins and signaling pathways involved in the MV/E were primarily associated with the immune-inflammatory response. This indicated that extracellular vesicles may serve an important role in the development of bilirubin-induced neurological neurotoxicity via immune-inflammatory proteins and signaling pathways. Moreover, extracellular vesicles in the CSF are derived from neurocytes; thus, their cargo (S100A9, S100A7, DEFA1 and LTF) represents the microenvironment of the brain and offers information about the pathophysiology of cerebral injury for patients with ABE (86). Altered DEP levels in extracellular vesicles of CSF results in changes to the microenvironment of the brain and the adaptation process of nerve cells due to hyperbilirubinemia. The results therefore suggested that extracellular vesicles and the identified proteins may serve as biomarkers and provide a critical advantage over traditional biomarkers in the early diagnosis of ABE. Currently, the diagnosis of ABE primarily depends on clinical manifestations and cranial MRI scans. Extracellular vesicles and proteins may serve as objective biomarkers. Furthermore, due to the lack of specific drug therapeutic options for ABE, extracellular vesicles may present a novel therapeutic target; the immune-inflammatory targets involved in extracellular vesicles may be engineered to regulate intercellular communication and improve the outcomes for patients with ABE, which could provide novel therapeutic perspectives. Thus, further investigations are required to determine the usefulness of MV/E as early diagnostic biomarkers and novel therapeutic targets for ABE.

Supplementary Material

Supporting Data

Acknowledgements

Not applicable.

Funding

The present work was supported by the Hunan Provincial Science and Technology Department, China (grant no. 2019JJ80059) and the Health and Family Planning Commission of Hunan Province, China (grant no. B20180259).

Availability of data and materials

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

Authors' contributions

BW and NT designed the study and wrote the manuscript. SH, ZW and BW wrote the manuscript. NT, SH, ZW and ZH performed the proteome experiments and western blotting analysis. NT, ZW and BW revised the manuscript. All authors read and approved the final manuscript.

Ethics approval and consent to participate

The study protocol was approved by the Ethics Committee of Affiliated First People's Hospital of Chenzhou, Southern Medical University (approval no. 2019049; Chenzhou, China). Written informed consent was obtained from parents/caregivers of all subjects prior to the start of the study. All experimental procedures were performed in accordance with the Declaration of Helsinki.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Glossary

Abbreviations

Abbreviations:

References