Characterization and microRNA Expression Analysis of Serum-Derived Extracellular Vesicles in Severe Liver Injury from Chronic HBV Infection

Background: Extracellular vesicle (EV) microRNAs have been documented in several studies to have significantly different expressions in hepatitis B virus (HBV)-related liver diseases, such as hepatocellular carcinoma (HCC). The current work aimed to observe the characteristics of EVs and EV miRNA expressions in patients with severe liver injury chronic hepatitis B (CHB) and patients with HBV-associated decompensated cirrhosis (DeCi). Methods: The characterization of the EVs in the serum was carried out for three different groups, namely, patients with severe liver injury-CHB, patients with DeCi, and healthy controls. EV miRNAs were analyzed using miRNA-seq and RT-qPCR arrays. Additionally, we assessed the predictive and observational values of the miRNAs with significant differential expressions in serum EVs. Results: Patients with severe liver injury-CHB had the highest EV concentrations when compared to the normal controls (NCs) and patients with DeCi (p < 0.001). The miRNA-seq of the NC and severe liver injury-CHB groups identified 268 differentially expressed miRNAs (|FC| > 2, p < 0.05). In this case, 15 miRNAs were verified using RT-qPCR, and it was found that novel-miR-172-5p and miR-1285-5p in the severe liver injury-CHB group showed marked downregulation in comparison to the NC group (p < 0.001). Furthermore, compared with the NC group, three EV miRNAs (novel-miR-172-5p, miR-1285-5p, and miR-335-5p) in the DeCi group showed various degrees of downregulated expression. However, when comparing the DeCi group with the severe liver injury-CHB group, only the expression of miR-335-5p in the DeCi group decreased significantly (p < 0.05). For the severe liver injury-CHB and DeCi groups, the addition of miR-335-5p improved the predictive accuracy of the serological levels, while miR-335-5p was significantly correlated with ALT, AST, AST/ALT, GGT, and AFP. Conclusions: The patients with severe liver injury-CHB had the highest number of EVs. The combination of novel-miR-172-5p and miR-1285-5p in serum EVs helped in predicting the progression of the NCs to severe liver injury-CHB, while the addition of EV miR-335-5p improved the serological accuracy of predicting the progression of severe liver injury-CHB to DeCi.


Introduction
Chronic hepatitis B (CHB) is a persistent inflammatory disease of the liver caused by hepatitis B virus (HBV) infection. Following infection with HBV, inflammation induces hepatocyte necrosis, which is a crucial pathophysiological process of disease progression. miRNA-122 (a major hepatic miRNA) and miRNA-192, have been found to increase with fibrosis stage in serum EVs [32][33][34]. Of note, as fewer than 10 patients were included in each study, this result requires further validation. miRNA-21 expression has been found to be higher in the circulating serum EVs of patients with HCC (n = 30) than in patients with CHB without cirrhosis and healthy controls (n = 30 in each group) [35]. The expression levels of miRNA-122 in circulating serum EVs can distinguish early-stage HCC from liver cirrhosis (LC) [36]. The search for EV miRNAs that can aid in predicting whether patients with CHB will rapidly progress to DeCi disease is essential. In this study, we collected samples from patients with severe liver injury-CHB and DeCi and performed a systematic observation and a comparative analysis of the EVs and their miRNAs for both diseases to provide ideas for subsequent research.

Participants and Sample Collection
A total of 58 patients with severe liver injury-CHB (HBeAg-positive chronic hepatitis B) and 33 patients with DeCi were enrolled in the present study. All patients were hospitalized in the Infectious Diseases Department of the Second Xiangya Hospital of Central South University (Changsha, China) from April 2015 to October 2020, and they were diagnosed according to the guidelines of Hepatitis B Virus-related Cirrhosis Clinical Management [37]. We established exclusion criteria to reduce the effects of other physical or pathological factors on the production of serum EV miRNAs. The exclusion criteria were as follows: patients with alcoholic cirrhosis, cirrhosis in the compensatory phases, chronic hepatitis C or D, other chronic liver diseases (such as autoimmune hepatitis, hepatolenticular degeneration, severe organ dysfunction, and hematological diseases), autoimmune diseases, malignant tumor diseases, and mental disorders. The clinical data of the patients are summarized in Table 1. We also selected 58 normal controls (NCs) who had visited the Second Xiangya Hospital of Central South University as the parallel controls.
Blood samples were collected from all patients and healthy controls. To remove cell debris, serum separation was accomplished via centrifugation at 300× g for 5 min at 4 • C within 2 h, and then the samples were stored at −80 • C until the EVs separated.

Serum EV Isolation
The serum from 5 mL of the blood sampled from each patient was centrifuged at 1500× g for 10 min to eliminate cell precipitation and debris, and then the cell-free supernatant was centrifuged at 15,000× g for 30 min to remove large EVs. Five pools, each from the NC group and the severe liver injury-CHB group, were used for NGS sequencing, and every pool contained a mixture of samples from 5 random individuals (400 µL/sample) of the same group. For the RT qPCR analysis, 1 mL of serum was taken from each sample for detection. For NTA, TEM, and WB analyses, 0.5 mL of serum was taken for detection. Each pool or sample was filtered through a 0.22 µm pore filter (EMD Millipore, Billerica, MA, USA) to remove any larger particles. Next, each sample or pool of samples was centrifuged in a 100 KD centrifuge concentrator tube (EMD Millipore) at 3000× g for 30 min at 4 • C. Eventually, the supernatant from the previous procedure was enriched with 1/4 volume of ExoQuick-5A™ (System Biosciences, Inc., Palo Alto, Santa Clara, CA, USA) according to the manufacturer's instructions. Finally, the EVs extracted were resuspended with a specific volume of phosphate-buffered saline (PBS) for different purposes and stored at −80 • C.

Characterization and Quantification of EVs
A nanoparticle tracking analysis (NTA) was carried out to track EV diameter and concentration. All samples were diluted with PBS before the NTA analysis, and then 100 µL of the sample was loaded into the EV analysis chamber of the Zetaview equipment PMX 110 (Particle Metrix, Meerbusch, Germany). Microsoft Excel 2013 (Microsoft Corp., Seattle, WA, USA) was used to handle the data obtained from Zetaview 8.04.02 SP2, whose parameters were optimized appropriately during the experiment.
For the transmission electron microscopy (TEM), 20 µL of the EV suspension was dropped into a carbon-coated 200 mesh copper grid in the form of a droplet for a period of time (more than 1 min). Subsequently, 2% uranyl acetate solution was added to yield negative staining, which was fixed in 10 min. The unnecessary liquid was removed using filter paper at room temperature, and then a Tecnai biological transmission electron microscope (model: Tecnai G2 Spirit, Thermo Fisher Scientific, Waltham, MA, USA) was used to obtain micrographs at 80 KV.
Further, we extracted EV proteins using RIPA Lysis Buffer (Beyotime, Shanghai, China) and a Protease Inhibitor Cocktail (Sigma-Aldrich, Darmstadt, Germany) for the Western blotting analysis. The EV proteins were then quantified using a Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific). Next, 15 µg of the EV proteins was loaded into each well of a 10% sodium dodecyl sulphate polyacrylamide gel for electrophoresis (SDS-PAGE) (Beyotime, Shanghai, China), followed by separation under the condition of 180 V (Tannon, Shanghai, China). Next, The EV proteins were transferred to a polyvinylidene difluoride (PVDF) membrane (Merck KGaA, Darmstadt, Germany) for 1.5 h with a 252 mA current. We purchased the following anti-bodies from Santa Cruz (Santa Cruz, CA, USA): anti-CD63 (sc-5275), CD9 (sc-13118), TSG101 (sc-7964), and calnexin (sc-23954). After blocking with 5% non-fat dry milk in 1 × TBST (TBS, 0.1% Tween 20) buffer for 1 h, the membrane was incubated in a primary antibody solution dissolved with 1 × TBST containing 5% non-fat dry milk (CD63, 1:2000; CD9, 1:2000; TSG101, 1:2000; and calnexin, 1:2000) at 4 • C overnight. The membrane was washed three times with 1 × TBST for 10 min. Next, the secondary antibody at a dilution of 1:10,000 was incubated at room temperature for 1.5 h and rewashed three times. The target protein was detected with an enhanced chemiluminescence kit (Super-Signal™ West Femto Maximum Sensitivity Substrate, Thermo Fisher Scientific, MA, USA) using a Bio-Rad imaging system according to the manufacturer's instructions.

RNA Extraction from Serum EVs
According to the manufacturer's instructions, the NGS sequencing sample's total RNA was extracted from the serum EVs using an miRNeasy Mini Kit (Qiagen, Valencia, CA, USA). As for the RT-qPCR analysis samples, cel-miR-39 (RiBoBioInc., Guangzhou, China) was used as an external reference during extraction, normalizing the process of RNA extraction and PCR. Briefly, 1000 fmol of external cel-miR-39-3p was added to the EVs and mixed well. Next, the miRNA was extracted from the serum-isolated EVs utilizing a miRNeasy Mini Kit and an RNeasy MinElute Cleanup Kit (Qiagen, Valencia, CA, USA) following the manufacturer's instructions. All RNAs were stored at −80 • C until further analysis.

RNA Sequencing and Data Analysis
Equal amounts of total RNA were taken for library preparation after identification and quantification using an Agilent 2100 bioanalyzer (Thermo Fisher Scientific, MA, USA) and then purified by electrophoretic separation on a 15% PAGE gel. Small RNAs of 18-30 nt were recovered and sequentially ligated to 3' and 5' RNA adapters. After transcription and amplification, a fragment of 110-130 bp was selected and purified. The library was quality assessed and quantitated using the Agilent 2100 bioanalyzer and QPCR, and then it was sequenced using the BGISEQ-500 platform (BGI, Shenzhen, China).
After filtering, clean tags were mapped to the reference genome with Bowtie 2 [38]. miRDeep2 was used to predict novel miRNA by exploring the secondary structure [39], and Cmsearch was performed for Rfam mapping [40]. DE miRNAs between two groups were performed using DEGseq. Next, we plotted volcano plots of DE miRNAs and the expression heatmap of 15 miRNAs online (https://www.bioinformatics.com.cn/, accessed on 20 March 2021).

RT-qPCR
Stem-loop primers and specific primers for miRNAs were designed using miRNA Design software (http://www.vazyme.com/companyfile/7/, accessed on 20 March 2021), as shown in Table S1. Reverse transcription for cDNA production was performed using a HiScript III 1st Strand cDNA Synthesis Kit (Vazyme, Nanjing, China). Next, we measured miRNA levels according to the manufacturer's instructions with miRNA Universal SYBR qPCR Master Mix (Vazyme, Nanjing, China).
In short, using RNA as a template, reverse transcription was carried out with the HiScript III 1st Strand cDNA Synthesis Kit and miRNA-specific RT primers to obtain cDNA. The cDNA was then amplified with miRNA Universal SYBR qPCR Master Mix (Vazyme, Nanjing, China) and specific primers with the following thermocycler protocol: 95 • C for 5 min + (95 • C for 10 s; 56 • C for 30 s; 72 • C for 30 s) for 40 cycles. The qPCR was run on a Bio-Rad CFX96 touch qPCR system, and a data analysis was performed using Bio-Rad CFX Manager software.
For this study, we calculated the expression levels of EV miRNAs in the samples by using the relative quantification method. For homogenization, equal quantities of cel-miR-39-3p were added as an internal control to each sample prior to RNA extraction. ∆Ct was calculated as Ct (miRNA of interest) − Ct (cel-miR-39-3p), and the relative expression values for the target microRNA was calculated as 2-∆Ct. All reactions were run in triplicate.

Statistical Analysis
GraphPad Prism 8.3 and SPSS 23.0 were utilized for data processing and statistical analyses. After calculating whether it was a normal distribution, the differences in clinical characteristics between the severe liver injury-CHB and DeCi groups were assessed using nonparametric tests and t-tests, respectively. We used mean (SE) to present miRNA expression values. The differences in expression levels between the two groups were compared using the nonparametric Mann-Whitney test. A receiver operating characteristic curve (ROC) analysis was carried out to assess the predictive value of the EV miRNAs and their different combinations. The predictive values of each combination were obtained by performing binary logistic regression before the ROC curve analysis, and area under the curve (AUC) > 0.7 and p < 0.05 were considered the criteria for evaluating the predictive markers. Spearman's analysis was used to evaluate the correlation between the miRNAs and clinical practices. A two-tailed t-test was used for p values, and p < 0.05 was regarded as statistically significant.
After calculating whether it was a normal distribution, the differences in clinical characteristics between the severe liver injury-CHB and DeCi groups were assessed using nonparametric tests and t-tests, respectively.

Study Population
In this study, 149 patients (severe liver injury-CHB = 58, DeCi = 33, and NC = 58) from the Second Xiangya Hospital were enrolled. There were 25 individuals in each group (NC and severe liver injury-CHB) for NGS, and there were 33 persons in each group for training and validation. The primary characteristics and a comparison between the patients with severe liver injury-CHB and the patients with DeCi are presented in Table 1. Even though all the patients were randomly selected, there were more male patients than female patients. The vast majority of the patients with DeCi were either on or had been on antiviral therapy prior to hospital admission. The patients from the severe liver injury-CHB, DeCi, and NC groups were aged 42 (interquartile range (IQR): 29-54), 53 (IQR: 48-61), and 37 (IQR: 30.5-49) years, respectively. As expected, the patients with severe liver injury-CHB had higher levels of ALT, AST, TBIL, DBIL, GGT, and AFP than the patients with DeCi. The serum levels of WBC, L, PLT, ALB, and CHE were lower, and INR, lgA levels were higher in patients with DeCi than in patients with severe liver injury-CHB. Additionally, there were no considerable differences between the patients with severe liver injury-CHB and the patients with DeCi in other clinical features. The patients who were enrolled were not diagnosed as having alcoholic cirrhosis, hepatitis C, hepatitis D, etc.

Characterization of Serum EVs
We visualized the serum-derived EVs by using TEM and NTA, and then we characterized them by using a Western blot for EV markers. In the representative images, the isolated EVs with diameters of 50-150 nm and their characteristic rounded membranebound morphology was observed at high-magnification views ( Figures S1 and 1B). Next, as shown in Figure 1A, the isolated EVs displayed markedly detectable expressions of EV markers, such as CD63, CD9, and TSG101, and had no expression of the endoplasmic reticulum calnexin, which is consistent with the previously reported EV characteristics.
Additionally, in our study, the patients with severe liver injury-CHB displayed the highest EV concentrations, which were approximately 2.6-fold greater than those of the patients in the DeCi group (p < 0.001, Figure 1D). The EV diameter significantly differed between the NC vs. severe liver injury-CHB and severe liver injury-CHB vs. DeCi groups, with sizes of 100. 23 Figure 1C). No difference in the concentrations and sizes of the EVs were found in the NTA analysis after stratification by gender ( Figure S2). patients in the DeCi group (p < 0.001, Figure 1D). The EV diameter significantly differed between the NC vs. severe liver injury-CHB and severe liver injury-CHB vs. DeCi groups, with sizes of 100.23 nm [83. 24-112.70], 126.77 nm [122. 30-134.57], and 115.07 nm [108.78-115.95] for the NC, severe liver injury-CHB, and DeCi groups, respectively ( Figure 1C). No difference in the concentrations and sizes of the EVs were found in the NTA analysis after stratification by gender ( Figure S2).  (D) Quantification of nanoparticles using NTA. In (C,D), the horizontal lines indicate the median and box, and the bars represent the 25-75th and the 5-95th percentiles, respectively. ** p < 0.01 and *** p < 0.001 (normal controls (NCs, n = 8), severe liver injury-chronic hepatitis B (CHB, n = 11), and decompensated cirrhosis (DeCi, (n = 8)).

Serum-Derived Exosome miRNA Screening
We sequenced NC and severe liver injury-CHB samples in order to screen for differences in the serum EV miRNAs. First, we constructed sample pools for the different groups. For each group, we took five samples and randomly pooled them into a single pool and finally obtained five pools. The sample pools of each group were sequenced and analyzed using a high-throughput sequencing technique (NC = 25 and severe liver injury-CHB = 25). A total of 364 miRNAs were identified as differentially expressed in the sequencing, of which, 268 miRNAs were significantly up-or down-regulated (|FC| > 2, p < 0.05; Figure 2A). Based on sequencing results and literature, 11 miRNAs were selected for detection in our training set (|FC| > 2, p < 0.0001, at least ≥50 copies; Table S1). In addition, four miRNAs previously reported to be associated with disease-related molecules (hsa-miR-1285-5p, hsa-miR-204-3p, hsa-miR-335-5p, hsa-miR-877-5p) were also selected for detection (|FC| > 2, p < 0.0001). Heat maps of the 15 miRNAs are shown in Figure 2B.  Figure 2A). Based on sequencing results and literature, 11 miRNAs were selected for detection in our training set (|FC| > 2, p < 0.0001, at least ≥50 copies; Table S1). In addition, four miRNAs previously reported to be associated with disease-related molecules (hsa-miR-1285-5p, hsa-miR-204-3p, hsa-miR-335-5p, hsa-miR-877-5p) were also selected for detection (|FC| > 2, p < 0.0001). Heat maps of the 15 miRNAs are shown in Figure 2B.

The Differences in the Expressions of miRNAs between the NC Group and the Severe Liver Injury-CHB Group
In this study, we focused on the 15 most viable candidate miRNAs for follow-up validation. Among these, miR-172-5p was a novel miRNA that was identified and sequenced (aggcuggagugcaguggcg). The primers are shown in Table S2. Ten samples each from the NC and severe liver injury-CHB groups were randomly enrolled, and serum EV RNA was extracted for an RT-qPCR assay. In all the groups screened, a total of six miRNAs were shortlisted based on valid Ct values (Ct < 35): miR-172-5p, miR-411-5p, miR-1285-5p, miR-335-5p, miR-877-5p, and miR-4433a-3p.
For a further evaluation of the potential of these miRNAs, the levels of these six individual miRNAs were assessed by RT-qPCR in an independent cohort of NCs (n = 23) and patients with severe liver injury-CHB (n = 23). The miRNA levels were normalized by cel-miR-39-3p. Interestingly, only miR-172-5p and miR-1285-5p were found to be significantly differentially expressed ( Figures 3A,B and S3).

The Differences in the Expressions of miRNAs between the NC Group and the Severe Liver Injury-CHB Group
In this study, we focused on the 15 most viable candidate miRNAs for follow-up validation. Among these, miR-172-5p was a novel miRNA that was identified and sequenced (aggcuggagugcaguggcg). The primers are shown in Table S2. Ten samples each from the NC and severe liver injury-CHB groups were randomly enrolled, and serum EV RNA was extracted for an RT-qPCR assay. In all the groups screened, a total of six miRNAs were shortlisted based on valid Ct values (Ct < 35): miR-172-5p, miR-411-5p, miR-1285-5p, miR-335-5p, miR-877-5p, and miR-4433a-3p.
For a further evaluation of the potential of these miRNAs, the levels of these six individual miRNAs were assessed by RT-qPCR in an independent cohort of NCs (n = 23) and patients with severe liver injury-CHB (n = 23). The miRNA levels were normalized by cel-miR-39-3p. Interestingly, only miR-172-5p and miR-1285-5p were found to be significantly differentially expressed ( Figures 3A,B and S3).

The Difference in the Expressions of miRNAs between the NC Group and the DeCi Group
We simultaneously examined the expressions of the six shortlisted miRNAs in the patients with DeCi (n = 33) and obtained results similar to those of patients with severe liver injury-CHB. Compared to the NC group, miR-172-5p and miR-1285-5p showed significantly decreased expression levels (based on the RT-qPCR results) in the DeCi group. Additionally, miR-335-5p was significantly downregulated in the DeCi group compared to that in the NC group.

The Difference in the Expressions of miRNAs between the NC Group and the DeCi Group
We simultaneously examined the expressions of the six shortlisted miRNAs in the patients with DeCi (n = 33) and obtained results similar to those of patients with severe liver injury-CHB. Compared to the NC group, miR-172-5p and miR-1285-5p showed significantly decreased expression levels (based on the RT-qPCR results) in the DeCi group. Additionally, miR-335-5p was significantly downregulated in the DeCi group compared to that in the NC group.
The ROC curve analysis of miR-335-5p showed an AUC area of 0.660 (95% CI: 0.5307-0.7893, p = 0.023). The low expression of miR-335-5p in serum EVs predicts a greater risk of further development of DeCi in patients with CHB and severe liver injury.
Additionally, we attempted to perform a joint analysis based on the available clinical indicators. We systematically analyzed the differences in the serological levels between the severe liver injury-CHB and DeCi groups and calculated their predictive sensitivity and specificity. The sensitivity and specificity of CHE, Ca 2+ , and AST/ALT exceeded expectations and showed better predictive properties (Table S3). The highest sensitivity and specificity (91.7% and 91.3%, respectively) were achieved after combining miR-335-5p with the aforementioned clinical indicators.
Additionally, to detect the correlation between miR-335-5p and disease, we calculated the correlation between miRNA concentrations and serological detection levels using Spearman's correlation analysis. The levels of miR-335-5p were remarkably related to the degree of inflammation (ALT and AST), GGT, and the index of hepatocyte regeneration (AFP) in a wide range of serological parameters (Table 2). However, miR-172-5p and miR-1285-5p displayed no statistical relevance to serological levels.  'rs' is the correlation coefficient between serum EV miR-335-5p and other variables. A p-value < 0.05 was considered statistically significant; *p < 0.05; **p < 0.01.  'rs' is the correlation coefficient between serum EV miR-335-5p and other variables. A p-value < 0.05 was considered statistically significant; * p < 0.05; ** p < 0.01.

Discussion
HBV infection is the primary cause of progressive liver disease. Liver biopsy has the potential to cause diagnostic mistakes attributable to heterogeneous distribution [9]. Blood tests include measurements of transaminases (ALT/AST), fibrosis-related factors, HBV DNA levels, and the presence or absence of HBV antigens. They have limited effectiveness in indicating the progression of liver disease [7]. In the past few decades, identifying biological phenotypes at the histological level has shown great potential for disease diagnosis. For example, clinical glycomic techniques used to distinguish patients with compensated cirrhosis from those with non-cirrhotic chronic liver disease have shown a sensitivity of 79% and a specificity of 86% [41]. Using genome-wide miRNA microarrays, miR-106b and miR-181b in serum were identified as specific biomarkers with an AUC of 0.774 for chronic HBV-associated liver cirrhosis (HBV-LC) and 0.915 for non-chronic HBV-LC [42]. Additionally, miR-101 in serum was reported to discriminate HBV-HCC from HBV-LC with a sensitivity of 95.5% and a specificity of 90.2% [43]. Serum microfibrillarassociated protein 4 (MFAP-4) can help distinguish patients with LC from individuals without liver disease (AUC = 0.97) using a proteomic approach [44].
EVs, as well as EV RNAs and EV proteins, are regarded as types of liquid biopsy, which boasts the advantages of non-invasiveness and dynamic monitoring, and they are able to overcome the limitations of tumor spatiotemporal heterogeneity. miRNAs are short non-coding RNA oligonucleotides that can increase mRNA degradation or decrease their translation by targeting specific mRNAs [45]. EVs with bilayer membranes in the circulatory system can protect their miRNA cargos and influence the target cells' behavior, making EVs a popular candidate for studying the mechanisms of diseases and intercellular communication [15,16]. The role of EV miRNAs in HCC has been studied in several articles. For instance, serum exosome miR-122 levels are lower in patients with HCC than in patients with CHB or DeCi [46,47]. In contrast to normal hepatocyte-derived exosomes, 49 significantly over-expressed miRNAs have been observed in HCC [48].
In this study, EVs were isolated from serum based on their physical size and the principle that hydrophobic protein and lipid molecules can bind to PEGs. EVs were identified based on their physical shape, diameter, and protein markers, such as CD63 and CD9 [16]. In the Western blot results, we observed a gradual increase in the expressions of CD63 and CD9 with disease progression. ScRNA-seq demonstrated the expansion of macrophages in a mouse model of NASH with a pro-inflammatory phenotype, and CD9 is overexpressed in macrophages [49]. Furthermore, Sabine et al. confirmed that macrophagemediated inflammation is critical in the pathogenesis of non-alcoholic steatohepatitis (NASH), and a higher level of CD63 and CD9 expression was detected in monocyteinfiltrating macrophages entering the liver [50]. The high expression of CD63 and CD9 found in the current research may be associated with macrophages, and more profound validation studies will be necessary shortly.
We observed that patients with severe liver injury-CHB had the largest EVs size, which became smaller when progressing to DeCi. It has been reported that the expression level of TSG101 can affect the quantity and size of EVs [51]. In the present study, we also detected higher expression levels of TSG101 in patients with severe liver injury-CHB than Life 2023, 13, 347 12 of 17 in patients with DeCi, which may contribute to the change in EV size. In addition, changes in the physiological environment may also affect the size of EVs. For example, increased micro-environment acidity led to significantly smaller EV sizes in patients with prostate cancer (PCA) compared to individuals with no signs of urological disease [52].
Many studies have demonstrated that higher EV concentrations are detected in patients with diseases when comparing multiple diseases with normal groups [53][54][55][56][57]. In the current study, we also confirmed that patients with severe liver injury-CHB had the most numerous EVs. Interestingly, in our study, it was also found that the Ca 2+ levels in peripheral blood were significantly higher in patients with severe liver injury-CHB than in patients with DeCi (p < 0.001), with a similar trend to the concentration of EVs. It has been demonstrated that exosomes are released in a Ca 2+ -dependent manner [58,59]. The production and shedding of red blood cell (RBC)-derived MV are associated with calcium and calcium carriers [60,61]. Increased intracellular Ca 2+ concentrations in erythrocytes and platelets promote micro-vesicular release [62,63]. In addition, Ca 2+ influx promotes cellular repair when a certain level of cellular damage has occurred, and the size or shape of the specific membrane damage prompts its removal via lysosomal cytosolic spitting or micro-vesicular shedding [64]. We speculate that the distribution of the number of EVs in the patients in this study may have some correlation with Ca 2+ levels.
The ROC analysis of the NC group compared with both the severe liver injury-CHB and DeCi groups indicated that the novel microRNA miR-172-5p was the best predictor, followed by miR-1285-5p. miR-335-5p might predict the risk of the further development of DeCi in patients with CHB and severe liver injury.
The novel-miR-172-5p was a newly identified miRNA in this study. miR-172-5p was predicted to target multiple disease-related genes using bioinformatics software. CAMP response element-binding protein 1 (CREB1) is a core transcription factor, and it may be a promising therapeutic target for liver diseases [65]. HBV DNA polymerase attenuates HBV replication by activating the CREB1-HOTTIP-HOXA13 axis [66]. However, in this study, Spearman's analysis did not find it to be associated with any clinical features (such as HBV DNA). Insulin-like growth factor binding protein 5 (IGFBP5) is also noteworthy, as it induced effects similar to those induced by TGFB1 [67]. IGFBP5 is a secretory protein associated with cell proliferation, adhesion, migration, systemic inflammatory response, and fibrosis. Furthermore, IGFBP is involved in the evolution of liver fibrosis, where an increase in IGFBP5 likely leads to a higher production of collagen type I in fibroblasts and enhanced tissue fibrosis [68]. novel-miR-172-5p is likely to be a helpful marker for liver diseases. Decreased expression levels of novel-miR-172-5p may promote the liver fibrosis process by affecting its target gene IGFBP, for which a more systematic study of this view is essential.
There is no clear evidence of a correlation between miR-1285 expression levels and chronic liver diseases, such as HCV infection, non-alcoholic steatohepatitis, and autoimmune liver disease. Several papers have reported their association with tumor processes. As a regulatory miRNA of p53, miR-1285-5p is a tumor suppressor that inhibits cell proliferation and migration. One study showed that miR-1285-5p expression was lower in tumor tissues than in normal tissues [69,70]. DAPK2, a known tumor-suppressor, mediates both the anti-proliferative and the pro-apoptotic effects of miR-1285 depletion [71]. Moreover, the over-expression of miR-1285 weakens the TGF-β2-induced EMT process [72]. Considering these reports, we speculate that miR-1285-5p may be essential in regulating the development of liver parenchymal cells into fibrotic cells. More functional studies on miR-1285-5p and its targets are expected. miR-335 regulates numerous genes and plays multifunctional roles. Through exosomes, miR-335-5p completes the transfer from donor to recipient cells and promotes the migration, invasion, and EMT of cancer cells [73]. Transcription Factor EB (TFEB), a major regulator of lysosomal biogenesis and autophagy, increases insulin receptor substrate 1 (IRS1) protein and modulates glucose tolerance by decreasing miR-335 levels [74]. Additionally, in our research, the expression of miR-335-5p was strongly correlated with the Life 2023, 13, 347 13 of 17 degree of inflammation (ALT, AST), GGT, and the index of hepatocyte regeneration (AFP) levels, which may facilitate an intensive study of the mechanisms of disease progression.
The diagnosis of HBV-related cirrhosis, which is the result of the development of CHB, involves the diagnosis of etiology, the assessment of compensated or decompensated status, and the evaluation of complication profiles. We observed the characteristics of the EVs in the disease phenotypes in the severe liver injury-CHB and DeCi samples in this study. The patients with severe liver injury-CHB had the highest number of EVs, which might be closely associated with Ca 2+ concentrations. The expression level of novel-miR-172-5p in serum EVs differed significantly between the healthy individuals and the patients with severe liver injury-CHB or DeCi. The expression level of miR-335-5p decreased gradually with the progression of HBV-associated liver disease. Although the difference in miR-335-5p used in this study could not satisfactorily distinguish between the severe liver injury-CHB and DeCi groups, the inclusion of EV miRNA improved its predictive accuracy. A more comprehensive serological analysis of the different phases of HBV disease progression holds promise for predicting disease progression and for diagnosing compensated cirrhosis. Furthermore, studies on the correlation of miRNAs with liver fibrosis progression, inflammation, and hepatocyte regeneration would be greatly appreciated.

Conclusions
We successfully isolated EVs and analyzed their concentrations and miRNA components. Patients with severe liver injury-CHB had the highest number of EVs, which might be closely related to Ca2+ concentrations. Furthermore, this study revealed the differential expression profiles of serum EV miRNAs in NC, severe liver injury-CHB, and DeCi groups, and it verified the candidate miRNA expression differences in the different groups using RT-qPCR. novel-miR-172-5p and miR-1285-5p in serum EVs provided a positive predictive accuracy for severe liver injury-CHB vs. NC. A combination of three miRNAs (novel-miR-172-5p, miR-1285-5p, and miR-335-5p) on a panel had the highest detection accuracy when differentiating the DeCi group from the normal group. Additionally, miR-335-5p was statistically different compared to the DeCi and severe liver injury-CHB groups, which may indicate a positive correlation with the level of inflammation. The addition of EV miR-335-5p improved the serological level of accuracy in predicting severe liver injury-CHB progression to DeCi. In summary, our research demonstrated the differences in EV characteristics and EV miRNA expressions between NC, severe liver injury-CHB, and DeCi groups, providing ideas for subsequent studies of predictive markers and studies of disease progression mechanisms.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/life13020347/s1. Figure S1: Representative electron micrograph of serum EVs; Figure S2: Regression lines and correlation coefficients between number of EVs per/mL at the NTA; Figure S3: Verification of serum EV miRNA expression levels evaluated by RT-qPCR; Table S1: Candidate miRNAs for validation set; Table S2: Primers in the present study involved; Table S3: Sensitivity and Specificity of clinical indicators in discriminating between different groups.