Study on plasma exosome biomarkers of pregnant women with intrahepatic cholestasis of pregnancy

Abstract Objective and aims Serum total bile acid (TBA) level as the main index for the diagnosis of intrahepatic cholestasis of pregnancy (ICP) has some limitations. The early diagnosis and new treatment of ICP still need to be further strengthened. Materials and methods Plasma samples were collected, and exosomes were isolated. Key differential proteins were screened by bioinformatics methods. ELISA method was used to detect the concentration of the key differential protein in plasma samples, and the receiver operating characteristic curve (ROC) curve was drawn to find out the best critical value. Results There were 138 differentially expressed proteins between the ICP and the normal groups by quantitative analysis. Cluster protein (CLU) was screened as a clinical validation index. The CLU concentration of plasma exosomes in the ICP group was significantly higher than that in the normal group (p < .0001). ROC curve analysis showed that the best critical point for diagnosing ICP according to the plasma exosomes CLU concentration of pregnant women was 255.28 ng/ml. In the ICP group, the best crucial point for predicting ICP with premature delivery is 286.72 ng/ml. Conclusions The plasma exosomes CLU in pregnant women with ICP is an important biomarker for clinical diagnosis and prediction of premature delivery of ICP.


Introduction
ICP is a particular complication during the second and third trimesters of pregnancy, mainly characterized by skin itching, abnormal liver function and elevated serum TBA. The pathogenesis of ICP is still unclear [1,2]. ICP primarily affects the growth and development of the fetus, including premature birth, fetal distress, neonatal asphyxia and even neonatal death [3,4]. Ursodeoxycholic acid (UDCA) and S-adenosylmethionine are the main therapeutic drugs. Although the serum TBA level of pregnant women with ICP decreased significantly after treatment, premature birth and perinatal death are still critical problems for obstetricians and pregnant women [5].
The maternal serum TBA level is the most commonly used laboratory parameter for the clinical diagnosis of ICP [6]. However, many studies have found that the accuracy of TBA in diagnosing ICP may be overestimated [7]. The use of TBA as the main indicator for the clinical diagnosis and prognosis of ICP has certain limitations. Therefore, it is very important to explore the pathogenesis of ICP and find more valuable biomarkers besides TBA for clinical diagnosis and prognosis judgment of pregnant women with ICP.
Exosomes were first found in sheep reticulocytes and were named 'Exosome' in 1987 [8]. Abels and Breakefield [9] found that cells can exchange genetic materials through RNA in exosomes, which are widely distributed in biological fluids, such as plasma, serum, saliva, urine, amniotic fluid, etc., and can maintain a stable state and biological characteristics at a long-term extremely low temperature, and are used to study many diseases [10]. Therefore, exosomes analysis has broad application prospects in the pathogenesis, diagnosis and treatment of diseases.
In this study, the plasma exosomes of pregnant women with ICP were isolated and analyzed by bioinformatics. The selected exosomes CLU was analyzed by ROC curve to preliminarily explore the significance of exosome CLU as a biomarker for clinical diagnosis of pregnant women with ICP.

Sample collection
ICP pregnant women who delivered in the First Affiliated Hospital of Chongqing Medical University from January 2020 to December 2020 were involved in this study. Before the termination of pregnancy (natural delivery or cesarean section), 5 ml of blood from the elbow vein was collected in an anticoagulant tube containing EDTA. After centrifugation (5000 r, 5 min), the plasma was stored in a -80 C refrigerator. This study was approved by the ethical committee of the First Affiliated Hospital of Chongqing Medical University.
The work was undertaken, and that it conforms to the provisions of the Declaration of Helsinki (as revised in Tokyo 2004). Written informed consent was obtained from all participants.
Inclusion criteria of clinical blood sample collection: ICP group (10 cases): 1) Maternal serum TBA level !10 mmol/l. 2) There were no other complications, such as gestational diabetes mellitus, gestational hypertension, hepatitis, pregnancy complicated with hyperthyroidism, etc.
3) The clinical symptoms and liver function of pregnant women returned to normal after delivery within 42 d. All pregnant women had no abnormal liver function before pregnancy and no history of liver-related operation, hepatitis B and hepatitis C. Normal group (10 cases): no other complications and the liver function were normal.

Extraction and ultrafiltration of exosomes
The plasma samples were taken out of the refrigerator at À80 C, thawed in the water bath at 25 C. According to the manufacturer's instructions, the qEV original kit (SP5, Izon Science, Christchurch, New Zealand) was used to enrich and purify plasma exosomes in the normal and ICP groups. According to the manufacturer's instructions, The 10 K ultrafiltration tube (Amicon ultra-0.5, Millipore, American) was used to ultrafiltrate exosomes.

Nanoparticle tracking detection (NTA)
Exosome samples were diluted, and the concentration and size distribution of exosomes were determined by nanoparticle tracking analysis (Zeta) (analysis software version Zeta View 8.04.02) on ZetaView S/N 17-310 (Particle Metrix, Inning am Ammersee, Germany).

Transmission electron microscope (TEM)
The samples were tested using a nano transmission electron microscope (TEM) at 80 kV of Tecnai G2 Spirit BioTwin (FEI, Hillsboro, OR).

Liquid chromatography-mass spectrometry (LC-MS) analysis
The polypeptide products were analyzed by QE-HF-X mass spectrometer and EASY-nLC 1200 nano-upgraded liquid chromatography (LC). Tandem mass spectrometry detection adopted data-dependent acquisition [DDA]) mode. The total scan resolution was 60,000 (FWHM), the range of mass-tocharge ratio was set to m/z 350-2000, and the collision energy was set to 28% in HCD fragmentation mode.

MaxQuant
The MaxQuant database was used to search for mass spectrometry data, and each original data and corresponding database were uploaded to max quant 1.5.8.3 institute for biochemistry, Germany. After searching the database, we removed the Reverse and Potential Contaminants proteins and made statistics on the search results.

Screening of differential proteins
According to the MaxLFQ algorithm embedded in MaxQuant software, this quantitative value (intensity) was calculated to obtain normalized quantitative value LFQ. Quantitative analysis of proteins in all samples was carried out according to LFQ values. In pairwise comparison between groups, the normalized signal mean of all samples in each group was calculated to calculate the ratio between groups. The p values of the two groups were calculated by Mann-Whitney rank-sum test. Proteins meeting the following two conditions were screened as differential proteins between groups: 1) ratio between groups > ¼ 1.5 or < ¼ 0.67 (i.e., 1/1.5); 2) p < .05.

Screening of differentially expressed proteins
For the experimental design with biological repetition (n ! 3), the data of samples detected by mass spectrometry were tested by T-test between groups, and two-parameter values, p value and Fold change, were obtained. The obtained differential proteins were processed by data to draw a Volcano plot.

Cluster analysis
Unsupervised hierarchical cluster analysis was performed for the differential proteins modulated between groups. Through protein expression data, the expression of differential proteins among multiple samples was calculated, and the expression of selected differential proteins calculated the direct correlation between samples.

GO enrichment analysis
The function of differentially expressed proteins between normal pregnant women and ICP pregnant women were annotated, and the primary biological process (BP), cellular component (CC), and molecular function (MF) of these differentially expressed proteins were analyzed by Gene Ontology enrichment.

Kegg pathway enrichment analysis
Kyoto encyclopedia of genes and genomes (Kegg) is the central database for systematic analysis of gene function, genome and proteome information. Different proteins performed their biological behaviors in coordination with each other, and protein and expression information were studied as a whole network.

Protein interaction network interaction analysis
Protein-protein interaction (PPI) is the most critical signal transduction and protein function in cells. PPI analysis of protein interaction network helps discover the protein which is in the core position of regulation among a large number of different proteins.

Screening and verification indicators
We consulted literature and Gene Cards, screened 15 differential proteins, and selected CLU, which was up-regulated as an index, to verify the experiment.

Maternal plasma sample collection of ICP
Inclusion criteria were the same as before (2.1 sample collection). Thirty pregnant women (aged 21-34 years) who visited the First Affiliated Hospital of Chongqing Medical University from March 2020 to December 2020 were randomly selected as research objects, including 10 cases of normal pregnant women and 20 cases of ICP pregnant women.

Extraction, identification and concentration of exosomes
The exosome was extracted, identified, and concentrated by the same experimental method as before, and finally, 35 concentrated exosome samples of 200 ul were obtained.
Detection of plasma exosome cluster protein CLU was detected by ELISA (R&D System, Minneapolis, MN, USA), reagents, samples and standards were prepared in turn according to the instructions, and read the OD value immediately at 450 nm (corrected at 570 nm).

Statistical analysis
GraphPad Prism version 6 (Graphpad Software Inc., La Jolla, CA) was used to calculate the t-test, and a histogram was drawn to compare the level of plasma exosome CLU between the normal and ICP groups. p < .05 indicates statistical significance. The results were statistically analyzed with IBM SPSS statistics version 23 software (IBM SPSS, Armonk, NY) and the ROC curve was drawn with 1specificity as abscissa and sensitivity as ordinate, and the area under the curve (AUC) was calculated.

Comparison of clinical data of subjects
The comparison of clinical data and laboratory examination is shown in Table 1. The statistical t-test showed that the gestational age of the ICP component was significantly smaller than that of the control group. TBA, ALT and AST in the ICP group were significantly higher than those in the control group. There was no significant difference in maternal age and TBIL between the two groups.

Extraction and identification of exosomes
To ensure the successful extraction of exosomes, we identified exosomes extracted from ICP and normal maternal plasma by TEM, NTA and Western Blot. The purified exosomes have a typical structure, and a clear vesicle structure can be seen under the electron microscope (Figure 1(A)). Conventional exosome markers CD63 and Hsp70 were detected in exosome samples (Figure 1(B)), Hsp70 and CD63 were positively expressed, and CD63 was relatively high, indicating that exosome extraction was successful. NTA showed that the particle size of the measured sample was about 100 nm, and the overall size of the sample met the standard (Figure 1(C)).

Data analysis of exosome proteins by mass spectrometry
According to MaxQuant search results, all samples' digestion efficiency was > 90%, which shows that the quality control of enzymatic hydrolysis was qualified. According to the mass spectrometry database, 591 non-redundant proteins, 5925 peptides (including redundant peptides), and 102,575 general spectrograms were detected. All identified proteins were mapped according to their relative molecular weight distribution (Figure 2(A)). Generally, it was considered that the protein with high credibility is a protein containing two unique peptide segments, and the distribution of the number of unique peptide segments is as follows (Figure 2(B)). The results showed that the number of proteins containing two unique peptide segments in this study is 463, accounting for 78.34% of the total protein. The distribution of peptide length (Figure 2(C)) showed that the maximum peptide length was 11, and the average length was 15.30. The protein coverage distribution (Figure 2(D)) showed that 31.13% of identified proteins were in the range of 0-10% coverage proteins, 48.9% identified proteins were in the range of

Differential protein analysis of exosomes
A total of 591 proteins were detected in this study. The results of the qualitative analysis showed that 34 specific proteins were detected in the ICP group, and 39 specific proteins were detected in the normal group. The protein quantitative analysis results showed that 51 proteins were up-regulated and 71 proteins were down-regulated. By combining qualitative analysis and quantitative analysis, 138 kinds of differential proteins were obtained in this study.

Bioinformatics analysis of mass spectrometry data
The screening results of exosome differentially expressed proteins were drawn into a Volcano plot (Figure 3(A)), and cluster analysis was carried out (Figure 3(B)).
GO enrichment analysis: ICP-related proteins were mainly enriched in biological functions, such as platelet degranulation, cell adhesion, etc. (Figure 4(A)), especially related to exosomes, extracellular regions (Figure 4(B)), primarily involved in protein binding, calcium ion binding and other MFs (Figure 4(C)).
KEGG pathway enrichment analysis: KEGG analysis found that differentially expressed proteins were mainly related to the PI3K-Akt signaling pathway, Rap1 signaling pathway, etc. The first 15 pathways were selected and plotted as follows (Figure 4(D)). Protein interaction analysis: 139 different proteins were analyzed by the protein interaction network in STRING online database. The results showed that ( Figure 5(A)) most of the proteins had an interaction relationship, and many proteins were located in the center of the network and were closely related to the surrounding proteins. The results were imported into Cytoscape ( Figure 5(B)). The size of the node represented the number of interacting proteins, the color represented FC value, the red represented up-regulation, the green represented down-regulation and the thickness of the connecting line represented the combined score. The centrality Degree was calculated in Cytoscape to evaluate the importance of GeneCards. The proteins were sorted according to the Degree value, and the Top15 proteins were selected for analysis ( Figure 6(A)). The functions of them through GeneCards are shown in Table 2.

Maternal plasma exosome CLU concentration of ICP pregnant women
The CLU concentration of plasma exosomes in the ICP group was significantly higher than that in the normal group (p < .0001) (Figure 6(B)). ROC curve analysis showed that the best critical point for diagnosing ICP according to the plasma exosome CLU concentration of pregnant women was 255.28 ng/ml, and its AUC was 0.995, sensitivity was 95%, and specificity was 100% (Table 3, Figure 6(C)). The best critical point of predicting ICP with premature delivery by plasma exosome CLU concentration in pregnant women with ICP is 286.72 ng/ml, with an AUC of 0.854, a sensitivity of 91.7%, and specificity of 72.2% (Table 3 and Figure 6(D)). NPY was significantly correlated with ALT. NPY had no correlation with TBA and AST (Figure 6(E-G)).

Discussion
The importance of exploring new biomarkers for clinical diagnosis of ICP ICP is a special complication during the second and third trimesters of pregnancy, which has a significant influence on perinatal outcomes. Early diagnosis and treatment of ICP may improve perinatal outcomes and reduce the risk of premature delivery and fetal death. At present, the commonly used indicator for clinical diagnosis of ICP is the serum TBA level of pregnant women. However, many studies have found that the accuracy of TBA in diagnosing ICP may be overestimated [7]. Taking TBA as the primary indicator of clinical diagnosis and prognosis of ICP has certain limitations. Many pieces of evidence showed that some pregnant women with ICP still suffered from fetal death, although maternal serum TBA level decreased to very low concentration or even normal after UDCA treatment. Therefore, it is very important to find more valuable biomarkers besides TBA for clinical diagnosis and prognosis judgment of pregnant women with ICP. One of the characteristics of exosomes is that they are enriched with cell-specific protein or miRNA, which can be used as a molecular marker for disease diagnosis and prognosis [11,12]. For example, Sardar et al. [13] found that neuron exosomes can be easily separated in blood and other body fluids and contain specific marker molecules, which can be used as diagnostic markers for Alzheimer's disease. Multiple studies [14,15] showed that exosome miRNA might be used as a marker for the diagnosis and prognosis of liver cancer. With the rapid development of gene detection technology and protein omics, it is possible to use exosomes in disease diagnosis. Therefore, this study explored the feasibility of exosomes as a diagnostic biomarker of ICP.
The signal pathway mediated by plasma exosome protein in ICP pregnant women may be involved in the pathophysiological process of ICP In this study, the exosome protein of ICP pregnant women was analyzed by KEGG. The results showed that the exosome protein of ICP pregnant women was closely related to the PI3K-Akt pathway and RaP1 signaling pathway. PI3K/Akt signaling pathway belongs to the casein kinase receptor transduction pathway. Many growth factors can activate PI3Kto produce lipid products and then activate downstream signal protein Akt. Akt regulates various cell functions through different downstream targets, including metabolism, protein synthesis, cell cycle progression, anti-apoptosis, tumor growth, and angiogenesis, and finally regulates cell proliferation, differentiation, migration and invasion. All of the above suggests that the PI3K-Akt pathway and RaP1 signal pathway may be involved in the pathophysiological process of ICP.
Maternal plasma exosome cluster protein can be used as a biomarker of clinical diagnosis and premature birth prediction of ICP pregnant women CLU is a multifunctional secretory glycoprotein, which exists in many different biological fluids, including urine, plasma, serum, semen, breast milk and cerebrospinal fluid. It plays an essential role in lipid metabolism regulation, cell proliferation, complement activity regulation and cell apoptosis. Scaltriti et al. and Lee et al. [16,17] found that CLU could promote cell apoptosis by preventing Bcl-xL from binding to Bax. In some cells, such as cardiomyocytes and neurons, CLU activates PI3K/AKT signaling pathway, which leads to cell survival or proliferation [18]. Oztas et al. [19] found that the increased expression of CLU in the placenta was related to IUGR. Shin et al. and Watanabe et al. [20,21] found that the expression of CLU in the placenta and serum of pregnant women with hypertensive disorder complicating pregnancy were significantly higher than those of normal pregnant women. Vac ınov a et al. [22] found that there was no significant difference in CLU level between gestational diabetes The abscissa is 20 sample names (10 normal samples and 10 ICP samples), and the ordinate is the screened differential protein. Red represents the high expression value of differential protein, while blue represents the low expression value of differential protein, which can show the expression level of different differential proteins among multiple samples.  mellitus and normal pregnant women. Musilova Ivana et al. [23] found that the concentration of CLU in the amniotic fluid of pregnant women with preterm premature rupture of membranes during pregnancy was lower than that of normal pregnant women. However, there is no previous report on the CLU level in ICP.
In this study, the ROC curve was used to evaluate the value of plasma exosome CLU in diagnosing ICP and predicting ICP with premature delivery. ROC curve analysis showed that the larger the Youden index or AUC indicated greater the diagnostic value. If AUC is greater than 0.9, it means a great diagnostic value. AUC 0.7-0.9 means medium diagnostic value; AUC 0.5-0.7 means poor diagnostic value. AUC less than 0.5 means no diagnostic value. Table 3 shows that the ROC curve AUC of plasma exosome CLU concentration in diagnosis of ICP was greater than 0.9, which meant good diagnostic value and high sensitivity and specificity. ROC curve AUC of maternal plasma exosome CLU concentration in diagnosis of ICP with premature delivery was between 0.7 and 0.9, which meant medium diagnostic value, high sensitivity, and low specificity. The results of this study showed that the best critical point for diagnosing ICP according to the plasma exosome CLU concentration of pregnant women was 255.28 ng/ml, and the best critical point for predicting ICP with premature delivery was 286.72 ng/ml.
In conclusion, plasma exosome analysis of ICP pregnant women demonstrated that the PI3K-Akt signal pathway and RaP1 signal pathway might be involved in the pathophysiological process of ICP. The plasma exosome CLU in pregnant women with ICP is an important biomarker for clinical diagnosis and prediction of premature delivery of ICP.