Liver Parenchymal Cells, Protein Expression Spectrum Construction And Function Analysis

The liver is the largest gland in the human body. Its main functions include metabolism,detoxification, defense, hormone system homeostasis, blood storage and pH regulation,hematopoiesis and immune regulation. The liver is mainly composed of hepatocytes (HC), liversinusoidal endothelial cells (LSEC), Kupffer cells (KC) and stellate cells (HSC). Cells are thebasis for complex functioning of the liver. Different types of hepatic cells perform distinctfunctions. However, the liver proteome have not been viewed from the cellular perspective yet.Thanks to the high-resolution mass spectrometer, high-throughput and large-scale identificationof eukaryotic proteome has become possible. Cell-based hepatic proteome profiling is anecessity for comprehensive understanding of cell-specific functions. It's one of the major aimsof proteomic research, and also an important part of human liver proteome project (HLPP).Our study aims to simultaneously isolate and purify the four types of cells within liver, andto construct protein expression profiles. As the main cell type in liver, the HC proteome wascomprehensively analyzed to gain insights into its function. We analyzed cell protein profilesthrough systems biology method, and discovered a set of molecules related to cell-type specificfunction. We compared the CD molecules identified in the liver cells to look for cell-specificsurface markers. We also compared KEGG pathways coverage between these cells, and foundthe pyruvate metabolism pathway that may be complementary between parenchymal andnon-parenchymal cells. This would help hint the cooperation of liver cells in metabolism.Moreover, several components of the pathway were further validated, providing evidence forour hypothesis. The cellular fractionation improved liver proteome coverage and many novelproteins are identified, some of which lack functional information. We utilized3state-of-the-artmethods to annotate their functions.Firstly, hepatic cells were isolated by a combination of collagenase-based density gradientcentrifugation and magnetic activated cell sorting (MACS) with high purity and yield for thefirst time. Cell purity was assessed by cytology and immunoblot methods. The successfulconstruction of cell sample preparation methods ensured enough pure cells for downstreamprotein mapping. We used C57BL/6J mouse liver to isolate HC, LSEC, KC and HSC at thesame time.Secondly, a label-free quantitative proteome procedure was established based onSDS-PAGE, in-gel digestion and LTQ-FT MS identification. After screening, bovine serumalbumin was chosen as internal standard protein. To a certain extent, the MS response signalcorrelated linearly with the amounts of lysate loaded onto SDS-PAGE gels at a complex samplebackground. A model for label-free quantitative proteome method was successfully established.APEX software was used to analyze the MS data. The above software ultilized normalizedspectral counting to estimate the relative abundance of proteins. The batch effect of experimentand ionization efficiency variation caused by difference in physicochemical property of peptideswere normalized using a machine learning algorithm, so proteome can be relative accuratelyquantified. Using the criteria of two-peptide identification and95%confidence level, a total of8,060proteins were found in HC with4,842proteins quantified. Thirdly, protein abundance distribution analysis revealed that metabolic enzymes were ofhigh abundance, indicating active metabolism ability of HC. The physiologically importantsub-proteomes including CYP450family proteins, glycogen-assciated proteins,polarity-associated proteins and kinases were analyzed by literature mining and GO annotation.4CYP450proteins and9polarity-related proteins were firstly identified in liver. Besides, thesesub-proteomes may be involved in many biological functions according to protein-proteininteraction network analysis. Compared with plasma proteome data, a list of proteins werehighly expressed in HC. These proteins can be serous candidates for liver injury. Integratedtranscriptome and proteome analysis discovered that mRNA and protein abundance correlationsbetween these core cellular genes that maintaining survival were relative high, while regulatorygenes were more lower. Transaminase, ribosomal and chromosomal genes showed nocorrelation.Fourthly, cell-specific gene clustes were identified from liver transcriptome and proteomedata. Analyzing these clusters by GO and KEGG, we found that HC play roles in metabolism,LSEC execute substance exchange, KC perfrom immune response and HSC participate inextracellular matrix formation, displaying a cell-specific pattern. Using Mammalian PhenotypeOntology (MPO) database, we found distinct physiological and pathological function betweendifferent cell types. The identified cluster differentiation molecules on cell surface werecompared between cells and a set of cell-specific proteins were found, which may serve asmembrane biomarkers deserve further verification.Sequently,3diferent function annotation methods named Blast2GO, weighted geneco-expression analysis (WGCNA) and Endeavour, which were based on protein sequence, geneexpression and deposited publicly available databases respectively, were applied to predict2,944functional unknown proteins. Among these,82were integral membrane proteins,59werenuclear proteins,40were cytosol proteins and35were mitochondrial proteins. Our analysisprovides a paradigm for annotation using various data resources.Lastly, pathway level of protein expression was compared in hepatic cells, and pyruvatemetabolic pathway may be performed by cooperation of HC and LSEC cells. These cell-specificcandidates were validated. Western blotting confirmed the cell-specific expression of APOA4,ACYP1, ACYP2, COLEC10, DOK2, IFNGR1and ME2. APOA4was HC specific protein.ACYP1,ACYP2and ME2were non-parenchymal specific proteins, DOK2and IFNGR1wereHC and KC specific proteins, COLEC10was mainly expressed in HC and HSC. Along withtranscriptome data, the specificity of these molecules were confirmed. These cell-specificmarkers reflect their specific function, which may provide cellular information for pathogenesisin liver disease.