Combination Of LCM And 2D-LC-MS/MS Analysis To Identify Protein Expression Profiles And Biological Pathways In Low Malignant BTCC

Objective:Bladder transitional cell carcinoma (BTCC) is the most common urologic tumor inChina. Recent studies have found that, from biological behavior to pathologicalcharacteristics, BTCC actually has two phenotypes: low malignant and aggressive, theso-called two type bladder transitional cell carcinomas or two-tie grading system. Thisnew classification provides the new molecular base for the study of bladder cancer.Further study may elucidate the molecular mechanism of bladder carcinogenesis andprovide potential clinical diagnosistic and therapeutic targets, and have importantsignificance. This study is for identifying the molecular mechanism, biomarkers,differentially expressed proteins and further elucidating the carcinogenesismechanism of BTCC on the basis of two-fie grading system, this is also one of urgentdifficulties in bladder cancer research area.Currently the main technique for high-flux proteomics research is two-dimensionalgel electrophoreus, (2-DE). 2DE technique can separate proteins by differentcharacters of the proteins. The first dimensional separation is pH gradientelectrophoresis (IEF), proteins are separated by different isoelectric point. Thesecond dimensional separation is sodium dodecyl sulfate polyacrylamide gelelectrophoresis (SDS); proteins are separated by different molecular weight. Aftertwo-dimensional separation, the spot on the 2-dimensional plane represent a protein.By this step, the proteins are separated, meanwhile, the theoretical molecular weight,isoelectric point and some other related information can be gained. Several two-dimensional electrophoresis(2-DE) based proteomic BTCC researches have beenreported, while only few differential expressed proteins were identified and can notpresent the high-flux of the proteomics technique. Additionally, several classes ofproteins including poorly soluble proteins, proteins with extremes of pI and molecularweight (MW), low abundance proteins and integral membrane proteins tend to bepoorly represented in 2-DE analyses. This suggests that alternative approaches areneeded to increase coverage of an expressed proteome. With the development oftechnology, some new emerging technique, such as two-dimensional colour spectrum,two-dimensional capillary electrophoresis, liquid chromatography-capillaryelectrophoresis was used as the supplement and displacement of traditional 2DE.The sample size used in these new techniques was fewer than 2DE, while theseparating efficient was higher than 2DE. After consider the shortcoming of 2DE andthe progression of proteomics technology carefully, we decided to applymass-spectrum shotgun method, namely two-dimensional liquid chromatography inconjunction with tandem mass spectrometer and to the directly identification ofproteins from mixture of polypeptides. Further reason for the shotgun strategy is theconsideration of LCM combined with proteomics technique. The smaller ample sizein the shotgun strategy makes the combination possible.Methods:1. Harvest the normal urothelium and cancer cells 250, 000shots by Pix CellⅡLaserCapture Micro dissection System, respectively. Extract 185.2μg and 147.6μgrespectively after digestion and make the mass spectroscopy analysis possible.2. Protein expression profiles of normal urothelium and cancer cells were identifiedby Thermo Finnigan linear ion trap tandem mass spectrometer and liquidchromatography. The experimental data were compared with presently testifieddifferentially expressed proteins between normal urothelium and cancer. 2differentially expressed proteins were selected to validation by immunoblot andimmunohistochemistry. 3. Bioinformatics tools analyze the identified proteins: the basic physico-chemicalproperty of the identified proteins was achieved by protein analysis tools, includeGRAVY, prediction of transmembrane, molecular weight and isoelectric point.4. In order to gain the feature of the protein biomarkers, evidence based biomedicineexploration was performed based on the proteins that identified in our research.5. Expression profiles of tumor and differentially expressed proteins were analyzed byGO software, as well as GO enrichment/depletion analysis. Cellular biology wasexplored based on the analysis outcome, as well as compared with the urine proteomeGO enrichment/depletion analysis. By the contrast, some proteins that trend to appearin urine were determined.5. Construction of database: open access proteomics database was constructed byDreamweave and XML editor software. The contents of our database include thereports about the identified proteins produced by TurboSequest software and results ofanalysis by Gene Ontology (GO). The identified proteins related data are in extensiblemarkup language (XML) format.6. Pathway analysis and visualization. Pathway analysis was performed byArrayTrack software. SWISSPORT-ACC-NUMBER of the differentially expressedproteins were input into the ArrayTrack V.3.3.0, Pathway visualization was finishedby GenMAPP software. The same procedure was performed in high malignant BTCC.Results:1. 440 proteins were identified in 4 tumor specimens. Among the 440 proteins, 24proteins were considered as hypothetical proteins through database searching,hydrophobic proteins occupied 5.5% (24 proteins). There were 76 (17.4%) proteinswith MW＜10 kDa or MW＞100 kDa, 84(19.1%) proteins with pI＞9.38 (8.6%)proteins with at least one transmembrane helices. 218 proteins were identified innormal specimens. 19 proteins were considered as hypothetical proteins, hydrophobicproteins occupied 5.3% (11 proteins). There were 42 (19.3%) proteins with MW＜10kDa or MW＞100 kDa, 40 (18.3%) proteins with pI＞9, 13 (6.0%) proteins with at least one transmembrane helics. There are 388 differentially expressed proteinsbetween tumor and normal tissue, among them 305 specifically expressed in tumorcells, 88 proteins specifically expressed in normal cells. The smallest and largest MWvalues observed in differentially expressed proteins were 7.86 and 1005.20 kDa, andthe proteins were distributed across a wide PI range (3.67-11.91).2. Immunoblot results: TPD52 was specific expressed in tumor (P＜0.01) and theexpression level of NHERF between normal and cancer tissue exist difference(P＜0.05). Immunohistochemistry results: Immunohistochemistry analysis on theexpression of TPD52 and NHERF shows that there are high positive expression ofTPD52 and NHERF compared with corresponding normal bladder mucosa (P＜0.05).Furthermore, the high expression of TPD52 has correlated with increased pathologicalgrade, while irrespective with pathological stage. The high expression of NHERFhave no correlated with pathological grade and stage.3. A total of 129 proteins were defined as biomarkers and the others were not. Themean value±standard deviation of PI, GRAVY, MW, and predicted transmembranehelices had no statistical significance between the two groups. If divided the proteinsinto two groups by PI, 71 proteins PI＞9 and 317 proteins PI≤9, 21 and 108 proteinswere defined as biomarkers in the two groups. The mean value±standard deviation ofPI of the biomarkers and other proteins in the according groups had no statisticalsignificance. Same outcome appeared in the groups by MW. Chi-square test showedthat the proteins had same chance to be a biomarker and irrespective with the PI andMW, (P＞0.05). Multivariate analysis of PI, gravy, mw transmembrane showed that allof these variables have no relation with the biomarker.4. In total, 41/22, 25/11, 23/20 exhibited significantly as overrepresented andunderrepresented terms compared with the entire list of International Protein Index(IPI) entries (IPI_Human, versions 3.13, 57050 protein sequences) in tumorproteomics expression profile. Of all the 388 differentially expressed proteins, 267proteins (68.8%) with biological process annotation; 256 proteins (66.0%) withcellular component annotation; 297 proteins (76.5%) with molecular functionannotation as to GO slim v1.8. Among them 10 proteins functional in cell adhesion; 13 proteins functional in cell proliferation; 9 proteins functional in cell cycle; 10proteins functional in cell differentiation; 37 proteins functional in cell signaltransduction. For further details of the analyses by GO, please refer to our database.5. The database includes experiment conditions; the list of identified proteins; rawmaterials for the identification; the analysis as to GO. Through identification ofprotein by MS/MS analyses and protein database searches, basic descriptions ofidentified proteins including protein names, theoretical MW, pI value, the theGRAVY value, the proteins with at least one transmembrane helices were presentedin the database as well as the accession IDs for protein database, include IPI,Swiss-port, TrEMBL, RefSeQ, Ensembl and H-Inv, etc. These data are in extensiblemarkup language (XML) format. Our GO database were presented in table form andconstructed to describe known molecular function, subcelluar location and biologicalrole of the identified proteins. The identified proteins' IPI accession numbers werecorrelated with Go terms in GO hierarchy by database retrieve. Therefore, the proteinssharing same GO term within the hierarchy were grouped into several clusters,functional distribution and the enrichment/depletion of the function classes weredemonstrated in the table too. From our GO database, users can specify the GO termsat any level within the hierarchy or specify the proteins they are interested in, and caneasily access GO terms deposited at QuickGO (http://www.ebi.ac.uk/ego/) or accessthe protein information deposited at European Bioinformatics Institute (http://ebi.ac.uk/swissprot/) by hyperlink. In order to facilitate other scientists refer to ourexpression profile, we offer a query interface in the database, users can query proteinsby the multiple accession IDs. Our database has been deposited in an accessible formto researchers at http://www.Proteome-NHTE.org.cn, andhttp://www.Proteome-SBTCC.org.cn.6. Functional clustering of the 129 proteins that differentially expressed in lowmalignant BTCC ArrayTrack software in the context of biological pathways revealedthat several pathways especially oxidative phosphorylation, ECM-receptor interaction:Toll-like receptor pathway, Cell Communication, Focal adhesion, MAPK signalingpathway, etc., are actively in the carcinogenesis and/or progression of the low malignant BTCC. Functional clustering of the 181 proteins that differentiallyexpressed in high malignant BTCC revealed almost the same changed pathways,while the proteins located in the same pathway were quite difference. The outcome ofthe pathway analysis give us some cues that multi-protein and multi-pathwayelucidate the molecular mechanism of bladder carcinogenesis, furthermore themechanism that drive the two type bladder transitional cell carcinoma is quitedifferent.Conclusion:1. Combining laser capture microdissection and proteomics technique is an effectiveapproach to identify protein expression profiles and differentially expressed profile.There were 388 differentially expressed proteins between normal urothelium andbladder cancer. Among them, 305 proteins were specifically expressed in cancer, and83 proteins were specifically expressed in normal urothelium. These differentiallyexpressed proteins indicate that many proteins abnormally expressed in thecarcinogenesis. The differentially expressed proteins offer numerous usefulinformation for the research of bladder cancer molecular mechanism.2. Two differentially expressed proteins were selected to validation by immunoblotand showed that there existed significant expression difference between normal andcancer tissue. Meanwhile these two proteins were identified only by one unipeptide.The outcome of immunoblot testified the combination of LCM and 2D-LC-MS/MSwas an effective approach to identify differential protein expression profiles.Immunohistochemistry analysis on the expression of TPD52 and NHERF show thatthere are high positive expressions of TPD52 and NHERF compared withcorresponding normal bladder mucosa. (P＜0.05). Furthermore, the high expressionof TPD52 has correlated with increased pathological grade, while irrespective withpathological stage. The high expression of NHERF have no correlated withpathological grade and stage.3. PI and MW can not be the predictive factor of the protein biomarker, some specialprocedure should be developed to guide the secretively cut the differentiallyexpressed spots in 2DE procedure. Shotgun strategy not only enlarges the coverage of the identified proteins but also enlarge the coverage of protein biomarkers. The basicphysical chemistry description include PI, MW, prediction of transmembrane heliceshave no predictive significant for the discovery of biomarker. To predict thepossibility of an unknown protein to be a biomarker, further bioinformatics researchshould be done.4. Theoretical evident for the further study on the cell biology and mechanism ofcarcinogenesis were achieved by the GO enrichment/depletion analysis. Protein typesthat tended to appear in the urine were discovered by document retrieve. Theoverlapped protein types in differentially expressed proteins and the urine proteomicsexpression profile provided evidents for the development of small protein chip fornoninvasive detection of BTCC and gave more witnesses for the reliability of ourexperiment.5. According to the convention of international proteomics study, we constructed theopen access database based on the protein identify information and functional analysis.Construction of database is decisive that enables laboratories to communicateefficiently and to compare data. From our GO database, users can specify the GOterms at any level within the hierarchy or specify the proteins they are interested in,and can easily access GO terms deposited at QuickGO (http://www.ebi.ac.uk/ego/) oraccess the protein information deposited at European Bioinformatics Institute (http://ebi.ac.uk/swissprot/) by hyperlink.6. Functional clustering of the differentially expressed proteins in low and highmalignant BTCC by ArrayTrack software in the context of biological pathwaysrevealed that almost the same pathways changed in the two type BTCC. Thesepathways include oxidative phosphorylation, ECM-receptor interaction, Toll-likereceptor pathway, Cell Communication, Focal adhesion, MAPK signaling pathway,etc. While the proteins located in the same pathway were quite difference betweentwo type BTCC. The changed pathways may be significant in the development andprogression of two type BTCC. The pathway analysis can promote there-consideration for oncology on the eye of systematic biology. Furthermore thedifferentially expressed proteins in the same pathway construct the basis for protein chip to monitor the progression and evaluate the prognosis of BTCC.7. Further study on these differentially expressed proteins and underlying biologicalpathways may elucidate the molecular mechanism of bladder carcinogenesis andprovide potential clinical diagnosistic and therapeutic targets.