| Background and Objectives Nephroblastoma(Wilms' Tumor,WT) is frequent one of pediatric malign tumor ,which incidence of children ,lower than 15 years old in the world, is 1/10000,wich it is first reported by wilms of German Doctor in 1899 ,so the authors named the disease as wilms' tumor. Nephroblastoma found by children malign tumor has risen from 8% to 10.7% by 1980s, which it has been the main mortuary of children. Clinical pathological stage and treatment methods is great relative with prognosis, NWTS reported in the summary of 2001: there is 7.0% rate of missed diagnosis by means of the current imaging (CT), in particular less than 3.0 cm in diameter of WT .Therefore, Surface Enhanced Laser Desorption /Ionization time of flight Mass Spectrometry(SELDI-TOF-MS) is an original technique of protein research ,which has a merit of fast ,convenient , high-flux and parallel-test multi-sample. Mass-spectrum and bioinformatics provide a utility way of sifting specific markers. What Ciphergen Company take pain to study the technique of SELDI-TOF-MS is a original proteomics investigation , which it applies a few natural protein extractive to analyses quickly and overcome the limit of traditional analysis, SELDI-TOF-MS has the merit of little sample ,simple operation and high sensitivity and high-flux, and succeed to apply the malignant tumor diagnosis of ovarian cancer ,prostate carcinoma, mammary adenocarcinoma, lung cancer and rectal cancer and sift the tumor markers and other proteomics study. Some cases apply SELDI-TOF-MS combined bioinformatics with Support Vector machine (SVM) to sift the specific protein markers of nephroblastoma and establish the serum protein finger printing of clinical stage in nephroblastoma and contrast clinical stage to CT stage.Nephroblastoma (Wilms' Tumor, WT) prognosis is closely relative with many factors such as different age of onset, pathologic type, disease local, clinical stage, treatment method etc, which cause to great influence of the prognosis. It is important factors to influence the prognosis, in particular clinical stage and pathologic type. The prognosis has nothing to do with the size of tumor. Because of the limit of diagnosis means and level, no means can accurately detect the survival tumor cell numbers after the tumor has cured completely and relieved finally by chemotherapy and radiotherapy today. Therefore, it is urgent to create a sensitive and specific detecting means to testify the survival tumor cells will evaluate the effect-rate accurately and avoid bringing about the toxic and side effect by treatment means and monitor the prognosis and predict the recurrence rate. But SELDI-TOF-MS have the character of fast, simple, high-flux, which can separate the different protein mass spectra by healthy tissue and body fluid. In the article, the technique of SELDI-TOF-MS is applied to sift specific protein markers combining bio-information and SVM technique can monitor the nephroblastoma prognosis and erect serum protein mass spectra model.Materials and methods1. The serum samples eighty cases come from pediatric surgery in the First Affiliated Hospital Medicine College of Zhengzhou University in China, including thirty preoperative nephroblastoma, thirty other malignant solid tumor (sixteen neuroblastoma, four hepatoblastoma, four renal rhabdosarcoma, four malignant teratoma and two pancreablastoma ), 20 healthy child of the kindergarten. Follow-ingsup: 2001.1-2005.12. All malignant tumors were confirmed by immuno-histochemistology and testified by several pathologic professors. Preoperative nephro -blastoma is made of twenty-one males and nine females(average age: 2.10±0.10 years old); other malignant solid tumors is made of eighteen males and twelve females(average age: 2.50±0.16 years old); Age and gender of control group is matched with nephroblastoma. All the whole blood preparations is taken suction in the empty stomach in the morning before being operated, and set one hour in 4℃and centrifuged with 3000rpm in ten minutes , extracted the serum and deported into -80℃deep freeze refrigerator. CT projection methods: Simens SCT ;scan range from anodic diaphragmatic muscle to inferior pole of kidney, including interval eight mm; thick five mm, important local interval five mm and five thick ;scan time :two seconds .Contrast medium: Omnipaque 1.5-2.0ml/kg.2.Main instrument and reagent: Ross Temp Inc, the Ciphergen PBS II SELDI-TOF- MS and WCX2 protein chip, ORBIT300,Nuaire, Bioprocessor, 1296-003, DELFIAPLATESHAKE of Wallac, which is purchased from the American company of Ciphergen, but Milli-Q Plus is bought from Millipore inc company; all the SPA( the Sinapinic acid) CHAPS, Urea, DTT, NaAc is purchased from Promega company in America.3The technique operation route of SELDI: First the samples melt on the ice,4°10000 rpm centrifuge 2 min; take 96 bore planks, put the ice box up, add the U9( the 9 M Urea,2% CHAPS,1% DTT)10μl and the serum 5μl on the each bore,4℃laminar analysis cabinet 600 rpm, swing 30 minutes; do the protein chip processing before end to swing in 15 minutes, build the chips into the Bio-processor and write down the chip orders ;the first step adds 200μl NaAC(100 Mms, pH4)into bore,600 rpm, 5 minutes in 4℃laminar analysis cabinet, repeat once; the processing 96 bores put the ice up, add 185ulNaAC quickly 6000 rpm in 4℃laminar analysis cabinet, swing 2 min ; add 100μl proceeded samples to chip, 4600 rpm one hour in 4℃laminar analysis cabinet, throw away and keep drying, repeat three times; Use the200μl hplc water 2 times clapping quickly; Repeat two times with 1μl 50% saturation SPA; start up the examination; put the chip into the Ciphergen reads to compose the instrument the examination.4. The data collect to use with processing: All data is revised by the protein chip software 3.1 to make the intensity and molecular mass of the total ion uniformed and analyzed by ZUCI-Protein Chip Data Analyze System software. Find out the sample with the method that part is worthy with each from of peak, and filter to believe the sound to compare small in 5 peaks. At this time, the peaks and m/ z values from each sample are dissimilarities. Cluster analysis takes 10% as MT, the lower 0.3% peak from each sample in the different m/z is uniformed. Support the vector machine (SVM) specific installing: Adopt the radial based kernel, Gamma: 0.6, punishing -dividing function(C):19.To choose character vector The selection of the characteristic vector the adoption statistics filter to combine the method establishment of the model dependence sieving discretion model, use stay a method to cross verification conduct and actions to evaluate the model discretion result.6. Statistics Analysis: The covariance sieves the quality lotus of to compare the peak to do the Wilconxon and examine to all first steps after learn analyze all proteins to compose the original data to pass by to filter to drop to believe the sound and gather an analysis processing, selecting 10 analysis with further peak of the p value least. 10 peaks combine arbitrarily (total 175 kinds of combinations)and input the support vector machine model, which it can stay the estimate result that a method evaluates the model(α=0.01)Results1. Nephroblastoma stage I group and nephroblastoma stage II group: nephroblastoma stage I group separate from nephroblastoma stage II group and filter in the first step and get 483 M/Z peaks , 5 M/ Z peaks (P<0.01) is obtained to the relative intensity by Wilconxon analysis. From the random combination of the obvious different protein peaks medium, the tallest grouping model of Youden index is sifted sift three markers M/ Z located in 7965.4,5022.4 and 8469.6 by SVM, nephroblastoma stage II is low expression, nephroblastoma stage I is high expression. The importation value is put two latent markers into, using to stay a method to cross the examination, the sensitivity is 80.00% and specificity is 100.00%.2. Nephroblastoma stage I group and nephroblastoma stage HI group: nephroblastoma stage I group separate from nephroblastoma stage III group and filter in the first step and get 496 M/ Z peaks ,4M/Z peaks (P<0.01) is obtained to the relative intensity by Wilconxon analysis. From the random combination of the obvious different protein peaks medium, the tallest grouping model of Youden index is sifted sift four markers M/ Z located in 4263.1,4122.8,4330.7and 4303.7 by SVM, nephroblastoma stage III is lower expression, nephroblasto -ma stage I is high expression. The importation value is put two latent markers into, using to stay a method to cross the examination, the sensitivity is 100.00% and specificity is 100.00%.3.Nephroblastoma stage I group and nephroblastoma stage IV group: nephroblastoma stage I group separate from nephroblastoma stage IV group and filter in the first step and get 565 M/ Z peaks , 4 M/ Z peaks (P<0.01) is obtained to the relative intensity by Wilconxon analysis. From the random combination of the obvious different protein peaks medium, the tallest grouping model of Youden index is sifted sift two markers M/ Z located in 10836.6 and 8179.1 by SVM, nephroblastoma stage IV is low expression , nephroblastoma stage I is high expres -sion. The importation value is put two latent markers into, using to stay a method to cross the examination, the sensitivity is 88.89% and specificity is 100.00%.4. Nephroblastoma stage II group and nephroblastoma stage III group: nephroblastoma stage II group separate from nephroblastoma stage III group and filter in the first step and get 490 M/ Z peaks ,3M/Z peaks (P<0.01) is obtained to the relative intensity by Wilconxon analysis. From the random combination of the obvious different protein peaks medium, the tallest grouping model of Youden index is sifted sift two markers M/ Z located in 5019.2 and 4143.2 by SVM, nephroblastoma stage III is low expression , nephroblastoma stage II is high expression. The importation value is put two latent markers into, using to stay a method to cross the examination, the sensitivity is 88.89% and specificity is 100.00%.5. Nephroblastoma stage II group and nephroblastoma stage IV group: nephroblastoma stage II group separate from nephroblastoma stage IV group and filter in the first step and get 508 M/ Z peaks , 20 M/ Z peaks (P<0.01) is obtained to the relative intensity by Wilconxon analysis. From the random combination of the obvious different protein peaks medium, the tallest grouping model of Youden index is sifted sift one markers M/ Z located in 7976.5 by SVM, nephroblastoma stage IV is low expression , nephroblastoma stage II is high expression. The importation value is put two latent markers into, using to stay a method to cross the examination, the sensitivity is 100.00% and specificity is 100.00%.6. Nephroblastoma stage III group and nephroblastoma stage IV group: nephroblastoma stage HI group separate from nephroblastoma stage IV group and filter in the first step and get 504 M/ Z peaks , 12 M/ Z peaks (P<0.01) is obtained to the relative intensity by Wilconxon analysis. From the random combination of the obvious different protein peaks medium, the tallest grouping model of Youden index is sifted sift one markers M/ Z located in 8194.4 by SVM, nephroblastoma stage IV is low expression, nephroblastoma stage III is high expression. The importation value is put two latent markers into, using to stay a method to cross the examination, the sensitivity is 93.75% and specificity is 100.00%.7.Nephroblastoma stage I+II group and nephroblastoma stage III+IV group: nephro -blastoma stage I+II group separate from nephroblastoma stage III+IV group and filter in the first step and get 519 M/ Z peaks ,3M/Z peaks (P<0.01) is obtained to the relative intensity by Wilconxon analysis. From the random combination of the obvious different protein peaks medium, the tallest grouping model of Youden index is sifted sift three markers M/ Z located in 4153.9, 3257.6 and 3290.7 by SVM, nephroblastoma stage III+IV is low expression , nephro -blastoma stage I+II is high expression. The important value is put two latent markers into, using to stay a method to cross the examination, the sensitivity is 83.33% and specificity is 93.75 %.8. Contrasting to all the clinical stage: The serum markers from the early diagnosis model of nephroblastoma in 6954.4 and 6455.5 m/z is analyzed that stage IV is lower expression than stage III; stage III is lower expression than stage II; stage II is lower expression than stage I; stage I is lower expression than healthy children; the latter is higher expression. The more latter clinical stage is, the lower expression is.9. Blind-checking the accuracy between CT and protein chip stage: protein chip stage as follows: stage I 6 cases; stage II 10 cases; stage III 10 cases; stage IV4 cases; all is concord with pathologic stage and the accuracy is 100%. However, stage CT: one of pathologic stage IV is misclassified to stage III; one of pathologic stage III is misclassified to stage II; all the stage accuracy is 100%; 80%; 80%; 75% ; respectively.ConclusionI. Many stage discrimination model located in 5022.4 Da,7965.4 Da,8469.6 Da,4303.7 Da,4122.8 Da,4330.7 Da,4785.6 Da,10836.6 Da,8179.1 Da,5019.2 Da,4143.2 Da,7976.5 Da,8194.4 Da,3290.7 Da,4153.9 Da and 3257.6 Da can solve the puzzle of clinical staging and make up a seasonable therapeutic and clinical stage accurately.II. Using SELDI-TOF-MS can detect the different protein markers between healthy children and nephroblastoma stage I,H,III,IV and accurately complete nephroblastoma clinical stage .The intensity of m/z is decrease by turns and the extent of decrease is obviously different(P<0.05). Healthy children and nephroblastoma stage I is high expression ,but the latter is more lower expression of healthy children .The more higher intensity ,the more lower stage and better prognosis.III .Blind-checking model: protein chip stage is more accurate than stage CT and concord with pathologic results, so it can make up for the defect and deficient of stage CT in the bioinformatics qualitation .Detecting the clinical stage model can succeed to complete the clinical staging and quotation before operation of nephrablastoma and perfect the clinical auxiliary examination.IV. Mass-spectrum combined with SVM, sifted the m/z peaks by high specificity and sensitivity protein markers from thousands of peaks, can provide a original way and according in the specific protein separation and purification of nephroblastoma.
|
PDF Full Text Request