| BackgroundNephroblastoma, also called Wilms tumor, is the most common renal malignant tumor occurring in children. Since the 1990s, the overall 5-year survival rate of nephroblastoma is about 80.0%; the surgical resection rate merely 100.0%. Wilms tumor is congenital, and children also have their own characteristics can not be timely and accurate perception illness currently. Existing ultrasound, CT, Intravenous urography impact (IVP) and other means of inspection only when the tumor grew to a certain size can be found. The existed test facilities could not meet the demands of non-invasive, simple, efficient, rapid diagnosis. There is an urgent need to search for more ideal biomarkers and seek new diagnostic approaches especially for early-stage nephroblastoma. Emergency of spectrum and bioinformatics technology could offer effective methods for selecting specific tumor markers. In this study SELDI-TOF-MS was used to select specific serum protein markers of nephroblastoma in children, and support vector machine (SVM) was combined with SELDI-TOF-MS to establish serum protein fingerprint model. We explored the application of these methods in preoperative diagnosis of nephroblastoma in children.ObjectivesTo find new biomarkers and to establish serum protein fingerprint models for early detection and diagnosis of nephroblastoma by SELDI-TOF-MS and bioinformatics tools.Subjects and methods 1. Subjects we collected 55 patients from Department of pediatric surgery of our hospital in 2005-2006. Among them, 30 patients had nephroblastoma. There are16 male and 14 female, the average age was 2.8±0.1 years. 25 patients had other abdominal solid tumors (9 patients had neuroblastoma, 5 hepatoblastoma, 2 pancreatoblastoma, 5 malignant teratoma, 3 rhabdosarcoma, 1 adrenocortical carcinoma). There are 11 male and 14 female, the average age was 3.5±0.1years. All tumor samples were confirmed by pathology. 20 healthy children selected from physical examination of pediatric outpatients' surgery were in the normal control group. The age and sex of the control group and nephroblastoma group were matched. All the tumor blood species were all drawn at the 15d before operation and the whole blood specimens are drawn on empty stomach in the morning, and after the specimens are placed under room temperature for 1-2 hours, they are turned centrifugally for 10 minutes, and then the serum are extracted and preserved under -80°C.2. Main reagents and devices CHAPS, Urea, DTT, NaAC, SPA (Sinapinic acid), they are all purchased from the Pomega Company of USA. Ciphergen PBS II ~+SELDI -TOF-MS and WCX2 protein chips are purchased from Ciphergen Company ofUSA.3. Protein chip technical route To unfreeze the serum specimens in ice bath, turn them centrifugally at 10000 rpm under 4°C for 2 minutes. Place a 96-hole plate on the ice box, add 10 ul of U9 (9MUrea, 2%CHAPS, 1%DTT) and 5 ul of serum to each hole, vibrate at 600 rpm under 4°C for 30 minutes in cold lab chamber. Before 15 minutes of finishing the vibration, pre-process the chip. The chip is placed in the Bioprocessor, note down the chip number, add 200 ul of NaNC (100 Mm, pH4) to each hole, vibrate at 600 rpm for 2 minutes in cold lab chamber, and repeat the above-mentioned operation once. Place the 96-hole plate being processed by U9 on the ice, use a medical gun to add 185 ul of NaNC, vibrate at 600 rpm under 4°C for 2 minutes in cold lab chamber. Add 100 ul of processed specimens to the chip, place them in the cold lab chamber under 4°C combining 600 rpm for 60 minutes, swing off the remaining liquid and dry out rapidly. Add 200 ul of NaAC, after vibrating at 600 rpm for 5 minutes, swing off and dry out, repeat this operation for three times. Wash each hole twice using 200 ul of deionized water, swing off the excessive water. After the chip is air dried, each hole is added 50% saturated SPA1 ul in two stages, after drying, place them on the device for testing.4. Data collection and processing Use a protein chip whose molecular weight is known to adjust the SELDI-TOF-MS system, until the tolerance of molecular weight is less than 0.1 %. Use mass spectrum reader to analyze the WCX2 protein chip combined with protein. Analysis parameters: laser strength is 170, sensitivity is 6, and the total number of each specimen collection is 140 times. The scope of data collection is 1000-30000 daltons; the optimized scope is 2000-20000 daltons. Use quality control serum to make repeated tests, the coefficients of variation of the peak value and the strength are 0.05% and 19.7% respectively.5. Support vector machine (SVM) Use linear SVM classifier. The selection of feature vector uses the method of statistical filtration combining with model dependent screening, to build discrimination model, use the method of leave-one-out crossing verification to assess the discrimination result of the model.This experiment also uses discrimination analysis method to process mass spectrum data, to verify the result being processed by SVM.6. Statistical analysis After the original data of mass spectrum have been filtered out the noise and after clustering analysis, the mass-charge-ratio peaks data being preliminary screened out is carried out Wilcoxon rank sum test, the testing standard is set atα=0.01.Results1. Tumor group and normal control groupAfter the mass spectrum data of tumor group and normal control group have been preliminary filtered and screened, 521 peaks of m/z were attained. After carrying out Wilcoxon rank sum test to the relative strength, 11 m/z peaks with P value less than 0.01 were attained. From the random combination of protein peaks with remarkable variation, used SVM to screen out the combination model with the maximum youden index of the predicted value, screened out 3 markers located at 4656.5, 5643.1 and 7777.4, in tumor group, it is high expression, while in normal control group, it is low expression. Combining three potential markers as input values used the method of leave-one-out to make crossing detection, in the test collection, the specificity of discrimination model was 100.0%, and its sensitivity is 96.3%.2. Nephroblastoma group and control groupSpectrum data from both groups were analyzed, and screened out 2 protein markers m/z located at 6984.5 and 6455.5, in nephroblastoma group, it is low expression, while in normal control group, and it is high expression. The specificity of discrimination model was 100.0% and its sensitivity was 100.0%.3. Nephroblastoma group and group of other abdominal solid tumorsSpectrum data from both groups were analyzed, and screened out 2 proteinmarkers m/z located at 6914.02 and 3256.7, in nephroblastoma group, it is high expression, while in normal control group, it is low expression. The results showed the specificity was 100.0% and sensitivity was 93.3%.ConclusionThe serum protein fingerprint models established by combination of SELDI-TOF-MS with bioinformatics tools are novel effective method for large-scale screening, differential diagnosis and qualitative diagnosis of nephroblastoma in children with high sensitivities and specificities. |
PDF Full Text Request