Clinical Application Of Proteomics In Children Patients With Wilms' Tumor

Backgroud:Nephroblastoma or Wilms'tumor is the most common pediatric tumor of the kidney. A study led by National Wilms' Tumor Study Group (NWTS) shows that by combining the nephrectomy with postnephrectomy chemotherapy, the 8-year relapse-free survival rate is within the range of 90.5% to 98.9% for stage I Wilms' tumor and is 73.4% to 88.7% for stageⅡandⅢWilms' tumor and decreases to 45.0% to 57.1% for stageⅣWilms'tumor, demonstrating the great significance of early detection for the treatment of Wilms'tumor. Early diagnosis, treatment and regular follow-up are important measures that should be taken to prevent recurrence and improve long-term survival rate. In another study conducted by NWTS in 2001, it is reported that the diagnosis of a noteworthy 7 percent cases is still missed mainly by CT scan, the majority of which were those having a tumor with a diameter less than 3 cm.So an effective, sensitive, specific diagnostic method is eagerly needed.The component and quantity of proteins in the cells will change prior to the occurrence of the pathological changes in each disease. So in theory, it is possible that the index and signs of diseases will be screened in early stage by dynamic observation of the protein. The precondition is to find the specific molecule for certain disease. The screening technology is high-throughput and can not be completed with the traditional techniques.Several genes have been implicated in the development of Wilms'tumour, including WT1 (11p13), WT2 (11p15) and abnormalities in 16p, 1p and 17p. However, these genes are not suitable as biomarkers for early screening and detection. For example, WT1 mutations are only found in 15% of WTs and therefore cannot be used alone either as a marker or a predictor of therapy or can not explain WT development in the other 85% of children.In recent years, the research and application of proteomics along with its related high technology make this large scale screening come true.the rapid development of proteomics has provided new technology platforms to find new tumor-markers and opens a fresh new stage of the molecular diagnosis technology for the clinical application.To date, the search for malignant tumor makers has attracted high attention. The SELDI-TOF-MS (Surface Enhanced Laser Desorption/Ionization Time of Flight-Mass Spectrometry) system developed by Ciphergen is a novel proteomic analysis tool. Based on the method,protein chips can non-specifically combine with various proteins, which will be converted to ions when bombarded by laser in the mass spectrometer. Because of their different flying time in the electric field, ions with different mass/charge ratio may be produced and detected by the detector, then different kurtoses are diplayed, thereby generating characteristic protein fingerprints. This technique has the advantages of small sample size, easy operation, high sensitivity and high throughput and has been successfully applied to the diagnosis, tumor marker screening and other proteomic researches of a variety of malignant tumors such as ovarian cancer, prostate cancer, breast cancer, lung cancer, liver cancer and thyroid carcinoma etc.In our study, SELDI-TOF-MS and support vector machine(SVM) were used to screen specific protein markers in serum taken from children with nephroblastoma. we examined serum protein in children patients with Wilms'tumor at various stages, healthy controls and post-nephrectomy patients and construct a serum protein fingerprint model for the early diagnosis, clinical staging and outcome monitoring of Wilms tumor.After that, the target proteins were separated and purified by multidimensional high performance liquid chromatography (HPLC). The peptide mass fingerprints (PMFs) were obtained after scanning with 2D-LC-LTQ-MS. Then the potential biomarkers were identified by searching PMFs in Bioworks database using the SEQUEST program.Objective:To identify specific serum protein markers in children with Wilms'tumor, to establish a serum protein fingerprint model for the early diagnosis, clinical staging and prognosis monitoring of Wilms'tumor. And identify the pecific target protein markers.Subjects and Methods:1.Subjects:130 serum samples were collected from the Department of Pediatric Surgery at the First Affiliated Hospital of Zhengzhou University during May 2006 to Dec.2007. The 130 samples included 30 samples from nephroblastoma patients before surgery (stage 1:6, stage II:10, stage III:10, stage IV:4),70 samples from . nephroblastoma patients after surgery, and 30 samples from healthy individuals. Thirty samples in the pre-surgical group were taken from 21 males and 9 females who were confirmed to have nephroblastoma by more than two pathologists. The age of patients in the pre-surgical group was 24 days to10 years, average 2.8±0.1 years. Among the 70 postsurgical samples,24 samples were taken two weeks after surgery (radical nephrectomy:21 cases, partial nephrectomy:3 cases),23 samples were taken three months after surgery and 23 samples were taken six months after surgery. In the pre-surgical group and post-surgical group,12 patients from each group (24 patients in total:16 males and 8 females,2.6±0.1years old) were enrolled in the comparison study. None of the 24 patients received radiotherapy or chemotherapy before surgery. Thirty samples from age and gender-matched children who were confirmed healthy in clinic physical examinations were used as control, including 21 males and 9 females (average age:2.6±0.1 years). Samples were taken from peripheral veins in the early morning on an empty stomach. Samples were incubated 1 hour at room temperature and then centrifuged at 3000rpm for 10 minutes. Serums were collected and stored at -80℃before use.2.Reagents and equipments:Trifluoroacetic acid (TFA) was purchased from Fluka (Stockholm, Sweden). Acetonitrile was purchased from Fisher Scientific Inc (New Jersey, USA). Cytochrome c (molecular weight:12361.96), insulin (molecular weight:5734.51), a-cyanoacrylate-4-hydroxy-cinnamic acid (CHCA) and trypsin were purchased from Promega Corporation (Wisconsin, USA). Iodine acetamide (IAM) was purchased from AppliChem Inc (Darmstadt, Germany).Dithiothreitol (DTT) was purchased from BIO-RAD (Germany). Ziptip C18 pipette tips were purchased from Millipore Inc (USA), SPD SpeedVac vacuum centrifuge enrichment system was from Thermo Electron Corporation (USA), liquid chromatography LC-10Avp was from Shimadzu Corporation (Japan), the new enhanced laser desorption/ionization time of flight mass spectroscopy (MALDI-TOF MS) AXIMA-CFRTMplus was from Kratos Analytical Inc (UK), and the LTQ linear ion trap LC-MS mass spectrometer was from Thermo Finnigan Corporation (USA).3.Preparing of serum samples before the PMFs were obtained3.1 Determination of the target proteinAnalysis of serum samples using SELDI protein chip was performed as follows: serum samples from -80℃were thawed on ice, then centrifuged at 10,000 rpm (centrifugal radius 0.5 m) at 4℃for 2 min. Chips were put on a 96-well plate and the plate was placed on ice, and 10μl U9 (9 mol/L urea,2% CHAPS,1% DTT) and 5μl serum were added into each well of the SELDI chip. The plate was oscillated in a chromatography cabinet at 600 rpm for 30 min. When the oscillation was begun for 15 minutes, a protein chip should be pre-processed:put the chip into the bioprocessor, record the chip number, add 200μl NaAC (100mmol/L, pH4) to each well, and oscillate in a chromatography cabinet at 600 rpm for 2 min, then the procedures were repeated once. After pre-processing the chip and the 96-well U9 plate was finished, the 96-well plate was placed on ice and 185μl of NaAC was added to each well. The plate was then placed on the oscillator in a chromatography cabinet at 600rpm for 2 min at 4℃.A 100μl serum sample that had been balanced with CHAPS and serum albumin was removed and added to the chip, and then oscillated in the chromatography cabinet at 600rpm for 1h at 4℃. After discarding the residue and drying the chip,200μl NaAC was added to the chip and the chip was oscillated at 600rpm for 5 min. Repeat the procedures of discarding residue, drying chips, addition of NaAC and oscillation three times. Each well was rinsed with 200μl deionized water twice, and then the chip was quickly dried. Finally,1μl 50% saturated Sinapinic acid (SPA) was added to each well and chips were dried before analysis.A protein chip with a known molecular weight was used to rectify SELDI-TOF-MS and make sure the error of the molecular weight was less than 0.1%. The WCX2 protein chip with bound proteins was analyzed by a mass spectrometer. High quality serum was used as control to check the repeatability. All data were rectified by Proteinchip Software 3.1, ensuring that the total ionic intensity and molecular weight were homogenized. Data were analyzed using the ZUCI-Protein Chip Data Analyze System software package. MS raw data were noise filtered and examined by cluster analysis. Results of m/z peaks obtained after preliminarily screening were analyzed using Wilconxon rank sum test (a=0.01). The linear support vector machine (SVM) classifier was also used in the study:RBF kernel function was adopted, with the gamma value at 0.6 and penalty function (c) at 19. Support vectors were selected by statistical filtering in combination with model-dependent screening. An evaluation model was established, evaluated and validated by leave-one-out cross-validation.3.2 Purification of the target proteins After potential biomarker proteins were found by SELDI-TOF-MS, the proteins were purified. Serum samples stored at -80℃were thawed in ice water. After thawing,100μl serum was mixed with 350μl ultra-pure water and 700μl pure CAN, and then incubated at -20℃for 30min, centrifuged at 13,000rpm for 10min. The supernatant was transferred into a new tube, and the tube was placed freeze-dried in an SPD SpeedVac for about 20min. Then the freeze-dried samples were analyzed by high-performance liquid chromatography (HPLC:1mm×100mm column), and the purified solution of different time courses was collected. The purified protein solution was freeze-dried in the SPD SpeedVac until the volume was about 20μl.1.5μl of the prepared sample was mixed with 1.5μl a-cyanoacrylate -4-hydroxy-cinnamic acid (CHCA), and then spotted onto the MALDI target plate. Before analysis by MALDI-TOF-MS, controls (cytochrome c+CHCA and Insulin+CHCA) were used to rectify the machine accuracy. Then the target plate was put onto MALDI-TOF-MS to identify purified samples whose m/z peak values were screened by SELDI-TOF-MS.3.3 Enzyme hydrolysisProteins were hydrolyzed before identification, which was performed as follows: ultra-pure water was added to the purified protein sample until the volume was 40μl. Then 4μl O.lmol DTT solution (5μl 11 mol DTT solution+45μl ultra-pure water) was added and incubated for at 37℃1 hour. Then 1.6μl Iodine acetamide (IAM) was added to the solution and the mixture was incubated in the dark for 1h. After that, 1.6μl 1mol DTT solution,150μl 0.1mol NH4HCO3 and 2μl parenzyme was subsequently add to the solution, then the mixture was incubated at 37℃overnight (6-8 h).3.4.Identification of proteinsAfter enzyme hydrolysis, purified protein samples were analyzed by 2D-LC-LTQ-MS using nitrogen laser with the following settings:laser wavelength-337nm, pulse duration-3ns, mass spectrum accumulation of signal from 50 single scan, accelerating voltage-30kv, absorbing voltage-9.3kv, detecting voltage-4.75kv, vacuum degree-1×10-6Pa, and detecting mode-positive ion. Peptide mass fingerprints (PMFs) were used as searching objects in Bioworks data based by SEQUEST. Potential biomarker proteins were identified, and their biological functions and characterization were analyzed using bioinformatics methods.3.5Statistical anaylsisData were presented as x±s. All statistic analysis was performed using SPSS 13.0 software. The t test was applied in the MS data analyses between the pre-surgery group and the control group, the pre-surgery and the post-surgery group, the post-surgery group and the control group. One-way Anova analysis was used to analyze mass spectrometry data of different stage. Rank sum test was used to analyze nonhomogeneous data. Difference was regarded as significantly when P<0.05.Results:1. Screening of serum protein biomarkers in child with Wilms'tumor by SELDI-TOF-MS1.1 Establishing experimental methods concerned with a serum protein fingerprint in Wilms'tumor and screening proteinchip types:In this study, we processed the serum samples by using the 0.5% CHAPS, and applied to Cibacron Blue 3GA specifically in order to remove serum albumin. Through the study of four kinds of different proteinchip:WCX2 proteinchip, SAX2 proteinchip, IMAC3 proteinchip and H4 proteinchip, we found that different chip can combine different number of serum proteins, and H4 chip in this study proved to be able to get more meaningful peak, which not only can be better applied to the serum protein fingerprinting studies of Wilms'tumor, but also has very good repeatability. 1.2 This study is the first reported to establish a serum protein fingerprint based on models of Wilms'tumor by use of SELDI-TOF-MS technology combined with SVM60 cases of serum samples protein fingerprinting were randomly divided into training group and test group, and the datas of which were collected by a mass spectrometer.efore each collection of experimental data we use known proteinchip with standard molecular weight to calibrate and make sure the error is less than 0.1%.Testing the quality control serum repetitively to be sure its peak size and intensity of the coefficient of variation (CVs) are controlled within the error range (0.05% and 15% or less), and applying ProteinChip Software 3.1 (Ciphergen Inc)software the same time as MS raw data were noise filtered and examined twice by cluster analysis ranging from 2000 to 30000. then, finding out M/Z peak of samples, defining 10% M/Z peak as minimum threshold to cluster. Using Wilcoxon rank sum test to analyze, and assess the relative importance of the various peaks to distinguish two types of samples according to P value. Combining the M/Z peak with significant difference and randomly inputting SVM, then screen markers to establish discriminant model. With the method of leave-one-out and cross-validation assessment model, we carried out blind test, increased the peak of M/Z successively according to the order. This study found that youden index of test set got best when the combination of M/Z is 5, and the results proved accuracy could reach up to 98.6%.This 5 M/Z peak is respectively 6455,6984,3221,5639,9190. With 30 cases of samples as the training group,30 cases of samples as a blind test group, the specificity of established model was 98.3%(95%confidence interval 85%～100%), a sensitivity of 98.9%(95% confidence interval 89%～100%), Youden index value of 0.87551, which all indicated that the model has a good diagnostic value in the aspect of differentiating Wilms'tumor.1.3 With M/Z peaks respectively 6455,6984,3221,5639,9190, the mass spectrum diagram showed that they were in low expression within Wilms' tumors,whereas in healthy children were highly expressed. At the same time, M/Z peak in 6455 and in 9190, their expression had a sequential decline manner followed healthy children and children with Wilma's tumor,so we supposed that M/Z 6455, M /Z 9190 protein may be playing an important role in the occurrence and development of Wilma's tumor.2. Application of serum protein markers in the clinical staging and prognosis monitoring of children with Wilms'tumor2.1483 M/Z peaks were obtained from mass spectral of Wilms'tumor stageⅠgroup andⅡgroup after a preliminary screening. Three M/Z peaks of 5022.4,7965.4,8469.6 were identified, whose expression levels were high in patients with Wilms'tumor stage I group but were low in Wilms'tumor stageⅡgroup.They have a sensitivity of 80.00% and a specificity of 100% for the discrimination model in the test set. And 496 M/Z peaks were obtained from mass spectral of Wilms'tumor stageⅠgroup andⅢgroup. Four M/Z peaks of 4330.7,4303.7, 4263.2 and 4122.8 were identified, whose expression levels were high in patients with Wilms'tumor stageⅠgroup but were low in Wilms' tumor stageⅢgroup. The sensitivity and specificity are both 100% for the discrimination model in the test set. Meanwhile 565 M/Z peaks were obtained from mass spectral of Wilms'tumor stageⅠgroup andⅣgroup. Two M/Z peaks of 8179.1 and 10836.6 were identified, whose expression levels were high in patients with Wilms'tumor stage I group but were low in Wilms'tumor stageⅣgroup. Discriminant model in the test set was 88.89% sensitivity and specificity of 100.00%.And 490 M/Z peaks were obtained from mass spectral of Wilms'tumor stageⅡgroup andⅢgroup. Two M/Z peaks of 4143.2 and 5019.2 were identified, whose expression levels were high in patients with Wilms'tumor stageⅡgroup but were low in Wilms'tumor stageⅢgroup. They have a sensitivity of 88.89% and a specificity of 100% for the discrimination model in the test set. And 508 M/Z peaks were obtained from mass spectral of Wilms'tumor stageⅡgroup and IV group. The selected M/Z peak 7976.5 has a high expression in Wilms tumor stageⅡgroup but low expression in stageⅣgroup. The sensitivity and specificity are both 100% for the discrimination model in the test set.And 504 M/Z peaks were obtained from mass spectral of Wilms'tumor stageⅢgroup andⅣgroup. The selected M/Z peak 8194.4 has a high expression in Wilms tumor stageⅢgroup but low expression in stageⅣgroup. Discriminant model in the test set was 93.75% sensitivity and specificity of 100.00%.And 519 M/Z peaks were obtained from mass spectral of Wilms'tumor stageⅠ+Ⅱgroup andⅢ+Ⅳgroup. Two M/Z peaks of 3257.6 and 4153.9 were identified, whose expression levels were high in patients with Wilms'tumor stageⅠ+Ⅱgroup but were low in Wilms' tumor stageⅢ+Ⅳgroup. They have a sensitivity of 83.33% and a specificity of 93.75% for the discrimination model in the test set.2.2 With the clinical staging in children increasing, M/Z peaks at 6455,9190 expressed in serum decreased in order of the staging.2.3 Wilms tumor in children with SELDI stages:I period of 6 cases,Ⅱperiod of 10 cases,Ⅲperiod of 10 cases,Ⅳperiod of 4 cases,30 cases of Wilms tumor staging and the pathological findings were consistent with the coincidence rate was 100.00%. Confirmed by pathology and surgery, CT diagnostic accordance rates followed by periods were 100.00%,85.00%,85.00%,75.00%.2.4 We screened protein markers to the pre-operative and post-operative in child with Wilms tumor.We found the difference was the most significant when m/z was at 6455.5,9190.8 department, which expressed lowly in the serum of pre-operative group, in the post-operative group and showed high expression in healthy children group.The difference was statistically significant (P<0.01); Between the two groups of post-operative and normal children,the result showed no significant difference (P> 0.01). The above all indicated that protein expression increased as inhibitory effect was weakened after tumor resection.2.5 Compared serum protein spectrum with post-operative children with Wilms tumor and healthy children group, and compared with the pre-operative and post-operative specimens in 21 cases of radical surgery children at the different time of 2 weeks,3 months and 6 months after operation,we found M/Z peaks at 6455.5, the protein expression in radical surgery group were respectively 2087.21±658.33,2189.67±856.42 and 2232.16±598.32 (compared with control group:P> 0.05; compared with preoperative group:P<0.0.1), in palliative surgery group were respectively 1044.86±533.21,677.86±435.26 and 431.65±158.31 (compared with control group:P<0.01; compared with preoperative group:P> 0.05);While M/Z-peaks at 9190.8,at the time of 2 weeks,3 months and 6 months after operation,the protein expression were respectively 1101.20±219.69,1078.21±322.43 and 1066.06±110.45 (compared with control group:P> 0.05; compared with preoperative group: P<0.01) in radical surgery group,in palliative group were respectively 328.11±210.26,289.10±108.09 and 276.49±92.15 (compared with control group:P<0.01; compared with preoperative group:P> 005).There were two protein markers in serum, which was in low-expression consecutively in palliative at 2 weeks,3 months, 6 months after operation; While in radical surgery group, we found that serum protein was no significant change at the two peaks by tracking review 3 months,6 months after operation.The lower stages, the better prognosis is, and M/Z intensity is higher.Considering the tumor still exist in the three cases of palliative resection and the inhibited protein continued to be inhibited, protein markers was in sustained low expression. Except for 1 patient died in 2 months after operation, expression of 2 cases continued to decline at 3 months and 6 months after operation, the result may be that the protein markers have been suppressed deeper.3. Identification analysis of serum protein biomarkers in child with Wilms' tumorIn this experiment, we established the serum samples for M/Z peaks at 6455.5,9190.8 as the target protein, By identifying their amino acid sequence, searching protein database, we found apolipoprotein C-Ⅰand haptoglobin was the two protein which corresponding to their protein molecular weight, so We speculated that apolipoprotein C-Ⅰand haptoglobin may be the candidate proteins to determine the extent of malignant treatment, prognosis, and development of Wilms tumor.Conclusion:The combined use of SELDI-TOF-MS and SVM to establish a serum protein fingerprint model for Wilms'tumor offers a novel, highly specific and sensitive method for the early diagnosis, clinical staging and prognosis monitoring of this condition. Apolipoprotein C-Ⅰand haptoglobin may be used as potential biomarkers to predict the degree of malignancy, therapeutic outcomes, and prognosis of nephroblastoma in children.