Analysis Of Proteomic Spectra On Cervical Cancer

1 Background and aimsCervical carcinoma is a common malignant tumors ranking second only to breast cancer as the most common malignancy among women worldwide. The estimate annual number of new cases of cervical cancer worldwide is more than 49 000, however, there are more than 80 percent of new cases in developing countries. The prognosis of cervical carcinoma is closely concerned with the clinical features such as clinical stage. The five-year survival rate in the early stage of cervical carcinoma is above 90 percent, while there are only about 10 percent for the middle and late stage of cervical cancer. The recurrent and metastasis of cervical carcinoma is not only relevant to the clinical stage but to the treatment scheme and so on. So it is important that how to improve the early diagnostic rate, how to make the treatment scheme and assessment the treatment effect, how to know the prognosis and prevent it from recurrence, which is the aims for our research. The cervical carcinogenesis is a multistage and progressive process. It begins from precancerous lesions, a group of lesions including the Cervical Intraepithelial Neoplasia (CIN)Ⅰ—mild dysplasia, CINⅡ—moderate dysplasia, CINⅢ—sever dysplasia and carcinoma in situ, to invasive cervical carcinoma. It is a progressive nature but some of cases can be regression. It gaves us a theory basis to early diagnosis and threatment for the cervical cancer because it would take about 5 to 10 years time for every step.Enormous efforts have been undertaken concerning the etiology of cervical carcinoma, yet it is still not very clearly solved. The most important thing is that infection with specific subtypes of Human Papilloma Virus (HPV) has been strongly implicated in cervical carcinogenesis. However, HPV infection alone is insufficient for malignant transformation. It is a basic factor for cervical carcinogenesis but has to act with others such as an abnormal sexual action. Biologically, the cervical carcinogenesis is a complex course with multigene changes. However, genetic changes may not reflect the stage and progression of disease directly and objectively because proteins carry out most of the cellular functions based on mRNA. Therefore, the direct measurement of protein levels and activities within the cell is the most determinant of overall cellular function. Recently, a novel proteomics technique, surface-enhanced laser desorption/ionization-time of flight-mass spectrometry (SELDI-TOF-MS), has been developed quickly. It can not only map protein profiles with a high throughput but also provide a protein pattern to discriminate the patients from healthy coupling with bio-information and needs low amounts of samples, such as serum, tissue and etc. It has been widely used in tumor research, such as cancers in digestive system, respiratory system, urinary and genital system, and etc, but there was a few report for cervical cancer, there was only one report by using tissue sample, nothing for using serum sample.To structure serum protein patterns for identifying cervical carcinoma and cervical precancerous and further to explore the mechanisms of cervical carcinogenesis, the SELDI-TOF-MS technique was employed to detect the changes of serum proteomics on cervical cancer and controls. The proteins that can discriminate the patients from normal were picked up and then compared them with the samples of cervical precancerous lesion, post-operation with radical hysterectomy and pelvic lymphadenectomy, and review patients by t test. It means that the whole course of cervical carcinogenesis and recover, a typical survival model—nontumortumor-nontumor of cervical cancer, was observed so that the biomarkers (proteins) those can be used for screening, early diagnosis, assessment treatment effect and prognosis would be found out.2 Material and Methods2.1 Study population The samples with pathological diagnosis and without radiotherapy or chemotherapy were obtained from the First Affiliated Hospital, Zhengzhou University.2.1.1 Serum sampleSerum samples enrolled in this study were ofâ‘ 49 patients with cervical cancer and 71 age-matched healthy women;â‘¡67 cases of cervical intraepithelial neoplasia(CIN) including 35 patients with CINâ… andâ…¡, and 32 patients with CINâ…¢;â‘¢24 patients with cervical cancer those serum samples were collected on the 10th day after radical hysterectomy and pelvic lymphadenectomy; 22 review patients at the time of 3-month after operation; 18 follow-up patients at the time of 1-year after operation.2.1.2 Tissue sample33 cervical cancerous tissues and 29 normal cervical epithelial tissues were obtained in this study.2.2 Sample collection2.2.1 Serum sample3 ml fast blood was collected and stood for 30 min, and then centrifuged at 2000rpm for 10 min. All serum samples were aliquoted into 100μl and stored at-80℃until use.2.2.2 Tissue sampleA piece of the cervical cancer specimen near the cancer floor was collected immediately after radical hysterectomy and pelvic lymphadenectomy. Normal cervical tissue was obtained at same time at the place not less 1cm from the tumor margin. They were cut and digested by 0.25% trypsinase. Then they were centrifuged at 900rpm after washing, repeated this step for 3 times. A cell smear slide was made for pathological diagnosis. The target cells, cervical carcinoma cells or normal cervical epithelial cells, were estimated about≥85%, the cell counting was to 1×10⁷/ml. The cell samples were stored at-80℃until use. 2.3 Methods2.3.1 Preparation of protein chipAn Eight-spot immobilized metal affinity capture array-Cu (IMAC-Cu) chip was put onto a bioprocessor. The spots were activated with 50μL of 100 mmol/L CuSO₄ and vortexed for 5 min, followed by a deionized water rinse for 5 times, then 50μl of 100 mmol/L sodium acetate buffer (pH 4.0) was added to each array and shaken for 5 min, followed by a deionized water rinse again. The activated array surfaces were equilibrated with 150μL of binding buffer (pH 7.0), agitated for 5 min, twice.2.3.2 Preparation of serum samplesSerum samples were thawed and diluted 10μl to 20μl (1:2) with U9 buffer, vortexed at 4℃for 30 min. Then it was diluted 1:12 in binding buffer, vortexed at 4℃for 30 min.2.3.3 Preparation of tissue sampleTissue specimen was thawed and add 50ul cellular splitting fluid, vortexed at 4℃for 30 min, centrifuged at 15000rpm at 4℃for 10min. The supernatants were mixed with U9 buffer (1:2), vortexed at 4℃for 30 min. Then it was diluted 1:12 in binding buffer, vortexed at 4℃for 30 min.2.3.4 Reaction of binding proteins50μl of diluted sample was applied onto the array surface and shaken at 4℃for 60 min. Then the chips were washed twice with 150μl of binding buffer for 5 min each wash cycle. The chips were removed from the bioprocessor, air-dried. Before SELDI-TOF MS analysis, 0.5μl of a saturated EAM solution was applied onto each spot twice and air-dried between each EAM application.2.3.5 Data collecting and analysis of bioinformationChips were placed into the Protein Biological Systemâ…¡mass spectrometer reader (PBSâ…¡, Ciphergen Biosystems, Inc). The spectra were calibrated by using the All-in-1 NP 20 protein molecular mass standard. In ProteinChip Software the protein reader program was set. The mass accuracy was calibrated to less than 0.1%. The coefficient of variance for peak height was less than 15%. Data were collected twice by averaging 90 laser shots, a detector sensitivity 6 and an optimized range of 2 000—20000 Da with a highest mass of 50 000 Da. Biomarker Wizard Software, similar as a statistic software, is used for analyzing the differences of the intensity for each labeled peak between the two samples by using t test. Biomarker Pattern Software, a data digging software, is for constructing a decision tree classification model with ten-fold cross validation. The model was set up to split the data set into two nodes based on the intensities of peaks. At each node a peak intensity threshold was set. If the peak intensity of a sample were lower than or equal to the threshold, this sample would be divided into the left-side node. Otherwise, the sample would go to the right-side node. The process would go on until a sample entered a terminal node. The model would be used that yielded the least classification error. Specificity and sensitivity were respectively calculated as the proportion of the number of disease samples correctly identified to the total number of control samples.2.3.6 t test for the significant expressive proteinsThe significant expressive proteins from samples of cervical carcinoma and controls were compared with samples of cervical precancerous lesion, post operation with radical hysterectomy and pelvic lymphadenectomy, and groups of review patients by t test with SPSS Software 11.0.3 Results3.1 Serum proteomic profiling analysis on cervical carcinoma47 proteins were detected with a significant level of P＜0.01 from cervical cancer patients and normal controls. 6 proteins with m/z value of M8929.31, M7930.52, M9127.31, M8141.01, M7963.06 and M9280.63 had high score(＞95%) in building a model of decision tree classification algorithm for cervical cancer detection. The relative splitting score was 100, 98.25, 98.25, 98.12, 97.35, and 97 respectively. The sensitivity and specificity of m/z value of M8929.31 were 97.96% (48/49) and 98.59% (70/71) respectively.3.2 Analysis of the 6 proteins on cervical precancerous lesionsMany proteins had changed in cervical cancer initiation and development. The 6 significant expressive proteins from cervical cancer with mass to charge value of M8929.31, M7930.52, M9127.31, M8141.01, M7963.06 and M9280.63 appeared to be down regulated in patients with cervical precancerous lesions. The intensities were orderly reducing from controls, CINâ… -â…¡,CINâ…¢to cervical cancer in a level of P＜0.01.3.3 Analysis of the 6 proteins on post-operation and review patients3.3.1 Analysis of the 6 proteins on post-operation (10th day)The levels of 6 proteins with mass to charge value of M8929.31, M7930.52, M9127.31, M8141.01, M7963.06 and M9280.63 were compared, the 10th day after operation to cervical cancer, and had gradually retrieved in a level of P＜0.01 after surgery, except the level of m/z value of 9280.63 which was no significant difference (0.6307±0.5789/0.4084±0.3098, P=0.083). However, they were still much lower than normal (P＜0.01).3.3.2 Analysis of the 6 proteins on the review patients at the time of 3 month post-operationComparing the intensities of the 6 proteins with mass to charge value of M8929.31, M7930.52, M9127.31, M8141.01, M7963.06 and M9280.63, 3 month post-operation with the 10th day, they had gradually retrieved in a level of P＜0.01 after surgery, including M9280.63.3.3.3 Analysis of the 6 proteins on review patients with cervical cancer at the time of 1 year post-operationThe 6 proteins with mass to charge value of M8929.31, M7930.52, M9127.31, M8141.01, M7963.06 and M9280.63 had continually climbed in a year after operation. The intensity of them was much higher than the time of 3 month post-operation (P＜0.01). It was almost same as normal (P＞0.05) but M8929.31(11.9831±8.1697/19.8839±13.3494,P=0.003)3.4 Tissue proteomic analysis on cervical cancer72 proteins were marked out with a significant level of P<0.01 from cervical cancer tissue and normal cervical epithelial tissue. 8 proteins with m/z value of M5929.87, M3630.52, M10335.39, M6793.00, M7365.78, M4498.06, M8856.45, M9586.27 had high score in building a model of decision tree classification algorithm for cervical cancer detection. M5929.87, M3630.52, M10335.39, M6793.00, M7365.78 and M4498.06 appeared to be down regulated but M8856.45 and M9586.27 were found at high intensity in patients with cervical carcinoma. The sensitivity and specificity of m/z value of M5929.87, M3630.52 and M10335.39 were 93.94% (31/33) and 96.55% (28/29) respectively.4 Conclusions4.1 Serum proteomic analysis on cervical cancerA group of proteins with m/z value of M8929.31, M7930.52, M9127.31, M8141.01, M7963.06 and M9280.63 are closely concerned with invasive cervical cancer. It suggests that they may be a group of proteins or peptides with physiological function and a group of protector in normal cervical epithelial tissues because they appeared to be down regulated in cervical carcinoma. Changes of the proteins would stand for changes in the level of cellular in cervical cancer. They will probably be a group of biomarkers for cervical cancer.4.2 Serum proteomic analysis on cervical precancerous lesionThe intensity of the 6 significant expressive proteins from cervical cancer with mass to charge value of M8929.31, M7930.52, M9127.31, M8141.01, M7963.06 and M9280.63 was orderly reducing from controls, CINâ… -â…¡,CINâ…¢to cervical cancer in a level of P＜0.01. It confirms the hypothesis that they may be a group of proteins or peptides with physiological function or a group of protector in healthy. They would be a group of potential screening and diagnostic biomarkers for precancerous lesion and cervical carcinoma.4.3 Serum proteomic analysis on post-operations and reviewsThe level of the 6 proteins with mass to charge value of M8929.31, M7930.52, M9127.31, M8141.01, M7963.06 and M9280.63 was orderly retrieved in a level of P＜0.01 from cervical cancer, the tenth day, 3 month to 1 year after radical hysterectomy and pelvic lymphadenectomy. It strongly suggests that they would be a group of biomarkers for assessment treatment effect and prognosis of cervical cancer after researching more cases.4.4 Tissue proteomic analysis on cervical cancerThere were 8 proteins with m/z value of M5929.87, M3630.52, M10335.39, M6793.00, M7365.78, M4498.06, M8856.45 and M9586.27 with a model of decision tree classification algorithm for cervical cancer detection. It suggests that the immunohistochemical biomarkers for cervical cancer would be found out from them.4.5 Comprehensive analysis for the dataThe intensity of the 6 significant expressive proteins from cervical cancer with m/z value of M8929.31, M7930.52, M9127.31, M8141.01, M7963.06 and M9280.63 was orderly reducing from controls, CINâ… -â…¡,CINâ…¢to cervical cancer in a level of P＜0.01 and retrieved in a level of P＜0.01 from cervical cancer, the tenth day, 3 month to 1 year after radical hysterectomy and pelvic lymphadenectomy in order. It strongly suggests that we should research them more for clinical use if there is no economic problem they should be looked for from the protein database house with protein analysis tools and then be identified. We really hope that the simple test boxes will be born for clinical use after they are confirmed by clinical research with enough samples.