| Background and aimsColorectal cancer (CRC) is one leading cause of cancer death worldwide, and its incidence is rising in a number of Asian countries. The five year survival rate for CRC diagnosed at early stage is higher than90%, while the five year survival rate for those diagnosed with widespread cancer stage is less than10%. So, it is critical to advance early diagnosis for CRC before metastasis to distant organs occurs. There are evidences that screening of average-risk individuals can increase the detection of CRC in early stage and result in a reduction of mortality.According to suggestions from Asia Pacific consensus recommendations for colorectal cancer screening, fecal occult blood test (FOBT), flexible sigmoidoscopy and colonoscopy are recommended for CRC screening. Although FOBT is the first choice for CRC screening in many resource-limited countries, it lacks sufficient sensitivity. Because of the visualization of precancerous and cancerous lesions, flexible sigmoidoscopy and colonoscopy are thought to be the most sensitive approaches for early detection. However, they are invasive, inconvenient and involve significant costs, which limited their use in CRC screening. It is necessary to find a non-or less-invasive approach, with improved safety, accuracy and patient compliance for CRC screening.Because serologic biomarkers could be analyzed relatively noninvasively and economically, they are far more acceptable than the current screening options and have the potential ability to increase the percentage of the population screened. Serum is expected to be an excellent source of protein biomarkers because it circulates through, or comes in contact with, all tissues. During this contact it is likely to pick up proteins secreted or shed by tissues, so human serum holds immense diagnostic potential. But serum is a complex body fluid containing a large diversity of proteins. More than10,000different proteins are present in human serum and many of them are secreted or shed by cells during different physiological or pathological processes. So it is difficult to identify a specific serum marker.In recent years, the developments in proteomics instrumentation and computational methodologies offer an unique chance to rapidly identify the new candidate markers or pattern of markers for cancer. Many studies show that matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS) is the most powerful relative method for separating complex mixtures of proteins and comparing variations of protein expression in normal and cancerous serum. ClinprotTM technology, which is based on the MALDI-TOF-MS, has many advantages for clinical application:it is sensitive, easy to use, and can perform high-throughput analysis. ClinprotTM technology has opened up new perspectives for establishing specific protein profiles and identifying new tumor biomarkers.The aim of this study was to investigate the use of ClinprotTM technology for the early diagnosis and monitoring of disease progression of CRC, and identify the valuable biomarkers for CRC. In the present study, serum protein profiles between35healthy volunteers,35advanced colorectal adenoma (ACA) patients and40CRC patients were established and the molecule Kininogen-1was identified. Kininogen-1was further studied by enzyme-linked immunosorbent assay (ELISA) and the immunohistochemistry to validate its value in CRC diagnosis and tumor carcinogenesis.Materials and Methods1. The serum proteome profiles of35healthy volunteers,35ACA patients and40CRC patients were compared by CLINPROTTM technology.2.After statistical analysis, the differently expressed peptides were identified with the Ultraflex MALDI-TOF/TOF mass spectrometer.3. ELISAs were conducted to confirm the protein identification and differential expression of kininogen-1in486sera from85healthy volunteers,80patients with ACA,143preoperative patients with CRC,58postoperative patients with CRC,50patients with UC,30patients with GC and40patients with LC. Meanwhile, serum CEA levels among healthy volunteers, ACA patients and CRC patients were determined using commercially available enzyme immunoassay kit.4. The immunohistochemistry test for kininogen-1in colorectal tissues was carried out.75normal colorectal mucosae,77ACAs and248CRCs were included in this test.5. The ClinProTools software2.2(Bruker, Daltonik) was used for analysis of all spectra data derived from serum samples of different groups. Significant different peptides were determined by means of Welch’s T-tests. Class prediction model was set up by QC algorithms. To learn the accuracy of the class prediction, a cross-validation was implemented. Statistical analysis for ELISA and immunohistochemistry test was performed with SPSS13.0software. To test the significance of differences in clinicopathological parameters between groups, the student’s t-test was used for age and tumor size, and the chi-square test was used for the remaining parameters. Correlations between the expression of kininogen-1(immunohistochemical scores) with clinicopathological parameters were evaluated by the Spearman rank-order correlation coefficient. The serum levels of kininogen-1and CEA were normally distributed, so the data were compared using a one-way analysis of variance (ANOVA) test, and multiple comparison were analyzed by LSD methods. Data were expressed as the mean±standard error of the mean (SEM). Receiver operating characteristic (ROC) curve analysis determined the cutoff values of serum kininogen-1and CEA for the diagnosis of colorectal tumor. Kaplan-Meier (Log-rank test) was used for survival analysis. All significance levels were defined as P<0.05.Resultsl.CLINPROTTMest(1). MALDI-TOF analysis of the healthy volunteers and ACA patients resulted in70distinguishable peaks in the1,000to10,000m/z range, out of which43peaks having differential expression and statistical significance (P<0.01). There were61distinguishable peaks between the healthy volunteers and CRC patients, with54peaks having statistical significance (P<0.01).(2). Genetic algorithms in Clinprot were trained with the detected peaks to generate cross-validated classification models between different groups. The best predicting models resulted in a recognition capability of98.96%,100%and100%between CRC and healthy volunteers, between ACA and healthy volunteers, between ACA and CRC, respectively. Every five MALDI-TOF peaks were used in these best classification models. The algorithm calculated the cross-validity estimate, which was95.49%,100%and98.19%, respectively.(3). In all the CRC cases, the overexpression of the peptides at m/z1943and2081was evident (P<0.0001). By MALDI-TOF/TOF tandem mass spectrometer, sequence analyses of them had been done. The results showed that MS/MS Fragmentation of m/z1943was NLGHGHKHERDQGHGHQ, and MS/MS Fragmentation of m/z2081was HNLGHGHKHERDQGHGHQ. It indicated that both of them belonged to the same protein, the kininogen-1.2. ELISA test(1). The serum concentrations of kininogen-1in different groupsThe serum kininogen-1concentrations were215.62±7.63μg/mL and153.22±8.43μg/mL in preoperative CRC patients and healthy volunteers (P<0.001), respectively. Likewise, the kininogen-1concentration in ACA patients was significantly higher than that in healthy volunteers (194.26±10.14μg/mL vs153.22±8.43μg/mL, P=0.003). There was no significant difference between preoperative CRCs and ACAs (P=0.082). The kininogen-1concentration decreased signifcantly in CRC patients after surgery (preoperative vs postoperative:215.62±7.63μg/mL vs188.04±11.70μg/mL, P=0.044). Interestingly, the kininogen-1concentration in GC patients was similar to that in healthy volunteers (171.25±17.17μg/mL vs194.26±10.14μg/mL, P=0.334). On the other hand, the kininogen-1concentrations in UC patients and LC patients were significantly higher than that in healthy volunteers (both P=0.000). There was no significant difference between preoperative CRCs and UCs and LCs.(2). Diagnostic value of serum kininogen-1levels for colorectal tumor patientsThe area under the ROC curve for serum kininogen-1in diagnosing CRC was0.706(95%CI,0.635-0.777); the area in diagnosing colorectal tumors including CRCs and ACAs was0.681(95%CI,0.613-0.749); the area in diagnosing enteric diseases including CRCs, ACAs and UCs was0.701(95%CI,0.636-0.766).According to the ROC curve, the serum kininogen-1concentration of173.96μg/mL was defined to be the optimal cutoff value for differentiating patients with CRC and healthy volunteers. With this cutoff value, the sensitivity, specificity, positive and negative predictive values, and accuracy were calculated to be63.6%,65.9%,75.8%,51.9%, and64.5%, respectively. Similarly, the serum kininogen-1concentration of161.69μg/mL was defined to be the optimal cutoff value for differentiating patients with colorectal tumor and healthy volunteers, with the sensitivity, specificity, positive and negative predictive values, and accuracy being62.3%,63.5%,81.8%,39.1%, and62.7%, respectively. The serum kininogen-1concentration of173.56μg/mL was defined to be the optimal cutoff value for differentiating patients with enteric diseases and healthy volunteers, with the sensitivity, specificity, positive and negative predictive values, and accuracy being61.2%,65.9%,85.2%,34.6%, and62.3%, respectively.(3). The serum concentrations of CEA in CRCs, ACAs and healthy volunteersSerum CEA concentrations were14.67±2.25μg/L,3.10±1.15μg/L and2.43±0.28μg/L in preoperative CRC patients, ACA patients and healthy volunteers, respectively (P<0.001).The area under the ROC curve for serum CEA in diagnosing CRC was0.695(95%CI,0.627-0.767), which was close to that for serum kininogen-1(0.706[95%CI,0.635-0.777]).5μg/L of serum CEA was used for cutoff value in this study, the sensitivity, specificity, positive and negative predictive values, and accuracy were38.5%,85.9%,82.1%,45.3%, and56.1%, respectively. Sensitivity, negative predictive value and accuracy were improved when patients with either serum kininogen-1or serum CEA positive were considered to have a positive test, while the specificity substantially decreased.Furthermore, the sensitivity, specificity, positive and negative predictive values, and accuracy of kininogen-1and CEA were assessed in Duke’s stage A and B CRC patients. The results for kininogen-1were70.1%,65.9%,65.1%,70.9%and67.9%, respectively. On the other hand, the values of aforementioned five parameters for CEA were39.0%,85.9%,71.4%,60.8%and63.6%, respectively. Except the specificity and positive predictive value, other parameters for kininogen-1were better than those for CEA.3. Immunohistochemistry(1). The expression levels of kininogen-1in normal colonic tissues, ACAs and CRCsNo significant difference was found in gender among three groups (P>0.05). The average ages of healthy volunteers, patients bearing ACA and patients bearing CRC were54.03±1.44,59.52±1.34and58.06±0.86years, respectively (P>0.05).Immunoreactivity of kininogen-1was found in the cytoplasm of ACA and CRC tumor cells. The expression level of kininogen-1was significantly higher in CRCs than that in ACAs and normal colorectal mucosae (48.39%vs15.58%vs0.00%, P<0.05).(2). Correlation of kininogen-1expression with clinicopathological features in advanced adenomas and colorectal carcinomasCytoplasmic accumulation of kininogen-1was significantly correlated with tissue histology in ACAs, and interestingly, this correlation was negative (rs=-0.250, P=0.029). However, it was not correlated with tumor location, tumor size and grade of intraepithelial neoplasia (all P>0.05).Cytoplasmic accumulation of kininogen-1was significantly correlated with Duke’s stage and status of lymph node metastasis in CRCs (rs=0.151, P=0.018for duke’s stage; rs=0.128, P=0.045for status of lymph node metastasis). It was not correlated with tumor location, tumor size, tumor cell differentiation and distant metastasis (all P>0.05).(3). Survival analysis for CRC patientsAll110CRC patients were further analyzed to determine the association of the immunoreactivity of kininogen-1with survival. Patients with negative kininogen-1expression had better survival than those who had positive kininogen-1expression, though the difference was not significant (P=0.166). The mean survival time of CRC patients with negative kininogen-1expression was45.21±3.17months, and the mean survival time of CRC patients with positive kininogen-1expression was38.15±3.07months.ConclusionsOur results showed CLINPROTTM is highly accurate and reproducible. kininogen-1appeared to be a potential CRC serum biomarker, which may be valuable for early detection of CRC. |
