Sandwich ELISA serum-based immunoassay for colon cancer detection

955

Colorectal cancer (CRC) is a major cause of cancer death. About 153,760 new diagnosed cases and 52,180 deaths are estimated in 2007 in the US. There are several recommended screening methods for CRC, including colonoscopy and FOBT. Colonoscopy is the current gold standard, yet it is invasive and expensive. It is recognized that CRC is curable by resection when localized, however, only 44% of US adults over the age of 50 undergo screening. Our aim is to develop an easy-to-use and more cost effective serum-based immunoassay for colon cancer detection.
 In recent years, we have discovered biomarkers for the detection of colon cancer (AACR 2006 #396; AACR 2007 #910). We have generated a battery of monoclonal antibodies (mAbs) to CRC serum biomarkers and tested them using our proprietary multiplex protein array technology (MPAT). Using an independent set of serum samples that include 68 CRC and 126 non-CRC patients, we selected several mAbs with high sensitivity for CRC.
 To further confirm mAb reactivity on CRC, we have applied the selected mAbs to in-house and commercial colon cancer tissue microarrays (75 cases). mAbs specifically stain CRC tissues, with membrane, nuclear and cytoplasmic localization.
 We have then tested the selected CRC specific mAbs by antibody capture immunoassay. First, mAbs are reacted with 32 clinical samples comprising 16 CRC and 16 non-CRC cases (colon adenoma and diverticulitis). mAb-biomarker reaction is detected by cheluminescence using a CCD camera, and quantification of mAb binding is performed from the scanned image. The 95th percentile of intensities in the non-CRC group is used as a cut-off value for positive signal. The best mAbs performers with binding intensities above the cut-off value, in at least 50% of CRC serum samples are selected for further studies.
 We have further validated the selected mAbs with 288 clinical samples comprising 32 CRC (12 stage I/II and 20 stage III/IV) and 264 non-CRC serum samples (50 normal, 50 cardiovascular, 64 lung cancer and 100 prostate cancer). Sensitivity and specificity are determined from receiver operating curves (ROC) using the statistical package GB-STAT (Dynamic microsystem). When comparing the CRC group (n=24) to the non cancer group (n=100), sensitivity for most mAbs ranges from 60 to 70% at 80% specificity. Furthermore, most antibodies showed a low sensitivity for lung cancer (n=64) or prostate cancer (n=64) when compared to non cancer samples (n=100) further confirming the CRC specificity of the biomarkers.
 The best CRC serum biomarkers are now being further validated by the antibody capture assay on over 700 serum samples, including early CRC (n=150), late CRC (n=150), benign/inflammatory colon (n=60), normal controls (n=50), as well as other cancers (n=290). We are also developing two-antibody sandwich assays using optimal capture/detecting antibody pairs.
 These immunodiagnostic assays would have tremendous clinical applications in CRC screening and diagnosis.