Background: Point mutations of the ras protooncogene, primarily within codons 12 and 13, are commonly identified in colorectal carcinomas and large adenomas. Despite data suggesting that ras genotyping may have clinical significance with respect to colorectal cancer screening and prognosis, more widespread use has been limited because of the lack of a suitable assay system. The principal objective of this study was to assess the feasibility and validity of a qualitative enzyme-linked immunosorbent assay (ELISA) for detecting the four most common ras mutations in human colorectal tumors at the protein (p21ras) level.
Methods: Tissue homogenates (11-121 micrograms) from endoscopically or surgically resected colorectal adenomas, carcinomas, and normal mucosae were evaluated by a commercially available ELISA (Oncogene Science, Inc. Cambridge, MA) for mutant p21ras containing arginine position 12 (arg12), valine position 12 (val12), aspartate position 12 (asp12), and aspartate position 13 (asp13) amino acid substitutions. Portions of the same tissue from an initial series of 27 specimens also were subjected to mutant-enriched polymerase chain reaction (PCR) and/or PCR amplification with subsequent DNA sequence analysis to validate the ELISA data.
Results: Forty-seven adenomas, 9 carcinomas, and 14 normal mucosae were assayed. Mutations were identified in 16 (34%) of the adenomas (7-asp12, 7-val12, 2-asp13), 3 (33%) of the carcinomas (1-asp12, 1-arg12, 1-asp13), and none of the normal mucosae by ELISA: Polymerase Chain Reaction and DNA sequencing analyses demonstrated identical results for 21 of the 23 (91%) and 14 of 16 (88%) homogenates tested, respectively. The ELISA demonstrated an overall sensitivity of 80-86%, specificity of 90-92%, positive predictive value of 86-100%, and negative predictive value of 86-91%.
Conclusions: The ELISA is a feasible and valid approach for identifying p21ras mutations in human colorectal adenomas and carcinomas.