Assessment of epidermal growth factor receptor (EGFR) signaling in paired colorectal cancer and normal colon tissue samples using computer-aided immunohistochemical analysis

The epidermal growth factor receptor (EGFR) plays a role in multiple tumor cell processes and is targeted by several anticancer therapies. Although EGFR mutations may determine tumor susceptibility in a small proportion of patients, knowledge of the EGFR signaling pathway status in tumors may help guide further drug development and hypothesis-driven combination studies. We aimed to validate and apply a novel computer-aided immunohistochemical (IHC) technique to characterize the status of EGFR signaling in matched colorectal tumor and normal colon tissue samples. Tissue Microarrays (TMA) were made from both cancerous and normal colorectal tissue in 18 patients and stained with antibodies against EGFR, phospho-EGFR (pEGFR), Akt, pAkt, MAPK, and pMAPK. TMA's were quantitatively scored using the Automated Cellular Imaging System (ACIS II, Chromavision, Inc). ACIS was compared against cell line Western blotting, ELISA, and visual scoring (0-3+) by a pathologist. We found that ACIS analysis was highly reproducible and results were well correlated with other techniques. A post-scan "image microdissection" technique of analyzing heterogeneous human samples showed good correlation between paired human samples [Pearson correlation for tumors, 0.922 (p < .001)]. Cancer samples had markedly higher staining of pEGFR, Akt, pAkt, MAPK, and pMAPK. We conclude that ACIS IHC of human tissue samples is quantitative, reproducible, and correlates with Western blots and ELISA in cell line pellets as well as pathologist's scores of human samples. Colorectal tumors show higher staining of pEGFR and downstream effectors compared to matched normal colorectal tissues.