Mesenteric lymph node (ln) metastasis is the single most important prognostic characteristic in colorectal cancer (CRC). The ln status is used for staging and is a decisive selection criterion for postoperative adjuvant therapy. However, it is difficult to accurately determine ln status by routine histopathology (H&E). Thus, ~25% of CRC patients, who by H&E are considered to lack tumor cells in their lns, i.e. stage I+II, die from CRC.
To explore the utility of biomarker mRNA analysis for staging and prognosis of CRC, lns were collected at surgery and mRNA levels for fourteen biomarkers, including carcinoembryonic antigen (CEA), kallikrein 6 (KLK6), cytokeratin 20 (CK20), guanylyl cyclase C (GCC), CEACAM1-S, CEACAM6 and mucin 2 (MUC2), were determined by quantitative RT-PCR with RNA copy standards. Results were compared to routine H&E analysis.
The biomarkers were analyzed for capacity to detect disseminated tumor cells in lns. mRNA levels were determined in CRC- and control lns, primary tumor, normal colon, immune cells and fibroblasts. Lack of expression in immune cells and fibroblasts and high and homogenous expression in primary tumors showed to be the determining factors. CEA fulfilled these criteria best, followed by KLK6, CK20, GCC, and MUC2.
Utility of the biomarker mRNAs for staging and prognosis was examined in 174 CRC patients. CEA was the best predictor of disease-free survival time after surgery with a 71 months difference between CEA(+) and CEA(-) patients and a hazard ratio of 5.1 for risk of recurrence for CEA(+) patients. CEA, CK20 and MUC2 were more sensitive than H&E in that these biomarkers identified patients who succumbed from recurrent CRC although H&E analysis had failed to detect the disseminated tumor cells. Combined analysis of CEA and MUC2 mRNAs improved prediction of outcome. Patients with high risk for recurrence had low MUC2/CEA ratios.
KLK6 mRNA was identified as a potential progression marker by genome-wide microarray analysis of gene expression. It was found to be ectopically expressed in CRC tumor cells. KLK6(+) lns was an indicator of poor prognosis (hazard ratio 3.7). Notably, the actual level was of importance for outcome. The higher the KLK6 mRNA levels the greater the risk of recurrence. At the 90
thpercentile the hazard risk ratio for KLK6(+) patients was 5.6. KLK6 positivity in lns with low numbers of tumor cells, as indicated by low CEA mRNA levels, indicated poor prognosis (hazard ratio 2.8). Thus, KLK6 adds prognostic information to CEA analysis.
Increased levels of mRNA for the proinflammatory cytokine interferon- and the down-regulatory cytokine interleukin-10 in lns of CRC patients suggested ongoing immune reactions against the infiltrating tumor cells. Elevated TGF-1 levels correlated weakly with survival, suggesting protection by the antiproliferative effect of TGF-1 in sporadic cases.
CEA mRNA was the best single biomarker for staging and prediction of disease-free survival time and risk of recurrence after surgery. In addition to CEA, KLK6 positivity and low MUC2/CEA ratio correlate with poor prognosis. Thus, CEA, MUC2 and KLK6 mRNAs form a strong "trio" for staging and prediction of outcome for CRC patients.