Critical evaluation of real-time reverse transcriptase-polymerase chain reaction for the quantitative detection of cytokeratin 20 mRNA in colorectal cancer patients

Abstract

We evaluated the usefulness of cytokeratin 20 (CK20) mRNA expression in the quantitative detection of circulating tumor cells in the blood of patients with colorectal cancer (CRC). Blood samples from healthy volunteers (HVs; n = 37), patients with localized (n = 42) and metastatic colorectal cancer (n = 40), and patients with chronic inflammatory bowel disease (CID; n = 15) were examined. After immunomagnetic enrichment using microbeads against human epithelial antigen, total RNA was extracted, reverse transcribed, and analyzed by real-time reverse transcriptase-polymerase chain reaction using the LightCycler instrument. CK20 expression in peripheral blood was found in 46 of 82 (56%) patients with CRC, 8 of 37 (22%) HVs, and 9 of 15 (60%) patients with CID. Levels of CK20 mRNA were significantly higher in blood samples from CRC patients (median 681) than in blood samples from HVs (median 0) (P = 0.001), whereas no difference could be detected between patients with CRC and CID. Although the present technique could not distinguish CRC from CID, the method warrants further efforts to improve sample preparation and tumor cell enrichment, which may render real-time CK20 reverse transcriptase-polymerase chain reaction a feasible technique in identifying circulating tumor cells in peripheral blood of cancer patients.

Figure 1

Box plots show that expression of PBGD was not significantly different among HVs, patients with CRC (Pat), and patients with CID (P = 0.100; Kruskal-Wallis test).

Figure 2

ROC curve for CK20 mRNA ratios to distinguish CRC patients (n = 82) from HVs (n = 37). ROC curves were constructed by plotting sensitivity and 100-specificity corresponding to each cutoff value for CK20 mRNA. The cutoff of 415 gave a sensitivity of 55% and a specificity of 78%. Area under the curve = 0.672; SE = 0.051; 95% confidence range, 0.580–0.756; P < 0.003.

Figure 3

Comparison of CK20 mRNA ratio between patients with CRC, LCC, and MCC; HVs; and patients with CID. Black bars, median; Cutoff value 415 is indicated by dashed horizontal line.

Figure 4

Representative results of a CK20 qRT-PCR using the LightCycler. Solid lines indicate amplification curves from the target gene (CK20) and dashed lines from the reference gene (PBGD). No product was amplified in the no-template sample or when reverse transcription was omitted (•).