High Percentage of False-Positive Results of Cytokeratin 19 RT-PCR in Blood: A Model for the Analysis of Illegitimate Gene Expression

Abstract

Cytokeratin 19 (CK19) RT-PCR is widely used in order to detect circulating tumor cells in peripheral blood and bone marrow. However, increasing amounts of information support the fact that it is also associated with a high percentage of false-positive results. In our study, we not only managed to demonstrate the significant limitations of this method, but were also able to clarify the reasons behind these limitations. We developed a completely novel RT-PCR for CK19 and sequenced an intron at nucleotide (nt) 980 of the CK19 mRNA to exclude DNA contamination. Tumor dilution experiments were performed in order to analyze the specificity and sensitivity of the method. Control experiments using the blood of healthy donors were performed. Tumor cell dilution experiments gave a detection limit of one tumor cell. If tumor cells were mixed with an equal volume of pure mononuclear cells, the detection limit was 1 tumor cell in 10⁵ mononuclear cells. RT-PCR of mononuclear cells from healthy blood donors gave false-positive results in 29% of the cases. We conclude that a significant decrease in the sensitivity of CK19 RT-PCR occurs if it is performed in blood cells and that the illegitimate CK19 gene expression in normal cells can lead to false-positive results. These limitations have to be taken into account if RT-PCR is to be used for the detection of tumor cells either in blood or in bone marrow in clinical practice.

Differential expression of carcinoembryonic antigen (CEA) splice variants in whole blood of colon cancer patients and healthy volunteers: implication for the detection of circulating colon cancer cells

“…Only two patients (6%) were responsible for this small and insignificant elevation of the mean value. Thereby, our detection rate was lower than in previous studies (Gerhard et al, 1994) but in accordance with studies which did not find significant differences between cancer patients and a control group (Ko et al, 1998). However, because we used a different method, i.e.…”

“…Previous studies indicate that both markers may serve as specific tumor cell markers in the analysis of whole blood samples (Gerhard et al, 1994;Burchill et al, 1995;Liefers et al, 1998;Mori et al, 1996;Soeth et al, 1997). However, according to our experience and reports from others (Mori et al, 1996;Ko et al, 1998;Goeminne et al, 2000) low CEA and CK20 levels are also present in normal blood samples. Thereby, standardization becomes difficult between various experiments and certainly between different research groups because a positive signal depends highly on minor assay condition in relation to cycle numbers.…”

“…It remains unclear why over a long postoperative period blood samples remained positive despite curative surgery. Other groups detected CEA in a significant number of normal blood samples (Ko et al, 1998) or in patients with inflammatory bowel disease (Castells et al, 1998) positive for CEA. It was also found that treatment with GM -CSF induced CEA expression in white blood cells (Jung et al, 1998).…”

“…Therefore, monitoring of the level of CEA expressing colon cancer cells is an attractive approach to determine therapeutic effectiveness in such trials. However, amplification has been reported in normal samples by using high cycle numbers indicating broad expression of these markers, in particular of CEA, at low levels (Mori et al, 1996;Ko et al, 1998;Goeminne et al, 2000). Quantification of the PCR signal would allow calculation of cell numbers.…”

Application of differential display to identify genes for lung cancer detection in peripheral blood

“…However, RT-PCR analysis of CK19 mRNA expression in blood can give rise to a significant number of false-positive results on the basis of nontissue-specific, constitutive low-level (illegitimate) transcription. 31,33 The success of using CK19 PCR can be improved by technical modifications to the assays. 11,14,22 Furthermore, CEA and CK19 are evaluated not as specific for lung cancer markers but for epithelial cells in general.…”