Comparison of fluorescein isothiocyanate- and Texas red-conjugated nucleotides for direct labeling in comparative genomic hybridization.

In this study, we investigated whether fluorescein isothiocyanate (FITC)-labeling of test DNA and Texas-red (TR) labeling of reference DNA in comparative genomic hybridization (CGH) experiments cause the results to differ from those obtained using the opposite combination (reverse labeling). Analysis was performed on a total of 20 DNA specimens consisting of 13 frozen bone marrow aspirates from patients with acute myeloid leukemia, and fresh peripheral blood samples from seven healthy donors. For CGH, one aliquot from each test DNA sample was labeled using nick-translation with FITC-dUTP and another with TR-dUTP. Afterwards, the FITC-dUTP and TR-dUTP-labeled test DNAs were hybridized to TR-dUTP- and FITC-dUTP-labeled normal reference DNAs, respectively. The results using the two combinations were compared with each other and with the results of G-banding karyotype analysis. Karyotype data was used to detect artifacts known to occur in some chromosome regions in CGH analysis. The control DNAs labeled with FITC or TR showed no DNA copy number changes. Regardless of the fluorochrome employed for labeling, no DNA copy number changes were detected using CGH in patients with normal karyotypes, nor in patients whose karyotype aberrations were present in less than 40% of cells. In the remaining patients, CGH revealed DNA copy number changes that coincided with the results of the G-banding analysis. Hybridization artifacts known to occur in CGH experiments affecting chromosome regions 1p33-pter, 16p, 17p, 19, and 22 were observed in 15-23% of the tumor samples labeled with FITC, but not in samples labeled with TR. In addition, other previously unreported overrepresentations affecting 7q21, 9q34, 16q, 17q, and chromosome 20 were observed at very low frequencies in up to 10% of the samples when FITC was used to label test DNA. However, when TR was used, overrepresentations were observed at 4q13-q21, 11q21-q23, 13q21-qter, and Xq21-q22, whereas 19p was underrepresented. The results demonstrate that TR-labeling confirms abnormalities detected using FITC-labeling and reduces hybridization artifacts in the known problematic regions of the human genome.