Automatic quantification of gene amplification in clinical samples by IQ-FISH.

BACKGROUND: The reliable detection and quantification of gene amplifications is crucial to clinical practice. Although there are different detection techniques, the fluorescence in situ hybridization (FISH) method has become highly accepted over past years because it is a reliable, robust, and quick method. Unfortunately, automatic quantification of gene amplification based on fluorescence intensities has not been possible thus far. Because current spot counting methods are reliable only when analyzing low amplification rates, we attempted to establish another method, i.e., to quantify the intensity of different FISH signals using an automatic fluorescence microscopical device on interphase nuclei: interphase quantitative FISH (IQ-FISH).
METHODS: We quantified the fluorescence intensities of the differently labeled FISH probes (MYCN and D2Z) hybridized to three different neuroblastoma cell lines, six peripheral blood (PB) samples, 10 spiked PB samples, and nine neuroblastoma samples using the Metafer4 system (MetaSystems, Altlussheim, Germany). To obtain the MYCN copy number per cell, the ratio between the fluorescence intensities of the MYCN gene and reference sequence (D2Z) was calculated. For automatic analysis of the HER-2/neu status in tumor cells, labeled FISH probes specific for HER-2/neu and a chromosome 17-specific probe were hybridized to peripheral blood and tumor specimens and analyzed using the automatic device.
RESULTS: When measuring the fluorescence intensity per cell for both probe pairs (MYCN/D2Z and HER-2/17p), amplified and non-amplified cells, showed distinct peaks with only little overlap. Whereas normal cells showed a fluorescence ratio peak for MYCN/D2Z between 200 and 800, cells with MYCN amplification clearly exceeded this ratio value (1000 to 25,000). When mixing a varying number of MYCN amplified cells (range 9-91%) to normal PB, the spiked tumor cells could be identified. Even one neuroblastoma tumor cell in 1000 mononucleated cells could reliably be detected using our device. In neuroblastoma patient samples, non-amplified cells were distinguished from amplified cells. Automatically and manually counted signals gave matching results in amplified and non-amplified samples. HER-2/neu-amplified cells were automatically detected in the breast cancer samples analyzed.
CONCLUSION: The automatic measurement of fluorescence signal intensities not only allows a reliable discrimination between non-amplified and amplified cells but also exact quantification of amplified sequences. This is the prerequisite for the following applications: detection of amplified cells in the bone marrow and second-look specimens; comparison between primary and relapse or pre- and post-chemotherapeutic specimens; detection of tumors with focal gene amplification; and quantification of elimination of amplified gene sequences. Copyright 2003 Wiley-Liss, Inc.