Changes in the transcriptome of human tumor cells following dissociation and culture

Abstract

4962

Primary cultures of cells isolated from human tumors have been used to test drug sensitivity in what have been termed in vitro phase 2 studies. However, the transport of the tumor tissues and dissociation of the cells represent potential stresses for them. Therefore, we have examined the effect of dissociation and culture on gene expression patterns in human lung cancer tissues. A portion of the tumor was placed in RNALater immediately after diagnostic specimens were taken. The remainder was placed in tissue culture medium. After transportation to the laboratory, the cells from the viable sample were dissociated, a portion retained for RNA analysis and the rest placed in culture for 1 or 5 days. RNA was prepared from all 4 samples and analyzed using Affymetrix HG-U133 Plus 2 arrays. The percentage of genes with significant signal levels (present calls) was 52.6 % in the tissue that had been put directly into RNALater, 47.2 % in the RNA from cells immediately after transportation in culture medium and dissociation (day 0), 48.8 % from cells cultured for 1 day, and 44.7 % from cells cultured for 5 days. The numbers of genes that are over or under expressed compared to the RNALater sample by 10-fold or more immediately after dissociation was 156. This dropped 57% in the first day of culture to 67 genes and to 30 genes by day 5 of culture. The numbers of genes differing from the sample placed in RNALater by 5-fold or more was 904 at day 0, falling to 528 at day 1 and 364 by day 5 of culture (40% of day 0). A similar trend is seen when the cutoff is genes differing by 3-fold or more. The most highly up regulated genes (> 10-fold) in dissociated cells at day 0 include carcinoembryonic antigen-related cell adhesion molecule 5 (CEACAM5), carcinoembryonic antigen-related cell adhesion molecule 6 (non-specific cross reacting antigen), Human heat-shock protein HSP70B, heat shock 70kD protein 6 (HSP70B) (HSPA6), growth arrest and DNA-damage-inducible 34 (GADD34), glutathione S-transferase A2 (GSTA2), interferon-gamma (HuIFN-gamma), interleukin 8 (IL8), bone morphogenetic protein 2, and interleukin 1, beta (IL1B). Those genes that were most down regulated (< 10 % of expression in intact tissue) in dissociated cells were S100 calcium-binding protein A2 (S100A2), laminin, beta 1 (LAMB1), tissue factor pathway inhibitor beta (TFPIbeta), aldo-keto reductase family 1, member C1 (dihydrodiol dehydrogenase 1; 20-alpha (3-alpha)-hydroxysteroid dehydrogenase), and pseudo-chlordecone reductase. These results indicate that some genes related to drug resistance or inflammation may by highly up regulated or down regulated when fresh tumor cells are put into culture. These changes decrease after time in culture suggesting that short-term stresses may drastically affect drug tests with primary tissues and a period acclimation to culture would lessen potential artifacts.