Erythroid differentiation: Murine erythroleukemia cell induction
Ontology highlight
ABSTRACT: Transcriptional profiling of murine erythroleukemia cells induced to differentiate Murine erythroleukemia (MEL) cells are derived from mice infected with the Friend erythroleukemia viral complex, which causes a lethal expansion of erythroid cells. The transformed, immortalized cells (MEL cells) isolated from infected mice are committed to erythroid differentiation, but are blocked roughly at the proerythoblast stage. Treatment with small, oxidized organic compounds like dimethyl sulfoxide or hydroxymethyl-bis-acetamide (HMBA) induce the cells to mature further, to roughly the erythroblast stage or beyond in some lines (Friend et al. 1971). During this induction, the amount of hemoglobin, some heme biosynthetic enzymes and other proteins characteristic of mature erythroid cells increases, largely as a result of changes in transcription (Aviv et al. 1976). A full description of changes in transcription of all mouse genes during this process is not available. To improve this situation, we have conducted transcriptional profiling of the RL5 line of MEL cells, using microarrays containing almost 7000 mouse genes. The RL5 line has a "random locus" containing a positive-negative selectable marker (HyTK) flanked by loxP sites, thus facilitating site-directed integration into the cells (Bouhassira et al. 1997; Feng et al. 1999; Molete et al. 2001). This particular line is not highly inducible, with less than half the cells accumulating large amounts of hemoglobin. However, they are advantageous for integration of expression constructs, which is our particular interest. MEL_RL5 cells were treated with 4 millimolar HMBA for 6 days to induce maturation, and RNA was isolated each day. Complementary DNA was synthesized and labeled with either Cy5 (red fluorescence) for the induced samples or Cy3 (green fluorescence) the untreated cells grown in parallel. For most time points, at least two independent experiments were run, some with dyes switched (SD). The microarrays were spotted at the PSU Microarray Facility. Hybridization was carried out in our laboratory, and hybridized chips were scanned on a GenePix scanner and quantified using GenePix software. The signals from each channel (red and green) were normalized so that the total red intensity equals the total green intensity. Data were filtered to remove "bad" spots, which are those in which the noise exceeded the signal. We calculated the median of the pixel intensities for each spot and standard deviation around the median, and with these values we computed a "modified Z-score" for each channel on each spot. A lenient threshold for the "modified Z-score" was applied to remove "bad" spots. After filtering, we performed global mean normalization for each chip, followed by lowess normalization for each chip to remove intensity-dependent effects.
ORGANISM(S): Mus musculus
SUBMITTER: Ross Hardison
PROVIDER: E-GEOD-2217 | biostudies-arrayexpress |
REPOSITORIES: biostudies-arrayexpress
ACCESS DATA