Quantitative real-time polymerase chain reaction (PCR) is an effective approach in investigating the effects of exogenous compounds on gene expression. This is often achieved by exploiting so-called 'housekeeping' genes as baseline controls to normalise expression levels, which have historically been assumed to have a relatively stable expression pattern. Recent non-in-vitro studies have questioned the validity of this, but previous in-vitro data were lacking following antidepressant treatment. We here investigated the stability of 12 housekeeping genes during treatment of the mouse L929 fibroblast cell line with escitalopram and nortriptyline. Cells were cultured in the presence of antidepressant at 1 microM or 10 microM for 30 min, 24 h or 48 h, and RNA subjected to quantitative PCR (qPCR). Stability of relative transcript expression values was assessed via gene-gene expression ratios and intra- and inter-group variation (using geNorm and NormFinder programs). The three most stable transcripts were adenosine triphosphate (ATP) synthase, H+ transporting mitochondrial F1 complex, beta subunit, beta-2 microglobulin and cytochrome c-1. The least stable were Gapdh, eukaryotic translation initiation factor 4A2 and Calnexin (Canx). In conclusion, care must be taken when choosing reference transcripts for analysis in qPCR. For in-vitro pharmacological studies, it should not be assumed that 'housekeeping' genes are stable.