Statistical analysis
Real time PCR data were evaluated with the REST software [65] that uses a randomisation test with a pairwise reallocation to assess the statistical significance of the differences in expression between each treatment-exposure group and the control. cAMP levels and PKA activities were analysed using the SigmaStat statistical package. Significant differences between samples were determined using one-way ANOVA followed by the multiple comparison Bonferroni's test. In any case, statistical difference was accepted when p<0.05.