ABSTRACT: Background: The Spemann/Mangold organizer is a transient tissue critical for patterning the gastrula stage vertebrate embryo and formation of the three germ layers. Despite its important role during development, there are still relatively few genes with specific expression in the organizer and its derivatives. Foxa2 is a forkhead transcription factor that is absolutely required for formation of the mammalian equivalent of the organizer, the node, the axial mesoderm and the definitive endoderm (DE). However, the targets of Foxa2 during embryogenesis, and the molecular impact of organizer loss on the gastrula embryo, have not been well defined. Results: To identify genes specific to the Spemann/Mangold organizer, we performed a microarray-based screen that compared wild-type and Foxa2 mutant embryos at late gastrulation stage (E7.5). We could detect genes that were consistently down-regulated in replicate pools of mutant embryos versus wild-type, and these included a number of known node and DE markers. We selected 314 genes without previously published data at E7.5 and screened for expression by whole mount in situ hybridization. We identified 10 novel expression patterns in the node and 5 in the definitive endoderm. We also found significant reduction of markers expressed in secondary tissues that require interaction with the organizer and its derivatives, such as cardiac mesoderm, vasculature, primitive streak, and anterior neuroectoderm. Conclusions: The genes identified in this screen represent novel Spemann/Mangold organizer genes as well as potential Foxa2 targets. Further investigation will be needed to define these genes as novel developmental regulatory factors involved in organizer formation and function. We have placed these genes in a Foxa2-dependent genetic regulatory network and we hypothesize how Foxa2 may regulate a molecular program of Spemann/Mangold organizer development. We have also shown how early loss of the organizer and its inductive properties in an otherwise normal embryo, impacts on the molecular profile of surrounding tissues. Experiment Overall Design: Microarrays: Experiment Overall Design: Tetraploid embryos were derived as previously described (Nagy et al. 1990; Nagy et al. 1993) using Foxa2 null ES cells (Dufort et al. 1998). Wild type embryos were collected from timed matings of ICR mice. Wild type or tetraploid embryos were collected at E7.5 and staged accordingly (Downs and Davies 1993). Embryos were between mid-streak and head-fold stages and the extra-embryonic regions were removed by dissection. The embryo proper, including extra-embryonic tetraploid-derived VE, was retained and multiple samples were pooled for total RNA isolation using Trizol reagent. Numbers of E7.5 embryos collected and pooled were as follows: wild type 1 (n=53), wild type 2 (n=50), Foxa2 null 1 (n=53), Foxa2 null 2 (n=48). Total RNA from each of the four embryo pools was processed for Affymetrix Murine Genome U74Av2 GeneChip analysis as previously described (Lickert et al. 2005). Experiment Overall Design: Data analysis: Experiment Overall Design: Affymetrix MAS 5.0 software was applied to the U74Av2 GeneChip data to target normalize the global expression level to 1000, and provide present, absent and fold change calls in a cross-comparisons of all replicate mutant samples over the two replicate wild-type samples. Next probe sets with absent calls in all samples were removed from the data set. As well, all samples with illogical calls were also removed, for example a probe set with an increase call that was called absent in the sample with the higher level of expression. Lastly, we filtered the data to include only probe sets that showed a statistical change call in two or more (2/4) of the cross comparisons. Experiment Overall Design: MOE430v2 GeneChips were analyzed as follows in the R statistical programming language (R Core Development Team, 2006; see Additional File 11 for details). Probeset summaries were calculated according to the MAS 5.0 algorithm. Logarithmic summary values were normalized by the loess smoother applied to the M-A scale transform. Statistical testing for differential expression between mutant and wild-type was performed on single probe level taking the duplicate chips into account employing the Wilcoxon paired rank test. The p-value threshold was adjusted by either family-wise error rate (FWER, Bonferroni procedure) or the false discovery rate (FDR, Bejamini-Hochberg technique). Significant probesets were tested for enrichment of GO terms by the hypergeometric distribution testing with the annotation as available from the GO consortium on March 25th, 2005.
