Abrogation Of Gap Junctional Communication In ES Cells Results In A Disruption Of Primitive Endoderm Formation In Embryoid Bodies

Gap junctional intercellular communication (GJIC) has been suggested to be involved in early embryonic development but the actual functional role remained elusive. Connexin (Cx) 43 and Cx45 are co-expressed in embryonic stem (ES) cells, form gap junctions and are considered to exhibit adhesive function and/or to contribute to the establishment of defined communication compartments. Here we describe the generation of Cx43/Cx45-double deficient mouse ES cells to achieve almost complete breakdown of GJIC. Cre-loxP induced deletion of both, Cx43 and Cx45, results in a block of differentiation in embryoid bodies without affecting pluripotency marker expression and proliferation in ES cells. We demonstrate that GJIC-incompetent ES cells fail to form primitive endoderm in embryoid body cultures, representing the inductive key step of further differentiation events in this system. Lentiviral overexpression of either Cx43 or Cx45 in Cx43/45 mutants rescued the observed phenotype, confirming the specificity and indicating a partially redundant function of both connexins. Upon differentiation GJIC-incompetent ES cells exhibit a strikingly altered subcellular localization pattern of the transcription factor NFATc3. Control EBs exhibit significantly more activated NFATc3 in cellular nuclei than mutant EBs suggesting that Cx-mediated communication is needed for synchronized NFAT activation to induce orchestrated primitive endoderm formation. Moreover, pharmacological inhibition of NFATc3 activation by Cyclosporin A, a well-described inhibitor of calcineurin, phenocopies the loss of GJIC in control cells. For this study we analyzed embryonic stem (ES) cells double deficient for connexin43 and connexin45 (Cx43/45 del ESC) as well as control ESCs (Cx43/45 flox ESC) and embryoid bodies at day 6 of differentiation derived thereof (Cx43/45 flox EB and Cx43/45 del EB). For the ESC samples 2 biological replicates were included. For the Cx43/45 flox EBs 3 biological replicates and for the Cx43/45 del EBs 4 biological replicates were used. RNA was isolated using the RNeasy-Kit (Qiagen, Hilden, Germany). Expression analysis had been performed following the Illumina (Illumina Inc., San Diego, CA, USA) Whole-Genome Gene Expression Direct Hybridization Assay analysis pipeline. mRNA transcription levels were evaluated using the MouseWG-6 (version 2, revision 3) array which queries 45281 probes in 30854 genes and described mRNA features. Data processing was performed using the GenomeStudio suite version 2011.1 and the Gene Expression module version 1.9.0 (both Illumina Inc., San Diego, CA, USA). The downstream analyses were performed using R/Bioconductor packages and Chipster, a graphical user interface for gene expression profiling (Kalio et al., 2011). The raw intensities were quantile normalized.