Altered signaling, but no cell fragility in mice lacking all type II keratin genes

Abstract

<p>Keratins are major cytoskeletal proteins of epithelia, organized in two large gene families of 28 (type I) and 26 (type II) members spanning 1.2 Mb and 0.68 Mb of genomic loci, respectively. Their expression is tightly regulated in a pairwise and differentiation-related manner in epithelial tissues starting from the 2-cell stage during embryogenesis. Fertilized mammalian eggs and embryonic stem cells are devoid of keratin filaments, raising the question whether they may be involved in stem cell differentiation. Studies from various keratin mutations leading to tissue fragility placed protection against mechanical stress as the primary function of keratins. Recent research has led to the hypothesis that keratins have major regulatory functions in cell proliferation, translation regulation, protection against apoptosis, organelle transport and regulation of signal transduction. The overlapping expression of several type I and II keratins in most epithelia has generated functional redundancy, and therefore has presented a major obstacle towards the analysis of keratin function. This thesis has addressed the function of keratins during development by constitutively and conditionally deleting all type II keratin genes in mice, thereby eliminating all problems related to redundancy. This is the first reported deletion of a large gene family in mammals.<br /> The <i>CrelloxP</i> system was adopted in order to flox the type II keratin gene cluster in mouse ES cells. This floxed allele contained all the type II keratins including the type I keratin Krt18, which is one of the first keratins to be expressed during embryonic development. Mice were generated from the floxed ES cells (for conditional deletion) as well as from the Cre expressed ES cells (for constitutive deletion), thereby carrying either the floxed or deleted keratin type II cluster, respectively. This thesis, however, has dealt only with the constitutive deletion of the keratin type II cluster, while the conditional deletion of the cluster in tissues of adult animals is underway. <br /> Mice heterozygous for the deleted keratin type II cluster were inbred to generate keratin type II null mice. However, no keratin type II null pups were recovered at birth. Mutant embryos were smaller than their wild type littermates from ~E9.0, and died at mid gestation (E9.5). Despite the absence of a keratin cytoskeleton, no cell fragility was found in contrary to many other individual keratin knockouts. This shows that the phenotype reported for previous keratin gene knockouts represents gain of toxic functions. No overt proliferation defects were observed in the mutants even around the time of death suggesting that no proliferation defects accounted for lethality. Mutant embryos did exhibit the expected formation of epithelial lineages during embryogenesis, indicating no apparent role of keratins during embryonic differentiation. Mutant embryos suffered from defective vasculogenesis and angiogenesis in the yolk sac, which are the most critical structures required to nourish the embryo, before the placenta can take over from E9.0. Reduced vasculogenesis was accompanied by the loss of endodermalmesothelial cell interactions. Moreover, the mutants failed to form an organized labyrinth and the spongiotrophoblast layer at E9.5, although the chorioallantoic attachment at E8.5 was normal. Following chorioallantoic attachment at E8.5, a subset of chorionic epithelial cells expressing Syncytin A, destined to commit to the branching morphogenesis, failed to elongate and inter-digitate to form a functional labyrinth, but rather remained flat. Elevated levels of apoptosis in non-epithelial cells were seen in smaller mutant embryos in comparison to the fairly larger mutants and the wildtype embryos, suggesting a positive correlation between the two. This indicated that apoptosis was not a direct consequence of the absence of keratins but a secondary effect due to nutritional deficiency. Unexpectedly, increased apoptosis did not elevate the pro-apoptotic signal transducer p38 MAP kinase. Infact, a reduced level of activated p38 MAP kinase was observed in the mutants. In agreement, the phenotype of the keratin type II null mutants showed a striking similarity to that observed in p38-α knockouts, although they died a day later than the keratin null mutants.<br /> The observations in this study have revealed major roles of keratins in epithelial-endodermal interactions and cell signaling. Lack of cell fragility in the absence of keratins suggests that mechanical support cannot be considered as their primary function, at least during embryonic development. Therefore, knockouts of individual keratins reported so far, represent dominant-negative, but no loss-offunction, mutations. It further questions previously reported roles of keratins in cell cycle, translational control and apoptosis. Keratins seem to mediate cell signaling, for which their topological organization within a cell is well suited. The data from the keratin type II null mutants suggests that keratins play a role upstream of the p38 MAP kinase pathway. The potential of these mice will be exploited to a great extent in the future by generating cell lines free of keratins for various functional assays as well as conditionally deleting the cluster in a tissue specific and spatially controlled manner. This will for the first time allow to study their role in tumor development and chronic diseases in genetically well-defined settings.</p>