Healing gets messy without a fibronectin isoform.

On page 149, Muro et al. describe an elegant strategy for studying the functions of alternatively spliced isoforms of fibronectin. The work reveals surprisingly subtle phenotypes in mice that cannot carry out one form of alternative splicing, and demonstrates a general approach that could be applied to virtually any alternatively spliced gene.Disruption of the fibronectin gene in mice is lethal to embryos, making it difficult to analyze the developmental functions of the protein's isoforms. The authors got around this obstacle by inserting optimized splice sites and loxP recombination sites in the introns adjacent to the EDA exon of fibronectin. In mice homozygous for this allele, the EDA exon is constitutively included in fibronectin. Crossing this strain with a CRE recombinase– expressing “deleter” mouse produced animals lacking the EDA exon.

In wild-type mice, most of the embryonic fibronectin includes the EDA exon, but most of the adult fibronectin lacks it. Despite this pattern, the new work shows that mice that constitutively include or exclude EDA develop normally. Constitutive inclusion of EDA, however, causes a substantial decrease in total fibronectin levels in adult tissues, whereas a lack of EDA causes abnormal cutaneous wound healing. Both mutant strains have significantly shorter lifespans than wild-type mice, possibly because of defects in tissue repair and regeneration. ▪