The current understanding of platelet-derived growth factor (PDGF) physiological functions in vivo is discussed in the context of mouse development. In particular, the review focuses on recent experiments in which genetic approaches have been applied in order to mutate the PDGF and PDGF receptor genes in the mouse. Thus, the PDGF-B and PDGF beta receptor (PDGFRb) genes were recently inactivated by homologous recombination in embryonic stem cells. Their phenotypes are highly similar, displaying cardiovascular, hematological and renal defects. The latter is particularly interesting since it consists of a specific cellular defect: the complete loss of kidnety glomerular mesangial cells. As such, the phenotype not only sheds light on the developmental importance of PDGF-B-PDGFRb interactions, but also reveals information about the function of mesangial cells. Based on detailed morphological studies of mutant glomeruli and the absence of urine collection in the urinary bladder, I propose that the mesangial cells function as interior" filter holders", the" filter" consisting of the glomerular basement membrane and associated cell types. The filter holder model would predict that glomerular filtration is critically dependent on an interior structural support of the filter, which is normally provided by the mesangial cells and the mesangial matrix. In addition to the mutants generated by gene targeting, the mouse patch mutation is discussed. This deletion encompasses the PDGFRa locus. The last part of the review focuses on the problems encountered when interpreting gene knockout phenotypes in the physiological functions of gene products.