Denys–Drash syndrome (DDS) is caused by dominant mutations of the Wilms' tumour suppressor gene, WT1 , and characterized by a nephropathy involving diffuse mesangial sclerosis, male pseudohermaphroditism and/or Wilms' tumourigenesis. Previously, we reported that heterozygosity for the Wt1^tmT396 mutation induces DDS in heterozygous and chimeric ( Wt1^tmT396 /+↔+/+) mice. In the present study, the fate of Wt1 mutant cells in chimeric kidneys was assessed by in situ marker analysis, and immunocytochemistry was used to re-examine the claim that glomerulosclerosis (GS) is caused by loss of WT1 and persistent Pax-2 expression by podocytes. Wt1 mutant cells colonized glomeruli efficiently, including podocytes, but some sclerotic glomeruli contained no detectable Wt1 mutant cells. The development of GS was preceded by widespread loss of ZO-1 signal in podocytes (even in kidneys where <5% of glomeruli contained Wt1 mutant podocytes), increased intra-renal renin expression, and de novo podocyte TGF-β1 expression, but not podocyte Pax-2 expression or loss of WT1, synaptopodin, α-actinin-4 or nephrin expression. However, podocytes in partially sclerotic glomeruli that still expressed WT1 at high levels showed reduced vimentin expression, cell cycle re-entry, and re-expressed desmin, cytokeratin and Pax-2. The results suggest that: (i) GS is not due to loss of WT1 expression by podocytes; (ii) podocyte Pax-2 expression reflects re-expression rather than persistent expression, and is the consequence of GS; (iii) GS is mediated systemically and the mechanism involves activation of the renin–angiotensin system; and (iv) podocytes undergo typical maturational changes but subsequently de-differentiate and revert to an immature phenotype during disease progression.