Viral vectors have been used in vitro and in vivo for more than a decade, with some significant results in specific situations, e.g., when recombinant adeno-associated virus is used for the long-term transduction of skeletal muscle in coagulation factor IX-deficient patients. However, the kidney has been quite difficult to transduce with any viral vector currently available. When viral transduction occurs, it is often heterogeneous, transient, and eventually associated with immune and toxic side effects. However, recombinant adeno-associated virus and lentiviral vectors remain to be fully evaluated in the kidney; the former is small enough to be filtered through the glomerular basement membrane. This may be critical, because glomerular filtration is required for DNA complex-mediated transduction of tubular cells. An alternative to in situ renal gene transfer is secretion of a therapeutic protein from a distant site, such as skeletal muscle. Several examples provide evidence that this could be a clinically relevant approach. It also may allow accurate determination of the pathophysiologic mechanisms involved in the establishment and maintenance of experimental glomerulonephritis.