Do circulating peptides contain 'zip codes' that allow them to home specifically to specific vascular beds? If so, one might imagine that discovering these tissue and/or organ-specific peptides might allow for, among other things, tissue-specific targeting of normal blood vessels and of angiogenesis-related targeting of tumor blood vessels. Indeed, Pasqualini and colleagues have shown that such a vascular address system exists in the mouse; a number of peptides that have been shown in mouse models to home specifically to certain blood vessels have been used to deliver cytotoxic drugs and peptides in a tissue-specific fashion. Moreover, coupling drugs to homing peptides, in general, has yielded target compounds that are more effective and less toxic than the parent one.

However, although certain ligands and receptors isolated in mouse models have been used to identify putative human homologs, targeted delivery may not always occur in humans using mouse-derived probes. Indeed, data from the Human Genome Project indicates that the greater complexity of humans derives in part from expression patterns of proteins at different sites and/or levels, rather than a greater number of genes per se. Moreover, there have been many examples of species-specific differences in gene expression within the human vascular network.

On pages 121–127, Pasqualini and colleagues now report the first step towards developing a ligand-receptor map of human vasculature—identifying proteins that might be specific to vascular beds within different organs. The authors injected a peptide library containing over 47,000 motifs into a terminally ill patient. They then took tissue samples to identify the distribution of the circulating peptides. They found that the distribution of the peptides is non-random, with some peptides homing to very specific vascular beds and others targeting multiple sites. Moreover, they were able to find certain peptide motifs in circulating ligands that home to specific vascular receptors. For instance, they show by phage overlay on human tissue sections that a prostate-homing page with an IL-11 peptide mimic specifically binds the endothelium of normal prostate (a) but not of skin (b).

By providing a rudimentary map of the molecular diversity of the vascular system, Pasqualini and colleagues have opened a new avenue to obtain biologically relevant information that could also be used to develop clinical applications. It may now be possible to use this technique to determine the molecular profiles of blood vessels in specific disease conditions, and to exploit this vascular address system to specifically target certain organs and disease states.