Blood, 15 November 2002, Vol. 100, No. 10, pp. 3443-3444
Phages that display P-selectin antagonism
Selectins mediate adhesion events among leukocytes, platelets,
and endothelial cells. Within the selectin family, P-selectin plays a
predominant role in the recruitment of polymorphonuclear neutrophils
into inflammatory sites and, thus, represents an important target for
future therapies against various inflammatory diseases. Despite intense
research efforts over the past 20 years, effective selectin inhibitors
have not yet become available to clinicians. Since all 3 selectins bind
sialylated and fucosylated oligosaccharides (for example, sialyl Lewis
X), much of the work has focused on the generation of specific
glycoconjugates that generally display relatively low binding
affinities when presented as mono- or oligovalent ligands. In its
natural environment, the chief P-selectin glycoprotein ligand (PSGL-1)
indeed requires both the specific carbohydrate decorations and sulfated
N-terminal tyrosine residues to bind P-selectin with high affinity
(nanomolar range).
Molenaar and colleagues (page 3570) used phage-display libraries to
isolate peptides containing the consensus sequence EWVDV, which
can specifically inhibit human P-selectin but not mouse P-selectin or
human E- or L-selectins. The binding affinity of an EWVDV-containing
peptide for P-selectin was greatly enhanced (200-fold) when presented
as a tetramer. Importantly, Molenaar et al also show that nanomolar
concentrations of the tetrameric peptide can efficiently inhibit the
adhesion of HL60 cells to P-selectin in static assays and increase
rolling velocities of HL60 cells in a flow system. How do such small
peptides do it? Because the EWVDV-containing peptide can inhibit the
binding of a carbohydrate selectin ligand (sulfated Lewis A), the
authors speculate that EWVDV may interact in close proximity of the
carbohydrate recognition domain of P-selectin. The fact that EWVDV
binding to P-selectin is only partially Ca++
dependent suggests that the peptide may also interact outside the
carbohydrate-binding pocket whose binding is strictly Ca++
dependent. Since the consensus sequence contains 2 acidic residues (glutamic acid and aspartic acid), its overall negative charge may also
promote interactions with the region of positive electrostatic potential previously shown to mediate high-affinity interactions with
the sulfotyrosines of PSGL-1 (Somers et al, Cell. 2000;103:467-479). This feature might also explain the dramatically enhanced avidity of
multimeric peptides. Beyond their physicochemical properties, the real
litmus test for the peptides will be to determine their capacity to
inhibit leukocyte adhesion on activated endothelium in vivo. This
remarkable study, however, provides a critical first step in the
generation of powerful selectin antagonists that have profound clinical relevance.
Paul S. Frenette
Mount Sinai School of Medicine
