ESL-1 Is a Major Physiological Leukocyte Ligand for E-Selectin That Cooperates with PSGL-1 and CD44, and Together Mediate All Binding Activity to Endothelial Selectins In Vivo.

Endothelial selectins mediate leukocyte (WBC) rolling on activated endothelium. ESL-1 was isolated by affinity purification more than 10 years ago but whether it is a physiological ligand for E-selectin remains unclear. Recent studies suggest that PSGL-1 and CD44 are physiological E-selectin ligands on neutrophils (PMNs)(J. Exp. Med. 2005 201:1183). To analyze the contribution of ESL-1, PSGL-1 and CD44, we knocked down ESL-1 expression using lentiviral transfer of short hairpin (sh) RNA into wild-type (WT), PSGL-1- or/and CD44-deficient bone marrow lineage-negative (Lin-) cells. Sorted GFP+ cells from mice transplanted with Lin- cells transduced with the shESL-1 exhibited >90% reduction in total ESL-1 levels by Q-PCR and Western blot analyses, and complete absence of ESL-1 on the PMN surface. shESL-1 transduced Lin- cells engrafted recipients mice with the same efficiency as control vector. Downregulation of ESL-1 slightly affected (~33%) the binding of soluble E-selectin to PMNs, but did not alter WBC rolling numbers in TNF-α-stimulated cremasteric venules. Absence of ESL-1 or CD44, but not PSGL-1, resulted in significant increases in mean rolling velocities (4.7 ± 0.4 μm/s for WT; 8.3 ± 0.4 μm/s for shESL-1*; 7.2 ± 0.2 μm/s for CD44−/−* and 5.4 ± 0.4 μm/s for PSGL-1−/−; *p<0.001). Moreover, WBCs lacking ESL-1 exhibited a characteristic “skipping” rolling motion, suggesting that ESL-1 plays a critical role in stabilizing WBC rolling. Remarkably, ESL-1 downregulation in the absence of PSGL-1 led to dramatic reductions in WBC binding to soluble E-selectin (~99%), and rolling (~93%) on TNF-α-stimulated venules. To determine the contribution of these 3 glycoproteins in WBC recruitment, mice transplanted with WT, PSGL-1- and/or CD44-deficient Lin- cells that were transduced with the shESL-1 or control vectors, were injected i.p. with thioglycollate and the number of peritoneal GFP+ PMNs determined. Only the absence of all 3 glycoproteins compromised PMN recruitment to the levels observed in P- and E-selectin double-deficient mice (82% reduction compared to WT; p<0.001), suggesting that ESL-1, PSGL-1 and CD44 comprise all endothelial selectin ligand activity on PMNs in vivo. To assess the role of E-selectin ligands in WBC activation, we monitored in vivo L-selectin clustering on the surface of over 4,000 rolling WBCs using high-speed multichannel fluorescence videomicroscopy. The majority of wild-type rolling WBCs exhibited polarization of L-selectin 150 to 230 min after TNF-α. L-selectin clustering was greatly reduced in E-selectin−/− mice (58 ± 4% polarized cells in WT vs 23.5 ± 1% in E-selectin−/− mice; p<0.001). To determine which E-selectin ligand(s) mediated L-selectin clustering in vivo, we evaluated L-selectin distribution on rolling WBCs deficient in ESL-1, CD44 or PSGL-1. Absence of CD44, but not ESL-1 or PSGL-1, reduced L-selectin redistribution to levels found in E-selectin−/− mice (20 ± 2% polarized cells), indicating that CD44 is the signalling E-selectin ligand required for L-selectin clustering. In addition, E-selectin-mediated CD44 signalling was p38-dependent, as assessed in WT mice treated with the p38 inhibitor SB203580 (77% reduction; p<0.001). Our results indicate that ESL-1, PSGL-1 and CD44 comprise all selectin ligand binding activity on WBCs, and suggest specialized contributions for each ligand where PSGL-1 mediates the initial tethering, ESL-1 is critical for the conversion into steady slow rolling, and CD44 is required for E-selectin-mediated signals contributing to WBC activation.