Monocytes constitute a heterogeneous cell population, comprising approximately 5–10% of total peripheral blood leukocytes. These cells arise from granulocyte–macrophage progenitors in the bone marrow and are subsequently released into peripheral blood, where they circulate for several days (5). At steady state (i.e., without any inflammatory cue), monocytes can enter non-lymphoid tissues, and there they either retain their blood monocytic behavior (6), or generate the immediate precursors of “monocyte-derived macrophages and DCs,” which constitute a small portion of tissue-resident macrophage and DC populations (7–9). On the other hand, under inflammatory conditions, monocytes transmigrate into injured tissues, where they then directly mediate antimicrobial activity or, depending on the local biochemical milieu, differentiate into inflammatory macrophages or monocyte-derived DCs (moDCs) (10) (Figure 1). Circulating monocytes, thus, function as a systemic reservoir of tissue-resident myeloid cells (11, 12).

There are three subsets of human monocytes, each of which display different functional and migratory abilities and can be distinguished based on their expression of specific chemokine receptors, CD14 [the lipopolysaccharide (LPS) receptor], and CD16 (Fcγ RIII) (13). “Classical” monocytes (CD14⁺⁺CD16⁻), which account for about 90% of circulating monocytes in healthy individuals, express high levels of the C-C chemokine receptor type 2 (CCR2), display high phagocytic and myeloperoxidase activities, generate reactive oxygen species, and produce inflammatory cytokines, such as interleukin (IL)-1β, IL-6, and tumor necrosis factor (TNF)-α (14). On the other hand, the “non-classical” monocytes (CD14⁺CD16⁺⁺) comprise a population that exhibits low phagocytic and myeloperoxidase activities (15, 16). Importantly, while classical monocytes are recruited preferentially to distressed tissues (17), non-classical monocytes are recruited to non-inflamed areas, where they patrol the microvasculature via the CX3C chemokine receptor 1 (CX3CR1) and leukocyte function-associated antigen (LFA)-1, monitoring the luminal surface of resting endothelium for signs of tissue damage or infection (18–20). In addition, non-classical monocytes are mainly responsive to virus-associated signals, via toll-like receptors (TLRs) 7 and 8, whereas classical monocytes respond mostly to bacteria-associated signals (21). An intermediate subset of monocytes, characterized as CD14⁺⁺CD16⁺, is viewed as being a transitional population between classical and non-classical monocyte subsets, displaying significant production of TNF-α and IL-1β, but low peroxidase activity (16, 22). While the migratory ability of the intermediate subset is controversial, they express the chemokine receptor CCR2, a feature supporting their ability to infiltrate sites of inflammation (23). Still, overall, the intermediate monocyte population reportedly displays weaker ability to migrate across resting endothelium compared to the other two monocytic subsets (24).

Dendritic cells are the antigen-presenting cells par excellence, showing a unique capacity to initiate immune responses. These specialized antigen-presenting cells constitute a unique leukocyte population that display high morphological and functional heterogeneity (25). DCs can be originated from common myeloid or lymphoid precursors and are divided into two main groups: conventional DCs (cDCs) and plasmacytoid DCs (pDCs) (26). After being released into the bloodstream, they are distributed to lymphoid organs (lymph nodes, spleen, and thymus) and various peripheral tissues. DC function is intrinsically related to their anatomical localization, and, therefore, a stringent DC functional-anatomical classification needs to be defined (Figure 1). At steady state, DCs are found to be immature (as indicated by high phagocytic and endocytic capacity and low expression of MHC and costimulatory molecules) and can be classified as either migratory or lymphoid-resident DCs (27, 28). Migratory DCs serve as immune sentinels screening peripheral tissues for signals of danger. They can also capture apoptotic cells or self-antigens in non-inflamed tissues and, after entering lymph nodes via afferent lymphatics, present these to T cells in the lymph nodes, thus playing a key role in antigen-mediated peripheral tolerance (29–31). On the other hand, lymphoid-resident DCs differentiate within lymphoid organs directly from blood DC precursors, and they function to continuously survey blood or lymph (27, 32). Both cDC and pDC hematopoietic progenitors contribute to the lymphoid-resident DC pool, whereas most migratory DCs arise from blood cDCs (33). Under infection or sterile inflammatory circumstances, both circulating cDCs and classical monocytes enter inflamed tissues, where they capture antigens and differentiate into highly functional mature DCs. The mature DCs migrate to the lymph nodes via the afferent lymph, initiating T cell-mediated immune responses.

In addition to cDCs and pDCs, a distinct subset of DCs are derived from monocytes (known as “moDCs”) which are considered to be “inflammatory DCs”; these cells prominently produce TNF-α, nitric oxide, and IL-23, and are potent inducers of TH17 cells (34–37). Interestingly, although pDCs are believed to be absent from peripheral tissues under steady-state conditions, a number of recent publications reported pDC extravasation into some inflamed tissues, where they secrete large amounts of type I interferon (38–42). In contrast to cDCs, pDCs do not enter reactive secondary lymphoid organs after trafficking from peripheral tissue via afferent lymphatics; instead, they apparently migrate directly from the bloodstream via high endothelial venules (HEVs) by an E-selectin-dependent mechanism (43–47).

Macrophages are a heterogeneous and versatile population of tissue-resident cells, mostly originating from self-renewing embryo-derived progenitors and from blood monocytes that have colonized tissues (48, 49). They exist virtually in every tissue throughout the body, where they survey for potential signs of infection/danger and perform phagocytic clearance of dying cells (50). In addition, macrophages play a role in adaptive immunity through antigen presentation and production of cytokines (51, 52).

There are two main macrophage subsets, the M1 and the M2 macrophages, with distinct responses to environmental signals. The M1 subset produces high amounts of pro-inflammatory cytokines and reactive oxygen and nitrogen species, thus playing a crucial role in Th1 polarization and promotion of cellular immunity. M2 macrophages are characterized by their ability to stimulate humoral immune responses, fight extracellular parasite infections, and promote tissue repair, angiogenesis, and tumor progression (53, 54). Whereas the major function of macrophages is to fight infections and kill target cells, they do not typically display hematogenous migration, nor leave sites of tissue injury (11, 55).

Recruitment of circulating cells from blood to inflamed tissue involves a sequential and coordinated series of molecular actions mediated by adhesive interactions between circulating sentinels and endothelial cells in post-capillary venules (56). Here, we review the molecular effectors that regulate the initial phagocyte–endothelial binding interactions, which are essential for transendothelial migration of blood monocytes and DCs to sites of injury.

To initiate the extravasation process, circulating phagocytes establish low-affinity and reversible interactions (tethering) on target endothelial cells, achieving low velocity “rolling” adhesive interactions (Step 1, Figure 2). Rolling exposes these cells to chemokines that are immobilized by glycosaminoglycans on the endothelial surface, and, in turn, facilitates engagement of G-protein-coupled chemokine receptors (GPCRs) expressed on the mononuclear phagocyte cell surface (Step 2, Figure 2), with resultant G-protein-driven integrin activation (3, 57). Activated integrins on phagocytes, principally very late activation protein 4 and LFA-1, bind to their respective endothelial receptors vascular cell adhesion molecule-1 and intercellular adhesion molecule-1 (ICAM-1), leading to firm adhesion of MPS cells on the endothelium (Step 3, Figure 2) (3, 57). The binding of activated integrins then allows diapedesis into the tissue (Step 4, Figure 2). Two distinct mechanisms enable diapedesis: (1) transient dismantling of endothelial junctions (paracellular migration) or (2) migration through individual endothelial cells (transcellular migration) (58, 59).

Although several cell-associated proteins are specialized at mediating the first step of cell migration, the selectins and their ligands are the most potent effectors of tethering and rolling adhesive interactions. These molecules are responsible for the initial low-affinity binding interactions of leukocytes on endothelial layer (60), a property related to the unique biophysics of lectin–carbohydrate interactions under fluid shear conditions.

The selectins are a family of three carbohydrate-binding proteins that can be expressed on endothelial cells, leukocytes and platelets (Figure 3). Due to their requirement of calcium ions for binding, all three selectins, E-selectin (CD62E), P-selectin (CD62P), and L-selectin (CD62L), belong to the C-type lectin family (61). Selectins share a common structure of five different domains: an N-terminal carbohydrate recognition domain (CRD), an epidermal growth factor-like domain (EGF), a varying number of short consensus repeats that have homology to complement regulatory domains (“CRs” of which there are 2, 6, and 9 within L-, E-, and P-selectin, respectively), a transmembrane region, and a C-terminal cytoplasmatic domain (Figure 3) (62–64). While the CRD and EGF domains are highly homologous between the three selectins, the structure of the transmembrane and cytoplasmic portions, as well as the extracellular CR domains are not conserved across the selectins, resulting in structural diversity and varying molecular weights between selectins (61, 65).

Despite sharing common elements, the three selectins have different functions in diverse pathological and physiological processes and vary in their distribution and binding kinetics.

The biology of L-selectin was first elucidated by use of an in vitro assay in which suspensions of lymphocytes were overlaid onto lymph node sections (66). This assay then allowed for creation of mAb that could interrupt this binding, such as the mAb known as “MEL-14” described by Gallatin and coworkers (67) in 1983, and thereafter led investigators to cloning of this structure (62). L-selectin is highly expressed on hematopoietic stem cells and mature leukocytes, including all myeloid cells, subsets of natural killer cells, naïve T and B cells, and central memory T cells. When leukocytes are activated, cell surface levels of L-selectin are downregulated by proteolytic cleavage via metalloprotease-dependent shedding of the extracellular domain (61, 68, 69).

P-selectin was described in 1984 by McEver and coworkers (70, 71) and Furie and coworkers (72) as a glycoprotein expressed on the cell surface of activated platelets. P-selectin is constitutively expressed by circulating platelets and endothelial cells, where it is stored in α-granules and Weibel–Palade bodies, respectively. Because it can be expressed on endothelial cells, P-selectin together with E-selectin (described below) are known as the “vascular selectins.” Following pro-inflammatory stimulus by molecules such as thrombin or histamine, P-selectin is rapidly translocated from the granules to the cell surface by fusion of intracellular storage compartments with the plasma membrane. In murine endothelial cells, inflammatory mediators, such as TNF-α, IL-1β, and LPS, induce P-selectin mRNA transcription, which requires the cooperative binding of the nuclear factor κ-light chain-enhancer of activated B cells (NF-κB) and activating transcription factor-2 (ATF-2) to their response elements within the P-selectin promoter (73–75). However, importantly, the promoter of P-selectin in humans and other primates lacks binding sites for NF-κB and ATF-2 (76). For this reason, in human endothelial cells, the only vascular selectin inducibly expressed by TNF-α, LPS, and IL-1β is E-selectin (77).

E-selectin was first reported by Bevilacqua and coworkers (63, 78) in 1980s as a leukocyte adhesion molecule on activated endothelial cells. Skin and bone marrow microvessels express E-selectin constitutively (79), however, in other tissues, endothelial cells do not constitutively express E-selectin but its expression is strongly upregulated by inflammatory cytokines, such as TNF-α and IL-1β. These cytokines potently induce transient transcription (within hours of exposure) of E-selectin mRNA in both human and mouse endothelial cells (80). Cytokine-dependent activation of E-selectin is mediated by NF-κB binding to regulatory domains in the E-selectin promoter (81). Functionally, E-selectin slows leukocyte rolling to much lower velocities than do either L- or P-selectin, favoring subsequent leukocyte arrest (4, 82). This capacity, along with the inability of human endothelial cells to upregulate P-selectin in the presence of IL-1β and TNF-α, is why E-selectin is considered to be the most important selectin for cell trafficking to sites of inflammation in humans, and it plays a critical role in the recruitment of immune effectors to target inflammatory sites.

E-selectin recognizes a range of structurally diverse glycan epitopes expressed by human leukocytes that typically contain α(1,3)-fucose (Fuc) and α(2,3)-sialic acid (Sia) modification(s) on a lactosamine backbone [consisting of galactose (Gal) linked to N-acetylglucosamine (GlcNAc)], as shown in Figure 4 (79). The terminal tetrasaccharide known as sialyl Lewis X (sLe^x—Siaα2-3Galβ1-4(Fucα1-3)GlcNAc) is the prototypical E-selectin-binding determinant (83–85). Some sLe^x-variant structures can also exhibit E-selectin binding activity, namely an internally fucosylated sLe^x-variant (VIM-2) and other polylactosamine structures, in which Fuc modifications occur at more than one GlcNAc residue along the polylactosamine chain (tri-fucosyl-sialyl Lewis^x and di-fucosyl-sialyl Lewis^x) (86–88). In addition, other glycan structures that are not natively expressed on leukocytes exhibit E-selectin binding activity, namely the sLe^x isomer, sialyl Lewis a (sLe^a—Siaα2-3Galβ1-3(Fucα1-4)GlcNAc) (89), some sulfated derivatives of Le^x and Le^a (3′-sulfo-Le^x and 3′-sulfo-Le^a, respectively) (90, 91), and a fucosylated glycoform of LacdiNac that displays a terminal N-acetylgalactosamine (GalNAc) instead of Sia (GalNAc-Lewis x) (92).

E-selectin binding determinants are typically displayed at the end of O-glycans, N-glycans, or glycolipid precursor structures, and require the coordinated and sequential action of specific glycosyltransferases localized within the lumen of the Golgi apparatus. Assembly of sLe^x is driven by the terminal addition of Sia (to Gal) and of Fuc (to GlcNAc) through the action of α(2,3)-sialyltransferases and α(1,3)-fucosyltransferases (FTs), respectively, on type 2 lactosamine (LacNAc) chains (i.e., Gal connected to GlcNAc through a β(1,4)-linkage) (Figure 4) (93, 94). The sialylated forms of Lewis antigens are synthesized by the action of the α(2,3)-sialyltransferases (ST3Gal isoenzymes). These enzymes transfer Sia residues to the Gal on the LacNAc chain, exclusively acting prior to fucosylation (95, 96). There are six members of the α(2,3)-sialyltransferase family (ST3Gal-I–ST3Gal-VI), but only ST3Gal-III, ST3Gal-IV, and ST3Gal-VI are reported to sialylate lactosamine chains (97). Importantly, ST3Gal-III exhibits preference for type 1 lactosamine chain acceptors (wherein Gal is connected to GlcNAc through a β(1,3)-linkage), whereas ST3Gal-IV and ST3Gal-VI preferentially act on type 2 polylactosamine chains (98–100). When Type 1 lactosamines are decorated with Sia in α(2,3)-linkage to Gal and with Fuc in β(1,3)-linkage to GlcNAc, this tetrasaccharide is known as sialyl Lewis A (sLe^A).

So far, six human FTs have been found to catalyze the addition of Fuc at α(1,3) linkage to GlcNAc with a type 2 lactosamine—FTIII, FTIV, FTV, FTVI, FTVII, and FTIX. Each enzyme exhibits specificity for acceptor substrates and, therefore, has the ability to generate distinct fucosylated structures (101, 102). Particularly, FTIII and FTV are unique in that they exhibit both α(1,3) and α(1,4) FT activity on both sialylated and unsialylated type 2 and type 1 lactosamines thereby creating (s)Le^x and (s)Le^a epitopes, respectively (103, 104). On the other hand, FTIV and FTVI fucosylate both sialylated and unsialylated type 2 lactosamine chains, with FTIV creating VIM-2 and Le^x (105, 106) and, modestly, sLe^x determinants (107, 108), and FTVI creating these structures as well as di-fucosyl-sLe^x (108–110). Uniquely, FTVII can only act on sialylated type 2 lactosamines, yielding sLe^x and di/tri-fucosyl-sLe^x-structures (111, 112), whereas FTIX is known to synthesize mostly Le^x (101, 106).

Most of the reports that assess the role of the different glycosyltransferases involved in selectin ligand biosynthesis in leukocytes have been performed using knock-out mouse models, with a small proportion of these studies using human leukocytes or human hematopoietic cell lines. Concerning the role of the α(2,3)-sialyltransferases, murine studies suggest that ST3Gal-IV and ST3Gal-VI collaborate together in murine E-selectin ligand biosynthesis, with ST3Gal-IV having an important role in the regulation of E-selectin-dependent rolling velocity (113, 114). Interestingly, ST3Gal-III does not seem to contribute to the synthesis of murine E-selectin ligand moieties, since deficiency of this enzyme did not affect E-selectin ligand expression or activity on murine leukocytes (113). Surprisingly, ST3Gal-IV is reportedly the only human α(2,3)-sialyltransferase involved in the biosynthesis of E-selectin ligands in human myeloid leukocytes, since ST3Gal-IV-silenced HL-60 cells (a human promyelocytic cell line), and neutrophils derived from stable ST3Gal-IV knockdown hematopoietic stem cells fail to engage in tethering and rolling interactions on E-selectin-bearing substrates (115). In case of α(1,3)-FTs, studies demonstrate that mostly FTVII, and to a lesser extent FTIV, are the key murine α(1,3) FTs that mediate leukocyte selectin ligand biosynthesis. In fact, murine leukocytes lacking FTVII show poor adhesive contacts with E- and P-selectin, indicating that this FT plays a prominent role in murine E-selectin ligand biosynthesis (116, 117). However, others reported that FTIV is crucial for slow murine leukocyte rolling velocity (118, 119). Importantly, E-selectin binding activity conferred by murine FTIV, but not by FTVII, apparently occurs mainly on glycolipids rather than glycoproteins (120). Conversely, in human leukocytes, there is evidence that FTVII, FTIV, and FTIX could each act in synthesis of E-selectin ligand determinants (121). Notably, the mouse genome encodes only FTIV, FTVII, and FTIX (122), whereas primates possess an additional three FT gene products—FTIII, FTV, and FTIV. These additional FTs provide for a much wider capacity to create sLe^X; in addition, the expression of FTIII and FTV in primates uniquely drives creation of sLe^A determinants. Human circulating monocytes express all the α(1,3)-FTs, with the exception of FTV, heightening the potential for variability in glycoconjugates bearing sLe^x among human and mouse cells (123). Notably, sLe^A is not expressed on any primate leukocytes as these cells do not synthesize Type 1 lactosamines (3).

Other glycosyltransferases involved in the biosynthesis of sLe^x have also been studied for their relevance in generating functional selectin ligands (Figure 5). Regarding sLe^x presentation on O-glycans, one study reported that leukocytes from mice deficient in the enzyme required for initiation of O-glycosylation, ppGalNAcT-1, showed impaired recruitment during inflammation due to a significant reduction in E- and P-selectin ligand levels (124). Mice lacking the O-glycan core 1 β3galactosyltransferase (C1GalT-I) showed dramatic loss of leukocyte rolling on E-selectin and, consequently, these leukocytes did not transmigrate into inflamed tissues (125). Transgenic mouse studies, where the O-glycan core 2 β6-N-acetylglucosaminyltransferase-I (C2GnT-I) was knocked out, also showed reduced E-selectin and P-selectin binding activity of leukocytes under static and shear-based rolling assays, with impaired leukocyte recruitment to sites of inflammation (126–128). In agreement, in human moDCs, the downregulation of C2GnT-I, with concurrent upregulation of ST3Gal-I and GalNAc α(2,6)sialyltransferase (ST6GalNAc)-II, results in a loss of the core 2 structures required for O-glycan display of sLe^x (Figure 5) (129). Furthermore, studies using HL-60 cells have revealed that the ST6GalNAc-II overexpression abrogates sLe^x cell surface display and reduces the number of adherent cells to E-selectin under flow conditions, reinforcing the notion that there exists a competition between ST6GalNAc-II and C2GnT-I for core 1 acceptors, affecting the biosynthesis of sLe^x-bearing core 2-O-glycan structures (Figure 5) (130). Interestingly, in mice, knockout of one of the β(1,4)galactosyltransferases (of the family of five isoenzymes) involved in Type 2 lactosamine synthesis, β(1,4)galactosyltransferase-I (β(1,4)GalT-I), showed reduced inflammatory responses and impaired P-selectin binding activity; however, the contribution of this enzyme in the synthesis of E-selectin counter-receptors remains to be elucidated (131).

One study has recently evaluated the contributions of N-glycans, O-glycans and glycosphingolipids (GSLs) to E-selectin binding by human myeloid cells under physiological flow conditions (132). To address this issue, O-glycan and GSL synthesis was abolished by, respectively, knocking-out the core 1 Gal transferase chaperone, i.e., the C1GalT-I-specific Molecular Chaperone (COSMC), β1,2 GlcNAc-transferase (MGAT-I), and UDP-glucose ceramide glucosyltransferase (UGCG). Notably, these studies indicate that while O-glycans are indispensable for myeloid cell binding to L- and P-selectins, N-glycans play the major role in the initial myeloid cell recruitment into E-selectin-bearing substrates, with O-glycans playing a more modest role. In addition, both glycolipids and N-glycans are responsible for the slowing down of rolling velocities that precede firm arrest (132).

Most studies that have assessed biologic modulators of E-selectin ligands in leukocytes have been performed using human and murine T cells. An array of cytokines has been shown to regulate E-selectin ligand expression via upregulation or downregulation of specific glycosyltrasferases that control sLe^x expression. Specifically, IL-2, IL-7, IL-15, and IL-12 increase the expression of glycosyltransferases involved in the biosynthesis of E-selectin ligand determinants, whereas IL-4 has the opposite effect (133). This selectin ligand upregulation in T cells in response to cytokine signaling was shown to be dependent on Th1 transcription factor T-bet (134) and on STAT4-mediated pathways (135). Interestingly, human myeloid cells treated with granulocyte-colony stimulating factor (G-CSF) show increased cell surface expression of E-selectin ligands associated with significant increases in gene expression of the glycosyltransferases ST3Gal-IV, FTIV, and FTVII (136).

The cutaneous lymphocyte antigen (CLA) is the E-selectin-reactive glycoform of PSGL-1. PSGL-1 is a transmembrane 240-kDa homodimeric, mucin-like glycoprotein expressed on leukocytes (and, reportedly, on some activated endothelial cells) that plays a crucial role in the homing of leukocytes into inflamed tissue (156, 157). E-selectin binding activity of PSGL-1 is conferred by sialylated and fucosylated core 2-based-O-glycans that are cluster-distributed along the stalk region of the PSGL-1 extracellular domain (158). A number of studies have identified PSGL-1 as one of the several scaffolds expressed by human classical monocytes displaying E-selectin-binding activity (123, 155). On the other hand, PSGL-1 is the only known scaffold that presents sLe^x determinants on human circulating cDCs (151, 155) and moDCs (129). Yet, Silva et al. observed that although PSGL-1 is essential for P- and L-selectin recognition by human moDCs under fluid shear conditions, it is not mandatory for tethering to E-selectin (153). Moreover, there are reports that extravasation of murine immature DC to inflamed tissues requires both E- and P-selectin, but not PSGL-1 (159). Together, these data suggest the expression of ligands for E-selectin in addition to CLA by human and murine DCs. Still, PSGL-1 is the dominant ligand for P- and L- selectin and is the only known glycoprotein that binds all three selectins (160, 161). Accordingly, circulating monocytes that have already bound to E-selectin on inflamed endothelium can also interact with L-selectin expressed by other circulating monocytes/DCs via PSGL-1 and support their secondary capture, potentiating mononuclear phagocyte recruitment to sites of inflammation (162).

HCELL is a sialofucosylated glycoform of CD44 that exhibits potent E-selectin (and L-selectin) binding activity. CD44 is a transmembrane protein that exists in a wide variety of protein isoforms due to alternative splicing and extensive post-translational modifications (with molecular weight ranging from 80 to 220 kDa). CD44 is expressed by most mammalian cells, where it serves as the principal receptor for hyaloronic acid and participates in a broad range of cellular activities, including lymphocyte activation, leukocyte trafficking, hematopoiesis, cell growth and survival, and tumor dissemination (163). Post-translational modifications along with extensive alternative splicing allow the formation of multiple protein isoforms, expressed in a tissue-specific manner (164, 165). The standard protein isoform of CD44 (CD44s or CD44H) is encoded by mRNA transcripts comprising exons 1–5 and 16–20 (“s1–s5 and s6–s10”). CD44s is ubiquitously expressed by mammalian cells and is the form most often displayed by hematopoietic-lineage cells. In addition to CD44s, non-hematopoietic cells characteristically display CD44 variant isoforms contain peptide products of variant exons (exons “v2–v10”) in addition to the standard exon peptide products.

CD44 post-translational modifications include the addition of different glycan structures, namely glycosaminoglycans, and N- and O-glycan substitutions (166). While previous studies indicated that HCELL was only expressed by human hematopoietic stem and progenitor cells (HSPCs) (167, 168), and some hematologic (167, 169) and solid malignancies (170), HCELL was recently reported to be expressed by classical human monocytes (123). Importantly, for human HSPCs, the sLe^x determinant is exclusively displayed on N-glycan lactosamines on CD44s, but classical monocytes express sLe^x on O-glycans of CD44s (123). Because of its ability to engage E- and L-selectin under relatively high fluid shear conditions (i.e., in excess of 20 dynes/cm² shear stress), HCELL is considered the most potent L- and E-selectin ligand expressed on mammalian cells (167, 171).

CD43, also known as sialophorin or leukosialin, is a cell surface glycoprotein expressed by nearly all hematopoietic cells and is involved in several important processes, including cell development, activation, survival, and migration (172–176). Glycosylation of CD43 molecules with either core 1 or core 2 O-glycan structures produces the ~115 and ~135 kDa glycoforms, respectively (177). Recently, we reported that human classical monocytes display E-selectin binding activity on CD43 (123). CD43 that displays sLe^x and binds E-selectin is known as “CD43E” and this protein harbors sLe^x on O-glycans (178).

In contrast to results in mouse leukocytes, reports using human cells have suggested that sialofucosylated determinants displayed on lipid scaffolds can mediate adhesion to E-selectin under static and flow conditions. In native human leukocytes and in the myelocytic leukemia cell line HL-60, the glycolipid structures that display ability to bind to E-selectin consist of sialylated lactosylceramides decorated with internal fucosylated GlcNAc structures (myeloglycans) (86, 87, 179). It has been reported that the disruption of myeloglycan biosynthesis via knockdown of ceramide glucosyltransferase (UGCG) in HL-60 cells disturbed stable cell rolling and impaired transmigration across inflamed endothelial cells (180). Finally, human classical monocytes treated with a broadly active protease did not show a complete abrogation of sLe^x staining, suggesting that although the majority of sialofucosylated moieties required for E-selectin binding are preferentially expressed on proteins rather than lipids, a considerable amount of sialofucosylated determinants are present on glycolipids (123).