Review Selectin Ligands

Home , CD22, Selectin

Proc. Nati. Acad. Sci. USA Vol. 91, pp. 7390-7397, August 1994 Review Selectin ligands Ajit Varki Gfro bg Program, Cancer Cente, and Division of Cellular and Molecular Medicine, Univert of California, San Diego, La Jolla, CA 92093

ABSTRACT The selectins initiate many critical interactions among blood cells. The volume of information and di- versity of opinions on the nature of the biologically relevant ligands for selectins is remarkable. This review analyzes the mat- ter and suggests the hypothesis that at least some of the specificity may involve recognition of "clustered saccharide patches." Five years ago, the cloning of three vascular proteins resulted in identification of the selectin family of cell adhesion molecules (1-9). A shared N-terminal carbohydrate-recognition domain homologous | L-SELECTIN yE-SELECTIN P-SELECTIN to other Ca2+-dependent (C-type) mammalian lectins (10) strengthened predic- FIG. 1. Location and topology of the selectin family of cell adhesion molecules and their tions from functional studies (for review, carbohydrate ligands. Expression of the selectins and their ligands may be constitutive or see ref. 1) that the cognate ligands for induced, depending upon the cell type, the tissue, and the biological circumstance. The selectins these receptors would be cell-surface can also be shed from cell surfaces by proteolysis. carbohydrates (Fig. 1). Several groups then reported that previously known lectin came from the inhibitory effects of (15, 18, 55-57). Taken together, these mammalian oligosaccharides bearing Fuc phosphorylated monosaccharides and a data indicated that specific carbohydrate and sialic acid (Sia) are recognized in a phosphorylated yeast mannan on the in- ligands are recognized by selecting. Ca2+-dependent manner by the selectins teraction of lymphocytes with high- Many Diverse Carbohydrates Are Rec- (11-24). Evidence for the function of se- endothelial venules (HEV) in lymph ognized by Selectins. A plethora ofsimple lectins in leukocyte trafficking, thrombo- nodes (1). These studies were inspired by and complex carbohydrates have been sis and inflammation, as well as the pos- work on a different class of mammalian reported as recognized by the selecting, sibility for therapeutic intervention in lectins, the mannose 6-phosphate recep- on the basis of either direct binding ex- reperfusion injury, inflammation, al- tors (52). However, sialidase treatment periments or inhibitory properties (see lergy, autoimmunity, and cancer (for re- of lymph node sections abolished lym- Table 1). Most, but not all, carry sialy- view, see refs. 1-9) attracted additional phocyte-HEV interaction, indicating a lated, sulfated, and/or fucosylated se- investigators (>500 articles in the last 5 critical role for Sia (53). Subsequently, quences normally found at the nonreduc- yr).* More recently, specific macromo- the selective inhibitory properties of var- ing termini of N-linked or O-linked oli- lecular ligands for the selectins have been ious phosphorylated and sulfated saccha- gosaccharides, or on glycosphingolipids. reported (25-38). Molecular cloning of rides (1) confirmed a lectin-like interac- Such sequences are presented in Table 2 the polypeptide backbones of some of tion. It is now evident that the selectin and contrasted with related sequences these molecules indicates that ligand for- ligands are not themselves phosphory- recognized by two other mammalian mation is indeed glycosylation-depen- lated (nor always sulfated) and that the blood cell adhesion molecules, CD22 and dent (28, 33, 34, 37, 39, 40). Some studies inhibitors worked either because of high sialoadhesin. The common feature of suggest that small sialylated, fucosylated most ofthose selectins is a oligosaccharides such as sialyl-Lewisx concentrations or high charge density. In recognized by (SLex) and sialyl-Lewisa (SLea) (11, 12, parallel, others (3) described monoclonal lactosamine backbone of either type 1 20, 22-24, 41-44) and/or their sulfated antibodies (mAbs) that recognized spe- (Galp1-3GlcNAc) or type 2 (Gal,81- equivalents (45-47) are the ligands, cific "addressins" involved in lympho- 4GlcNAc). When present, Sia is usually whereas others indicate that complex- cyte trafficking. The peripheral node ad- in a2-3 linkage, Fuc is in either al-3 or clustered saccharide motifs on specific dressin recognized by mAb MECA-79 al-4 linkage, and the location of the Fuc glycoconjugates are required for biolog- (54) is a family of glycoproteins (gps), residue relative to the Sia residue (deter- ically relevant recognition (28, 31). Some some ofwhich interact with L-selectin in mined by fucosyltransferase and cell have even reported selectin ligands that a Sia-dependent manner (25). Likewise, type) is important (43, 58-60) (see Table lack Sia or Fuc (36, 43, 48-51). Because the mAb HECA-452 recognizes Sia- 2). Conformational analyses suggest that of recent excellent reviews on other as- dependent epitopes of a proposed cuta- a specific face of the SLeX tetrasaccha- pects of selectin biology (3-9), this dis- neous lymphocyte antigen, which medi- cussion will focus exclusively upon the ates recognition of skin-seeking lympho- Abbreviations: SLex, sialyl-Lewisx; SLea, sialyl-Lewisa; HEV, high-endothelial venule(s); nature of the selectin ligands. cytes by E-selectin (27). Others defined mAbs, monoclonal antibodies; Sia, sialic acid; Selectin Ligands: Historical Back- the ligands operationally, showing loss of gp, glycoprotein. ground. Early evidence for Ca2+-depen- selectin-mediated binding upon sialidase *Because of space constraints the bibliogra- dent carbohydrate recognition by L-se- or protease treatment of the target cells phy presented is necessarily incomplete. 7390 Downloaded by guest on September 29, 2021 Review: Varki Proc. Natl. Acad. Sci. USA 91 (1994) 7391

ride involved in recognition is shared in an in vitro assay and what does actually to reveal biologically relevant binding of with the isomer SLea (22, 24, 61). Signif- bind in vivo. specific molecules. Likewise, cell inter- icant differences in binding can be gen- Importance ofAssay Conditions Used to action studies incorporating biologically erated by modifications ofthe GlcNAc or Identify Potential Ligands. In vitro, cer- relevant shear forces (e.g., in flow cham- Sia N-acetyl groups (24, 62) or specific tain assay conditions can show selectin bers or on rocker platforms) are more OH groups on the Fuc or Gal residues binding that may not be biologically rel- likely to reveal specific binding. (63). In contrast, truncation of the Sia evant. For example, the use of TLC to When Is Recognition by Selectins Bio- side chain has little effect upon recogni- detect potential ligands has the advan- logically Relevant? Selectin ligands must tion (24, 64). tage that natural or artificial glycolipids have both the opportunity (be located at Such data suggest a high specificity for can be separated into distinct molecular the right place at the right time), as well the interaction with tetrasaccharides species, all of which can be probed si- as adequate affinity or avidity (bind such as SLeX. However, sulfate esters or multaneously, with a visual readout. strongly enough to mediate specific bio- uronic acids can substitute for Sia resi- However, apotential disadvantage is that logical recognition under natural circum- dues, and some neutral fucosylated the oligosaccharides might be presented stances). The selectins may not normally chains can be recognized, at high densi- in an unnaturally high density, giving have the opportunity in vivo to recognize ties or under specific experimental con- avid binding that might not occur if the all structures listed in Table 1. Thus, ditions (13, 20, 43, 65). Also, some se- same molecules were presented as minor some candidate oligosaccharides (e.g., quences that lack Fuc residues are rec- components of a natural lipid bilayer. SLea) have not been found on any of the ognized (see Table 2), and P-selectin may Likewise, the use of highly multivalent relevant vascular cells (although type 1 recognize Sia residues in a2-6 linkage arrays of either selectins (e.g., as immu- chains could generate pathological li- (63, 66). Thus, there now appears to be a nocomplexes) or of potential oligosac- gands in carcinomas). Regarding binding wide and diverse range of candidate oli- charide ligands (e.g., as neoglycopro- strength, most studies of monomeric li- gosaccharide ligands. However, selec- teins) can give a positive result that may gands (Table 1) report IC50 values (rough tive inhibition ofN- or O-linked glycosy- not be relevant to the natural situation. indicators of binding constants) in the lation (38, 66-68) suggests that the bio- The same is true ofthe use oftransfected PM-mM range (20, 51, 62, 63, 69, 70). logically relevant ligands may not be cell lines expressing very high densities One reported exception (71) might be carried on all glycoconjugate classes. A of selectins. Also, static cell adhesion explained by a 60g force used to induce further difficulty is that if all three selec- assays with long interaction times can initial cell binding. In contrast, the ability tins recognized the same carbohydrate show in vitro binding that may not be to use stringent techniques such as prob- structure (e.g., SLex), the situation could relevant in the dynamic flow situations ing of blots, flow cytometry analysis, be impractical because selectin and li- encountered in vivo. In contrast, assays precipitation, and affinity chromatogra- gand might appear on the same cell (see that use soluble molecules and/or strin- phy (26, 29-33, 35, 36, 38, 40, 69, 72, 73) Fig. 1). To explain these confusing data, gent washing (e.g., direct binding of mo- indicates that interactions with certain one must carefully consider the differ- nomeric selectins, immunoprecipitation, ligands involve much tighter binding. In- ence between what can bind to a selectin affinity chromatography) are more likely deed, in the few direct-binding studies Table 1. General classes of molecules reported to be recognized by selectins Recognized by Class of molecules Sources E-selectin P-selectin L-selectin Comments Sialylated fucosylated Widespread on N- and + + + + + See Table 2; free lactosamine oligosaccharides O-linked oligosaccharides, oligosaccharides inhibit and on glycolipids at high concentrations Sulfated fucosylated Tumor mucins, some + ++ ++ See Table 2; free lactosamine oligosaccharides normal mucins oligosaccharides inhibit at high concentrations 3-Sulfated glucuronosyl lipids Brain glycolipids; some - + + + See Table 2; probably gps? low-affinity ligands Sulfatide Various sources - ++ + See Table 2; not Ca2+ dependent? Heparin (commercial) Pig gut or bovine lung - ++/- ++/- Batch-to-batch variation in inhibitory effects? Heparan sulfate proteoglycans Cultured calf and human - ++++ ++++ See Table 3 endothelial cells Chondroitin sulfate Various sources - Weak inhibition at high concentrations Polyphosphomannan Phosphorylated yeast - - + + Also ligand for mannose polysaccharide 6-phosphate receptors Fucoidan Sulfated fucosylated - + + + + Sulfate and Fuc required seaweed polysaccharide for inhibitory action Dextran sulfate Semi-synthetic sulfated - + + + + Sulfate required for polysaccharide inhibitory action Sialylated fucosylated Human and murine + + + + + + See Table 3 mucin-type gps neutrophils, HL-60 cells Sialylated, sulfated, fucosyl- Mouse lymph nodes + + + + + + See Table 3 ated mucin-type gps Sialoglycoproteins, type Human, bovine, and murine ++++ ++++ See Table 3 unspecified leukocytes Recognition (based on binding or inhibitory capacity) by the selectins is indicated on an arbitrary scale of + + + + (best) to none (-). Downloaded by guest on September 29, 2021 7392 Review: Varki Proc. Natl. Acad. Sci. USA 91 (1994) Table 2. Some anionic oligosaccharide sequences recognized by mammalian lectins Recognized by Common name Structure E-selectin P-selectin L-selectin CD22 Sialoadhesin Sialyl Tn Siaa2-6GalNAcal-O-Ser/Thr Sialyl T Siaa2-3Gal#1-3GalNAcal-Rl/P1-R2 ++ 6' Sialyl-LN Siaa2-6Gall-4GlcNAcl-R3 - +/- - ++ 3' Sialyl-LN Siaa2-3Ga4Il-4GlcNAc#1-R3 - - - - SLex Siaa2-3Galf81-4(Fucal-3)GlcNAcl-R3 + + + + + VIM-2 (CD65) Siaa2-3Gal81-4GlcNAc,81-3Galpl-4(Fucal-3)GlcNAc81-R3 +/- + ?+ Sialyl dimeric Lewisx Siaa2-3Gal,81-4(Fucal-3)GlcNAc,81-3Gal1-4(Fucal-3)- + + ++ ?+ GlcNAc,81-R3 SLea Siaa2-3Gal81-3(Fucal4)GlcNAc,81-R3 + + + + + 3' Sulfo-Lewisx SO4-3Gal81-4(Fucal-3)GlcNAc,81-R3 + ++ ++ 3' Sulfo-Lewisa SO4-3Galf81-3(Fucal4)GlcNAc,81-R3 + ++ ++ Sulfatide SO4-3Gal,91-'1-ceramide - + ++ 3' Sulfo-LN S04-3Gal#1-4GlcNAc,81-R3 SGNL (HNK-1) S04-3GlcA81-4GalIl-4GlcNAc#l-R3 + GNL GlcA,1-4Gal,14GlcNAc31-R3 + Recognition by the individual selectins is indicated on an arbitrary scale of + + (best) to none (-). R1, O-linked oligosaccharide; R2, ceramide-linked oligosaccharide; R3, O-linked, N-linked, or ceramide-linked oligosaccharide; GicA, glucuronic acid. with soluble selectins, affinities in the explanation is that the nonhuman oligo- generated by a polypeptide backbone low nanomolar range are reported (18, saccharide ligands are masked from mAb (Fig. 2A). This has indeed been suggested 48, 73). Although the data set is incom- recognition by Sia modifications or by for some high-affinity selectin ligands plete, it seems to predict that biologically other unknown subtleties of structure that are mucins (28, 31, 33, 34) which relevant recognition involves higher or- (86, 87). However, these experiments have many oligosaccharides, closely der structures that can generate a suffi- were done with an extensive mAb panel spaced to one another. However, the cient affinity (or avidity) to function in of overlapping specificities (82). A better IC50 of chemically synthesized dimeric vivo. explanation is that during evolution the SLex for E-selectin showed only a 5-fold Sialylated Fucosylated Lactosamines binding pocket ofa selectin (derived from improvement over the monomeric form Are Necessary Components of Selectin Li- amino acids) is less likely to have di- (70). Further, P- and L-selectin do not gands in Hmns. In a human genetic verged than the ligands which (based on bind to some cell types that express con- disorder called leukocyte adhesion defi- oligosaccharide sequences) are under in- siderable amounts of SLex, and myeloid ciency type II (74), a general failure of dependent selective pressures. Since cell recognition by P-selectin is de- fucosylation reduces expression of all sugar chains are more flexible than poly- stroyed by the enzyme O-sialoglycopro- fucosylated lactosamines and is associ- peptides, they may also be more capable tease (31, 73, 89), even though the vast ated with markedly diminished selectin- of accommodating structural changes, majority of surface SLex remains intact mediated binding both in vitro and in vivo while still preserving critical three- (31). Thus, simple ligand multivalency (74, 75). While the patients have other dimensional features for recognition. seems insufficient to explain biologically unrelated abnormalities probably due to Thus, the conserved binding pockets of relevant selectin binding. Another possi- the Fuc deficiency, this experiment of the selectins could cross-react with dif- bility is multivalent aggregation of the nature indicates that fucosylated lac- ferent carbohydrate ligands in different selectin (Fig. 2B). However, high-affinity tosamines are critical components of species, each of which can generate the binding of soluble monomeric P-selectin many biologically relevant selectin li- same three-dimensional recognition to cell surfaces (73) indicates that this is gands in humans. However, the clinical structure. These observations indicate not essential. Also, there is no published abnormality is not complete, and some that the biologically relevant selectin li- evidence for naturally occurring multi- leukocyte trafficking can still occur in gands are unlikely to be linear, well- meric selectins. Ligand multivalency these patients, possibly due to alternate defined oligosaccharide sequences and provided by cell surfaces or specific poly- ligands that lack Fuc (36, 49-51). may be complex structures that can be peptides could also be recognized by Lack of Conservation of Selectin Li- generated in more ways than one. multiple binding sites within a single se- gands Among Species. Recognition of li- Mechanisms by Which Monovalent Oi- lectin lectin domain (Fig. 2C). However, gands by selectins is preserved across a accharide Chains Could Generate Se- this is negated by studies of the lectin variety of mammalian species (76, 77). lectin Linds With Enhnc Affinity or domains of the selectins by epitope map- Thus, L-selectin from one species can Avidity. Studies ofmany plant and animal ping, mutagenesis, and homology criteria recognize lymph node HEV ligands in lectins (10, 88) indicate that monovalent (90-93), which indicate binding sites for other species (36, 64, 76, 78), and cells carbohydrate ligands usually do not have carbohydrates that are quite small; this from one species can show selectin- high affinity. Similarly, most reported has been recently confirmed for E-selec- dependent interactions within the vascu- IC5o values for monovalent selectin li- tin by x-ray crystallography (94). Fig. 3 lature of others (75, 79, 80). However, gands are rather high (20, 51, 62, 63, 69, shows that there is barely enough room in studies using mAbs indicate that sialy- 70). There are several ways in which the E-selectin lectin-binding domain to lated fucosylated lactosamines such as monovalent oligosaccharide chains could accommodate one copy of a SLex tet- SLex are not expressed on the neutro- generate the enhanced binding seen with rasaccharide, let alone many closely phils of many mammals, including some biologically relevant selectin ligands spaced units. Thus, simple ligand or se- rodents and nonhuman primates (81, 82). (Fig. 2). Simple multivalency of both lectin multivalency does not appear suf- Despite this, in vivo blocking studies with oligosaccharide and selectin by presen- ficient to explain high-affinity recogni- SLex-based oligosaccharides (83-85) in tation on intact cell surfaces could en- tion. Anotherpossibility is a combination rats indicate that their endogenous selec- hance avidity (Fig. 2A). Oligosaccharide of both a basic oligosaccharide sequence tins can recognize these ligands. One multivalency could be more efficiently and an adjacent peptide-recognition site Downloaded by guest on September 29, 2021 Review: Varki Proc. Nadl. Acad. Sci. USA 91 (1994) 7393

(Fig. 2D). Some studies involving muta- first case (Fig. 2 F), a common oligosac- would be distinct from improved binding genesis, domain swapping, or deletion of charide with low affinity for the selectin of certain lectins to multiantennary N- the other protein domains of selectins (e.g., SLe') mf ght be forced by other linked oligosaccharides or to glycopep- seem to support this (95-97). However, surrounding sugawr chains into an unusual tides with widely spaced O-linked oligo- the persistence of high-affinity recogni- conformation favored fQr high-affinity saccharides (88). This results from cor- tion after extensive denaturation or pro- recognition. Thihs could also be aided or rect spacing of terminal sugar residues, teolytic digestion of some biologically stabilized by pe,ptide interactions with each of which are still recognized as relevant selectin ligands (28, 29, 31, 36, the adjacent sug,ar chains. The adjacent distinct components of individual chains. 69) is against this. These observations sugars might also induce a favorable con- The hypothesis also predicts that struc- also make it unlikely that the peptide formational chamige in the selectin lectin tural analyses of O-linked oligosaccha- carrier forces a monovalent oligosaccha- domain itself. II the second case (Fig. rides released from a specific selectin ride into an unusual conformation fa- 2G), the binding site would be generated ligand might not uncover unique se- vored for recognition (Fig. 2E). By ex- by combinationis of side chains and quences required for recognition. There- clusion, the remaining possibility is that groups from mulltiple oligosaccharides- fore, it may be necessary to completely high-affinity binding involves recognition i.e., a "discontLiguous saccharide epi- elucidate both the polypeptide chain se- of a "clustered saccharide patch" (31) tope." Thus, the patch recognized by the quence and the oligosaccharide struc- created by a peptide carrying multiple selectin might ilnclude, for example, a tures, as well as their specific sites of oligosaccharides (Fig. 2 F and G). hydroxyl group fFrom one sugar chain, an attachment on the protein. Ultimately, Free oligosaccharides in solution or at acetyl group froim another sugar chain, one must reconstruct the ligand and de- single attachment sites have significant and a carboxyl ior sulfate group from a fine the position of each of its compo- freedom of mobility and show limited third chain. In bBoth cases, the polypep- nents in three-dimensional space. interactions with associated proteins tide need not tbe part of the selectin There is, in fact, much prior evidence (98). However, oligosaccharides packed recognition site. However, it would be for recognition of "clustered saccharide closely together (particularly in mucins) crucial for preseinting the sugar chains in patches." The enzyme O-sialoglycopro- can form rigid rod-like structures (99, the correct arr; angement. Thus, high- tease (101), which has been of particular 100) that must have less mobility and yet affinity recognitiion would be lost if the value in identifying high-affinity selectin are relatively unaffected by denaturation arrangement is (disrupted, and the indi- ligands, will only recognize and cleave of the polypeptide backbone. Such clus- vidual sugar chaiins may not have detect- mucins with large numbers of closely tering of common oligosaccharides could able affinity for the selectin. This might spaced 0-linked oligosaccharides (31). present uncommon "clustered saccha- explain why mo: st reported high-affinity Many mAbs that detect specific glyco- ride patches" generated by multiple oli- selectin ligands .are heavily glycosylated proteins have "sialidase-sensitive" or gosaccharides that are closely spaced macromolecules , and why free oligosac- "Sia-dependent" epitopes-e.g., cancer enough to restrict their motion (31). The charides released from such ligands do mucins, CD24, CD45, and CD43 (see ref. patch recognized by a selectin could be not bind with rec ognizable affinity to the 31 for citations). With some heavily gly- generated in two possible ways. In the same selectin (346, 69). Such recognition cosylated O-linked gps, almost all avail- able mAbs require glycosylation for spe- A C cific recognition. Many such proteins B with sialidase-sensitive epitopes are either mucins with clustered 0-linked oligosaccharides or have large numbers of N-linked oligosaccharides. However, when the sialyl-oligosaccharides on such proteins are examined, they are not unusual in structure compared with those from other proteins of the same cells. In a well-documented example, the recognition of a common sialylated oligosaccharide sequence by different mAbs dif- fered depending upon the type of glyco- G conjugate attached (102). Other examples F include the reactivity of a mucin-specific lectin from Sambucus sieboldiana (103), -] recognition of clustered O-linked oligosaccharides on glycophorin A by the in- D vasion receptor of Plasmodium falci- .I

...... parum (104), the specific binding of ro- taviruses to intact mucins (105), and recognition of the sperm receptor on the mucin-type zonapellucidaglycoprotein 3 ofthe mouse egg (106). In such instances, unique clustering of relatively common oligosaccharides can explain the specificity of recognition. * Basic _Polypeptide Hypothesis: A Clustered S ride riz Selectin ol igosaccha ride Patch Could Be Shared by Seiy Dis- ligand prate GlycocouJuptes. Selectins can FIG. 2. Models that could explain improved binding of selectiins to their natural ligands also recognize glycoan chains relative to the basic oligosaccharide ligands they can recognize. As discussed in the text, in a Ca2+-dependent manner (23, 36, 48, possibilities A-E cannot explain many observations concerning selectin binding. F and G 51, 107). Glycosaminoglycans have re- present the concept of "clustered saccharide patches" that could reconcile these data. peating disaccharide structures that do Downloaded by guest on September 29, 2021 7394 Review: Varki Proc. Nati. Acad. Sci. USA 91 (1994) tion of these other ligands. It is also difficult to know whether to include specific glycosphingolipids among the biologically relevant ligands. Molecules such as sulfatide show binding (47, 49) to some selectins, but this is not always Ca2+- dependent, and a large number of unrelated adhesion proteins are known to bind to this small sulfated glycolipid (121). SLex-bearing glycosphingolipids can 1W97 Y -1 97o1 e 9 7 clearly support E-selectin-based binding when purified and immobilized on artificial surfaces. However, whether they can attain similar densities on cell surfaces is unknown. FIG. 3. Size comparison of the E-selectin lectin domain and SLex (courtesy of B. Graves, Nomenclature Difficulties. The nomen- based on ref. 94). Stereo diagram of a "top" view ofthe E-selectin lectin domain, alongside an clature for the low-affinity monomeric SLeX tetrasaccharide. The protein is depicted as a ribbon and tube diagram, with explicit atoms oligosaccharide structures recognized by shown for the side chains ofthe eight amino acid residues known to affect carbohydrate binding selectins is straightforward and is simply and for the calcium ion (with a van der Waals dot surface around it). a description oftheir structures (Table 2). There is also no problem with the nomen- not typically contain Sia or Fuc (108) but matography, or detection by blotting. clature of ligands defined by mAbs have variable numbers of sulfate esters Most, but not all, of these ligands are recognizing multiple and/or unknown and uronic acids. Competition experi- mucin-type gps. Less well-characterized species with unusual carbohydrate epi- ments and Ca2+ dependence suggest that ligands not listed in the table include mul- topes-e.g., peripheral node addressin the binding site is the same or overlap- tiple sulfated gps found in a rat-HEV (MECA-79) (25), cutaneous lymphocyte ping with that for sialylated fucosylated derived cell line Ax (30), 230-kDa and antigen (HECA-452) (27), and SLex-var oligosaccharides. How could two such 130-kDa P-selectin ligands reported on (2F3) (86). These antibodies were impor- disparate structures be recognized by the murine myeloid cells and HL-60 cells (38), tant in understanding selectin-mediated same binding site? One explanation is a sialidase-resistant ligand found on cen- recognition-e.g., MECA-79 reacts with that a "clustered saccharide patch" rec- tral nervous system myelin (110), siali- cells that bear L-selectin ligands (25, 34) ognized by a selectin can be generated by dase-sensitive ligands on activated endo- and with ligands from murine lymph different types ofglycoconjugates. Thus, thelial cells (55-57), and a variety of car- nodes (26, 69). However, they may also the patch generated by a cluster of cinoma-associated ligands (44, 111-113). react with related carbohydrate epitopes O-linked oligosaccharides on a mucin Also not included are various neutrophil on other molecules that are not ligands type glycoprotein might be mimicked by gps bearing SLeX that are recognized by for the selectins. Regarding the discrete a proper arrangement of carboxyl, hy- E- and/or P-selectin in vitro, including biological ligands for the selectins (Table droxyl, and sulfate groups on a heparan LFA-1 (114, 115), nonspecific cross- 3), there are several difficulties with cur- sulfate glycosaminoglycan chain. In the reacting antigens CD66/CD67 (116), rent nomenclature. (i) The different latter case, the chain must generate the LAMP-1 (117), and L-selectin itself (118). names and terms do not relate well to recognition site itself, because release Like most otherneutrophil proteins, these each other. (ii) A polypeptide that acts as from the polypeptide backbone does not polypeptides have SLex-bearing oligosac- a scaffolding for generating a selectin cause loss of binding (36). This hypoth- charides. Thus, when purified in quantity ligand in one tissue may not do so in esis could also explain why relatively and immobilized on artificial surfaces, another tissue (39), and some may have high-affinity binding of selectins is seen they can support cell adhesion via E- completely independent functions (34). when glycosphingolipids are immobilized and/or P-selectin. However, their role in (iii) It appears likely that there are sig- at high density on plates or on TLC plates presenting high-affinity, biologically rele- nificant differences in functionally active (43, 46). Here, the clustered saccharide vant ligands in the natural situation is ligands among different animal species. patch might be artificially recreated on unclear. A particularly confusing case is (iv) Different classes of glycoconjugates the immobile surface because of a high that ofthe L-selectin molecule on neutro- (N-linked gps, mucin-type gps, proteo- density not normally found on cell sur- phils, which has been suggested as a major glycans) appear to function as ligands. faces. However, because glycosphin- presenter of ligands for P- and E-selectin The issue of nomenclature for these li- golipids can coaggregate in patches on (118). Indeed, anti-L-selectin antibodies gands deserves more attention. cell surfaces (109), this could also be a partially block binding in systems where Is Weak Bindig Also Bologlcally Im- natural mechanism for generating spe- E- and P-selectin recognition are impor- portant? This discussion has favored the cific selectin recognition. tant (119, 120). Also, L-selectin isolated notion that macromolecules with com- Selectin Ligands That Are Probably Bi- from neutrophils carries SLex and can be plex oligosaccharide arrangements are ologically Relevant. From the foregoing, recognized by E- and P-selectin when the critical high-affinity ligands for the most, ifnot all ofthe biologically relevant purified and immobilized on a surface selectins. However, in vitro studies show selectin ligands appear to be complex (118). However, neutrophils that have that high-copy numbers of oligosaccha- macromolecules. Table 3 lists such can- shed L-selectin continue to bind well (18), rides such as SLeX on cell surfaces can didate ligands chosen based on expres- as do many cells (e.g., HL-60) that lack give weak, but detectable, binding be- sion in the right cell type and demon- L-selectin. Thus, an alternative explana- tween cells. Is such binding biologically strated ability to bind with high affinity to tion may lie in the unusual distribution of relevant in vivo? For example, the high selectins in a Ca2+- and carbohydrate- L-selectin on the tips of neutrophil level of SLeX on neutrophils could actu- dependent way. The definition of "high pseudopodia (118). Perhaps the true li- ally be responsible for the initial recog- affinity" is somewhat arbitrary and gands for E- and/or P-selectin are also nition of a neutrophil by E- or P-selectin based upon ligand binding surviving ex- located in the same favored site for inter- on an endothelial cell. This could be tensive washing during procedures such actions; if so, anti-L-selectin antibodies necessary, but not sufficient, to initiate as direct binding to cells, affinity chro- may act by locally hindering the recogni- rolling, serving as the first recognition Downloaded by guest on September 29, 2021 Rev iew: Varki Proc. Nail. Acad. Sci. USA 91 (1994) 7395

Table 3. Natural selectin ligands likely to be of biological relevance Recognition Polypeptide Natural requires Original name Source name ligand for Sia Fuc S04 Comments Refs. sgp5O HEV L-selectin Mouse peripheral lymph GlyCAM-1 L-selectin + +? + Primarily secreted-a natural 26, 28 ligand node antiadhesin? Susceptible to O-sialoglycoprotease 120-kDa myeloid Human neutrophil, HL-60 PSGL-1 P-selectin + + Cell surface homodimer-minor 29, 31, 37 P-selectin ligand cells component of total surface gps. Susceptible to O-sialoglycoprotease. Murine homologue? 150-kDa myeloid Mouse bone marrow cells Polypeptide E-selectin + N-linked oligosaccharides required for 32, 38 E-selectin ligand (80%o neutrophils) unknown binding. Resistant to O-sialoglycoprotease. Human homologue? sgp90 sulfated Mouse peripheral lymph CD34 L-selectin + + Polypeptide also present in 33 L-selectin ligand node nonlymphoid endothelium-not a ligand in these locations MAdCAM-1 Mouse mesenteric lymph MAdCAM-1 L-selectin + Polypeptide also present in Peyers 34 node patch HEV-not a major ligand in this location because glycosylation is not appropriate 250-kDa E-selectin Bovine y/8 lymphocytes Polypeptide E-selectin + Isolated by affinity chromatography 35 ligand unknown Heparan sulfate Bovine and human Core L-selectin - + Isolated by affinity chromatography 36 endothelial cells protein from cultured cells-subcellular not location and function in intact required tissues not proven 160-kDa myeloid Mouse bone marrow cells Polypeptide P-selectin + Susceptible to O-sialoglycoprotease 32, 38 P-selectin ligand (80%o neutrophils) unknown N-linked oligosaccharides also required for binding? Homologue in human HL-60 cells? sgp, Sulfated gp; PSGL-1, P-selectin gp ligand 1; MAdCAM, mucosal addressin cell adhesion molecule; GlyCAM-1, glycosylation-dependent cell adhesion molecule 1. step that permits subsequent interactions E-selectin (27, 72, 87, 125, 126). The peutic use. The small SLex-based thera- with the higher-affinity (but lower copy nature of these ligands, their structural peutics currently entering clinical trials number) macromolecular ligands on the relationship to the neutrophil ligands, (J. Paulson, personal communication) same cell. Such recognition could also be and their importance in the in vivo situ- have the potential to partially block the important pathologically in carcinoma ation remain uncertain. If the ligands are functions ofall three selectins. Studies of cells, which have high levels of SLex shared among different leukocytes, the the complex macromolecular ligands and/or SLea on their cell surfaces (44, differential trafficking ofcell types might listed in Table 3 could lead to design of 111-113). be explained by the multistep combina- specific analogs that discriminate be- Are Some "Ligands" Actuafly "Antiad- torial nature of leukocyte extravasation tween selectins and permit selective hesins"? In at least two instances a por- (127). blockade in specific disease states. Ad- tion of the ligand synthesized by cells in Most studies have focused upon the ditional considerations include a better vitro is found in the culture medium-i.e., interaction of a single selectin with a half-life in the circulation and enough glycosylation-dependent cell adhesion single ligand from a particular cell type. affinity to use small amounts for sus- molecule (28) and heparan sulfate proteo- However, the overlap in recognition of tained blockade of selectin function in glycans (36) (K. Norgard-Sumnicht and small oligosaccharides by the selectins vivo. A.V., unpublished work). These soluble (20, 41, 42, 51, 62) predicts that overlaps Future Directions. There are many ligands could act as inhibitors of selectin may also exist with regard to the macro- otherunexplained mysteries and interest- binding rather than as adhesion mole- molecular ligands. While some overlaps ing questions regarding selectin ligands cules. Whether their potential antiadhe- have been seen (38, 66, 73, 91, 128), that deserve investigation. The fact that sive function is biologically relevant re- interactions with a particular ligand are selectin ligands survive treatments such mains to be seen. not comparably strong between the se- as heating, ionic detergents, protease Are the Lgands Shared Between Differ- lectins. Detailed comparisons of the rec- treatment, and organic extraction (26, 28, ent Selectins and Between Different Leu- ognition of specific ligands by the three 29, 31, 36-, 69) suggests that primary rec- kocyte Types? To date, most attention has selectins are needed. ognition involves carbohydrates alone. been directed toward the recognition of Therapeutic Inhibition of Selectins in However, other domains of the selectins high-endothelial venule ligands by L-se- Vivo by Lignd Analogs Problems and could be recognizing other as-yet- lectin on lymphocytes and of neutrophil Prospects. The development of selectin unidentified polypeptides via protein- ligands by endothelial E- and P-selectin. ligand analogs that can interfere in vivo protein interactions. These could be im- However, monocytes and eosinophils has obvious therapeutic potential in a portant in facilitating or modulating some possess ligands for E-selectin, which are variety of pathological processes (3-9). recognition events mediated by the se- functional in interaction with immobi- The use ofsugar analogs or related struc- lectins. The epidermal growth factor-like lized selectins or activated endothelium tures is attractive also because of their and complement-regulatory-like domains in vitro (56, 122-124), and distinct subsets relative lack of immune reactivity, as of the selectins are candidates for such of lymphocytes carry ligands for P- or well as their ease of handling for thera- interactions. The "variant" forms of Downloaded by guest on September 29, 2021 7396 Review: Varki Proc. Natl. Acad. Sci. USA 91 (1994) SLex on blood cell surfaces detected by I acknowledge many investigators for shar- McEver, R. P. (1992) J. CeU Biol. 118, 445- specific mAbs such as HECA-452 (22) ing work in press and thank H. Freeze, M. 456. N. J. and T. Kogan 30. Tamatani, T., Kuida, K., Watanabe, T., and 2F3 (86) remain unexplained. An- Ginsberg, Varki, Paulson, Koike, S. & Miyasaka, M. (1993) J. Immunol. other mystery is the fact that concentra- for helpful comments and for their critical 158, 1735-1745. tions of small oligosaccharides that seem review of the manuscript. 31. Norgard, K. E., Moore, K. L., Diaz, S., Stults, N. L., Ushiyama, S., McEver, R. P., inadequate to block selectin function in 1. Yednock, T. A. & Rosen, S. D. (1989) Adv. Cummings, R. D. & Varki, A. (1993) J. Biol. vitro are surprisingly successful in abro- Immunol. 44, 313-378. Chem. 268, 12764-12774. gating some selectin-mediated patholog- 2. Stoolman, L. M. (1989) Cell 56, 907-910. 32. Levinovitz, A., Mdhlhoff, J., Isenmann, S. & ical reactions in the intact animal (83, 84, 3. Picker, L. J. & Butcher, E. C. (1992) Annu. Vestweber, D. (1993) J. Cel Biol. 121, 449- Rev. Immunol. 10, 561-591. 459. 129). They may function by an alternative 4. Lasky, L. A. (1992) Science 258, 964-969. 33. Baumhueter, S., Singer, M. S., Henzel, W., mechanism in vivo-e.g., by inducing 5. Varki3 A. (1992) Curr. Opin. Cell Biol. 4, Hemmerich, S., Renz, M., Rosen, S. D. & down-regulation of selectin expression 257-266. Lasky, L. A. (1993) Science 262, 436-438. on cell surfaces (130, 131). Regarding the 6. Rosen, S. D. (1993) Semin. Immunol. 5, 237- 34. Berg, E. L., McEvoy, L. M., Berlin, C., Bar- 247. gatze, R. F. & Butcher, E. C. (1993) Nature high-affinity cell-specific macromolecu- 7. Bevilacqua, M. P. & Nelson, R. M. (1993) J. (London) 366, 695-698. lar ligands, it is possible that they have Clin. Invest. 91, 379-387. 35. Walcheck, B., Watts, G. & Jutila, M. A. other biological functions mediated by 8. McEver, R. P. (1994) Curr. Opin. Immunol. 6, (1993) J. Exp. Med. 178, 853-863. their ligation-e.g., the binding of P-se- 75-84. 36. Norgard-Sumnicht, K. E., Varki, N. M. & 9. Lefer, A. M., Weyrich, A. S. & Buerke, M. Varki, A. (1993) Science 261, 480-483. lectin to neutrophils causes biological (1994) Cardiovasc. Res. 28, 289-294. 37. Sako, D., Chang, X.-J., Barone, K. M., responses (132). Another area of confu- 10. Drickamer, K. (1993) Curr. Opin. Struct. Biol. Vachino, G., White, H. M., Shaw, G., Veld- sion is the description of "partially Ca2+- 3, 393-400. man, G. M., Bean, K. M., Ahern, T. J., Fb- 11. Phillips, M. L., Nudelman, E., Gaeta, rie, B., Cumming, D. A. & Larsen, G. R. dependent" ligands, seen particularly in F. C. A., Perez, M., Singhal, A. K., Hako- (1993) Cel 75, 1179-1186. the interactions of P- and L-selectin with mori, S. & Paulson, J. C. (1990) Science 250, 38. Lenter, M., Levinovitz, A., Isenmann, S. & sulfated ligands. Thus, addition ofexcess 1130-1132. Vestweber, D. (1994) J. Cell Biol. 12S, 471- EDTA to chelate Ca2+ sometimes may 12. Walz, G., Aruffo, A., Kolanus, W., Bevilac- 481. qua, M. & Seed, B. (1990) Science 250, 1132- 39. Dowbenko, D., Kikuta, A., Fennie, C., Gil- not completely reverse a selectin-based 1135. lett, N. & Lasky, L. A. (1993) J. Clin. Invest. binding. This may occur despite the fact 13. Larsen, E., Palabrica, T., Sajer, S., Gilbert, 92, 952-960. that the primary binding requires Ca2+. G. E., Wagner, D. D., Furie, B. C. & Funie, 40. Imai, Y., Lasky, L. A. & Rosen, S. D. (1993) B. (1990) CeU 63, 467-474. Nature (London) 361, 555-557. One explanation is that ligand-dependent 14. Lowe, J. B., Stoolman, L. M., Nair, R. P., 41. Berg, E. L., Magnani, J., Warnock, R. A., activation or an "induced-fit" might oc- Larsen, R. D., Berhend, T. L. & Marks, Robinson, M. K. & Butcher, E. C. (1992)Bio- cur, such that Ca:2+ is required for initial R. M. (1990) Cell 63, 475-484. chem. Biophys. Res. Commun. 184, 1048- binding but not for subsequent mainte- 15. Corral, L., Singer, M. S., Macher, B. A. & 1055. Rosen, S. D. (1990) Biochem. Biophys. Res. 42. Foxall, C., Watson, S. R., Dowbenko, D., nance of the complex. Alternatively, a Commun. 172, 1349-1356. Fennie, C., Lasky, L. A., Kiso, M., Hase- Ca2+ could form a very tight complex 16. True, D. D., Singer, M. S., Lasky, L. A. & gawa, A., Asa, D. & Brandley, B. K. (1992) J. with the ligand after initial binding, which Rosen, S. D. (1990) J. Cell Biol. 111, 2757- Cell Biol. 117, 895-902. to remove 2764. 43. Larkin, M., Ahern, T. J., Stoll, M. S., Shaf- might then be very difficult 17. Goelz, S. E., Hession, C., Goff, D., Griffiths, fer, M., Sako, D., O'Brien, J., Yuen, C.-T., without prolonged EDTA exposure. Re- B., Tizard, R., Newman, B., Chi-Rosso, G. & Lawson, A. M., Childs, R. A., Barone, gardless, these and other data (47, 49, 93, Lobb, R. (1990) Cell 63, 1349-1356. K. M., Langer-Safer, P. R., Hasegawa, A., 97) suggest that P- and L-selectin may 18. Moore, K. L., Varki, A. & McEver, R. P. Kiso, M., Larsen, G. R. & Feizi, T. (1992) J. (1991) J. CeU Biol. 112, 491-499. Biol. Chem. 267, 13661-13668. have a sulfate-recognition site whose 19. Tiemeyer, M., Swiedler, S., Ishihara, M., 44. Takada, A., Ohmori, K., Yoneda, T., function is not strictly Ca2+-dependent. Moreland, M., Schweingruber, H., Hirtzer, P. Tsuyuoka, K., Hasegawa, A., Kiso, M. & Finally, most studies on the role of se- & Brandley, B. K. (1991) Proc. Natl. Acad. Kannagi, R. (1993) Cancer Res. 53, 354-361. lectins in neutrophil traffic have focused Sci. USA 88, 1138-1142. 45. Green, P. J., Tamatani, T., Watanabe, T., 20. Polley, M. J., Phillips, M. L., Wayner, E., Miyasaka, M., Hasegawa, A., Kiso, M., upon pathological situations. It is of in- Nudelman, E., Sighl, A. K., Hakomori, S. Yuen, C.-T., Stoll, M. S. & Feizi, T. (1992) terest to know if the selectins are in- & Paulson, J. C. (1991) Proc. Natl. Acad. Sci. Biochem. Biophys. Res. Commun. 18, 244- volved in normal turnover and/or "mar- USA 88, 6224-6228. 251. of neutrophils, which have 21. Zhou, Q., Moore, K. L., Smith, D. F., Varki, 46. Yuen, C.-T., Lawson, A. M., Chai, W., Lar- gination" A., McEver, R. P. & Cummings, R. D. (1991) kin, M., Stoll, M. S., Suart, A. C., Sulivan, very short half-lives in circulation. J. CeU Biol. 115, 557-564. F. X., Ahern, T. J. & Feizi, T. (1992) Bio- Summary and Conclusions. Much evi- 22. Berg, E. L., Robinson, M. K., Mansson, O., chemistry 31, 9126-9131. dence suggests that the biologically rele- Butcher, E. C. & Magnani, J. L. (1991) J. 47. Suzuki, Y., Toda, Y., Tamatani, T., Wata- vant ligands for the selectins are diverse Biol. Chem. 266,14869-14872. nabe, T., Suzuki, T., Nakao, T., Murase, K., 23. Handa, K., Nudelman, E. D., Stroud, M. R., Kiso, M., Hasegawa, A., Tadano-Anitomi, K., and complex macromolecules that share Shiozawa, T. & Hakomori, S. (1991) Biochem. Ishizuka, I. & Miyasaka, M. (1993) Biochem. in common certain types of anionic car- Biophys. Res. Commun. 181, 1223-1230. Biophys. Res. Commun. 19, 426-434. bohydrate structures. While details re- 24. Tyrrell, D., James, P., Rao, N., Foxall, C., 48. Skinner, M. P., Lucas, C. M., Bums, G. F., garding these molecules begin to emerge, Abbas, S., Dasgupta, F., Nashed, M., Hase- Chesterman, C. N. & Berndt, M. C. (1991) J. gawa, A., Kiso, M., Asa, D., Kidd, J. & Biol. Chem. 266, 5371-5374. it remains to be seen which ligands are Brandley, B. K. (1991) Proc. Natl. Acad. Sci. 49. Aruffo, A., Kolanus, W., Walz, G., Fredman, important for specific biological pro- USA 88, 10372-10376. P. & Seed, B. (1991) CeU 67, 35-44. cesses involving the selectins. High- 25. Berg, E. L., Robinson, M. K., Warnock, 50. Needham, L. K. & Schnaar, R. L. (1993) affinity recognition of seemingly dispar- R. A. & Butcher, E. C. (1991) J. Cell Biol. Proc. Natl. Acad. Sci. USA ", 1359-1363. 114, 343-349. 51. Nelson, R. M., Cecconi, O., Roberts, W. G., ate oligosaccharides may involve the for- 26. Imai, Y., Singer, M. S., Fennie, C., Lasky, Aruffo, A., Linhardt, R. J. & Bevilacqua, mation of similar "clustered saccharide L. A. & Rosen, S. D. (1991) J. Cell Biol. 113, M. P. (1993) Blood 82, 3253-3258. patches." If so, fully understanding se- 1213-1222. 52. Kormfeld, S. (1992) Annu. Rev. Biochem. 61, lectin-ligand interactions will require de- 27. Berg, E. L., Yoshino, T., Rott, L., Robinson, 307-330. M., Warnock, R., Kishimoto, T. K., Picker, 53. Rosen, S. D., Singer, M. S., Yednock, T. A. fining the precise three-dimensional L. J. & Butcher, E. C. (1991) J. Exp. Med. & Stoolman, L. M. (1985) Science 228, 1005- structures ofboth partners. Such knowl- 174, 1461-1466. 1007. edge could also permit the design of a 28. Lasky, L. A., Singer, M. S., Dowbenko, D., 54. Streeter, P. R., Rouse, B. T.&Butcher, E. C. more specific and efficacious therapeutic Imai, Y., Henzel, W. J., Grimley, C., Fennie, (1968)J. Cel Biol. 107, 1853-1862. C., Gillett, N., Watson, S. R. & Rosen, S. D. 55. Spertini, O., Luscinskas, F. W., Kansas, ligand analog to be used for intervention (1992) Cell 69, 927-938. G. S., Munro, J. M., Griffin,J. D., Gimbrone, in the many disease processes involving 29. Moore, K. L., Stults, N. L., Diaz, S., Smith, M. A., Jr., &Tedder, T. F. (1991)J. Immunol. selectins. D. F., Cummings, R. D., Varki, A. & 147, 2565-2573. Downloaded by guest on September 29, 2021 Rkeview: Varki Proc. Natl. Acad. Sci. USA 91 (1994) 7397

56. Spertini, O., Luscinskas, F. W., Gimbrone, K.-E. & Butcher, E. C. (1993) Blood 82, 182- 106. Wassarman, P. M. (1988) Annu. Rev. Bio- M. A., Jr., & Tedder, T. F. (1992) J. Exp. 191. chem. 57, 415-442. Med. 175, 1789-1792. 81. Thorpe, S. J. & Feizi, T. (1984)Biosci. Rep. 4, 107. Skinner, M. P., Fournier, D., Andrews, R., 57. Brady, H. R., Spertini, O., Jimenez, W., 673-685. Gorman, J. J., Chesterman, C. N. & Berndt, Brenner, B. M., Marsden, P. A. & Tedder, 82. Ito, K., Handa, K. & Hakamori, S. (1994) M. C. (1989) Biochem. Biophys. Res. Com- T. F. (1992) J. Immunol. 149, 2437-2444. Glycocoijugate J., in press. mun. 164, 1373-1379. 58. Lowe, J. B., Kukowska-Latallo, J., Nair, R., 83. Mulligan, M. S., Paulson, J. C., De Frees, S., 108. Kjelln, L. & Lindahl, U. (1991) Annu. Rev. Larsen, R., Marks, R., Macher, B. A., Kelly, Zheng, Z.-L., Lowe, J. B. & Ward, P. A. Biochem. 60, 443-475. R. J. & Ernst, L. K. (1991)J. Biol. Chem. 26, (1993) Nature (London) 364, 149-151. 109. van Meer, G. (1993) Curr. Opin. Cell Biol. 5, 17467-17477. 84. Mulligan, M. S., Lowe, J. B., Larsen, R. D., 661-673. 59. Yago, K., Zenita, K., Ginya, H., Sawada, M., Paulson, J., Zheng, Z., DeFrees, S., Mae- 110. Huang, K., Geoffroy, J. S., Singer, M. S. & Ohmori, K., Okuma, M., Kannagi, R. & mura, K., Fukuda, M. & Ward, P. A. (1993)J. Rosen, S. D. (1991) J. Clin. Invest. 88, 1778- Lowe, J. B. (1993) CancerRes. 53,5559-5565. Exp. Med. 178, 623-631. 1783. 85. Asako, H., Kurose, I., Wolfe, R., DeFrees, 111. Majuri, M.-L., Mattila, P. & Renkonen, R. 60. Goelz, S., Kumar, R., Potvin, B., Sundaram, (1992) Biochem. Biophys. Res. Commun. 182, S., Brickelmaier, M. & Stanley, P. (1994) J. S., Zheng, Z.-L., Phillips, M. L., Paulson, 1376-1382. Biol. Chem. 269, 1033-1040. J. C. & Granger, D. N. (1994) J. Clin. Invest. 112. Nakamori, S., Kameyama, M., Imaoka, S., 61. Ichikawa, Y., Lin, Y.-C., Dumas, D. P., 93, 1508-1515. Furukawa, H., Ishikawa, O., Sasaki, Y., Kab- Shen, G.-J., Garcia-Junceda, E., Williams, 86. Ohmori, K., Takada, A., Ohwaki, I., Taka- uto, T., Iwanaga, T., Matsushita, Y. & M. A., Bayer, R., Ketcham, C., Walker, hashi, N., Furukawa, Y., Maeda, M., Hase- Irimura, T. (1993) Cancer Res. 53, 3632-3637. L. E., Paulson, J. C. & Wong, C.-H. (1992) J. gawa, A., Kannagi, M. & Kannagi, R. (1993) 113. Stone, J. P. & Wagner, D. D. (1993) J. Clin. Am. Chem. Soc. 114, 9283-9298. Blood 82, 2797-2805. Invest. 92, 804-813. 62. Nelson, R. M., Dolich, S., Aruffo, A., Cec- 87. Pinola, M., Renkonen, R., Majuri, 114. Asada, M., Furukawa, K., Kantor, C., Gah- coni, 0. & Bevilacqua, M. P. (1993) J. Clin. M.-L., Tiisala, S. & Saksela, E. (1994) J. mberg, C. G. & Kobata, A. (1991) Biochem- Invest. 91, 1157-1166. Ininunol. 1S2, 3586-3594. istry 30, 1561-1571. 63. Brandley, B. K., Kiso, M., Abbas, S., Nikrad, 88. Lee, Y. C. (1992) FASEB J. 6, 3193-3200. 115. Kotovuori, P., Tontti, E., Pigott, R., Shep- P., Srivasatava, O., Foxall, C., Oda, Y. & 89. Steininger, C. N., Eddy, C. A., Leimgruber, herd, M., Kiso, M., Hasegawa, A., Renkonen, Hasegawa, A. (1993) Glycobiology 3, 633-639. R. M., Mellors, A. & Welply, J. K. (1992) R., Nortamo, P., Altieri, D. C. & Gahmberg, 64. Norgard, K. E., Han, H., Powell, L., Krie- Blochem. Biophys. Res. Commun. 188, 760- C. G. (1993) Glycobiology 3, 131-136. gler, M., Varki, A. & Varki, N. M. (1993) 766. 116. Kujpers, T. W., Hoogerwerf, M., Van der Proc. Nad. Acad. Sci. USA ", 1068-1072. 90. Erbe, D. V., Wolitzky, B. A., Presta, L. G., Laan, L. J. W., Nagel, G., Van der Schoot, 65. Kojima, N., Handa, K., Newman, W. & Ha- Norton, C. R., Ramos, R. J., Burs, D. K., C. E., Grunert, F. & Roos, D. (1992) J. Cell komori, S. (1992) Biochem. Biophys. Res. Rumberger, J. M., Rao, B. N. N., Foxall, C., Biol. 118, 457-466. Commun. 189, 1686-1694. Brandley, B. K. & Lasky, L. A. (1992)J. Cell 117. Sawada, R., Lowe, J. B. & Fukuda, M. (1993) 66. Larsen, G. R., Sako, D., Ahern, T. J., Shaf- Biol. 119, 215-227. J. Biol. Chem. 268, 12675-12681. fer, M., Erban, J., Sajer, S. A., Gibson, 91. Erbe, D. V., Watson, S. R., Presta, L. G., 118. Picker, L. J., Warnock, R. A., Burns, A. R., R. M., Wagner, D. D., Furie, B. C. & Furie, Wolitzky, B. A., Foxall, C., Brandley, B. K. Doerschuk, C. M., Berg, E. L. & Butcher, B. (1992) J. Biol. Chem. 267, 11104-11110. & Lasky, L. A. (1993) J. Cell Biol. 120, 1227- E. C. (1991) Cell 66, 921-933. 67. Kojima, N., Handa, K., Newman, W. & Ha- 1235. 119. Smith, C. W., Kishimoto, T., Abbass, O., komori, S. (1992) Biochem. Biophys. Res. 92. Hollenbaugh, D., Bajorath, J., Stenkamp, R. Hughes, B., Rothlein, R., McIntire, L., Commun. 182, 1288-195. & Anrffo, A. (1993) Biochemistry 32, 2960- Butcher, E. & Anderson, D. C. (1991)J. Clin. 68. Leeuwenberg, J. F. M., Tan, A., Jeunho- 2966. Invest. 87, 609-618. mme, T. M. A. A., Ploegh, H. L. & Buur- 93. BaJorath, J., Hollenbaugh, D., King, G., 120. Kishimoto, T. K., Warnock, R. A., Jutila, man, W. A. (1991) Eur. J. Immunol. 21, 3057- Harte, W., Jr., Eustice, D. C., Darveau, R. P. M. A., Butcher, E. C., Lane, C., Anderson, 3059. & Aruflo, A. (1994) Biochemistry 33, 1332- D. C. & Smith, C. W. (1991) Blood 78, 805- 69. Imai, Y., Lasky, L. A. & Rosen, S. D. (1992) 1339. 811. Glycobiology 2, 373-381. 94. Graves,B. J., Crowther, R. L., Chandran, C., 121. Roberts, D. & Ginsburg, V. (1988) Arch. Bio- 70. DeFrees, S. A., Gaeta, F. C. A., Lin, Y.-C., Rumberger, J. M., Li, S., Huang, K.-S., chem. Biophys. 267, 405-415. Ichikawa, Y. & Wong, C.-H. (1993) J. Am. Presky, D. H., Familletti, P. C., Wolitzky, 122. Weller, P. F., Rand, T. H., Goelz, S. E., Chi- Chem. Soc. 115, 7549-7550. B. A. & Burns, D. K. (1994) Nature (London) Rosso, G. & Lobb, R. R. (1991) Proc. Natd. 71. Yuen, C.-T., Bezouska, K., O'Brien, J., Stoll, 367, 532-538. Acad. Sci. USA 88, 7430-7433. M., Lemoine, R., Lubineau, A., Kiso, M., 95. Watson,S. R.,Imai, Y., Fennie, C., Geoffrey, 123. Carlos, T., Kovach, N., Schwartz, B., Rosa, Hasegawa, A., Bockovich, N. J., Nicolaou, J., Singer, M., Rosen, S. D. & Lasky, L. A. M., Newman, B., Wayner, E., Benjamin, C., K. C. & Feizi, T. (1994)J. Biol. Chem. 269, (1991) J. Cell Bidol. 115, 235-243. Osborn, L., Lobb, R. & Harlan, J. (1991) 1595-1598. 96. Li, S. H., Burns, D., Rumberger, J., Presky, Blood 77, 2266-2271. 72. Moore, K. L. & Thompson, L. F. (1992) Bio- D., Wilkinson, V. L., Anostario, M., Jr., 124. Leeuwenberg, J. F. M., Jeunhomme, chem. Biophys. Res. Commun. 186, 173-181. Wolitzky, B. A., Norton, C. R., Familletti, T. M. A. A. & Buurman, W. A. (1992) Scand. 73. Ushiyama, S., Laue, T. M., Moore, K. L., P. C., Kim, K. J., Goldstein, A. L., Cox, J. Immunol. 35, 335-341. Erickson, H. P. & McEver, R. P. (1993) J. D. C. & Huang, K.-S. (1994) J. Biol. Chem. 1125. Picker, L. J., Kishimoto, T. K., Smith, Biol. Chem. 268, 15229-15237. 269, 4431-4437. C. W., Warnock, R. A. & Butcher, E. C. 74. Etzioni, A., Frydman, M., Pollack, S., Avi- 97. Kansas, G. S., Saunders, K. B., Ley, K., (1991) Nature (London) 349, 79-799. dor, I., Phillips, M. L., Paulson, J. C. & Ger- Zakrzewicz, A., Gibson, R. M., Furie, B. C., I126. Shimizu, Y., Shaw, S., Graber, N., Gopal, shoni-Baruch, R. (1992) N. Engl. J. Med. 327, Furie, B. & Tedder, T. F. (1994) J. Cell Biol. T. V., Horgan, K. J., Van Seventer, G. A. & 1789-1792. 124, 609-618. Newman, W. (1991) Nature (London) 349, 75. Von Andrian, U. H., Berger, E. M., 98. Carver, J. P., Michnick, S. W., Imberty, A. & 799-802. Ramezani, L., Chambers, J. D., Ochs, H. D., Cumning, D. A. (1989) Ciba Found. Symp. 1127. Butcher, E. C. (1991) Cell 67, 1033-1036. Harlan, J. M., Paulson, J. C., Etzioni, A. & 145, 6-18. 128. Mebius, R. E. & Watson, S. R. (1993) J. Im- Arfors, K.-E. (1993) J. Clin. Invest. 91, 2893- 99. Jentoft, N. (1990) Trends Biochem. Sci. 15, munol. 151, 3252-3260. 2897. 291-294. 1129. Buerke, M., Weyrich, A. S., Zheng, Z., Ga- 76. Stoolman, L. M., Yednock, T. & Rosen, S. D. 1LOO. Butenhof, K. J. & Gerken, T. A. (1993) Bio- eta, F. C. A., Forrest, M. J. & Lefer, A. M. (197) Blood 70, 1842-1850. ch/mistry 32, 2650-2663. (1994) J. Clin. Invest. 93, 1140-1148. 77. Wu, N. W., Jalkanen, S., Streeter, P. R. & 1LO1. Sutherland,D. R., Abdullah, K. M., Cyopick, 1L30. Von Asmuth, E. J. U., Smeets, E. F., Ginsel, Butcher, E. C. (1988) J. Cell Biol. 1W7, 1845- P. & Mellors, A. (1992) J. Immunol. 148, L. A., Onderwater, J. J. M., Leeuwenberg, 1851. 1458-1464. J. F. M. & Buurman, W. A. (1992) Eur. J. 78. Watson, S. R., Imai, Y., Fennie, C., Geof- 1102. Saito, S., Levery, S. B., Salyan, M. E. K., Immunol. 22, 2519-2526. froy, J. S., Rosen, S. D. & Lasky, L. A. Goldberg, R. I. & Hakomori, S. (1994)J. Biol. 1:31. Smeets, E. F., De Vries, T., Leeuwenberg, (1990) J. Cell Biol. 110, 2221-2229. Chem. 269, 5644-5652. J. F. M., Van den Eijnden, D. H., Buurman, 79. Ley, K., Gaehtgens, P., Fennie, C., Singer, 1103. Devine, P. L. & Harada, H. (1991) Biol. W. A. &Neeoes, J. J. (1993)Eur. J. Immunol. M. S., Lasky, L. A. & Rosen, S. D. (1991) Chem. Hoppe Seyler 372, 935-942. 23, 147-151. Blood 77, 2553-2555. 1(04. Orlandi, P. A., Klotz, F. & Haynes, J. D. 1 32. Wong, C. S., Gamble, J. R., Skinner, M. P., 80. Von Andrian, U. H., Chambers, J. D., Berg, (1992) J. Cell Biol. 116, 901-909. Lucas, C. M., Berndt, M. C. & Vadas, M. A. E. L., Michie, S. A., Brown, D. A., Karolak, 1()5. Willoughby, R. E. (1993) Glycobiology 3, 437- (1991) Proc. Natd. Acad. Sci. USA 88, 2397- D., Ramezani, L., Berger, E. M., Arfors, 445. 2401. Downloaded by guest on September 29, 2021