multi-recognition capability of e-selectin in a dynamic flow system, as evidenced by differential...

Vol. 189, No. 3, 1992 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS December 30, 1992 Pages 1686-l 694

MULTI-RECOGNITION CAPABILITY OF E-SELECTIN IN A DYNAMIC FLOW SYSTEM, AS EVIDENCED BY DIFFERENTIAL EFFECTS OF SIALIDASES AND ANTI-CARBOHYDRATE ANTIBODIES ON SELECTIN-MEDIATED CELL ADIIE-

SION AT LOW VS. HIGH WALL SHEAR STRESS: A PRELIMINARY NOTE

Naoya Kojimal, Kazuko Handal, Walter Newmana, and Sen-itiroh Hakomorit

‘The Biomembrane Institute, 201 Elliott Ave. W., Seattle, WA 98119; and Dept. of Pathobiology, University of Washington, Seattle, WA 98195

*Otsuka Pharmaceutical Co., Maryland Research Laboratories, 9900 Medical Center Dr., Rockville, MD 20850

Received November 13, 1992

SUMMARY: E-selectin has a “multi-recognition” capability in terms of epitope binding specificity, depending on adhesion conditions (static vs. low- or high-shear stress dynamic systems). Specifically, (i) adhesion based on expression of a2-+3 sialylated Lex (SLex) is prominent under static or low shear stress dynamic conditions; (ii) adhesion under high shear stress dynamic conditions does not depend on the known SLex species, but rather on Lex with an adjacent unidentified sialosyl substitution, which shows different susceptibility to sialidases and antibodies compared to known SLex. 0 1992 Academic Pn?SS, 1°C.

The carbohydrate (CHO) epitope recognized by E-selectin (1) in static adhesion

systems was initially claimed to be a2+3 sialosyl-LeX (SLex) (2,3), and later (additionally)

a2+3 sialosyl Lea (SLea) (4,5). More recently, Larkin et al. reported that Lex, Lea, and

even Leb bind to E-selectin-expressing cells, and proposed that these non-sialylated

structures at high density may also be epitopes of E-selectin (6). We previously studied

CHO-dependent as compared to lectin- or integrin-dependent adhesion in a dynamic

flow system (7), using a parallel-plate laminar flow chamber and videotape recording

method as described by Lawrence et al. (8). Results from the dynamic system were

very different from those from static systems for some adhesion molecules (7).

Therefore, we systematically examined HL60 cell adhesion to E-selectin-coated plates in

The abbreviations used are: AU, Arthrobacter ureafaciens; BSA, bovine serum albumin; CHO, carbohydrate; EC, endothelial cell; HUVEC, human umbilical vein endothelial cell; MoAb, monoclonal antibody; NDV, Newcastle Disease Virus; PBS, phosphate-buffered saline; SLea, sialosyl-Lea; SI..ex, sialosyl-LeX; VC, Vibrio cholerae.

0006-291X/92 $4.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved. 1686

Vol. 189, No. 3, 1992 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

static vs. dynamic systems, with emphasis on the effects of various sialidases and anti-

SLex and anti-LeX MoAbs.

MATERIALS AND METHODS

Adhesion of HI60 cells to E-selectin-coated mates. Promyelocytic leukemia HL60 cells were metabolically labeled by incubation with [3H]thymidine and added to E-selectin-coated plates (1x106 cells/well). The 96-well plates (Probind, Falcon, Lincoln Park, NJ) were coated with 0.2 &ml of truncated recombinant E-selectin lacking transmembrane and cytoplasmic domains (9) for 18 hr. Binding of E-selectin to the plates was confirmed by reactivity with anti-E-selectin MoAb 3B7 (10). Plates were then coated with 1% BSA/PBS for 1 hr, washed once with PBS, and used for adhesion assays. After incubation (10 min in most cases), plates were washed with PBS, and adherent cell number was estimated based on radioactivity count. These numbers were used as “baseline” values for subsequent studies as described below.

Effects of various sialidases and anti-CHO MoAbs on E-selectin-deoendent cell adhesion. HL60 cells (5x106) were treated with 0.2 U/ml of Newcastle Disease Virus (NDV) sialidase (prepared from allantoic-amniotic fluid of virus-infected chicken egg, or purchased from Oxford Glycotech Inc., Oxford, U.K.) or Kbrio cholerue (VC) sialidase, or 1 U/ml of Arthrobacter ureafuciem (AU) sialidase in PBS, incubated for various durations up to 90 min at 37°C. The reaction was terminated by adding FCS to reaction mixture then immediately centrifuged at 1000 rpm for 3 min and washed with RPM1 medium containing 10% FCS (x2).

For studies of MoAb effect on cell adhesion, cells were mixed with culture supernatant of various hybridomas containing 5-10 pg IgM or IgG antibody per ml. MoAbs tested were anti-SLeX SNH3 (IgM) and SNH4 (IgG,) (ll), and anti-LeX SHl (IgG,) (12) and FH2 (IgM) (13). Non-specific mouse IgG and IgM (10 pg/ml) were used as control antibodies. [3H]Thymidine-labeled cells (1~10~) in PBS were treated with 200 ~1 of MoAb solution for 90 min at 4°C washed with RPM1 medium containing 10% FCS, and used for adhesion assay. For combination treatment of sialidase plus MoAb, cells were treated with sialidase for 90 min at 37 OC and washed as described above. The treated cells were then added with MoAbs incubated at 40c for 90 min. The MoAb concentrations (5-10 &ml) used in these experiments gave the same fluorescence intensity as 4-fold higher (20-40 pg/ml) concentrations. Cytofluorometric analysis of cells with various concentrations of MoAbs indicated that 5-10 &ml Ig concentration completely saturated cell surface epitopes.

E-selectin-denendent adhesion assav in a dvnamic flow system. A parallel-plate laminar flow chamber connected to an infusion pump was constructed as described by Lawrence et al. (8), and used as described in our previous report (7). Briefly, the flow chamber consisted of a glass plate to which a parallel, transparent plastic surface (coated with E-selectin) was attached; there was a 114 pm gap between the two surfaces. A laminar flow with defined rate and wall shear stress was achieved by manipulation of the pump. A suspension of HMO cells, with or without treatment by sialidases or MoAbs, was passed through the chamber. Cell movements were observed under inverted phase-contrast microscope and recorded by time-lapse videocassette recorder. Adhesion was observed as rolling followed by stopping of cells. Number of cells bound during 3 min at different shear stresses (0.76 to 15.5 dynes/cm2) was counted from several fields recorded on videotape. Shear stress (T) was calculated by the equation of Lawrence et al. (8). Detachment of once-adhered cells was observed at defined shear stress values in the presence of MoAb, and grade of detachment was

1687


expressed as speed of rolling of once-adhered cells as estimated from videotape recording.

For E-selectin coating, 50 ~1 of truncated recombinant E-selectin (9) (0.2 &ml) was placed on a marked area (0.5 cm diameter) of “non-adherent” plastic plates (25x75 mm) and incubated for 18 hr at 4°C. Plates were then coated with 1% BWPBS for 1 hr at room temp and washed with PBS. In some experiments, E-selectin-coated plates were treated with 10 &ml of anti-E-selectin MoAb 3B7 or Fab fragment of MoAb 7A9 for 1 hr at room temp. Plates coated as above were affixed in a flow chamber, a suspension of HL60 cells (lxl@ /ml) was passed through the chamber as described in the preceding section, and adhesion at various shear stresses was measured.

RESULTS AND DISCUSSION

SLex expression and E-selectin-deoendent adhesion of HL60 cells. NDV, VC,

and AU sialidases were equally effective in eliminating SLex expression, as probed by

flow cytometry using MoAbs SNH3 and SNH4 (Fig. 1A and its inset). Only 10 m in

incubation was required to eliminate >95% of SLex expression with all types of

sialidase. In contrast, E-selectin-dependent HL60 adhesion was only ~50% inhibited by

NDV sialidase (which cleaves a2+3 sialosyl linkage only), even after prolonged incuba-

tion. VC or AU sialidases (which cleave both (r2+3 and a2+6 sialosyl linkages) inhibited

HL60 adhesion by >95% within l-2 hr (Figs. lB, 2A). Therefore, the well-known a2+3

sialosyl Lex (SLex) structure is not the sole epitope for E-selectin-dependent adhesion

of HL60 cells. HL60 cells do not express SLea or any other type 1 chain structure, and

these structures can therefore be excluded as adhesion epitopes.

Effects of sialidases and MoAbs on HI.60 adhesion in static and dvnamic

In a static system, E-selectin-dependent HL60 adhesion was effectively systems.

inhibited by anti-LeX MoAbs SHl or FH2, to a similar degree as by anti-SLeX MoAb

SNH4 (in fact, in some experiments the effect of SNH4 was much weaker than that of

SHl). However, adhesion was strongly (~-80%) inhibited by combinations of SHl plus SNH4 or SNH3 (Fig. 2B) or NDV sialidase plus SHl or FH2 (Fig. 2C). However, this

inhibitory effect was weaker than the effect of VC or AV sialidases (Fig. 2A).

Adhesion was almost completely blocked by anti-E-selectin MoAbs 7A9 or 3B7 (data

not shown).

In a dynamic system, NDV sialidase had an inhibitory effect at low shear stress,

whereas VC or AU sialidase reduced adhesion greatly at both low and high shear stress

(Fig. 3A). Anti-LeX IgG MoAb SHl strongly inhibited adhesion even at high shear

stress, whereas the effect of anti-SLeX IgG3 MoAb SNH4 was m inimal (Fig. 3B). A

strong inhibition was produced by a combination of NDV sialidase plus anti-LeX MoAb

SHl (Fig. 3C). A m ixture of anti-LeX plus anti-SLeX MoAbs produced stronger

1688

vol. 189, No. 3, 1992 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

6

0 0 20 40 60 80 100

Time (min)

Figure 1. Effects of sialidases on SLex exoression (A) and E-selectin-deuendent adhesion (B) of HL60 cells

Panel A: SLex expression on HL60 cells incubated with 0.2 U/ml NDV sialidase (A), 0.1 U/ml VC sialidase (a), or 0.2 U/ml AU sialidase (0) for various durations (abscissa). Values for VC and AU sialidases were identical and are therefore shown as single points. 1x106 cells were treated with 0.5 ml PBS (for VC or AU sialidase) or 0.1 ml PBS (for NDV sialidase).

Panel A inset: Flow cytometry of HL60 SLex expression at 90 min incubation. Abscissa, log fluorescence intensity. Ordinate, relative cell number. A: Solid line, cells stained with MoAb SNH4 as primary antibody. Dotted line, control cells (stained with mouse IgG plus IgM [lo &ml] as primary antibody). B: MoAb SNH3 as primary antibody; control as in A. C: Solid line, cells treated with Newcastle Disease Virus (NDV) sialidase and then stained with MoAb SNH4. Dotted line, control cells (as in A, after sialidase treatment). D: NDV sialidase followed by MoAb SNH3; control as in C. E: fibrio choierne (VC) sialidase followed by MoAb SNH4. F: VC sialidase followed by MoAb SNH3. Note that expression of SLex (defined by both SNH3 and SNH4) was completely abolished by both NDV and VC sialidases.

Panel B: & cells treated with NDV sialidase (0.2 U/ml). 0, VC sialidase (0.1 U/ml). 0, AU sialidase (0.2 U/ml). Abscissa, incubation time. Each point represents the mean of 3 experiments.

inhibitory effect than either MoAb alone (Fig. 3D). These inhibitory effects in the

presence of antibodies were, however, weaker than the effect of VC or AV sialidases

(Fig. 3A).

ComDarative effects of sialidases and MoAbs on E-selectin-deDendent HL60

adhesion at various shear stresses. At low shear stress (~4 dynes/cm2), adhesion was

significantly inhibited by NDV sialidase or by anti-SL,eX MoAb SNH4, whereas these

1689


Sialidase MAb Treotment NDV Sialldase

:. A Treatment B C

- MAb Treatment

Treotmeni

Figure 2. Adhesion of HL60 cells to E-selectin-coated elates in a static system Abscissa, type of treatment. Ordinate, percent cell adhesion (relative to

untreated control cells). A: effects of various sialidases. B: effects of anti-LeX and anti-SLeX MoAbs alone and in combinations (incubated 90 min at 4°C). C: effects of NDV sialidase plus MoAb. Each value represents the mean of 6 experiments.

reagents had no effect at high shear stress (8-16 dynes/cm2). In striking contrast, VC

sialidase completely abolished adhesion at high shear stress (Fig. 4). HL,60 adhesion at

high shear stress depends partially on Lex and fully on some still-unidentified sialylated

200

“E E

; 100 E 0) 6 a O

Wall Shear Stress (dynes/cm2)

Fieure 3. Adhesion of HL60 cells to E-selectin-coated nlates in a dynamic flow system Truncated E-selectin was coated onto marked areas (diameter 0.5 cm) of nlastic

plates, and adhesion under defined wall shear stresses was issayed as described h- Methods. Abscissa, shear stress (dynes/cm2). Ordinate, number of cells adhered within 3 min. A: 0, control (untreated) cells; A, cells treated with NDV sialidase; 0, VC sialidase; A, AU sialidase. B: 0, control; A, cells cultured in medium containing anti- SLex IgG3 MoAb SNH4; l , anti&+ IgM MoAb FH2; A, anti-LeX 1gG~ MoAb SHI. C: 0, control; A, NDV sialidase; 0, MoAb SHl; & NDV sialidase plus MoAb SHI. D: 0, control; 0, mixture (1:l) of MoAbs SNH4 and FH2. A, mixture (1:l) of MoAbs SNH4 and SHl. Each point represents the mean of 3-5 experiments.

1690


0 2 4 6 8 IO 12 14 16

Shear Stress I” dynes /cm’

Figure 4. Effects of MoAbs and sialidases on HL60 cell binding in a dvnamic flow system at various shear stresses

Ordinate: percent HL60 cell binding (relative to untreated control cells) to E-selectin-coated plates after treatment of cells with sialidases or anti-f33 or anti-LeX MoAbs, or after treatment of plates with anti-E-selectin MoAb. Abscissa: shear stress (dynes/cm2). A, anti-33 MoAb SNH4 (15 PgIml). A, anti-LeX MoAb SHl (15 pg/ml). 0, NDV sialidase (0.2 U/ml). 0, VC sialidase (0.1 U/ml). v, Fab fragment of anti-E-selectin MoAb 7A9 (10 &ml). Reasons for the relative ineffectiveness of MoAb 7A9 and VC sialidase at low shear stress (l-8 dynes/cm2) are unknown. Each point represents the mean of 3 experiments. Average numbers of untreated cells bound at shear stresses of 15.5, 7.75, 3.13, 1.56, and 0.78 dynes/cm2 were, respective&. 4.5+1.0, 27.Ok5.2, 109.629.6, 206.2k27.9, and 283.8G8.1 per mm2.

structure (see Discussion). Adhesion is consistently dependent on E-selectin, since it

was abolished at both low and high shear stress by treatment of plates with anti-

E-selectin MoAb 7A9 or intact MoAb 3B7. MoAb SHl inhibited adhesion more

strongly at high than at low shear stress, but the difference was relatively small.

Discussion. Two important observations and conclusions were made: (i) expres-

sion of SLex with a2+3 sialylated structure was eliminated within 10 min by NDV sialidase, whereas E-selectin-dependent HL60 adhesion was reduced only ~50% by this

treatment. Reduction of adhesion by >90% required VC or AU sialidases, which

cleave both 1~2-3 and a2+6 sialosyl linkages. Therefore, E-selectin-dependent HL60

adhesion seems to require as epitope not only the well-known ~2-3 sialylated SLex, but also Lex with unknown sialylated structure which can be cleaved by VC or AU sialidase

but not by NDV sialidase. (ii) NDV sialidase and anti-SLeX MoAb SNH4 produced

only moderate inhibition of E-selectin-dependent HL60 adhesion at low shear stress,

and had no inhibitory effect at high shear stress, indicating that the well-known SLex is

not an efficient adhesion epitope at high shear stress. Strong inhibition, particularly at

high shear stress, required treatment with VC or AU sialidase. Anti-LeX MoAb was a

better inhibitor than anti-&Y MoAb in both static and dynamic systems, and strongest

1691


r Adhesion in - static system n

m

I Adhesion in

E-selecti? - ~~~$,$stem- SA2-.3Le’-0

n

\ shear stress L&O

m

Adhesion in dynamic system - at high wall shear stress

Figure 5. Multi-recognition of various CHO eDitoDes bv E-selectin under dvnamic (high and low shear stress) and static conditions

The scheme summarizes findings on differential effects of sialidases and anti- CHO MoAbs on E-selectin-dependent adhesion under various conditions. Type size used for various structures reflects their relative importance in adhesion. Under static conditions, the major epitope is the well-known ~~2-3 SLP multiply O-linked structure (upper right). Under high shear stress dynamic conditions, this structure becomes unimportant, and the major epitope is an unidentified structure which is insensitive to NDV sialidase or anti-Sk% MoAb, but completely inhibitable by VC or AU sialidase or partially inhibitable by anti-LeX MoAb. This unidentified high-affinity epitope could be Lex with branched structure sialylated at an unknown x position (but most likely 6-position) which is insensitive to NDV sialidase (lower right). Multiply O-linked LeX may play some role in E-selectin-dependent cell adhesion in both low and high shear stress dynamic systems.

inhibition was produced by a combination of anti-L.eX plus anti-SL,eX MoAbs, or NDV

sialidase plus anti-LeX MoAb. Above results (i) and (ii) taken together suggest that the

major high-affinity epitope of E-selectin may consist of some unknown Lex structure

with an adjacent sialylated substituent. Such structure could be (i) highly clustered

O-linked SL.ex (14) or bi- or multivalent SLex whose sialosyl residues are sterically

inaccessible to NDV sialidase or MoAb SNH3 or SNH& (ii) Lex having adjacent a2+6

sialosyl branching or internal a%6 sialosyl residue. This structure, however, is different

from “6-C ganglioside” (SAa6Galp4GlcN/?3Galp4[Fuc~3]GlcNp3Galp4GlcCer) (15),

which does not bind to E-selectin. An analogous situation may exist for adhesion via

type 1 chain epitopes Lea, a2+3 SLea, and a2+6 SLea (Kojima N, Handa K, Hakomori

S, unpubl.)

In a static system or low shear stress dynamic system, E-selectin preferentially

recognizes (~2+3 SLex, whereas in a high shear stress dynamic system an unidentified

structure consisting of Lex plus an adjacent sialosyl substitution resistant to NDV

1692

Vol. 189, No. ‘3, 1992 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

sialidase (as described above) comprise the preferred high-affinity epitope, and the role

of a2+3 SLex becomes negligible. This idea (presented schematically in Fig. 5) is

supported by the findings that (i) NDV sialidase or anti-SLeX MoAb were inhibitory

only at low shear stress or in a static system, (ii) anti-LeX MoAb was equally inhibitory

at low and high shear stress, (iii) adhesion at high shear stress was completely abolished

by VC sialidase, which cleaves both a2+3 and a2+6 sialosyl linkages, and (iv) all

observed cell adhesion events depend totally on E-selectin, since Fab of anti-E-selectin

MoAb greatly reduced or completely abolished the adhesion.

Results of the present study partially support the observation of Larkin et al.

that Lex also binds to E-selectin (6). However, Lex alone is clearly not sufficient as

E-selectin epitope in high shear stress dynamic systems; rather, Lex with unidentified

sialylated structures resistant to NDV sialidase but sensitive to AU and VC sialidases

comprise the high-affinity binding site under such conditions. Interestingly, the same

pattern of susceptibility to sialidases and antibodies has been observed for HL60 cell

adhesion to activated human endothelial cells in static vs. dynamic systems (data to be

presented elsewhere). It is possible that E-selectin adaptively changes its conformation?

and thereby accommodates different binding specificities, in a dynamic flow environ-

ment.

ACKNOWLEDGMENTS. The authors thank Dr. Stephen Anderson for scientific editing and preparation of the manuscript. This study was supported by funds from The Biomembrane Institute, in part under a research contract with Otsuka Pharmaceu- tical Co., and by National Cancer Institute Outstanding Investigator Grant CA42505.

REFERENCES

1.

2.

3.

4.

5.

6.

7.

Bevilacqua, M.P., Stengelin, S., Gimbrone, M.A.Jr., and Seed, B. (1989) Science 243, 1160-l 165. Phillips, M.L., Nudelman, E.D., Gaeta, F.C.A., Perez, M., Singhal, A.K., Hako- mori, S., and Paulson, J.C. (1990) Science 250, 1130-1132. Walz, G., Aruffo, A., Kolanus, W., Bevilacqua, M.P., and Seed, B. (1990) Science 250, 1132-1135. Berg, E.L., Robinson, M.K., Mansson, O., Butcher, EC., and Magnani, J.L. (1991) J. Biol. Chem. 266, 14869-14872. Takada, A., Ohmori, K., Takahashi, N., Tsuyuoka, K., Yago, A., Zenita, K., Hasegawa, A., and Kannagi, R. (1991) Biochem. Biophys. Res. Commun. 179, 713-719. Larkin, M., Ahern, T.J., Stall, M.S., Shaffer, M., Sako, D., O’Brien, J., Yuen, C.-T., Lawson, A.M., Childs, R.A., Barone, K.M., Langer-Safer, P.R., Hasegawa, A., Kiso, M., Larsen, G.R., and Feizi, T. (1992) J. Biol. Chem. 267, 13661-13668. Kojima, N., Shiota, M., Sadahira, Y., Handa, K., and Hakomori, S. (1992) J. Biol. Chem. 267, 17264-17270.

1693


8.

9.

10.

11.

12.

13.

14.

1.5.

Lawrence, M .B., Smith, C.W., E&in, S.G., and McIntire, L.V. (1990) Blood 75, 227-237. Shimizu, Y., Shaw, S., Graber, N., Gopal, T.V., Horgan, K.J., Van Seventer, G.A., and Newman, W. (1991) Nature 349, 799-802. Graber, N., Gopal, T.V., Wilson, D., Beall, L.D., Polte, T., and Newman, W. (1990) J. Immunol. 145, 819-830. Muroi, K., Suda, T., Nojiri, H., Ema, H., Amemiya, Y., M iura, Y., Nakauchi, H., Singhal, A.K., and Hakomori, S. (1992) Blood 79, 713-719. Singhal, A., Orntoft, T.F., Nudelman, E.D., Nance, S., Schibig, L., Stroud, M .R., Clausen, H., and Hakomori, S. (1990) Cancer Res. 50, 1375-1380. Fukushi, Y., Hakomori, S., Nudelman, E.D., and Cochran, N. (1984) J. Biol. Chem. 259, 4681-4685. Kojima, N., Handa, K., Newman, W., and Hakomori, S. (1992) Biochem. Biophys. Res. Commun. 182, 1288-1295. Hakomori, S., Nudelman, E.D., Levery, S.B., and Patterson, C.M. (1983) Biochem. Biophys. Res. Commun. 113, 791-798.

1694

multi-recognition capability of e-selectin in a dynamic flow system, as evidenced by differential...

Documents