binding patterns of monoclonal anti-b, anti-h and anti-(leb+ y) on erythrocytes, imaged in the...

Eur J Haematol 1989;43:226-234

Key words: human erythrocytes - monoclonal anti-8 - monoclonal anti-H - monoclonal anti-(Leb+ Y) - surface labelling - colloidal gold particles - scanning electron microscopy - backscatter electron imaging

Binding patterns of monoclonal anti-B, anti-H and anti-(Leb + Y) on erythrocytes, imaged in the scanning electron microscope

Hans Erik Heier' & Ellen Namork'

National Institute of Public Health, Departments of 'Immunology and of 'Methodology, Oslo, Norway

Binding patterns of monoclonal IgM anti-B, anti-H and anti-(Leb + Y) antibodies on erythrocytes were visualized in the backscatter electron imaging (BEI) mode of a scanning electron microscope by indirect immunogold labelling. The results were independent of secretor status, supporting the view that type 2 oligosaccharides predominate on erythrocytes. There were considerable cell-to-cell variations of amounts of the antigens, presumably because of differences of antigen development. The amount of H antigens may have been overestimated in previous studies. B and H antigens seemed linked mostly to mobile structures, probably lipid, on the convex parts of the cells and to immobile structures, probably protein, on the concave parts. Y antigens appeared bound to mobile structures also on the concave part of the cell membrane. The results therefore confirm that B oligosaccharides on erythrocytes develop from both lipid- and protein-bound H, but suggest that Y develops mainly from lipid-bound H.

Accepted for publication March 3, 1989

The binding of anti-A antibodies to erythrocytes has been studied previously in immunofluores- cence (1, 2), transmission electron microscopy (3-6) and in scanning electron microscopy (SEM) (7). Both polyclonal (1-6) and mouse monoclonal anti-A (7) have been investigated. These studies have revealed that amount and composition of the antigen A vary between and within the various subgroups of A, between single cells, and between the convex and concave parts of the cells. Different proportions of A oligosaccharides bound to protein and lipid, respectively, as well as different amounts of mono- and difuco-

sylated A oligosaccharides may account for the variations (7).

While the binding of anti-A to erythrocytes has been studied extensively, few results have been published on the binding of antibodies to B, H and other erythrocyte antigens related to the ABO system (2). It seems of interest to perform such studies, since they may increase our insight into the development and interrelationship of the various antigens (8-10).

The antigens in question are biochemically highly complex (8-lo), and the human alloanti- bodies to A, B, H and related antigens may vary

BINDING OF ERYTHROCYTE ANTIBODIES 227

in their fine specificities (11). It seems advisable, therefore, to study binding patterns of monoclonal antibodies reacting with biochemically de- fined epitopes. SEM studies are of special interest, since SEM permits comparison of binding patterns on various parts of morphologically in- tact cells. Bound antibodies are revealed in SEM by labelling with colloidal gold particles, which are easily visualized by atomic number contrast in the backscatter electron imaging (BEI) mode (12-14). This mode of imaging has been used in the present study to characterize binding patterns of monoclonal IgM anti-monofucosylated B, anti-H and anti-(Leb + Y) antibodies to erythrocytes by indirect irnmunogold labelling. The binding patterns observed are discussed in relation to structure and development of the antigens.

Material and methods Blood samples: Fresh ACD anticoagulated blood was obtained from 26 individuals: 11 of blood group B, 10 of blood group 0 and 5 of blood group A,. ABH grouping and A subgrouping had been carried out

previously in at least two independent blood samples using standard techniques with polyclonal reagents. Lewis typing was carried out using polyclonal allo-anti- Lea and anti- Leb. In Le (a-b-) cases secretor status was determined by agglutination inhibition assay with sali- va.

Agglutination reactions were performed by standard tile technique at room temperature and read micro- scopically.

Monoclonal antibodies (moabs): Mouse monoclonal IgM anti-B (code B 003) specific for monofucosylated B oligosaccharide chains (15), IgM anti-H (code 647/9A2) specific for H type 2 chains (16), and IgM anti-(Leb + Y) (code H W), reacting similarly with type 1 (Leb) and type 2 (Y) difucosylated chains (16), were applied. The reactivity of H 004 is very similar to that of another, recently published moab (code 64/4D8) from the same producer (17). Mouse monoclonal IgM anti-A (code A 003) (7, 15) was used for control pur- poses. All moabs were kindly donated by BioCarb AB, S-223 70 Lund, Sweden.

Immunolabelling and preparation for electron microscopy: For details, see (7). Erythrocytes were prefixed in 0.06% glutaraldehyde (GA) to prevent agglutination without reducing the antibody binding capacity of the antigens (3, 18-20), washed in 0.5% glycine and

Figure 1 . Blood group B erythrocytes sensitized with mouse monoclonal IgM anti-B (B 003). labelled with 40 nm colloidal gold probe and visualized in the BE1 mode. a): cell-to-cell variation of labelling strength, and some clustering mainly on the convex part of the membrane (4OOO x). Picture width 17.5 pm. b) higher mangincation of the strongly labelled cell in a), showing individual labelling particles and more extensive clustering on convex than on concave part of the membrane (12500 x). Picture width 5.5 pm .

228 HEIER & NAMORK

Figure 2. Blood group 0 (a) and A, (b) erythrocytes sensitized with mouse monoclonal IgM anti-H (647/9A2), labelled and visualized as in Figure 1. Note cell-to-cell variation of labelling strength, more pronounced clustering on convex part of the membrane, and ‘pit accumulation’ (arrow) (4OOO x). Picture widths 17.5 wm.

incubated for 30 min with a 1 + 1 dilution of moab. After washing, the cells were incubated overnight with goat anti-mouse antibodies conjugated to 40 nm gold particles (CAM IgM G40, Janssen Life Sciences Prod- ucts, Belgium) at a final concentration of 1 + 4. The incubations were carried out at room temperature. All dilutions and washings of cells were performed in 0.1 mol/l Na-cacodylate in 0.1 mol/l saccharose, pH 7.2, containing 0.02% Na-azide. The dilutions of moabs and gold probe were chosen from results in preliminary experiments. Cells incubated with non-reactive antibody (A 003 with B and 0 cells, B 003 with A, cells) and cells incubated with buffer instead of antibody, served as negative controls. A2 cells sensitized with A 003, as investigated previously (7), were used as positive control.

The labelled cells were sedimented onto chips of mica, previously coated with poly-L-lysine, and fixed in 2.5% GA. The specimens were then dehydrated, critical point dried and coated with a thin layer of carbon before mounting and examination in the electron microscope.

Electron microscopy: The specimens were observed in the BE1 mode of a JSM 840 scanning electron microscope equipped with a solid state backscatter de- tector and a lanthanium hexaboride (LaB,) emitter. The microscope was operated at 15 keV using a probe current of 1 - 10.’’ Amp and a working distance of 7-8

mm. Micrographs of 150-200 cells were recorded for each specimen at a primary magnification of 3 0 0 0 ~ .

Interpretation of micrographs: A cell was recorded as labelled if at least 10 gold particles were visible on its surface. Differential counting was carried out inde- pendently by each of the authors and included rec- ording of unlabelled cells (-), weakly to moderately labelled cells (+) and strongly labelled cells (+ +). The percentage of cells showing labelling mainly on the concave part of the membrane (‘pit accumulation’ of labelling, illustrated in Figure 2a (arrow)) was also recorded. The term cluster was used to designate closely adjacent gold particles surrounded by unlabelled membrane (see Figure lb).

Results The immunolabelling and agglutination results are summarized in Tables 1-3. Electron micrographs of labelled erythrocytes are shown in Fig- ures 1-3. The labelling was considered specific, since no labelling was observed in the negative controls. Labelling of A, cells sensitized with A 003, used as positive control, showed the expected labelling pattern (7).

Immunolabelling of B cells sensitized with anti-B (B 003): A mean of 75% of the cells was


labelled, but with considerable variation between donors. Cell-to-cell variation of labelling strength was noted in all cases, but the percentage of strongly labelled cells was low in most cases (Figure la, Table 1).

Cells with ‘pit accumulation’ of labelling were usually weakly labelled and were seen in all individuals, although at variable percentages (Table 1). Some clusters of labelling particles occurred, mostly on strongly labelled cells and on the convex part of the cell membrane (Figure l a and b).

The immunolabelling results were similar in secretors and non-secretors (Table 1).

Immunolabelling of 0 and A , cells sensitized with anti-H (647/9A2): Means of 40% of the 0 cells and about 30% of the A, cells were labelled, but with considerable variation between donors (Table 2). Cell-to-cell variation of labelling strength was apparent, although few strongly labelled cells were seen (Figure 2).

Cells with ‘pit accumulation’ of labelling were observed in all individuals, although at variable percentages (Table 2). Clustering was generally more pronounced as compared to B cells sensitized with anti-B, but the clusters were larger and surrounded by more extensive unlabelled areas

on the convex than on the concave parts of the cells (Figure 2). Clustering was more pronounced on strongly than on weakly labelled cells.

The immunolabelling results were similar in secretors and non-secretors (Table 2).

Immunolabelling of 0 and A , cells sensitized with anti-(L$+ Y) (H 004): Means of about 35% of the 0 cells and less than 30% of the A, cells were labelled, but with considerable variation between donors (Table 3). Few strongly labelled cells were seen, but cell-to-cell variation of labelling strength was nevertheless observed (Figure 3).

Cells with ‘pit accumulation’ of labelling were seen in all individuals, although at variable percentages (Table 3). Clusters occurred frequently, also on rather weakly labelled cells, and were equally apparent on the convex and concave parts of the cells (Figure 3).

Both in 0 and A,, the mean percentage of labelled cells was a little higher in non-secretors than in secretors (Tables 3a and b), while no difference was found between 0 Le(b+) and 0 Le(b-) samples (Table 3c).

Agglutination (Tables 1-3): Agglutination ti- ters and scores of anti-B with B cells were slightly weaker than those of anti-H and anti-(Leb + Y)

Figure 3. Blood group 0 (a) and A, @) erythrocytes sensitized with mouse monoclonal IgM anti-(Leb+Y) (H 004). labelled and visualized as described in Figure 1. Note cell-to-cell variation of labelling strength, equally pronounced clustering on convex and concave parts of the membrane, and ‘pit accumulation’ (4000 x). Picture widths 17.5 pm.

230 HEIER & NAMORK

TABLE 1 Differential counting and agglutination results with anti-B (B 003) in blood group B. M = mean, R = range

Differential counts (%) Agglutination Secretor status n ‘Pit acc’ + + + Titer Score

Secretor 5’ M 24.4 M 75.1 M 0.5 M 28.3 M 59.3 R 12.0 R 50.5 R 0.0 R 5.0 R 64’ R 55.0 -48.0 -87.5 -1.0 -47.0 -62.0

Non-secretor 62 M 28.3 M 67.1 M 4.4 M 16.5 M 56.2 R 7.5 R 45.2 R 0.5 R 10.0 R 32 R 53.0

-52 -78 -22.5 -25 - 128‘ -60.0

B total 11 M 26.8 M 70.8 M 2.5 M 25.8 57.6

I : AU cases Le(a-b+). ’: 5 cases Le(a+b-), 1 case Le(a-b-). ’: 64 in all cases. ‘: 32 in 2 cases, 64 in 3 cases and 128 in 1 case.

weaker than those of anti-H and anti-(Leb+Y) with both 0 and A, cells. The 2 latter moabs showed similar agglutination potencies and agglutinated A, cells as strongly as 0 cells. Agglut- ination results were similar in secretors and non- secretors. Anti-(Leb + Y) agglutinated 0 Le(b +) and 0 Le(b-) cells with equal strength (Table 3c).

Discussion A, B, H and Le antigens develop by sequential addition of monosaccharide units to common oligosaccharide precursors. The existence of two series of chain isomers, type 1 and type 2, has been firmly established in blood groups B, 0 and A, (8-10). A and B originate from H type 1 or 2, whereas Leb and Y are isomers which originate from H type 1 and 2, respectively (Figure 4). Type 1 H, A and B, and Leb can only be formed in secretors and are thought to be adsorbed to the erythrocyte membrane, while type 2 chains are produced intrinsically in erythrocyte precursors in nearly all individuals. The oligosaccharides are linked to lipids and proteins in the erythrocyte membrane (9).

Our findings support the view that type 2 B and H chains predominate on erythrocytes (7-10, 21), since labelling with anti-B and anti-H were similar in secretors and non-secretors. The results also support the notion that the number of Lewis active molecules is low on erythrocytes (22), since the findings with anti-(Leb+Y) were similar in Le(b+) and Le(b-) individuals. The binding pat-

terns observed therefore reflect the distribution of three type 2 oligosaccharide structures, one of which is the precursor of the two others.

Cell-to-cell variations of immunogold labelling strength were found with all 3 moabs. The amounts of all type 2 structures therefore vary from cell to cell (1-7). An equilibrium appears to exist between amounts of A and H antigens in the various subgroups of A (23), suggesting that the variations reflect different degrees of antigen development and not differences of total amounts of type 2 oligosaccharides. Preliminary results with double immunogold labelling in A, support this assumption (24). Nakajima et a1 (25) found that the amount of ABH antigens in secre- tory granulae in salivary glands varied between acini, but appeared quite similar in the cells of single acini. It could be, therefore, that the de- gree of ABH antigen development differs between erythrocytes originating from different erythroid progenitor clones.

We found that fewer A, than 0 cells were labelled when sensitized with anti-H, as might be expected from the concept of equilibrium between the amounts of H and A antigens. No such difference was found in agglutination, however, and the immunogold labelling therefore appears more sensitive to epitope number differences than agglutination on tiles. The immunolabelling results with anti-(Leb + Y) indicate similarly that there are more Y epitopes on 0 than on A, cells.

Results with enzymatic conversion of H to A (26) and by radioimmunoassay with polyclonal

BINDING OF ERYTHROCYTE ANTIBODIES 23 1

anti-B (23) indicate that there are at least as many H epitopes on 0 cells as B epitopes on B cells. The consistently weaker immunolabelling results with anti-H in blood group 0 compared with anti-B in B might therefore indicate that anti-H had a low affinity for its epitope. Howev- er, one would then not have expected stronger agglutination with anti-H than with anti-B. It seems more reasonable, therefore, to suggest that anti-H bound with high affinity to a relatively low number of epitopes, while anti-B bound more weakly to a higher number of epitopes. This is supported by the more pronounced clustering observed with anti-H than with anti-B, since clustering is thought to represent cross- linking of mobile antigens (14, 27) and may increase with increasing antigen-antibody affinity (6, 7). Our study therefore indicates that there are fewer H epitopes on 0 cells than B epitopes

on B cells. The discrepancy with previous results suggests that the enzyme conversion method is less specific than binding of monoclonal anti-H. The number of H epitopes may not differ much from that of Y, but from the antigen equilibrium it might be hypothesized that H and Y tend to be expressed on different cells. This should be studied in double immunogold labelling.

We have suggested previously that ABH-related antigens which remain mobile on erythrocytes prefixed with GA, are carried by lipids (7). If so, the proportions of lipid-carried B and H chains may be higher on the convex than on the concave part of the cell, while lipid-carried Y chains may predominate throughout the membrane. This supports the idea that B develops from both lipid- and protein-carried H, but suggests that Y develops mainly from lipid-carried H. The possi- bility seems remote that Y antigens in the con-

TABLE 2a Differential counting and agglutination results with anti-H (647/9A2) in blood group 0. M = mean, R = range

Differential counts (To) Agglutination Secretor status n ‘Pit acc’ - + + + Titer Score

Secretor 5’ M 59.2 M 39.1 M 1.9 M 29.7 M 68.4 R 44.0 R 23.0 R 0.0 R 7.5 R 6 4 R 60.0 -77.0 -50.5 -5.3 -52.0 -256 -72.0

Non-secretor 5’ M 54.0 M 45.6 M 0.6 M 28.6 M 71.2 R 44.5 R 22.5 R 0.0 R 7.0 R 128 R 61.0 -64.4 -55.0 -1.5 -59.5 -256 -83.0

0 total 10 M 56.6 M 42.4 M 1.3 M 29.2 M 69.8

I : 3 cases Le(a-b+), 2 cases Le(a-b-). ’: 3 cases Le(a-b+), 2 cases Le(a-b-).

TABLE 2b Differential counting and agglutination results with anti-H (647/9A2) in blood group A,. M = mean, R = range

Differential counts (To) Agglutination Secretor status n ‘Pit acc’ + + + Titer Score

Secretor 3 M 66.8 M 31.9 M 1.0 M 41.2 M 69.0 R 63.4 R 26.8 R 0.0 R 18.5 R 128’ R 60.0 -72.0 -36.6 -1.9 -58.4 -81 .O

Non-secretor 2 M 68.5 M 30.8 M 1.1 M 38.5 M 65.5 R 61.1 R 17.9 R 0.3 R 29.0 R 6 4 R 62.0 -75.9 -33.5 -1.5 -48.0 -256 -69.0

A, total 5 M 67.5 M 31.5 M 1.0 M 40.1 M 67.6

I: Titer 128 in all cases.

232 HEIER & NAMORK

TABLE 3a Differential counting and agglutination results with anti-(L@ + Y) (H 004) in blood group 0. M = mean, R = range

Secretor status Differential counts (70)

n i + + ‘Pit acc’

Agglutination

Titer Score

Secretor 5’ M 67.6 M 31.8 M 0.7 R 50.0 R 21.0 R 0.0 -79.0 -47.5 -2.5

Non-secretor 5 M 59.6 M 37.9 M 0.4 R 50.5 R 24.9 R 0.0 -75.2 -48.0 -1.5

0 total 10 M 63.6 M 34.9 M 0.6

M 39.6 R 14.5 -75.5

M 28.6 R 7.0 -59.5

M 34.1

M 69.0 R 128 R 66.0 -256 -74.0

M 71.2 R 128 R 61.0 -256 -83.0

M 70.1

TABLE 3b Differential counting and agglutination results with anti-(L8 + V (H 004) in blood group A,. M = mean, R = range

Differential counts (To) Secretor status n + + +

Agglutination

Titer Score ‘Pit acc’

Secretor 3 M 76.9 M 23.1 M 0.0 R 73.7 R 19.4 R -80.6 -26.3

Non-secretor 2 M 66.9 M 32.9 M 0.3 R 62.8 R 29.0 R 0.0 -71.0 -36.7 -0.5

A, total 5 M 72.9 M 27.0 M 0.1

M 54.2 M 69.0 R 6.2 R64 R 60.0 -84.0 -512 -81.0

M 30.2 M 64.0 R 12.0 R 128’ R 60.0 -48.4

M 44.2 M 67.0

I : 128 in both cases.

TABLE 3c Differential counting and agglutination results with onti-(L@ i Y) (H 004) in blood group 0 in relation to L$ status. M = mean, R = ranze

Differential counts (%) Agglutination Lewis (b) status n ‘Pit acc’ + i+ Titer Score

+ (POS) 3 M 65.6 M 33.3 M 1.1 M 41.7 M 69.0 R 50.0 R 23.5 R 0.3 R 14.5 R 128 R 66.0 -76.0 -47.5 -2.5 -75.5 -256 -74.0

- (neg) 7 M 64.8 M 28.1 M 2.3 M 43.6 M 70.0 R 50.5 R 21.0 R 0.0 R 7.0 R 128 R 61.0

-79 -40.0 -9.5 -74.0 -256 83

0 total 10 M 63.6 M 34.9 M 0.6 M 34.1 M 70.1

In our previous study with anti-A (A 003) (7) we found a clustering pattern on cells from mem- bers of an A, family which resembled closely that observed with anti-(Leb + Y) in the present study. We suggested that the A antigen on the erythrocytes of these individuals was characterized by a high percentage of difucosylated A (‘AY’ or

‘ALeY’) (28) chains. One may hypothesize, therefore, that the enzymes transferring the subtermi- nal fucose to type 2 chains act preferentially on lipid-bound H and A oligosaccharides (Figure 4) (29, 30).

With anti-B and anti-H, the ‘pit accumulation’ phenomenon might indicate that protein-carried


oligosaccharides vary less between cells than do lipid-carried ones. However, one would then not have expected ‘pit accumulation’ to occur with anti-(Leb+Y), since we have suggested that Y is bound mainly to lipid throughout the membrane. At present we therefore cannot explain this rather variable phenomenon.

In conclusion, this study has given indications that protein- and lipid-carried type 2 B and H antigens tend to be distributed differently on single erythrocytes. Y antigens, however, may be bound mainly to lipid throughout the membrane. The results thus add to the complexity of the ABO/H/Le antigen system by suggesting that differences may exist between the development

I

tdGal( l-O).iGIcNAc-R

dGal(1-4) dGlcNAc-R 3 1

crFuC

dGaI( 1 4 ) JGlcNAc-R 2

~ F u c

2 I nGaIUAC(1-3) liGbl(1-4)dGlcUAc-R

1 n FUC I

I 4 I

I ~rG~1(1-3)dG.1(~~4)11G1CNAC-R I 2 I 1 I ~ F U C I I I V

“V” ,rG.lNAc( 1-3)dGaI( 1 -4 ) JGICNAC.R

2 3 1 1

~ F Y C lrFuc

Figure 4. Biosynthesis of type 2 ABH and related structures. The genes which code for the transferases are underlined. X is formed from i and Y from H by action of the same enzyme. The probable development of ‘AY’ is shown by a dotted line. A similar scheme can be drawn for type 1 structures, sub- stituting H by Se, X by Le, X by Lea, Y by Leb and ‘AY’ by ‘ALeb’. Modified, in part, from (8).

of protein- and lipid-carried ABH/Le antigen- carrying oligosaccharides on erythrocytes. Fur- ther studies are needed to clarify the development and distribution of the various difucosylated type 2 chains. The interrelationship of the various antigens should be studied at the single cell level by double immunogold labelling (24, 31, 32).

Acknowledgments We are grateful to Karin Low and her colleagues at BioCarb AB, S-223 70 Lund, Sweden, and to Lisbeth Messeter, MD, Blood Centre, S-221 85 Lund, for gift of moabs and for their continuous interest in this work. Elisabeth Falleth, MT, provided excellent tech- nical assistance. This study was supported by a grant from the Norwegian Council for Science and the Hu- manities (NAVF).

References 1. Cohen F, Zuelzer WW. Interrelationship of the various

subgroups of the blood group A: Study with immunofluo- rescence. Transfusion 1965;5:223-9.

2. Dunstan RA, Simpson MB. Heterogeneity of antigen den- sity for major blood group systems on human erythrocytes. Transfusion 1984;24:418.

3. Matsukura Y. Demonstration of ferritin-labelled antibodies bound to human erythrocytes fixed with glutaraldehyde. Vox Sang 1972;22:549-53.

4. Williams MA, Voak D. Studies with ferritin-labelled Do/- ichos bifrorus lectin on the numbers and distribution of A sites on A, and A, erythrocytes, and on the nature of its specificity and enhancement by enzymes. Br J Haematol 1972;23:427-41.

5 . Reyes F, Gourdin MF, Cartron JP, Breton-Gorius J, Drey- fus B. The heterogeneity of erythrocyte antigen distribution in human normal phenotypes: an immunoelectron microscopy study. Br J Haematol 1976;34:613-21.

6. Romano EL, Stolinski C, Hughes-Jones NC. Distribution and mobility of the A, D and c antigens on human red cell membranes: Studies with a gold-labelled antiglobulin re- agent. Br J Haematol 1975;30:507-16.

7. Heier HE, Namork E, Falleth E. A study of immunogold labelled blood group A erythrocytes in the scanning electron microscope. Eur J Haematol 1988;41:17-24.

8. Oriol R, Le Pendu J , Mollicone R. Genetics of ABO, H, Lewis, X and related structures. Vox Sang 1986;51:161-71.

9. Samuelsson B, Breimer ME. ABH antigens: Some basic aspects. Transpl Proc 1987;21(6):4401-7.

10. Clausen H, Hakomori S. ABH and related histo-blood group antigens; Immunochemical differences in carrier iso- types and their distribution. Vox Sang 1989;56:1-20.

11. Lundblad A, Kabat EA. Immunochemical studies on blood groups. XLVIII. Activity of fucose-containing oligo-

234 HEIER & NAMORK

9. Samuelsson B, Breimer ME. ABH antigens: Some basic aspects. Transpl Proc 1987;21(6):4401-7.

10. Clausen H, Hakomori S. ABH and related histo-blood group antigens; Immunochemical differences in carrier iso- types and their distribution. Vox Sang 1989;56:1-20.

11. Lundblad A, Kabat EA. Immunochemical studies on blood groups. XLVIII. Activity of fucose-containing oligosaccharides isolated from human urine and from human ovarian cyst A and B blood group substances. J Immunol 197 1 ; 106(6): 1572-7.

12. Nava MT, Soligo D, Pozzoli E, Lambertenghi-Deliliers G. Backscattered electron imaging for immunogold viewing by scanning electron microscope. J Immunol Methods 1984;70:269-71.

13. de Harven E, Soligo D. Scanning electron microscopy of cell surface antigens labeled with colloidal gold. Am J Anatomy 1986; 175:277-87.

14. de Harven E, Leung R, Christensen H. A novel approach for scanning electron microscopy of colloidal gold-labelled cell surfaces. J Cell Biol 1984;99:53-7.

15. Chen HT, Kabat EA. Immunochemical studies on blood groups. The combining site specificities of mouse monoclonal hybridoma anti-A and anti-B. J Biol Chem 1985;260: 13208-17.

16. Messeter L, Low K. Personal communication. 17. Brodin T, Chester MA, Karlsson KA, Messeter L, Zopf D,

Lundblad A. A monoclonal antibody that recognizes both Leb and Y (Ley) antigens. Glycoconjugate J 1987;4:399- 406.

18. Reyes F, Lejonc JL, Gourdin MF, That HT, Breton- Gorius J . Human normoblast A antigen seen by immunoelectron microscopy. Nature (Lond) 1974;247:461-2.

19. van Ewijk E, Coffman RC, Weissman IL. Immunoelec- tron microscopy of cell surface antigens: a quantitative analysis of antibody binding after different fixation proto- cols. Histochemical J 1980;12:349-61.

20. van Ewijk E, van Soest PL, Verkerk A, Jongkind JF. Loss of antibody binding to prefixed cells: fixation parameters for immunocytochemistry. Histochemical J 1984; 16: 179- 93.

21. Clausen H, Levery SB, Dabelsteen E, Hakomori S. Blood Group ABH Antigens: A New Series of Blood Group A-Associated Structures (Genetic Regulation and Tissue Distribution). Transpl Proc 1987;19(6):4408-12.

22. Hanfland P. Isolation and purification of Lewis blood- group active glycosphingolipids from the plasma of human 0 Leb individuals. Eur J Biochem 1978;87:161-70.

23. Mollison PL. Blood transfusion in clinical medicine. 7th ed. Oxford/London/Edinburgh/Boston/Melbourne: Blackwell Scientific Publications 1983: 274-5.

24. Namork E, Heier HE, Falleth E. Double labelling of weak blood group A erythrocytes imaged with backscattered electrons in SEM. J Ultrastruct Molec Struct Res 1988;98:326.

25. Nakajima M, Ito N, Nishi K, Okamura Y, Hirota T. Immunogold labeling of blood-group antigens in human salivary glands using monoclonal antibodies and the strep- tavidin-biotin technique. Histochemistry 1987;87:539-43.

26. Schenkel-Brunner H. Blood-group-ABH antigens of human erythrocytes. Quantitative studies on the distribution of H antigenic sites among different classes of membrane components. Eur J Biochem 1980;104:529-34.

27. Greaves MF, Owen JJT, Raff MC. T and B lymphocytes: Origins, properties and roles in immune responses. Am- sterdam: Excerpta Medica/New York: American Elsevier Publishing Co., Inc. 1974.

28. Clausen H, McKibbin JM, Hakomori S. Monoclonal antibodies defining blood group variants with difucosyl type 1 chain (ALeb) and difucosyl type 2 chain (ALeY). Biochem- istry 1985;24:6191-4.

29. Kobata A, Ginsburg V. Uridine diphosphate-N-acety1-D- galactosamine: D-galactose a-3-N-acetyl-D- galactosami- nyltransferase, a product of the gene that determines blood type A in man. J Biol Chem 1970;245(6):1484-90.

30. Schwyzer M, Hill RL. Porcine A blood group-specific N- acetylgalactosaminyltransferase. J Biol Chem 1977;252(7): 2346-55.

31. Namork E, Heier HE, Falleth E. Double labelling of cell surface antigens imaged with backscattered electrons. J Electron Micr Techn 1987;6:87-8.

32. Namork E, Heier HE. Silver enhancement of gold probes (5-40nm): Single and double labeling of antigenic sites on cell surfaces imaged with backscattered electrons. J Elec- tron Micr Techn 1989:11:102-8.

Correspondence to: Hans Erik Heier, MD Department of Immunology National Institute of Public Health Geitmyrsveien 75 N-0462 Oslo 4 Norway

binding patterns of monoclonal anti-b, anti-h and anti-(leb+ y) on erythrocytes, imaged in the...

Documents