Vox Sang. 30: 81-90 (1976)

A Lymphocyte Immunogenetic System, Atri, Associated with the ABO Blood Group and the ABH Secretor System A. MARCELLI-BARGE, 3. C. POIRIER,A. BENAJAM and 3. DAUSSET Laboratoire d’Immuno-Htmatologie, INSERM U93, Centre Hayem, HGpital Saint-Louis, Paris

Abstract. The antigen Atri can be detected on the lymphocytes of 5.19% of A, ABH secretor individuals, The Atri substance is present in the plasma of the same individuals and can be fixed on group 0 erythrocytes. It is also found in the saliva of 30 out of 32 A non-secretor individuals. Its existence in the saliva of A secretor individuals, however, cannot be proved due to the presence of both the A and Atri antigens. The A and Atri antigens have been shown to be distinct on lymphocytes by capping and by blocking of the A sites. The fact that in several families the Atri antigen is not expressed by the parents shows that its expression on lymphocytes requires the intervention of at least one gene in addition to ABO and Se.

In 1972, two independent reports described non-HL-A lymphocytotoxic antibodies reacting against some, but not all, A lymphocytes. DAUSSET et al. [1973] showed that serum Atri and 11 others, all containing an immune anti-A, were active against the lymphocytes of 3.46% of A individuals. JEANNET et al. [1973] described another serum (W100) reacting against the lymphocytes of 8.44% of A individuals. The two sera [1973]. A third serum, seem to have a contrasting distribution DAUSSET RB, studied later by MAYRand MAYR[1974], behaved similarly, but reacted with 31.4% of A lymphocytes. It is possible, but not yet proven, that these three specificities are part of the same system. A preliminary communication MARCELLI-BARGE et al. [19741 showed that the Atri substance is secreted and can be absorbed by 0 erythrocytes. This study extends these observations and tries to give a general survey of this system, which is ossociated with ABO and probably with ABH secretor and another gene controlling its expression on lymphocytes.

Received: March 14, 1975; accepted: March 25, 1975.

82

MARCEUI-BAR~E/POIRIER/BENAJAM/DAUSSET

Material and Methods Antibodies. Most of the study was done using the original serum from Mrs. Atri, non-transfused mother of four children, The same specificity was detected in other sera, four from multiparous women, out of 2,030 tested, and seven from polytransfused dialysis patients. Four antibodies were also detected in sera from 480 healthy male group 0 donors. Population and family studies. A first series of 431 unrelated individuals, chosen at random, was typed for the antigen Atri as well as for ABO, rhesus, MNSs, P, Kell, Duffy, Lewis and HL-A. A second series of 559 selected A, or A, individuals was typed for Atri and for ABH secretor status as determined on saliva. 85 selected families were studied. The Atri antigen was present in ten of them, totalling 26 children. Serological methods: detection of the Atri antigen on lymphocytes. The Atri antigen was detected by the two-stage lymphocytotoxicity method [ M r r r ~ et . al., 19681 at 37 O C . Other serological methods, such as agglutination and lymphocyte complement fixation, were inactive. Neutralization of the anti-Atri antibody by saliva or the A-soluble substance. Three volumes of anti-Atri serum were added to two volumes of each geometric saliva dilution and the lymphocytotoxicity test was performed with the mixture. Soluble A substances from pigs’ stomachs (Pfizer) were used similarly. Absorption tests. Anti-Atri serum was absorbed by erythrocytes, lymphocytes and platelets. Erythrocytes and lymphocytes were obtained from defibrinated blood. Erythrocytes were first separated from lymphocytes by centrifugation on a FicollIsopaque gradient [HARRIS and UKAEJIOPO, 19701 and then practically all leucocytes were eliminated by several successive sedimentations on dextran (160,000 molec. wt). This technique ensures that even in the most unfavourable conditions the number of residual lymphocytes is only l/mm3. Lymphocytes were isolated on FicollIsopaque. Platelets were obtained by centrifugation of EDTA blood. 0.5 ml of anti-Atri serum or a dilution thereof was absorbed for 1h at 37OC by 1.1010 erythrocytes, 6.107 erythrocyte-free lymphocytes and 6.108 platelets, respectively. Blocking tests. DAUSSET and LEGRAND’S[1971] technique was used to block Atripositive and Atri-negative lymphocytes with immune anti-A. 6.106 lymphocytes were incubated three times for 1 h at 37 OC with 0.2ml of anti-A. After one washing, 0.2 ml of immune anti-Atri was added for 1h at 37 O C . The supernatants were tested for remaining anti-Atri activity. Redistribution of A and Atri sites on lymphocytes. Redistribution tests were done using a modification of the technique of BERNOCO et al. [1973]. 5.100 washed Atri-positive lymphocytes, suspended in 200 p l of Eagle’s minimum essential medium, were sensitized with an equal volume of undiluted immune anti-A serum, titrating 1512 in AB serum. After 2 h of incubation at room temperature, the cells were washed, resuspended in 0.05 ml of an anti-human IgG, marked with fluorescein isothiocyanate (FITC),and incubated for 30 min at 0 O C . In order to ensure that all the A antigens were capped and that none were widespread on the cell surface, the lymphocytes were washed three times and incubated again for 2 h at 0 OC with the same immune anti-A and subsequently with an anti-human IgG labelled with tetra-

A Lymphocyte Immunogenetic System

83

methylrhodamine isothiocyanate (TRITC). Checking was performed by examination with a Zeiss standard GFL fluorescence microscope. Anti-Atri was then added to determine the presence or absence of the Atri determinant by the dye exclusion lymphocytotoxicity technique. Controls, consisting of Atri-negative lymphocytes, were tested at the same time. Fixation of the Atri substance on group 0 eryrhrocyres. 0.10ml of packed erythrocytes were added to 1ml of Atri-positive or Atri-negative heparinized plasma containing 0.10 ml of 2% streptomycin. The mixture was incubated a t 37 OC for 48 h, the plasma being renewed every 12 h. The erythrocytes were washed once in saline and the agglutinating ability was checked after incubation for 1h a t room temperature with an anti-Atri o r an anti-A and an anti-Lea.

Results Characteristics of the antibody. The lymphocytotoxic antibody usually has a low titre (1:l to 1:4), It is more active at 37 "Cthan at room temperature, at which it is always weak and sometimes undetectable. All the anti-Atri sera studied possessed an immune anti-A. Four of the 480 normal 0 sera studied possessed both an anti-Atri and an immune anti-A, and 64 possessed an immune anti-A only. Filtration of three different sera on Sephadex G-200 showed that antiAtri activity occurs in the IgM only and anti-A in both the IgG and the IgM. Anti-Atri activity can be destroyed by heating the sera to 70 "C for 10 min. It is possible to absorb anti-Atri activity on all A, or A, erythrocytes, irrespective of their Lewis status or of the Atri status of the corresponding lymphocytes. Anti-Atri activity is also absorbed by the platelets of all A individuals, but only by Atri-positive lymphocytes (6.107/0.5ml of serum). It is also completely neutralized by the soluble A substance (dilution 1:2). Secretion of the Atri substance in saliva. The anti-Atri antibody can easily be neutralized by the saliva of ABH secretors, even at a dilution of 1:128 or 1:256. The saliva of non-secretors is also able to neutralize anti-Atri activity, but at a much lower dilution. Great variations in titre were observed. In only three cases was there an inhibition over 1:32 and in two cases there was no inhibition at all. These quantitative variations are independent of the A, or A, and Lewis status of the subjects (table I). By means of a technique similar to that used to fix the Lewis plasma substance on erythrocytes, it was possible to fix specifically the Atri substance present in plasma onto 0 erythrocytes. These erythrocytes are

a4

MARCELLI-BARGE/POIRRfBENAJAdDAUSSET

Table I. Relationship between the Atri factor (salivary and/or lymphocyte) and the ABO and ABH secretor systems

ABO and AF3H secretor type

Anti-Atri inhibition by saliva (titre)

167l

A, (or A,) Se A, (or A,) se 0 Se or se

3

Presence of Atri on lymphocytes

11

10

17

2 10

0 0

Number of unrelated individuals observed in each category. -

Table II. Fixation of the Atri plasmatic substance on 0 erythrocytes

Number of cases

Plasma from individuals of group

Agglutination of 0 erythrocytes1 by Anti-Lea

A, Le(a - b -) Atri A, Le(a - b +) Atri

+ +

Anti-A (titre 1:512)

Anti-Atri

++ ++

A, Le(a - b -) A, Le(a+b-) A, Le(a - b +) 0 Le(a-b-) 0 Le(a+b-) 0 Le(a-b+)

Ten different varieties of 0 erythrocytes were used.

then able to be agglutinated specifically by an anti-Atri antibody, but not by an anti-A from B individuals, even one which is strong and immune, nor by an anti-Lea when the plasma was from a Lea-negative individuals (table 11). Differentiation of A and Atri sites on lymphocytes. Blocking lymphocyte A sites using an immune anti-A showed that A and Atri antigenic sites are distinct from one another. Atri-positive lymphocytes were saturated with a high-titre (1:256) immune anti-A. The anti-Atri was then absorbed onto these coated lymphocytes. Anti-Atri activity disappeared

A Lymphocyte Immunogenetic System

85

Table 111. Anti-Rtri absorption on Atri-positive and Atri-negative lymphocytes after blocking of the A lymphocyte sites Anti-Atri lymphocytotoxicity on Atri-positive lymphocytes after absorption by A, Atri-negative lymphocytes Al Atri-positive lymphocytes

-__

unblocked

blocked' by an immune anti-A

unblocked

blocked' by an immune anti-A

+1:8

+1:8

-

-

The titre of the anti-A contained in the anti-Atri serum remained intact, showing that the A sites were blocked correctly.

Table I V . Association between Se status and the presence of the antigen Atri on lymphocytes

Se se

Atri +lymphocyte

Atri -lymphocyte

29 0

435 95

p10.01.

Table V. A comparative study of the Atri, WlOO and RB specificities on a panel of 87 individuals Number of individuals

Atri

Wloo

RB

0 0

2 0 71 5

2 7

+

87

15

MARCELLI-BARGWPOIIUERIER/BENAJAM/DAUSSET

86

Table VZ. Segregation of A and Atri reactivity, family Duv. [family completed from DAUSSET and LEGRAND, 19711

n

n Atri . a W28. XiW32. HL-A12 b

U T U

w

Atri c W31. HL-A13/HL.A11.5 d

R, r MSIMS P,

ATRl HL.A

kk

kk

Fy(a-b+) Lela-b+) SW

Fy(a-M) Lela-b+l

Sere

+

+

+

ac

ac

ac

A, 0 R, R, MSlMr

A,O

A,O R2R, MSlNS

R,r MSIMs

bd A,O R,R, MSIMr

bd

A,O R,r MSIMs

bd A10 RzRI MSlMr

ad AtO RIRI MSlNS

from the supernatant only where the lymphocytes were Atri-positive; where they were Atri-negative, anti-Atri activity remained unchanged. The associated anti-A antibody also contained in the Atri serum remained at its original titre (table 111). Indirect immunofluorescence studies gave similar results. Large spots of fluorescence were visible, corresponding to the FITC labelling of the A sites on the membrane. Apparent complete redistribution of the A antigenic sites was obtained, since no rhodamine labelling was seen outside the green A spots in the presence of another anti-A revealed by TRITC anti-human Ig. However, lymphocytes remained sensitive towards the four different anti-Atri sera used in lymphocytotoxicity. They also remained sensitive to an anti-HL-A antibody. Population studies. In a random population, Atri-positive lymphocytes were found only among A (A, or A,) individuals (5.19%, i.e. 2.33% of all individuals, irrespective of their ABO group, in the French population). It was also noted that all subjects with Atri-positive lymphocytes are ABH secretors (table IV). The significance of this association is p < 0.01. No positive or negative associations with the following specificities were found: MNSs, Rhesus, P, Kell, Duffy, HL-A. A comparison of the reactivities of the three sera WlOO, RB and Atri, on a panel of 87

A Lymphocyte Immunogenetic System

87

Table V11. Family Sag.

Atri + HL-AZ, W27/HL-All. W5 A, 0

-0

Atri

-

H L - A Not tested

w ri

NdNr

MslMs

PI

P, kk kpla-h+) Fylatl Lu(a-bt1 Jkla.b+l Lehib-I rese

kk kplabtl Fy(a+l Lula-btl Jklattrel Le(a-I Sese

__’

Atri cHL-A10(W26).W18 H L A l l . i h S d

Atri a HL-AlOlW261, W21iHL-AZ. HL-A12 b

W

A, 0 rr

R2 r MSiMn

MsINr P, kk kpla-b+) Fyia+b+l Lula.btI Jkla-btl Lelatl rese

Atri + HL-A A, 0

p* P I kk kpla-ht) Fylatbtl Lula-btl Jkla+b+l Le(a I Se

Atri + .3C

Atri t bd

rr

MdNs p, p, Fyla-b+l kk kpla-bt) Lula-b+l Jk(atbt1 Lela 1 Sere

P, p, Fyia+b-1 kk kpla-br) Luia.b+i Jkia+b+l Lela-) Sese

p, p, Fy(a-b+)

kk kpla-b+) Luls-btJ Jk(atbt1 Lela-I Sese

unrelated group A individuals, showed the absence of individuals possessing all three specificities (table V). Family studies. In the ten families in which the Atri specificity was found, 17 of the 26 children were both A and Atri-positive. Segregation does not, however, follow the A gene, and is independent of Rhesus, MNSs, and HL-A. Two of the families were especially interesting because of the absence of the antigen Atri from the lymphocytes of both parents (tables VI, VII), one of which is A and the other 0. In family Duv. the

88

MARCELLI-BARCE/POIRIER/BENAJAM/DAUSSET

mother is A, Se and it is impossible to know whether it was she who transmitted the Atri gene. On the other hand, in family Sag. the mother, A, Se, probably received the Atri gene from her father.

Discussion The specificities Atri, WlOO and RB could belong to the same immunogenic system, provisionally known as Atri. All three specificities share the characteristic of being present on a certain percentage of A lymphocytes, 5.19, 8.44 and 31.4%, respectively. The distribution of these antigens in a panel of 87 unrelated individuals of blood group A is compatible with the hypothesis that they are three alleles. The relationship with the ABN secretor system is very close. All Atripositive lymphocytes belong to ABH secretor individuals (p < 0.01). Similarly, all RB-positive individuals are ABH secretors. A few exceptions were noted for WlOO individuals. Moreover, the Atri system appears to be a secretory system. The saliva of A ABH secretors neutralizes anti-Atri at a very high titre. The saliva of non-A ABH secretors also has a neutralizing effect, but at a lower titre, and in two cases out of 32 this effect was apparently absent. Similarly, the Atri substance is only found in the plasma of those individuals who possess this antigen on their lymphocytes. The other A individuals do not possess this inhibiting action in their plasma, even if they do in their saliva. In this connection, it is possible to put forward the hypothesis that true expression on lymphocytes does not exist, but is due to absorption of the plasmatic substance. This possibility cannot be completely ruled out, although positive lymphocytes are not rendered insensitive to antibody action after up to 10 washings. Thus, at the current stage of our knowledge it may be said that Apositive ABH secretor subjects may be subdivided into two according to whether the Atri substance is found on their lymphocytes and in their plasma, or only in their saliva. A-positive non-ABH secretor individuals have a smaller amount of the Atri substance in their saliva. A few people lack this substance altogether. The Atri substance has never yet been found on the lymphocytes or in the plasma or saliva of 0 or B individuals. The relationship of this system with ABO should be discussed in greater detail. Anti-Atri, as well as anti-RB, is absorbed by all A, or A,

A Lymphocyte Immunogenetic System

89

RBC and platelets, whether or not the corresponding lymphocytes are Atri-positive. Similarly, the Atri substance is neutralized by the A substance found in pigs’ stomachs. The most simple explanation is that there is a cross-reaction between the A and Atri determinants. Nevertheless, although anti-Atri could be fixed on the antigen A, the converse does not seem to be true, since blocking the A sites with an anti-A does not prevent the further fixation of an anti-Atri (table 111). A unidirectional cross-reaction‘ could thus account for the observed facts. If this is the case, it is not possible to prove that the Atri substance is in fact present in the saliva of A secretor individuals, although this is possible for A non-secretors. Blocking and immunofluorescence tests seem to indicate that the Atri and A determinants are separate on the lymphocyte membrane, which does not exclude the existence of a unidirectional cross-reaction. An important argument in favour of the individuality of the A and Atri deter[personal commun., 19741 studies on minants is provided by FELLOUS’ fibroblasts in culture. He found out that whereas after three or four passages in vitro the A determinant is no longer expressed, the Atri determinant still remains. On a genetic basis, it has now been established that the Atri system is functionally associated with the ABO and secretor systems. In order to express the antigen Atri, and probably the antigens WlOO and RB on one’s lymphocytes, it is necessary to have the blood group A and to be an ABH secretor, but this in itself is not sufficient. The fact that in several families the antigen is not expressed in the parents shows that its expression on lymphocytes requires the intervention of at least one gene in addition to ABO and Se. It is difficult to disprove the intervention of the Lewis gene because of the well-known difficulty of determining the Leb type of A, erythrocytes. However, it is nevertheless possible that the Se and the Le genes have to be involved simultaneously. Using the best anti-Leb sera, both RUTHSANGERand ourselves found at least one A Le(a-b-) Atri-positive individual, which seems to indicate that the Le gene is not a necessary intermediary. If it is not involved, Atri is definitely distinct from the Siedler factor, present on A Le(a-b+) erythrocytes, and from the Magard factor, present on A Le(a-b-) erythrocytes. In any case, conversely to 1 Similar examples of unidirectional cross-reactions are well known in the HL-A system [COLOMBANI et fll., 19721.

90

MARCELLI-BARGE/POIRIER/BENAJAM/DAUSSET

both these factors, Atri is expressed on all A Le(a-b+) or Le(a-b-) erythrocytes. Only specifically inhibiting biochemical tests using polysaccharides of known composition will be able to clarify the relationships existing between the A molecules and those of this new system. Acknowledgements We should like to thank Drs. J. Y.M ~ E R J., P. CAGNARD, P. GUICHOUX, HOSPICE from the Centre National de Transfusion Sanguine and Drs. I. REVIRON and J. F. ROBERTfrom the Blood Bank of the HBpital St-Louis for their generous and invaluable help in collecting blood and saliva samples. We should also like to thank Miss C. DEHAY,Mrs. M. MARCEAU and Mrs. J. ROULLIER for their technical assistance, and Miss A. M. SOMERSET for the preparation of the English manuscript. A. GENTILand P.

References BERNOCO, D.; CULLEN, S.; SCUDELLER, G.; TRINCHIERI, G., and CEPPELLINI, R.: HL-A molecules at the cell surface. Histocompatibility testing 1972, p. 527 (Munksgaard, Copenhagen 1973). COLOMBANI, J.; COLOMBANI, M.; DEGOS,L.; TERRIER, E.; GAUDY,E., and DASTOT, H.: Effect of cross reactions on HL-A antigen immunogenicity. Tissue Antigens 4: 136 (1 974). DAUSSET, J.; COLOMBANI, J.; LEGRAND, L.; LEPAGE, V.; MARCELLI-BARGE, A., and DEHAY, C.: Population and family studies in a French population with special reference to non-HL-A antibodies. Histocompatibility testing 1972, p. 107 (Munksgaard, Copenhagen 1973). DAUSSET, J. et LEGRAND, L. : La structure anti&ne du systhme HL-A Human genetics. Roc. 4th Int. Congr. of Human Genetics, Paris. Excepta med. 150: 331 (1971). HARRIS,R. and UKAEJIOFO, E. 0.: Tissue typing using a routine one-step lymphocyte separation procedure. Br. J. Haemat. 18: 229 (1970). JEANNET, M.; BODMER, J. G.; BODMER, W. F., and SCHAPJRA, M.: Lymphocytotoxic sera associated with the ABO and Lewis red blood cell groups. Histocompatibilitytesting, p. 493 (Munksgaard, Copenhagen 1973). MARCELLI-BARGE, A.; POIRIER, J. C.; BENAJAM, A.; TILZ,G. P. aud DAUSSET, J. : Skr6tion et absorption sur hkmaties 0 de la substanceAtri. C. r. Stanc. SOC.Biol. 168: 199(1974). MAYR,W.R. and MAYR,D.: A lymphocytotoxic antibody associated with ABO blood groups and ABH secretor status. Preliminary report. J. Immunogenet. I: 49 (1974). M ~ A LK. , K.; MICKEY,M. R.; SINGAL,D. P., and TRASAKI, P. I.: Refinement of microdroplet lymphocyte cytotoxicity test. Transplantation 6: 916 (1968). Dr. A. MARCELLI-BARGE, Laboratoire d’Immuno-H6matologie. INSERM U93, Centre Hayem, Hdpital Saint-Louis, 2, place du Dr Fournier, F-75474Paris CPdex 10 (France)