Eur. J. Immunol. 1991. 21: 715-719

Isoforms of the E2 molecule

715

Catherine Gelin, Didier Zoccola, HBlkne Valentin, Brigitte Raynal and Alain Bernard

Isoforms of the E2 molecule: D44 monoclonal antibody defines an epitope on E2 and reacts differentially with T cell subsets*

Laboratoire d’lmmunology des Tumeurs de I’Enfant, Institut Gustave-Roussy, Villejuif

Human T cell rosetting with erythrocytes is clearly dependent on the CD2-CD58 interaction. We have previously demonstrated that other T cell molecules are involved in the rosette phenomenon, including the E2 molecule, a 32-kDa transmembrane glycoprotein. In this report we show that the D44 monoclonal antibody (mAb), previously shown to subdivideT cells into subpopulations with distinct functional repertoires and to identify 70% of lymphocytes from bronchoalveolar lavage from HIV+ patients, defined a new epitope on the E2 molecule. This was illustrated by the reactivity of the D44 mAb in Western blot experiments performed with the immunoaffinity purified E2 molecule. Moreover, doublelabeling experiments showed that the E2 molecule exhibited varying epitopes when expressed in different cell types. The D44 and 12E7 epitopes were restricted to distinct subpopulations of T cells and, more importantly, the D44 expression was limited to the CD29+ population including the memory subset.The 0662 and L129 epitopes are present on all T cells.Thus, the E2 molecule has both common and variable epitopes in its extracellular domain, and only the common epitopes seem to be involved in Tcell adhesion processes.

1 Introduction Human T cell rosetting with erythrocytes has been very useful to identify humanT lymphocytes. More recently, this phenomenon proved to be particularly informative to identify important T cell surface molecules involved in T cell functions and Tcell adhesion processes as exemplified by the T cell molecule CD2 [l-31 and its ligand CD58 (LFA-3) [4,5]. We have demonstrated that other T cell surface molecules are involved in the lymphocyte-erythrocyte adhesion, including the T cell homing receptor CD44 [6,7], and recently theT cell molecule CD59 [8,9]. In this context, we have previously identified a T cell surface antigen, termed E2, which is involved in rosette formation [lo]. E2 is a highly glycosylated protein with no N-linked sugar residues; all sugar residues are O-linked and account for 14 kDa in the molecular mass. An intracellular precursor protein had been identified as a 28-kDa species which corresponds to the unsialylated, fully glycosylated protein [ll].The amino acid sequence has revealed that E2 is a proline-rich protein which displays an organization typical of an integral membrane molecule [12]. Finally, given its wide cellular distribution, E2 might be involved in T cell adhesion to many other cell types in addition to erythrocytes. Furthermore, we have described an mAb, termed D44, which reacts with cortical thymocytes and, to a lesser

extent, with medullary cells and peripheral blood T cells. This mAb is able to remove, in the presence of C, both NK activity and alloreactive T cell cytotoxicity without affecting most of the mixed lymphocyte reactivity and PHA proliferative responses [13-151. This mAb recognizes a 28-32-kDa molecule. We demonstrate here that D44 defines a fourth epitope on the E2 molecule, different from the three epitopes already identified, i.e. the 0662, L129 and 12E7 epitopes. Moreover, we show the differential expression of those epitopes on different functional subsets of T cells, and their differential involvement in rosette formation.

2 Materials and methods 2.1 Cells Human thymi were obtained from children undergoing cardiac surgery; single thymocyte suspensions were prepared by gently teasing cells from thymic samples. BM cells were obtained from normal volunteers for the purpose of allogeneic BM graft. Purified T lymphocytes were obtained from healthy donors by leukophoresis and prepared as previously described [3]. Sheep and human erythrocytes were prepared by removing the buffy coat after centrifugation and washing three times in PBS. All cell lines were grown in RPMI 1640 supplemented with 10% FCS, 100 U/ml of penicillin and 50 pg/ml of streptomycin.

[I 85321

*

This work was supported by grants from Association pour la Recherche Medicale, Institut Gustave Roussy (CRC No. 87D6). Institut National de la SantC et de la Recherche Mtdicale (CRE No. 873005) and Electricit6 de France (No.BOOL18).

Correspondence: Catherine Gelin, INSERM U 93, Centre Hayem, HBpital Saint-Louis, 1 av. Claude Vellefaux, F-75475 Pans Cedex 10, France 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1991

2.2 mAb

The 0662 and L129 mAb were prepared as previously described [ll].The 12E7 mAb was kindly provided by Dr. R. Levy [16] and the 2H4 (CD45R) mAb by S. E Schlossman [17]. The N4 [18], 0275 (CD2), D44 [13] and K20 (CD29) [19] mAb were obtained in our laboratory.The OO14-2980/91/0303-07 15$3.50+ .25/0

Eur. J. Immunol. 1991. 21: 715-719

C. Gelin, D. Zoccola, H.Valentin et al.

716

directly PE-conjugated anti-CD4 (Leu-3) and anti-CD8 (Leu-2) mAb, and FITC-conjugated anti-CD45 (Hle 1) mAb were purchased from Becton Dickinson (Mountain View, CA). 2.3 lmmunofluorescence assays

In double-labeling immunofluorescence experiments, purified T cells (2 x lo6cells) were incubated for 30 min with a first-step mAb. After two washes with PBS 0.2% sodium azide, cells preincubated with 2H4 or K20 mAb were stained for 30 min with F(ab’)2goat anti-mouse IgG FITC conjugate, while cells preincubated with biotinylated mAb were stained with FITC-conjugated avidin. After two washes, cells were incubated for 20 min with excess mouse IgG to block free goat anti-mouse IgG-binding sites. In the second-step labeling, cells were incubated for 30 min either with the directly PE-conjugated mAb CD4 or CD8, or with biotin-conjugated mAb. In the latter case, PE-conjugated avidin was added for 30 min after two washes.

2.5 Biochemical analysis The E2 molecule was purified from thymocytes as described elsewhere [ 1 ll. Western blotting experiments were performed as described [ 111. Briefly, purified E2 molecule was electrophoresed on SDS-PAGE under reducing conditions and then electrophoretically transferred to nitrocellulose. Nitrocellulose sheets were next incubated with mAb (10 kg/ml) and 1251-labeledF(ab’)2 anti-mouse Ig fragment before autoradiography.

3 Results 3.1 The D44 mAb defines an epitope on the E2 molecule, which is not involved in rosette formation

In view of similarities in mol. mass and cellular distribution of E2 and D44 antigen, the question of their relationships was raised.The E2 molecule was purified from thymocytes by immunoaffinity using E2 mAb covalently linked to protein A-Sepharose CL4B, as described elsewhere [ 111. In competitive binding experiments, thymocytes were The purified material was then subjected to Western blot incubated with the first mAb (full range of dilution was analysis. Fig. 1 shows that D44 mAb reacted with the investigated, up to 1 mg/ml) for 30 min. Then the second purified E2 molecule to the same extent as did the 0662 antibody, coupled to biotin, was added and the fixation mAb. Similar results were obtained by using immunoprerevealed with FITC-conjugated avidin as described [lo]. cipitates from thymocytes with the 0662 mAb. All incubations and washes were performed at 4°C. Cells were analyzed by using FCM (Ortho-Becton Dickinson, Thus, those results demonstrated that D44 mAb recognizes the purified E2 molecule. The next question was whether System 50, Mountain View, CA). the D44 mAb defined a peculiar eptiope on E2, distinct from the three epitopes already described. Competitive binding experiments were performed and typical results, 2.4 T cell E binding (rosette) assay presented in Fig. 2, showed that D44 mAb did not compete Rosette formation and rosette inhibition assays have been for binding with any E2 mAb tested (0662, L129 or 12E7). previously described [ 101. Rosettes were counted as thymo- It must be noted that 12E7 mAb partially inhibited cytes tightly bound to at least three red cells and 300 subsequent binding of D44 mAb, and two hypotheses could mononuclear cells were counted in each test. Each experi- have been raised to explain this result: (a) 12E7 and D44 mAb define the same epitope on the E2 molecule and the ment was performed at least five times. results obtained could be explained by a difference of binding affinity between the two mAb, and (b) 12E7 and D44 mAb define two different epitopes.

32

-

Double-labeling experiments carried out on purified Tcells enable distinction between these two hypotheses. Results of a typical experiment are shown in Fig. 3, and indicate that the vast majority of purified Tcells carried either the 12E7 epitope or the D44 epitope (Fig. 3), only 10% of T-PBL being 12E7+D44+. Most cells (70%) displayed only 12E7 while 5% T-PBL displayed only D44. Thus, these double-labeling experiments, demonstrating an exclusive expression of 12E7 and D44 epitopes on some T cell subpopulations, ruled out the first hypotheses and indicated that D44 mAb recognized a distinct epitope on a subset of E2 molecules. Figurel. The 12E7, L129, 0662 and D44 mAb recognize the same 32-kDa protein. Western blot analysis of the purified E2 molecule probed with 0662 mAb (lane l ) , L129 mAb (lane 2), 12E7 mAb (lane 3), 0275 mAb (lane 4) and D44 mAb (lane 5).

As E2 has been first identified as a Tcell surface antigen involved in rosette formation [lo], we looked for the involvement of the D44 epitope in this phenomenon. The blocking effect of E2 mAb was tested on T cell rosettes made with sheep and human erythrocytes. As demonstrated inTable 1, the D44 mAb was unable to block rosette formation with any type of erythrocytes, whatever the amount of antibody added, in contrast to the 0662 and L129 mAb which strongly inhibited rosettes. Thus, among

Eur. J. Immunol. 1991. 21: 715-719 OImZ b

Isoforms of the E2 molecule Ll2Ob

IZE7b

717

044b

N4

Figure 2. The D44 mAb define a fourth epitope on the E2 molecule. Fluorescence histograms of competitive binding experiments between 12E7, L129, 0662 and D44 mAb. Target cells were thymocytes. The second antibody (X-axis) was coupled to biotin, as indicated by “ b and revealed by fluoresceinated avidin. A full range of first antibody (Y-axis) dilution was investigated, up to 1 mg/ml.

oM.2

LllO

lPL7

044

the four epitopes now defined on the E2 molecule - 0662, L129, 12E7 and D44 - only two seem to be involved in rosettes, pinpointing characteristics domains on the E2 molecule. 3.2 The D44 epitope expression is restricted to the CD29+ T cell subset

(10%and 5% of the wholeTcel1 population, respectively). This result confirmed our previous results obtained by cytotoxicity [15] which have shown that D44 was expressed on a limited CD4+ and CD8+ subpopulations. The 12E7 expression was also restricted to a Tcell subpopulation; it was present on 80% of CD4+ cells and 70% of the CD8+ cells (50% and 20% of the whole T cell population, respectively).

The expression of the E2 epitopes onT cells vanes with the type of cell tested. Thus, while they were all expressed on 95% of thymocytes, D44 was restricted to a subpopulation of 15% of peripheral Tcells and 12E7 only expressed on 80%of such cells. It must be noted that among the cell lines tested, only two major phenotypes could be defined: 0662+L129+12E7+DMf and 0662+L129+12E7+D44(data not shown), and we wondered which T cell subsets express the D44 and 12E7 epitopes. Two-color FCM analyses were performed to assess the expression of the restricted E2 epitopes on the Tcell subpopulations. Results summarized in Fig. 5 revealed that while 0662, L129 were present on all CD4+ and CD8+ cells, the D44 epitope was expressed on 15% of the CD4+ cells and 15% of the CD8+

As the T lymphocyte subpopulations can be further subdivided into distinct functional subsets based on the reactivities with various mAb including 2H4 (CD45R) (201 and K20 (CD29) [21], a second series of double-labeling experiments were performed to investigate the expression of the D44 and 12E7 epitopes among the CD29+ and CD45R+ Tcell subsets. It can be seen from Fig. 4 that the 12E7 epitope was present on the CD29+ and CD45R+ subsets but not on all such cells: its expression was limited to 85% of the CD29+subset and 90% of the CD45R+cells. By contrast, the D44 epitope expression was restricted to the CD29+ subpopulation: no D44+CD45Rf cells could be detected, while 19% of the CD29 subset (15y0 of the whole Tcell population) are stained with the D44 mAb.

Table 1. Ability of E2 mAb to inhibit spontaneousTcel1 rosettes with sheep or human erythrocytes

Antibody added to rosette assay’)

D44

7011

10%

Cells used in rosette assays

rnYmocYt=+ sheep E None (control) 0662 L129 12E7 CD2 (0275)

I

Thymocytes + human E

1W)

55

43 21 85 0 100

30 14 51 2 50

a) mAb were used at 50 pg/ml. D44 and 12E7 used at a concentration range from 0 to 1 mg/ml produced no inhibition. b) Data are expressed as % Tcells binding E and represent results obtained in a typical experiment. Five independent experiments gave identical results.

044

Figure3. fro-color FCM analysis of purified T cells for D44 (X-axis) vs. 12E7 (Y-axis). Biotinylated D44 mAb was revealed by FlTC-conjugated avidin, and the 12E7 mAb was directly conjugated to PE.

718

Eur. J. Immunol. 1991. 21: 715-719

C. Gelin, D. Zoccola, H.Valentin et al.

shown by Plata et al. [23] that 65%-87% of the lymphocytes found in the broncho-alveolar lavage from HIV+ patients were double positives for the CD8 and D44 markers, and accounted for the alveolar CTL activities. Nevertheless, it will be interesting to find out whether the CD29+D44+ and CD29+D44- subpopulations represent different stages in the Tcell activation pathway or separate sublineages of T cells.

The absence of the D44 epitope from cell surface of someT cell subsets could be due t o different sugar residues on the E2 molecule present on D44- or D44+ cell populations.We

12E7

I

04 4

Figure 4. Concordance of Tcell subsets identified by staining with CD29 or CD45R and E2 mAb. Two-color FCM contour plot of purified Tcells comparing CD29 (K20) or CD45R (2H4) staining (Y-axis) with 12E7 or D44 staining (X-axis).The CD29 (K20) and CD45R (2H4) mAb were ascitis and revealed by F(ab’)z goat anti-mouse IgG FITC-conjugate. The D44 and 12E7 mAb were used as biotinylated mAb and revealed by PE-conjugated avidin.

CD45 CD45R (2H4)

have previously determined that the LD129 epitope was masked by sialic acid residues on erythrocytes from Xg(a+) women [ 121 and we have tested if it was also the case for the D44 epitope o n T cells. In fact, no difference in the D44 epitope expression was noted between neuraminidasetreated or untreated T cells. Thus, the differential expression of D44 might be due t o a difference in the proteic core of the molecule resulting from a differential splicing in the E2 gene. We have checked that it was impossible to distinguish mRNA and their products from cell lines exhibiting either the D44+ or D44- phenotype, solely based on their electrophoreitc mobilities (data not shown). Consequently, if differential splicing occurs it must involve exons with very similar sizes.

The generation of multiple structure of mRNA and proteins by differential splicing is not uncommon among eukaryotic genes [24], but it is interesting to note that if the hypothesis of a differential splicing is proved for E2, the E2 gene would be, like the CD45 gene [22],very unusual in that the diversity occurs in the extracellular domain.

CD29 (K2O)

E2RB (1 2E7)

We wish to thank Michel Samson for PE- and FITC-coupling mAb, Dr. Laurence Boumsell and Dr. Stephen Shaw for very helpful discussion and Dr. Johanna Kubar for critical reading of the manuscript.

€2 (0662, L129)

Received May 14, 1990; in revised form November 9, 1990.

EZRA (D44)

Figure 5. T cell subsets definition according to the E2 epitope expression.The data presented in this figure summarize the results obtained in two-color FCM analysis for the expression of E2 epitopes within either the CD4+ and CD8+ or the CD45R+ or CD29+ T cell subsets. The D44 and 12E7 epitopes were noted, respectively, E2RA and E2RB for E2-restricted epitopes.

4 Discussion Here we demonstrate that the four E2 epitopes are differentially involved in rosette formation, and differentially expressed o n T cells. Results obtained from doublelabeling experiments permit us t o further subdivide the CD29+ population into two distinct subsets depending on the expression of the D44 epitope, as represented in Fig. 5. It appears that, like CD45 [22], the E2 molecule has both common and variable epitopes in its extracellular domain. The 0 6 6 2 and L129 epitopes are expressed on the wholeT cell population; by contrast, the D44 and 12E7 epitopes are restricted t o T cell subsets, the D44 expression being exclusive with the CD45Rexpression as it is only present on a subpopulation of the CD29+ subset.We already know that the D44 epitope expression within the CD8 subpopulation is restricted to the cytotoxic subset [15] and it has been

5 References 1 Kamoun, M . , Martin, F? J., Hansen, J. A., Brown, M. A., Siadak, A. W. and Nowinski, R. C., J. Exp. Med. 1981. 153: 207. 2 Meuer, S. C., Hussey, R. E., Fabbi, M., Fox, D., Acuto, 0.. Fitzgerald, K. A., Hodgon, J. C., Potentis, J. E, Schlossman, S. E and Reinherz, E. L., Cell 1984. 36: 897. 3 Huet, S., Wakasugi, H., Sterkers, G . , Gilmour, J., Boumsell, L.and Bernard, A . , J. Immunol. 1986. 137: 1420. 4 Shaw, S . , Luce, G. E. G . ,Quinones, R., Gress, R. E., Springer, T. A. and Sandres, M. E., Nature 1986. 323: 262. 5 Vollger, L.W.,Tuck, D.T., Springer,T. A., Haynes, B. F. and Singer, K . H., J. Immunol. 1987. 138: 358. 6 Huet, S . , Groux, H., Caillou, B.,Valentin, H . , Prieur, A. M . and Bernard, A., J. Immunol. 1989. 143: 798. 7 Jalkanen, S., Reichert, R. A., Gal1atin.W. M . , Bargatze, R. F., Weissman, I. L. and Butcher, E. C., Immunol. Rev. 1986. 91: 39. 8 Stefanova, I., Hilgert, I . , Kristofova, H., Brown, R . . Low, M. G . and Horejsi,V, Mol. Immunol. 1989. 26: 153. 9 Groux, H., Huet, S., Aubrit, F., CongTran. H., Boumsell, L. and Bernard, B., J. Immunol. 1989. 142: 3013. 10 Bernard, A., Aubrit, E , Raynal, B., Pham, D. and Boumsell, L., J. Immunol. 1988. 140: 1802.

Eur. J. Immunol. 1991. 21: 715-719 11 Aubrit, F., Gelin, C., Pham, D., Raynal, B. and Bernard, A . , Eur, J. Immunol. 1989. 19: 1431. 12 Gelin, C., Aubrit, F., Phalipon, A., Raynal, B., Cole, S., Kackzorek, M. and Bernard, A,, EMBO J. 1989. 8: 3253. 13 Bernard, A . , Gay-Bellile,V, Amiot, A., Caillou, B., Chrbord, P. and Boumsell, L., J. Immunol. 1984. 132: 2338. 14 Calvo, C. F., Boumsell, L., Laffy, B., Bernard, A . and Senik, A , , J. Immunol. 1984. 132: 2345. 15 Calvo, C. F., Bernard, A . , Huet, S., Leroy, E., Boumsell, L. and Senik, A . , J. Immunol. 1986. 136: 1144. 16 Levy, R., Dilley, J., Fox, R. I. and Warnke, R., Proc. Natl. Acad. Sci. USA 1979. 76: 6552. 17 Morimoto, C., Distaso, J. A., Aldrich,W. R. and Schlossman, S. E , J. Immunol. 1985. 134: 1508. 18 Bernard, A., Cong Tran, H. and Bournsell, L., J. Immunol. 1987. 139: 18.

Isoforms of the E2 molecule

719

19 Amiot, M., Bernard, A . , CongTran, H., Leca, G., Kanellopoulos, J. and Boumsell, L., Scand. J. Immunol. 1986. 136: 1752. 20 Reinherz, E. L. and Schlossman, S. F., Cell 1980. 19: 821. 21 Amiot, M., Huet, S., Azogui, O., Dastot, H., Bernard, A. and Boumsell, L., in Dupont, B. (Ed.), Immunobiology of HLA, vol. II: Immunogenetics and Histocompatibility, Springer Verlag, New York 1988, p. 556. 22 Streuli, M., Hall, L. R., Saga,Y., Schlossman, S. F. and Saito, H., J. Exp. Med. 1987. 166: 1548. 23 Plata, F., Autran, B., Martins, L. P.,Wain-Hobson,S., Raphael, M., Mayaud, C., Denis, M., Guillon, J. and Debrt, P.. Nature 1987. 328: 348. 24 Hemperly, J. J., Murray, B. A . , Edelman, G. M. and Cunningham, B. A., Proc. Natl. Acad. Sci. USA 1986. 83: 3037.