Early Expression of Glycophorin C During Normal and Leukemic Human Erythroid Differentiation1

Home , Glycophorin, Glycophorin C, Glycoprotein

(CANCER RESEARCH 49. 2626-2632, May 15, 1989] Early Expression of Glycophorin C during Normal and Leukemic Human Erythroid Differentiation1

Jean-Luc Villeval, Caroline Le Van Kim, Ali Bettaieb, Najet Debili, Yves Colin, Bouchra El Maliki, Dominique Blanchard, William Vainchenker,2 and Jean-Pierre Cartron

Unite Inserm U 91, HÃ´pital Henri Mondar, 94010 CrÃ©Ãeil[J-L.K, A. B., N. D., W. V.], and UnitÃ©Inserm U 76, Institut National de Transfusion Sanguine, 6 rue Alexandre Cabanel, 75015 Paris [C. L. V. K., Y. C., B. E. M., D. B., J-P. C.J, France

ABSTRACT red cell membrane. They carry the MN and Ss blood group antigens, respectively, and have extensive structural homolo- Glycophorins C and D (GPC and GPD) are two erythrocyte glycoproteins which originate from the same gene but differ in their NHj-terminal gies, suggesting that they have derived by duplication of a common ancestral gene. GPC and GPD are minor components residues. The cell surface expression of these glycoproteins during normal and erythroid differentiation has been investigated with monoclonal and of the red cell membrane carrying the blood group Gerbich polyclonal antibodies and has been compared to the expression of gly- determinants. Several lines of evidence indicate that these GPs cophorin A (GPA), the major sialoglycoprotein of human red cells. Using play a critical role in regulating membrane deformability and glycosylation-independent antibodies (monoclonal or polyclonal), GPC or mechanical stability of red cells (2) by interacting via protein GPD was detected in erythroid and nonerythroid cell lineages. However, 4.1 with the membrane skeleton (3). Recently, it was suggested a glycosylation-dependent monoclonal antibody (MR4-130) detected an that GPC and GPD are encoded by the same gene (4) and that epitope on GPC which appears to be erythroid specific, suggesting that GPD might be a shortened form of GPC (5). Indeed, these two lineage specificity of this glycoprotein is related to some carbohydrate GPs differ in their NH2-terminal region (4) but are immunolog structures. During normal erythroid differentiation, GPC was expressed ically (6) and structurally (5) indistinguishable in their COOH- early at the level of erythroid progenitors (part of erythroid burst-forming unit and erythroid colony-forming unit) as detected with a glycosylation- terminal domain. independent monoclonal antibody (APO 3), whereas GPA is only present Since the red cell membrane glycophorins are now well during terminal erythroid differentiation. The MR4-130 epitope was not characterized, they represent useful markers to follow the se coordinately expressed on the cell surface with the GPC molecule in the quential steps of erythroid differentiation, and GPA has been erythroid differentiation, since it was detected at the level of the more already used for this purpose (7). GPA is synthesized during mature erythroid colony-forming unit slightly later than the GPC poly- terminal erythroid differentiation but is detected slightly earlier peptide. with PoAbs which recognize peptidic epitopes than with MoAbs In four erythroleukemic patients, blast cells blocked at discrete stages which react with carbohydrate-dependent determinants (8). of the erythroid differentiation were also investigated with antibodies and However, even with a PoAb, GPA is not detected on the surface complementary DNA probes for GPA and GPC. GPA was immunologi- of BFU-E and CFU-E (9). Only preliminary data, however, cally detected in three of four cases, and its cell surface expression was have been obtained concerning the expression of GPC during correlated with the amount of specific mRNA in the cells, as seen by Northern blot analysis. GPC was immunologically detected on the blast normal (10) and leukemic differentiation (11). In the present cells of all four patients. However, in two cases including one with positive study, we have compared the expression of GPA and GPC expression of GPA, the MR4-130 epitope was absent from the GPC during normal and leukemic erythroid differentiation using molecule. By Northern blot analysis, we found that the GPC/GPD mRNA specific antibodies, and we have investigated their biosynthesis was present at a high level in all four patient samples. Western blot by Northern blot analysis of their mRNA in four patients analysis of GPC and GPD in two of these patients revealed that these suffering from erythroleukemia. mRNAs were mostly translated into the GPD molecule, suggesting that these glycoproteins might be differently processed in certain cases of erythroleukemia. MATERIALS AND METHODS

Samples. Normal bone marrow and blood samples obtained from INTRODUCTION normal donors after informed consent were collected on heparin. In Human erythrocytes carry at least four main species of sial- addition, the blast cells from 4 patients presenting an early erythroid oglycoproteins known as GPA,3 GPB, GPC, and GPD, which leukemia were analyzed; 3 of these 4 cases have been previously de scribed (11). have been extensively investigated and have been used as models Isolation of the Cells. Cells from blood and normal marrows were of integral membrane proteins (for review, see Ref. l). These separated by a Ficoll-metrizoate gradient centrifugation (Lymphoprep; GPs have a high content in sialic acid which contributes to the Nyegaard, Norway). Adherent and nonadherent cells were separated by negative charge of red cells. Moreover, they carry blood group a 2-h incubation. T-cells were separated from other mononuclear cells determinants and receptors for viruses, bacteria, and parasites. by rosette formation with 2-amino-ethylisothiouronium bromide- GPA and GPB are major and well-known constituents of the treated sheep red blood cells (12). Granulocytes were isolated by sedi mentation on dextran followed by a Ficoll-metrizoate gradient centrif Received 10/17/88; revised 1/27/89; accepted 2/9/89. ugation. The remaining red cells in the granulocyte cell pellet were The costs of publication of this article were defrayed in part by the payment removed by lysis with ammonium chloride (13). Platelets were obtained of page charges. This article must therefore be hereby marked advertisement in from platelet-rich plasma. accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ' This work was partly supported by grants from la Ligue Nationale sur le Antibodies. The antiserum directed against GPC and GPD (L 857) Cancer, from ARC (Association pour la Recherche sur le Cancer) and from the was produced by immunization of rabbits with purified GPD.4 Mono GEFLUG. clonal antibodies anti-GPC, MR4-130, and APO 3 were obtained from 2To whom requests for reprints should be addressed. 3The abbreviations used are: GPA. glycophorin A; GPB, glycophorin B; GPC, P. Rouger (14) and W. Dahr (15), respectively. Two previously de glycophorin C; GPD, glycophorin D; GP, glycoprotein; MoAb. monoclonal scribed antibodies against GPA, a murine MoAb (LICR-LON-R18) antibody; PoAb, polyclonal antibody; CA 1, carbonic anhydrase 1; BFU-E, erythroid burst-forming unit: CFU-E, erythroid colony-forming unit; CFU-GM, 4 B. El Maliki, D. Blanchard, W. Dahr, K. Beyreuther, and J-P. Carton. granulocyte-macrophage colony-forming unit; CFU-MK, megakaryocyte colony- Structural homology between the glycophorins C and D of the human erythrocyte forming unit. membrane, submitted for publication. 2626

(16,17) and a rabbit PoAb ( 18) (generous gifts of Dr. P. A. W. Edwards tion. The monoclonal antibody APO 3 reacted with a trypsin- and Dr. C. G. Gahmberg, respectively), have also been used. A murine sensitive but neuraminidase-resistant determinant located in MoAb against myeloid cells (80H5) (19) and several rabbit PoAbs the NH2-terminal region of GPC (15). Western blot studies against myeloperoxidase (20). CA 1 (21), hemoglobin (22), and von carried out with red cell membrane preparations further indi Willebrand factor (Dakopatts. Copenhagen, Denmark) have been used cated that APO 3 reacted with GPC but not with GPD (Fig. in double staining experiments. 1). It may be deduced from these data that MR4-130 and APO Fluorescent Labeling. Indirect immunofluorescence for murine 3 are both directed against GPC without cross-reactivity for MoAbs was performed on unfixed cells with an appropriate dilution of the antibody (1/50 to 1/500 depending upon the antibody). Binding GPD. However, the MR4-130 epitope was glycosylation de was revealed by incubation with a goat anti-mouse IgG Fab'2 fragment pendent, whereas the APO 3 epitope was not. The rabbit conjugated to fluorescein 500-fold diluted (Bioart, Meudon, France). antibody L 857 was obtained by immunization with a purified Cells were either spread on a slide and examined with a Zeiss micro GPD preparation.4 The immune serum was absorbed on intact scope (Oberkochen, Federal Republic of Germany) equipped with flu erythrocytes in order to select glycosylation-independent anti orescence epiillumination or analyzed by flow cytometry. Rabbit antibodies directed against the intracellular or intramembranous serum against either GPA or GPC was applied on cytospinned cells after 1-min fixation by methanol; their binding was revealed by a goat domain of the minor glycophorins. As expected, the absorbed anti-rabbit IgG Fab'2 fragment conjugated to rhodamine (Cappel, Coch- antibody reacted with both GPC and GPD in Western blot analysis (Fig. 1) since both glycoproteins share a common C- ranville, PA). Normal rabbit serum was used as a control. Fluorescence terminal region.4 We subsequently investigated the reactivity of labeling was performed on the different types of separated hematopoietic cells as well as on hematopoietic colonies grown in plasma clot these antibodies directed against GPC and GPD and those (21). against GPA with hematopoietic cells. Cell Sorting. Flow cytometric analyses and cell sorting were per Expression of GPA and GPC by Normal Hematopoietic Cells. formed on fresh light-density marrow cells stained by MoAbs MR4- Among peripheral blood cells, the MoAb or PoAb directed 130, APO 3, and R 18 using a FACS 440 (Becton Dickinson, Moutain against GPA stained only red cells as previously reported (7, 9, View, CA) or an ATC 3000 (Odam, Wissembourg, France). Marrow cells labeled by a fluorescent conjugated Fab'i fragment or by an 10). In contrast, a different pattern was observed with antibodies irrelevant IgGl followed by the fluorescent conjugated Fab'2 fragment directed against GPC and GPD. MR4-130 stained only red cells by indirect immunofluorescence, whereas APO 3 stained were used as a negative control. Cells were sorted under sterile condi tions and collected into positive and negative fractions. Unfractionated leukocytes as well. APO 3 gave significant fluorescence labeling with all T-cells (rosette-forming cells) and very weak staining, marrow cells and cells sorted into positive and negative fractions were assayed for hematopoietic progenitors or were cytocentrifuged for May- just above background with B-cells (nonadherent, non-roset- Griinwald-Giemsa staining. ting-forming cells) and monocytes (adherent cells). Granulo- Clonogenic Assays. Clonogenic assays for BFU-E, CFU-E, CFU- cytes and platelets were negative. The reactivity of the rabbit GM, and CFU-MK were performed by either the methylcellulose serum (L 857) was similar to that of APO 3. technique (23) or the plasma clot technique (21). Stimulating factors When tested on marrow cells, the antibodies against GPA were a combination of 5% supernatant of the Mo cell line and of l IU/ ml of human erythropoietin (Terry Fox Laboratory, Vancouver, Can ada) for erythroid progenitors, 5% supernatant of the 5637 cell line for CFU-GM, or 2.5% phytohemagglutinin leukocyte-conditioned medium for CFU-MK. Complementary DNA Probes and Northern Blot Analysis. Blood cells (>90% of blasts) from the 4 erythroleukemic patients were obtained by separation on a Ficoll gradient, and total RNAs were prepared accord ing to published procedures (24, 25). Aliquots of 10 Mgwere resolved by electrophoresis on 6% formaldehyde/1% (w/v) agarose gels and transferred onto nitrocellulose membranes according to the method of Thomas (26). Hybridizations with 32P-labeled complementary DNA inserts for GPC (27) and GPA (28) were performed as previously described (27). Western Blot Analysis. The presence of the GPC and GPD molecules on the blasts from Patients 2 and 3 were investigated by immunoblotting technique using the polyclonal and monoclonal (APO 3) antibodies A- according to Towbin et al. (29). C

RESULTS Specificity of Antibodies against Human Glycophorins C and B D. The specificity of three antibodies directed against the erythrocyte glycophorins, GPC and GPD, was examined in detail. The structure recognized by the monoclonal antibody MR4- 130 has been originally described as directed against the blood group Gerbich antigens (14) which are carried by GPC and GPD (for review, see Ref. l). This antibody, however, recog nized the NH2-terminal residues of GPC and did not bind to Fig. 1. Western blot analysis of the specificity of APO 3 MoAb and L 857 GPD as deduced from inhibition studies with purified glyco- antiserum on human red cells. Sixty Â¿igofproteins solubilized in the presence of sodium dodecyl sulfate and protease inhibitors were run in a 10% polyacrylamide protein fractions (15). Furthermore, the binding to GPC in gel and blotted onto nitrocellulose sheets (29). The binding of APO 3 (Lane I) volved sialic acid residues. We found that MR4-130 antibody and L 857 (Lane 2) was revealed with '"I-Protein G or '"I-labeled rabbit IgG directed against mouse IgG, respectively, and subsequent autoradiography for 1 could be used for binding studies on intact cells, but it did not day at â€”¿80Â°C.The letters A, B.C. and D on the left refer to the migration position react properly on either Western blots or by immunoprecipita- of the monomeric species of glycophorins from normal erythrocytes. 2627

Downloaded from cancerres.aacrjournals.org on October 2, 2021. © 1989 American Association for Cancer Research. EXPRESSION OF GLYCOPHORIN C DURING ERYTHROID DIFFERENTIATION stained only erythroblasts which were characterized morpho logically under the light microscope and by double labeling experiments using the anti-CA 1 or anti-hemoglobin polyclonal antibodies. Identical results were obtained with the MR4-130 MoAb. The specificity of the anti-GPA antibodies and MR4- 130 MoAb for the erythroid lineage was further confirmed by the analysis of CFU-GM-, BFU-E-, and CFU-MK-derived col onies in indirect immunofluorescence experiments. All these antibodies stained only erythroid colonies as ascertained by their morphology and by double labeling experiments. Further more, no staining was observed when erythropoietin was omit ted from the culture medium. APO 3 also gave a bright fluorescence on erythroblasts. However, other marrow cells (negative for hemoglobin or CA 1) were also stained, but with a very weak intensity of labeling. These cells could not be identified. The L 857 antiserum labeled erythroid cells strongly as well as some myeloid precursors which were identified by a staining with 80H5 (anti-CD15: X hapten, 3-fucose-/V-acetyl lactosamine). In order to more precisely investigate the reactivity of APO 3 and MR4-130 on hematopoietic precursors, marrow cells were analyzed after indirect immunofluorescence labeling by flow cytometry. At first, using R 18 as a positive control, two cell populations were clearly identified by their fluorescence intensity. Ninety-six % of the cells were negative and superim posed with the negative controls, while 4% of marrow cells were clearly positive which, after sorting, were identified as erythroblasts. In contrast, a complex curve was observed when using MR4-130 and APO 3 without clear-cut cell populations. Cells were subsequently analyzed for their fluorescence intensity and their forward angle light scattering (Fig. 2). Cells with the maximum autofluorescence (myeloid precursors) have a high forward angle light scattering and can be easily removed by selecting gates as shown in Fig. 2. Accordingly, an irrelevant IgGl MoAb gave less than 0.4% positive cells in this gate. Morphological analysis of the cells sorted with MR4-130 and APO 3 is reported in Table 1. MR4-130 positive population (3.7%) was constituted by 80% of erythroblasts. In addition, numerous naked nuclei from acidophilic erythroblasts were present but were not enumerated. A minor population (14%) Fig. 2. Correlation between fluorescence intensity of GPC antibodies ( Y-axis) of myeloid precursors was present which could be either truly and forward angle light scattering (X-axis) of low density human marrow cells. positive cells or a cell contaminant since, after reanalysis, the A, staining with MR4-130 antibody (glycosylation dependent); B, staining with sorted population was 90% positive for MR4-130. The APO 3 the APO 3 antibody (glycosylation independent) of the same marrow cells. Sorted populations are gated. positive population (6%) included a large fraction of blast cells (Table 1: myeloblasts plus blasts, 46%) besides the erythroid 10. A double staining was also performed between MR4-130 cells. These immature cells showed a low intensity of fluores and the rabbit serum against GPA at Day 7 of culture. The cence labeling and an intermediate or high forward angle light majority of BFU-E-derived colonies stained by MR4-130 re scattering. Some of these blast cells were myeloblasts; others mained negative with the rabbit anti-GPA, indicating that GPC had the morphological appearance of early erythroid cells. can be detected on the membrane of erythroid cells earlier than Therefore, the first part of this study indicated that, in GPA in the erythroid differentiation even when investigations contrast to GPA, GPC or molecules immunologically related are carried out with a PoAb. to GPC are expressed not only on the erythroid lineage but also In a second set of experiments, MR4-130 was used in cell- on other cell lineages such as T-cells, although at low antigenic sorting analyses. Clearly positive and negative bone marrow density. However, the carbohydrate-dependent epitope identi cells were sorted. All the morphologically identifiable erythro fied by MR4-130 was restricted to the erythroid lineage. blasts were found in the positive fraction containing from 8% GPC Is Expressed Earlier Than GPA during Erythroid Dif to 10% of marrow cells. In addition, all typical proerythroblasts ferentiation. In one set of experiments, the reactivity of MR4- and blast cells classified as erythroid cells by their strong 130 and anti-GPA antibodies with blood BFU-E-derived colo basophilia with a developed ergastoplasm were present in the nies from Day 5 to Day 12 of culture was investigated. The positive fraction. Clonogenic assays were performed, and the erythroid origin of these colonies was deduced from double results are summarized in Table 2. Hematopoietic progenitors labeling with the anti-CA 1 antibody. MR4-130 stained eryth of all cell lineages (BFU-E, CFU-GM, and CFU-MK) were roid colonies as early as Day 5 of culture. In contrast, R 18 did found in the negative fraction, and a significant percentage not recognize the BFU-E-derived colonies until Day 8; labeling (31%) of CFU-E was found in the positive fraction. These was significant from Day 9 of culture and was complete at Day CFU-E corresponded to progenitors which gave small erythroid 2628

Table 1 Morphological analysis ofunsorted versus sorted MR4-130 and APO 3 positive marrow cells Unsorted MR4-130 APO 3 cells positive cells positive cells Cell types (100%) (3.7%) (6%) Blasts 1

Myeloid cells Myeloblasts/promyelocytes Myelocytes Metamyelocytes 82% - Polymorphs 10

Erythroid cells Proerythroblasts Basophilic erythroblasts 2% 80% 46% Polychromatophilic erythroblasts Acidophilic erythroblasts

Lymphocytes Monocytes Plasma cell

Table 2 Sorting of human bone marrow cells with the MR4-130 MoAb A cell sorting was performed only on the fluorescence intensity. Bright fluorescent cells were sorted in the positive fraction. Cells were cultured at a concentration of 5.103 cells/ml for MR4-130 positive cells and 1.10* cells/ml for MR4-130 negative cells in the CFU-E, BFU-E, and CFU-GM assays, whereas they were plated at a concentration of 15.103 cells/ml and 2.10* cells/ml, respectively, for the CFU-MK assay. Results are expressed as the number of progenitors per 1.105 cells. CFU-E CFU-GM %of Cell fraction cells Total 8-16 cells >16 cells BFU-E Day 7 Day 14 CFU-MK MR4-130 positive cells 2251 Â±30Â°(31)* 2100 Â±290 (81) 151 Â±10(3) 15 Â±7(0.5) 285 Â±21(3) 255 Â±21 (3) 18(1)

MR4-130 negative cells 96 213 Â±26(69) 21 Â±10(19) 192 Â±16(97) 142 Â±22 (99) 410 Â±42(97) 293 Â±46 (97) 74 Â±4(99) Â°Mean Â±SD. * Numbers in parentheses, percentage.

Table 3 Sorting of human bone marrow cells with MR4-130 and APO 3 MoAbs A biparametric cell sorting (fluorescence intensity and forward angle light scattering) (see Fig. 2) was performed. Cells were plated at a concentration of 5.10' cells/ ml for MR4-130 positive cells, 35.10' cells/ml for APO 3 positive cells, and 1.10* cells fro all the other cell fractions. Results are expressed as the number of progenitors per 1.10* cells.

CellfractionUnsortedMR4-130 cells1003.7 cells157 cells165 7978 14216 Â±25"7360 Â±66500 Â±18860 Â±59HI Â±741180 Â±4780

positive cells Â±900 (55)* Â±1280(80) + 360(16) Â«10(4) + 360(6) Â±28(1) cellsAPOMR4- 130 negative 96.35.894.2Total322241 Â±1(45)4214 64 Â±6(2)2031 177 Â±4(84)2457 281Â±30(96)2686 743 Â±38(94)1857 269+17(99)296+

3 positive cells Â±214 (78) Â±286(88) Â±314(74) Â±161 (47) Â±157(12) 10(7) APO 3 negative cells%of 71 Â±1 (22)CFU-E8-1617 Â±(12)CFU-GM>1654 Â±26 (26)BFU-E305187.5 Â±23(53)Day 865 Â±111 (88)Day243.5 Â±26 (93) Â°Mean Â±SD. * Numbers in parentheses, percentage. colonies (8 to 16 cells), whereas CFU-E-progenitor cells of Expression of GPA, GPC, and GPD in Erythroleukemias. The larger colonies were in the negative fraction. cell surface expression of GPs on the blasts from the 4 erythro- Previous studies from other investigators have shown that leukemic patients was investigated, and the results are sum GPA is not present on erythroid progenitors (7, 9, 30). marized in Table 4. In 3 patients, leukemic cells resembled APO 3 Epitope of GPC Is Expressed Earlier Than the MR4- normal erythroid cells by their immunophenotype. Patient 4 130 Epitope. Cell sorting with these two MoAbs was performed showed a more mature phenotype with the expression of GPA as shown in Table 1 in order to isolate all positive cells. Results with MR4-130 (Table 3) were identical to previous experiments detected by MoAb and PoAb on the majority of the cells. In Patient 2, 60%, 1%, and 35% of the blast cells were stained by (Table 2) with the exception that a higher number (55% versus MR4-130, R 18, and the PoAb directed against GPA, respec 31%) of CFU-E was found in the positive fraction, suggesting that CFU-Es have a low antigenic density of the MR4-130 tively. In Patient 1, no significant staining was observed with anti-GPA antibodies and MR4-130; the rare positive cells cor epitope. Other progenitors including BFU-E were negative. In contrast, APO 3 identified a significant fraction (47%) of BFU- responded to erythroblasts. In contrast, APO 3 antibody as well E and a majority (78%) of CFU-E. These BFU-Es corresponded as the L 857 serum faintly stained all blast cells. Two patterns to colonies with a low number of subcolonies, i.e., Day 12 or of labeling were observed in this case with the PoAb, i.e., a mature BFU-E. In addition, a higher number of CFU-GM weak membrane staining presumably on the majority of cells (12% versus 6% for Day 7 CFU-GM) was present in the APO but difficult to count on these fixed preparations and a more 3 positive fraction than in the MR4-130 positive fraction. This intense diffuse cytoplasmic labeling in about 40% of the cells. percentage was slightly higher than the contaminant cells The immunophenotypes of the leukemic cells from these three (<10%). Results of the expression of GPA and GPC during cases overall corresponded to that of a basophilic erythroblast, erythroid differentiation are schematically summarized in Fig. a proerythroblast or a late CFU-E, and an erythroid progenitor 3. (CFU-E or Day 12 BFU-E), respectively. 2629

p BFU-E m BFU-E CFU-E CFU-E PrtwylllroblÂ«Â« ErythroMists . It nib) I Â«*â€¢<Â»M. < It C4b

Fig. 3. Schematic diagram of the expres sion of GPA and GPC during normal erythroid GLYCOPHORIN C (APO3 EPITOPE) differentiation, p BFU-E, primitive BFU-E (Day 16); m BFU-E, mature BFU-E (Day 12).

GLYCOPHORIN C (MR4 130 EPITOPE)

GLYCOPHORIN A

Table 4 Phenotype of the blast cells in four cases of erythmleukemia B Indirect immunofluorescent labeling was performed on unfixed cells for the anti-GPA and anti-GPC MoAb and on methanol-fixed cells for the three PoAbs. Cells were quantitated under a fluorescent microscope. Results are expressed as 2 3 the percentage of positive cells. GPA GPC GPC/ GPDMoAb MoAb MoAb PoAb Patient857)111 (R 18) PoAb (MR4-130) (APO 3) (L PoAb91 1 100 40* 2 1 35 60 ND ND 87 3 23 63 0 42 40* 85 ND"4 58 82 60 ND 90of C- CAI, carbonic anhydrase 1; ND, not determined. 0 A diffuse and intense cytoplasmic staining was observed in 40% the cells;c-D- in addition, a weak membrane staining was observed on the cells.CAI"

In the last patient (Patient 3), leukemic cells differed from a t normal erythroid cell by its phenotype. Indeed, about half of Fig. 4. Detection of GPC and GPD by Western blot in protein extracts from normal red cells and blasts from Patients 2 and 3. Electrophoresis and blotting the blast cells from Patient 3 express GPA, but GPC was of membrane proteins as in Fig. 1. A, immunoblotting with L 857 polyclonal undetectable with MR4-130. In contrast, L 857 serum and antibody 50-fold diluted followed by a detection with '"I-labeled Protein G. Lane 1, normal red blood cells; Lane 2, Patient 2; Lane 3, Patient 3. The high- APO 3 faintly stained the blast cells as observed in Patient 1. molecular-weight material corresponds to IgG bound to the cell surface and Our results suggest an abnormal processing of the GPC/GPD revealed by the '"I-labeled Protein G. B, immunoblotting with APO 3 MoAb molecules, especially of the GPC glycosylation, in these leu (culture supernatant 500-fold diluted). Binding was revealed by phosphatase alkaline-conjugated goat IgG directed against mouse immunoglobulin. Lane 1, kemic cells. In order to further discriminate between the expres normal red blood cells; Lane 2, Patient 3. A similar but weaker band which could sion of GPC and GPD, immunostaining of the proteins ex not be photographed was obtained in Patient 2. tracted from the blast cells of Patients 2 and 3 was performed using the L 857 antiserum and APO 3 (Fig. 4). In both patients, GPD was clearly detected with the PoAb. The apparent molec ular weight of the GPD was slightly higher in the two patients than in control red cells, suggesting an abnormal glycosylation. On Western blots, APO 3 revealed low amounts of GPC on the -3.Â« erythroblasts of Patient 3.

To investigate whether the antibody staining in blast cells -2.2 was correlated with gene transcription, the level of mRNAs specific of GPA and GPC was determined by Northern blot analysis (Fig. 5). The GPC mRNA was sized at 1.1 kilobases as in other tissues (4) and was detectable at a high level in blasts .0.9 from all patients, including those which express no significant amount of the MR4-130 epitope (Patients 1 and 3). Using the same Northern blot, the GPA mRNAs were detected in signif icant amounts only in cells from patients in which GPA had been previously detected by immunological techniques (Patients GPC probe GPA probe 3 and 4). A very low level of GPA mRNAs was present in Fig. 5. Level of GPC/GPD and GPA mRNAs in the blast cells from Patients Patients 1 and 2. The size of the GPA mRNAs (1.1, 1.7, and 1 to 4. After electrophoresis and transfer to nitrocellulose (see "Materials and 2.4 kilobases) was similar to that previously reported in fetal Methods"), 10 pg of total RNAs were hybridized first with the GPA complemen and adult erythroblasts and in the K562 cell line (28, 31). The tary DNA probe and then rehybridized with the GPC complementary DNA probe. Stringent washes were done in 1 mM NaH2PO4, pH 7.7, containing 18 3.6-kilobase mRNA detected with the GPA probe could repre mM NaCl, 0.1 mm EDTA, and 0.1% (w/v) sodium dodecyl sulfate at 60Â°Cfor 10 sent an unspliced precursor, since it was detected only when min. The number of each lane corresponds to that of each patient in Table 2. 2630

Downloaded from cancerres.aacrjournals.org on October 2, 2021. © 1989 American Association for Cancer Research. EXPRESSION OF GLYCOPHORIN C DURING ERYTHROID DIFFERENTIATION total RNAs were analyzed by electrophoresis and not when distribution than GPC. In fact, the PoAb directed against both purified polyadenylate-containing RNA was used. GPC and GPD had a cellular reactivity very close to that of a GPC-specific antibody, suggesting that GPD and GPC have a coordinate expression. However, further studies using contin DISCUSSION uous cell lines are necessary to determine whether GPC and GPD may have a dissociated expression in certain cell types In this paper, we have investigated the expression of GPA since no antibody is specific for GPD alone. and GPC/GPD during normal human erythroid differentiation Finally, we compared the expression of GPA and GPC/GPD and in erythroleukemias. Several lines of evidence indicate that in four cases of erythroleukemia undiagnosed by morphological expressions of GPA and GPC/GPD during hematopoietic dif criteria (11) with blast cells blocked at different stages of the ferentiation clearly differ. We found that GPC, in contrast to erythroid maturation. Important differences in the expression GPA (7, 9, 30), was not an erythroid-specific protein. When of these glycophorins have been found. Based on the immuno- using the APO 3 MoAb which recognizes a glycosylation- logical detection of GPA, these cases have been classified into independent epitope of GPC, a molecule identical or immuno three groups corresponding to discrete stages of the normal logically related to GPC was found on lymphocytes, monocytes, erythroid differentiation (11). Northern blot analysis indicated and presumably on myeloid precursors but with a much lower a direct relationship between the cell surface expression of GPA antigenic density than on the erythroid lineage. It is well estab and the level of its specific mRNAs. In contrast, while the GPC lished that this molecule is GPC, since the expected mRNA was recently detected in all these cell types.5 However, further molecule was detected in all cases with the APO 3 antibody, the cell surface expression of the MR4-130 epitope on blast analyses are necessary to characterize the GPC molecule of cells did not perfectly correlate with the classification estab nonerythroid cells. Some glycosylation-dependent epitopes of lished previously (11), since in one patient this MoAb was not the GPC molecule may be lineage specific. Indeed, the MR4- reactive with the blast cells expressing GPA. In fact, this result 130 MoAb which defines a sialic acid-dependent epitope on can be explained by an incomplete glycosylation of the GPC GPC has an erythroid-restricted reactivity. From the present molecule and probably also by a posttranscriptional control of study, it cannot be firmly concluded whether this phenomenon GPC/GPD expression. Two lines of evidences support this last is related to an erythroid-specific glycosylation of GPC or to a hypothesis. First, Western blot analysis indicated that the GPC/ glycosylation determinant which is predominant on the eryth GPD mRNA is preferentially translated into an incompletely roid GPC. The presence of some myeloid precursors in the glycosylated GPD molecule. Second, the large amounts of MR4-130 positive cell fraction in sorting experiments suggests GPC/GPD mRNA present in blast cells from all patients did that the second hypothesis is more likely. Cell type-specific O- not correlate with the level of these molecules. linked carbohydrate structures are well documented for leuko- In conclusion, our studies indicate that GPA and GPC/GPD sialin, another sialoglycoprotein of the hematopoietic series are useful markers to follow the differentiation of the erythroid (32). It was shown that leukosialin is an ubiquitous molecule cell lineage. GPC is expressed earlier than GPA, but evidences among hematopoietic cells which displays cell lineage-specific are accumulating that GPC has a wider distribution than GPA O-glycosylation. Supporting this hypothesis a MoAb called on hematopoietic cells. Interestingly, however, erythroid spec GA3 was produced recently which identifies an O-glycosylation- ificity of GPC appeared to be acquired by maturation-dependent dependent leukosialin epitope which, like the MR4-130 epitope carbohydrate structures. Investigations on the cell surface on GPC, appears to be erythroid restricted (33). expression and transcription of GPA and GPC/GPD molecules GPC is expressed early during erythroid differentiation, from erythroleukemic cells suggest an abnormal processing of whereas GPA is only synthesized during terminal erythroid the GPC/GPD molecules. Further studies should indicate differentiation (7, 9, 30). Indeed, GPC could be detected with whether this is due to the leukemic process or to a posttran APO 3 on a part of BFU-E and on CFU-E. This early synthesis scriptional control of GPC/GPD production in immature cells. of GPC during erythroid differentiation is in agreement with a recent report which indicates the presence of GPC on normal proerythroblasts (34). In addition, the expression of the MR4- ACKNOWLEDGMENTS 130 epitope was found to be maturation dependent, since it was We are indebted to Z. Mishal and A. Katz for their help in cell- detected later in the erythroid differentiation than the APO 3 sorting experiments. The authors wish also to thank Dr. P. A. W. epitope. MR4-130 did not react with BFU-E but bound to a Edwards for providing R 18, Dr. W. Dahr for APO 3, Dr. C. G. fraction of CFU-E. The MR4-130 positive CFU-E gave rise to Gahmberg for the anti-GPA PoAb, Dr. P. Mannoni for 80H5, and Dr. small colonies, suggesting that they correspond to the more P. Rouger for MR4-130. We are grateful to Drs. J. London and C. differentiated types of these progenitors (35). Maturation-de Rahuel for providing the complementary DNA probe for GPA and to pendent glycosylation of other erythroid sialoglycoproteins has J. M. Masse for photographic assistance. been previously reported. Gahmberg et al. (8, 36) have shown, using PoAb and MoAb, that O-glycosylation of GPA gradually REFERENCES increases during erythroid differentiation, its glycosylation being complete at the stage of the polychromatophilic erythro- 1. Dahr, W. Immunochemistry of sialoglycoproteins in human red cell mem branes. In: V. Vengelen-Tyler and W. J. Judd (eds.). Recent Advances in blast. However, these results were not confirmed by a recent Blood Group Biochemistry, pp. 23-65. Arlington, VA: American Association report using two MoAbs, including one which is not sugar of Blood Banks, 1986. 2. Reid, M. E., Chasis, J. A., and Mohandas, N. Identification of a functional dependent (37). In contrast, leukosialin has been clearly dem role for human erythrocyte sialoglycoproteins ÃŸand >. Blood, 69: 1068- onstrated to exhibit a maturation-dependent O-glycosylation 1072, 1987. (33). 3. Mueller, T. J., and Morrison, M. Glycoconnectin (PAS 2), a membrane attachment site for human erythrocyte cytoskeleton. In: W. C. Kruckenberg, We have also investigated whether GPD, which represented J. W. Eaton, and G. J. Brewer (eds.), Erythrocyte Membrane 2: Recent a shortened variant of GPC, could have a different tissue Clinical and Experimental Advances, pp. 95-112. New York: Alan R. Liss, 1981. ' L. V. Kim el al., manuscript in preparation. 4. Le Van Kim, C., Colin, Y., Blanchard, D.. Dahr, W., London, J., and 2631

Cartron, J.-P. Gerbich blood group deficiency of the Ge:-I,-2,-3 and Ge:-l,- neutrophils by immunogold. Blood, 65: 423-432, 1985. 2,3 types: imtnunochemical study and genomic analysis with cDNA probes. 21. Villeval, J.-L., Testa, U., Vinci, G., Tonthat, H., Bettaieb, A., Titeux, M., Eur. J. Biochem., 165: 571-579, 1988. Cramer, P., Edelman, L., Rochant, H., Breton-Gorius, J., and Vainchenker, 5. Blanchard, D., El Maliki, B., Hermand, P., Dahr, W., and Cartron, J.-P. W. Carbonic anhydrase I is an early specific marker of normal human Structural homology between glycophorins C and D, two minor glycoproteins erythroid differentiation. Blood, 66: 1162-1170, 1985. of the human erythrocyte membrane carrying group Gerbich antigens. In: 22. Henri, A., Testa, U., Tonthat, H., Riou, J., Titeux, M., Vainchenker, W., Proceedings of the IXth International Symposium on Glycoconjugates, July Feuihade, F., Galacteros, F., and Rochant, H. Disappearance of I Ibi and i 6-11, Lille, p. F54, 1987. antigens during the first year of life. Am. J. Hematol., 9: 161-170, 1980. 6. Reid, M. E., Anstee, D. J., Tanner, M. J. A., Ridgwell, K., and Nurse, G. T. 23. Aye, M. T., Seguin, J. A., and McBurney, J. P. Erythroid and granulocytic Structural relationships between human erythrocyte sialoglycoproteins ,1and colony growth in cultures supplemented with human serum lipoproteins. J. y and abnormal sialoglycoproteins found in certain rare human erythrocyte Cell. Physiol., 99: 233-238, 1979. variants lacking the Gerbich blood-group antigen(s). Biochem. J., 244: 123- 24. Favaloro, J., Treisman, R., and Kamen, R. Transcription maps of polyoma 128, 1987. virus-specific RNA: analysis by two-dimensional nuclease SI gel mapping. 7. Robinson, J., Sieff. C., Delia, D., Edwards, P. A. W., and Greaves, M. F. Methods Enzymol., 65: 718-749, 1980. Expression of cell surface HLA-Dr, HLA-ABC, and glycophorin during 25. Maniatis, T., Fritsch, E. F., and Sambrook, J. Molecular cloning. In: A erythroid differentiation. Nature (Lond.), 289:68-71, 1981. Laboratory Manual, p. 191. Cold Spring Harbor, NY: Cold Spring Harbor 8. Gahmberg, C. G., Ekblom, M., and Andersson, L. C. Differentiation of Laboratory, 1982. human erythroid cells is associated with increased O-glycosylation of the 26. Thomas, P. S. Hybridization of denatured RNA and small DNA fragments major sialoglycoprotein, glycophorin A. Proc. Nati. Acad. Sci. USA, 81: transferred to nitrocellulose. Proc. Nati. Acad. Sci. USA, 77: 5201-5205, 6752-6756, 1984. 1980. 9. Greaves, M. F., Chan, L. C., Furley, A. J. W., Watt, S. M., and Molgaard, 27. Colin, Y., Rahuel, C., London, J., Romeo, P.-H., d'Auriol, L., Galibert, F., H. V. Lineage promiscuity in hemopoietic differentiation and leukemia. and Cartron, J.-P. Isolation of cDNA clones and complete amino acid Blood, 67: 1-11, 1986. sequence of human erythrocyte glycophorin C. J. Biol. Chem., 261: 229- 233, 1986. 10. Sieff, C. A., Caine, G., and Davis, S. Cell surface antigens expressed on 28. Rahuel, C., London, J., d'Auriol, L., Mattei, M. G., Tournamille, C., Skrynia, haematopoietic cells and the use of monoclonal antibody combinations to enrich for multipotent and committed progenitors (Abstract). Br. J. 1lac C., Lebouc, Y., Galibert, F., and Cartron, J.-P. Characterization of cDNA matol., 58:180, 1984. clones for human glycophorin A. Use for gene localization and for analysis 11. Villeval, J.-L., Cramer, P., Lemoine, F., Henri, A., Bettaieb, A., Bernaudin, of normal and deficient (Finnish type) genomic DNA. Eur. J. Biochem., 772: F.. Beuzard, Y., Berger, R., Flandrin, G., Breton-Gorius, J., and Vainchenker, 147-153, 1988. W. Phenotype of early erythroblastic leukemia. Blood, 68:1167-1174,1986. 29. Towbin, H., Staehelin, T., and Gordon, J. Electrophoretic transfer of proteins 12. Pellegrino, M. A., Ferrane, S., and Theofilopoulos, A. N. Isolation of human from polyacrylamide gels to nitrocellulose sheets: procedure and some appli T- and B-lymphocytes by rosette formation with 2-amino-ethylisothiouron- cations. Proc. Nati. Acad. Sci. USA, 76:4350-4354, 1979. ium bromide (AET)-treated sheep red blood cells and with monkey red blood 30. Loken, M. R., Shah, V. O., Dattilio, K., and Civin, C. I. Flow cytometric cells. J. Immunol. Methods, //: 273-282, 1976. analysis of human bone marrow. I. Normal erythroid development. Blood, 69: 255-263, 1987. 13. Verheught, F. W. A., von dem Borne, A. E. G. K., Prins, H. K., and Engelfriet, C. P. The detection of granulocyte antibodies in relation to granulocyte 31. Siebert, P. D., and Fukuda, M. Isolation and characterization of human transfusion. Exp. Hematol., 5 (suppl.): 151-155, 1977. glycophorin A cDNA clones by a synthetic oligonucleotide approach: nucleo- tide sequence and mRNA structure. Proc. Nati. Acad. Sci. USA, 83: 1665- 14. Rouger, P., Lee, H., Juszczak, G., Doinel, C., and Edelman, L. Murine monoclonal antibodies against Gerbich antigens. J. Immunogenet., 10: 333- 1669, 1986. 32. Carlsson, S. R., Sasaki, H., and Fukuda, M. Structural variations of O-linked 335, 1983. oligosaccharides present in leukosialin isolated from erythroid, myeloid, and 15. Dahr, W., Blanchard, D., Kiedrowski, S., Poschmann, A., and Cartron, J.- T-lymphoid cell lines. J. Biol. Chem., 261: 12787-12795, 1986. P. Characterization of Gerbich blood group related monoclonal antibodies. 33. Bettaieb, A., Farace, F., Mitjavila, M.-T., Mishal, Z., Dokhelar, M.-C, Tursz, In: P. Rouger and C. Salmon (eds.), Monoclonal Antibodies against Human T., Breton-Gorius, J., Vainchenker, W., and Kieffer, N. Use of a monoclonal Red Blood Cells and Related Antigens, pp. 289-292. Paris: Librairie Arnette, antibody (GA3) to demonstrate lineage restricted O-glycosylation on leuko 1987. sialin during terminal erythroid differentiation. Blood, 71:1226-1233,1988. 16. Anstee, D. J., and Edwards, P. A. W. Monoclonal antibodies to human 34. Telen, M. J., Scearce, R. M., and Haynes, B. F. Human erythrocyte antigens. erythrocytes. Eur. J. Immunol., 12: 228-232, 1982. III. Characterization of a panel of murine monoclonal antibodies that react 17. Mawby, W. J., Anstee, D. J., and Tanner, M. J. A. Immunochemical evidence with human erythrocyte and erythroid precursor membranes. Vox Sang., 52: for hybrid sialoglycoproteins of human erythrocytes. Nature (Lond.), 291: 236-243, 1987. 161-162, 1981. 35. Ouellete, P. L., and Monette, F. C. Erythroid progenitors forming clusters 18. Gahmberg, C. G., Jokinen, M., and Andersson, L. C. Expression of the in vitro demonstrate high erythropoietin sensitivity. J. Cell. Physiol., 105: major sialoglycoprotein (glycophorin) on erythroid cells in human bone 181-184, 1980. marrow. Blood, 52: 379-387, 1978. 36. Ekblom, M., Gahmberg, C. G., and Andersson, L. C. Late expression of M 19. Mannoni, P., Janowska-Wieczorek, A., Turner, A. R., McGann, L., and and N antigens on glycophorin A during erythroid differentiation. Blood, 66: Turc, J. M. Monoclonal antibodies against human granulocytes and myeloid 233-236, 1985. differentiation antigens. Hum. Immunol., 5: 309-323, 1982. 37. Loken, M. R., Civin, C. I., Bigbee, W. L., Langlois, R. G., and Jensen, R. 20. Cramer, E., Pryzwansky, K. B., Villeval, J.-L., Testa, U., and Breton-Gorius, H. Coordinate glycosylation and cell surface expression of glycophorin A J. Ultrastructural localization of lactoferrin and myeloperoxidase in human during normal human erythropoiesis. Blood, 70:1959-1961, 1987.

2632

Downloaded from cancerres.aacrjournals.org on October 2, 2021. © 1989 American Association for Cancer Research. Early Expression of Glycophorin C during Normal and Leukemic Human Erythroid Differentiation

Jean-Luc Villeval, Caroline Le Van Kim, Ali Bettaieb, et al.

Cancer Res 1989;49:2626-2632.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/49/10/2626

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/49/10/2626. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.