HLA-DR molecules on all cells are electrophoretically unique

Leukemia Research Vol. 13, No. 9, pp. 851--862, 1989. Printed in Great Britain.

HLA-DR

0145-2126/89 $3.00 + .00 Pergamon Press plc

MOLECULES

ON ALL CELLS

ELECTROPHORETICALLY

ARE

UNIQUE*

JOHN M. PESANDO, PATRICIA HOFFMAN and MONA ABED STUCKI Fred Hutchinson Cancer Research Center and The Biomembrane Institute, Seattle, WA, U.S.A. (Received 13 January 1989. Revision accepted 3 June 1989) Abstract--Previous SDS-PAGE studies of autologous clonal ALL and normal B cell lines indicated that HLA-DR molecules on leukemic cells have an extra fl chain band. We now show that these results are not an artifact of cell lines, EBV transformation, or cell growth in culture. Leukemic cells from four ALL patients were surface labeled with lzsI and their HLA-DR molecules compared with those on autologous normal peripheral blood B cells. The electrophoretic patterns of HLA-DR molecules on these in vivo cell populations are identical to those on the corresponding cell lines. Moreover, EBV-transformation does not alter the electrophoretic appearance of HLA-DR molecules on normal tonsil B cells. Cell lines can now be used to study the chemical basis for these electrophoretic differences. Key words: Autologous B cells, HLA-DR molecules, SDS-PAGE.

INTRODUCTION

DR molecules from surface iodinated clonal malignant B cells versus only one fl chain band in H L A - D R molecules from autologous clonal B lymphoblastoid cells. This same pattern is observed using multiple HLA-DR-specific monoclonal antibodies (MoAb) and in the H L A - D R molecules of both parental haplotypes. Prior to undertaking a detailed analysis of the chemical basis for these electrophoretic differences using cell lines, it is first necessary to determine if similar differences characterize the H L A - D R molecules on in vivo cell populations. Accordingly, immune precipitation and SDS-PAGE were used to compare surface iodinated H L A - D R molecules on cryopreserved pre-B leukemic cells obtained directly from patients with common acute lymphoblastic leukemia antigen ( C A L L A / C D 1 0 ) and CD19 positive A L L with those on autologous normal peripheral blood B cells. The latter cells were obtained from children in long-term first clinical remission, and studied either directly or following EBV transformation to increase cell number. These studies confirm the previously reported electrophoretic differences between H L A - D R molecules on leukemic pre-B and mature normal B cells. Moreover, similar comparisons of H L A - D R molecules on EBVtransformed versus nontransforrned normal (nonmalignant) tonsillar B cells from the same individual indicate that EBV transformation itself does not significantly alter the electrophoretic appearance of H L A - D R molecules.

CLASS II ANTIGENS of the major histocompatibility complex are polymorphic cell surface glycoproteins that play a central role in antigen presentation and self-recognition. In man these antigens are encoded by genes of the H L A - D R , -DQ, and -DP loci. Each antigen is composed of two noncovalently linked glycosylated polypeptides having molecular weights of approximately 34 (tr) and 28 (fl) kilodaltons [1-3]. Class II antigens are found primarily on normal and malignant hematopoietic cells, but they can be expressed by numerous additional cell types [2, 3]. In a previous study of H L A - D R molecules on normal and malignant human B cell lines established from same individual, we reported that H L A - D R molecules on pre-B acute lymphoblastic leukemia (ALL) and Epstein-Barr virus (EBV)-transformed normal B cells are electrophoretically and hence chemically distinct [4]. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis indicates that there are two fl chain bands in HLA* This work was supported by National Institutes of Health grants CA15704 and CA34206, and by a grant from the Biomembrane Institute. Abbreviations: SDS-PAGE, sodium dodecyl sulfatepolyacrylamide gel electrophoresis; EBV, Epstein--Barr virus; ALL, acute lymphoblastic leukemia; MoAb, monoclonal antibodies; CALLA, common acute lymphoblastic leukemia antigen; sIg, surface immunoglobulin. Correspondence to: Dr John M. Pesando, the Biomembrane Institute, 201 Elliott Avenue West, Seattle, WA 98119, U.S.A. 851

J.M. PESANDOet al.

852

The unique H L A - D R molecules on normal and leukemic B cells m a y reflect leukemia-specific processes or simply the different stages of B cell maturation represented by these two cell populations. In either case it m a y p r o v e possible to exploit these differences for diagnostic and therapeutic purposes.

Monoclonal antibodies The HLA-DR-specific p4.1 and ISCR3 MoAb were used to identify HLA-DR molecules [8-10]. Similarly, MoAb HK-19 (Mallinckrodt, St Louis, MO) and B7/21 were used to identify HLA-DQ and HLA-DP molecules, respectively [11, 12]. MoAb were obtained either as supernatants from overgrown cell cultures or as ascites from tumor-bearing mice, and were ultracentrifuged and filtered before use.

MATERIALS AND METHODS

Radiolabeling and immune precipitation Cells were surface labeled with 125I using the lactoperoxidase method [4, 10, 13]; I mCi was used to label 25 x 106 cells. Washed labeled cells were lysed in TrisHCl-buffered Triton X-100 (Sigma, St Louis) during a 30 min incubation. Nuclei were pelleted by centrifugation, and the lysate was applied to a Sephadex G-25 column (Pharmacia Fine Chemicals, Piscataway, NJ) to isolate labeled proteins. Lysates were precleared by overnight incubation with protein A-Sepharose (Pharmacia) followed by centrifugation. Immune precipitation was performed by incubating lysates with washed protein A-Sepharose that had been incubated overnight with the appropriate MoAb. Samples were washed and individual antigens eluted by heating for 2 rain at 95°C in reducing sample buffer.

Leukemic cell populations Leukemic cells were obtained from the peripheral blood of patients at the Children's Hospital Medical Center in Seattle at the time of diagnosis. All patients had CALLA/ CD10, CD19 and HLA-DR (Ia) positive ALL. Patient age ranged from 4 to 12 years. More than 95% of the HLADR positive cells in each sample had the morphologic appearance of leukemic blasts. Mononuclear cells were purified by Ficoll-Hypaque density gradient centrifugation [5] and cryopreserved in liquid nitrogen until use. Cell viability on thawing ranged from 80 and 90%. When necessary, thawed samples were again purified by Ficoll-Hypaque density gradient centrifugation to remove nonviable cells. HLA typing was not performed. Normal B cell populations Normal B cells were obtained from the peripheral blood of ALL patients at the Children's Hospital Medical Center who remained in long-term continuous first hematologic remission. Erythrocytes and granulocytes were removed by Ficoll-Hypaque density gradient centrifugation. Adherent cells were removed by incubating the sample for 1 h at 37°C. The small size of these samples, and the need to minimize cell loss, precluded T cell depletion when B cells were to be used directly for radiolabeling. The majority of the HLA-DR positive cells in these samples should be normal B cells. Viability of purified peripheral blood B cells was >95%. Human tonsils were obtained from patients undergoing elective tonsillectomy for tonsillitis at the Swedish Hospital Medical Center or at the Children's Hospital Medical Center in Seattle. Patient ages ranged from 2 to 23 years. Tonsils were mechanically teased apart to release mononuclear cells into sterile media containing a mixture of penicillin, streptomycin and amphotericin. Mononuclear cells were then purified by Ficoll-Hypaque density gradient centrifugation. T cells were removed from washed cells by rosette depletion using sheep red blood cells treated with 2-aminoethylisothiouronium bromide (AET) at 4°C [6]. Adherent cells were removed by incubating the sample for 1 h at 37°C. Viability of cryopreserved normal tonsil B cells was between 80 and 90%. EBV-transformed B cell lines EBV-transformed B cell lines were established from purified normal peripheral blood or tonsillar B cells using standard procedures [7]. Briefly, T cell and monocytedepleted mononuclear cells are seeded in 96-well plates at 1 x l0 s cells per well. An equal volume of virus-containing supernatant from an overgrown culture of the EBV-transformed Marmoset B958 cell line is added at 100 ~tl per well. Fresh medium is added weekly, and EBV-transformed human B cell lines are expanded after 2-4 weeks. Viability of those cells used for radiolabeling was approximately 90%.

Surface marker analysis The surface antigens on individual cells were detected using an indirect immunofluorescence assay monitored on the fluorescence-activated cell sorter as described [5]. Individual MoAb having the desired specificity were used together with fluoresceinated anti-mouse immunoglobulins. Gel electrophoresis Antigens were resolved by SDS-PAGE using a 25 cm 10-15% gradient slab gel and visualized by autoradiography. To avoid problems associated with potential differences in resolving power between individual polyacrylamide slab gels, HLA-DR molecules of all B cells from the same individual were purified and analyzed in parallel. Thus the singlet and doublet HLA-DR fl chain patterns reported below for samples from the same individual were observed on the same slab gel.

RESULTS

Cell populations Cryopreserved leukemic cells obtained at the time of diagnosis from the peripheral blood of over 50 patients with A L L were screened to identify those samples c o m p o s e d largely of malignant cells. More than 90% of cells from 31 of these patients were H L A - D R and C A L L A / C D 10 positive [5], expressed the B cell-specific CD19 surface antigen, and had the morphologic appearance of leukemic blasts. O v e r a 15-month period, peripheral blood specimens (510 ml) were obtained from four of these children who remained in first clinical remission m o r e than 12 months after diagnosis. Mononuclear cells were isolated by F i c o l l - H y p a q u e density gradient centrifugation, and monocytes were r e m o v e d by

Electrophoretically distinct HLA-DR molecules on autologous B cells

adherence at 37°C. To avoid excessive cell loss on handling, HLA-DR negative T lymphocytes were not removed. Mononuclear cells from two of these patients were directly radiolabeled with 1251and used in immune precipitation studies. Cell yields from two additional patients were too low for direct analysis, and B cells in these samples were therefore transformed with the Epstein-Barr virus and propagated in tissue culture prior to study. Viability of all cells used for radiolabeling was >90%. Following iodination, HLA-DR molecules from normal and malignant cells from these four patients were isolated by immune precipitation and compared by slab gel electrophoresis. Antigens from the paired cell populations of each patient were run on the same slab gel to ensure the identity of electrophoretic conditions.

Electrophoretic studies SDS-PAGE analysis reveals consistent differences between HLA-DR molecules on normal and leukemic B cells obtained directly from the same individuals. Two low molecular weight or/3 chain bands are observed with iodinated HLA-DR molecules on cry0preserved A L L cells from four out of four patients, while only a single/3 chain band is observed with HLA-DR molecules on corresponding autologous normal peripheral blood B cells. Identical results are obtained using EBV-transformed (Figs 1 and 2) or nontransformed (Figs 3 and 4) normal peripheral blood B cells, suggesting that EBV transformation itself does not alter the electrophoretic appearance of HLA-DR molecules. These reproducible electrophoretic findings imply chemical differences, possibly minor, between the HLA-DR molecules on these two cell populations. The small size of at least one sample in each pair, and difficulty in working with cryopreserved leukemic samples (low cell viability), precluded more detailed electrophoretic analysis on two-dimensional gels. One of the two HLA-DR/3 chain bands observed on leukemic cells has an apparent molecular weight comparable to that of the single /3 chain band observed on normal B cells. The apparent molecular weight of the second/3 chain band on leukemic cells is higher than the first. The high molecular weight or chain band of HLA-DR molecules on ALL cells is greater than that on autologous normal B cells. While HLA-DQ molecules are detected by indirect immunofluorescence assay on both leukemic and EBV-transformed normal B cells from these four patients, they are barely detectable by immune precipitation on leukemic pre-B cells. This problem may reflect the low quantitative expression of HLADQ on these latter cells. Because of this difficulty it is not possible to compare HLA-DQ molecules on

853

these autologous cell populations electrophoretically. Like HLA-DR molecules, HLA-DP antigens are detected by immune precipitation on all samples studied. However, while some electrophoretic differences are observed between HLA-DP molecules from autologous normal and malignant B cells, no pattern emerged. For example, HLA-DP molecules are comparable on normal and leukemic B cells from two patients. In a third patient, two bands are observed in the HLA-DP/3 chain region of EBVtransformed normal B cells, but not leukemic ones. In the fourth patient a broad band is observed in the c~chain region of EBV-transformed normal B cells, but not autologous leukemic B cells. Comparable heterogeneity of HLA-DP electrophoretic patterns is observed in the autologous normal and leukemic B cell lines studied previously [4].

Effects of EBV transformation HLA-DR molecules on autologous untreated and EBV transformed nonmalignant tonsillar B cells were directly compared to assess the potential contribution of EBV transformation on the electrophoretic properties of these molecules. Tonsillar B lymphocytes from nine individuals were purified and cryopreserved. Small aliquots of cells from each individual were transfected with EBV and grown as immortal cell lines in tissue culture. HLA-DR molecules were then isolated from both cryopreserved nontransformed and continuously growing EBV-transformed cells from each individual following surface labeling with 1251.HLA-DR molecules were again compared by SDS-PAGE on slab gels (Figs 5 and 6). In seven of these nine individuals, identical numbers of HLA-DR/3 chain bands are observed in the two B cell populations from each individual, although the number of/3 chain bands detected varied from one donor to another (Table 1). For example, in two of seven individuals a single HLA-DR/3 chain band is observed on both cell populations while two such bands are observed on the other five pairs. This observation contrasts with detection of a single HLADR/3 chain band on normal peripheral blood B cells obtained directly from two of two patients. An additional/3 chain band is observed in HLADR molecules from autologous nontransformed versus EBV-transformed tonsillar B cells from two of nine individuals. In one individual, three HLA-DR /3 chain bands are observed on untreated cells versus two such bands on cells from the autologous EBVtransformed cell line. In a second individual, two closely spaced HLA-DR/3 chain bands are observed on untreated cells versus a single predominant /3

J. M. PESANDOet al.

854

TABLE 1. H L A - D R MOLECULES ON AUTOLOGOUS UNTREATED AND EBV-TRANSFORMED TONSILLAR B-CELLS

Patient no. 1 2 3 4 5 6 7 8 9

B Cell type

No. of fl chains

No. of a~chains

Untreated Transformed Untreated Transformed Untreated Transformed Untreated Transformed Untreated Transformed Untreated Transformed Untreated Transformed Untreated Transformed Untreated Transformed

Two Two One One Two Two Two Two One One Two Two Two Two Two One Three Two

One One One Two One One One Two One Two One Two One Two One Two One One

Cells were surface labeled with 1251and HLA-DR molecules identified by immune precipitation using the HLADR-specific MoAb p4.1 and ISCR3. Antigens were visualized by SDS-PAGE and autoradiography. Aliquots of purified tonsillar B cells were transformed with EBV using standard procedures while the rest of the sample was cryopreserved until the B lymphoblastoid cell line was established. All samples from the same individual were radiolabeled and studied at the same time. HLA-DR molecules on autologous cell populations were analyzed in parallel on the same polyacrylamide slab gel.

chain band accompanied by a small satellite band on cells from the autologous EBV-transformed B cell line. The EBV-transformed cells from these two individuals may fail to reflect all of the B cell populations, and hence all of the H L A - D R molecules, contained within the original specimens. In addition, two discrete H L A - D R a~ chain bands are observed on EBV-transformed tonsillar B cells from six of these nine individuals while single but broad a~chain bands are observed on autologous untreated tonsillar B cells. The electrophoretic patterns of H L A - D Q and H L A - D P molecules from untreated and EBV-transformed normal tonsillar B cells appear to be comparable. DISCUSSION These studies indicate that H L A - D R molecules on in vioo normal and leukemic B cells from the same

patient are electrophoretically unique. H L A - D R molecules on leukemic cells from patients with

C A L L A and CD19 positive A L L have an extra fl chain band when compared under identical conditions with H L A - D R molecules on normal mature B cells from the same individual. Identical results are obtained when H L A - D R molecules on freshly isolated leukemic B cells from patients or leukemic B cells lines grown in tissue culture are compared with those on similarly obtained autologous normal B cells or B cell lines. These electrophoretic differences therefore do not appear to be unique to cell lines, or to the conditions used for cell growth in oitro. The process of E B V transformation does not appear significantly to alter the electrophoretic patterns of H L A - D R molecules. The electrophoretic patterns of H L A - D R molecules of nontransformed and EBV-transformed normal peripheral blood B cells from different individuals are comparable. H L A - D R molecules on nontransformed and EBVtransformed nonmalignant tonsiUar B cells from the same individual contain identical numbers of fl chain bands in seven of nine individuals. Moreover, these unique H L A - D R fl chain band patterns do not arise from comparison of polyclonal leukemic versus clonal normal B cells, since all leukemic cells used in these studies are also of clonal origin. The chemical basis for these electrophoretic differences remains to be determined. H L A - D R molecules on autologous B cells were studied to control for differences in the peptide chains of these molecules that usually occur when samples from different individuals are compared [1-3]. Since these electrophoretic differences are observed between normal and malignant B cells from the same individual, they may reflect unique post-transitional processing of identical proteins. Zacheis et al. [19] report two distinct populations of H L A - D R fl chain molecules on tonsilar B cells from a single individual. One of these H L A - D R fl chains contains a high mannose oligosaccharide, and the other contains a complex oligosaccharide. Similarly, Alexander et al. [20] find that molecular weight variations between the a'chains of H L A - D R molecules on autologous human melanoma and B lymphoblastoid cell lines reflect unique carbohydrate structures on identical proteins. Alternatively, however, the electrophoretic differences that we have observed may reflect cell-tocell variation in the quantiative expression of the two H L A - D R fl chain gene products found on most cells (ill and f13). We suspect that differences in fl chain glycosylation account for these molecular weight variations. Our previous studies of autologous normal and leukemic B cell lines suggest that sialic acid residues contribute to these differences but are not the sole basis for them [4]. The unique H L A - D R molecules on normal and

A 1 FIG. 1. Comparison of H L A - D R molecules on cryopreserved pre-B A L L cells (lanes A - D ) and EBV-transformed normal B cells (lanes A ' - D ) from the same patient. Lanes A and A ' include the anti-CALLA-1 (J5) MoAb as a control. H L A class II antigens were identified with the HLA-DR-specific MoAb p4.1 (lanes B and B'), the HLADP-specific MoAb B7/21 (lanes C and C'), and the HLADQ-specific MoAb HK-19 (lanes D and D'). All samples were run on the same gel. Note the two bands in the fl chain region of H L A - D R molecules on leukemic cells in lane B but not on EBV-transformed normal cells in lane B'. Note also the differences in the H L A - D P molecules on these autologous B cell populations.

855

0-A

B

C

A~ B t

Ct

2 FIG. 2. Comparison of HLA class II molecules on autologous cryopreserved pre-B ALL (lanes A-C) and EBVtransformed normal B cells (lanes A ' - C ' ) from a second patient. Antigens were identified by the HLA-DR-specific MoAb p4.1 (lanes A and A'), the HLA-DP-specific MoAb B7/21 (lanes B and B'), and the HLA-DQ-specific MoAb HK-19 (lanes C and C'). Again note the differences in the /3 chain bands of HLA-DR molecules on leukemic and normal B cells in lanes A and A'.

856

FIO. 3. Comparison of HI,A-DR molecules on cryopreserved pre-B ALL cells (lanes A-D) and normal peripheral blood cells (lanes A ' - D ' ) from the same patient. The latter cells were collected when the patient was in hematologic remission and have been depleted of monocytes. The HLA-DR-specific MoAb p4.1 and ISCR3 were used in lanes B and C and in lanes B' and C', respectively. The HLA-DP-specific MoAb B7/21 was used in lanes D and D'. The CALLA-specific MoAb anti-CALLA-1 (J5) was used in lanes A and A'. Comparison of short and long exposures of this autoradiograph indicates that there are two bands in the ~ chain region of the HLA-DR molecules on leukemic but not normal cells.

857

46--

0--

4

A

B

C

D

E

At

Bt Ct

Dw Et

FIG. 4. Comparison of HLA-DR molecules on cryopreserved pre-B ALL cells (lanes A-E) and fresh peripheral blood cells (lanes A'-E') from a second patient. Normal peripheral blood cells have been depleted of monocytes. Antigens are identified by the HLA-DR-specific MoAb p4.1 (lanes A and A'), ISCR3 (lanes B and B'), and I-2 (lanes C and C'), the HLA-DP specific MoAb B7/ 21 (lanes D and D'), and the HLA-DQ-specific MoAb HK-19 (lanes E and E'). Again note the two bands in the /3 chain region of HLA-DR molecules on malignant but not normal B cells.

858

N

5

B

C

FIG. 5. Comparison of HLA-DR molecules on untreated (lanes A through E) and EBV-transformed (lanes A' through E') nonmalignant tonsillar B lymphocytes from the same patient. The following MoAb were employed: anti-CALLA-1 (J5), lanes A and A'; the HLA-DR-specific MoAb p4.1, lanes B and B'; the HLA-DR-specific MoAb ISCR3, lanes C and C'; the HLA-DP-specific MoAb B7/ 21, lanes D and D'; and the HLA-DQ-specific MoAb HK19, lanes E and E'. The HLA-DQ and -DP molecules are better visualized with longer exposures.

859

A

B

C

D

E

E a B I C'

D i Ai

6 FIG. 6. Comparison of HLA-DR molecules on untreated (lanes A through E) and EBV-transformed (lanes A' through E') nonmalignant tonsillar B lymphocytes from a second patient. The following MoAb were employed: antiCALLA-1 (J5), lanes A and A'; the HLA-DR-specific MoAb p4.1, lanes B and B'; the HLA-DR-specific MoAb ISCR3, lanes C and C'; the HLA-DP-specific MoAb B7/ 21, lanes D and D'; and the HLA-DQ-specific MoAb HK19, lanes E and E'.

860

Electrophoretically distinct HLA-DR molecules on autologous B cells leukemic peripheral blood B cells m a y reflect leukemia-specific processes, or simply the different stages of B lymphocyte maturation represented by these two cell populations. For example, the surface and cytoplasmic m a r k e r s of C A L L A / C D 1 0 and CD19 positive A L L cells closely resemble those of a small population of early B cells in normal bone m a r r o w [21,22,24, 26-28]. These cells also lack surface immunoglobulin (slg) and are usually CD20 negative. In contrast, the normal peripheral blood B cells used in these studies possess slg and other surface markers (e.g. CD20) characteristic of m o r e mature B cells [21-25]. Since even small changes in the protein or carbohydrate portions of H L A - D R molecules can profoundly affect their function [14-18], these electrophoretically distinct H L A - D R molecules may be biologically significant. These different H L A - D R molecules m a y signal a n d / o r participate in unique stage-specific cell--cell interactions. REFERENCES 1. Benacerraf B. (1981) Role of MHC gene products in immune regulation. Science 212, 1229. 2. Giles R. C. & Capra J. D. (1985) Structure, function, and genetics of human class II molecules. Adv. lmmun. 37, 1. 3. Shackelford D. A., Kaufman J. F., Korman A. J. and Strominger J. L (1982) HLA-DR antigens: structure, separation of populations, gene cloning, and function. Immun. Rev. 66, 133. 4. Graf L. & Pesando J. M. (1987) Evidence for chemical differences in HLA-DR molecules on autologous acute lymphoblastic leukemia and B-lymphoblastoid cell lines. Blood 69, 7. 5. Pesando J. M., Ritz J., Lazarus H., Costello S. B., Sallan S. E. & Schlossman S. F. (1979) Leukemiaassociated antigens in ALL. Blood 43, 1240. 6. Lum L. G., Muchmore A. V., Keren D., Decker J., Koski I., Strober W. & Blaese R. M. (1979) A receptor for IgA on human T lymphocytes. J. lmmun. 122, 65. 7. Schwaber J., Lazarus H. and Rosen F. S. (1978) Bone marrow-derived lymphoid cell lines from patients with agammaglobulinemia. J. clin. Invest. 62, 302. 8. Nepom G. T., Nepom B. S., Antonelli P., Mickelson E., Silver J., Goyert S. M. & Hansen J. A. (1983) The HLA-DR4 family of hapiotypes consists of a series of distinct DR and DS molecules. J exp. Med. 159, 394. 9. Watanabe M., Suzuki T., Taniguchi M. & Shinohara N. (1983) Monoclonal anti-Ia murine alloantibodies crossreactive with the Ia-homologues of other mammalian species including humans. Transplantation 36, 712. 10. Pesando J. M. & Graf L. (1986) Differential expression of HLA-DR, -DQ, and -DP antigens on malignant B cells. J. Immun. 136, 4311. 11. Shipp M. A., Schwartz B. D., Kannapell C. C., Griffith R. C., Scott M. G., Ahmed P., Davie J. M. & Nahm M. H. (1983) A unique DR-related B cell differentiation antigen. J Immun. 131, 2458. 12. Watson A. J., DeMars R., Trowbridge I. S. & Bach F. H. (1983) Detection of a novel human class II HLA antigen. Nature, Lond. 304, 358.

861

13. Pesando J. M., Nadler L. M., Lazarus H., Tomaselli K. J., Stashenko P., Ritz J., Levine H., Yunis E. J. & Schlossman S. F. (1981) Human cell lines express multiple populations of Ia-like molecules. Hum. Immun. 3, 67. 14. McKenzie I. F. C., Morgan G. M., Sandrin M. S., Michaelides M. M., Melvold F. W. & Kohn H. I. (1979) A new H-2 mutation in the I region of the mouse. J. exp. Med. 150, 1323. 15. Maclntyre K. A. & Seidman J. G. (1984) Nucleotide sequence of mutant I-Abml2 gene is evidence for genetic exchange between immune response genes. Nature, Lond 308, 551. 16. Cohn L. A., Glimcher L. H., Waldmann R. A., Smith J. A., Ben-Nun A., Seidman J. G. & Choi E. (1986) Identification of functional regions on the I-Ab molecule by site-directed mutagenesis. Proc natn. Acad. Sci. U.S.A. 83, 747. 17. Glimcher L. H., Kim K.-Y., Green I. & Paul W. E. (1982) Ia antigen-bearing B cell tumor lines can present protein antigen and alloantigen in a major histocompatibility complex-restricted fashion to antigenreactive T cells. J. exp. Med. 155, 445. 18. Cowing C. & Chapdelaine J. M. (1983) T ceils discriminate between Ia antigens expressed on allogeneic accessory cells and B cells: A potential function for carbohydrate side chains on Ia molecules. Proc. natn. Acad. Sci. U.S.A. 80, 6000. 19. Zacheis M., Giacoletto K. & Schwartz B. D. (1986) Analysis of normal human tonsil class II antigen glycosylation by lectin affinity chromatography. J. biol. Chem 261, 17004. 20. Alexander S., Hubbard S. C. & Strominger J. L. (1984) HLA-DR antigens of melanoma and B-lymphoblastoid cell lines: Differences in glycosylation but not protein structure. J. Irnmun. 133, 315. 21. Vogler L. B., Crist W. M., Bockman D. E., Pearl E. R., Lawton A. R. & Cooper M. D. (1978) Pre-B-cell leukemia: a new phenotype of childhood lymphoblastic leukemia. New Engl. J. Med. 298, 872. 22. Janossy G., Bollum F. J., Bradstock K. F., McMichael A., Rapson N. & Greaves M. F. (1979) Terminal transferase-positive human bone marrow cells exhibit the antigenic phenotype of common acute lymphoblastic leukemia. J. lmmun. 123, 1525. 23. Nadler L. M., Ritz J., Hardy R., Pesando J. M. & Schlossman S. F. (1981) A unique cell surface antigen identifying lymphoid malignancies of B cell origin. J. clin. Invest. 67, 134. 24. Nadler L. M., Korsmeyer S. J., Anderson K. C., Boyd A. W., Slaughenhoupt B., Park E., Jensen J., Coral F., Mayer R. J., Sallan S. E., Ritz J. & Schlossman S. F. (1984) B cell origin of non-T cell acute lymphoblastic leukemia: a model for discrete stages of neoplastic and normal pre-B cell differentiation. J. clin. Invest. 74, 332. 25. Korsmeyer S. J., Arnold A., Bakhshi A., Ravetch J. V., Siebenlist, U., Hieter P. A., Sharrow S. O., LeBien T. W., Kersey J. H., Poplack D. G., Leder P. & Waldmann T. A. (1983) Immunoglobin gene rearrangement and cell surface antigen expression in acute lymphocytic leukemias of T cell and B cell precursor origins. J. clin. Invest. 71, 301. 26. Greaves M. & Janossy G. (1978) Patterns of gene expression and the cellular origins of human leukemias. Biochem. biophys. Acta 516, 193.

862

J.M. PESANDOet al.

27. Pesando J. M., Tomaselli K. J., Lazarus H. & Schlossman S. F. (1983) Distribution and modulation of a human leukemia-associated antigen (CALLA). J. lmmun. 131, 2038.

28. Hokland P., Nadler L. M., Griffin J. D., Schlossman S. F. & Ritz J. (1984) Purification of common acute lymphoblastic leukemia antigen positive cells from normal human bone marrow. Blood 64, 662.