Formation to Leukemic Cell Lines Binding of ...

Binding of Monoclonal Antibodies That Inhibit Spleen Colony Formation to Leukemic Cell Lines Stephen J. Ralph, Peter McCaffery, An Seng Tan, et al. Cancer Res 1984;44:3825-3830. Published online September 1, 1984.

Updated Version

E-mail alerts Reprints and Subscriptions Permissions

Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/44/9/3825

Sign up to receive free email-alerts related to this article or journal. To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at [email protected]. To request permission to re-use all or part of this article, contact the AACR Publications Department at [email protected].

Downloaded from cancerres.aacrjournals.org on July 14, 2011 Copyright © 1984 American Association for Cancer Research

[CANCER RESEARCH 44,3825-3830,

September 1984]

Binding of Monoclonal Antibodies That Inhibit Spleen Colony Formation to Leukemic Cell Lines1 Stephen J. Ralph, Peter McCaffery, An Seng Tan, and Michael V. Berridge2 Wellington Cancer and Medical Research Institute, Wellington Clinical School of Medicine, Wellington Hospital, Wellington 2, New Zealand

ABSTRACT Clonogenic tumor cells and normal stem cells share the prop erty of extensive proliferative potential. Normal stem cells are under stringent growth restraint and respond to appropriate differentiation signals, whereas tumor stem cells have lost the ability to respond normally to these controls. In an attempt to define cell surface molecules involved in the control of hemopoietic cell proliferation and differentiation, we have produced 5 monoclonal antibodies against antigens held in common between hemopoietic stem cells and the Abelson virus-induced pre-Blymphoma cells from which they were derived. Four of these monoclonal antibodies produced greater than 90% reduction of spleen colony-forming cells, whereas the other bound to a subpopulation (60 to 70%) of spleen colony-forming cells at plateau values. The expression of antigens recognized by these and two other anti-stem cell monoclonal antibodies has been shown to correlate with the differentiation status of a panel of tumor cell lines, with greater expression being observed on cells more closely resembling the pluripotent stem cell than mature hemo poietic cells. Immunoperoxidase staining of bone marrow showed that these antigens are mainly expressed by monocytes and blast cells. Treatment of bone marrow cells with those antibodies which extensively inhibited spleen colony formation and with rabbit complement abolished the ability of progenitor cells to form colonies in soft agar. Quantitative absorption studies distinguished the antigens recognized by two of the anti-stem cell monoclonal antibodies from those detected by anti-H-2k 114.1 monoclonal antibody. These observations suggest that the antigens involved may play a role in the regulation of growth and differentiation of stem cells and undifferentiated leukemic cells. INTRODUCTION The concept that tumors arise as a result of altered control of normal cell differentiation processes has been developed in several models involving blocked ontogeny (22), blocked differ entiation (17), or an uncoupling of proliferation and differentiation (26). In these models, cells may accumulate in a particular state of differentiation because a differentiation-inducing signal or its cell surface receptor is absent or unable to transmit the appro priate message. In support of this model, lymphoid tumors with cell surface antigens characteristic of several stages in the differentiation pathway have been described in both human (7, 17,19) and animal (15,31) systems. Furthermore, several tumor cell types can be forced to bypass their characteristic differentia tion state with appropriate stimulation (8, 25, 26). In general, however, the cellular alterations which result in altered growth 1This work was supported by the Malaghan Trust, the Cancer Society of New Zealand, and the Wellington Branch of the Cancer Society of New Zealand. 2 To whom requests for reprints should be addressed. Received May 2,1983; accepted June 7,1984.

SEPTEMBER

and arrested differentiation are largely unknown. Hemopoietic stem cells and tumor stem cells share common properties. Both show extensive self-renewal and an ability to respond to differentiation inducing signals (8, 18, 25). Changes in cell surface receptors for growth- and differentiation-inducing signals may be central to tumorigenesis. Recently, cell surface antigens which may be involved in the control of stem cell proliferation and differentiation have been described (1-3). Thus, heterologous antisera against differentiated hemopoietic cells have demonstrated the presence of a new class of cell surface antigens, the cell lineage antigens, the presence of which re stricted to a particular differentiation pathway. These lineagespecific antigens are expressed on the pluripotent stem cell (CFU-s)3 and are thought to become progressively restricted to one pathway of differentiation as differentiation proceeds. Al though the role of these antigens is not known, their unique occurrence on hemopoietic cells suggests a possible function in stem cell commitment and differentiation. Attempts to further characterize the nature of these antigens with monoclonal anti bodies have met with limited success (24). Two monoclonal antibodies were found to cross-react with CFU-s, but lineage specificity was difficult to demonstrate because of low binding to mature cells. It is possible that less mature cells may express higher concentrations of cell lineage antigens and thus be better immunizing cell types. Because of the difficulties in obtaining pure populations of precursor cells, tumor cells containing a more homogeneous, undifferentiated phenotype were used as suggested by Lanier ef al. (14,15). In the present study, several further monoclonal antibodies against the stem cell have been derived by immunizing rats with the Abe 8.1 pre-B-lymphoma, a tumor cell line closely related to the stem cell. Monoclonal antibodies which inhibit spleen colony formation are shown to bind more strongly to tumor cells blocked in an early stage of differentiation. MATERIALS AND METHODS Animals. CBA or CBA x C57BL/6 F, mice (2 to 3 months old) and Fischer rats were bred at Wellington Clinical School. BALB/c mice were obtained from the National Health Institute, Wellington, New Zealand. CBA mice were used as donors in the spleen colony assay and for preparing erythrocytes, platelets, B-lymphocytes, thymocytes, spleen cells, and bone marrow cells for absorption or binding studies. Male or female CBA or CBA x C57BL/6 F, mice were used as recipients in the spleen colony assay without noticeable difference when used inter changeably. The Abe 8.1 BL cell line was passaged every 2 weeks in BALB/c mice by s.c. injection in the shoulder with 2 x 10s cells in PBS. EL-4 was passaged in C57BL/6 mice by i.p. injection of 2 x 10* cells in PBS. Cell Preparation

and Absorption. MSBSS (3) or PBS was used

3 The abbreviations used are: CFU-s, spleen colony-forming unit; Abe 8.1 BL, Abelson virus-induced pre-B-lymphoma; MSBSS, mouse isoosmotic buffered bal anced salt solution; PBS, phosphate-buffered saline.

3825

1984


>.-'*â€¢&â€¢-â€¢*. ?, I*

S. J. Ralph et al. throughout. Platelets, thymocytes, and erythrocytes were prepared as described previously and were pure cell populations (>99.9%). Bone marrow cells were flushed from the femur and tibia as a single-cell subpopulation and washed with MSBSS. Spleen cells were prepared by teasing spleen through a fine-meshed gauze into MSBSS, dispersing cells by pipeting, allowing debris to settle for 5 min, and then washing the cell suspension 3 times with MSBSS. B-lymphocytes were purified from spleen cells that had been depleted of erythrocytes on IsopaqueFicoll (p = 1.09). These cells were adsorbed on to plastic Retri dishes that had been coated with an IgG fraction of rabbit anti-mouse IgG (Miles

Table 1 Effect of thymocytes on anti-stem cell activity

of toss of antiof loss Anticytes/ spleen colo of spteen stem cell mouse2x nies/mouse28.6 colonies087789378958588906364% activity101610-2-2 bod/ControlserumControlserumABrSABrSATSATSAPSAPS9F69F613C613C6Dilution1:161:16 2.3*28.7 Â± 1072x

Â±1.93.8

horse serum (Grand Island Biological Co.). Abe

0.86.4 Â± 1072X1072x1072x1072X107Av. 1.12.0 Â± 0.66.2 Â± Â±1.51.4 Â±0.74.2 0.93.4 Â± 1.13.0 Â± 0.510.6 Â± Â±1.310.2 Â±1.0% * Antibrain serum (ABrS), antithymocyte serum (ATS), and antiplatelet serum

8.1 BL was obtained from Dr. J. McCarthy (Walter and Eliza Hall Institute), while EL-4 and P815 were obtained from Dr. J. Marbrook (Auckland University). Abe 8.1 BL and EL-4 were passaged in vivo, while

(APS) were absorbed with an equal volume of CBA mouse erythrocytes prior to use. Mean Â±S.E.

other tumor cell lines were grown in vitro. All cells were washed in MSBSS prior to use. Absorptions were carried out at 4Â°for 45 min using undiluted hybridoma supernatants or monoclonal anti-H-2Kk clone 11-

were: platelets, 0.07; B-lymphocytes, 1.3; Abe 8.1 BL, 1.6; 416B, 1.6; EL-4,1.7; BW5147, 2.4; P815, WEHI-3, and NS-1, 3.8. 125l-lgGrabbit anti-rat IgG (15 to 25 x 103 cpm/well, 4 to 8 x 107 com/

Laboratories, Elkhart, IN) according to the method of Wysocki and Sato (32). By Â¡mmunofluorescence, B-lymphocytes prepared in this way were 90 to 94% pure and contained less than 2% Thy-1-positive cells. The hemopoietic cell lines 416B, BW5147, WEHI-3, and NS-1 were obtained from Drs. G. Johnson, A. Harris, and P. Bartlett (Walter and Eliza Hall Institute, Melbourne, Australia) and were cultured in Dulbecco's modified Eagle's medium (Grand Island Biological Co., Grand Island, NY) contain ing 10% heat-inactivated

4.1 (Becton-Dickinson,

Sunnyvale, CA) diluted 1:16. After absorption,

antibodies were diluted to a selected point close to their titration range (12 to 32 times the 50% tier for stem cells) prior to assay for spleen colony inhibition. Immunization and Fusion. Fischer rats were inoculated i.m. with 2 to 3 x 107 Abe 8.1 BL cells emulsified in complete Freund's adjuvant and thereafter After 4 to anti-stem inoculated of immune

injected i.p. at weekly intervals with a similar number of cells. 6 weeks, the rats were bled, and the serum was tested for cell activity. Those rats showing the greatest activity were daily for 3 to 4 days prior to fusion. A single-cell suspension spleen cells was fused with NS-1 plasmacytoma cells accord

ing to the method of Galfre ef al. (9) as modified by Claflin and Williams (5). Selected hybridomas were cloned in soft agar, retested, and stored in liquid nitrogen. Spleen Colony Inhibition Assay and Hybridoma Screening. Growthpositive cells were initially selected on the basis of high growth and high levels of IgG production in hybridoma supernatants using IgG rabbit antirat IgG (Miles Laboratories) adhered to plastic microtiter wells (32). Hybridoma screening for inhibition of spleen colony formation was mea sured by the method of Till and McCulloch (30) as described previously (1). Up to 4 hybridoma supernatants were combined and incubated with an equal volume of bone marrow cells (3x107/rnl) prior to dilution and injection into single lethally irradiated mice (850 R; 300 R/min). Mice showing greater than 40% loss of spleen colonies after 7 days indicated hybridoma supernatants with anti-stem cell activity. Individual superna tants were then screened for anti-stem cell activity in groups of 3 to 5 mice. That the anti-stem cell activity observed in the spleen colony assay when bone marrow cells were treated with rabbit antisera or monoclonal antibodies was not caused by an indirect effect of antibody on accessory cells required for spleen colony formation as has been suggested by others (20, 23) is shown in Table 1. Thymocytes injected into lethally irradiated mice 1 hr prior to antibody-treated bone marrow cells had little effect on the inhibition of spleen colony formation. Furthermore, extensive absorption analysis of the anti-stem cell activity in anti-thymocyte serum, anti-platelet serum, and anti-macrophage serum whereby only the inoc ulating cell type removed the anti-stem cell activity (1-3) argues against an indirect effect of antibody on accessory cells. Radioimmunoassay. Hybridoma supernatants were screened against cells in a solid-phase radioimmunoassay (28) using a standardized num ber of cells per flat-bottomed microtiter well. The relative number of cells plated per well was determined according to cell surface area which was calculated from the average cell diameter measured microscopically or from published data (1). Relative to thymocytes other cell surface ratios

MQ)was used to detect rat monoclonal antibodies bound to immobilized cells, background binding being determined in each experiment with an inappropriate hybridoma supernatant. Immunoperoxidase Staining and Fluorescent Antibody Binding Analysis. The binding of anti-stem cell monoclonal antibodies to bone marrow cells was determined (a) by indirect staining with horseradish peroxidase conjugated to rabbit anti-rat IgG (Miles Laboratories) as outlined by Swirsky ef al. (29) and (b) by indirect fluorescent analysis using affinity-purified mouse anti-rat IgG which had been absorbed with an equal volume of mouse spleen cells after conjugation with fluorescein (10). Fluorescent antibody binding was detected on both viable and fixed cell preparations using a Zeiss Photomicroscope II with epifluorescence or a fluorescence-activated cell sorter (FACS IV; Becton-Dickinson, Sunnyvale, CA).

RESULTS Monoclonal Antibodies with Anti-Stem Activity. Rats were immunized with Abe 8.1 BL cells, and those which exhibited the strongest anti-stem cell activity were chosen for fusion. In 2 separate experiments involving 3500 growth-positive hybrido mas. 1108 were selected on the basis of either immunoglobulin production or strong growth for screening against the stem cell. The efficiency of selecting hybridomas secreting monoclonal antibodies with anti-stem cell activity was greatly increased by pooling up to 4 supernatants for assay against bone marrow stem cells. Injecting relatively large numbers of antibody-treated stem cells (producing 25 to 50 spleen colonies in controls) into single mice enabled effective isolation of hybridomas which were then retested individually. Of 8 supernatants exhibiting greater than 40% spleen colony inhibition, 5 were isolated and cloned as stable cell lines. Each of these monoclonal antibodies was titrated against the stem cell. Table 2 summarizes the plateau level of spleen colony inhibition and the titer of each monoclonal antibody and compares these values with data for several other monoclonal antibodies known to inhibit spleen colony formation and with 30G12 (16) which binds to the T200 family of glycoproteins. 1A9,13C6, and 1C10 resulted in 46 to 72% loss of spleen colonies at plateau values whereas 9F6, 9F9, 15E10, and 16F7 produced 90 to 97% spleen colony inhibition.

3826

CANCER RESEARCH


VOL. 44

Surface Antigens on Stem Cells and Leukemic Cell Lines Table 2 Characteristics of monoclonal antibodies exhibiting anti-stem cell activity of loss of spleen colo globulin (rangeat nies isotype"lgG2b )92-97 plateau

clonal an typeAbe Inoculating cell tibody9F6 B-lymphomaThymocytes 8.1 NDC 9F9 03 15E10 F8C11 ND ND 16F7 1A9 B8 lgG2b 13C6" A1 lgG2b 1C10" lgG2b C5 Platelets H-2K" " 11-4.1tmmunolgG2a Spleen cells 30G12CloneH2 lgG2a% ThymocytesMono

observed with a non-immunoglobulin-secreting hybridoma whereas 14.8 bound only to B-lymphocytes. Binding of anti-stem cell monoclonal antibodies to bone mar

row and spleen cells was compared with binding to thymocytes for CFU-s2048and platelets using immunoperoxidase-conjugated rabbit anti-rat IgG and fluorescein-conjugated

90-97 104 91-95 32 92-97 675 59-71 1024 52-72 74 46-71 23 78-84 1024 0-24Trter 0

mouse anti-rat IgG. The results

presented in Table 5 show that, with the exception of 15E10 which exhibited a low titer for CFU-s (see Table 2), each of the anti-stem cell monoclonal antibodies bound significantly to each

cell type as determined by fluorescence intensity profiles relative to cells treated with control ascites or NSI culture supernatant. Immunoperoxidase staining showed a higher fraction of positive 8 Antibodies typed using rat subclass typing sera (Miles Laboratories). cells in the bone marrow than in the spleen with each antibody 6 Data summarized from titration curves and from assays carried out at plateau tested, with 9F6, 9F9, and 16F7 showing greater binding than values. c ND, not determined. did 13C6, 1C10, and 1A9. Except for 13C6 and 1C10, each " Data from Ralph ef al. (24). monoclonal antibody showed a similar pattern of binding indi cated by the percentage of the different cell types stained in Table3 bone marrow. Monocytes and blast cells bound more antibody Binding of monoclonal antibodies to bone marrow progenitor cells than did lymphocytes, followed by myeloblasts and myelocytes bone of reduction of and finally by metamyelocytes, polymorphonuclear cells, and treatment*Control Antibody marrow cells238 colonies-16-71133 erythroblasts. With 13C6 and 1C10, blast cells appeared to stain Â± 8C Control RC'13C6 + more weakly than did other cell types. Strong intracellular stain 241 7223Â± ing was observed when monocytes and blast cells were labeled Â±10 with 9F6. RC'1C10 13C6 + 256 Â± 14213 Binding of Monoclonal Antibodies to Hemopoietic Tumor Â±18 Cell Lines. Table 6 summarizes the plateau binding of each RC'1A9 1C10 + 245 Â±12230 monoclonal antibody to a panel of leukemic cell lines. The highest level of binding with each of the monoclonal antibodies with antiÂ±10 RC'9F6 1A9 + 213 Â± 4238 110 stem cell activity regardless of the immunizing cell type was observed with 416B cells. Abe 8.1 BL, BW5147, and WEHI-3 Â± 9 9F6 RC'H-2K* +

1235 1 Â±

1001100

11-4.1 H-2K* 11-4.1 + RC'Colonies/105

Â±10 0% '' Hybridoma supematants and control NSI culture medium were used undiluted. Rabbit complement was used at a final dilution of 1:3. The H-2K" 11-4.1 cell line was obtained from the American Type Culture Collection. b Total colonies in agar were determined as described previously (3) using optimum concentrations of pokeweed mitogen-stimulated spleen-conditionedme dium. c Mean Â±S.E.

Effect of Monoclonal Antibodies on in Vitro Colony Forma tion by Bone Marrow. Using a complement-dependent cytotoxic assay, 9F6 hybridoma supernatant completely abolished colony formation by progenitor cells in bone marrow, as did anti-H-2Kk culture supernatant (Table 3). 13C6, 1C10, and 1A9 culture supematants, however, had no significant effects on colony growth. More recently, we have demonstrated that ascites fluid from 1C10 and 1A9 hybridomas inhibits colony formation in a complement-dependent manner by 36 and 100%, respectively but have not been successful in raising ascites with 13C6 hybri doma cells. Binding of Monoclonal Antibodies to Normal Hemopoietic Cells. Mybridoma supematants with anti-stem cell activity were titrated against B-lymphocytes, thymocytes, and platelets in a solid-phase radioimmunoassay. Plateau levels of binding are summarized in Table 4. With the exception of 30G12 and 14.8, which recognizes B220 (12), each of the monoclonal antibodies exhibited weak but positive binding to B-lymphocytes. Binding to thymocytes and platelets was either weakly positive or neg ative. 30G12 bound in high levels to B-lymphocytes and thymo cytes but showed no binding to platelets above the background SEPTEMBER

Table 4 Plateau binding of monoclonal antibodies to normal hemopoietic cells determined by radioimmunoassay Monoclonal B-lymphocytes13C6 antibody +*1C10 +1A9 Â±9F6 ++9F9 +15E10 +16F7 30G12-T200" 14.8-8220Â°

+ ++++ ++++Thymocytes+++Â±Â±+Â±Â¿Â»Platelets-++Â±+Â±-

' -, binding not above background; Â±,0 to 50 cpm; +, 50 to 100 cpm; ++, 100 to 200 cpm; +++, 200 to 300 cpm; ++++, over 300 cpm above background. " Obtained from Dr. I. McKenzte, University of Melbourne. '' Obtained from Dr. P. Kincade, Sloan-Kettering Institute, New York, NY.

TaWe5 Binding of anti-stem cell monoclonal antibodies to hemopoietic cells determined by IF* staining and fluorescence analysis cellsMonoclonal

% of positive marrowIP5

antibodyControl

(FACS)>90(14)

(9) 13C6 39 1C10 >95 (20) >95 (13) >90 (7) 51 23 7 >90(11) >95(16) >95 (15) >90 (9) 1A9 18 33 8 >90(12) >90 (4) 9F6 >95 (21) >95 (23) >95 (14) 90 49 10 NDC >95 (23) >95 (25) >90 (5) 9F9 ND ND 70(16) 0 (1) 15E10 13 9 >90 (7) 6 >90 (3) 0 (0) >95 (21)SpleenIP232 >90(14) 16F7Bone 90FACS>95(15f 45FACS>95(14) >95 (25)ThymocyteIP27 10FACS>90 >90 (5)Platelets * IP, immunoperoxidase. FACS, fluorescence-activated cell sorter. " Numbers in parentheses, logarithm of the peak fluorescence intensity relative to controls (arbitrary scale). platelets > erythrocytes, the absorption order with anti-H-2K" was thymo

to 416B cells, which show properties of an erythroleukemia although originally being able to form multipotential spleen colo nies (6). Quantitative Absorption Studies with 9F6,13C6, and H-2K"

cytes > erythrocytes > platelets.

Monoclonal Antibodies. The binding assays described above have involved cell fixation, procedures which may influence an tibody binding. An alternative approach to the problem not involving cell fixation is to examine the ability of mature cells to absorb the anti-stem celt activity. Chart 1 shows that extensive absorption with thymocytes removed most of the anti-stem cell

DISCUSSION Hemopoietic stem cells and tumor stem cells hold in common the properties of extensive self-renewal. However, differences in their ability to respond to differentiation-inducing stimuli may be fundamental to the tumor state. The possibility that normal and tumor stem cells may express cell surface antigenic determinants in common relating to their proliferative and differentiation poten tial was investigated in this study. Thus, antigenic determinants held in common between a primitive tumor line, Abe 8.1 BL, and the hemopoietic stem cell were investigated by producing mon oclonal antibodies common to these 2 cell types. All 5 of these anti-stem cell antibodies and 2 others derived previously by

activity from 9F6 and 1A9 supematants and marginally reduced the anti-stem cell activity in 13C6. Platelets and erythrocytes Table 6 Plateau binding of monoclonal antibodies to hemopoietic tumor cell lines determined by radkxmmunoassay Monoclonal antibodies

416B*

BW5147

Abe 8.1 BL

WEHI-3

EL-4

13C6

P815

NSI

inoculation of rats with platelets and thymocytes (24) were found to be expressed in greater concentration on teukemic cell lines more closely resembling the hemopoietic stem cell than more mature hemopoietic cells. Improved methods for screening large numbers of hybridoma supematants against stem cells increased the detection fre quency of hybridomas exhibiting anti-stem cell activity to 4 to 5/ 1000 screened for spleen colony inhibition. That the monoclonal antibodies described in this study bind directly to hemopoietic stem cells is indicated by several lines of evidence: (a) cy totoxici ty studies with bone marrow progenitor cells (Table 3) suggest direct binding of 9F6 to primitive cells in the bone marrow; (b)

Â± Â± Â±

1C10

1A9 9F6 9F9

Â± Â± Â±

15E10

16F7 30G12-T200

*416B erythroleukemia was obtained from Dr. G. Johnson, BW5147 pre-Tlymphoma and WEHI-3 myelomonocytic leukemia were from Dr. A. Harris, Abe 8.1 pre-B-tymphoma was from Dr. J. McCarthy, and NS-1 was from Dr. P. Bartlett. Walter and Eliza Hall Institute, Melbourne. EL-4 T-lymphoma and P815 mastocytoma were from Dr. J. Marbrook, Department of Cell Biology, University of Auck land. 0 For definition of symbols, see Table 4, Footnote a.

Chart 1. Quantitative absorption of anti-stem cell monoclonal antibodies with normal hemopoietic cells. Hybridoma supematants were absorbed consecutively with an equal volume (IV) of packed cells and samples analyzed for anti-stem cell activity in the spleen colony assay using groups of 5 mice. The average standard error was Â±7%. A, 9F6; 8, 1A9; C, 1C10; D. 13C6. O, platelets; â€¢, thymocytes; A, erythrocytes.

u. u

20

-20

-

-

1Â«IV

2Â»W

3 Â»IV

IxlV

3828

1xW

2Â«IV

3Â«IV

CANCER RESEARCH


IxlV

VOL. 44

-f'.ff

V OÃ->~'

Surface Antigens on Stem Cells and Leukemic Cell Lines lymphocytes, thymocytes, or platelets (i.e., to their inoculating cells or nontransformed cell counterparts (Table 4). In contrast, all 7 monoclonal antibodies showed high binding to tumor cells blocked in an early stage of differentiation. Although the exact differentiation state of 416B is unknown, the fact that immedi ately following derivation it was able to form spleen colonies (6) suggests that it may be the most primitive cell line tested. Both BW5147, a T-cell lymphoma (27), and Abe 8.1 pre-B-lymphoma,

20

2Â«IV

3xlV

Absorbing cells Chart 2. Quantitative absorption of antJ-H-2IC monoclonal antibody with normal hemopoietic cells. Purified stock antibody (0.5 mg/ml) was diluted 1:16 and ab sorbed as described in Chart 1. The average standard error was Â±7%.O, platelets; â€¢,thymocytes; A, erythrocytes.

immunoperoxidase staining and immunofluorescence analysis indicate binding to most bone marrow cells including blast cells (Table 5); (c) supplementation of irradiated mice with accessory thymocytes prior to antibody-treated bone marrow cells did not alter the anti-stem cell activity of 9F6 and 13C6 monoclonal antibodies (Table 1). (d) Absorption analysis of the anti-stem cell activity in cell lineage antisera indicate a direct effect of antibody on stem cells (1-3). Four of the anti-stem cell monoclonal antibodies produced greater than 90% reduction of CFU-s, whereas 1A9 bound to a subpopulation (60 to 70%) of CFU-s at plateau values and in this regard was similar to 2 other monoclonal antibodies described previously (24). Attempts to increase the plateau anti-stem cell values of 13C6, 1C10, and 1A9 with 10- to 20-fold ammonium sulfate concentrates of hybridoma supernatants, 100-fold con centrates from serum-free culture supernatants, ascites produc tion with titers between 104 and 105, and use of developing rabbit anti-rat IgG serum and/or 17-day-old rabbit complement did not markedly alter spleen colony inhibition. Initial immunoprecipitation experiments with 9F6, 1A9, 13C6, and 1C10, using 125I-and ^S-methionine-labeled cells, failed to identify specific cell surface molecules on polyacrylamide gels. It is possible that molecules expressing these antigens may be devoid of tyrosine residues, have low turnover, be bound covalently to the cytoskeletal structure, or be nonpolypeptidic. In the present studies, we have used a sensitive radioimmunoassay based on the trace assay system of Morris and Williams (21) which is capable of detecting a few hundred antigenic determinants per cell (50 to 100 cpm above background) to measure monoclonal antibody binding to purified cell populations and tumor cell lines. Because the 125l-rabbit anti-rat IgG probe is of high specific activity and is used in trace amounts, it becomes limiting above a few thousand antlgenic sites per cell and is not quantitative above this level. Except for 30G12-T200 and 14.8-B220, which were used as positive antibody-binding controls, most other binding results are within the sensitivity range of the radioimmunoassay (Tables 4 and 6). None of the anti-stem cell monoclonal antibodies exam ined by radioimmunoassay showed high levels of binding to BSEPTEMBER

which has undergone heavy and light chain rearrangement but does not secrete immunoglobulin or express surface immuneglobulin (4, 15), are thought to be derived from early in their respective lineages whereas NS-1, EL-4, and P815 express surface antigens characteristic of mature cells (15, 27). WEHI-3 myelomonocytic leukemia which is FcR positive and Thy-1.2 positive (27) is probably characteristic of cells intermediate in the myelomonocytic lineage. These binding results show trends similar to other results obtained with xenoantiserum raised against a mast cell tumor line FMPI.1 (11), which showed higher binding to immature cells of the bone marrow including granulocyte-macrophage colony-forming cells, erythroid bursts and CFU-s. Quantitative absorption studies clearly distinguished 9F6 and 1A9 from 13C6 and 1C10 according to their inoculating cell derivation. Thus, 13C6 and 1C10 supernatants showed little loss of anti-stem cell activity even after 3 to 4 consecutive absorptions with an equal volume of packed erythrocytes, platelets, or thy mocytes whereas the anti-stem cell activity in 9F6 and 1A9 supernatants was removed by extensive absorption with thy mocytes and partially removed with platelets and erythrocytes (Chart 1). Thus, although the quantitative absorption results confirm the general conclusions of the radioimmunoassay that mature hemopoietic cells express low levels of antigens recog nized by 9F6,1A9,1C10, and 13C6 monoclonal antibodies, it is not clear why the sensitive radioimmunoassay did not detect significant levels of binding of 9F6 and 1A9 to thymocytes and platelets as would have been predicted from the absorption results (Chart 1). Also, the differential absorption of 9F6, 1A9, 13C6, and 1C10 by thymocytes and platelets was not evident in immunoperoxidase staining or immunofluoresence results (see Table 6). These results suggest that cell fixation may be affecting antibody binding in some of these assays. The quantitative absorption profiles of 9F6 and 1A9 showed clear differences from those of anti-H-2Kk 11-4.1 [Charts 1 and 2; see also the paper of Klein (13)], indicating that these anti-stem cell mono clonal antibodies are not binding to known specificities of the Class I major histocompatibility complex. Immunoperoxidase staining of bone marrow cells treated with anti-stem cell monoclonal antibodies showed higher binding to monocytes than to lymphocytes, maturing granulocytes, and erythroid cells. Undifferentiated blast cells were strongly stained with 9F6 and 16F7 but not with 13C6 and 1C10. Using the more sensitive technique of fluorescence-activated cell sorter analysis, significant binding of anti-stem cell monoclonal antibodies (ex cept 15E10) was demonstrated to all of the hemopoietic cells tested (Table 6), with higher binding being observed to bone marrow and spleen cells than to thymocytes and platelets. The nature of the antigens held in common between primitive tumor cells and stem cells which are recognized by the anti-stem cell monoclonal antibodies described in this study remains unknown, although their widespread presence on immature cell types suggests a possible role in self-renewal and differentiation. Mon3829

1984


S. J. Ralph et al. oclonal antibodies against tumor cells and stem cells may prove to be useful for immunologically manipulating the hemopoietic system in vivo and for immunotherapy of leukemias, because their stem cell-binding properties may be advantageous for the elimination of clonogenic tumor cells. Other monoclonal antibod ies currently being used therapeutically recognized tumor cell differentiation antigens and may thus spare the tumor stem cell. ACKNOWLEDGMENTS The authors thank Drs. G. Johnson, A. Harris, P. Bartlett. J. McCarthy, and J. Marbrook for the tumor ceÂ«lines; Dr. I. McKenzie for the 30G12-T200 hybndoma cell line; and Or. P. Kincade for the 14.8-B220 monoclonal antibody.

REFERENCES 1. Berndge. M W. A new class of cell surface antigens. Quantitative absorption studies defining cell-lineage-specific antigens on hemopoietic cells. J. Exp. Med., 150: 977-986,1979. 2. Berndge. M. V. Surface antigens of murine hemopoietic stem cells. IV. Char acteristics of antisera exhibiting cell-lineage specificity and differentiating ability of stem cells surviving treatment with antisera against cell-lineage antigens. Stem Cells, 2: 229-245, 1982. 3. Berndge, M V., and O'Kech, N. Surface antigens of murine hemopoietic stem cells. 1. Cross reactivity of antisera against differentiated hemopoietic cells with bone marrow stem cells. Exp. Hematol . 7: 452-468,1979. 4. Boyd, A. W., and Schrader, J. W. Derivation of macrophage-like lines from the pre-B lymphoma ABLS 8.1 using 5-azacytidine. Nature (Lond.), 297:691-694, 1982. 5. Clallin, L , and Williams, K. Mouse myeloma-spleen cell hybrids. Enhanced hybridization frequencies and rapid screening procedures. Curr Top. Micrabiol lmmunol.,8i: 107-109,1979. 6. Dexter, T. M., Allen, T. D., Scott, D., and Teich, N. M. Isolation and character ization of a bipotential haematopoietic cell line. Nature (Lond,). 277:471-474, 1979. 7. Foon, K. A., Schroff, R. W., and Gale, R. P. Surface markers on leukemia and lymphoma cells: recent advances. Blood, 60:1-19,1982. 8. Friend, C. Murine virus-induced erythroleukemic cell lines. In: G. Sato (ed.), Functionally Differentiated Cell Lines, pp. 235-250. New York: Alan R. Liss, 1981. 9. Galfre. G., Howe, S. C . Milstein, C , Butcher, G. W., and Howard, J. C. Antibodies to major histocompatibility antigens produced by hybrid cell lines. Nature (Lond.), 266: 550-552,1977. 10. Coding, J. W. Conjugation of antibodies with fluorochromes: modifications to the standard methods. J. immunoi. Methods, 73: 215-226,1976. 11. Hasthorpe, S., and Rogerson, J. Tumor-associated surface membrane anti gens on CFU-s: characterization using the fluorescent activated cell sorter. Exp. Hematol., 10: 9-19,1982. 12. Kincade, P. W., Lee, G., Watanabe, T., Sun, L . and Scheid, M. P. Antigens displayed on murine B lymphocyte precursors. J. Immunoi., 727: 2262-2268,

1981. 13. Klein, J. Biology of the Mouse: HJstocompatibility-2 Complex, pp. 330-331. Berlin: Springer-Verlag, 1975. 14. Lamer, L. L, Walker, E. B., Richie, E. R., and Warner, N. L. Murine hematopoietic cell tumours: models for analysis of cellular differentiation. In: E. S. Vrtetta and C. F. Fox (eds.), B and T Cell Tumours, pp. 29-42. 15. Lamer, L. L., Warner, N. L., Ledbetter, J. A., and Herzenberg, L. A. Quantitative immunof Â¡uorescentanalysis of surface phenotypes of murine B cell lymphomas and plasmacytomas with monoclonal antibodies. J. Immunoi., 727: 16911697,1981. 16. Ledbetter, J. A., and Herzenberg, L. A. Xenogeneic monoclonal antibodies to mouse lymphoid differentiation antigens. Immunoi. Rev., 47: 64-90,1969. 17. Magrath. I. T. Lymphocyte differentiation: an essential basis for the compre hension of lymphoid neoplasia. J. Nati. Cancer Inst., 67: 501-514,1981. 18. Metcalf, D., Johnson, G. R., and Burgess, A. W. Direct stimulation by purified GM-CSF of the proliferation of multipotential and erythroid precursor cells. Blood, 55:138-147, 1980. 19. Mmowada. J., Sagawa, K., Trowbridge, I. S. Kung, P. D., and Goldstein, G. Marker profiles of 55 human leukemia-lymphoma cell lines. In: S. A. Rosenberg and H. S. Kaplan (eds.), Malignant Lymphomas, pp. 53-77. New York: Aca demic Press, Inc., 1982. 20. Monelle, F. C., and Wassef, W. Y. Thymocyte involvement in spleen colony formation by subpopulations of hematopoietic stem cells. Exp. Hematol., 2: 1011-1019,1981. 21. Morris, R. J., and Williams, A. F. Antigens on mouse and rat lymphocytes recognized by rabbit antiserum againsl ral brain: the quantitative analysis of a xenogeneic antiserum. Eur. J. Immunoi., 5: 274-281,1975. 22. Potter, V. R. The present status of the blocked ontogeny hypothesis of neoplasia: the thalassemia connection. Oncodev. Biol. Med., 2: 243-266, 1981. 23. Poverenny, A. M., Semina, O. V., Semenels, T. N., and Yarilin, A. A. Probable mechanism of spleen colony formation suppression with rabbit antimouse brain antiserum. Exp. Hematol.. 8:1216-1221,1980. 24. Ralph, S. J.. Tan, S. A., and Berndge, M. V. Monoclonal antibodies detect subpopulations of bone marrow stem cells. Stem Cells, 2: 88-107,1982. 25. Sachs, L. Control of normal cell differentiation and the phenotypic reversion of malignancy in myetoid leukaemia. Nature (Lond.), 274: 535-539,1978. 26. Sachs, L. Constitutive uncoupling of the controls for growth and differentiation in myetoid leukemia and the development of cancer. J. Nati. Cancer Inst., 65: 675-679,1980. 27. Shimizu, S., Smilh, R. T., Norcross, M. A., and Maino, V. C. Mechanisms controlling TCGF production by cloned subÃ®mes of EL-4 azcf in response to stimulation by anti-Thy-1 antibody. J. Immunoi., 728: 296-301,1982. 28. Stocker, J. W., and Heusser, C. H. Methods for binding cells to plastic. Application to a solid-phase radioimmunoassay for cell-surface antigens. J. Immunoi. Methods, 26: 87-95,1979. 29. Swirsky, D. M., Walt, S. M., Gilmore, D. J.. Hayhoe, F. G. J., and Waldmann, H. The characterization of monoclonal antibodies against haemopoietic cells: comparison of an immunoperoxidase method with fluorescence activated cell sorting. J. Immunoi. Melhods, 67:171-182,1983. 30. TiH,J. E., and McCulkxh, E. A. A direct measurement of the radiation sensitivity of normal mouse bone marrow cells. RadiÃ¢t.Res., 14:213-222,1961. 31. Warner, N. L. Neoplasms of immunoglobulin producing cells in mice. Recent Results Cancer Res., 64: 316-324.1978. 32. Wysocki, L. J., and Sato, V. L. Panning for lymphocytes: a method for cell selection. Proc. Nati. Acad. Sci. USA, 75: 2844-2848,1978.

3830

CANCER RESEARCH


VOL. 44