Human Monoclonal Antibodies Reactive with Human ...

(CANCER RESEARCH 49, 1665-1670, April 1, 1989)

Human Monoclonal Antibodies Reactive with Human Myelomonocytic Leukemia Cells1 Marshall R. Posner,2 David J. Santos, Hillary S. Elboim, Marea B. Tumber, and A. Raymond Frackelton, Jr. From the Division of Hematology/Oncology, Department of Medicine, Roger Williams General Hospital, The Roger Williams Cancer Center, and the Brown University Medical School, Providence, Rhode Island 02908

ABSTRACT Peripheral blood mononuclear cells from a patient with chronic myelogenous leukemia (CML), in remission, were depleted of CDS-positive T-cells and cultured with Epstein-Barr virus. Four of 20 cultures (20%) secreted human IgG antibodies selectively reactive with the cell surfaces of certain human leukemia cell lines. Three polyclonal, Epstein-Barr virus-transformed, B-cell lines were expanded and fused with the humanmouse myeloma analogue HMMA2.11TG/O. Antibody from secreting clones HL 1.2 (IgG,), HL 2.1 (IgG3), and HL 3.1 (IgG,) have been characterized. All three react with HL-60 (promyelocytic), RWLeu4 (CML promyelocytic), and U937 (monocytic), but not with KG-1 (myeloblastic) or K562 (CML erythroid). There is no reactivity with T-cell lines, Burkitt's cell lines, pre-B-leukemia cell lines, or an undifferentiated CML cell line, BV173. Leukemic cells from two of seven patients with acute myelogenous leukemia and one of five with acute lymphocytic leukemia react with all three antibodies. Normal lymphocytes, monocytes, polymorphonuclear cells, red blood cells, bone marrow cells, and platelets do not react. Samples from patients with other diverse hematopoietic malignancies showed no reactivity. Immunoprecipitations sug gest that the reactive antigen(s) is a lactoperoxidase iodinatable series of cell surface proteins with molecular weights of 42,000-54,000 and a noniodinatable protein with a molecular weight of 82,000. Based on these data these human monoclonal antibodies appear to react with myelomonocytic leukemic cells and may detect a leukemia-specific antigen or a highly restricted differentiation antigen.

INTRODUCTION Autoantibody responses to autologous tumors have been noted in a variety of human malignancies. Of particular interest, serological studies of patients with diverse acute and chronic leukemias have demonstrated the presence of human antibodies on fresh leukemic cells. Reactivity of these antibodies was frequently found to extend to allogeneic leukemias but not normal cells (1-4). These studies suggest that tumor-specific antigens are present on the cell surfaces of human leukemic blasts and that the patient can recognize these antigens as foreign or abnormal. Such tumor-specific or tumor-related an tigens may represent virally encoded antigens such as those seen with human T-cell leukemia virus I-associated lymphomas and leukemias, altered growth factor receptors, oncofetal anti gens, or normal cellular antigens presented in an abnormal context. The role of these antibodies and the antigens with which they react may be important in understanding the pathogenesis of these diseases and in developing therapeutic strategies. How ever, with the exception of human T-cell leukemia virus I antigens and antibodies, evaluation of these responses has not proceeded (5, 6). In part this may be due to the polyclonal nature of the human humoral response and the difficulty in Received 6/7/88; revised 11/29/88; accepted 12/29/88. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 'This work was supported by a grant (RO1-CA38687) from the National Cancer Institute. 2To whom requests for reprints should be addressed, at Division of Oncology/ Hematology, Roger Williams General Hospital, Brown University Medical School, 825 Chalkstone Avenue, Providence, Rl 02908.

performing reproducible studies with these polyclonal sera and heterogenous leukemias. It is apparent that to study these antibodies it would be important to obtain them in large quan tities and pure form. Human monoclonal antibody technology can provide these reagents but until recently has not been available (7). In the present study we describe the production and initial characterization of three human monoclonal antibodies ob tained from a single patient with CML.3 These antibodies react with leukemic cells of myelomonocytic origin, recognizing antigenic determinants on the cell surfaces of several human myelomonocytic cell lines and leukemic blasts from patients with certain acute leukemias, but not with normal hematopoietic cells of diverse origin. MATERIALS

AND METHODS

Cell Culture. Cell lines and established hybridomas were grown in aminimum essential medium lacking nucleosides, supplemented with: 1 HIMsodium pyruvate, 2 HIML-glutamine, 1% (v/v) nonessential amino acids, 10% (v/v) fetal bovine serum (high cloning efficiency and growth promotion; GIBCO, Grand Island, NY), 0.22% (w/v) sodium bicarbon ate, and 50 ^g/ml gentamicin. All other cell culture was performed with the same media containing 20% fetal bovine serum. Other addi tives were included as indicated. Cultures growing in flasks were flushed, sealed with a 5% COz/air (v/v) mixture and maintained at 37Â°C. Repeated gassings with the CU2 were performed as needed to maintain proper pH. Cultures in microtiter plates or multiwell plates were incubated in a 5% CO2 atmosphere at 37Â°Cin a humidified incubator. Cell Lines. The HMMA2.11TG/O cell line, a nonsecreting human mouse myeloma analogue, developed in this laboratory, was used for fusions (HB 9583, ATCC, Rockville, MD) (7). The B95-8 marmoset cell line was used as a source of EBV for cell transformation. The OKT8 hybridoma was obtained from the ATCC (CRL 8014, ATCC, Rockville, MD). The origins of the leukemic cell lines used in these studies are described in several reviews (8-10). These cell lines were kindly supplied by a number of investigators.4 Volunteer and Patient Cells. After informed consent, PBM and bone marrow mononuclear cells were obtained by venipuncture or bone marrow aspiration in preservative-free heparin and separated from contaminating cells by density gradient separation as previously de scribed (7). If not used immediately, cells were stored by cryopreservation in liquid nitrogen after resuspension in media containing 10% dimethyl sulfoxide. Cell Fusion. Fusions were performed according to our standard procedure (7). In brief, the human-mouse myeloma analogue HMMA 2.11TG/O was fused in a minimum ratio of 2:1 with the other parental cells. Fused cells were distributed in 96-well microtiter plates 3The abbreviations used are: CML, chronic myelogenous leukemia; EBV, Epstein-Barr virus; HAT, hypoxanthine, aminoptermin, thymidine; HT, hypoxanthine, thymidine; HuMoab, human monoclonal antibody; SDS-PAGE, so dium dodecyl sulfate-polyacrylamide gel electrophoresis; ALL, acute lymphocytic leukemia; PBM, peripheral blood mononuclear cells; PSG, Puck's saline G; I'M. WO, Puck's saline G without calcium and magnesium; PBS, phosphate buffered saline; PSG2.5, Puck's saline G with 2.5% fetal bovine serum; RIPA, radio immunoprecipitation assay. 4 Cell lines were kindly supplied by Dr. H. Lazarus (Centecor Corporation, Malvern, PA), Dr. R. Todd (University of Michigan, Ann Arbor, MI), Dr. J. Ritz (Dana Farber Cancer Institute, Boston, MA), Dr. J. Pesando (Biomembrane Institute, Seattle, WA), and Dr. M. Weimann (Roger Williams General Hospital, Providence, RI).

1665

Downloaded from cancerres.aacrjournals.org on October 3, 2017. © 1989 American Association for Cancer Research.

HUMAN MONOCLONAL

ANTIBODIES AND LEUKEMIA

in HAT-containing media. The number of cells seeded into each well in any experiment was based on the maximum number of potential hybrids given a hypothetical 1/1 fusion efficiency. The final volume in each well was 200 /il. In all experiments 10 MMouabain in 50 n\ of media with HT was added to yield a 2 /i\i ouabain concentration in each well 24 h after fusion to prevent the growth of normal or trans formed cells. A concentration of 2 ?M ouabain was maintained in the wells for 1 week, after which routine feeding was performed. Fusions were fed at 4-7-day intervals by removal of 100-150 ^1 of media and replacement with an equal volume of media containing HT. After selection for expansion, cells were transferred to 24-well multiwell plates in media containing HT and were maintained in HT until they were passaged once in flasks. Hybridomas were cloned by resuspension in media containing HT and distributed in microtiter plates such that an average of 1 cell/100 Â¿Â¿I/well was obtained. Cloned cells were fed at weekly intervals with media lacking HT. Complement Depletion of Selected T-Cells and EBV Transformation. PBM were selectively depleted of CDS-positive T-cells by antibodymediated complement lysis. Cells were suspended at a density of 10 x 106/ml in media and 100 ^1 of culture supernatant from an OKT8 culture were added. The cells were incubated for 30 min at room temperature, rabbit complement (Pel Freeze, Brown Deer, WI) was added at a final concentration of 1:5, and the cells incubated for l h at 37"C. The cells were then washed three times with Puck's Saline G without calcium or magnesium. For EBV transformation, PBM were placed in 24-well plates such that there were 0.2-0.3 x 10* cells per well, in 1.0 ml of media containing a 1:10 dilution of stock B95-8 supernatant according to the prelysis cell number. EBV-transformed cultures were maintained by feeding weekly with the addition of small amounts of media until cell growth was observed and then every 3-4 days. Growing cultures were treated as above. Indirect Immunofluorescence of Cell Surface Antigens. Indirect immunofluorescence for the detection of cell surface antigens was per formed as previously described with some modifications (11). In brief, 1.0 x IO6 live cells per sample were washed with PBS. To block Fc binding, cells were incubated with 100 /.Aof nonspecific mouse ascites for 15 min at 4"( ' and then washed once. 100 //I of culture fluid containing monoclonal antibody were added to the cell pellet and incubated at 4Â°Cfor 30 min. The cells were then washed twice with PBS, followed by the addition of 100 p\ of fluorescein-conjugated F(ab')2 goat anti-human IgG or polyspecific immunoglobulins (Tago,

onine by culture in methionine-free media supplemented with 20% cell culture medium described above and 0.3 mCi of [35S]methionine/5 x IO5cells for 14 h. The cells were centrifuged and lysed with RIPA and the lysates immunoprecipitated with sepharose conjugated to goat antihuman IgG (Caltag, San Francisco, CA) that had been preincubated with HuMoabs. After unbound lysate protein had been removed by five washes in RIPA buffer, the bound antigen was eluted by boiling in Laemmli buffer and subjected to SDS-PAGE under reducing condi tions. Unreduced eluates were diluted in RIPA buffer to reduce the concentration of SDS and then reprecipitated with the identical HuMoabs. These secondary immunoprecipitates were washed, eluted, reduced, and analyzed by SDS-PAGE as above. Determination of Antibody Isotype and Secretion. Supernatants from test wells or bulk cultures were tested for IgG isotype using a modifi cation of a microelisa assay as described previously (7). In brief, test wells (Immulon 2; Dynatech, Alexandria, VA) were coated with 50 M' of mouse monoclonal antibody to isotypes IgGi, IgGz, IgGs, or IgG4 (Caltag Laboratories) and incubated for at least 2 h. Plates were then blocked with PSG with 2.5% fetal bovine serum (v/v) (PSG 2.5%) for a minimum of 2 h, washed twice with PBS with 0.5% tween 20 (v/v) (PBS-tween) and twice with PBS. Then 100 Â¿il of test supernatant were added. The wells were incubated for 2 h, washed as above, and 75 p\ of peroxidase-conjugated goat anti-human immunoglobulins (IgM, IgG, and IgA) or specific peroxidase-conjugated goat anti-human IgG, IgM, or IgA (Tago, Inc., Burlingame, CA) diluted 1:3000 in PSG2.5% were added and incubated for 1 h. The wells were washed three times with PBS-tween and three times with PBS after which 100 n\ of O-phenylenediamine in citrate buffer were added. Plates were read at 5-45 min by observing a color change and scoring from negative to 4~. Quantitation of secretion was performed by a modification of the assay described above. Wells were coated with 75 n\ of a 30-^g/ml solution of goat anti-human immunoglobulins. Supernatants from rou tinely growing cultures were collected every 3-4 days and pooled. Dilutions in PSG2.5% were tested as above using peroxidase-conju gated goat anti-human IgG antibodies. A standard curve was generated by serial dilution of a purified preparation of Â¡sotypespecific human monoclonal antibody of known antibody content, and the antibody content of the experimental Supernatants determined simultaneously. Optical density was quantitatively determined on a microelisa reader and used to quantify the reaction.

RESULTS

Inc., Burlingame, CA) diluted in PSG 2.5%. The sample was incubated for 30 min at 4Â°Cand then washed once in PBS. Following this, the

Production of Human Monoclonal Antibodies. Peripheral

pellet of live cells was resuspended in 250 n\ PBS and then 250 M' of blood mononuclear cells were obtained from a patient with 1% (v/v) formaldehyde in PBS was added to fix the sample. The fixed CML in hematological remission following one cycle of inten cells were resuspended and stored up to 5 days until analysis at 4Â°C. sive chemotherapy. The PBM were depleted of CDS-positive Analysis was performed on an Epics C cell sorter (Coulter, Hialeah, FL). Cells fixed with formaldehyde after labeling with antibodies have been shown to retain, without significant alteration, prefixation findings upon flow cytofluorimetric analysis. Fixation permits storage of the sample for delayed evaluation and neutralizes human pathogens that might be present in the sample. Our own experiments and those of others have shown no effect of fixation on analysis under these condi tions (12, 13). Negative controls for these experiments consisted of normal human sera diluted 1:1000 (approximately 10-20 jig/ml of IgG) and the monoclonal IgG immunoglobulins and antibodies 7C7 (IgGj immunoglobulin) and I I 11)1 2 (IgGi antitetanus human monoclonal antibody) (7). All negative controls gave the same results in multiple assays. Radioimmunoprecipitation of Antigens. U937 cells were surface la beled by the lactoperoxidase method with '"I (1.0 mCi/107 cells) and lysed with RIPA buffer (14). After centrifugation to remove nuclei and debris, the lysates were incubated with 25 n\ of protein A-Sepharose4B CL (Pharmacia, Uppsala, Sweden) that had been preincubated with 100 //I of goat anti-human immunoglobulins (Tago, Burlingame, CA), and 500 /il of HuMoab supernatant. After washing away unbound proteins, bound antigens were eluted by boiling in Laemmli buffer and subjected to SDS-PAGE under reducing conditions. RWLeu4 and U937 cells were metabolically labeled with [35S]methi-

T-cells and then distributed into 20 wells of a 24-well culture plate at 250,000 cells/well with EBV. Within 3 weeks all 20 of the cultures showed evidence of transformation. Supernatants from the EBV-transformed cultures were initially screened for reactivity with the human leukemic cell lines RWLeu4, HL60, and K562 by cell surface indirect immunofluorescence. Nega tive controls included monoclonal IgM and IgG immunoglob ulins, sera, and antitetanus human monoclonal antibodies of IgG and IgM classes. Positive controls were not available. Four presumed polyclonal EBV-transformed B-cell cultures reacted with the acute promyelocytic cell line HL60 and the myeloid CML cell line RWLeu4, but were negative with the CML erythroleukemia cell line K562. The remaining 16 Supernatants were negative with all three cell lines. All four of the polyclonal EBV-transformed B-cell lines se creting reactive antibodies were expanded and three were fused with the human-mouse myeloma analogue HMMA2.11TG/O. The results of these fusions are shown in Table 1. The fusion efficiency was high and, in two fusions, the majority of the hybridomas secreted antibody reactive with the two myeloid leukemic cell lines. These results are in contrast to our previous

1666


HUMAN MONOCLONAL ANTIBODIES AND LEUKEMIA

Table 1 Results of the fusion ofHMMA2.11TG/O with antileukemic antibody secreting polyclonal B-cell lines Fusion

Density"

Wells

Growth

AB

AB

A

Cloned

F37F38F4012,00010,0005,2003843841921009921,92*3211001.652212

" Fused EBV transformants/well. * 192 wells each from fusions F37 and F38 were tested by indirect immunofluorescence. Table 2 Reactivity of human monoclonal antibodies with cell lines of hematopoietic origin Antibody lineHL60RWLeu4BV173K562KG1U937CEMMOLT4HPB-ALLRAMOSJD39NALM1NALM6HL1.13V2+0004+0000000HL2.13+2+0004+0000000HL3.13+2+0004+0000000 Cell

B

" Cell surface reactivity by indirect Â¡mmunofluorescence above background. 0 â€¢¿â€¢ 0-10%, 1+ = 11-25%, 2+ = 26-50%, 3+ = 51-75%, 4+ = 76-100%.

studies with tetanus toxoid in which fusion efficiency with EBV transformants is slightly lower and the percentage of antibodysecreting wells substantially lower. Antibody-secreting hybridomas from each fusion were randomly selected, cloned and recloned. Recloned hybridomas were then used to generate culture supernatants containing human monoclonal antibodies which were used for subsequent studies. Antibody secretion by recloned hybridomas has been stable for periods of up to 6 months without recloning. Cultures were terminated at that time without loss of secretion for reasons of economy. The three human monoclonal antibodies have been designated HL1.1, HL2.1, and HL3.1. Isotype analysis indicates that they are of the IgGi, IgG3, and IgGi isotypes, respectively. Routinely growing cultures were found to yield 11, 4, and 3.5 Â¿Â¿g/ml of monoclonal antibody respectively in pooled supernatants. Reactivity with Human Hematopoietic Cells. The three human monoclonal antibodies demonstrated identical patterns of cell surface reactivity with established human hematopoietic cell lines of diverse origin. As shown in Table 2, they react with a promyelocytic cell line (HL60), a promyelocytic CML cell line (RWLeu4), and a monoblastoid cell line (U937), but not with other myeloid or undifferentiated cell lines (KG1, BV173), an erythroleukemia CML cell line (K562), T-cell lines (CEM, MOLT4, HPB-ALL), Burkitt's lymphoma cell lines (JD39, RAMOS), or pre-B-leukemia cell lines (NALM1, NALM6). Thus, the reactivity of these antibodies appears to be restricted to cell lines of myelomonocytic origin. Examples of the cytofluorographic reactivity of each of the antibodies with one of the cell lines, HL60, RWLeu4, or U937, are shown in Fig. 1. The antibodies were further tested for reactivity with normal hematopoietic cells and leukemic cells from patients with di verse forms of chronic and acute leukemia. As shown in Table 3, no reactivity was observed with normal peripheral blood cells or bone marrow cells. Leukemic cells from two of seven patients with acute myelogenous or myelomonocytic leukemia and one patient with ALL were reactive with all three antibodies while leukemic cells from patients with other forms of leukemia were unreactive. Interestingly, the antibodies did not react with the

C

Fig. 1. Representative cytofluorographic determinations of indirect immunofluorescent cell surface reactivity of the human monoclonal antibodies on the cell lines U937,1II 60, and RWLeu4. One antibody is shown with each cell line since all react in the same fashion. Each figure has a negative control antibody to the left of the arrow, overlapping (darkly shaded area) of the negative and positive antibody, and positive antibody to the right of the arrow. The curves are expressed as log green fluorescence (X-axis) by cell number (Y-axis). /4, U937 and HL1.1 (85% positive); B, HL60 and HL2.1 (50% positive); and C, RWLeu4 and HL3.1 (28% positive). The negative control antibody in A-C is F11DE2, an IgGi antitetanus human monoclonal antibody, and is adjusted to 7% positivity in the assays.

chronic phase PBM of the patient from whom the hybridomas were derived, perhaps because of modulation of antigenic de terminants by preexisting serum antibody or the later stage of differentiation of the circulating cells. Blast phase cells were not available from this patient. Analysis of the ALL patient's cells with monoclonal antibodies demonstrated reactivity with a typical pre-B-cell leukemia pattern of CD 10, CD 19, and CD20 positivity. In addition, however, a fraction of the cells expressed MY7, a myeloid marker, indicating some lineage infidelity (15). Radioimmunoprecipitation of Antigen. In order to assess the cell surface expression of these proteins, U937 cells were surface labeled with I25I, immunoprecipitated, and the immunoprecipitates resolved by SDS-PAGE. In addition to the known M, 72,000 Fc receptor, precipitated by the negative control IgG human monoclonal antitetanus antibody (16), the human anti-

1667


HUMAN MONOCLONAL


Table 3 Reactivity of human monoclonal antibodies with patient samples Number positive/number tested (range of reactivity).

B

D

SourceNormalPBL"PMNMonocytesRBC*PlateletsBone

marrowAcute leukemiaAMLAPLALLAULT-Cell (1-2+)0/31/5(1+)0/40/20/10/30/20/2HL3.10/30/30/30/30/30/22/7 (1+)0/31/5(1+)0/40/20/10/30/20/2AntibodyHL2.10/30/30/30/30/30/22/6 (1+)0/31/5(1+)0/40/20/10/30/20/2

leukemiaCLLSezaryChronic

leukemiaCML phase)CLLCML (chronic (blast phase)HL1.10/30/30/30/30/30/22/7 " Populations judged 95-99% pure by murine monoclonal antibodies. * The abbreviations used are: RBC, red blood cells; AML, acute myelogenous leukemia; APL, acute promyelocytic leukemia; ALL, acute lymphocytic leukemia; AUL, acute undifferentiated leukemia; CLL, chronic lymphocytic leukemia.

leukemia antibodies precipitate a series of proteins at approxi mately A/r 42,000-54,000. A protein at approximately A/r 54,000 appeared to be well defined in HL2.1 and was not present in the control lane. Diffuse bands at A/r 42,000-54,000 were precipitated by HL 1.1 and 3.1 in association with a great deal of heavy chain protein (Fig. 2). In addition, light chains from all three human monoclonal antibodies migrated different distances on the SDS-gels in this experiment and the experi ment described below (data not shown) confirming that anti bodies HL1.1 and HL3.1 were uniquely different from one another although both are Igd isotype and have the same pattern of reactivity. U937 cells, which expressed the largest cell surface amount of antigen, were metabolically labeled with [35S]methionine. Subsequent immunoprecipitation and reimmunoprecipitation of these metabolically labeled U937 cells demonstrated A/r 82,000 and 54,000 proteins specifically immunoprecipitated by all three antibodies and a Afr 63,000 protein precipitated by HL1.1 and HL3.1. Because of the low expression of these antigens, background was high despite reimmunoprecipitation (data now shown) and other approaches will be necessary to define these antigens. DISCUSSION In the present study we describe three human monoclonal antibodies that react with cell surface antigens on human myelomonocytic leukemia cell lines and blasts but not with hematopoietic cells from normal individuals or other leukemic pa tients and not with cell lines of other diverse hematopoietic origins. This highly restricted pattern of reactivity suggests that these antibodies react with differentiation or tumor-specific antigens present on myelomonocytic leukemia cells. All three antibodies, although derived from separate EBV transformants and having different heavy chain isotypes or light chains, seem to react with the same or highly associated antigens as shown by the pattern of cell surface reactivity and immunoprecipita tion. Auto-antileukemia antibody production has been observed

Fig. 2. Radioimmunoprecipitation of 125Icell surface-labeled U937 cells. Lanes A-D, immunoprecipitations with HL1.1, HL2.1, HL3.1, and 7C7 (an IgG3negative control) monoclonal antibodies, respectively. The thick arrow indicates a putative M, 72,000 FcRl receptor protein and the thin arrow the M, 54,000 protein specifically precipitated by HL2.1 track B.

and evaluated in a number of serological systems. Metzgar et al. and Garrett et al. described autoantibody production in leukemics and showed serological reactivity with both autologous and allogeneic leukemic blasts (2,3). Mitchell et al. showed that cytophilic antibodies were present in the sera of leukemics and that specificity was easily demonstrable using these sera (1). Gutterman et al. demonstrated blastogenic responses to autologous leukemic blasts in association with membranebound immunoglobulin (4). Since these serological studies rep resent evaluations of polyclonal responses, the antigenic targets may be heterogenous. Targets of these polyclonal sera could include viral antigens, altered growth factor receptors, oncofetal antigens, or abnormally expressed normal antigens. Alterna tively, these antibodies could represent a phenomenon similar to systemic lupus erythematosis in which self antigens become targeted by the human humoral response possibly as a result of

1668


HUMAN MONOCLONAL


immune dysregulation induced by the disease or the therapy (17). Study of these antigens and antibodies might be important in understanding the pathogenesis of these diseases, the role of the host immune response in disease progression, and in the development of potentially therapeutic reagents. The develop ment of murine monoclonal antibody technology shifted the emphasis away from studies of these autochthonous systems. Unfortunately, murine monoclonal antibody technology does not address the repertoire of the human auto-antitumor re sponse. This is inherent in the murine system because the murine host preferentially reacts with differentiation and hu man specific antigens (7). Thus human monoclonal antibodies will be useful in capturing these antibodies so that we may study the repertoire of these responses and the impact of these antibodies and antigens on the progression of these diseases. Solid tumor-reactive human monoclonal antibodies from pa tients with cancer have been reported by several investigators (18-23). Frequently, these human monoclonal antibodies, par ticularly IgM antibodies, have been found to react with a broad range of normal and abnormal cells. Occasionally, a normal intracellular component, such as cytokeratin, which is not nor mally antigenic, is found to be the target of antibody reactivity and confers an element of tumor tissue specificity (18, 22). Although IgG human monoclonal antibodies have been rare, the antisolid tumor human monoclonal IgG antibodies have had similar patterns of reactivity (20, 23). Our own studies have shown that IgM human monoclonal antibodies with diffuse cell surface reactivity to neoplastic and normal hematopoietic cells are relatively easy to obtain from both diseased and healthy individuals5'6 while tumor-specific antibodies may be both dif ficult to obtain and difficult to prove as being tumor specific. Relatively specific IgG human monoclonal antibodies reac tive with cell surface antigens on human leukemia cell lines and leukemic blasts have been reported by Andreasen and Olsson. They were able to prepare two separate IgG HuMoabs with similar reactivities from a patient with acute myelogenous leu kemia and a patient with CML (24). The pattern of reactivity may have been a function of their screening method which eliminated reactivity with a variety of other hematopoietic cell lines. The antibody described by these coworkers reacted with an intracellular antigen in normal bone marrow cells and an abnormal A/r 18,000 antigen expressed on the cell surfaces of leukemic cells. The antibodies described in the present study have a signifi cantly different pattern of reactivity with cell lines and less broad reactivity with AML patients than those described by Andreasen and Olsson. In addition, all three antibodies immunoprecipitate proteins of M, 82,000 and 54,000 from metabolicaly labeled leukemic cells and a series of iodinatable cell surface-associated proteins in the range of M, 42,000-55,000. The data from these experiments are consistent with the M, 80,000 protein being an intracellular antigen while the M, 42,000-54,000 proteins are surface antigens. Alternatively, these antigens may be unrelated and share epitopes or may be disulfide-linked chains of a heterooligomeric protein. The dif ferences also suggest that HL1.1 and HL3.1 are detecting identical antigens, while HL2.1 has a different but highly re lated specificity. 5 M. R. Posner, H. J. Barrach, H. S. Elboim, K. Nievens, D. J. Santos, C. O. Chichester, and E. V. Lally. The generation of hybridomas secreting human monoclonal antibodies reactive with type II collagen. Hybridoma, in press. 6 M. R. Posner, H. S. Elboim, and M. Tumbor. Human monoclonal antibodies from a patient with acute lymphocytic leukemia that react with cell surface antigens on normal and leukemic cell populations. Manuscript in preparation.

Unlike the system described by Andreasen and Olsson, in the presently described system for HuMoab production, preselec tion of EBV transformants on the basis of specificity, for fusion, is possible. Fusions were performed with three of the four antibody-producing transformants, but all four had the same reactivity pattern in the initial screening against HL60, K562, and RWLeu4. If reactivity patterns had varied, preselection for fusion would have been performed. This is an important poten tial advantage of the EBV transformation/fusion method with the high fusion efficiency cell line HMMA2.11TG/O. Of inter est, fusion efficiency was 25-50% higher than that seen in our tetanus toxoid or collagen experimental systems,5 and equiva lent to that seen in other antileukemia antibody experiments.6 Furthermore, compared to the tetanus or collagen systems, the frequency of hybridomas producing the desired antibody is 1020 times greater in the antileukemia experiments. Taken to gether, these data support the notion that a large fraction of circulating B-cells in this leukemic patient were able to produce antibodies to these antigens and indicate that these antigens are highly immunogenic. One may speculate an important biolog ical role for either the antigens or the humoral response. In the present study, we have demonstrated that auto-antileukemia human monoclonal antibodies were readily obtainable from a patient with CML. These antibodies have highly re stricted patterns of reactivity with established human leukemia cell lines and leukemic blasts, and lack significant reactivity with normal hematopoietic cells, suggesting that they react with tumor-specific antigen(s) on myelomonocytic leukemia cells, or highly restricted differentiation antigen(s). The biological im portance and potential therapeutic use of these antibodies and antigens will be the subject of future studies. REFERENCES 1. Mitchell, M. S., Mokyr, M. B., Aspnes, G. T., and Mclntosh, S. Cytophilic antibodies in man. Ann. Intern. Med., 79: 333-339, 1973. 2. Metzgar, R. S., Mohanakumar, T., and Miller, D. S. Membrane-bound immunoglobulins on human leukemic cells. J. Clin. Invest., 56: 331-338, 1975. 3. CarreÂ«, T. J., Takahashi, T., Clarkson, B. D., and Old, L. J. Detection of antibody to autologous human leukemia cells by immune adherence assays. Proc. Nati. Acad. Sci. USA, 74:4587-4590, 1977. 4. ( Â¡interinan.J. U., Rossen, R. D., Butler, W. T., McCredie, K. B., Bodey, G. P., Freireich, E. Jâ€žand Hersh, E. M. Immunoglobulin on tumor cells and tumor-induced lymphocyte blastogenesis in human acute leukemia. N. Engl. J. Med., 288: 169-175, 1973. 5. Posner, L. E., Rogert-Guroff, M., Kalyanaraman, V. S., Poiesz, B. J., Ruscelli, F. W., Fossieck, B., Bunn, Jr., P. A., Minna, J. D., and Gallo, R. C. Nalural antibodies to Ihe human T cell lymphoma virus in patienls wilh culaneous T cell lymphomas. J. Exp. Med., 154: 333-346, 1981. 6. Mann, D. L., DeSantis, P., Mark, G., Pfeifer, A., Newman, M., Gibbs, N., Popovic, M., Sarngadharan, M. G., Gallo, R. C., Clark, J., and Blallner, W. HTLV-1-Associated B-cell CLL: Indirecl role for relrovirus in leukemogenesis. Science (Wash. DC), 236: 1103-1106, 1987. 7. Posner, M. R., Elboim, H., and Sanios, C. The conslruction and use of a human-mouse myeloma analogue suitable for ihe routine production of hybridomas secreting human monoclonal anlibodies. Hybridoma, 6: 611625, 1987. 8. Minowada, J., Koshiba, H., Sagawa, K., et al. Marker profiles of human leukemia and lymphoma cell lines. J. Cancer Res. Clin. Oncol., 101:91-100, 1981. 9. Wiemann, M. C., Hollmann, A., Posner, M., Arlin, Z., Friedland, M., and Calabresi, P. Eslablishment of a new Philadelphia chromosome posilive human chronic myeloid leukemia cell line. Clin. Res., 33:461 A, 1985. 10. Posner, M. R., Antoniou, D., Griffin, J., Schlossman, S. F., and Lazarus, H. An enzyme-linked immunosorbent assay (elisa) for ihe detection of mono clonal antibodies to cell surface antigens on viable cells. J. Immunol. Meth ods, 4Â«:23-31, 1982. 11. Posner, M. R., Schlossman, S. F., and Lazarus, H. Novel approach to the construclion of human "myeloma analogues" for ihe produclion of human monoclonal antibodies. Hybridoma, 2: 369-381, 1983. 12. Posner, M. R., Reinherz, E. L., Breard, J., Nadler, L. M., Rosenlhal, D. S., and Schlossman, S. F. Lymphoma subpopulalions of peripheral blood and spleen in untreated Hodgkin's disease. Cancer (Phila.), 48:1170-1176,1981. 13. Lifson, J. D., Sasaki, D. T., and Engelman, E. G. Utilily of formaldehyde

1669


HUMAN MONOCLONAL

14. 15. 16. n 17.

18.

19.


fixation for flow cytometry and inactivation of the AIDS associated retrovirus. J. Immunol. Methods, 86: 143-149, 1986. Gielkins, A. L. J., Van Zaane, D., Bloemers, H. P. J., and Bloemendal, H. Synthesis of Rauscher murine leukemia virus-specific polypeptides in vitro. Proc. Nati Acad. Sci USA, 73' 356-360 1976 Foon, K. A., and Todd, III, R. F. Immunologie classification of leukemia and lymphoma. Blood, 68:1-31, 1986. Liesveld, J. L., Abboud, C. N., Looney, R. J., Ryan, D. H., and Brennan, J. K. Expression of IgG Fc receptors in myeloid leukemic cel, lines. J. ImmuM. K ju? r M T'n .-r , t.t. , , Mmota, S., and Wmfield, J. B. Identification of three major target molecules of IgM antilymphocyte autoantibodies m systemic lupus erythematosus. J. Immunol., 139: 3644-3651, 1987. Skaletsky, E., Oh, E., Rulot, C., Baird, S. M., Burnett, K. G., Masuho, Y., Astanta, R. W., Haghighi, P., Collins, H., Starr, R., Carlo, D. J., Royston, I., and Soboi, R. E. A human monoclonal antibody to cytokeratin intermediate filament antigens derived from a tumor draining lymph node. Hybridoma, 7: 367-376, 1988. Haspel, M. V., McCabe, R. P., Pomato, N., Janesch, N. J., Knowlton, J. V.,

20. 21" 22.

23.

24.

Peters, L. C, Hoover, H. C, and Hanna, M. G. Generation of tumor cellreactive human monoclonal antibodies using peripheral blood lymphocytes from actively immunized colorectal carcinoma patients. Cancer Res., 45: 3951-3961, 1985. Kan-Mitchell, J., Imam, A., Kempf, R. A., Taylor, C. R., and Mitchell, M. s- Human monoclonal antibodies directed against melanoma tumor associated antigens. Cancer Res., 46: 2490-2496, 1986. Smith, L. "â€¢'Y'". A." Â»"ber, M., and Teng, N. N. H. Generation of human ---â€¢.Â£*--* y-ggg^ Ho, M-K, Geist, C, Murray, J., Kalo, K., DeLellis, R., Wolfe, H., Carney, w Rabm â€ž¿and Dur^ >p Distribution ^ immunochemical characterÂ¡zationof a keratjn.iike antigen Â¡nepithelial tumors using mouse and human monoclonal antibodies. Cancer Res., 48:4969-4975, 1988. Kjeldsen, T. B., Rasmussen, B. R., Rose, C., and Zeuthen, J. Human-human hybridomas and human monoclonal antibodies obtained by fusion of lymph â€ž¿ode lymphocytes from breast cancer patients. Cancer Res., 48: 3208-3214, 1988. Andreasen, R. B., and Olsson, L. Antibody-producing human-human hybridomas. J. Immunol. Methods, 137:1083-1090, 1986.

1670


Human Monoclonal Antibodies Reactive with Human Myelomonocytic Leukemia Cells Marshall R. Posner, David J. Santos, Hillary S. Elboim, et al. Cancer Res 1989;49:1665-1670.

Updated version

E-mail alerts Reprints and Subscriptions Permissions

Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/49/7/1665

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].
