Immunobiol., vol. 183, pp. 374-385 (1991) 1 Department
of Pediatrics, Friedrich Schiller University, 2 Physico-Technical Institute, Jena, Germany, and 3 Laboratoire Pierre Weiss, Universite Louis Pasteur, Strasbourg, France
Model Experiments for Immunomagnetic Elimination of Leukemic Cells from Human Bone Marrow. Presentation of a Novel Magnetic Separation System BERND GRUHN1, RALF HAFER 1, ANGELIKA MULLER \ WILFRIED ANDRA2 , HENRI DANAN3 , and FELIX ZINTL 1 Received February 11,1991 . Accepted in Revised Form June 19, 1991
Abstract Optimal conditions for removing leukemic cells from human bone marrow with monoclonal antibodies (mAb) and magnetic immunobeads were investigated. Monodisperse 3 [!m polystyrene micro spheres containing magnetite were coated with affinity-purified rabbit antimouse IgG at 4°C, pH 9.6 for ISh. SKW-3 cells (T-CLL cell line) were marked with the supravital DNA stain Hoechst 33342, seeded into normal human bone marrow, and then incubated with the mAb CD1, CD6, and CDS at 4°C for 30 min. In preliminary experiments REH cells (cALL cells) and mouse anti-REH cell antibodies were used to find the most favorable conditions for the binding of magnetic beads to tumor cells. Optimal formation of cell-bead rosettes was achieved by rotating beads and tumor cells together at room temperature at a concentration of 1 x 10 7 cells/ml, a bead: tumor cell ratio of 100:1 and an incubation time of one hour. The novel magnetic separation apparatus consists of three polystyrene chambers connected by silicone rubber tubing. The chambers contain four steel inserts each equipped with 32 nickel wires, which are magnetized by permanent magnets in such a way that the inhomogeneous high gradient magnetic field could be established within the cell suspension containing the cells to be depleted. The fluid flow was established by a peristaltic pump. At a flow rate of 1.5 mlimin and a field strength of 160 kA/m, no beads could be detected in the purged marrow. A cocktail of the three mAb was more effective than any single antibody in forming bead-cell rosettes. Two sequential purging cycles were superior to one. The marrow recovered was highly viable as assessed by trypan blue dye exclusion and by growth of CFUGM.
Introduction Most tumors show a dose-related response to anti-neoplastic drugs and radiation. This means that higher doses result in increased tumor cell death (1). However, toxicity for normal tissues, especially for bone marrow, is dose-limiting (2). This problem can be overcome by autologous bone marrow transplantation. However, there is the risk that in tumors which have their origin in the bone marrow or which metastasize to the bone Abbreviations: mAb = monoclonal antibody; CLL = chronic lymphoblastic leukemia; cALL = common acute lymphoblastic leukemia; CD = cluster of differentiation; CPU-GM = colony-forming units - granulocyte - macrophage
Immunomagnetic Elimination of Cells from Bone Marrow . 375
marrow occult tumor cells may occur in the harvested bone marrow. These cells are capable of causing a relapse after reinfusion (1, 3). To avoid this effect, methods were sought for to remove tumor cells in vitro from bone marrow. Physical (4, 5), pharmacological (6-8) and immunological (9-11) purging procedures have already been described. One modern immunological method is the immunomagnetic elimination of tumor cells.
Materials and Methods Principle
Primary antibodies directed against tumor cell antigens are bound to the surface of tumor cells. Secondary antibodies of another species which detect primary antibodies are coated onto the surface of magnetic beads. The subsequent joint incubation allows the binding of magnetic beads to tumor cells on the basis of antigen-antibody reaction with the help of this indirect method. After incubation, cells with bound beads are retained in the magnetic system. Magnetic beads
The magnetic beads used were produced by Dr. J. UGELSTAD (Norwegian Institute of Technology, University of Trondheim, Norway). They have a diameter of 3 [Am, consist of polystyrene and contain magnetite (12, 13). 50 mg magnetic beads were suspended in 100 ml isotonic sodium chloride solution for the separation experiments and dispersed by means of an ultrasonic disintegrator (HR-30 model of Elektromat, Dresden, Germany) and an agitator. Magnetic separation system
One electromagnet, four permanent magnets and three chambers situated between the magnets were available for the magnetic separation of the beads from the suspension. The chambers were connected by silicone rubber tubing, and the fluid flow was established by a
Figure 1. Polystyrene chamber with 4 inserts and a single insert with nickel wires on the left side.
376 . B. GRUHN, R. HAFER, A. MOLLER, W. ANDRA, H. DANAN, and F. ZINTL peristaltic pump. The electromagnet used consisted of a soft iron core, two coils with a total resistance of 22 ohms and soft iron pole pieces between which a magnetic field strength of 40 kA/m ... 185 kA/m could be generated. As permanent magnets, we used rectangle magnets having a size of 50 x 30 x 24 mm} (Keramische Werke, Hermsdorf, Germany). These were hard ferrites between which a magnetic field strength of 160 kA/m could be achieved after adequate remagnetization. The three chambers consist of polystyrene, have an internal volume of 2 ml each and contain four stainless-steel inserts with the external size of 20 x 10 x 5 mm} (Fig. 1). Each insert was divided into four planes, and in each plane there were nickel wires with a diameter of 0.2 mm. One chamber insert contained 32 nickel wires in total, the number of wires per plane being 5, 7, 9, and 11. During the experiments the chambers were so installed that the number of wires per plane of an insert increases in flow direction to obtain a filter effect. The axis of the nickel wires stands perpendicular to the magnetic field, so that an optimal magnetization of the wires is achieved. The wires themselves induce a highly effective inhomogeneous magnetic field. As a result, the beads moving with the flow of the fluid get directly to the wires and are retained there. Detection of passed magnetic beads after separation
The concentration of passed beads was determined with the aid of the Neubauer counting chamber. We were able to increase the sensitivity of the bead-counting method essentially by centrifugation of the suspension at 3000 rim in for 5 min. Thus, the volume could be reduced 1000-fold, and the concentration could be increased 1000-fold. A further increase in sensitivity was obtained by counting all large squares of the counting chamber. By these means, the detection limit of the concentration of beads could be reduced to 1 bead/m!. Incubation of magnetic beads with antibodies
The micro spheres were coated with affinity-purified rabbit anti-mouse IgG (Dr. K.-H. VOGT, Institute of Medical Microbiology, Friedrich Schiller University, Jena, Germany). This antibody was dissolved in sodium carbonate buffer (0.015 M Na2C03, 0.D35 M NaHC0 3), pH 9.6 at a concentration of 200!!g protein/m!. 1 ml antibody buffer solution was added to 1 mg (7.9 X 107) beads. The incubation of beads with antibodies was performed for 18 hat 4°C by means of an agitator. Then the suspension was washed twice with isotonic sodium chloride solution and once again with RPMI 1640 medium (Institut fUr Immunpriiparate und Niihrmedien, Berlin, Germany) to remove unbound antibodies. Cell material
Two different permanent cell lines were available for the experiments: REH cells (cALL cells) and SKW-3 cells (T-CLL cells). In addition, a mononuclear bone-marrow cell fraction, harvested according to a modified version of BOYUM'S method (14) was used. Normal human bone marrow was obtained from the iliac crest of healthy volunteers. After addition of preservative-free heparin, bone marrow was added to a mixture of Infukoll 6 % (Serumwerk, Bernburg, Germany) and Visotrast 370 (Fahlberg-List, Magdeburg, Germany) with the density of 1.09 g/cm 3 • After sedimentation for 1 h, the upper erythrocyte-poor fraction was taken and centrifuged for 10 min at 1500 r/min. The sediment was resuspended in isotonic sodium chloride solution and added to a mixture of Ficoll 400 (Pharmacia Fine Chemicals AB, Uppsala, Sweden) and Visotrast 370 of the density of 1.13 g/cm 3. After density centrifugation for 20 min at 1500 r/ min, the bulk of the polymorphonuclear cells was separated. The mononuclear cell fraction was taken from the borderline and washed three times with RPMI 1640 medium. Incubation of cells with antibodies
To bind REH cells with antibodies, mouse anti-REH cell serum (non-absorbed crude serum) was used. Three mAb VIT-6 (CDl), VIT-12 (CD6) and VIT-8 (CD8) (kindly provided
Immunomagnetic Elimination of Cells fro m Bo ne Marrow . 377 by Prof. Dr. W. KNAPP, Institute fo r Immunology, Universi ty of Vi en na, Au stria) were available for the binding to the SKW -3 cells. The incubatio n of cells with antibodies directed against the cells was performed for 30 min at 4 °C a t aconcentration of immunoglobulin having at least 30 Jlg/ I x107 cells/ m!. The unbound antibodies were removed by washing twice with RPM I 1640.
Staining of tumor cells To differentiate SKW-3 cells fro m bone marrow cells, SKW-3 cells were stained with the fluorescent supravital stain Hoechst 33342 (H 33342) (SERVA, Heidelberg, Germany) that links with the chromosomal DNA of the cells (15).
Statistics Data represent the ari thmetic mean from 4 independent experiments. The statistical analysis was carried out with the help of Wilcoxon's paired rank sum test (16).
./ ./ . //
."',,,'" ./
'"
'"
.,.-
'//'"
-- 40 k A / m
./
.'1/
12 0 k A / m
.'1/
18Sk A /m
.II
/1
./
I
~1
I
L-~~
____
1.01~ 2.5
~
__________
5.0
~
______________________
10.0
~~
20.0
Flow rote Iml / mln I
Figure 2. Concentration of beads passed (beads/ml) depending on flow rate (mUm in) and magnetic field strength (kA/ m). Each point represents the arithmetic mean of 4 experiments.
378 . B. GRUHN, R. HAFER, A. MOLLER, W. ANDRA, H. DANAN, and F.
ZINTL
Results U sing an electromagnet and one chamber we first investigated under what conditions the beads were retained in the magnetic field. It could be shown that with decreasing flow rate and increasing magnetic field strength significantly (p < 0.025) fewer beads passed the chamber (Fig. 2). By means of three chambers connected in series and situated between four permanent magnets a further improvement of the separation efficiency was achieved. A flow rate of 1.5 mllmin proved to be optimal - no beads could be detected after passing through the last chamber. REH cells with bound mouse antiREH cell antibodies were added to the microspheres coated with rabbit anti-mouse antibodies. The incubation was performed at room temperature under gentle rotation. In each of the 4 experiments at least 200 REH cells were analyzed for bound beads. The optimal formation of rosettes was obtained at an incubation time of 1 h (Fig. 3), at a cell concentration of 1 x 107 cells/ml (Fig. 4) and a bead: tumor cell ratio of 100:1 (Fig. 5). Under these conditions a rosetting rate of 99.9 ± 0.1 % was reached. In 4 separation experiments we observed a reduction from (4.0 ± 0.4) x 105 REH cells/ml to (4.1 ± 0.3) x 102 REH cells/ml after magnetic separation. Depending on the rosette formation, a removal of 3 logs of tumor cells was achieved. For the immunomagnetic elimination of SKW-3 cells from human bone marrow, it was necessary to use mAb against these tumor cells. By means of the rosette formation we observed that the three mAb CD1, CD6, and CDS were bound to a different extent by the SKW-3 cells. A cocktail of the three mentioned mAb resulted in a significantly (p < 0,01) higher rosetting rate (Fig. 6). 100 80
r----
I i
20
0.25
0.5
to
2.0 t
Time (h)
Figure 3. Percentage of rosettes formed (%) depending on incubation time (h) (REH cell concentration 1.6 x 106 cells/ml, bead:tumor cell ratio 5: 1). Each point represents the arithmetic mean of 4 experiments, and vertical bars denote standard error of the mean.
Immunomagnetic Elimination of Cells from Bone Marrow . 379
100
80 60 ~
'" 2!
40
Qj
'" 0
a:
20
1cf> Cell concentration (cells I ml)
Figure 4. Percentage of rosettes formed (%) depending on REH cell concentration (cells/ml) (incubation time 15 minutes, bead:tumor cell ratio 5:1). Each bar represents the arithmetic mean ± SE from 4 experiments.
The formation of rosettes was significantly (p < 0.01) lower at 4°C than at 20°C or 37 °C, while between the latter temperatures no significant difference occurred (Fig. 6). In our cell separation experiments, the rosette formation was carried out at room temperature.
100
80 ~
- 60
III
QI
:s: ~
fi.
40
20
2: 1
5:1
10 :1
Bead : tumor cell ratio
Figure 5. Percentage of rosettes formed (%) depending on bead:tumor cell ratio (incubation time 15 min, REH cell concentration 4 x 105 cells/ml). Each bar represents the arithmetic mean ± SE from 4 experiments.
~ CO S
§
ESl
C01
C06
mID Coc..ktml
100
so ~ 60
20
20 Temperature (O e I
37
Immunomagnetic Elimination of Cells from Bone Marrow' 381
T""T=- - O
•
•
•
Figure 7. H 33342-stained SKW-3 cells with bound magnetic beads (a) and unstained bone marrow cells which in part bind (b) or phagocytize (c) magnetic beads. Magnification x 600.
The concentration of bone marrow cells was reduced to (5.5 ± 0.2) x 106 cells (55 %) after the first cycle and (3.0 ± 0.2) x 106 cells (30 %) after the second cycle of magnetic separation. However, viability of the recovered bone marrow always exceeded 90 % by trypan blue dye exclusion. As a measure of the depletion of hematopoiTable 1. Rate of phagocytosis of bone marrow cells (%) depending o nincubation temperature CC) Incu bation temperature
(0C)
4
20
37
• b
Rate of phagocytosis
Counted bone marrow celts
Bone marrow celts with phagocytosis
228 242 216 223
12 13 10 13
5.3 5.4 4.6 5. 8
219 227 205 223
29 35 36 33
13.2 15.4 17.6 14.8
211 216 225 207
63 68 75 62
29.9 31.5 33 .3 30.0
Arithmetic mean Standard error
x' ± SE b (0;', )
5.3 ± 0.5
15.2 ± 1.8
31.2±1.6
382 . B.
GRUHN,
R.
HAFER, A. MOLLER,
W.
ANDRA, H. DANAN,
and F.
ZINTL
etic stem cells that occurs with this procedure, CFU-GM were performed on marrow after two cycles of purging. An average of 82 % (3.2 x 104 out of 3.9 x 10 4 ) of the total CFU-GM was recovered after immunomagnetic purging. In each experiment 4 culture dishes were counted.
Discussion The system of magnetic separation used is based on the high-gradient magnetic separation described by OWEN (17). For the purging of human bone marrow, it is necessary that not only tumor cells but also all beads are removed from the suspension to avoid the risk of micro embolism in the patient (18). In the first experiment it was shown that the concentration of passed beads was reduced with decreasing flow rate and increasing magnetic field strength (Fig. 2). Using three chambers connected in series, no beads could be detected after passing through the last chamber. This result could be confirmed in all experiments. During the formation of cell-bead rosettes a plateau was achieved after an incubation time of 1 h (Fig. 3). Therefore, this time proved to be optimal, especially since longer incubation promotes the death of the cells. The same result was obtained by SEEGER et al. (19) in their experiments. We found that with rising concentration of cells, the rosetting rate increased, while only at a concentration of 1 x 10 7 cells/ml was a sufficient rosette formation achieved (Fig. 4). The reason for this phenomenon is that with rising concentration of cells the volume of incubation becomes smaller and the probability of beads becoming attached to cells increases. The same effect was observed by FAVROT et al. (20) in their model experiments. Furthermore, we could point out that with rising bead:tumor cell ratio the rosetting rate increased. In our experiments the bead:tumor cell ratio of 100:1 proved to be optimal, because here an almost complete rosette formation was achieved (Fig. 5). On the basis of their results KVALHEIM et al. (21) recommended a bead:tumor cell ratio of at least 75:1. A cocktail of three mAb resulted in a significantly (p < 0.01) higher rosette formation than with any single antibody (Fig. 6). This phenomenon is due to the intratumor heterogeneity in antigen expression and antigen density and has been described by many authors (18-20, 22). This heterogeneity is best understood for the hemopoietic malignancies that are thought to arise from the arresting of maturation during differentiation. Cells blocked at different stages of maturation express different profiles of intracellular or cell-membrane antigens (22). To differentiate leukemic cells from bone marrow cells, we stained SKW3 cells with the fluorescent DNA stain H 33342. REYNOLDS et al. (15) could reproducibly show that this staining allows the detection of a single tumor cell in 1 million bone marrow cells. Thus, in model experiments, a very small tumor cell infiltration can be detected. Therefore, the use of this dye is
Immunomagnetic Elimination of Cells from Bone Marrow . 383
very well suited for assessing of the efficiency of immunomagnetic elimination of tumor cells from bone marrow. Because of an incomplete antigen expression on the surface of SKW-3 cells with respect to the antibodies used, a maximal rosetting rate of 90.0 % (Fig. 6) was achieved under optimal conditions (cocktail of three mAb, room temperature). Therefore, a depletion of only 1 log of SKW-3 cells could be obtained after one cycle of magnetic separation. After a second cycle of treatment with both mAb and immunobeads, we observed a removal of an additional 1 log of SKW -3 cells. Therefore, we recommend two complete cycles of magnetic separation for an optimal elimination of tumor cells. The necessity of two sequential cycles is pointed out by other authors (19-21, 23). REYNOLDS et al. (23) supposed that the second cycle removes tumor cells that bind antibodies less efficiently than those removed in the first cycle. With the immunomagnetic elimination of SKW-3 cells from human bone marrow, we observed that beads attached not only to SKW-3 cells but also to 25 % bone marrow cells. This is accounted for by the presence of T cell antigens in bone marrow against which the mAb CD1, CD6, and CDS used are directed. Furthermore, we found 15 % bone marrow cells phagocytizing polystyrene beads. This phenomenon is based on the presence of monocytes, macrophages and granulocytes. The phagocytosis of the beads is undesirable, because the phagocytizing cells are retained in the magnetic field and so an endogenous source of colony-stimulating factor in the bone marrow is diminished. With decreasing incubation temperature the rate of phagocytosis was reduced. This phenomenon suggests that phagocytosis is an active energy-dependent process. Because rosette formation was significantly (p < 0.01) lower at 4°C, the temperature of 20°C proved to be optimal for the incubation of cells with beads. The loss of bone marrow cells was caused by washing procedures, binding and phagocytosis of magnetic beads. The marrow recovered after purging of leukemic cells was highly viable as assessed by both trypan blue dye exclusion and growth of CFU -GM. The results suggest that the purged marrow retains its proliferative capability which is necessary for hematological and immunological reconstitution. The percentage of normal cells and CFU-GM recovered indicates that this method is very well suited for the immunomagnetic elimination of tumor cells and can be used for autologous bone marrow transplantation.
References 1.
GALE, R. P. 1987. Bone marrow purging: current status, future directions. Bone Marrow Transplant 2 (Supp!. 2): 107. 2. GOLDIN, A. 1969. Factors pertaining to complete druginduced remission of tumor in animals and man. Cancer Res. 29: 2285.
384 . B. GRUHN, R. HAFER, A. MOLLER, W. ANORA, H. DANAN, and F. ZINTL 3. KAIZER, H., M. D. WHARAM, R. J. JOHNSON, J. G. ECONOMON, H. S. SHIN, G. W. SANTOS, G. J. ELFENBEIN, P. J. TUTSCHKA, H. G. BRAIN, L. L. MUNoz, and B. G. LEVENTHAL. 1980. Requirements for the successful application of autologous bone marrow transplantation in the treatment of selected malignancies. Haematol. Blood Transfus. 25: 285. 4. ELKINS, W. L. 1985. Preliminary studies of agglutination of metastatic neuroblastoma by soy bean lectin. In: EVANS, A. E., G. J. D'ANGIO, and R. C. SEEGER (eds.), Advances in Neuroblastoma Research. Alan R. Liss, Inc., New York. 405 pp. 5. FIGDOR, C. G., P. A. VOUTE, J. DE KRAKER, L. N. VERNIE, and W. S. BONT. 1985. Physical cell separation of neuroblastoma cells from bone marrow. In: EVANS, A. E., G. J. D'ANGIO, and R. C. Seeger (eds.), Advances in Neuroblastoma Research. Alan R. Liss, Inc., New York. 459 pp. 6. DEGLIANTONI, G., L. MANGONI, and V. RIZZOLI. 1985. In vitro restoration of polyclonal hematopoiesis in a chronic myelogenous leukemia after in vitro treatment with 4Hydroperoxycyclophosphamide. Blood 65: 753. 7. SANTOS, G. W., and O. M. COLVIN. 1986. Pharmacological purging of bone marrow with reference to autografting. Clin. Haematol. 15: 67. 8. SIEBER, F., S. ROA, S. D. ROWLEY, and M. SIEBER-BLUM. 1986. Dye-mediated photolysis of human neuroblastoma cells: implications for autologous bone marrow transplantation. Blood 68:32. 9. STEIN, J., S. STRANDjORD, U. SAARINEN, P. WARKENTIN, S. GERSON, H. LAZARUS, D. VON HOFF, P. COCCIA, and N.-K. V. CHEUNG. 1988. In vitro treatment of autologous bone marrow for neuroblastoma patients with anti G D2 monoclonal antibody and human complement: a pilot study. In: EVANS, A. E., G. J. D'ANGIO, A. G. KNUDSON, and R. C. SEEGER (eds.), Advances in Neuroblastoma Research 2. Alan R. Liss, Inc., New York, 237 pp. 10. CASELLAS, P., C. CANAT, A. A. FAUSER, O. GROS, G. LAURENT, P. PONCELET, and F. K. JANSEN. 1985. Optimal elimination of leukemic T cells from human bone marrow with T lOl-ricin A-chain immunotoxin. Blood 65: 289. 11. EMBLETON, M. J., G. F. ROWLAND, R. G. SIMMONDS, E. JACOBS, C. H. MARSDEN, and R. W. BALDWIN. 1983. Selective cytotoxicity against human tumour cells by a vindesinmonoclonal antibody conjugate. Br. J. Cancer 47: 43. 12. UGELSTAD, J., K. H. KAGGERUD, F. K. HANSEN, and A. BERGE. 1979. Absorption of low molecular weight compounds in aqueous dispersions of polymer-oligomer particles: A two step swelling process of polymer particles giving an enormous increase in absorption capacity. Makromol. Chern. 180: 737. 13. UGELSTAD, J., L. SODERBERG, A. BERGE, and J. BERGSTROM. 1983. Monodisperse polymer particles - a step forward for chromatography. Nature 303: 95. 14. BOYUM, A. 1968. Isolation of mononuclear cells and granulocytes from human blood. Scand. J. Clin. Lab. Invest. 21: 77. 15. REYNOLDS, C. P., T. J. Moss, R. C. SEEGER, A. T. BLACK, and J. N. WOODY. 1985. Sensitive detection of neuroblastoma cells in bone marrow for monitoring the efficacy of marrow purging procedures. In: EVANS, A. E., G. J. D'ANGIO, and R. C. SEEGER (eds.), Advances in Neuroblastoma Research. Alan R. Liss, Inc., New York. 425 pp. 16. WEBER, E. 1980. GrundriG der biologischen Statistik, 8. Aufl. VEB Gustav Fischer Verlag Jena. 340 pp. 17. OWEN, C. S. 1983. Magnetic cell sorting. In: PRETLOW, T. G., and T. P. PRETLOW (eds.), Cell separation: methods and selected applications. Academic Press Inc., New York. 127 pp. 18. TRELEAVEN, J. G., F. M. GIBSON, J. UGELSTAD, A. REMBAUM, T. PHILIP, G. D. CAINE, and J. T. KEMSHEAD. 1984. Removal of neuroblastoma cells from bone marrow with monoclonal antibodies conjugated to magnetic microspheres. Lancet 1: 70. 19. SEEGER, R. C., C. P. REYNOLDS, D. D. Yo,]. UGELSTAD, and]. WELLS. 1985. Depletion of neuroblastoma cells from bone marrow with monoclonal antibodies and magnetic immunobeads. In: EVANS, A. E., G.]. D'ANGIO, and R. C. SEEGER (eds.), Advances in Neuroblastoma Research. Alan R. Liss, Inc., New York. 443 pp.
Immunomagnetic Elimination of Cells from Bone Marrow . 385 20. FAVROT, M. C, I. PHILIP, V. COMBARET, O. MARITAZ, and T. PHILIP. 1987. Experimental evaluation of an immunomagnetic bone marrow purging procedure using the Burkitt lymphoma model. Bone Marrow Transplant. 2: 59. 21. KVALHEIM, G., O. FODSTAD, A. PIHL, K. NUSTAD, A. PHARO, J. UGELSTAD, and S. FUNDERUD. 1987. Elimination of B-lymphoma from human bone marrow: model experiments using monodisperse magnetic particles coated with primary monoclonal antibodies. Cancer Res. 47: 846. 22. KEMSHEAD, J. T., A. GOLDMAN, J. FRITSCHY, J. S. MALPAS, and J. PRITCHARD. 1983. Use of panels of monoclonal antibodies in the differential diagnosis of neuroblastoma and lymphoblastic disorders. Lancet 1: 12. 23. REYNOLDS, C P., R. C SEEGER, D. D. Vo, A. T. BLACK, J. WELLS, and J. UGELSTAD. 1986. Model system for removing neuroblastoma cells from bone marrow using monoclonal antibodies and magnetic immunobeads. Cancer Res. 46: 5882. Dr. BERND GRUHN, Department of Pediatrics, Friedrich Schiller University, Kochstralle 2, 0-6900 Jena, Germany
of Pediatrics, Friedrich Schiller University, 2 Physico-Technical Institute, Jena, Germany, and 3 Laboratoire Pierre Weiss, Universite Louis Pasteur, Strasbourg, France
Model Experiments for Immunomagnetic Elimination of Leukemic Cells from Human Bone Marrow. Presentation of a Novel Magnetic Separation System BERND GRUHN1, RALF HAFER 1, ANGELIKA MULLER \ WILFRIED ANDRA2 , HENRI DANAN3 , and FELIX ZINTL 1 Received February 11,1991 . Accepted in Revised Form June 19, 1991
Abstract Optimal conditions for removing leukemic cells from human bone marrow with monoclonal antibodies (mAb) and magnetic immunobeads were investigated. Monodisperse 3 [!m polystyrene micro spheres containing magnetite were coated with affinity-purified rabbit antimouse IgG at 4°C, pH 9.6 for ISh. SKW-3 cells (T-CLL cell line) were marked with the supravital DNA stain Hoechst 33342, seeded into normal human bone marrow, and then incubated with the mAb CD1, CD6, and CDS at 4°C for 30 min. In preliminary experiments REH cells (cALL cells) and mouse anti-REH cell antibodies were used to find the most favorable conditions for the binding of magnetic beads to tumor cells. Optimal formation of cell-bead rosettes was achieved by rotating beads and tumor cells together at room temperature at a concentration of 1 x 10 7 cells/ml, a bead: tumor cell ratio of 100:1 and an incubation time of one hour. The novel magnetic separation apparatus consists of three polystyrene chambers connected by silicone rubber tubing. The chambers contain four steel inserts each equipped with 32 nickel wires, which are magnetized by permanent magnets in such a way that the inhomogeneous high gradient magnetic field could be established within the cell suspension containing the cells to be depleted. The fluid flow was established by a peristaltic pump. At a flow rate of 1.5 mlimin and a field strength of 160 kA/m, no beads could be detected in the purged marrow. A cocktail of the three mAb was more effective than any single antibody in forming bead-cell rosettes. Two sequential purging cycles were superior to one. The marrow recovered was highly viable as assessed by trypan blue dye exclusion and by growth of CFUGM.
Introduction Most tumors show a dose-related response to anti-neoplastic drugs and radiation. This means that higher doses result in increased tumor cell death (1). However, toxicity for normal tissues, especially for bone marrow, is dose-limiting (2). This problem can be overcome by autologous bone marrow transplantation. However, there is the risk that in tumors which have their origin in the bone marrow or which metastasize to the bone Abbreviations: mAb = monoclonal antibody; CLL = chronic lymphoblastic leukemia; cALL = common acute lymphoblastic leukemia; CD = cluster of differentiation; CPU-GM = colony-forming units - granulocyte - macrophage
Immunomagnetic Elimination of Cells from Bone Marrow . 375
marrow occult tumor cells may occur in the harvested bone marrow. These cells are capable of causing a relapse after reinfusion (1, 3). To avoid this effect, methods were sought for to remove tumor cells in vitro from bone marrow. Physical (4, 5), pharmacological (6-8) and immunological (9-11) purging procedures have already been described. One modern immunological method is the immunomagnetic elimination of tumor cells.
Materials and Methods Principle
Primary antibodies directed against tumor cell antigens are bound to the surface of tumor cells. Secondary antibodies of another species which detect primary antibodies are coated onto the surface of magnetic beads. The subsequent joint incubation allows the binding of magnetic beads to tumor cells on the basis of antigen-antibody reaction with the help of this indirect method. After incubation, cells with bound beads are retained in the magnetic system. Magnetic beads
The magnetic beads used were produced by Dr. J. UGELSTAD (Norwegian Institute of Technology, University of Trondheim, Norway). They have a diameter of 3 [Am, consist of polystyrene and contain magnetite (12, 13). 50 mg magnetic beads were suspended in 100 ml isotonic sodium chloride solution for the separation experiments and dispersed by means of an ultrasonic disintegrator (HR-30 model of Elektromat, Dresden, Germany) and an agitator. Magnetic separation system
One electromagnet, four permanent magnets and three chambers situated between the magnets were available for the magnetic separation of the beads from the suspension. The chambers were connected by silicone rubber tubing, and the fluid flow was established by a
Figure 1. Polystyrene chamber with 4 inserts and a single insert with nickel wires on the left side.
376 . B. GRUHN, R. HAFER, A. MOLLER, W. ANDRA, H. DANAN, and F. ZINTL peristaltic pump. The electromagnet used consisted of a soft iron core, two coils with a total resistance of 22 ohms and soft iron pole pieces between which a magnetic field strength of 40 kA/m ... 185 kA/m could be generated. As permanent magnets, we used rectangle magnets having a size of 50 x 30 x 24 mm} (Keramische Werke, Hermsdorf, Germany). These were hard ferrites between which a magnetic field strength of 160 kA/m could be achieved after adequate remagnetization. The three chambers consist of polystyrene, have an internal volume of 2 ml each and contain four stainless-steel inserts with the external size of 20 x 10 x 5 mm} (Fig. 1). Each insert was divided into four planes, and in each plane there were nickel wires with a diameter of 0.2 mm. One chamber insert contained 32 nickel wires in total, the number of wires per plane being 5, 7, 9, and 11. During the experiments the chambers were so installed that the number of wires per plane of an insert increases in flow direction to obtain a filter effect. The axis of the nickel wires stands perpendicular to the magnetic field, so that an optimal magnetization of the wires is achieved. The wires themselves induce a highly effective inhomogeneous magnetic field. As a result, the beads moving with the flow of the fluid get directly to the wires and are retained there. Detection of passed magnetic beads after separation
The concentration of passed beads was determined with the aid of the Neubauer counting chamber. We were able to increase the sensitivity of the bead-counting method essentially by centrifugation of the suspension at 3000 rim in for 5 min. Thus, the volume could be reduced 1000-fold, and the concentration could be increased 1000-fold. A further increase in sensitivity was obtained by counting all large squares of the counting chamber. By these means, the detection limit of the concentration of beads could be reduced to 1 bead/m!. Incubation of magnetic beads with antibodies
The micro spheres were coated with affinity-purified rabbit anti-mouse IgG (Dr. K.-H. VOGT, Institute of Medical Microbiology, Friedrich Schiller University, Jena, Germany). This antibody was dissolved in sodium carbonate buffer (0.015 M Na2C03, 0.D35 M NaHC0 3), pH 9.6 at a concentration of 200!!g protein/m!. 1 ml antibody buffer solution was added to 1 mg (7.9 X 107) beads. The incubation of beads with antibodies was performed for 18 hat 4°C by means of an agitator. Then the suspension was washed twice with isotonic sodium chloride solution and once again with RPMI 1640 medium (Institut fUr Immunpriiparate und Niihrmedien, Berlin, Germany) to remove unbound antibodies. Cell material
Two different permanent cell lines were available for the experiments: REH cells (cALL cells) and SKW-3 cells (T-CLL cells). In addition, a mononuclear bone-marrow cell fraction, harvested according to a modified version of BOYUM'S method (14) was used. Normal human bone marrow was obtained from the iliac crest of healthy volunteers. After addition of preservative-free heparin, bone marrow was added to a mixture of Infukoll 6 % (Serumwerk, Bernburg, Germany) and Visotrast 370 (Fahlberg-List, Magdeburg, Germany) with the density of 1.09 g/cm 3 • After sedimentation for 1 h, the upper erythrocyte-poor fraction was taken and centrifuged for 10 min at 1500 r/min. The sediment was resuspended in isotonic sodium chloride solution and added to a mixture of Ficoll 400 (Pharmacia Fine Chemicals AB, Uppsala, Sweden) and Visotrast 370 of the density of 1.13 g/cm 3. After density centrifugation for 20 min at 1500 r/ min, the bulk of the polymorphonuclear cells was separated. The mononuclear cell fraction was taken from the borderline and washed three times with RPMI 1640 medium. Incubation of cells with antibodies
To bind REH cells with antibodies, mouse anti-REH cell serum (non-absorbed crude serum) was used. Three mAb VIT-6 (CDl), VIT-12 (CD6) and VIT-8 (CD8) (kindly provided
Immunomagnetic Elimination of Cells fro m Bo ne Marrow . 377 by Prof. Dr. W. KNAPP, Institute fo r Immunology, Universi ty of Vi en na, Au stria) were available for the binding to the SKW -3 cells. The incubatio n of cells with antibodies directed against the cells was performed for 30 min at 4 °C a t aconcentration of immunoglobulin having at least 30 Jlg/ I x107 cells/ m!. The unbound antibodies were removed by washing twice with RPM I 1640.
Staining of tumor cells To differentiate SKW-3 cells fro m bone marrow cells, SKW-3 cells were stained with the fluorescent supravital stain Hoechst 33342 (H 33342) (SERVA, Heidelberg, Germany) that links with the chromosomal DNA of the cells (15).
Statistics Data represent the ari thmetic mean from 4 independent experiments. The statistical analysis was carried out with the help of Wilcoxon's paired rank sum test (16).
./ ./ . //
."',,,'" ./
'"
'"
.,.-
'//'"
-- 40 k A / m
./
.'1/
12 0 k A / m
.'1/
18Sk A /m
.II
/1
./
I
~1
I
L-~~
____
1.01~ 2.5
~
__________
5.0
~
______________________
10.0
~~
20.0
Flow rote Iml / mln I
Figure 2. Concentration of beads passed (beads/ml) depending on flow rate (mUm in) and magnetic field strength (kA/ m). Each point represents the arithmetic mean of 4 experiments.
378 . B. GRUHN, R. HAFER, A. MOLLER, W. ANDRA, H. DANAN, and F.
ZINTL
Results U sing an electromagnet and one chamber we first investigated under what conditions the beads were retained in the magnetic field. It could be shown that with decreasing flow rate and increasing magnetic field strength significantly (p < 0.025) fewer beads passed the chamber (Fig. 2). By means of three chambers connected in series and situated between four permanent magnets a further improvement of the separation efficiency was achieved. A flow rate of 1.5 mllmin proved to be optimal - no beads could be detected after passing through the last chamber. REH cells with bound mouse antiREH cell antibodies were added to the microspheres coated with rabbit anti-mouse antibodies. The incubation was performed at room temperature under gentle rotation. In each of the 4 experiments at least 200 REH cells were analyzed for bound beads. The optimal formation of rosettes was obtained at an incubation time of 1 h (Fig. 3), at a cell concentration of 1 x 107 cells/ml (Fig. 4) and a bead: tumor cell ratio of 100:1 (Fig. 5). Under these conditions a rosetting rate of 99.9 ± 0.1 % was reached. In 4 separation experiments we observed a reduction from (4.0 ± 0.4) x 105 REH cells/ml to (4.1 ± 0.3) x 102 REH cells/ml after magnetic separation. Depending on the rosette formation, a removal of 3 logs of tumor cells was achieved. For the immunomagnetic elimination of SKW-3 cells from human bone marrow, it was necessary to use mAb against these tumor cells. By means of the rosette formation we observed that the three mAb CD1, CD6, and CDS were bound to a different extent by the SKW-3 cells. A cocktail of the three mentioned mAb resulted in a significantly (p < 0,01) higher rosetting rate (Fig. 6). 100 80
r----
I i
20
0.25
0.5
to
2.0 t
Time (h)
Figure 3. Percentage of rosettes formed (%) depending on incubation time (h) (REH cell concentration 1.6 x 106 cells/ml, bead:tumor cell ratio 5: 1). Each point represents the arithmetic mean of 4 experiments, and vertical bars denote standard error of the mean.
Immunomagnetic Elimination of Cells from Bone Marrow . 379
100
80 60 ~
'" 2!
40
Qj
'" 0
a:
20
1cf> Cell concentration (cells I ml)
Figure 4. Percentage of rosettes formed (%) depending on REH cell concentration (cells/ml) (incubation time 15 minutes, bead:tumor cell ratio 5:1). Each bar represents the arithmetic mean ± SE from 4 experiments.
The formation of rosettes was significantly (p < 0.01) lower at 4°C than at 20°C or 37 °C, while between the latter temperatures no significant difference occurred (Fig. 6). In our cell separation experiments, the rosette formation was carried out at room temperature.
100
80 ~
- 60
III
QI
:s: ~
fi.
40
20
2: 1
5:1
10 :1
Bead : tumor cell ratio
Figure 5. Percentage of rosettes formed (%) depending on bead:tumor cell ratio (incubation time 15 min, REH cell concentration 4 x 105 cells/ml). Each bar represents the arithmetic mean ± SE from 4 experiments.
~ CO S
§
ESl
C01
C06
mID Coc..ktml
100
so ~ 60
20
20 Temperature (O e I
37
Immunomagnetic Elimination of Cells from Bone Marrow' 381
T""T=- - O
•
•
•
Figure 7. H 33342-stained SKW-3 cells with bound magnetic beads (a) and unstained bone marrow cells which in part bind (b) or phagocytize (c) magnetic beads. Magnification x 600.
The concentration of bone marrow cells was reduced to (5.5 ± 0.2) x 106 cells (55 %) after the first cycle and (3.0 ± 0.2) x 106 cells (30 %) after the second cycle of magnetic separation. However, viability of the recovered bone marrow always exceeded 90 % by trypan blue dye exclusion. As a measure of the depletion of hematopoiTable 1. Rate of phagocytosis of bone marrow cells (%) depending o nincubation temperature CC) Incu bation temperature
(0C)
4
20
37
• b
Rate of phagocytosis
Counted bone marrow celts
Bone marrow celts with phagocytosis
228 242 216 223
12 13 10 13
5.3 5.4 4.6 5. 8
219 227 205 223
29 35 36 33
13.2 15.4 17.6 14.8
211 216 225 207
63 68 75 62
29.9 31.5 33 .3 30.0
Arithmetic mean Standard error
x' ± SE b (0;', )
5.3 ± 0.5
15.2 ± 1.8
31.2±1.6
382 . B.
GRUHN,
R.
HAFER, A. MOLLER,
W.
ANDRA, H. DANAN,
and F.
ZINTL
etic stem cells that occurs with this procedure, CFU-GM were performed on marrow after two cycles of purging. An average of 82 % (3.2 x 104 out of 3.9 x 10 4 ) of the total CFU-GM was recovered after immunomagnetic purging. In each experiment 4 culture dishes were counted.
Discussion The system of magnetic separation used is based on the high-gradient magnetic separation described by OWEN (17). For the purging of human bone marrow, it is necessary that not only tumor cells but also all beads are removed from the suspension to avoid the risk of micro embolism in the patient (18). In the first experiment it was shown that the concentration of passed beads was reduced with decreasing flow rate and increasing magnetic field strength (Fig. 2). Using three chambers connected in series, no beads could be detected after passing through the last chamber. This result could be confirmed in all experiments. During the formation of cell-bead rosettes a plateau was achieved after an incubation time of 1 h (Fig. 3). Therefore, this time proved to be optimal, especially since longer incubation promotes the death of the cells. The same result was obtained by SEEGER et al. (19) in their experiments. We found that with rising concentration of cells, the rosetting rate increased, while only at a concentration of 1 x 10 7 cells/ml was a sufficient rosette formation achieved (Fig. 4). The reason for this phenomenon is that with rising concentration of cells the volume of incubation becomes smaller and the probability of beads becoming attached to cells increases. The same effect was observed by FAVROT et al. (20) in their model experiments. Furthermore, we could point out that with rising bead:tumor cell ratio the rosetting rate increased. In our experiments the bead:tumor cell ratio of 100:1 proved to be optimal, because here an almost complete rosette formation was achieved (Fig. 5). On the basis of their results KVALHEIM et al. (21) recommended a bead:tumor cell ratio of at least 75:1. A cocktail of three mAb resulted in a significantly (p < 0.01) higher rosette formation than with any single antibody (Fig. 6). This phenomenon is due to the intratumor heterogeneity in antigen expression and antigen density and has been described by many authors (18-20, 22). This heterogeneity is best understood for the hemopoietic malignancies that are thought to arise from the arresting of maturation during differentiation. Cells blocked at different stages of maturation express different profiles of intracellular or cell-membrane antigens (22). To differentiate leukemic cells from bone marrow cells, we stained SKW3 cells with the fluorescent DNA stain H 33342. REYNOLDS et al. (15) could reproducibly show that this staining allows the detection of a single tumor cell in 1 million bone marrow cells. Thus, in model experiments, a very small tumor cell infiltration can be detected. Therefore, the use of this dye is
Immunomagnetic Elimination of Cells from Bone Marrow . 383
very well suited for assessing of the efficiency of immunomagnetic elimination of tumor cells from bone marrow. Because of an incomplete antigen expression on the surface of SKW-3 cells with respect to the antibodies used, a maximal rosetting rate of 90.0 % (Fig. 6) was achieved under optimal conditions (cocktail of three mAb, room temperature). Therefore, a depletion of only 1 log of SKW-3 cells could be obtained after one cycle of magnetic separation. After a second cycle of treatment with both mAb and immunobeads, we observed a removal of an additional 1 log of SKW -3 cells. Therefore, we recommend two complete cycles of magnetic separation for an optimal elimination of tumor cells. The necessity of two sequential cycles is pointed out by other authors (19-21, 23). REYNOLDS et al. (23) supposed that the second cycle removes tumor cells that bind antibodies less efficiently than those removed in the first cycle. With the immunomagnetic elimination of SKW-3 cells from human bone marrow, we observed that beads attached not only to SKW-3 cells but also to 25 % bone marrow cells. This is accounted for by the presence of T cell antigens in bone marrow against which the mAb CD1, CD6, and CDS used are directed. Furthermore, we found 15 % bone marrow cells phagocytizing polystyrene beads. This phenomenon is based on the presence of monocytes, macrophages and granulocytes. The phagocytosis of the beads is undesirable, because the phagocytizing cells are retained in the magnetic field and so an endogenous source of colony-stimulating factor in the bone marrow is diminished. With decreasing incubation temperature the rate of phagocytosis was reduced. This phenomenon suggests that phagocytosis is an active energy-dependent process. Because rosette formation was significantly (p < 0.01) lower at 4°C, the temperature of 20°C proved to be optimal for the incubation of cells with beads. The loss of bone marrow cells was caused by washing procedures, binding and phagocytosis of magnetic beads. The marrow recovered after purging of leukemic cells was highly viable as assessed by both trypan blue dye exclusion and growth of CFU -GM. The results suggest that the purged marrow retains its proliferative capability which is necessary for hematological and immunological reconstitution. The percentage of normal cells and CFU-GM recovered indicates that this method is very well suited for the immunomagnetic elimination of tumor cells and can be used for autologous bone marrow transplantation.
References 1.
GALE, R. P. 1987. Bone marrow purging: current status, future directions. Bone Marrow Transplant 2 (Supp!. 2): 107. 2. GOLDIN, A. 1969. Factors pertaining to complete druginduced remission of tumor in animals and man. Cancer Res. 29: 2285.
384 . B. GRUHN, R. HAFER, A. MOLLER, W. ANORA, H. DANAN, and F. ZINTL 3. KAIZER, H., M. D. WHARAM, R. J. JOHNSON, J. G. ECONOMON, H. S. SHIN, G. W. SANTOS, G. J. ELFENBEIN, P. J. TUTSCHKA, H. G. BRAIN, L. L. MUNoz, and B. G. LEVENTHAL. 1980. Requirements for the successful application of autologous bone marrow transplantation in the treatment of selected malignancies. Haematol. Blood Transfus. 25: 285. 4. ELKINS, W. L. 1985. Preliminary studies of agglutination of metastatic neuroblastoma by soy bean lectin. In: EVANS, A. E., G. J. D'ANGIO, and R. C. SEEGER (eds.), Advances in Neuroblastoma Research. Alan R. Liss, Inc., New York. 405 pp. 5. FIGDOR, C. G., P. A. VOUTE, J. DE KRAKER, L. N. VERNIE, and W. S. BONT. 1985. Physical cell separation of neuroblastoma cells from bone marrow. In: EVANS, A. E., G. J. D'ANGIO, and R. C. Seeger (eds.), Advances in Neuroblastoma Research. Alan R. Liss, Inc., New York. 459 pp. 6. DEGLIANTONI, G., L. MANGONI, and V. RIZZOLI. 1985. In vitro restoration of polyclonal hematopoiesis in a chronic myelogenous leukemia after in vitro treatment with 4Hydroperoxycyclophosphamide. Blood 65: 753. 7. SANTOS, G. W., and O. M. COLVIN. 1986. Pharmacological purging of bone marrow with reference to autografting. Clin. Haematol. 15: 67. 8. SIEBER, F., S. ROA, S. D. ROWLEY, and M. SIEBER-BLUM. 1986. Dye-mediated photolysis of human neuroblastoma cells: implications for autologous bone marrow transplantation. Blood 68:32. 9. STEIN, J., S. STRANDjORD, U. SAARINEN, P. WARKENTIN, S. GERSON, H. LAZARUS, D. VON HOFF, P. COCCIA, and N.-K. V. CHEUNG. 1988. In vitro treatment of autologous bone marrow for neuroblastoma patients with anti G D2 monoclonal antibody and human complement: a pilot study. In: EVANS, A. E., G. J. D'ANGIO, A. G. KNUDSON, and R. C. SEEGER (eds.), Advances in Neuroblastoma Research 2. Alan R. Liss, Inc., New York, 237 pp. 10. CASELLAS, P., C. CANAT, A. A. FAUSER, O. GROS, G. LAURENT, P. PONCELET, and F. K. JANSEN. 1985. Optimal elimination of leukemic T cells from human bone marrow with T lOl-ricin A-chain immunotoxin. Blood 65: 289. 11. EMBLETON, M. J., G. F. ROWLAND, R. G. SIMMONDS, E. JACOBS, C. H. MARSDEN, and R. W. BALDWIN. 1983. Selective cytotoxicity against human tumour cells by a vindesinmonoclonal antibody conjugate. Br. J. Cancer 47: 43. 12. UGELSTAD, J., K. H. KAGGERUD, F. K. HANSEN, and A. BERGE. 1979. Absorption of low molecular weight compounds in aqueous dispersions of polymer-oligomer particles: A two step swelling process of polymer particles giving an enormous increase in absorption capacity. Makromol. Chern. 180: 737. 13. UGELSTAD, J., L. SODERBERG, A. BERGE, and J. BERGSTROM. 1983. Monodisperse polymer particles - a step forward for chromatography. Nature 303: 95. 14. BOYUM, A. 1968. Isolation of mononuclear cells and granulocytes from human blood. Scand. J. Clin. Lab. Invest. 21: 77. 15. REYNOLDS, C. P., T. J. Moss, R. C. SEEGER, A. T. BLACK, and J. N. WOODY. 1985. Sensitive detection of neuroblastoma cells in bone marrow for monitoring the efficacy of marrow purging procedures. In: EVANS, A. E., G. J. D'ANGIO, and R. C. SEEGER (eds.), Advances in Neuroblastoma Research. Alan R. Liss, Inc., New York. 425 pp. 16. WEBER, E. 1980. GrundriG der biologischen Statistik, 8. Aufl. VEB Gustav Fischer Verlag Jena. 340 pp. 17. OWEN, C. S. 1983. Magnetic cell sorting. In: PRETLOW, T. G., and T. P. PRETLOW (eds.), Cell separation: methods and selected applications. Academic Press Inc., New York. 127 pp. 18. TRELEAVEN, J. G., F. M. GIBSON, J. UGELSTAD, A. REMBAUM, T. PHILIP, G. D. CAINE, and J. T. KEMSHEAD. 1984. Removal of neuroblastoma cells from bone marrow with monoclonal antibodies conjugated to magnetic microspheres. Lancet 1: 70. 19. SEEGER, R. C., C. P. REYNOLDS, D. D. Yo,]. UGELSTAD, and]. WELLS. 1985. Depletion of neuroblastoma cells from bone marrow with monoclonal antibodies and magnetic immunobeads. In: EVANS, A. E., G.]. D'ANGIO, and R. C. SEEGER (eds.), Advances in Neuroblastoma Research. Alan R. Liss, Inc., New York. 443 pp.
Immunomagnetic Elimination of Cells from Bone Marrow . 385 20. FAVROT, M. C, I. PHILIP, V. COMBARET, O. MARITAZ, and T. PHILIP. 1987. Experimental evaluation of an immunomagnetic bone marrow purging procedure using the Burkitt lymphoma model. Bone Marrow Transplant. 2: 59. 21. KVALHEIM, G., O. FODSTAD, A. PIHL, K. NUSTAD, A. PHARO, J. UGELSTAD, and S. FUNDERUD. 1987. Elimination of B-lymphoma from human bone marrow: model experiments using monodisperse magnetic particles coated with primary monoclonal antibodies. Cancer Res. 47: 846. 22. KEMSHEAD, J. T., A. GOLDMAN, J. FRITSCHY, J. S. MALPAS, and J. PRITCHARD. 1983. Use of panels of monoclonal antibodies in the differential diagnosis of neuroblastoma and lymphoblastic disorders. Lancet 1: 12. 23. REYNOLDS, C P., R. C SEEGER, D. D. Vo, A. T. BLACK, J. WELLS, and J. UGELSTAD. 1986. Model system for removing neuroblastoma cells from bone marrow using monoclonal antibodies and magnetic immunobeads. Cancer Res. 46: 5882. Dr. BERND GRUHN, Department of Pediatrics, Friedrich Schiller University, Kochstralle 2, 0-6900 Jena, Germany
