Vol. 15, No. 1, pp. 29-36, 1991. Printedin Great Britain.
0145-2126/91 $3.00 + 0.00 PergamonPresspie
Leukemia Research
I N VITRO EFFECTS OF RECOMBINANT HEMOPOIETIC G R O W T H FACTORS ON P R O G E N I T O R CELLS FROM PATIENTS WITH MYELODYSPLASTIC SYNDROMES XAVIER THOMAS,* DENIS G U Y O T A T , t LYDIA CAMPOS,* ZHI HUA SHI,* LILIANA VILA,~ ALAIN EHRSAM* a n d DENIS FIERE~
*Service de Cryobiologie, Centre de Transfusion Sanguine, Lyon, France, i'D6partement d'H6matologie, H6pital Nord, St Priest en Jarez, France and ~tService d'H6matologie, H6pital Edouard Herriot, Lyon, France
(Received 23 February 1990. Revision accepted 17 July 1990) Abstract--The effects of four recombinant hemopoietic growth factors (HGF) and of the impure factor HTB9 on proliferation and maturation of marrow myeloid (CFU-GM) and erythroid (BFU-E) progenitor cells were studied in 22 cases of myelodysplastic syndromes (MDS). In most cases, IL-3, GM-CSF and G-CSF increased significantly the number of myeloid colonies, the best combination being 1L-3 + GM-CSF. A significant increase in the myeloid colony/cluster ratio was also noted, but cytological examination of colony cells showed little maturation. The analysis of myeloid colony surface markers with four monoclonal antibodies (to CD13, CD15, CD33 and CD34) showed minor modifications with an increase of CD13 and CD15 in about one third of cases when compared to control without HGF. Erythroid colonies were obtained in one case with erythropoietin alone, and in 19 cases with the addition of GM-CSF and/or IL-3. In short-term liquid cultures, IL-3, GM-CSF and G-CSF increased 3H-thymidine incorporation. We conclude that progenitor cells of most MDS are able to proliferate in the presence of HGF, with wide case-to-case variations. However, the pattern of growth remains abnormal when compared to normal marrow. Although the combination of IL-3 and GM-CSF is the most efficient, there is a large overlap in the stimulating effects of all factors studied.
Key words: Hemopoietic growth factors, IL-3, GM-CSF, myelodysplastic syndromes, progenitor cells.
erogeneous group of stem-cell diseases characterized by maturation defects with ineffective hemopoiesis and peripheral blood cytopenias [1]. The evolution to overt acute leukemia is frequently observed, particularly in refractory anemia with excess of blasts ( R A E B ) , or R A E B in transformation ( R A E B - T ) [1, 2]. The in vitro study of hemopoietic progenitors using conditioned medium as stimulating factor frequently shows an abnormal growth pattern [1-4]. Although abnormalities are not directly correlated with FAB sub-type, a growth of a low number of myeloid colonies with or without excess of clusters is frequently predictive of leukemic evolution [3, 4]. Hemopoietic growth factors ( H G F ) have recently been produced by recombinant D N A techniques and
their in vitro activities on proliferation and maturation of normal marrow cells have been reported [5-7]. Granulocyte-macrophage stimulating factor (GM-CSF) is able to support the growth of CFUGM in most MDS cases, whereas its effect on BFUE growth is inconstant [8, 9]. Interleukin 3 (IL-3), GM-CSF, granulocyte-CSF (G-CSF) and macrophage-CSF (M-CSF) alone were not able either to restore a normal growth pattern of erythroid, myeloid or mixed colonies in a series of ten cases of R A E B or R A E B - T [10]. In this study, we compared the effects of three recombinant H G F s (IL-3, GMCSF and G-CSF) on granulocyte-macrophage and erythroid colony growth in 22 patients with different types of MDS. In addition, we studied the cytological and immunological differentiation of myeloid colonies, and the proliferation of MDS cells in short-term liquid culture.
Correspondence to: Dr D. Guyotat, D6partement d'H6matologie, H6pital Nord, Avenue Albert-Raimond, 42277 St Priest en Jarez Cedex, France.
Patients Bone marrow was obtained from 22 patients with MDS
INTRODUCTION THE MYELODYSPLASTICsyndromes (MDS) are a het-
MATERIALS AND METHODS
29
30
X. THOMASet al.
(12 males and 10 females) by sternum or posterior iliac crest aspiration. The diagnosis was based on cell morphology and cytochemical staining according to the French--American-British (FAB) Cooperative Group criteria [11]. The FAB typing was: refractory anemia (RA): four cases; refractory anemia with ringed sideroblasts (RAS): two cases; refractory anemia with excess of blasts (RAEB): nine cases; RAEB in transformation (RAEB-T): four cases; chronic myelomonocytic leukemia (CMML): three cases. Two patients (7 and 13) had received prior therapy with methyl-prednisone and retinoic acid, respectively, and one (case I) had received chemotherapy for previous malignant disease. These patients had not been treated for at least 6 weeks at the time of study. Only four patients (patients 13, 14, 15 and 20) had required previous red-cell transfusions. Normal marrow samples used as controls were obtained from healthy donors for marrow transplantation. All patients gave informed consent to donate bone marrow.
Bone marrow cells Mononuclear cells were separated by centrifugation (433g, 30min) on Ficoll-Hypaque density gradient (Eurobio, Paris), washed twice in RPMI 1640 (Flow Laboratories, Irvine, U.K.) and finally resuspended in Iscove's modified Dulbecco's medium (IMDM). Individual cell fractions were then separated for cultures, thymidine incorporation, morphological (May-Griinwald-Giemsa staining) and surface marker analysis. Hemopoietic growth ]'actors Recombinant GM-CSF from E. coil expression system (biological activity of 5 × 107CFU/mg protein), and recombinant IL-3 and G-CSF from yeast expression system (biological activity of 1 x 107 CFU/mg protein) were generously provided by BehringWerke AG (Marburg, F.R.G.). Recombinant erythropoietin (Terry Fox Laboratories, Vancouver, Canada) was used at the concentration of 2 U/ml. Conditioned medium of 5637 human bladder carcinoma cell line (HTB9) was used as positive control. HTB9 is a rich source of interleukin-1, G-CSF and GM-CSF, and stimulates the proliferation of normal progenitors [12] and of leukemic [13] and MDS cells [8]. Myeloid progenitor (CFU-GM) assay A total of 15 × 104 cells were plated per well (24 Multidish Nunc, Ruskilde, Denmark) containing 0.5 ml IMDM supplemented with 20% fetal calf serum (FCS, GIBCO, Paisley, U.K.), 0.8% methylcellulose (Methocel MC, Fluka, Switzerland), and HGFs used alone or in combination. HGFs were used at a concentration of 50 U/ml. Negative and positive controls were performed without HGF and with 10% HTB9, respectively. Granulocyte-monocyte colony-forming units (CFUGM) were counted after 7-10 days incubation at 37°C in humidified atmosphere with 5% CO2. Colonies (more than 50 cells) and clusters (20-50 cells) were scored under an inverse microscope. Colony and cluster numbers represent means of triplicate cultures. After culture, individual colonies were collected using fine pipettes, pooled, washed in phosphate-buffered saline (PBS) and analyzed by morphology and immunology to evaluate the effects of HGFs on cell maturation in vitro.
Erythroid progenitor (BFU-E) assay Cells (0.5 × 105) were cultured in 1 ml IMDM containing 0.3% agar supplemented with 1% deionized bovine serum albumin (fraction V, Sigma), 20% FCS, 5 × 10-5 mercaptoethanol and erythropoietin (Epo) (2 U/ml). Erythroid colonies (>50 hemoglobinized cells) were counted after 14 days of culture at 37°C in 5% CO2 humidified atmosphere. Immunological study Surface markers of diagnosis cells and CFU-GM cells were analyzed by indirect immunofluorescence. The following monoclonal antibodies (MoAb) were used: MY10 (CD34), MY7 (CD13), MY9 (CD33) from Coulter Immunology (Hialeah, FL, U.S.A.); FMyl5a (CD15) (Biosys, Meudon). The cells washed twice in PBS were incubated with the MoAb at 4°C for 30 min. FITC-labelled sheep anti-mouse Fab~ fragments (Bioart, Meudon) were used as second layer (4°C, 30 min). Controls were performed using the MslgG and MslgM kits (Coulter Immunology) as first antibody. Cell fluorescence was assessed using a Leitz microscope. The staining was considered positive when 20% cells more than the control were stained. Cytological analysis In eleven cases (patients, 4, 5, 7, 8, 11, 12, 13, 14, 15, 18 and 20) cytocentrifuge smears stained with MayGriinwald-Giemsa were prepared from culture cells. The percentages of immature cells (myeloblasts, monoblasts, promyelocytes, myelocytes) and mature cells (metamyelocytes, segmented forms and mature monocytesmacrophages) were separately scored to calculate the maturation index (ratio mature/immature cells). Proliferation in liquid culture A total of 5 x 10~ cells were resuspended in I ml IMDM with 20% FCS, 2 txmol/ml L-glutamine, penicillin (100 U/ ml) and streptomycin (I00 ~tg/ml), with or without the addition of HGFs. Cells were cultured in 96 wells (200 Ixl/ well) round-bottom polystyrene microtiter plates (Greiner, Nfirtingen, F.R.G.) in a fully humidified atmosphere at 37°C for 72 h. Twenty hours before harvesting, 1 ~tCi tritiated thymidine (3H-TdR, 5 Ci/mmol, Amersham, U.K.) was added into each well. All experiments were performed in triplicate. Cells were harvested on nitrocellulose paper. The filters were dried, transferred to vials containing scintillation fluid (Toluene scintillator, United Technologies, Packard). Radioactivity was determined with a Betamatic Basic liquid scintillation counter (Kontron Analytical). Data are expressed as percentages of 3H-TdR incorporation by comparison to control without HGF. Statistical analysis Mean values were compared by the paired rank-sum test. RESULTS
Clonogenic assays Overall analysis of C F U - G M experiments shows that the presence of all factors studied significantly increased the colony n u m b e r as c o m p a r e d to the control (Table 1). The best results were obtained by
TABLE 1. MYELOIDCOLONY/CLUSTERENUMERATIONAFTERCULTUREIN SEMI-SOLIDMEDIUM Case No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Mean Normal marrowt
Growth factor FAB*
Control
RA RA RA RA RAS RAS RAEB RAEB RAEB RAEB RAEB RAEB RAEB RAEB RAEB RAEB-T RAEB-T RAEB-T RAEB-T CMML CMML CMML
0/0 0/0 0/0 0/9 0/0 36/167 0/29 37/>500 4/23 0/50 8/310 0/60 0/125 0/0 0/2 0/3 0/1 >500 0/0 2/48 2/18 16/21 5/43
5/6
HTB9 1/0 0/2 0/0 3/45 59/337 43/172 13/20 176/>500 10/218 50/262 154/>500 174/220 53/100 172/170 14/10 2/14 0/56 >500 0/0 58/452 11/38 22/46 48/133 73/71
Mean calculated after excluding case 18. * French-American-British classification [11]. t Mean of ten experiments. nd = not determined.
IL-3
GM-CSF
IL-3 + GM-CSF
G-CSF
G + GM-CSF
0/0 0/4 0/50 1/3 10/152 9/249 5/22 61/>500 24/48 5/400 90/>500 nd 60/98 22/136 2/8 0/4 0/6 >500 0/0 226/>500 2/24 10/40 26/117
0/(3 0/8 0/35 4/32 0/26 8/25 6/23 154/>500 10/70 3/430 4/350 10/42 94/•02 94/342 9/10 1/5 0/6 >500 1/6 44/407 1/8 1/40 21/110
1/1 0/2 0/54 2/31 26/154 93/50 19/29 183/>500 10/10 0/322 160/>500 nd 129/100 212/326 20/12 0/0 0/16 >500 1/2 36/450 13/52 0/35 27/142
1/1 0/8 0/9 4/46 nd 0/56 11/5 54/>500 2/20 12/450 16/>500 146/220 43/120 64/246 5/6 0/1 1/74 >500 4/5 60/152 4/18 0/21 21/123
0/0 0/2 nd 19/2 nd 24/31 10/2 97/>5 2/3 17/4 2/10 2/1 55/6 8/1 11/4 0/0 2/1 >50 2/1 138/5 3/1 0/1 20/1
65/84
95/97
O
O
B O e-,
32
X. THOMASet al.
the combination of IL-3 + GM-CSF and by HTB9 ( P < 10 -4 with control), which were also significantly more efficient than IL-3, GM-CSF, or G-CSF alone (P < 0.05). The combination of G- and GM-CSFs was not more efficient than each factor alone. There were wide individual variations in the number of colonies and in the responses to HGFs. An absence of colony formation (less than ten colonies) in controls and in the presence of any factor was observed in six cases. In one case, more than 500 colonies per 1.5 × 105 cells were recovered with and without growth factor. Three other cases spontaneously gave a low number of colonies. Although the best results were obtained with HTB9 and the combination of IL-3 and GM-CSF, cases responding to one factor usually responded to all factors studied. The highest stimulation was obtained in R A E B patients, whereas the cases which did not respond to any factor were R A (3/4 cases) and R A E B - T (3/4 cases). The ratio colony/cluster was studied in eleven cases giving more than ten clusters, and compared to that of normal marrow. The mean number of clusters was increased from 43 -+ 77 in controls without H G F to 133 + 155 with HTB9, 117 + 170 with IL-3, 109 + 151with GM-CSF, 142 + 187with IL-3 + GMCSF, 105 + 153 with G-CSF and to 93 ± 151 with G + GM-CSF. Although significant, the increase in the number of clusters was less than the increase in the number of colonies. Therefore, the colony to cluster ratio was very low in controls without HGFs (0.12 ± 0.05) and increased significantly in the presence of IL-3 (0.31-+0.15, P < 0 . 0 5 ) , GM-CSF (0.29 -+ 0.13, P < 0.05), IL-3 + GM-CSF (0.42 + 0.21, P
0145-2126/91 $3.00 + 0.00 PergamonPresspie
Leukemia Research
I N VITRO EFFECTS OF RECOMBINANT HEMOPOIETIC G R O W T H FACTORS ON P R O G E N I T O R CELLS FROM PATIENTS WITH MYELODYSPLASTIC SYNDROMES XAVIER THOMAS,* DENIS G U Y O T A T , t LYDIA CAMPOS,* ZHI HUA SHI,* LILIANA VILA,~ ALAIN EHRSAM* a n d DENIS FIERE~
*Service de Cryobiologie, Centre de Transfusion Sanguine, Lyon, France, i'D6partement d'H6matologie, H6pital Nord, St Priest en Jarez, France and ~tService d'H6matologie, H6pital Edouard Herriot, Lyon, France
(Received 23 February 1990. Revision accepted 17 July 1990) Abstract--The effects of four recombinant hemopoietic growth factors (HGF) and of the impure factor HTB9 on proliferation and maturation of marrow myeloid (CFU-GM) and erythroid (BFU-E) progenitor cells were studied in 22 cases of myelodysplastic syndromes (MDS). In most cases, IL-3, GM-CSF and G-CSF increased significantly the number of myeloid colonies, the best combination being 1L-3 + GM-CSF. A significant increase in the myeloid colony/cluster ratio was also noted, but cytological examination of colony cells showed little maturation. The analysis of myeloid colony surface markers with four monoclonal antibodies (to CD13, CD15, CD33 and CD34) showed minor modifications with an increase of CD13 and CD15 in about one third of cases when compared to control without HGF. Erythroid colonies were obtained in one case with erythropoietin alone, and in 19 cases with the addition of GM-CSF and/or IL-3. In short-term liquid cultures, IL-3, GM-CSF and G-CSF increased 3H-thymidine incorporation. We conclude that progenitor cells of most MDS are able to proliferate in the presence of HGF, with wide case-to-case variations. However, the pattern of growth remains abnormal when compared to normal marrow. Although the combination of IL-3 and GM-CSF is the most efficient, there is a large overlap in the stimulating effects of all factors studied.
Key words: Hemopoietic growth factors, IL-3, GM-CSF, myelodysplastic syndromes, progenitor cells.
erogeneous group of stem-cell diseases characterized by maturation defects with ineffective hemopoiesis and peripheral blood cytopenias [1]. The evolution to overt acute leukemia is frequently observed, particularly in refractory anemia with excess of blasts ( R A E B ) , or R A E B in transformation ( R A E B - T ) [1, 2]. The in vitro study of hemopoietic progenitors using conditioned medium as stimulating factor frequently shows an abnormal growth pattern [1-4]. Although abnormalities are not directly correlated with FAB sub-type, a growth of a low number of myeloid colonies with or without excess of clusters is frequently predictive of leukemic evolution [3, 4]. Hemopoietic growth factors ( H G F ) have recently been produced by recombinant D N A techniques and
their in vitro activities on proliferation and maturation of normal marrow cells have been reported [5-7]. Granulocyte-macrophage stimulating factor (GM-CSF) is able to support the growth of CFUGM in most MDS cases, whereas its effect on BFUE growth is inconstant [8, 9]. Interleukin 3 (IL-3), GM-CSF, granulocyte-CSF (G-CSF) and macrophage-CSF (M-CSF) alone were not able either to restore a normal growth pattern of erythroid, myeloid or mixed colonies in a series of ten cases of R A E B or R A E B - T [10]. In this study, we compared the effects of three recombinant H G F s (IL-3, GMCSF and G-CSF) on granulocyte-macrophage and erythroid colony growth in 22 patients with different types of MDS. In addition, we studied the cytological and immunological differentiation of myeloid colonies, and the proliferation of MDS cells in short-term liquid culture.
Correspondence to: Dr D. Guyotat, D6partement d'H6matologie, H6pital Nord, Avenue Albert-Raimond, 42277 St Priest en Jarez Cedex, France.
Patients Bone marrow was obtained from 22 patients with MDS
INTRODUCTION THE MYELODYSPLASTICsyndromes (MDS) are a het-
MATERIALS AND METHODS
29
30
X. THOMASet al.
(12 males and 10 females) by sternum or posterior iliac crest aspiration. The diagnosis was based on cell morphology and cytochemical staining according to the French--American-British (FAB) Cooperative Group criteria [11]. The FAB typing was: refractory anemia (RA): four cases; refractory anemia with ringed sideroblasts (RAS): two cases; refractory anemia with excess of blasts (RAEB): nine cases; RAEB in transformation (RAEB-T): four cases; chronic myelomonocytic leukemia (CMML): three cases. Two patients (7 and 13) had received prior therapy with methyl-prednisone and retinoic acid, respectively, and one (case I) had received chemotherapy for previous malignant disease. These patients had not been treated for at least 6 weeks at the time of study. Only four patients (patients 13, 14, 15 and 20) had required previous red-cell transfusions. Normal marrow samples used as controls were obtained from healthy donors for marrow transplantation. All patients gave informed consent to donate bone marrow.
Bone marrow cells Mononuclear cells were separated by centrifugation (433g, 30min) on Ficoll-Hypaque density gradient (Eurobio, Paris), washed twice in RPMI 1640 (Flow Laboratories, Irvine, U.K.) and finally resuspended in Iscove's modified Dulbecco's medium (IMDM). Individual cell fractions were then separated for cultures, thymidine incorporation, morphological (May-Griinwald-Giemsa staining) and surface marker analysis. Hemopoietic growth ]'actors Recombinant GM-CSF from E. coil expression system (biological activity of 5 × 107CFU/mg protein), and recombinant IL-3 and G-CSF from yeast expression system (biological activity of 1 x 107 CFU/mg protein) were generously provided by BehringWerke AG (Marburg, F.R.G.). Recombinant erythropoietin (Terry Fox Laboratories, Vancouver, Canada) was used at the concentration of 2 U/ml. Conditioned medium of 5637 human bladder carcinoma cell line (HTB9) was used as positive control. HTB9 is a rich source of interleukin-1, G-CSF and GM-CSF, and stimulates the proliferation of normal progenitors [12] and of leukemic [13] and MDS cells [8]. Myeloid progenitor (CFU-GM) assay A total of 15 × 104 cells were plated per well (24 Multidish Nunc, Ruskilde, Denmark) containing 0.5 ml IMDM supplemented with 20% fetal calf serum (FCS, GIBCO, Paisley, U.K.), 0.8% methylcellulose (Methocel MC, Fluka, Switzerland), and HGFs used alone or in combination. HGFs were used at a concentration of 50 U/ml. Negative and positive controls were performed without HGF and with 10% HTB9, respectively. Granulocyte-monocyte colony-forming units (CFUGM) were counted after 7-10 days incubation at 37°C in humidified atmosphere with 5% CO2. Colonies (more than 50 cells) and clusters (20-50 cells) were scored under an inverse microscope. Colony and cluster numbers represent means of triplicate cultures. After culture, individual colonies were collected using fine pipettes, pooled, washed in phosphate-buffered saline (PBS) and analyzed by morphology and immunology to evaluate the effects of HGFs on cell maturation in vitro.
Erythroid progenitor (BFU-E) assay Cells (0.5 × 105) were cultured in 1 ml IMDM containing 0.3% agar supplemented with 1% deionized bovine serum albumin (fraction V, Sigma), 20% FCS, 5 × 10-5 mercaptoethanol and erythropoietin (Epo) (2 U/ml). Erythroid colonies (>50 hemoglobinized cells) were counted after 14 days of culture at 37°C in 5% CO2 humidified atmosphere. Immunological study Surface markers of diagnosis cells and CFU-GM cells were analyzed by indirect immunofluorescence. The following monoclonal antibodies (MoAb) were used: MY10 (CD34), MY7 (CD13), MY9 (CD33) from Coulter Immunology (Hialeah, FL, U.S.A.); FMyl5a (CD15) (Biosys, Meudon). The cells washed twice in PBS were incubated with the MoAb at 4°C for 30 min. FITC-labelled sheep anti-mouse Fab~ fragments (Bioart, Meudon) were used as second layer (4°C, 30 min). Controls were performed using the MslgG and MslgM kits (Coulter Immunology) as first antibody. Cell fluorescence was assessed using a Leitz microscope. The staining was considered positive when 20% cells more than the control were stained. Cytological analysis In eleven cases (patients, 4, 5, 7, 8, 11, 12, 13, 14, 15, 18 and 20) cytocentrifuge smears stained with MayGriinwald-Giemsa were prepared from culture cells. The percentages of immature cells (myeloblasts, monoblasts, promyelocytes, myelocytes) and mature cells (metamyelocytes, segmented forms and mature monocytesmacrophages) were separately scored to calculate the maturation index (ratio mature/immature cells). Proliferation in liquid culture A total of 5 x 10~ cells were resuspended in I ml IMDM with 20% FCS, 2 txmol/ml L-glutamine, penicillin (100 U/ ml) and streptomycin (I00 ~tg/ml), with or without the addition of HGFs. Cells were cultured in 96 wells (200 Ixl/ well) round-bottom polystyrene microtiter plates (Greiner, Nfirtingen, F.R.G.) in a fully humidified atmosphere at 37°C for 72 h. Twenty hours before harvesting, 1 ~tCi tritiated thymidine (3H-TdR, 5 Ci/mmol, Amersham, U.K.) was added into each well. All experiments were performed in triplicate. Cells were harvested on nitrocellulose paper. The filters were dried, transferred to vials containing scintillation fluid (Toluene scintillator, United Technologies, Packard). Radioactivity was determined with a Betamatic Basic liquid scintillation counter (Kontron Analytical). Data are expressed as percentages of 3H-TdR incorporation by comparison to control without HGF. Statistical analysis Mean values were compared by the paired rank-sum test. RESULTS
Clonogenic assays Overall analysis of C F U - G M experiments shows that the presence of all factors studied significantly increased the colony n u m b e r as c o m p a r e d to the control (Table 1). The best results were obtained by
TABLE 1. MYELOIDCOLONY/CLUSTERENUMERATIONAFTERCULTUREIN SEMI-SOLIDMEDIUM Case No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Mean Normal marrowt
Growth factor FAB*
Control
RA RA RA RA RAS RAS RAEB RAEB RAEB RAEB RAEB RAEB RAEB RAEB RAEB RAEB-T RAEB-T RAEB-T RAEB-T CMML CMML CMML
0/0 0/0 0/0 0/9 0/0 36/167 0/29 37/>500 4/23 0/50 8/310 0/60 0/125 0/0 0/2 0/3 0/1 >500 0/0 2/48 2/18 16/21 5/43
5/6
HTB9 1/0 0/2 0/0 3/45 59/337 43/172 13/20 176/>500 10/218 50/262 154/>500 174/220 53/100 172/170 14/10 2/14 0/56 >500 0/0 58/452 11/38 22/46 48/133 73/71
Mean calculated after excluding case 18. * French-American-British classification [11]. t Mean of ten experiments. nd = not determined.
IL-3
GM-CSF
IL-3 + GM-CSF
G-CSF
G + GM-CSF
0/0 0/4 0/50 1/3 10/152 9/249 5/22 61/>500 24/48 5/400 90/>500 nd 60/98 22/136 2/8 0/4 0/6 >500 0/0 226/>500 2/24 10/40 26/117
0/(3 0/8 0/35 4/32 0/26 8/25 6/23 154/>500 10/70 3/430 4/350 10/42 94/•02 94/342 9/10 1/5 0/6 >500 1/6 44/407 1/8 1/40 21/110
1/1 0/2 0/54 2/31 26/154 93/50 19/29 183/>500 10/10 0/322 160/>500 nd 129/100 212/326 20/12 0/0 0/16 >500 1/2 36/450 13/52 0/35 27/142
1/1 0/8 0/9 4/46 nd 0/56 11/5 54/>500 2/20 12/450 16/>500 146/220 43/120 64/246 5/6 0/1 1/74 >500 4/5 60/152 4/18 0/21 21/123
0/0 0/2 nd 19/2 nd 24/31 10/2 97/>5 2/3 17/4 2/10 2/1 55/6 8/1 11/4 0/0 2/1 >50 2/1 138/5 3/1 0/1 20/1
65/84
95/97
O
O
B O e-,
32
X. THOMASet al.
the combination of IL-3 + GM-CSF and by HTB9 ( P < 10 -4 with control), which were also significantly more efficient than IL-3, GM-CSF, or G-CSF alone (P < 0.05). The combination of G- and GM-CSFs was not more efficient than each factor alone. There were wide individual variations in the number of colonies and in the responses to HGFs. An absence of colony formation (less than ten colonies) in controls and in the presence of any factor was observed in six cases. In one case, more than 500 colonies per 1.5 × 105 cells were recovered with and without growth factor. Three other cases spontaneously gave a low number of colonies. Although the best results were obtained with HTB9 and the combination of IL-3 and GM-CSF, cases responding to one factor usually responded to all factors studied. The highest stimulation was obtained in R A E B patients, whereas the cases which did not respond to any factor were R A (3/4 cases) and R A E B - T (3/4 cases). The ratio colony/cluster was studied in eleven cases giving more than ten clusters, and compared to that of normal marrow. The mean number of clusters was increased from 43 -+ 77 in controls without H G F to 133 + 155 with HTB9, 117 + 170 with IL-3, 109 + 151with GM-CSF, 142 + 187with IL-3 + GMCSF, 105 + 153 with G-CSF and to 93 ± 151 with G + GM-CSF. Although significant, the increase in the number of clusters was less than the increase in the number of colonies. Therefore, the colony to cluster ratio was very low in controls without HGFs (0.12 ± 0.05) and increased significantly in the presence of IL-3 (0.31-+0.15, P < 0 . 0 5 ) , GM-CSF (0.29 -+ 0.13, P < 0.05), IL-3 + GM-CSF (0.42 + 0.21, P
