Leukemic Colony-forming Cells Research Articles

Malignant Progenitors From Patients With Acute Myelogenous Leukemia Are Sensitive to a Diphtheria Toxin–Granulocyte-Macrophage Colony-Stimulating Factor Fusion Protein

AbstractWe have previously demonstrated that human granulocyte-macrophage colony-stimulating factor (GM-CSF) fused to a truncated diphtheria toxin (DT388-GMCSF) kills acute myelogenous leukemia (AML) cell lines bearing the GM-CSF receptor. We now report that exposure of malignant cells from 50 different patients with AML for 48 hours in culture to DT388-GMCSF reduces by a median of 1.6 logs (range, 0 to 3.7 logs) the number of leukemic cells capable of forming colonies in semisolid media (leukemic colony-forming cells [CFU-L]) with a median IC50 of 3 × 10−12 mol/L (range, 5 to >4,000 × 10−12 mol/L). Furthermore, the cell kill is dependent on the presence of high-affinity GM-CSF receptors on leukemic blasts, because CFU-L from 27 of 28 AML samples expressing ≥35 GM-CSF receptors per cell were inhibited by the toxin, whereas the colony growth from all 4 leukemic samples (2 AML, 1 acute lymphoblastic leukemia [ALL], and 1 prolymphocytic leukemia [PLL]) that had less than 35 receptors per cell was unaffected by the drug. Sensitivity of CFU-L to DT388-GMCSF was seen regardless of the clinical responsiveness of the patient's leukemia to standard chemotherapy agents. In contrast, clonogenic cells from normal bone marrow formed colonies at near control numbers after exposure to much higher toxin concentrations (4 × 10−9 mol/L) than those required to kill CFU-L from most patients. Thus, leukemic progenitors isolated directly from the peripheral blood of most AML patients show the same sensitivity to DT388-GMCSF as previously demonstrated for AML cell lines. Under the same conditions of exposure, normal hematopoietic progenitors are relatively unaffected by DT388-GMCSF, suggesting its potential as a therapeutic agent in AML.

Malignant progenitors from patients with acute myelogenous leukemia are sensitive to a diphtheria toxin-granulocyte-macrophage colony-stimulating factor fusion protein.

We have previously demonstrated that human granulocyte-macrophage colony-stimulating factor (GM-CSF) fused to a truncated diphtheria toxin (DT388-GMCSF) kills acute myelogenous leukemia (AML) cell lines bearing the GM-CSF receptor. We now report that exposure of malignant cells from 50 different patients with AML for 48 hours in culture to DT388-GMCSF reduces by a median of 1.6 logs (range, 0 to 3.7 logs) the number of leukemic cells capable of forming colonies in semisolid media (leukemic colony-forming cells [CFU-L]) with a median IC50 of 3 x 10(-12) mol/L (range, 5 to >4,000 x 10(-12) mol/L). Furthermore, the cell kill is dependent on the presence of high-affinity GM-CSF receptors on leukemic blasts, because CFU-L from 27 of 28 AML samples expressing > or = 35 GM-CSF receptors per cell were inhibited by the toxin, whereas the colony growth from all 4 leukemic samples (2 AML, 1 acute lymphoblastic leukemia [ALL], and 1 prolymphocytic leukemia [PLL]) that had less than 35 receptors per cell was unaffected by the drug. Sensitivity of CFU-L to DT388-GMCSF was seen regardless of the clinical responsiveness of the patient's leukemia to standard chemotherapy agents. In contrast, clonogenic cells from normal bone marrow formed colonies at near control numbers after exposure to much higher toxin concentrations (4 x 10(-9) mol/L) than those required to kill CFU-L from most patients. Thus, leukemic progenitors isolated directly from the peripheral blood of most AML patients show the same sensitivity to DT388-GMCSF as previously demonstrated for AML cell lines. Under the same conditions of exposure, normal hematopoietic progenitors are relatively unaffected by DT388-GMCSF, suggesting its potential as a therapeutic agent in AML.

In vitro response of blasts to IL-3, GM-CSF, and G-CSF is different for individual AML patients: factors that stimulate leukemic clonogenic cells also enhance Ara-C cytotoxicity.

In vivo, growth factors are currently investigated for their capacity to trigger leukemic stem cells into cycle and thus overcome kinetic drug resistance. In this study, the susceptibility of leukemic clonogenic cells to individual growth factors was related to cytosine-arabinoside (Ara-C) sensitivity. The effects of interleukin-3 (IL-3), granulocyte-macrophage colony-stimulating factor (G-CSF), granulocyte colony-stimulating factor (G-CSF), and combinations of these recombinant hematopoietic factors were tested on blast cells of nine acute myeloid leukemia (AML) patients. Growth factor responses were assessed in semi-solid clonogenic assay and in a 10-day liquid culture followed by clonogenic assay. Heterogeneity in growth factor response was observed in both test systems, resulting in a variable pattern for individual leukemias. In the majority of cases (six of nine) the response patterns in the semi-solid and liquid cultures were divergent. To test the Ara-C sensitivity, leukemic blasts were exposed in liquid to various concentrations of Ara-C in the absence and presence of preselected growth factors. After 10 days, the number of surviving leukemic colony-forming cells (CFU-L) was assessed. Exposure to Ara-C in the presence of optimal stimulatory factor(s) resulted in a 3- to 1000-fold increase of the Ara-C toxicity in seven patients. The Ara-C concentrations resulting in 50% inhibition of clonogenicity (ID50) were 0.48-123 x 10(-8) M Ara-C in the absence of stimulatory growth factors, versus only 0.12-0.40 x 10(-8) M Ara-C in the presence of these factors. In two patients, addition of one or more factors neither increased the number of CFU-L in liquid nor enhanced the Ara-C toxicity. Even in the absence of growth factors the ID50 values in these cases were as low as 0.20 and 0.28 x 10(-8) M Ara-C and in the same range as the ID50 values observed with maximum growth factor stimulation in the other seven patients. These results indicate that Ara-C cytotoxicity can be enhanced by individually selected, clonogenic cell growth-promoting hematopoietic factors.

Suppression of Juvenile Chronic Myelogenous Leukemia Colony Growth by Interleukin-1 Receptor Antagonist

Bone marrow (BM) and peripheral blood (PB) cells from patients with juvenile chronic myelogenous leukemia (JCML) exhibit spontaneous in vitro proliferation. Several cytokines including granulocyte-macrophage colony-stimulating factor (GM-CSF), interleukin-1 beta (IL-1 beta), and tumor necrosis factor alpha (TNF alpha) have been implicated in supporting the growth of leukemic monocyte-macrophage colonies either by autocrine or paracrine pathways. In seven untreated JCML patients, we investigated the role of IL-1 in the spontaneous growth of these cells by specifically blocking IL-1 receptors. The IL-1 receptor antagonist (IL-1 Ra) was added to the clonogenic assays, and in each case significant (mean = 63%, range = 35% to 82%) inhibition of spontaneous proliferation was observed. Uncultured circulating cells from PB or BM of four out of five patients expressed IL-1 beta-specific mRNA and secreted the protein into the culture supernatants. Moreover, by means of reverse transcriptase-polymerase chain reaction (RT-PCR), we demonstrated that most of the spontaneously growing leukemic colony-forming unit cells (CFU-C) obtained from BM cells of two patients were positive for the presence of the IL-1 beta-specific mRNA. Despite the presence of a measurable amount of GM-CSF in JCML cell culture supernatants, GM-CSF-specific mRNA in CFU-C cells of four cases was not detected by RT-PCR. These data further support a central role for IL-1 beta in the pathogenesis of JCML and suggest that the use of IL-1 Ra could represent a novel therapeutic strategy against this disorder.

Modulation of Cytotoxicity and Differentiation-Inducing Potential of Arabinofuranosylcytosine in Myeloid Leukemia Cells by Hematopoietic Cytokines

Hematopoietic growth factors may be useful in improving the clinical effectiveness of arabinofuranosylcytosine (ara-C). In vitro studies have indicated that interleukin 3(IL-3) and, to a lesser extent, granulocyte-macrophage colony-stimulating factor (GM-CSF), but not G-CSF or M-CSF, may be capable of specifically augmenting the ability of ara-C to kill leukemic myeloid cells by pharmacological and cytokinetic mechanisms including increase of intracellular ara-CTP/dCTP pool ratios and enhanced ara-C DNA incorporation in leukemic blast cells, decrease of IC 90 of ara-C for leukemic colony-forming cells (CFC) as compared with normal CFC growth, and recruitment of quiescent leukemic cells into the cell cycle. In contrast, the combination of ara-C with M-CSF or with the leukemia inhibitory factor (LIF) appears to be useful in overcoming the block in differentiation of leukemic blast, while the effects of GM-CSF and IL-3 on ara-C-induced differentiation appear limited. The combined treatment of human myeloid leukemia cells by ara-C and LIF is associated with down-regulation of c-myc gene expression, transcriptional activation of jun/fos gene expression, and features of functional differentiation (e.g., the capability to reduce nitroblue tetrazolium, to express lysozyme, or to display differentiation-related surface receptors including C3bi and the c-fms protein). On the basis of these in vitro studies first clinical trials are underway that are examining the efficacy of ara-C combinations with these molecules for the treatment of myeloid disorders.

Growth modulation of human leukemic, preleukemic, and myeloma progenitor cells by L-ascorbic acid

L-Ascorbic acid (LAA) was shown to modulate the in vitro growth of colonies of human and mouse myeloma progenitor-stem cells through use of a unique cell-culture assay. LAA was also shown to modulate the in vitro growth of leukemic colony-forming cells (L-CFC) from bone marrow of patients with acute myelocytic leukemia. LAA enhanced the growth of L-CFC in 35% of patients and suppressed the growth of L-CFC in 15% of patients. The minimum effective concentration was 0.03 mmol/L. The modulating effect is specific to LAA because other redox compounds are without effect. From the cell kinetic standpoint, the LAA effect is cytostatic rather than cytocidal. Similar LAA effects have prognostic value in patients with myelodysplastic syndromes (MDS), with LAA-sensitive patients displaying shorter survival than LAA-insensitive patients. MDS appears to be the ideal disease for clinical trials involving in vivo LAA manipulation to control the disease process.

Participation of the cytokines interleukin 6, tumor necrosis factor-alpha, and interleukin 1-beta secreted by acute myelogenous leukemia blasts in autocrine and paracrine leukemia growth control.

Autonomous in vitro growth of myeloid leukemic colony-forming cells may in part result from autocrine production of colony-stimulating factors (CSF). Some acute myeloid leukemia (AML) samples, however, fail to synthesize CSF despite growing autonomously in agar, and are therefore believed to bypass CSF requirements. Cytokines such as IL-6, tumor necrosis factor (TNF)-alpha, and IL-1, products of cells of the myeloid lineage, are known to be involved in growth control of myeloid progenitor cells. Since these molecules may also contribute to autocrine and paracrine growth regulation of myeloid leukemias, we screened a series of AML for cytokine production. In addition, possible roles of IL-6, TNF-alpha, and IL-1 in growth control of AML were investigated in vitro. We show that a substantial proportion of AML cells produce IL-6, TNF-alpha, and IL-1-beta and use these mediators to stimulate their growth by disparate mechanisms: IL-6 acts as a costimulator to enhance CSF-induced clonogenicity of AML blasts. TNF-alpha induces CSF production by endothelial cells and may therefore provide a paracrine loop to support leukemia growth.

Prolonged survival of mice with myeloid leukemia by subcutaneous injection of recombinant human G-CSF

We studied the effect of recombinant human granulocyte colony-stimulating factor (rhG-CSF) on leukemia development and survival of mice with leukemia by using a radiation-induced myeloid leukemia cell line (C2M) with A-type phosphoglycerate kinase (PGK) as marker isoenzyme. C3H/He mice with B-type PGK were inoculated with 2 × 10 6 C2M cells through the tail vein. From the next day, they received a daily subcutaneous injection of 1 μg rhG-CSF or control solution. The survival of rhG-CSF-treated recipients of C2M was significantly longer than control-solution-treated recipients. In rhG-CSF-treated recipients, not only spleen weight but also the number of blasts in hemopoietic organs was less than those in control recipients. While there was a remarkable decrease in bone marrow content of CFU-GM in control recipients, the content in rhG-CSF-treated recipients was comparable to that in normal mice. A reduced bone marrow and spleen content of leukemic colony-forming cells (L-CFU) was observed in rhG-CSF-treated recipients in comparison with control recipients. The electrophoretic analysis of PGK phenotypes of hemopoietic organs indicated the delayed appearance of A-type PGK from C2M cells in rhG-CSF-treated recipients. From these findings, we concluded that the injection of rhG-CSF delayed the onset of leukemia and improved the survival and several hematological parameters by suppressing C2M cells in recipient mice.

Leukemic colony-forming cells in acute myeloblastic leukemia: maturation hierarchy and growth conditions.

Despite their primitive morphological appearance, the majority of leukemic blasts in acute myeloblastic leukemia (AML) are end-stage, nonproliferating cells. Only a small subset of AML blasts are capable of a sufficient number of divisions to form colonies in semisolid medium [1, 2]. It has been suggested that these leukemic colony-forming cells (L-CFC) may act in vivo as progenitor cells to maintain the rest of the leukemic cell population [3, 4]. L-CFC share several properties with normal myeloid progenitor cells, including self-renewal potential and high thymidine suicide index [2, 3]. As in the case of normal myeloid progenitor cells (NMPC), colony growth of L-CFC from most patients requires exogenous colony-stimulating factors (CSF) which are routinely supplied by the addition of media conditioned by activated T cells, placental tissue, or media derived from various tumor cell lines, including GCT [5], MO [6], or 5637 [7].KeywordsLeukemic BlastAcute Myeloblastic LeukemiaQuadruplicate CultureLeukemic Cell PopulationAgar UnderlayerThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Effects of Recombinant Human GM-CSF on Proliferation of Clonogenic Cells in Acute Myeloblastic Leukemia

Proliferation of acute myeloblastic leukemia (AML) cells in vitro is limited in most cases to a small subset of blasts that have several properties of stem cells. These leukemic colony-forming cells (AML-CFU) generally require addition of exogenous growth factors for proliferation in agar or methylcellulose. These factors can be supplied by media conditioned by phytohemagglutinin-stimulated normal leukocytes or by CSF-secreting tumor cell lines. However, the exact factor or factors required for stimulation of AML-CFU growth have not been defined. We compared the AML-CFU stimulatory activity of a human recombinant GM-CSF with that of GCT-CM, Mo-CM, and the PHA-leukocyte feeder system in 15 cases of AML. In each of the 12 cases that required exogenous growth factors for maximum AML-CFU growth, recombinant GM-CSF could replace either GM-CSF or Mo-CM, and could partially replace the PHA-leukocyte feeder system. These results indicate that this GM-CSF is a growth promoter of AML-CFU in these culture systems.

Effects of recombinant human GM-CSF on proliferation of clonogenic cells in acute myeloblastic leukemia

Proliferation of acute myeloblastic leukemia (AML) cells in vitro is limited in most cases to a small subset of blasts that have several properties of stem cells. These leukemic colony-forming cells (AML-CFU) generally require addition of exogenous growth factors for proliferation in agar or methylcellulose. These factors can be supplied by media conditioned by phytohemagglutinin-stimulated normal leukocytes or by CSF-secreting tumor cell lines. However, the exact factor or factors required for stimulation of AML-CFU growth have not been defined. We compared the AML-CFU stimulatory activity of a human recombinant GM-CSF with that of GCT-CM, Mo-CM, and the PHA-leukocyte feeder system in 15 cases of AML. In each of the 12 cases that required exogenous growth factors for maximum AML-CFU growth, recombinant GM-CSF could replace either GM-CSF or Mo-CM, and could partially replace the PHA-leukocyte feeder system. These results indicate that this GM-CSF is a growth promoter of AML-CFU in these culture systems.

On the Maturation Order of AML Cells: A Distinction on the Basis of Self-Renewal Properties and Immunologic Phenotypes

Abstract To investigate the heterogeneous cellular structure of human acute myeloid leukemia (AML), subpopulations of cells were distinguished by two combined criteria: proliferation and differentiation. Purified blast cells were fractionated from blood or bone marrow of patients with newly diagnosed AML and colonies and clusters grown in phytohemagglutinin (PHA)-leukocyte feeder cultures. Large colonies, small colonies, macroclusters, and microclusters were recloned separately to assess the replicative capacities as a function of clone size. Large colonies showed higher proliferative capacities than did small ones, etc. Anti-Ia and an antigranulocyte (B4.3) monoclonal antibody (MoAb) were then employed to evaluate the stage of differentiation of AML cells in two patients before and following colony culture. Alterations of the immunologic phenotypes appeared during colony formation. This suggested differentiation of cells to more mature B4.3 granulocyte antigen-positive stages. MoAb-dependent cell lysis with the two antibodies was subsequently performed to assess the phenotypes of the precursors of the colonies and clusters. Leukemic colony-forming cells were Ia-positive and B4.3-negative and different from cluster-forming cells, which were largely Ia-negative and B4.3- negative. These data suggest that the cell organization of AML fits a maturation scheme containing immature cells with relatively high proliferative capacities, intermediate cells with low proliferative capacities, and end cells that are nonreplicative, and each with specific phenotypes.

On the maturation order of AML cells: a distinction on the basis of self-renewal properties and immunologic phenotypes

To investigate the heterogeneous cellular structure of human acute myeloid leukemia (AML), subpopulations of cells were distinguished by two combined criteria: proliferation and differentiation. Purified blast cells were fractionated from blood or bone marrow of patients with newly diagnosed AML and colonies and clusters grown in phytohemagglutinin (PHA)-leukocyte feeder cultures. Large colonies, small colonies, macroclusters, and microclusters were recloned separately to assess the replicative capacities as a function of clone size. Large colonies showed higher proliferative capacities than did small ones, etc. Anti-Ia and an antigranulocyte (B4.3) monoclonal antibody (MoAb) were then employed to evaluate the stage of differentiation of AML cells in two patients before and following colony culture. Alterations of the immunologic phenotypes appeared during colony formation. This suggested differentiation of cells to more mature B4.3 granulocyte antigen-positive stages. MoAb-dependent cell lysis with the two antibodies was subsequently performed to assess the phenotypes of the precursors of the colonies and clusters. Leukemic colony-forming cells were Ia-positive and B4.3-negative and different from cluster-forming cells, which were largely Ia-negative and B4.3-negative. These data suggest that the cell organization of AML fits a maturation scheme containing immature cells with relatively high proliferative capacities, intermediate cells with low proliferative capacities, and end cells that are nonreplicative, and each with specific phenotypes.

A monoclonal antibody reactive with normal and leukemic human myeloid progenitor cells

Anti-MY9 is an IgG2b murine monoclonal antibody selected for reactivity with immature normal human myeloid cells. The MY9 antigen is expressed by blasts, promyelocytes and myelocytes in the bone marrow, and by monocytes in the peripheral blood. Erythrocytes, lymphocytes and platelets are MY9 negative. All myeloid colony-forming cells (CFU-GM), a fraction of erythroid burst-forming cells (BFU-E) and multipotent progenitors (CFU-GEMM) are MY9 positive. This antigen is further expressed by the leukemic cells of a majority of patients with AML and myeloid CML-BC. Leukemic stem cells (leukemic colony-forming cells, L-CFC) from most patients tested were also MY9 positive. In contrast, MY9 was not detected on lymphocytic leukemias. Anti-MY9 may be a valuable reagent for the purification of hematopoietic colony-forming cells and for the diagnosis of myeloid-lineage leukemias.

Use of surface markers to identify a subset of acute myelomonocytic leukemia cells with progenitor cell properties

The expression of differentiation-associated surface antigens was used to identify subsets of human acute myelomonocytic leukemia cells. The leukemic colony-forming cells expressed antigens characteristic of very immature myeloid cells (Ia, MY7), while the majority of leukemic cells also expressed an antigen (MY4) characteristic of later myeloid cells. The highly proliferative MY4 negative colony-forming cells may serve as progenitor cells for the less proliferative MY4 + leukemic cells.

Use of Surface Markers to Identify a Subset of Acute Myelomonocytic Leukemia Cells With Progenitor Cell Properties

The expression of differentiation-associated surface antigens was used to identify subsets of human acute myelomonocytic leukemia cells. The leukemic colony-forming cells expressed antigens characteristic of very immature myeloid cells (Ia, MY7), while the majority of leukemic cells also expressed an antigen (MY4) characteristic of later myeloid cells. The highly proliferative MY4 negative colony-forming cells may serve as progenitor cells for the less proliferative MY4 + leukemic cells.

Acute myeloid leukemia colony growth in vitro: differences of colony- forming cells in PHA-supplemented and standard leukocyte feeder cultures

Bone marrow or blood of patients with acute myeloid leukemia was subjected to cell separation and the cells investigated for in vitro colony growth. Discontinuous albumin density gradient centrifugation and depletion of E-rosette-forming cells resulted in purified fractions of acute myeloid leukemia cells. From these fractions, growth of large leukemic colonies was obtained in the PHA-leukocyte feeder (PHA-LF) colony technique in 12 of 14 patients. The standard double agar layer techniques with a leukocyte feeder for granulocyte-macrophage colony forming cells (GM-CFC) supported colony formation in only four cases. The PHA-LF leukemic colony-forming cells (CFC) were found to be of low buoyant density (always less than or equal to 1.062 g.ml-1) when compared to normal marrow GM-CFC (peak at 1.065 g.ml-1). The density profile of PHA-LF CFC paralleled the distribution of the nucleated cells in 8 cases, but in 4 patients, the cFC peak was found at a distinctly lower density; this suggested that a specific leukemic subpopulation had a colony-forming capacity. In three of the four patients with colony growth in the double layer agar technique, it was evident that these CFC had density properties different from those of PHA-LF CFC. These findings suggest that cells giving rise to large colonies in the PHA-LF and double layer agar assays represent distinct leukemic subpopulations.

Acute Myeloid Leukemia Colony Growth In Vitro: Differences of Colony-Forming Cells in PHA-Supplemented and Standard Leukocyte Feeder Cultures

Bone marrow or blood of patients with acute myeloid leukemia was subjected to cell separation and the cells investigated for in vitro colony growth. Discontinuous albumin density gradient centrifugation and depletion of E-rosette-forming cells resulted in purified fractions of acute myeloid leukemia cells. From these fractions, growth of large leukemic colonies was obtained in the PHA-leukocyte feeder (PHA-LF) colony technique in 12 of 14 patients. The standard double agar layer techniques with a leukocyte feeder for granulocyte-macrophage colony forming cells (GM-CFC) supported colony formation in only four cases. The PHA-LF leukemic colony-forming cells (CFC) were found to be of low buoyant density (always less than or equal to 1.062 g.ml-1) when compared to normal marrow GM-CFC (peak at 1.065 g.ml-1). The density profile of PHA-LF CFC paralleled the distribution of the nucleated cells in 8 cases, but in 4 patients, the cFC peak was found at a distinctly lower density; this suggested that a specific leukemic subpopulation had a colony-forming capacity. In three of the four patients with colony growth in the double layer agar technique, it was evident that these CFC had density properties different from those of PHA-LF CFC. These findings suggest that cells giving rise to large colonies in the PHA-LF and double layer agar assays represent distinct leukemic subpopulations.

Studies on chronic myeloid leukemia cell populations with colony-forming abilities in PHA-leukocyte feeder and robinson assays

Investigation of leukemic colony-forming cells (CFC) in PHA-supplemented cultures requires removal of T lymphocyte precursors prior to culture. Using a method of discontinuous density gradient centrifugation with concurrent depletion of E-rosette forming cells, T lymphocytes were effectively separated from light density CML bone marrow and blood cell fractions. Consequently, in light density fractions (1.056 and 1.059 g/ml) pure leukemic colony growth was obtained in the PHA-leukocyte feeder (PHA-l.f.) assay. Fraction 1.062 g/ml also yielded pure leukemic colonies in most experiments. Comparison of the density distributions of leukemic PHA-l.f. CFC and Robinson CFC revealed that both CFC populations had congruent density profiles in most patients. In others PHA-l.f. CFC were found to be of somewhat higher density than Robinson CFC. The most striking divergence was apparent in a patient in blast crisis. The findings suggest that different subsets of precursor cells within the CML population proliferate in PHA-l.f. and Robinson colony methods. Both colony techniques are thus potentially useful for discriminating subpopulations of colony-forming cells in chronic myeloid leukemia.

In vitro sensitivity of CFU-L to cytosine arabinoside (Ara-C).

The in vitro sensitivity of leukemic colony-forming cells to Cytosine arabinoside (Ara-C) was assessed in acute myeloid leukemia, by means of a clonogenic assay using methyl-cellulose with phyto-hemagglutinin (PHA)-leukocyte conditioned medium. Bone marrow clonogenic cells were more sensitive to Ara-C than blood clonogenic cells. Also a continuous exposure to Ara-C induced a higher reduction of the leukemic colonies than one hour exposure. Patients were, according to the in vitro results, divided into 3 groups: eight patients had a dose-responsive curve for all the tested drug concentrations and were characterized by a marked sensitivity to the induction treatment combining Ara-C and Adriamycin. Three patients had a = 50% reduction of colonies with the 3 concentrations; two achieved complete remissions. Seven patients were in vitro insensitive to the drug; four of them had a relative or absolute resistance to the induction treatment.