Six patients with recurrent and/or refractory acute nonlymphocytic leukemia (ANLL) were treated with recombinant human granulocyte colony-stimulating factor (G-CSF) and cytotoxic agents administered simultaneously. Neither of the two patients who received cytosine arabinoside (ara-C) in combination with G-CSF achieved complete remission. The other four patients, who received multi-drug combination therapy together with G-CSF, all achieved complete remission. No major side effects due to G-CSF were observed. These results demonstrated that the effects of G-CSF in enhancing the sensitivity of leukemic cells to cytotoxic agents and accelerating the recovery of leukocytes could lead to its possible use in the treatment of ANLL.

Two proteins forming the receptor for human granulocyte-macrophage colony-stimulating factor (GM-CSF)1 were identified and characterized. One with apparent Mr of about 80,000 was defined as alpha-chain and has Kd of 0.7-2.8 nM. The other binding molecule with apparent Mr of about 135,000 was defined as beta-chain and is related to the high-affinity binding with Kd of 10-40 pM. The binding kinetic studies confirmed that the 125I-GM-CSF associated slower to and dissociated more rapidly from the alpha-chain than the beta-chain. The alpha-chain is expressed not only on hemopoietic cells but also on full-term placental tissues, choriocarcinoma cells, and other solid tumor cells. In contrast, the distribution of the beta-chain is restricted on hemopoietic cells. The alpha-chain probably corresponds to the low-affinity GM-CSF receptor whose cDNA has been cloned and sequenced.

Receptors for granulocyte-macrophage colony-stimulating factor (GM-CSF) were identified on 9 of 35 (26%) human nonhematopoietic tumor cell lines including non-small cell lung cancer, stomach cancer, colon cancer, and osteosarcoma cells. GM-CSF receptors distributed on these human tumor cells were low affinity types with an equilibrium dissociation constant of 1.5-10.0 nM. Cross-linking studies revealed that the molecular weights of the low affinity GM-CSF receptors were 65-85 kilodaltons. The high affinity receptors identified on hematopoietic cells were not detected on human nonhematopoietic tumor cells which we studied, and we could detect no effects of GM-CSF on cell growth of these tumor cells.

Recombinant latent form of transforming growth factor-β1 (L-TGF-β1) is activated by various chemical treatments, including acidification and heating. However, cellular mechanisms that release transforming growth factor-β (TGF-β) in an active form have not been fully elucidated. Investigated herein are the effects of L-TGF-β1 on various leukemic cell lines. Heat-activated L-TGF-β1 inhibited colony formation of U937, KG-1 and HL-60, whereas untreated L-TGF-β1 had only a marginal effect on these cells. In contrast, colony formation of human erythroleukemia cell line (HEL) was markedly inhibited by both heat-activated and untreated L-TGF-β1. In vitro incubation of L-TGF-β1 with HEL cells did not release the active form in the culture supernatants. These results suggest that HEL cells are capable of activating L-TGF-β1, but only in a cell-associated manner. Since HEL cells produce L-TGF-β1, it may act as an autocrine negative growth factor on these cells.

Human granulocyte colony-stimulating factor (G-CSF) rapidly loses the biological activity and the receptor binding capacity following radioiodination. We have made a mutein of human G-CSF, KW-2228, in which Thr-1, Leu-3, Gly-4, Pro-5, and Cys-17 were respectively substituted with Ala, Thr, Tyr, Arg, and Ser; showed more potent G-CSF activity; and retained full biological activity and receptor binding capacity at least 2 weeks of radioiodination. G-CSF is an effective growth factor for the blasts of myeloid leukemia. Radioiodinated KW-2228 was prepared using solid-phase glucose oxidase-lactoperoxidase. Human leukemia cell lines and the blast cells from leukemia patients were examined for binding. High affinity binding sites were identified on myeloid cell lines and on the blasts obtained from acute myeloid leukemia patients. Scatchard analysis showed that a single binding site for G-CSF was observed (361-1688 receptors/cell; Kd 128-1400 pM). In contrast, specific binding of 125I-KW-2228 was not demonstrated on lymphoblastic cell lines or the blast cells of acute lymphoid leukemia or lymphoma. This difference was reflected in the effectiveness of G-CSF to stimulate colony formation in acute myeloid leukemia blasts, while G-CSF did not stimulate colony formation of the blast cells from acute lymphoid leukemia.

Harringtonine was discovered as an anticancer agent in China. It has been shown to be effective against myeloid leukemia. In this report, we have demonstrated that harringtonine inhibited the growth of human myeloid leukemia cells in vitro at low concentrations. Together with the clinical data in which 28% of the patients could be induced into complete remission only by harringtonine, this agent may be used as a first choice of antileukemia agents in the treatment of myeloid leukemia. The mechanism of the antitumor action of harringtonine is considered to be an effect on protein synthesis and is characterized by breakdown of polysomes to monosomes. The mechanism of action appears to be different from those of the other antileukemia agents, such as cyclophosphamide, daunorubicin, vincristine, or cytosine arabinoside. Harringtonine could be added to the other anti-leukemia agents used routinely in treatment of leukemia, and the combination of harringtonine with the other agents is expected to improve the therapy of myeloid leukemia.

We have established a novel cell line, designated as TF-1, from a patient with erythroleukemia, which showed complete growth dependency on granulocyte-macrophage colony-stimulating factor (GM-CSF) or on interleukin-3 (IL-3) and carried a homogeneous chromosomal abnormality (54X). Erythropoietin (EPO) also sustained the short-term growth of TF-1, but did not induce erythroid differentiation. These three hematopoietic growth factors acted on TF-1 synergistically. Transforming growth factor-beta and interferons inhibited the factor-dependent growth of TF-1 cells in a dose-dependent fashion, and monocyte-colony stimulating factor and interkeukin-1 enhanced the GM-CSF-dependent growth of TF-1. Ultrastructural studies revealed some very immature features in this cell line. Although TF-1 cells do not express glycophorin A or carbonyl anhydrase I, the morphological and cytochemical features, and the constitutive expression of globin genes, indicate the commitment of TF-1 to erythroid lineage. When induced to differentiate, TF-1 entered two different pathways. Specifically, hemin and delta-aminolevulinic acid induced hemoglobin synthesis, whereas TPA induced dramatic differentiation of TF-1 into macrophage-like cells. In summary, TF-1 is a cell line of immature erythroid origin that requires GM-CSF, IL-3, or EPO for its growth and that has the ability to undergo differentiation into either more mature erythroid cells or into macrophage-like cells. TF-1 is a useful tool for analyzing the human receptors for IL-3, GM-CSF, and EPO or the signal transduction of these hemopoietic growth factors.

Binding of radiolabeled human granulocyte-macrophage colony-stimulating factor (GM-CSF) was studied with blast cells from eight patients with acute myeloblastic leukemia (AML), and neoplastic lymphoid cells from one patient with acute lymphoblastic leukemia (ALL), two patients with chronic lymphocytic leukemia (CLL) and one patient with undiagnosed B cell neoplasia. In all AML cases studied, Scatchard graphs of the direct binding data were curvilinear, and were best fitted by curves derived from a two-binding-site model; one site with high affinity (Kd1 = 12-71 pM; 174-602 sites/cell) and the other with low affinity (Kd2 = 0.5-2.7 nM; 1137-6020 sites/cell). A cross-linking study on blast cells from one AML patient demonstrated specific bands which were similar to those reported for peripheral blood neutrophils. Furthermore, blast colony assays for the same preparations showed remarkable proliferative response to GM-CSF in the concentration range from 0.3 nM to 7.0 nM (ED50 greater than 0.7 nM). This concentration range is approximately one order of magnitude higher than that which is effective for colony formation from normal bone marrow progenitors (ED50 in equilibrium 0.1 nM). No significant correlation could be observed between the responsiveness of blast progenitors to GM-CSF, and the numbers or affinities of GM-CSF binding sites demonstrated on blast cells. In studies with neoplastic lymphoid cells from four patients, 125I-GM-CSF also specifically bound in two cases, while response to GM-CSF was not observed in these cases. These results indicate that the expression of GM-CSF receptor is not restricted to the GM-CSF-responsive AML blast cells, but can be observed in other AML blast cells and even in neoplastic lymphoid cells.