We have isolated a subline of the M-07 human megakaryoblastic leukemia cell line, designated M-07e, that requires either interleukin-3 (IL-3) or granulocyte macrophage colony-stimulating factor (GM-CSF) for growth, even in the presence of fetal calf serum. This cell line will not grow long term in any other cytokine although it responds slightly to IL-2, IL-4, IL-6, IL-9, and interferon-gamma. We have used the M-07e subline to develop a quantitative bioassay for the measurement of levels of either GM-CSF or IL-3. This assay is as sensitive to either factor as the human bone marrow colony assay (CFU-GM) or the chronic myelogeneous leukemic (CML) blast cell proliferation assay for these factors and is much more convenient and reliable than either. With this assay, as little as 25-50 pg/ml of either IL-3 or GM-CSF can be detected, a level that should render the assay useful for analysis of these molecules in samples from patients undergoing colony-stimulating factor therapy and from conditioned media from natural sources of the factors. In these cases, neutralizing antisera to each cytokine are required to demonstrate the specificity of the assay. This assay, in combination with quantitative immunoassays, should greatly facilitate the analysis of the roles of IL-3 and GM-CSF in regulating hematopoiesis both in patients and in natural sources of the cytokines.

Flt4 is a receptor protein tyrosine kinase that is expressed in the adult lymphatic endothelium and high endothelial venules. We have used a BIAcore assay to identify rodent and human cell conditioned media containing the ligand of Flt4 (Flt4-L). Receptor-based anity chromatography was used to purify this growth factor, followed by amino acid sequencing and molecular cloning of the murine cDNA, the orthologue of human vascular endothelial growth factor-C and vascular endothelial growth factor related protein. The murinē t4-L gene was localized to chromosome 8 and demonstrated to be widely expressed. Flt4-L was found to have a hydrophobic signal sequence and a pro-peptidelike sequence that is removed to generate the mature Nterminus. In addition, the C-terminal region of Flt4-L has four repeats of a cysteine-rich motif that is presumably also proteolytically processed to generate the 21 000 M r polypeptide subunit of the Flt4-L homodimer. Recombinant Flt4-L activated Flt4 as judged by induction of tyrosyl phosphorylation, and induced mitogenesis in vitro of lymphatic endothelial cells.Keywords: Flt4; Flt4 ligand; VEGF-C; BIAcoreTo identify a source of the Flt4 ligand, we used a Flt4-Fc fusion protein. This consists of the extracellular region of murine Flt4 (Finnerty et al., 1993) fused to the hinge-C H 2-C H 3 domains of human immunoglobulin g1. The conditioned media (CM) of over 100 dierent cell lines were screened for binding to Flt4-Fc using a BIAcore 2000 instrument. The media conditioned by the murine embryonic liver cell line, BNL 1NG A.2, human bladder carcinoma 5637, and bualo rat BRL 3A each showed binding to Flt4-Fc, but not human IgG1 (Figure 1). The binding was speci®c because when the CM was coinjected with excess Flt4-Fc, but not human IgG1 or KIT-Fc, the signal was inhibited (Figure 1). The presence of Flt4 ligand (Flt4-L) in BNL 1NG A.2 CM was con®rmed by demonstrating that it activated the Flt4 receptor protein tyrosine kinase. CHO cells stably transfected with a murine Flt4 cDNA expression vector were treated with 50-fold concentrated BNL 1NG A.2 CM for 10 min at 378C. This resulted in a marked increase in tyrosyl phosphorylation of Flt4 (Figure 2a, lane 1) relative to mock stimulated samples (Figure 2a, lane 2).Six micrograms of Flt4-L were puri®ed from 15L of BNL 1NG A.2 CM using Flt4-Fc anity chromatography. Flt4-Fc (100 mg/ml) or KIT-Fc (200 mg/ml) were directly immobilized to 1 mL Pharmacia HiTrap NHS columns. A volume of 50 ml of concentrated BNL 1NG A.2 CM was loaded ®rst onto the KIT-Fc column to adsorb nonspeci®c binding, then onto the Flt4-Fc column. The columns were washed and separately eluted with 24 mM hydrochloric acid. As expected, only fractions from the Flt4-Fc anity column had binding to Flt4-Fc on the BIAcore assay. The speci®c activity in response units (RU)/mg was determined using RU that were in the linear range of RU versus concentration of Flt4-L puri®cation fractions. The speci®c activity of Flt4-L from BNL 1NG A.2 was 4.2610 6 RU/mg, an increase of 98 455-fold after t...

We have investigated the mechanism of action of the thrombopoietic cytokine, recombinant human interleukin-11 (rhIL-11), on megakaryocytopoiesis in vitro. We have shown that rhIL-11–induced murine and human megakaryocytopoiesis are not mediated by thrombopoietin (Tpo). Murine megakaryocytes (MKs) were produced from bone marrow (BM) mononuclear cells cultured with rhIL-11, IL-3, and a combination of the two cytokines. Conditioned media (CM) were collected and assayed for the presence of biologically active Tpo. Tpo activity was not detected in any of the CMs tested. Next, human BM CD34+ cells were cultured in serum-free fibrin clot medium with rhIL-11, IL-3, or rhIL-11 plus IL-3 and an antibody that neutralizes human Tpo activity. No inhibition of either burst-forming unit-MK– or colony-forming unit-MK–derived colony formation was observed. The antibody did partially inhibit steel factor-induced MK-colony formation, suggesting that the actions of this cytokine are mediated, in part, by Tpo. We determined that MKs can be direct targets of rhIL-11 by showing the expression of functional IL-11 receptor on these cells. Total RNA was prepared from cultured human BM CD41+CD14− cells (MKs) and IL-11 receptor α chain mRNA was detected in the MKs by reverse transcription-polymerase chain reaction. Analysis of single-sorted CD41+CD14− cells confirmed that the observed IL-11 receptor expression was not due to contaminating CD41− cells in the pool. The presence of rhIL-11 receptor α chain protein in the cells was established by Western blot analysis. After a short exposure of purified BM MKs to rhIL-11, enhanced phosphorylation of both its signal transduction subunit, gp130, and the transcription factor, STAT3 was detected, showing a direct activation of receptor signaling by the cytokine. Consistent with the lack of effect of rhIL-11 on platelets in vivo, IL-11 receptor α chain mRNA and protein were not detected in isolated human platelets. These data indicate that rhIL-11 acts directly on MKs and MK progenitors but not on platelets.

Human rIL-7 was studied for its effects on myeloid and erythroid progenitors from human bone marrow cells. IL-7 did not support the granulocytic/monocytic or erythroid lineage but exclusively stimulated eosinophil colony formation (CFU-Eo) (4 +/- 3 vs 48 +/- 17 CFU-Eo/10(5) nonadherent fraction-non-T cell (NAF-NT) cells). This supportive effect was not mediated by T cells or monocytes because similar results were obtained with or without T cell or adherent depleted cell fractions. In addition, it was shown that CD34+ sorted cells could be stimulated by IL-7 (0 vs 15 +/- 9 CFU-Eo/3 x 10(3) CD34+ cells) Furthermore studies with IL-3 or granulocyte-macrophage CSF (GM-CSF) demonstrated an additive effect on the IL-7 supported colony formation. Finally, experiments were performed with anti-IL-3, anti-GM-CSF, anti-IL-1, and anti-IL-5 to exclude the possibility that IL-7 indirectly stimulated the eosinophil progenitor cell. Anti-GM-CSF, anti-IL-1, or anti-IL-3 did not influence the supportive effects of IL-7. However, anti-IL-5 did abolish the effects of IL-7 on the eosinophil colony formation (69 +/- 15 vs 3 +/- 2 CFU-Eo/10(5) NAF-NT, n = 3). Similar results were obtained with CD34+ sorted cells. Moreover, IL-5 mRNA expression could be demonstrated in IL-7-stimulated NAF-NT cells. These data suggest that the supportive effects of IL-7 on eosinophil precursors are mediated by the endogenous release of IL-5.