Factors that induce proliferation of the human hematopoietic stem cell are ill-defined. Primitive hematopoietic progenitors can be maintained and differentiate in stroma-dependent, long-term bone marrow cultures (LTBMC), originally described by Dexter et al. (Dexter, T. M., L. H. Coutinho, E. Spooncer, C. M. Heyworth, C. P. Daniel, R. Schiro, J. Chang, and T. D. Allen. 1990. Molecular Control of Haemopoiesis). However, 70-80% of primitive progenitors capable of reinitiating secondary stromal cultures (LTBMC-initiating cells [IC]) are lost over a period of 5 wk in such cultures. We have recently described a novel "stroma-noncontact" culture system, in which hematopoietic progenitors are separated from the stromal layer by a 0.4-/~m microporous filter membrane. Primitive progenitors in such cultures can not only differentiate into committed progenitors, but are also maintained to a greater extent than in "Dexter" cultures. However, still only 50% of the originally seeded LTBMC-IC are recovered at week 5. Since maintenance of primitive progenitors may depend not only on growth-promoting factors but also on factors that inhibit differentiation and/or proliferation, we evaluated the effect of macrophage inflammatory protein 1 ol (MIP-lo 0 or "stem cell inhibitor'" in combination with the growth-inducing factor interleukin 3 (IL-3) on the recovery of LTBMC-IC from stroma-noncontact cultures. We demonstrate that addition of MIP-lot alone to stroma-noncontact cultures does not change the number of LTBMC-IC present after 8 wk, indicating that this factor may not directly inhibit or stimulate proliferation of primitive progenitors. Addition of the growth stimulatory cytokine, IL-3, alone results in exhaustion of LTBMC-IC after 8 wk of culture, possibly as a result of their terminal differentiation. However, LTBMC-IC can be maintained for at least 8 wk when grown in stroma-noncontact cultures supplemented with both MIP-lc~ plus IL-3. This effect depends on soluble (ill-defined) stromal factors, and results from a direct interaction of these cytokines with the progenitor population or its progeny, but not the stroma.

Stromal cells of the bone marrow can provide the growth-promoting and differentiation-inducing molecules which are necessary for haemopoiesis. While the nature of these stimuli is largely unknown, the development of haemopoietic cells in association with stromal cells requires intimate cell contact. Molecules of the extracellular matrix, such as heparan sulphate, are able to bind growth factors and in this way the stromal cells may form microenvironmental niches which preferentially promote development of multipotent and committed cells along discrete lineages. Cells from some patients with acute and chronic myeloid and lymphoid leukaemias are defective in their ability to interact with stromal cells and consequently cannot survive in stromal cell-mediated long-term marrow cultures. We have exploited this phenomenon to obtain normal haemopoietic cells from patients with leukaemia, and to use these cells for successful autografting in patients with acute and chronic myeloid leukaemias.

Haemopoietic colony stimulating factors (CSFs) are defined by their ability to stimulate the clonal expansion of bone marrow cells in semi-solid medium. Four main CSFs, interleukin-3 (IL-3), granulocyte macrophage-CSF (GM-CSF), macrophage-CSF (M-CSF) and granulocyte-CSF (G-CSF), have been described. Although all four CSFs are glycoproteins, they are otherwise unrelated. M-CSF is the only dimer, the others being monomers with disulphide bridges. Each factor binds to its own distinct receptor which, in the case of M-CSF, is similar to the product of the c-fms oncogene [1]. The responses mediated by these receptors in haemopoietic cells are survival, proliferation, lineage commitment and activation of end cell function.

Mono-ADP ribosylation is a post-transcriptional modification of proteins which can alter their biological properties. Particular substrates for this reaction are the GTP-binding proteins involved in the adenylate cyclase and phospholipase C second messenger pathways. Consequently, mono-ADP ribosylation may be an important element in intracellular signaling. Cholera toxin is a potent mono-ADP-ribosyl transferase, while benzylaminododecylguanine hydrochloride (BADGH) is an inhibitor of cholera toxin-induced ADP ribosylation. We have used these compounds to modulate the effects of inducers of differentiation in HL-60 cells. Cholera toxin, although unable to induce differentiation itself, synergized with inducers of both granulocytic and monocytic differentiation. In contrast, BADGH selectively inhibited the growth of undifferentiated cells. These effects imply regulatory roles for substrates for mono-ADP ribosylation both in the proliferation of undifferentiated cells and in the early stages of differentiation.