Differentiation-linked secretion of urokinase and tissue plasminogen activator by normal human hemopoietic cells.

Previous studies have shown that the response of patients with acute myeloid leukemia to induction chemotherapy can be predicted by the species of plasminogen activator that their cells secrete. Patients whose cells secreted tissue plasminogen activator (tPA) only failed to respond to combination chemotherapy. Individuals whose leukemic cells display features of the early progenitor phenotype also respond poorly to therapy. This suggested that the two species of plasminogen activator secreted by leukemic cells might be produced by normal cells at distinct stages of differentiation. These results indicate that the secretion of the two enzyme types is a differentiation-linked property of normal cells with tPA being produced by granulocyte/macrophage progenitors and urokinase by more differentiated cells and by mature neutrophils and macrophages.

Retinoic acid and phorbol ester induced hyperphosphorylation of topoisomerase II-alpha is an early event in HL-60 human leukaemia cell differentiation: effect on topoisomerase activity and etoposide sensitivity.

Treatment of HL-60 with phorbol myristate acetate (PMA) for 30 min, or all-trans retinoic acid (RA) for 60 min, results in hyperphosphorylation (3-5x) of topoisomerase II (p170, topo II) in vivo. RA and PMA activate a coprecipitating kinase, respectively inducing 1.6 and 2.7-fold increases in phosphorylation of topo II in immunoprecipitates. The activity of the co-precipitating kinase is inhibited by heparin and unlabelled GTP suggesting that casein kinase II (CKII) is, at least in part, responsible for the topo II hyperphosphorylation in response to differentiation signals. Although following dephosphorylation of the enzyme with alkaline phosphatase there was virtual abrogation of activity, the differentiation associated hyperphosphorylation had little impact on the decatenation activity of topo II in nuclear extracts. There were, however detectable changes in topo II function in vivo which affected the formation of the etoposide stabilised cleavable complex, but only after PMA treatment. PMA resulted in a rapid reduction in etoposide induced cleavage, 30 min treatment with PMA reducing cleavage by 20%. However, treatment with RA for 1 or 2 h when hyperphosphorylation was maximal did not affect cleavage. Immunoband depletion assays suggested that differentiation associated changes in chromatin structure rather than alterations in the enzyme per se are responsible for the reduction in cleavable complex formation following PMA treatment. Etoposide cytotoxicity was significantly reduced following just 30 min PMA treatment, but not reduced and even possibly enhanced by retinoic acid treatment. These findings are relevant not only to the dissection of the role of topo II in differentiation but also to its exploitation as a therapeutic target.

Dissociation of the proliferation and differentiation stimuli of granulocyte colony-stimulating factor (G-CSF).

One proposed therapeutic application of granulocyte colony-stimulating factor (G-CSF) is in differentiation induction therapy of myelodysplastic states (MDS) or acute myeloid leukemia (AML). G-CSF however has a substantial growth including effect which limits its potential as a differentiation inducing agent. We have therefore made a systematic search for agents which might restrain the proliferative effects of G-CSF whilst retaining the differentiation stimulus. Of all the agents we have tested on human bone marrow progenitor cells: (6-thioguanine, all-trans retinoic acid, vincristine, recombinant human alpha-2b and gamma-interferon) only the latter abolished the stimulation of cell growth and retained, or possibly increased, the differentiation effect of G-CSF. The antiproliferative drugs 6-thioguanine and vincristine both antagonized the neutrophil-granulocyte differentiation inducing action of G-CSF. Retinoic acid and alpha-2b interferon both had weak effects on proliferation and failed to enhance differentiation. These results suggest that it may be possible, by combining G-CSF with a suitable second agent, to utilize its substantial differentiation inducing effect without incurring the potentially hazardous effects of increased leukemic cell growth.

Effects of DNA topoisomerase II inhibitors on human bone marrow progenitor cells.

Topoisomerase II (topo II) is a target for many cytotoxic agents. Two observations, however, warrant caution in their therapeutic use: first, these agents can inhibit differentiation and second, perturbations in function render the enzyme error-prone. Illegitimate recombination events occurring at sites where topo II acts in differentiation could be particularly important in the development of secondary malignancies (relatively frequent after therapy with agents that target topo II). Topo II inhibitors are heterogeneous in mechanisms of action; in site-specificity of cleavable complex 'entrapment' (where present) and in the relative potency against the two topo II isoforms, all potentially influencing the site of maximum DNA damage. The object of this study was to examine the effect of topo II inhibitors on human haemopoietic precursor cells, to determine which have most impact on differentiation. We selected two which act via cleavable complex entrapment, but with different site preferences (m-AMSA and VP-16), and two acting via other mechanisms (merbarone and fostriecin). VP-16 and m-AMSA showed similar patterns with low dose stimulation of granulocyte-macrophage colony formation and high dose inhibition of all colony types. The stimulation was accompanied by an increase in colony size and blast content, consistent with a low dose inhibition of differentiation. Forstriecin, in contrast, stimulated predominantly mixed and erythroid colonies. Merbarone failed to increase colony formation. Neither produced substantial inhibition of colony formation. The effects on granulocyte-macrophage progenitors were confirmed using 7-day suspension cultures, using nitroblue tetrazolium (NBT) reduction and 3-4,5,dimethylthiazol 2,5-diphenyl tetrazolium bromide (MTT) assays for differentiated cells and total cell mass, respectively. These results demonstrate that the effects of topo II inhibitors on haemopoietic cell proliferation and differentiation are agent-specific and can involve lineage-restricted partial inhibition of differentiation.

Acidic isoferritin stimulates differentiation of normal granulomonocytic progenitors.

Acidic isoferritin (AIF) has been shown to be released by cells from patients with leukemia and to have an inhibitory effect on the growth of normal granulomonocytic (GM) progenitors (leukemia-inhibitory activity) during the S phase of the cell cycle. AIF is also produced by normal mature cells of the monocyte-macrophage lineage. We studied the effects of AIF on the differentiation of normal GM progenitors and found an increase in the number of mature cells in AIF-exposed cultures. This increase did not occur when AIF was pretreated with anti-heart ferritin antiserum or when basic isoferritin was used in the place of AIF. The influence of AIF was not mimicked by removing S phase cells by pretreatment with a pulse of high specific activity tritiated thymidine. Thus, the apparent differentiation-stimulating effect of AIF is not likely to be due to selective removal of immature dividing cells. The results suggest that AIF-inhibitory activity on the proliferation of GM progenitors might at least in part be mediated by a stimulus for differentiation of the target cells, thus regulating the number of mature cells which might be formed by a single progenitor cell.

Differentiation stage-specific expression of a gene during granulopoiesis.

Differential screening of a recombinant cDNA library using cDNAs transcribed from poly(A)+ RNA of normal or leukemic leukocytes revealed a number of recombinants homologous to mRNAs characteristic of particular leukemias. The occurrence of one of these (pCG14) in high abundance was shown to be sufficiently characteristic of the circulating leukocyte population of chronic granulocytic leukemia (CGL) patients to distinguish them from all other populations of leukocytes. We have now characterized the gene encoding this mRNA and shown that its expression is specific to the granulocyte lineage in hemopoietic cells and is, moreover, limited to a narrow stage of differentiation during granulopoiesis. Our results explain why high levels of pCG14 RNA are characteristic of chronic granulocytic leukemia peripheral blood leukocytes.

Evidence for the involvement of DNA topoisomerase II in neutrophil-granulocyte differentiation.

Agents that slow cellular proliferation usually stimulate myeloid differentiation. The demonstration in this report of an anomalous inhibitory behavior of the epipodophyllotoxin VP16-213, an agent known to inhibit the enzyme DNA topoisomerase II, prompted us to investigate the role of this enzyme in both changes in DNA supercoiling and in DNA strand breakage and reunion events occurring during the induction of neutrophil-granulocyte differentiation. We recently reported that retinoic acid, an inducer of granulocytic differentiation, stimulates transient relaxation of DNA supercoiling. We now show that this is associated with the formation of small numbers of protein-linked DNA breaks (a characteristic of topoisomerase reactions). Both events are perturbed by VP16-213, and since this agent inhibits subsequent differentiation, these observations raise the possibility of a role for DNA topoisomerase II in granulocytic differentiation. The possible relevance of these findings to mechanisms of leukemogenesis is discussed.

A bioassay system for two types of colony stimulating factor in human serum.

The biological assay of factors with granulocyte-macrophage colony stimulating activity in human serum poses special problems. Assays based on colony counts suggest, often erroneously, that serum lacks colony stimulating factors (CSF) but interacts with other materials to potentiate or produce factors. Although the scoring of total clone numbers reveals the presence of CSF in serum, this is still no sufficient for the interpretation of the effects of mixing materials containing CSF, because increments in clone numbers are not directly proportional to increments in CSF. This is particularly important in serum assays because two types of activity are present, one which stimulates progenitor cells directly, and another which results from the interation of serum and bone marrow adherent CSF-producing) cells or peripheral blood leucocytes, indicating the presence of both direct acting and "adherent cell dependent" CSF. A biological assay is described which uses analysis of dose-response curves of clone formation in agar culture, and allows simultaneous assay of both types of activity. The criteria for the selection of suitable target progenitor cell populations are discussed.

Stochastic model of human granulocyte-macrophage progenitor cell proliferation and differentiation. I. Setting up the model.

A mathematical model is constructed for the proliferation and differentiation of granulocyte-macrophage progenitor cells in response to the specific proliferation/differentiation stimulus granulocyte-macrophage colony stimulating activity (gm-CSA). The major objective of this model was to test an earlier conceptual model in which we proposed that clone size potential and sensitivity to gm-CSA are functional properties of granulocyte-macrophage progenitor cells, properties that gradually change as the cells differentiate down the granulocyte-monocyte pathway. Another aim was to provide a tool for further analysis of the regulation of granulopoiesis and granulocyte-macrophage progenitor cell proliferation and differentiation. To formulate and then test the mathematical model, available experimental data were divided, one part being used to construct the model, and the other, the results of different and independent experiments, being used for model validation. This report describes the mathematical model and the estimation of model parameters using in vitro experimental results on the relationship between clone number and gm-CSA concentration and on the clone size distributions obtained under conditions of maximal stimulation by gm-CSA. The accompanying article shows how the model was then tested using data from three other types of experiment.

Stochastic model of human granulocyte-macrophage progenitor cell proliferation and differentiation. II. Validation of the model.

In the accompanying paper we constructed a model for the proliferation and differentiation of granulocyte-macrophage progenitor cells in response to the specific proliferation/differentiation stimulus granulocyte-macrophage colony-stimulating activity (gm-CSA). Here we have tested the model without altering the model parameters, which were estimated on the basis of two types of experiment (clone number versus gm-CSA dose-response curves and clone size distributions with maximal gm-CSA) by using these parameters to simulate additional independent experiments. These included data from separation experiments in which progenitor cells were separated on a basis of their density and were then tested for their sensitivity to gm-CSA and their proliferative capacity. In addition, other sets of experimental data (e.g., mean clone size versus time and colony versus gm-CSA dose-response curves and colony-to-cluster ratios) were also examined to detect whether the mathematical model provides an adequate representation of the data. The model gave a successful simulation of all experimental data.

DNA-protein interaction sites in differentiating cells. I. Two-dimensional mapping of modulated sites.

We have previously shown that DNA-protein attachment sites form during the induction of hematopoietic cell differentiation. Affinity phase-partitioning studies of DNA/protein complexes demonstrated that the DNA involved is not randomly distributed throughout the genome. The object of this study was to use filter binding followed by two-dimensional (2D) polyacrylamide gel electrophoresis (using a neutral 6% gel in the first dimension and a denaturing gradient gel in the second dimension) to gain insight into changes in DNA-protein interactions during induced granulocytic and monocytic differentiation of HL60 cells. Nitrocellulose filter-binding enriched samples for protein-associated DNA sufficiently to change the pattern of DNA spots on 2D gels. The patterns of spots obtained was reasonably reproducible between experiments and highly reproducible within experiments. Gels obtained from cells induced to differentiate by either phorbol ester or all-trans retinoic acid (RA) showed identical patterns for the majority of spots but changes in a small proportion of spots with respect to uninduced controls. Both intensification and reduction/disappearance of spots was observed, demonstrating the existence of both invariant and variant DNA/protein attachment sites during the early stages of hematopoietic cell differentiation. Previous studies have implicated DNA topoisomerase II in chromatin structural changes that are necessary for induction of granulocytic differentiation. We therefore examined the filter-binding DNA preparation by 5'-exonuclease digestion (since topoisomerase II is known to bind covalently to the 5'termini on either side of its cleavage sites). The filter-associated DNA exhibited increased 5' exonuclease protection (with respect to filter flow-through DNA), and the degree of protection increased significantly with exposure to phorbol ester and less markedly with retinoic acid. However, since not all filter DNA was 5' protected, it remains unresolved whether the specific differentiation-associated DNA-protein interactions revealed here involve DNA topoisomerase II or some other protein.

Brief report: qualitative differences in the biological activity of two sources of colony stimulating factor: GCT conditioned medium and leucocyte feeder layers.

A most stringent test for the non-identity of biologically active materials is provided by the techniques used for parallel line type bioassay of these materials. We have studied commercially available GCT conditioned medium using a parallel line type biological assay for granulocyte-macrophage colony stimulating factors (GM-CSF) and mononuclear leucocytes as a reference source of GM-CSF. Only test and reference GM-CSF sources which have active constituents with identical dose response characteristics can produce parallel line displacement in this assay system. GCT conditioned medium failed to produce parallel displacements in five different assays. We must therefore conclude that one or more active factors from the test and reference sources of GM-CSF (GCT conditioned medium and mononuclear leucocytes) are not identical. This indicates a need for caution when substituting GCT conditioned medium for leucocyte feeder layers in human bone marrow cultures.

Bone marrow endogenous colony stimulating factor(s): relation to granulopoiesis in vivo.

Granulocyte-macrophage colony stimulating factors (CSF) are produced both in peripheral tissues and within the bone marrow. Stimulators from either source might regulate granulocyte and monocyte production in vivo. The purpose of this study was to devise an assay for bone marrow endogenous CSF so that its role in regulation of granulopoiesis might be assessed. There was a significant positive correlation between the endogenous CSF level and the subsequent trend in peripheral blood neutrophil count in both normal and infected patients, suggestive of a regulatory role. In addition, granulopoietic recovery after neutropenia was associated with increased endogenous CSF levels. There are two candidate regulators of endogenous CSF production, a stimulatory humoral factor and a neutrophil derived inhibitor which inhibits endogenous CSF production in vitro. Our results suggest that the marrow neutrophil level had a much more significant inhibitory effect than the neutrophil level in the culture.

DNA-protein interaction sites in differentiating cells. II. A subset of alphoid repetitive sequences with retinoic acid induced protein attachment and an unusual purine-pyrimidine 'signature'.

Two independent methods were used to recover the tightly or covalently attached DNA-protein complexes which form during the differentiation induction of HL60 cells by all-trans retinoic acid. One method employed nitrocellulose filter binding (described in the accompanying paper) to recover these protein-DNA complexes, following by representational difference analysis (RDA). RDA is an extremely powerful new technique for cloning the difference between complex DNA samples, exploiting a combination of kinetic enrichment during PCR amplification, and subtractive hybridization. In this case, we used filter-bound DNA from undifferentiated cells for subtraction of equivalent DNA from differentiating cells. A second approach used a combination of sodium dodecyl sulfate (SDS)/KCI precipitation and affinity phase partitioning for purification of DNA bound to proteins, followed by selection of clones showing differentiation-specific attachment of proteins (by differential hybridization to protein-attached DNA from differentiating and undifferentiated HL60 cells). Both procedures yielded a high proportion of alphoid repetitive sequences, although slightly higher in the RDA sample than that of the other method (approximately 50% and approximately 30%, respectively). Plots of purine-pyrimidine composition showed that the alphoids recovered by both techniques were unusual and remarkably similar. Although the clones were related to each other, they differed more from the alphoid consensus than examples of alphoid sequences selected from data banks. There was also a contrast between the purine-pyrimidine composition plots of archetypal alphoid monomers and this subset. The principal difference observed related to two adjacent homopyrimidine tracts present in the archetypal monomers but altered in the subset. The possible significance of these differences is discussed with respect to the established roles of alphoid sequences and known/putative protein binding sites including that for the centromeric binding protein (CENB-P box) and response elements for retinoic acid receptors. Following induction of differentiation with retinoic acid, protein attachment (monitored by filter binding) was rapid, maximal at 40 minutes, and still elevated at 165 minutes. This was specific to the alphoid subset, and induced protein binding at a nonrepetitive site in DNA had different kinetics. Changing protein attachment at this subset of alphoid repetitive sequences contributes to differentiation-associated chromatin structural changes.

Polyethylene glycol (PEG)-modified granulocyte-macrophage colony-stimulating factor (GM-CSF) with conserved biological activity.

Polyethylene glycol (PEG) modification improves the pharmacological properties of proteins, usually extending plasma half-life and concomitantly increasing in vivo bioactivity, reducing both antigenicity and immunogenicity, and increasing solubility and resistance to proteolysis. Despite these established benefits, few PEG proteins are in use. Current coupling methods are either traumatic for the protein or involve lengthy and difficult procedures to activate monomethoxyPEG (MPEG). We have applied a new coupling method that allows coupling of MPEG directly to proteins under physiological conditions. Using this method with recombinant human (rh)granulocyte-macrophage colony-stimulating factor (GM-CSF) we were able to construct biologically active PEG-GM-CSF. Fast protein liquid chromatography (FPLC) and phase-partitioning confirmed the presence of PEG modification, and the former was used to fractionate modified and unmodified material. Bioactivity was measured in colony assays of normal human bone marrow cells and by tritiated thymidine uptake (of chronic myeloid leukemia cells and TF-1 cells). With both uptake and colony assays, using unfractionated material, we observed only a modest reduction in biological activity. Assays of FPLC-fractionated material confirmed that much of the bioactivity of the PEG-GM-CSF preparations was due to the modified species and any residual unmodified GM-CSF. Species uncontaminated by tresylmonomethoxyPEG (TMPEG; which was somewhat inhibitory in the thymidine uptake assay and eluted over a broad region of the FPLC profile) had no significant reduction in activity, but we cannot rule out the possibility that PEG-GM-CSF species eluting elsewhere in the profile had modest reduction of activity. Subcutaneous injection into mice confirmed the anticipated improved half-life in vivo and demonstrated a longer uptake from the injection site. This is, as far as we are aware, the first successful construction of PEG-GM-CSF with conserved biological activity.

CFUGEMM, CFUGM, CFUMK: analysis by equilibrium density centrifugation.

A modification of the technique of Fauser and Messner was used for the culture of human multipotent hematopoietic progenitors. This modified technique shows a linear relationship between mixed granulocytic-erythrocytic colonies and the number of cells plated, even at extremely low cell doses (10(4) cells/dish) and is therefore a more suitable assay system for cell separation studies than the original non-linear method. The buoyant density of CFUGEMM (colony forming unit granulocytic-erythrocytic-megakaryocytic-macrophage) was determined using equilibrium density centrifugation. CFUGEMM were of lower buoyant density than the majority of nucleated marrow cells. Substantially enriched populations of CFUGEMM could be obtained with a single density separation procedure. The density distribution profile for CFUGEMM was also distinct from the density distribution of granulocyte-macrophage colony forming cells (CFUGM), the latter being of somewhat greater buoyant density than the former. Cells which formed clones containing only megakaryocytes in culture (CFUMK) had an intermediate density between that of CFUGEMM and CFUGM. The morphological characteristics of these progenitor cells were studied using correlation analysis. Results suggested that the CFUGEMM correspond to transitional cells without granules, the CFUGM to transitional cells with 1-4 granules and the heterogeneous group of GM-cluster forming cells to a broad category including myeloblasts, promyelocytes, myelocytes and metamyelocytes.

The identification of mixed granulocytic-erythrocytic colonies in vitro.

The identification of mixed granulocytic-erythrocytic colonies in culture poses special problems. First, methods based on the detection of peroxidase-like activity of red-cell hemoglobin using benzidine can give false-positive results with eosinophilic cells in colonies. Secondly, the use of the color of hemoglobin itself to identify mixed colonies prevents the detection of those colonies where the erythrocytic component is small or contains relatively little hemoglobin. We present here a new method using O-Dianisidine (O-D) to detect red-cell peroxidase activity. Subsequent staining with Luxol fast blue showed that eosinophils did not give a positive reaction with O-D. The technique also has the advantage that the whole culture gel is processed and that there is no need to pick off individual clones for identification. Using the combined stain O-D and Luxol fast blue most lineages within clones can be discriminated.

PEGylation of adenovirus with retention of infectivity and protection from neutralizing antibody in vitro and in vivo.

Replication-defective recombinant adenovirus (Ad) vectors are under development for a wide variety of gene therapy indications. A potential limiting factor associated with virus gene therapy requiring repeated treatment is the development of a humoral immune response to the vector by the host. In animal models, there is a dose-dependent rise in neutralizing antibodies after primary vector administration, which can preclude effective repeat administration. The strategy we have developed to circumvent the neutralization of adenovirus vectors by antibodies is to mask their surface by covalent attachment of the polymer polyethylene glycol (PEG). Covalent attachment of PEG to the surface of the adenovirus was achieved primarily by using activated PEG tresylmonomethoxypolyethylene glycol (TMPEG), which reacts preferentially with the epsilon-amino terminal of lysine residues. We show that the components of the capsid that elicit a neutralizing immune response, i.e., hexon, fiber, and penton base, are also the main targets for PEGylation. Several protocols for PEGylation of an adenovirus vector were evaluated with respect to retention of virus infectivity and masking from antibody neutralization. We show that covalent attachment of polymer to the surface of the adenovirus can be achieved with retention of infectivity. We show further that PEG-modified adenovirus can be protected from antibody neutralization in the lungs of mice with high antibody titers to adenovirus, suggesting that PEGylation will improve the ability to administer Ad vectors on a repeated basis.

Immunological characterization of hemopoietic cells in the common marmoset, rhesus monkey, and man. In search of a model for human marrow transplantation.

Bone marrow, lymphoid, and peripheral blood cells from the common marmoset and rhesus monkey have been tested with a panel of heterologous and monoclonal antibodies, and their reactivity pattern has been compared with that of blood and bone marrow cells from human donors. Conventional antibodies reveal extensive cross-reactivity within the B cell, T cell, and granulocytic systems in all three species, however, some important differences have been exposed. Only the monoclonal antibodies to HLA-A,B,C and Ia-like antigens react with marmoset cells, and we have exploited this finding to show that the vast majority of colony-forming units (CFU-c) in the marmoset bone marrow (as in man) are Ia positive. The use of the common marmoset as a suitable model for human bone marrow transplantation is discussed in the light of these findings.

The uses and properties of PEG-linked proteins.

Poly(ethylene glycol) (PEG) is a water soluble polymer that when covalently linked to proteins, alters their properties in ways that extend their potential uses. PEG-modified conjugates are being exploited in many different fields. The improved pharmacological performance of PEG-proteins when compared with their unmodified counterparts prompted the development of this type of conjugate as a therapeutic agent. Enzyme deficiencies for which therapy with the native enzyme was inefficient (due to rapid clearance and/or immunological reactions) can now be treated with equivalent PEG-enzymes. PEG-adenosine deaminase has already obtained FDA approval. PEG-modified cytokines have been constructed and, interestingly, one of the conjugates, PEG-modified granulocyte-macrophage colony-stimulating factor, showed dissociation of two biological properties. This novel observation may open new horizons to the application of PEGylation technology. The biotechnology industry has also found PEG-proteins very useful because PEG-enzymes can act as catalysts in organic solvents, thereby opening the possibility of producing desired stereoisomers, as opposed to the racemic mixture usually obtained in classical organic synthesis. Covalent attachment of PEG to proteins requires activation of the hydroxyl terminal group of the polymer with a suitable leaving group that can be displaced by nucleophilic attack of the epsilon-amino terminal of lysine residues (other nucleophilic groups can also interact). Several chemical groups have been exploited to activate PEG, thereby giving rise to a variety of PEG-proteins. Some of these varieties retain part of the activating group as a coupling moiety between PEG and protein and others provide a direct linkage. For each particular application, different coupling methods provide distinct advantages.(ABSTRACT TRUNCATED AT 250 WORDS)