Induced differentiation of murine erythroleukemia cells: cellular and molecular mechanisms

Transforming growth factor beta 1 is an inducer of erythroid differentiation

Normal human bone marrow cells, highly enriched for burst-forming units-erythroid (BFU-E), were cultured in serum-free medium, in the presence and absence of various factors, to investigate the mechanisms involved in regulating erythroid differentiation. In cultures containing interleukin 3 (IL-3), Steel factor (SF), and erythropoietin (Ep), benzidine-positive erythroblasts first became detectable on day 6. Their numbers then rapidly increased until, by day 16, > 99% of the cells, which were 20,000-fold amplified over input numbers, were benzidine-positive. It is interesting to note that omission of either SF or Ep from this assay markedly enhanced the rate of differentiation and reduced total cell numbers, whereas omission of IL-3 had no effect on the rate of differentiation and only slightly reduced cell numbers. Of various agents tested, the most potent erythroid differentiation inducer (and inhibitor of cell proliferation) was found to be transforming growth factor beta 1 (TGF-beta 1). This cytokine stimulated both the rapid appearance of hemoglobin-positive cells and an early cessation of cell proliferation. Using fluorescently tagged antibodies to glycophorin A and fluorescence-activated cell sorter (FACS) analysis, this phenomenon was shown to be due to an early induction of erythroid differentiation rather than an aberrant production of hemoglobin. Methylcellulose assays indicated that the well-documented reduction of BFU-E colony numbers observed with TGF-beta 1 may actually be due to a TGF-beta 1-induced "conversion" of BFU-E into colony-forming units-erythroid (CFU-E). Thus, in vivo, TGF-beta 1 might serve, in part, to decrease the number of mature erythrocytes by stimulating BFU-E to skip a number of cell divisions and differentiate early.

Antagonism of cadmium cytotoxicity by differentiation inducers

Studies on the antagonism of toxicity can provide information about toxic mechanisms and suggest chemotherapeutic strategies. A rapid cell growth assay that measures the effects of test agents on the accumulation of cell protein (Shopsis and Eng, Toxicol. Lett. 1985;26:1) has been applied to studies of the antagonism of the cytotoxicity of cadmium. Exposure of Balb/c mouse 3T3 cells to 15 mumol/L Cd2+ for 24 h or 7 mumol/L Cd2+ for 48 h caused a 50% decrease in total cell protein. Zn2+ and selenite ion, antagonists of Cd toxicity in vivo, antagonized Cd2+ cytotoxicity when added in micromolar concentrations at the initiation of exposure to Cd2+. A diverse group of chemicals that can induce differentiation in vitro in cultured erythroleukemia and other cells were also found to antagonize the cytotoxic effects of Cd2+ to 3T3 cells. Dimethyl sulfoxide (DMSO), hexamethylene bisacetamide, N,N-dimethyl formamide, N-methyl formamide, dimethyl acetamide, hypoxanthine, hemin, ouabain, and sodium butyrate, when added to cultures simultaneously with Cd2+, each antagonized Cd2+ toxicity. These agents were used at concentrations equal to or lower than the concentrations at which they induce cellular differentiation. Other cytotoxicity assays and morphological studies confirmed these observations. DMSO added as much as 6 h after the initiation of a 24-h exposure to Cd2+ still protected cells; conversely, pretreatment of cultures with butyrate or DMSO for 24 h followed by their removal did not confer protection against subsequent Cd2+ challenge. Ethanol and methanol (noninducers of differentiation) did not antagonize Cd2+ cytotoxicity, and differentiation-inducing agents did not protect the cells from Zn(2+)- or Hg(2+)-induced cytotoxicity. DMSO treatment does not induce an increase in the concentrations of metallothionein or glutathione in these cells.

Analysis by cell hybridization of mechanisms that regulate beta-adrenergic responses in reticulocytes and in differentiating erythroid cells

In intact reticulocytes, but not in fragmented membranes, the loss of adenylate cyclase activity during cell maturation followed a biphasic time course. A rapid phase (t1/2 approximately 2 h) during which the initial activity was reduced by 40-50% was followed by a slow phase with t1/2 close to 3 days. The fast decay seemed to occur on the adenylate cyclase level since (-)isoprenaline- or forskolin-stimulated activities behaved similarly and bacterial toxin-monitored Gs and Gi proteins remained stable. The mechanism of the initial decrease in hormonal responsiveness was further analysed in hybrid cells prepared by fusing reticulocytes with Friend erythroleukemia (MEL) cells. The hybrids contained reticulocyte-derived beta-adrenoceptors and MEL cell-derived adenylate cyclase and G proteins. Fusion of reticulocytes to native MEL cells caused adenylate cyclase activity to drop by 30% at 2 h and 45% at 18 h after fusion. By contrast, hybrids prepared after dimethylsulfoxide-induced differentiation of MEL cells showed stable or increasing rates of receptor-coupled cAMP formation between 2 and 18 h after fusion, concomitant with the enhanced activity of the Gs protein in these cells. A cyclase-stimulating factor present in the cytosol of MEL cells and of reticulocytes appeared not to be involved in short-term regulation of hormonal responsiveness. We conclude that the strength of beta-adrenergic responses in erythroid progenitor cells is primarily regulated by modulating G protein-mediated receptor cyclase coupling while reticulocytes, during early maturation, seem to rely on direct inactivation of adenylate cyclase, probably via a cytosolic proteolytic pathway.

Regulation of beta-adrenergic responses during in vitro differentiation of mouse erythroleukemia cells

Dimethyl sulfoxide (DMSO)-induced erythroid differentiation of Friend mouse erythroleukemia (MEL) cells is associated with a marked transient modulation of catecholamine sensitivity. Within 24 h after induction and well before the onset of hemoglobin synthesis, we observed a 3-fold increase in beta-receptor density and a more than 10-fold increase in receptor-coupled cAMP formation. During the following 4 days, in parallel with the development of normoblast-like cells, receptor numbers returned to preinduction levels while catecholamine-dependent cAMP formation remained significantly elevated. Simultaneously, the apparent potency of the beta-adrenoceptor agonist isoprenaline increased 10-fold. Improved receptor-cyclase coupling is probably due to a major shift in the expression of Gi and Gs regulatory proteins. Bacterial toxin-mediated ADP-ribosylation of membrane proteins suggests that the dominating species in native cells is Gi (Gsa:Gia = 1.7). By contrast, Gs predominates in differentiated cells (Gsa:Gia = 1.8:1). Receptor-independent forskolin-stimulated cAMP formation showed a pronounced, albeit transient, decrease during differentiation. We suggest that these changes in cellular cAMP responses may be important for transient positive or negative cooperative interactions between hormones and growth factors in the course of erythroid cell development.

Artificial reversion of acute myeloid leukemia cells into normal phenotype

1. Induction of tumor cell differentiation could reverse transformed cells into normal, mature cells. Important question is whether these malignant-to-normal reversed cells are really normal ones. 2. We have developed an experimental model based on the examination of three different levels of human acute myeloid leukemia cell properties before and after induction of differentiation: morphological (percentage of undifferentiated blast cells), functional (DNA ploidy, Fc receptors, phagocytic activity, clonogenic assay in soft agar, oxidative metabolism which accompanies phagocytosis in mature granulocytes) and genetical (expression of oncogene p53). 3. Several inducers have been employed: dimethylsulfoxide (DMSO) granulocyte-macrophage colony stimulating factor (GM-CSF); tunicamycin, interferon gamma, tumor necrosis factor and lipopolysaccharide. 4. Our results indicate that the reversion of leukemic cells into mature normal ones with some inducers (DMSO, GM-CSF) could be a complete process.

Temporal changes in phosphoamino acid phosphatase activities in murine erythroleukaemic cells

1. The activities of phosphotyrosine, phosphothreonine and phosphoserine phosphatases were measured at various time periods in Friend murine erythroleukaemic (MEL) cells. 2. Effects of DMSO (dimethyl sulphoxide) and HMBA (hexamethylene bisacetamide), inducers of differentiation, were examined. 3. The activities of all three enzymes showed cyclic variation when measured on a daily basis over a period of several days; this was also the case when phosphothreonine phosphatase was determined in cells treated with DMSO or HMBA and when phosphoserine phosphatase was assayed after stimulation of the cells with HMBA. 4. Evidence for rhythmic variation in phosphotyrosine phosphatase activities was also obtained when the activities were determined at hourly intervals both in control cells and those treated with HMBA. 5. The time-dependent changes observed could be significant in that control of dephosphorylation may possibly be achieved by altering the rhythms.

Induction of HL-60 monocytic cell differentiation promoted by a perturbation of DNA synthesis: hydroxyurea promotes action of TPA

Control of terminal cell differentiation was studied using the human promyelocytic leukemia cell line, HL-60. HL-60 cells are known to undergo terminal monocytic differentiation when continuously exposed to 1.6 nM tetradecanoylphorbol acetate (TPA). The dose-response relationship between TPA concentration and induced differentiation is relatively steep. TPA (1.1 nM) induces little G1/0 specific growth inhibition or phenotypic differentiation. In contrast, pretreating the cells with a pulse exposure to hydroxyurea promotes their capability to terminally differentiate in response to TPA. Initially exponentially proliferating cells exposed for 20 h, approximately one doubling time, to 0.3 mM hydroxyurea, a subcytotoxic dose, underwent rapid G1/0 specific growth arrest and cell differentiation in response to subsequent exposure to 1.1 nM TPA. The extent of terminal differentiation was comparable to that induced by 1.6 nM TPA. The results support the hypothesis that early events in induction of terminal HL-60 cell differentiation depend on an S phase-specific process which may involve gene amplification.

Hydroxyurea indices precommitment during retinoic induced HL-60 terminal myeloid differentiation: possible involvement of gene amplification

Control of terminal cell differentiation was studied in the HL-60 human promyeloctyic leukemia cell line. Retinoic acid is known to induce myeloid differentiation associated with GO arrest in these cells. In this case, onset of terminal differentiation occurs after an exposure period corresponding to two division cycles. This is preceded by acquisition of a precommitment memory state occurring by one division cycle. Cells in precommitment undergo accelerated onset of terminal differentiation upon reexposure to inducer. The present report shows that the precommitment state can be induced by a pulse exposure to hydroxyurea. While the hydroxyurea exposure does not itself induce terminal differentiation, the treated cells undergo accelerated onset of phenotypic differentiation and GO arrest upon exposure to retinoic acid. Thus a perturbation of S-phase specific cellular metabolism induces the early precommitment regulatory state in the course of induced HL-60 terminal myeloid differentiation. The results support a model in which initiation of a cellular program of terminal differentiation depends on an S-phase specific event associated with replication of cellular DNA and possibly involving gene amplification. Significantly, the results indicate that a conventional chemotherapeutic agent such as hydroxyurea can synergistically interact with a differentiation inducing agent such as retinoic acid. This indicates the significance of investigating the interaction between conventional S-phase specific chemotherapeutic agents and differentiation inducing agents as a potential treatment modality.

Regulation of proliferation and differentiation of respiratory tract epithelial cells by TGF beta

In this paper we examined the effects of transforming growth factor beta (TGF beta) on the proliferation and differentiation of rabbit tracheal epithelial cells in primary culture. Treatment of these cells with TGF beta inhibits cell proliferation in a time- and dose-dependent manner; concentrations as low as 1 pM are able to inhibit cell growth. Concomitantly, TGF beta causes cells to accumulate in the G0/G1 phase of the cell cycle and a sharp reduction in the ability of the cells to form colonies after subculture at clonal density. These results indicate that TGF beta induces terminal cell division in these cells. The inhibition of cell growth is accompanied by changes in cell morphology and a stimulation of the formation of cross-linked envelopes. TGF beta enhances the levels of transglutaminase activity and cholesterol sulfate, two markers of squamous differentiation. Our results indicate that TGF beta induces terminal squamous cell differentiation in rabbit tracheal epithelial cells. Retinoic acid (RA) does not affect the commitment to terminal cell division induced by TGF beta, but inhibits the expression of the squamous phenotype. Growth of normal human bronchial epithelial cells was affected by TGF beta in a way similar to that of rabbit tracheal epithelial cells. Several carcinoma cell lines tested were quite resistant to TGF beta, whereas growth of one carcinoma cell line was stimulated by TGF beta. These results indicate that a modified response to TGF beta could be one mechanism involved in the aberrant growth control of malignant cells.

Distribution of interdivisional times in proliferating and differentiating Friend murine erythroleukaemia cells

The interdivisional times of Friend murine erythroleukaemia cells which are growing continuously, or during terminal erythroid differentiation after exposure to dimethyl sulphoxide (DMSO), were determined by time lapse video photography. The median interdivisional times were found to increase from 11.75 hr before exposure to DMSO, to 24.0 hr at 72 hr after exposure. This increase in median interdivisional time was accompanied by an increase in heterogeneity of interdivisional times (%CV = 8.5----40.8), by an increase in the similarity of sister interdivisional times (ryy = 0.622----0.925), and by a decrease in the fraction of cells observed to divide (F = 1.0----0.807). Cells exposed to DMSO for 72 hr can be induced to divide at least once with nearly normal interdivisional times, if they are resuspended at a tenfold higher cell concentration. Computer simulations of cell cycle regulation, based on the opposing reactions model of Murphy, generate interdivisional time distributions which resemble the experimental data better than the single transition probability model of Smith and Martin.

The proportions of hemoglobin types induced in mouse erythroleukemia cells vary with the inducer or combination of inducers, the inducer concentration and the time of induction

The relative amounts of hemoglobin (Hb) major and Hb minor accumulated during induction of erythrodifferentiation in mouse erythroleukemia (MEL) cells were studied. The ratio of major to minor was found to depend not only upon the inducer tested (as reported previously by others), but also upon the concentration of the inducer and the time of exposure to the inducer as well as the specific cell line of MEL cells studied. At concentrations required for optimal induction of differentiation, certain agents led to the accumulation of predominantly Hb major, but suboptimal concentrations of the same inducers led to predominantly Hb minor accumulation. After a relatively short induction time (2 da) utilizing a given inducer either the level of Hb minor was higher than that of Hb major or the levels of the two Hb's were approximately equal, but after longer induction periods (3-7 da) Hb major was more abundant than Hb minor. In addition, it was found that the three proteases tested induced predominantly Hb minor. The addition of suboptimal concentrations of low molecular weight inducers acted synergistically with a given protease to produce a high yield of Hb-containing cells. When these agents were added singly they induced relatively low Hb major/Hb minor ratios, but when a low molecular weight inducer was added together with a protease in a "synergistic" combination, elevated ratios were induced. The proportions of hemoglobin types induced in MEL cells may be related in part to the intensity of the induction response. In view of these data, classifications of inducers based solely on the ratios of Hb types produced must be guarded.