Macrophages function as a ferritin iron source for cultured human erythroid precursors

Iron is essential for the survival as well as the proliferation and maturation of developing erythroid precursors (EP) into hemoglobin-containing red blood cells. The transferrin-transferrin receptor pathway is the main route for erythroid iron uptake. Using a two-phase culture system, we have previously shown that placental ferritin as well as macrophages derived from peripheral blood monocytes could partially replace transferrin and support EP growth in a transferrin-free medium. We now demonstrate that in the absence of transferrin, ferritin synthesized and secreted by macrophages can serve as an iron source for EP. Macrophages trigger an increase in both the cytosolic and the mitochondrial labile iron pools, in heme and in hemoglobin synthesis, along with a decrease in surface transferrin receptors. Inhibiting macrophage exocytosis, binding extracellular ferritin with specific antibodies, inhibiting EP receptor-mediated endocytosis or acidification of EP lysosomes, all resulted in a decreased EP growth when co-cultured with macrophages under transferrin-free conditions. The results suggest that iron taken up by macrophages is incorporated mainly into their ferritin, which is subsequently secreted by exocytosis. Nearby EP are able to take up this ferritin probably through clathrin-dependent, receptor-mediated endocytosis into endosomes, which following acidification and proteolysis release the iron from the ferritin, making it available for regulatory and synthetic purposes. Thus, macrophages support EP development under transferrin-free conditions by delivering essential iron in the form of metabolizable ferritin.     

Effect of extracellular hemin on hemoglobin and ferritin content of erythroleukemia cells

Mouse (MEL) and human (K-562) erythroleukemia cell lines can be induced to undergo erythroid differentiation, including hemoglobin (Hb) synthesis, by extra cellular hemin. In order to study the effect of extracellular hemin on intracellular ferritin and Hb content, we have used Mossabauer spectroscopy to measure the amount of 57Fe incorporated into ferritin or Hb and a fluorescent enzyme-linked immunosorbent assay (ELISA) to measure the ferritin protein content. When K-562 cells were cultured in the presence of a 57Fe source either as transferrin or citrate, in the absence of a differentiation inducer, all the intracellular 57Fe was detected in ferritin. When the cells were cultured in the presence of 57Fe-hemin, 57Fe was found in both ferritin and Hb. 57Fe in ferritin increased rapidly, and after 2 days it reached a plateau at 5 X 10(-14) g/cell. 57Fe in Hb increased linearly with time and reached the same value after 12 days. Addition of other iron sources such as iron-saturated transferrin, iron citrate, or iron ammonium citrate caused a much lower increase in ferritin protein content as compared to hemin. When K-562 cells were induced by 57Fe-hemin in the presence of 56Fe-transferrin, 57Fe was found to be incorporated in equal amounts into both ferritin and Hb. However, when the cells were induced by 56Fe-hemin in the presence of 57Fe-transferrin, 57Fe was incorporated only into ferritin, but not into Hb, which contained 56Fe iron. These results indicate that in K-562 cells, when hemin is present in the culture medium it is preferentially incorporated into Hb, regardless of the availability of other extra- or intracellular iron sources such as transferrin or ferritin. In MEL cells induced to differentiate by dimethylsulfoxide (DMSO) a different pattern of iron incorporation was observed; 57Fe from both transferrin and hemin was found to incorporate in ferritin as well as in Hb.     

Autoinduction of differentiation in myeloid leukemic cells: restoration of normal coupling between growth and differentiation in leukemic cells that constitutively produce their own growth-inducing protein.

Growth and differentiation of normal myeloid haematopoietic cells are regulated by a family of macrophage- and granulocyte-inducing (MGI) proteins. Some of these proteins (MGI-1) induce cell growth and others (MGI-2) induce cell differentiation. Addition of MGI-1 to normal myeloid cells induces growth and also induces the endogenous production of MGI-2. This induction of differentiation-inducing protein by growth-inducing protein then ensures the coupling between growth and differentiation found in normal cells. There are myeloid leukemic cells that constitutively produce their own MGI-1, but the cells do not differentiate in culture medium containing horse or calf serum. By removing serum from the medium, or in medium with mouse or rat serum, these leukemic cells are induced to differentiate to mature cells, which like normal mature cells, then no longer multiply. Leukemic cells with constitutive production of MGI-1 continuously cultured in serum-free medium with transferrin were also induced to differentiate by removing transferrin. This induction of differentiation was in all these cases associated with the endogenous production of MGI-2 by the cells. The results indicate that changes in specific constituents of the culture medium can result in autoinduction of differentiation in these leukemic cells due to restoration of the induction of MGI-2 by MGI-1, which then restores the normal coupling of growth and differentiation.     

Expression of transferrin receptors and intracellular ferritin during terminal differentiation of human monocytes

Human blood monocytes when cultured on hydrophobic Teflon membranes differentiate into mature macrophages. The expression of transferrin receptors was monitored by monoclonal antibody (OKT9) binding as detected by immunoperoxidase staining. Whereas monocytes were negative, an increasing percentage of macrophages, starting from day 2 in culture, labelled with the antitransferrin receptor antibody as these cells undergo differentiation. After completion of maturation more than 90% of macrophages expressed transferrin receptors. While 90-95% of macrophages from broncho-alveolar lavage fluids labelled with the OKT9 antibody, only a minor portion of macrophages obtained from peritoneal and pleural cavities did so. In parallel, intracellular ferritin in cells of the monocyte-macrophage lineage increased from 10 ng/10(6) cells to 350-1,500 ng/10(6) cells during maturation in vitro. Alveolar macrophages proved to have the highest ferritin content which ranged from 355-8,400 ng/10(6). The results may indicate that iron uptake and storage is a function of cells at late stages of macrophage maturation and that the occurrence of surface receptors for transferrin can be regarded as differentiation dependent marker.     

Growth of human tumor cell lines in transferrin-free,low-iron medium

Summary Iron is essential for tumor cell growth. Previous studies have demonstrated that apart from transferrin-bound iron uptake, mammalian cells also possess a transport system capable of efficiently obtaining iron from small molecular weight iron chelates (Sturrock et al., 1990). In the present study, we have examined the ability of tumor cells to grow in the presence of low molecular weight iron chelates of citrate. In chemically defined serum-free medium, most human tumor cell lines required either transferrin (5 μg/ml) or a higher concentration of ferric citrate (500 μM) as an iron source. However, we have also found that from 13 human cell lines tested, 4 were capable of long-term growth in transferrin-free medium with a substantially lower concentration of ferric citrate (5 μM). When grown in medium containing transferrin, both regular and low-iron dependent cell lines use transferrin-bound iron. Growth of both cell types in transferrin medium was inhibited to a certain degree by monoclonal antibody 42/6, which specifically blocks the binding of transferrin to the transferrin receptor. On the contrary, growth of low-iron dependent cell lines in transferrin-free, low-iron medium (5 μM ferric citrate) could not be inhibited by monoclonal antibody 42/6. Furthermore, no autocrine production of transferrin was observed. Low-iron dependent cell lines still remain sensitive to iron depletion as the iron(III) chelator, desferrioxamine, inhibited their growth. We conclude that low-iron dependent tumor cells in transferrin-free, low-iron medium may employ a previously unknown mechanism for uptake of non-transferrin-bound iron that allows them to efficiently use low concentrations of ferric citrate as an iron source. The results are discussed in the context of alternative iron uptake mechanisms to the well-characterized receptor-mediated endocytosis process.     

The in vitro response of human tumour cells to desferrioxamine is growth medium dependent

Abstract. Iron chelating agents have been demonstrated to inhibit tumour cell growth. However, in vitro and in vivo results using desferrioxamine a hexadentate iron chelating agent, for anti-cancer treatment are not always in agreement. Therefore, we have studied the response of three human tumour cell lines (HL-60 promyelocytic leukaemia, MCF-7 breast cancer and HepG2 hepatoma), grown in culture medium supplemented with either human pooled (HPS) or fet al bovine serum (FBS), to desferrioxamine. Desferrioxamine, at micromolar concentrations, induced severe cytotoxicity in all tumour cell lines grown in FBS medium. When grown in HPS medium, comparable desferrioxamine cytotoxicity was observed in the millimolar range. The addition of 50% saturated human transferrin to FBS medium resulted in protection against desferrioxamine cytotoxicity. HL-60 cells were further studied for iron metabolism characteristics. HL-60 cells, grown in medium with FBS, were found to have an 8.4 fold increase in surface transferrin receptor (TfR) expression ( P < 0.001) as compared with HL-60 cells grown in medium with HPS. However, iron uptake of HPS cultured HL-60 cells, after incubation with saturated human transferrin, was higher, resulting in a higher concentration of iron in HPS cultured HL-60 cells as compared with FBS cultured cells (1.72 ± 0.02 μmol/g protein v. 1.32 ± 0.14 μmol/g protein; P < 0.001). Using desferrioxamine it was shown that TfR expression is dependent on the biological availability of iron in the cell. Consistent with the lower iron content in FBS cultured cells, we conclude that the cytotoxicity of desferrioxamine is dependent on the ability of cells to replenish cellular iron stores from the culture medium. Cells grown in FBS medium lack this ability and are therefore more susceptible to desferrioxamine.     

Transferrin receptor and ferritin expressions of a human-human hybridoma in protein-free media

Summary The expressions of transferrin receptor and ferritin in a human-human hybridoma in protein-free media were examined. The regulation of the expressions of the two proteins appeared to be decoupled in protein-free cultures. Such cultures also exhibited much higher ferritin expression and endogenous iron pool than a culture in serum-free medium containing transferrin.     

Receptor binding and mitogenic effects of insulin and insulinlike growth factors I and II for human myeloid leukemic cells

Insulin and insulinlike growth factors I and II (IGF-I and IGF-II) influence mesodermal cell proliferation and differentiation. As multiple growth factors are involved in hemopoietic cell proliferation and differentiation, we assessed the receptor binding and mitogenic effects of these peptides on a panel of mesodermally derived human myeloid leukemic cell lines. The promyelocytic cell line HL60 had the highest level of specific binding for these 125I-labeled ligands, with lower binding to the less differentiated myeloblast cell line KG1 and undifferentiated blast variants of these cell lines (HL60blast, KG1a). Insulin binding affinity and receptor numbers were reduced significantly by chemically induced granulocytic differentiation of HL60 cells and was unchanged following induced monocytic differentiation. No substantial alteration in IGF-I or -II binding occurred with induced HL60 cell differentiation. Insulin and IGF-I demonstrated cross competition for receptor binding and down-regulated their homologous receptors without detectable cross modulation of the heterologous receptors on HL60 cells. IGF-I and insulin increased HL60 cell proliferation, as assessed by 3H-thymidine uptake, IGF-I greater than insulin. IGF-I binding and mitogenic effects were blocked by the monoclonal anti-IGF-I receptor antibody IR3, indicating that IGF-I-induced proliferative effects were mediated via its homologous receptor. In contrast, insulin binding and mitogenesis displayed blocking by both anti-IGI-I and anti-insulin receptor antibodies, indicating mediation of its activity through both receptors. These data demonstrate specific binding and mitogenic interactions between insulin, IGFs, and hemopoietic cells which are associated with their state of differentiation.     

Expression of transferrin receptors in phytohemagglutinin-stimulated human T-lymphocytes. Evidence for a three-step model

Resting human T-lymphocytes show an elevated intracellular concentration of ferritin, whereas transferrin receptors are not detectable. Stimulation by phytohemagglutinin markedly lowers their ferritin content, while inducing the synthesis of transferrin receptors. Addition of iron salts (ferric ammonium citrate) in activated T-lymphocyte cultures causes a marked enhancement of both [3H]uridine and [3H]thymidine incorporation. Nevertheless, it also induces a concentration-dependent decrease in transferrin receptor synthesis, associated with a marked rise of ferritin production. Hemin treatment exerts the same effects. Addition of picolinic acid in phytohemagglutinin-stimulated cultures causes a decrease of [3H]thymidine incorporation, whereas transferrin expression is markedly enhanced. The action of iron salts and chelators is specific for transferrin receptors, since the expression of other membrane markers of activated human T-lymphocytes (interleukin-2 receptor, insulin receptor, and HLA-DR antigen) is not modified by treatment with iron or picolinic acid. These observations suggest that expression of transferrin receptors in activated T-lymphocytes is specifically modulated by their intracellular iron level, rather than their proliferative rate. Addition of picolinic acid to resting T-lymphocytes in the absence of mitogen induces a marked decrease of their ferritin content, but not the appearance of transferrin receptors. On the basis of these results, we suggest a three-step model: (a) in resting T-lymphocytes, the gene for transferrin receptor is apparently "closed," in that it is not expressed under both normal conditions and following iron deprivation. (b) After mitogen stimulus, T-lymphocytes are reprogrammed into cell cycle progression, which necessarily entails synthesis of transferrin receptors (c) Expression of these receptors is modulated by the intracellular iron level, rather than the rate of proliferation per se.     

Replacement of serum by insulin and transferrin supports growth and differentiation of the human promyelocytic cell line, HL-60

The human promyelocytic leukemia cell line HL-60 can be grown in serum-free synthetic medium supplemented with insulin and transferrin alone. Growth of HL-60 in this defined medium is at a rate approx. 80% of that in medium containing serum. Moreover, the distinct morphological and histochemical myeloid characteristics of HL-60 are maintained in such serum-free medium. The HL-60 promyelocytes are induced by DMSO to differentiate to mature granulocytes equally well in both serum-supplemented and serum-free medium. However, this defined medium does not support colony growth of HL-60 in semi-solid medium such as methylcellulose.     

Development of calcium and secretory responses in the human promyelocytic leukemia cell line HL60

We have begun to characterize the development of the excitation-response coupling sequence in the human promyelocytic leukemia cell line HL60. Using the recently developed fluorescent calcium probe quin-2, it was found that DMSO induced myeloid differentiation of the HL60 cells is accompanied by the development of a calcium response to the addition of the chemotactic factors fMet-Leu-Phe and leukotriene B4. The characteristics (time course, concentration dependence, stereospecificity, and metabolic dependence) of the calcium response are extremely similar to those previously described in human neutrophils. These results imply that functional receptors for leukotriene B4 appear in HL60 cells upon the induction of differentiation and also lend strong support to the use of these HL60 cells as a model of human myeloid differentiation. We have also characterized the emergence of a secretory response to fMet-Leu-Phe and leukotriene B4 in cytochalasin B treated HL60 cells. In addition, it is found that differentiation was required for the calcium ionophore A23187 to express its secretory activity toward the HL60 cells. This last set of results implies that differentiation is accompanied by the coordinated appearance of surface receptors and cytoplasmic factors required for the expression of cellular responsiveness.     

Mechanism of phorbol diester-induced regulation of surface transferrin receptor involves the action of activated protein kinase C and an intact cytoskeleton

Phorbol diesters are tumor-promoting agents that cause differentiation of HL60 human leukemic cells and concomitantly regulate surface transferrin receptors. Regulation of transferrin receptors by phorbol diesters involves receptor internalization in association with increased receptor phosphorylation (hyperphosphorylation). The intracellular mechanism of action of phorbol diester involves binding to and activation of the Ca2+-phospholipid-dependent protein kinase (protein kinase C). Present studies comparing results obtained with whole cells and those from a cell-free system reconstituted from purified protein kinase C and transferrin receptor components have revealed that the transferrin receptor is phosphorylated by protein kinase C activated by phorbol esters. Following tryptic digestion and two-dimensional separation of phosphopeptides of phosphorylated transferrin receptors, two major and several minor phosphoserine-containing fragments are resolved. These fragments are identical whether transferrin receptor is phosphorylated in whole cells incubated with phorbol diesters or following phosphorylation of affinity immobilized transferrin receptor in the in vitro reconstitution system. Phosphoamino acid analysis of these fragments indicates that serine is the only amino acid phosphorylated in whole cells or in the cell-free system. In addition, colchicine is shown to inhibit in a dose-dependent manner phorbol diester-induced internalization but not hyperphosphorylation of the surface transferrin receptor in whole cells. This inhibition is specific for colchicine since inactive beta- and gamma-Lumicolchicine have no such effect, while taxol reverses the inhibition. These results indicate that the phorbol diester-mediated process of down-regulation of the surface transferrin receptor is associated with phosphorylation of the receptor by activated protein kinase C and requires an intact cytoskeleton to affect receptor internalization.     

Effect of iron and retinoic acid on the control of transferrin receptor and ferritin in the human promonocytic cell line U937.

The effect of changes in iron availability and induction of differentiation on transferrin receptor expression and ferritin levels has been examined in the promonocytic cell line U937. Addition of iron (as 200 micrograms/ml saturated transferrin) or retinoic acid (1 microM) both caused approx. 70% reduction in the average number of surface transferrin receptors, while the iron chelator desferrioxamine caused an 84% increase. Comparable changes also occurred in the levels of transferrin receptor mRNA. Neither iron nor retinoic acid significantly altered the half-life of transferrin receptor mRNA in the presence of actinomycin D (approx. 75 min) but a 10-fold increase in stability occurred in the presence of desferrioxamine. Iron and retinoic acid both caused an increase in intracellular ferritin levels (approx. 4-and 3-fold, respectively), while desferrioxamine reduced ferritin levels by approx. two-thirds. The effect of iron and retinoic acid added together did not differ greatly from that of each agent alone. None of the treatments greatly affected levels of L-ferritin mRNA. Virtually no H-ferritin mRNA was detected in U937 cells. These results show that changes in ferritin and transferrin receptor caused by treatment with retinoic acid are similar to those induced by excess iron, and suggest that changes in these proteins during cell differentiation are due to redistribution of intracellular iron into the regulatory pool(s), rather than to iron-independent mechanisms.     

HL60 cells halted in G1 or S phase differentiate normally

Differentiating agents regulate the proliferation and myeloid maturation of HL60 cells by mechanisms that are at least partly independent (Drayson et al., (2001), Exp. Cell Res. 266, 126-134). We have investigated whether halting HL60 cells in G1 or S phase influences their commitment to or maturation along the neutrophil and monocyte pathways. Early G1 and S phase cells were isolated separately by elutriation. Quinidine was used to block the cell cycle progression of G1 cells and aphidicolin to greatly retard the progression of S phase cells. Neutrophilic (in response to all-trans-retinoic acid) or monocytic (to 1 alpha,25-dihydroxyvitamin D(3)) differentiation were assessed by induction of CD11b, M-CSF receptor and CD14 expression, acquisition of granulocyte-colony stimulating factor responsiveness, capacities to phagocytose yeast and reduce nitroblue tetrazolium, and down-regulation of CD30 and transferrin receptor expression. The cell-cycle-blocked cells differentiated at normal rates, mostly without incorporating bromodeoxyuridine. These observations establish: (a) that neither transit through the cell cycle nor a cell's position in the cell cycle substantially influences execution of the neutrophilic and monocytic differentiation programs by HL60 cells; and (b) that individual HL60 cells are genuinely bipotent.     

Mitogenic agents induce redistribution of transferrin receptors from internal pools to the cell surface.

Incubation of serum-growth HeLa cells in serum-free medium causes a rapid (t1/2 3 min) 30-60% decrease in the binding of 125I-diferric transferrin to the cell surface. Addition of fetal bovine serum to cells in serum-free medium results in a rapid (t1/2 3 min) and concentration-dependent increase in binding activity. The loss or gain in ligand binding is a result of changes in surface receptor number rather than an alteration in ligand-receptor affinity. A variety of hormones (insulin, insulin-like growth factor, interleukin 1 and platelet-derived factor) were found to mimic the effect of serum on receptor number. The alteration in surface receptor number was found to be calcium-dependent. Changes in surface receptor number were independent of either receptor biosynthetic rate or the absolute cellular content of receptors. The effect of insulin or serum on Hela cell transferrin receptor distribution was unaffected by the presence of transferrin, demonstrating that receptor distribution in this cell type is ligand-independent. The ability of serum or insulin to modify surface transferrin receptor number was also observed in mouse L-cells, human skin fibroblasts, and J774 macrophage tumour cells. However, transferrin receptors on K562 and Epstein-Barr virus-transformed human lymphoblasts were unaltered by these agents. The quantities of receptors whose distribution is predominantly on the surface (i.e. epidermal growth factor or low density lipoprotein receptor) were unaltered by addition of the mitogenic agents. These results extend our previous studies [H.S. Wiley & J. Kaplan (1984) Proc. Natl. Acad. Sci. U.S.A. 81, 7456-7460] demonstrating that mitogenic agents can induce redistribution of receptor pools in selected cell types.     

A granulosa cell differentiation-stage-specific cytotoxic activity in bovine and rat sera

Heat-inactivated serum is cytotoxic to granulosa cells from preantral follicles but not to cells from preovulatory follicles. A dominant feature of the granulosa cells of preovulatory follicles is the presence of luteinizing hormone (LH) receptors on the surface of the cells. In the present study, we have examined the relationship between the process of LH receptor induction and the acquisition of serum tolerance in granulosa cells in vitro. Granulosa cells from the ovaries of immature rats primed with diethylstilbestrol (DES) were cultured in a 1:1 mixture of Ham's F-12 and Dulbecco's modified Eagle's medium containing 30 ng of ovine follicle-stimulating hormone (oFSH; NIH-15). At either 0, 24, or 48 h of culture, heat-inactivated fetal bovine serum (FBS) was added (10% by volume) to separate groups of culture tubes. All cells were cultured for a total of 72 h, at which time the cultures were assessed for LH receptor (specific 125I-human chorionic gonadotropin [hCG] binding) and DNA content. LH receptors were induced in all FSH-containing serum-free cultures by 48 h. Receptors were not induced, however, when serum was added after either 0 or 24 h of culture. Furthermore, serum addition at these times resulted in a cell loss (assessed by DNA) of 40-60%. Serum addition at 48 h to FSH-containing cultures resulted in an inability to detect LH receptors at 72 h and with no significant effect on the culture DNA content. Addition of a protein extract of FBS at the initiation of cell culture prevented FSH-stimulated LH receptor induction and was cytotoxic. A lipid extract of FSH did not interfere with receptor induction and was not cytotoxic.(ABSTRACT TRUNCATED AT 250 WORDS)     

Clonal growth and self-renewal of human myeloid leukemia cell lines (BRM and DD) in serum-free semi-solid culture

Primary and secondary colony formation of two new human myeloid leukemia cell lines (BRM and DD) were studied in serum-free semisolid cultures. The results indicate that bovine serum albumin and transferrin were essential for clonal growth in chemically defined medium. Insulin contributed only moderately beneficial effects. Initial cell density was also a major modulator of plating efficiency. Positive cooperation between the leukemia cells was shown by using autologous conditioned media. This is the first serum-free culture method that allows self-renewal of human myeloid leukemia cell lines in terms of secondary colony formation in methylcellulose cultures.     

Cyclic AMP-dependent and -independent protein kinases and protein phosphorylation in human promyelocytic leukemia (HL60) cells induced to differentiate by retinoic acid

The human leukemia cell line HL60 which resembles promyelocytes can be induced to differentiate to cells displaying features of the mature myeloid phenotype by a variety of agents including retinoic acid (RA) and agents that elevate intracellular adenosine 3:5 cyclic monophosphate (cyclic AMP) levels, e.g., 8-bromo-cyclic adenosine 3:5 monophosphate (8-Br-cyclic AMP), cholera toxin. Since most, if not all the effects of cyclic AMP, are mediated by adenosine 3:5 cyclic monophosphate-dependent protein kinase (cyclic AMP-dPK), we investigated the role of cyclic AMP-dPK and adenosine 3:5 cyclic monophosphate-independent protein kinase (cyclic AMP-iPK) in the induced differentiation of HL60 cells. Marked stimulation of cyclic AMP-dPK and cyclic AMP-iPK appears to be intimately involved with and specific for HL60 myeloid differentiation as evidenced by: (1) Stimulation of cyclic AMP-dPK and cyclic AMP-iPK early during HL60 myeloid differentiation and prior to phenotypic changes. (2) RA and dimethylformamide (DMF), agents that induce differentiation along the myeloid pathway, cause a marked increase in the type I cytosolic cyclic AMP-dPK and cyclic AMP-iPK (protamine kinase) while no such increases are noted in cells treated with 12-0-tetradecanoyl-phorbol-13-acetate (TPA) which induces differentiation along the monocyte/macrophage pathway. (3) Both native polyacrylamide gel electrophoresis as well as photoaffinity labeling with 8-azido-cyclic AMP demonstrate marked increases in type I cyclic AMP-dPK in the cytosols of cells exposed to agents that induce myeloid differentiation but no increase in TPA-differentiated cells. (4) The appearance and disappearance of specific cyclic AMP-dependent and -independent protein phosphorylations are associated with the induced myeloid differentiated state.     

Retinoic acid acylation (retinoylation) of a nuclear protein in the human acute myeloid leukemia cell line HL60

all-trans-Retinoic acid is a potent inducer in vitro of the differentiation of the human acute myeloid leukemia cell line HL60 and of fresh cells from patients with acute promyelocytic leukemia. The recent discovery of nuclear retinoic acid receptors provides a basis for understanding how retinoic acid acts at the genetic level. We have now found that retinoic acid is incorporated into HL60 cells in a form that is not removed by extraction with CHCl3:CH3OH. About 90% of this labeled retinoic acid is trichloroacetic acid-soluble after digestion with proteinase K or after hydrolysis with either NH2OH or CH3OH:KOH under mild conditions. Methyl retinoate is the major product of hydrolysis with CH3OH:KOH. These results are consistent with retinoylation of protein with the formation of an ester, probably thioester, bond. The extent of the retinoylation of HL60 protein is dependent on both time and retinoic acid concentration. A major fraction of the retinoylation is of protein that has a molecular mass of 55 kDa after reduction with dithiothreitol. On two-dimensional gels, the retinoylated protein has a pI of about 4.9 and a molecular mass of 55-60 kDa. These characteristics and its localization in the cell nucleus are consistent with retinoylation of the HL60 nuclear retinoic acid receptor or a closely related protein.