Abstract
Conflicting results have been reported regarding the effect of TNF-α on the growth of human primitive hemopoietic cells. In this study, we have examined the effect of TNF-α on the proliferation of several CD34+/CD38+ (KG-1, TF-1) and CD34+/CD38− (KG-1a, TF-1a) myeloid leukemic progenitor cell lines. Our data show that TNF-α markedly inhibits the growth of these cells in both liquid and soft agar cultures. Addition of GM-CSF or IL-3 does not prevent TNF-α-induced growth inhibition. Flow cytometry analyses of propidium iodide-stained cells demonstrated cell death of all four cell lines, as judged by the presence of cells with hypodiploid DNA content after exposure of cells to TNF-α for 4 days. Annexin V assays detected apoptosis in TF-1, but not in TF-1a, KG-1, and KG-1a cells in terms of translocation of phosphatidylserine shortly after TNF-α treatment. Neutralizing anti-TNF receptor type I (TNFR-I; p55) Ab almost completely reversed TNF-α-induced growth inhibition in both liquid and soft agar cultures, whereas anti-TNFR-II (p75) Ab had only a marginal effect. TNF-α rapidly induced marked activation of nuclear transcription factor NF-κB in all 4 cell lines. The majority of this effect was abolished by the type I receptor Ab, whereas the type II receptor neutralizing Ab had no effect. Our data also show that TNF-α is incapable of inducing activation of the mitogen-activated protein kinase pathway in these leukemic cell lines.
Tumor necrosis factor α is a protein synthesized and secreted by mononuclear phagocytes in response to stimulation with bacterial endotoxin and other agents (1). TNF-α affects a wide range of biological activities of many cell types, including endothelial cells (2), fibroblasts (3), and hemopoietic cells (4, 5, 6, 7). Conflicting results have been reported with regard to the effects of TNF-α on the proliferation of human hemopoietic stem cells and myeloid progenitor cells, from potent inhibition (5, 6, 7, 8) to stimulation (9, 10, 11). Because TNF-α induces the production of a variety of human growth factors, including IL-1, IL-6, M-CSF, G-CSF, and GM-CSF (12, 13, 14, 15), its growth-stimulatory effects have been correlated with the production of these growth factors (16, 17). A number of studies from several laboratories suggest that the direct effects of TNF-α on bone marrow multipotential progenitor cells are only inhibitory (5, 7). In contrast, some recent reports argued that TNF-α is a potential growth stimulator of human CD34+/CD38− cells, or IL-3/GM-CSF-responsive CD34+ bone marrow cells (9, 18, 19). One of the fundamental unanswered questions with respect to TNF-α action on CD34+ cells is whether this conclusion is true for leukemic hemopoietic cells with CD34+/CD38− or CD34+/CD38+. Because leukemic cell lines have been widely used as models for the study of the mechanisms regulating cell growth, an understanding of the role that TNF-α plays in mediating growth of human myeloid leukemic cells might have physiological importance and clinical relevance.
Two TNF receptors, with molecular masses of 55 kDa (TNFRI-p55) and 75 kDa (TNFRII-p75) have been identified and these cDNAs were cloned (20, 21, 22, 23, 24). p55 exclusively mediates the stimulating effects of TNF-α on GM-CSF- and IL-3-induced colony formation derived from human bone marrow cells, as well as the inhibition of committed erythroid progenitor cells and G-CSF-induced colony growth (25, 26), whereas the activities mediated through p75 prominently support the proliferation of T cells (27, 28). Conversely, Heller et al. (29) and Jacobsen et al. (30) propose that p75 mediates TNF-α-induced cytotoxicity/growth inhibition in human HeLa cells and non-T hemopoietic myeloid progenitors, respectively. In an attempt to reconcile these conflicting data and to test the effects of TNF-α on human myeloid leukemic cells, we have investigated the effect of TNF-α on NF-κB activation and proliferation of several CD34+ human myeloid leukemic cell lines. The primary aim of this study was to determine whether TNF-α stimulates or inhibits the growth of these human myeloid leukemic progenitors and to examine the possible role (s) that TNF receptors play in TNF-α-induced biological activities.
Materials and Methods
Reagents
3 and mitogen-activated protein/extracellular signal-related kinase kinase (MEK) Abs were purchased from Promega (Madison, WI) and New England Biolabs (Beverly, MA), respectively. Anti-NF-κB (p50) Ab was purchased from Santa Cruz (Santa Cruz, CA).
Maintenance of leukemia cell lines
TF-1 is a growth factor-dependent human cell line that originally was isolated from the bone marrow cells of a patient with erythroleukemia (31). The growth factor-independent variant (TF-1a) was isolated from TF-1 cells in our laboratory (32). Phenotypically, TF-1a is CD34+/CD38− and TF-1 is CD34+/CD38+. These cells were maintained in RPMI 1640 supplemented with 10% FCS in the presence (for TF-1) or absence (for TF-1a) of GM-CSF (5 ng/ml) at 37°C in humidified air containing 5% CO2. KG-1 (33) and KG-1a (34) cell lines were maintained in IMDM with 20% FCS. They are CD34+/CD38+ and CD34+/CD38−, respectively (35). Media and serum were purchased from Life Technologies (Gaithersburg, MD).
Assays of cell proliferation
Cell proliferation was examined by directly counting cells with a hemocytometer or by indirect colorimetric immunoassay (MTT). Tetrazolium salt MTT is metabolized by NAD-dependent dehydrogenase to form a colored reaction product, and the amount of dye formed directly correlates with the number of cells. Briefly, cells were grown in microtiter plates in a final volume of 100 μl RPMI 1640 or IMDM in the presence or absence of various cytokines. After 24 h incubation, the MTT labeling reagents (10 μl) were added to the cells, and the cells were incubated for 4 h at 37°C in a humidified atmosphere air containing 5% CO2. After solubilizing, the formazan dyes were quantitated using a microtiter plate (ELISA) reader at a wavelength of 550 nm, as recommended by the manufacturer. Proliferation of progenitors was measured by counting colony numbers under an inverted microscope (see below).
Soft agar cultures
Exponentially growing cells were harvested and resuspended in serum-free medium. After cell number was adjusted, the cells were plated into 35-mm petri dishes in a final volume of 2 ml containing 0.3% agar (Difco, Detroit, MI) and 20% FCS in the presence or absence of GM-CSF. Each dish contained 5,000 cells for KG-1, TF-1, and TF-1a cell lines and 10,000 cells for the KG-1a cell line. Colonies containing >50 cells were enumerated with an inverted microscope on day 10 of the culture for the KG-1, TF-1, and TF-1a cells and on day 14 for the KG-1a cells.
Flow cytometric analysis of cells by propidium iodide and annexin V staining
For propidium iodide (PI) staining, cells were incubated in six-well plates in the presence or absence of TNF-α (10 ng/ml) for 3 days, after which the cells were removed from plates, washed twice with PBS, and fixed with 80% ethanol/PBS. Subsequently, the medium was removed by centrifugation, and the pellets were resuspended in 500 μl PBS. The cells were incubated in the dark for 30 min at room temperature in the presence of PI (0.05 mg/ml) and DNase-free RNase A (1 mg/ml). Thereafter, cell cycle status and apoptosis were determined using a FACScan flow cytometer (Becton Dickinson, San Jose, CA). For the annexin V staining assay, cells treated with or without TNF-α were collected and resuspended in 1× binding buffer (0.01 M HEPES/NaOH, pH 7.4, 0.14 mM NaCl, 2.5 mM CaCl2) at a concentration of 1 × 106 cells/ml. Subsequently, 100 μl of the cell suspension was transferred to a 5-ml tube and annexin V (5 μl) and PI (10 μl) were added, as recommended by the manufacturer (PharMingen, San Diego, CA). The cells were incubated at room temperature for 15 min, after which 400 μl of 1× binding buffer was added, and apoptosis was determined by flow cytometry analysis.
Receptor studies
For detection of TNF receptors, cells in log phase were collected, washed three times with PBS contained 0.5% BSA, and resuspended in the same buffer to a final concentration of 4 × 106 cells/ml. Subsequently, 25 μl of cells was transferred to a 5-ml tube for staining with PE-conjugated anti-p55 or p75 receptor Ab for 45 min at 4°C. After this incubation, unreacted Ab was removed by washing cells twice with the PBS buffer. Finally, cells were resuspended in 200 μl of the PBS buffer for flow cytometric analysis. For neutralization studies, the cells in log phase were collected, washed once with PBS, and resuspended in RPMI 1640 or IMDM. The cells then were incubated with either TNF p55 (15 μg/ml) or p75 receptor (20 μg/ml)-neutralizing Ab for 1 h at 37° in a humidified atmosphere containing 5% CO2. The concentrations of the Abs used here were the optimal doses that we determined, to obtain a maximum neutralizing effect. Equal amounts of IgG were added to control cells. Subsequently, liquid and clonogenic cultures were performed for the time noted previously. For the study of NF-κB, cells were first treated with p55 or p75 receptor Ab for 1 h as described above, after which TNF-α (10 ng/ml) were added to cells and incubation was continued for 15–30 min. The activity of NF-κB was then detected by electrophoretic mobility shift assay (EMSA) as described below.
Electrophoretic mobility shift assay
Cells (1 × 107) treated with TNF-α at 37°C for various times were washed in PBS and lysed in buffer A (10 mM HEPES (pH 7.4), 1.5 mM MgCl2, 10 mM KCl, 3 mM DTT, 1% Nonidet P-40, 1 mM vanadate, 1 mM NaF, 1 mM PMSF, 5 μg/ml leupeptin, and 100 μM pepstatin). Nuclear pellets were incubated in a small volume of buffer C (20 mM HEPES (pH 7.4), 1.5 mM MgCl2, 3 mM DTT, 25% glycerol, 450 mM NaCl, 0.2 mM EDTA, with vanadate, NaF, PMSF, leupeptin, pepstatin as in buffer A), and the extracted material was diluted to give concentrations of the buffer ingredients as specified for buffer D (20 mM HEPES (pH 7.4), 60 mM KCl, 3 mM DTT, 19% glycerol, 0.2 mM EDTA with vanadate, NaF, PMSF, leupeptin, and pepstatin as in buffer A) (36). Nuclear extracts were normalized for protein and incubated at 4°C with anti-NF-κB (p50) Ab for 1 h in a buffer containing 10 mM Tris (pH 7.5), 50 mM NaCl, 1 mM EDTA, 1.8 μg/ml salmon sperm DNA, 0.2% Nonidet P-40, and 20 μg/ml BSA. Double-stranded radiolabeled oligonucleotide (1 ng/reaction) containing the κB sequence of the mouse κ light chain gene (37) was added, and the mixture was incubated for 20 min at room temperature. Samples were electrophoresed on 6% polyacrylamide gels containing 25 mM Tris, 22 mM borate, and 0.25 mM EDTA. After electrophoresis, the gels were dried and exposed to x-ray film overnight.
Western blot analysis
Exponentially growing cells were washed free of serum and growth factors and incubated in serum-free RPMI 1640 for 24 h at 37°C in a humidified atmosphere air containing 5% CO2. Before stimulation, the cells were centrifuged and resuspended in serum free medium without growth factors. These cells were then exposed to TNF-α (5 ng/ml) at 37°C for 0–36 h, after which the cells were washed once with cold PBS and lysed in SDS sample buffer (62.5 mM Tris-HCl, 2% SDS, 10% glycerol, 50 mM DTT, 0.1% bromophenol blue). The lysates were cleared by centrifugation at 12,000 × g for 10 min, and the supernatants were transferred to a fresh tube. Lysates in sample buffer were heated at 100°C for 4 min before SDS-PAGE. Proteins were separated by a 10% SDS-PAGE gel and electrophoretically transferred to a nitrocellulose membrane (Amersham, Arlington Heights, IL). The proteins in the membrane were then immunoblotted with anti-phosphorylated MAP kinase or MEK1/2 Ab at 4°C overnight, after which the first Ab was removed, and the blot was washed three times with TBST buffer (20 mM Tris, 137 mM NaCl (pH 7.6), 0.1% Tween 20). Subsequently, the blot was incubated for 1 h with HRP-conjugated anti-rabbit secondary Ab at room temperature, and the expression of MAPK and MEK were detected by the chemiluminescence reaction (New England Biolabs).
Statistical analysis
All results were expressed as the mean ± SD of data obtained from three or more separate experiments. The statistical significance of differences between group means was determined by Student’s t test.
Results
TNF-α inhibits growth of CD34+/CD38− and CD34+/CD38+ leukemic cell lines in liquid culture
Recombinant TNF-α was first evaluated for its effect on the growth of KG-1 (CD34+/CD38+), KG-1a (CD34+/CD38−), TF-1 (CD34+/CD38+), and TF-1a (CD34+/CD38−) cells in liquid culture. Because TF-1 is a factor-dependent myeloid cell line, for all studies this cell line was maintained in culture supported by GM-CSF (5 ng/ml) as described in Materials and Methods. Exponentially growing cells (2 × 105) were collected by centrifugation and resuspended in culture medium containing 5 ng/ml TNF-α. After 4 days, cell proliferation was evaluated by counting cell numbers with a hemocytometer. Addition of TNF-α inhibited the growth of all four cell lines, with 26% (p < 0.05), 37% (p < 0.05), 30% (p < 0.05), and 22% (p < 0.05) reductions being observed in KG-1, KG-1a, TF-1, and TF-1a cells, respectively (Fig. 1⇓). Because it has been shown that the effects of TNF-α (stimulatory or inhibitory) depend on its concentration in culture, which varies with different cell types, we examined the dose-response effect of TNF-α-induced growth inhibition. As shown in Fig. 2⇓, a low concentration of TNF-α (0.1 ng/ml) had no effect on the growth of these leukemic cells. However, growth inhibition was apparent in TF-1 and KG-1a cells exposed to 0.5 ng/ml TNF-α, and a further increase in the amount of TNF-α (2 ng/ml) caused growth inhibition in all four cell lines. The maximal inhibitory effects were observed at a concentration of 2 ng/ml or more. At 10 ng/ml, TNF-α-induced growth inhibitions were 44, 28, 37, and 47% for TF-1, TF-1a, KG-1, and KG-1a cells, respectively. These results demonstrate that TNF-α significantly inhibits growth of these human CD34+ leukemic cells in culture.
TNF-α inhibits growth of CD34+ human myeloid leukemia cells. Cells were seeded in six-well plates in the presence or absence of TNF-α (5 ng/ml). After 4 days of culture at 37°C, the cells were collected and the cell numbers were counted using a hemocytometer under a microscope. Values reported are means ± SD from three independent experiments.
Growth inhibition induced by TNF-α is dose dependent. Cells (2 × 104) were seeded in 96-well microplates containing 100 μl RPMI 1640 (TF-1, TF-1a) or IMDM (KG-1, KG-1a) supplemented with 10% or 20% FBS in the presence or absence of TNF-α (0–10 ng/ml). After 4 days of culture at 37°C, the cells were labeled with MTT labeling agents for 4 h, and cell proliferation was determined by MTT color reaction with an ELISA plate reader at 550 nm, as described in Materials and Methods. Values reported are means ± SD from three to four independent experiments.
TNF-α inhibits growth of CD34+/CD38− and CD34+/CD38+ myeloid leukemic cell lines in clonal culture
Because TNF-α has been reported to be a potent growth stimulator for IL-3 and GM-CSF-induced human CD34+ hemopoietic progenitor cells (9) and very primitive CD34+/CD38− myeloid pro-genitors (19), we conducted soft agar clonal cultures to determine whether TNF-α stimulates or inhibits the growth of the progenitor cells within the KG-1, KG-1a, TF-1, and TF-1a cell populations. First, we tested whether KG-1a cells would grow in soft agar culture, because previous studies suggest that this cell line is unable to form colonies in semisolid media (34). Surprisingly, the KG-1a cells formed colonies in the system we used after 14 days of culture, with a plating efficiency of ∼1.5% in the absence of growth factors. The colony formation in agar was linearly related to the number of KG-1a cells plated (data not shown). Compared with the other three cell lines, KG-1a progenitor cells grew more slowly and formed smaller colonies in soft agar (data not shown). Next, we investigated possible effects of TNF-α on the proliferation of the progenitors within all four cell lines described above. After 10 days (KG-1, TF-1, and TF-1a) or 14 days (KG-1a) in culture, TNF-α (2 ng/ml) markedly reduced their colony numbers, with 38, 63, 73, and 36% inhibition being observed in KG-1, KG-1a, TF-1, and TF-1a cells, respectively (Table I⇓). The responses were dose dependent, with maximal inhibition at concentrations of 2 ng/ml TNF-α or more. A low concentration of TNF-α (0.5 ng/ml) inhibited colony formation of TF-1 and KG-1a cells but not the other two cell lines. Because TF-1a and KG-1a are CD34+/CD38− and TF-1 and KG-1 are CD34+/CD38+ (34), the present studies demonstrate that TNF-α is a potent growth inhibitor not only for the CD34+/CD38+ but also for the CD34+/CD38− myeloid leukemic progenitor cell lines tested. Subsequently, we asked whether some growth factors, such as GM-CSF and IL-3, would affect TNF-α-induced growth inhibition. This study was conducted in TF-1a and KG-1 cells, because both cell lines are factor independent but still respond to GM-CSF and IL-3. As shown in Table II⇓, TNF-α suppressed ∼60% of colony formation in TF-1a cells and 51% in KG-1 cells, respectively, in the absence of any growth factors. Addition of GM-CSF or IL-3 did not reverse TNF-α-induced colony reduction at any concentrations up to 100 ng/ml for each (data not shown).
Inhibitory effects of TNF-α on colony-forming cellsa
Growth factors do not reverse TNF-α-induced growth inhibitiona
TNF-α-induced inhibition of the CD34+ myeloid leukemic cells is a consequence of both apoptosis and necrosis
Our preliminary data suggest that TNF-α does not induce cell cycle arrest in the myeloid leukemic cells tested, although TNF-α-induced G1 arrest has been reported in some types of cell (7). Therefore, the growth inhibition by TNF-α reported here must be either from apoptosis or from necrosis. To clarify this question, we used PI and annexin V staining to examine cell cycle status and apoptosis, the latter as measured by translocation of phosphatidylserine (PS) in the CD34+ myeloid cell lines. Flow cytometry analysis of PI-stained cells demonstrated TNF-α-induced cell killing in all four cell lines, as judged by the presence of the cells with hypodiploid DNA content (sub-G1 population) after exposure of the cells to TNF-α (10 ng/ml) for 4 days (Fig. 3⇓). Because PI staining does not distinguish apoptosis-caused cell death from necrosis-induced cell killing, we used annexin V to determine the presence of PS in the cell membrane shortly after TNF-α treatment. Translocation of PS from cytoplasm to cell membrane is one of the earliest features of apoptosis. PS has a high affinity to annexin V, resulting in an positive annexin V staining. As shown in Fig. 4⇓, as little as 4 h of exposure to TNF-α caused marked apoptosis of TF-1 cells, with a significant population of annexin V-FITC-positive cells (8.2% vs 21.9%) being observed, whereas KG-1a and TF-1a cells show primarily annexin V negative as compared with the cells without TNF-α treatment. Addition of TNF-α to KG-1 cells induced a small level of apoptosis (5.9% vs 8.7%).
Cell cycle analysis by flow cytometry of PI-stained cells. Cells were incubated in six-well plates for 3 days in the presence or absence of TNF-α (10 ng/ml), after which the cells were fixed, stained with propidium iodide, and analyzed as described in Materials and Methods. Similar results were obtained in two more repeated experiments.
Flow cytometric analysis of annexin V staining. Cells were incubated in 24-well plates in the presence or absence of TNF-α for 4 h, after which the cells were collected and resuspended in 1× binding buffer at 1 × 106/ml. Subsequently, 100 μl cell suspension were incubated with 5 μl annexin V and 10 μl PI for 15 min at room temperature as recommended by manufacturer. Thereafter, 400 μl 1× binding buffer were added, and cell death and apoptosis were analyzed by flow cytometry. The annexin V-positive cells are the cells undergoing apoptosis and are present in the lower right quadrant (LR).
TNF-α rapidly induces NF-κB activity in CD34+ myeloid leukemic cells
Because NF-κB activation is known to be a major mediator of TNF-α action (38, 39, 40, 41, 42), we investigated the effect of TNF-α on the activation of NF-κB in these leukemic cell lines. As shown in Fig. 5⇓, incubation of cells for 15 min with as little as 2 ng/ml TNF-α activated transcription factor NF-κB in all four cell lines, as detected by EMSA. The expression of activation was varied with cell lines, with TF-1a cells expressing the lowest NF-κB activity. The NF-κB activation induced by TNF-α appears to be persistent, with NF-κB activation still detectable for 2–3 days in cells treated with TNF-α (data not shown).
TNF-α rapidly induces activation of NF-κB cells treated with or without TNF-α for 15 min at 37°C. Thereafter, the cells were collected and washed once in PBS. Nuclear extracts were prepared and tested for activation of NF-κB, as described in Materials and Methods.
TNF p55 but not p75 receptor predominantly regulates TNF-α-induced apoptosis and activation of NF-κB
To help clarify the functional role of p55 and p75 receptors in TNF-α-mediated growth inhibition we examined the potential involvement of these two receptors in signals initiated by TNF-α binding by using receptor specific neutralizing Abs. First, we examined whether these cells express both p55 and p75 receptors. As show in Fig. 6⇓, both p55 and p75 receptors are expressed on KG-1, KG-1a, TF-1, and TF-1a cells determined by analysis of fluorescent stained cells, using PE-conjugated anti-TNF p55 and TNF p75 receptor Abs. A high level of p75 receptor was detected in all of these myeloid leukemic cell lines, which is in contrast to what is seen in endothelial cells (43, 44). The high expression of p75 receptors may suggest an important role of these receptors in TNF-α-induced signaling that is responsible for growth inhibition. To determine whether this hypothesis is true, we performed neutralization receptor studies. Addition of TNF-α inhibits the growth of these myeloid leukemic cell lines, as described above, by a range of 20–40% (depending on cell types). Cells treated with neutralizing p75 receptor Ab (20 μg/ml) only slightly (<10%, p > 0.05) reduced TNF-α-induced inhibition in TF-1 and KG-1 cells and had barely any effect on TF-1a and KG-1a cells, even at a high concentration; However, neutralizing p55 receptor Ab (15 μg/ml) reversed ∼60–70% of TNF-α-induced inhibition in these human myeloid leukemic cells (Fig. 7⇓). This response is dose dependent. Fig. 8⇓ shows does-response curve of TNF receptor Abs on TF-1 cells. Data obtained with the other three cell lines were similar to the TF-1 cell results (data not shown). Because a number of reports regarding the effect of TNF-α were from the study of human hemopoietic progenitors, we thus investigated the role of TNF receptor on colony-forming cells within these leukemic cell populations. As we expected, TNF-α at a concentration of 5 ng/ml inhibited colony formation by 54, 47, 54, and 61% for TF-1, TF-1a, KG-1, and KG-1a cells, respectively. Cells treated with anti-p55 receptor Ab markedly reduced the TNF-α-induced inhibition of colony growth (by 73–100%). Again, an anti-p75 receptor Ab had only a marginal effect on KG-1a and TF-1a and no effect on the other two cell lines (Table III⇓). TNF-α is a highly pleiotropic cytokine and thus displays multiple cellular responses. It is possible that the lack of a significant role of p75 receptor may be specific only for cell proliferation. To explore this possibility, we examined the activation of NF-κB in response to TNF-α treatment. Incubation of cells for 30 min with 10 ng/ml TNF-α markedly induced NF-κB activation, as detected by EMSA. Preincubation of p75 receptor Ab (1 h before addition of TNF-α) had no effect on TNF-α-induced NF-κB activation, whereas cells treated with p55 receptor Ab almost completely blocked activation of NF-κB induced by TNF-α. Thus, our results indicate that the activation of NF-κB occurs through the TNF p55 receptor and is barely influenced by the p75 receptor (Fig. 9⇓). Taken together, these results suggest that the TNF p55 receptor is primarily responsible for the growth inhibition and NF-κB activation induced by TNF-α.
Expression of TNF receptors in human myeloid leukemic cells. Cells in log phase were collected and resuspended in PBS supplemented with 0.5% BSA. Cells (25 μl) were incubated with PE-conjugated anti-TNF p55 or p75 receptor Ab (10 μl for each) for 45 min at 4°C. After this incubation, cells were washed three times with PBS buffer and resuspended in PBS buffer (200 μl) for final cytometric analysis. As a negative control (CTL), cells were incubated with PE-conjugated mouse IgG at the same concentration. Three independent experiments show identical profiles.
Neutralizing Ab against p55 but not p75 receptor reverses TNF-α-induced leukemic cell inhibition in liquid culture. Cells were pretreated with anti-p55, or anti-p75 neutralizing Ab, or mouse IgG for 1 h at 37°C. Subsequently, TNF-α (10 ng/ml) was added to cells (in the presence of mouse p55, p75 receptor, or IgG) and incubation was continued for 4 days. Thereafter, the cells were collected, and cell proliferation was determined by MTT assay, as described in Materials and Methods. The results shown are means ± SD of three separate experiments.
Dose-response effect of p55 receptor Ab on blocking TNF-α-induced growth inhibition. Cells were pretreated with or without p55 or p75 Ab at the concentrations indicated for 1 h at 37°C, after which TNF-α (10 ng/ml) was added and incubation was continued for 4 days in the presence of p55 or p75 Ab. Thereafter, the cells were collected and washed once in PBS, and cell proliferation was performed by the MTT method. Control cells were treated with or without TNF-α in the presence of mouse IgG. The blocking effect of p55 or p75 neutralizing Ab on TNF-α-induced growth inhibition was calculated as [MTT reading (TBF-α+, Ab+) − MTT reading (TNF-α+, IgG+, Ab−)]/[MTT reading TNF-α+, Ab−, IgG+)]. The differences of MTT reading between the cells treated with and without p55 or p75 Ab in the presence of TNF-α reflects the degree of reversal of growth inhibition by the Abs.
p55 but not p75 receptor neutralizing Ab reverses TNF-α-induced activation of NF-κB. Cells were pretreated with or without p55 or p75 Ab for 1 h at 37°C, after which TNF-α was added and incubation was continued for 30 min in the presence or absence of the Abs. Subsequently, nuclear extracts were prepared and tested for activation of NF-κB (top band) as described in Materials and Methods.
TNF type I (p55) receptors predominantly reverse TNF-α-induced growth inhibition of human myeloid leukemic progenitorsa
TNF-α does not activate the MAPK pathway
The addition of GM-CSF or IL-3 to serum-free cell cultures of the leukemic cells rapidly activates 42- and 44-kDa MAPK in KG-1, TF-1, and TF-1a cells (KG-1a cells are not sensitive to GM-CSF and IL-3), as detected by the appearance of their phosphorylated forms (Fig. 10⇓a). In contrast, TNF-α treatment did not significantly induce the phosphorylation of 42/44-kDa MAPKs in all four leukemic cell lines at any time up to 24 h (Fig. 10⇓, B and C). An Ab that recognizes both phosphorylated and nonphosphorylated forms of the MAPK showed identical 42- and 44-kDa bands at all time points (Fig. 10⇓C, top).
TNF-α does not induce activation of MAP kinase. TF-1, TF-1a, KG-1, and KG-1a cells starved in serum-free medium for 24–48 h were stimulated with GM-CSF for 10 min or TNF-α for 0–60 min, after which the cells were collected by centrifugation, washed once in PBS, and lysed in 1× sample buffer. The expression of MAPK was detected by Western blot, probed with anti-phosphorylated MAPK Ab. A, Cells were treated with GM-CSF(10 ng/ml). B, Cells were treated with TNF-α (20 ng/ml). C, Serum-starved TF-1 cells were stimulated with TNF-α for 0–24 h, after which the cells were collected by centrifugation, washed in PBS, and lysed in 1× sample buffer. The expression of MAPK was detected by Western blot, probed with anti-MAPK (top) or anti-phosphorylated MAPK Ab (bottom), respectively. PCRL, positive control from PMA-stimulated lysate or phosphorylated p44/42 MAPK protein from New England Biolabs.
Discussion
Although a number of reports have demonstrated that TNF-α inhibits the growth of hemopoietic myeloid progenitors in human and murine systems, some other recent studies argued that TNF-α is a growth-stimulatory factor for human primitive myeloid progenitors, such as some isolated CD34+ or with CD34+/CD38− bone marrow cells (9, 18, 19). Because TNF-α can up-regulate the expression of GM-CSF and IL-3Rs in human bone marrow cells, the possibility that the TNF-α-induced growth stimulation was a result of release of growth factors from accessory cells cannot been excluded. TNF receptors lack a “hemopoietic receptor motif” (∼200 aa) that have been found in many other hemopoietic growth factor receptors that are capable of transducing a proliferative signal after they are bound to their ligands (45). In this study, we have clearly demonstrated an inhibitory effect of TNF-α on the growth of several CD34+/CD38+ and CD34+/CD38− myeloid leukemic cell lines. Through the studies of the colony-forming cells in soft agar, we showed here that TNF-α inhibits the progenitor cells not only from CD34+/CD38+ cell lines (KG-1, TF-1) but also from CD34+/CD38− cell lines (KG-1a, TF-1a), supporting an inhibitory model of TNF-α action, which might have important physiological and pathological role and clinical relevance.
In contrast with a previous report that the inhibitory role of TNF-α was a result of retaining normal hemopoietic progenitor cells in G0/G1 (7), we are unable to detect any TNF-α-induced G1 arrest of cell cycle in the myeloid leukemic cells that we tested but demonstrate that the TNF-α-induced growth inhibition was a consequence of cell death, which was judged by the presence of the cells with hypodiploid DNA content after 4 days of incubation with TNF-α at a concentration of 10 ng/ml. To determine whether the cell death was induced by apoptosis or by necrosis, we conducted annexin V staining to detect apoptosis of the cells after treatment with TNF-α. This approach confirmed that TNF-α caused a marked annexin V-positive population in TF-1 and a low level of apoptosis in KG-1 cells, with no significant apoptosis observed in TF-1a and KG-1a cells exposed to TNF-α. Therefore, it is most likely that major cell killing induced by TNF-α in TF-1a, KG-1a, and KG-1 cells was a result of TNF-α-induced necrosis.
In humans and mice, two TNF receptors have been identified, with molecular masses of 55 kDa and 75 kDa. It has been shown that p55 receptors are predominantly expressed in human cells of epithelial origin. Subsequent experiments demonstrate that expression of these two receptors on the cell surface is independently regulated. The cytoplasmic portion of the p55 receptor has a domain ∼80 aa long, referred to as the “death domain” for its role in TNF-mediated cell death, whereas the cytoplasmic portion of the p75 receptor lacks the death domain (46). Even so, p75 receptor also has been implicated in TNF-dependent cell death (30). In addition, the relative roles of the p55 and p75 receptors in the activation of NF-κB is not fully understood. Accordingly, the role of TNF receptors in signaling TNF-mediated growth inhibition remained to be clarified, especially in human hemopoietic cells. In contrast to what was observed on epithelial cells, p75 but not p55 receptors are predominantly expressed in human KG-1, KG-1a, TF-1, and TF-1a cells. Similar finding on HL-60, U937, and K562 cell lines (43) has been reported. It is possible that this type of expression is common for human myeloid leukemic cells, although more broad studies have not been done. This raises a interesting question of whether the expression of TNF receptors paralleled their biological activities, i.e., whether p75 receptor is a major factor that is responsible for TNF-α-induced growth inhibition. Through several independent approaches, we have clearly demonstrated that the inhibitory effects of TNF-α on the leukemic progenitor cells are mediated predominantly through the p55 but not the p75 receptor. This is based on the following evidence: 1) addition of p55 but not p75 neutralizing Ab blocked most of the growth-inhibitory effect induced by TNF-α in the liquid cultures; 2) the p55 receptor but not the p75 receptor neutralizing Ab reverses the majority of TNF-α-induced growth inhibition of colony-forming cells in soft agar culture; 3) treatment of the myeloid leukemic cells with the p55-specific Ab prevented TNF-α-induced activation of NF-κB, whereas the p75-specific Ab only had no blocking effect. The inability of p75 receptor Ab to reverse TNF-α-induced growth inhibition cannot be explained by the notion that p75 Ab has a weaker binding activity than p55 Ab, because the Abs used here are the same as that used for the detection of expression of TNF receptors in the cells where a high level of p55 receptors were detected (Fig. 6⇑). The demonstration by Chainy et al. (39) of a cytotoxic response to the p55-specific TNF mutein is consistent with our results. Other p55 receptor-regulated signals induced by TNF-α have been reported that also display growth-inhibitory responses. In cell lines examined thus far, only anti-TNF-p55 Abs (nonneutralizing) are able to mimic the cytotoxic activity of TNF, even in cell lines expressing both receptors (47, 48). Moreover, it has been shown that the TNF-α-dependent activation of NF-κB is abolished in T lymphocytes isolated from p55 receptor-deficient mice (49). All of these results are against a signaling role of TNF p75 receptor in cell killing and suggest that activation of NF-κB and growth inhibition by TNF-α in a variety of cells is predominantly regulated through TNF p55 receptors. Nevertheless, there are some data showing that the Abs that block binding of TNF to p75 receptor partially inhibit TNF killing. In this study, we also found that anti-p75 receptor polyclonal Ab had a very modest contribution to TNF-α-induced growth inhibition in TF-1a and KG-1a cells. Our results suggest that whereas the binding of TNF to p75 receptor may be important for facilitating cell killing, the activation of p75 receptor and the generation of a p75 receptor signal are unlikely be involved. The role of the apparently high levels of p75 receptor in these myeloid leukemic cells is an unresolved issue. Tartaglia et al. (48) suggest that p75 receptor regulates the rate of TNF associated with p55 receptor, possibly by increasing the local concentration of TNF at the cell surface through rapid ligand association and dissociation. It is also possible that both receptors are capable of activating signal-transducing pathways that ultimately lead to cell death. However, signals from the TNF p55 receptor alone can result in cell death, whereas signals from the p75 receptor alone are insufficient.
Involvement of NF-κB in TNF-α-mediated apoptosis and HIV-induced T cell apoptosis has been suggested (42). More recently, NF-κB has been shown to play a critical role in preventing cell death from TNF-α-induced apoptosis (50, 51, 52). The results reported here demonstrate that TNF-α is able to induce activation and nuclear translocation of nuclear NF-κB in all four myeloid leukemic cell lines that we have tested. The activation is rapid, dose dependent and persistent, at least for 2–3 days (data not shown). Although the activation of NF-κB parallels the TNF-α-induced growth inhibition, whether TNF-α-induced NF-κB is directly involved in apoptosis/necrosis in our system remained to be explained. It is well known that NF-κB is a critical regulator of TNF-α-inducible gene expression (53). Genes regulated by nuclear NF-κB include cytokines, chemokines, and leukocyte adhesion molecules. It is possible that any of these NF-κB-induced genes could play a role in cell death induced by TNF-α (54).
Although the present report supports previous studies proposing p55 receptor modulation as one potential mechanism of TNF-α-induced inhibition of hemopoiesis, the pathway beginning with p55 receptor is not clear. Thus, we examined possible effects of the MAPK pathway because down-regulation of this pathway has been suggested to play a role in growth factor withdrawal-induced apoptosis in rat PC 12 pheochromocytoma cells (55). Our data suggest that the MAPK pathway is not involved in TNF-α-induced growth inhibition/apoptosis in the human myeloid leukemic cells tested, which was demonstrated by applying two separate approaches. First, by using anti-MAPK and anti-phosphorylated MAPK Abs we were unable to detect any activation of MAPK after TNF-α treatment at any time point up to 24 h. Second, GM-CSF and IL-3, which are well known as activators of the MAPK signal transduction pathway, had no effect on TNF-α-induced growth inhibition. Thus, the precise mechanism(s) of activation of the p55 receptor and the cytoplasmic signaling pathway for TNF-α-induced growth inhibition in these human myeloid leukemic cells remains to be elucidated.
In summary, our results demonstrate that: 1) TNF-α-induced growth inhibition in human CD34+ myeloid leukemic progenitor cells is a consequence of apoptosis and necrosis. Treatment with TNF-α induces apoptosis in TF-1 cells, whereas the cell killing that occurs in KG-1, KG-1a, and TF-1a mostly results from TNF-α-induced necrosis. There are no such reports on these cells previously; 2) the inhibitory effect of TNF-α is almost exclusively through its interaction with the p55 (type I) receptor; and (3) p75 (type II) receptors play little, if any role in the inhibitory effects of TNF-α on leukemic cell growth, even though these cells express high levels of p75 receptors.
Acknowledgments
We thank Jodi Kroeger and Matt Morrow for the FACS analysis.
Footnotes
-
↵1 This work was supported by National Cancer Institute Grants CA56072 and P30 CA76292.
-
↵2 Address correspondence and reprint requests to Dr. Xiaotang Hu, Division of Medical Oncology and Hematology, University of South Florida College of Medicine, H. Lee Moffitt Cancer Center and Research Institute, 12902 Magnolia Drive, Tampa, FL 33612. E-mail address: hu{at}moffitt.usf.edu
-
↵3 Abbreviations used in this paper: MAPK, mitogen-activated protein kinase; MEK, mitogen-activated protein/extracellular signal-related kinase kinase; PI, propidium iodide; EMSA, electrophoretic mobility shift assay; PS, phosphatidylserine.
- Received March 19, 1999.
- Accepted July 6, 1999.
- Copyright © 1999 by The American Association of Immunologists
References
In this issue
