Jun N-terminal kinase promotes proliferation of immature erythroid cells and erythropoietin-dependent cell lines

Erythropoietin (EPO) is the glycoprotein hormone necessary for the production of mature erythroid cells, a process known as erythropoiesis. During erythropoiesis, erythroid progenitors at the erythroid colony-forming unit (CFU-e) and proerythroblasts stage of development respond to EPO with increased proliferation and survival; withdrawal of these cells from EPO results in apoptosis or programmed cell death.^1-3  EPO exerts its effects on erythroid progenitors by binding to its receptor (EPOR), which then directs the appropriate cellular response. The EPOR exerts its effects via activation of cytoplasmic kinase pathways such as the Janus kinase JAK2/signal transduction activators of transcription-5 (STAT5)⁴  and phosphatidylinositol-3 kinase (PI3-kinase),⁵  as well as activation of the adaptor protein SHC,^6-8  the Src homology 2–containing inositol phosphatase (SHIP),^9,10  and the MAP kinase pathways Jun N-terminal kinase (JNK),^11,12  extracellular signal–related kinase (ERK),^13,14  and p38.^11,15,16  Activation of ERK has been associated predominantly with proliferation, whereas the role of JNKs in cellular processes is more complex. JNK1 was first identified as a serine/threonine kinase that phosphorylated the activator protein 1 (AP1) family member cJun in response to ultraviolet light; these phosphorylation events increased the transactivation potential of cJun.¹⁷  AP1 activity has long been associated with cell-cycle progression and tumor promotion and proliferation.¹⁸  More recently, increases in c-jun expression have been implicated as both positive and negative regulators of apoptosis. AP1 has been implicated in the induction of apoptosis observed when growth factors are removed from growth factor–dependent cells.¹⁹  However, AP1 activation has also been implicated in the repression of apoptosis in growth factor–dependent cells,²⁰  and expression of c-jun delays apoptosis induced in murine erythroleukemia cells, indicating that c-jun may be cytoprotective in erythroid cells.²¹  Therefore, AP1 may exhibit diverse effects on the regulation of apoptosis.

There are 3 isotypes of JNK that have been identified: JNK1¹⁷  and JNK2,²²  which are expressed in a wide variety of tissues, and JNK3, which is expressed primarily in neural tissues.^23,24  JNK proteins phosphorylate serines 63 and 73 of cJun and increase the ability of cJun to activate transcription.^17,25  Although JNK proteins were first identified as proteins that were activated by stress and apoptosis-inducing agents,¹⁷  it is now clear the JNKs have very diverse roles in the regulation of cell proliferation and survival. JNK is known to be activated by a number of growth factors that promote proliferation and survival of hematopoietic cells, including EPO and stem cell factor (SCF),^12,26  as well as thrombopoietin, interleukin-3, and granulocyte macrophage–colony-stimulating factor. The role of JNK activation on growth factor–related signals is not, however, clear. JNK has been shown to promote proliferation in response to platelet-derived growth factor²⁷  and during liver regeneration.²⁸  Targeted disruption of JNK1 results in a defect in T-cell differentiation²⁹  and in T-cell–mediated immunity.³⁰  Knockout of both JNK1 and JNK2, by contrast, results in embryonic lethality between days 11 and 12 of gestation.³¹  It is apparent, therefore, that JNK activation may be an inducer of apoptosis or may protect from apoptosis depending on the cell system. A role for JNK in EPO-induced and stress-induced erythroid differentiation of murine cell lines has been proposed.^15,16  A clear role for JNK activity in the proliferation and survival signals in erythropoiesis has not yet been elucidated.

Our laboratory uses the murine erythroleukemia cell line HCD57 as a model system to study the regulation of EPO-induced erythroid cell proliferation and survival. HCD57 cells require EPO for proliferation and survival; removal of EPO from these cells induces apoptosis. We have previously demonstrated that EPO induces sustained activation of JNK1 and JNK2 in HCD57 cells.¹¹  We report here on the activation of JNK in murine and human erythroleukemia cells and in primary erythroid cells, which indicates an important role for JNK signaling in EPO-induced proliferation.

SP60125 inhibitor, bromodeoxyuridine (BrdU), propidium iodide, MTT reagent (3-(4,5-Dimethyl-2-thiozol)-2,5-diphenyl-2H-tetrazolium bromide), and cytosine arabinoside (ara-C) were purchased from Calbiochem (La Jolla, CA). The SP600125 inhibitor was reconstituted in dimethyl sulfoxide (DMSO; Fisher Scientific, Piscataway, NJ) to make a 20-mM stock solution. Phospho-specific antibodies against JNKs (Thr183/Tyr185), ERKs (Thr/Tyr204), c-jun (Ser 63/Ser73), AKT (Ser473), and p38 (Thr180/Tyr182) were obtained from Cell Signaling Technologies (Beverly, MA). The antibodies that recognize both the phosphorylated and nonphosphorylated forms of JNK1 (C-17), ERK, and p38 were obtained from Santa Cruz Biotechnologies (Santa Cruz, CA). Fluorescein isothiocyanate (FITC)–anti-BrdU was purchased from BD Pharmingen (San Diego, CA).

HCD57 cells were cultured in Iscove modified Dulbecco medium (IMDM; Invitrogen, Carlsbad, CA), 30% fetal calf serum (FCS; Hyclone, Logan, UT), 1 U EPO, 10 μg/mL gentamicin (Invitrogen) and maintained in 2 U EPO/mL (EPOGEN; Amgen, Thousand Oaks, CA) of media unless otherwise noted. UT7-EPO cells were cultured in IMDM, 10% FCS, 10 μg/mL gentamicin, 1 U EPO/mL. TF-1 cells were cultured in RPMI 1640 media (Invitrogen), 10% FCS, 10 μg/mL gentamicin, 1 U EPO/mL. All cells were cultured at 37° C in a 5% CO₂ environment. For the SP600125 inhibitor studies, SP600125 or 0.1% DMSO vehicle was added to cells cultured in EPO for the times indicated in the figure legends. For EPO deprivation studies, the cells were washed 3 times in media without serum or growth factors and incubated in IMDM, 30% FCS, 10 μg/mL gentamicin with no EPO for 4 hours prior to stimulation with 10 U EPO/mL for the times indicated in the figure legends. Primary murine erythroid progenitors infected with the anemia-inducing strain of the friend virus (FVA) were isolated as previously described³²  and 10⁶ FVA cells per time point were cultured in either the presence or absence of 1 U EPO/mL for 0, 1, 2, 4, 8, or 12 hours prior to lysis in 1X sample buffer (0.05 M Tris, pH 8; 2% sodium dodecyl sulfate; 0.1% bromophenol blue; 10% glycerol; 10% β-mercaptoethanol). The protein extracts were prepared in the laboratory of Dr Maurice Bondurant at the Vanderbilt University Medical Center. Primary erythroid progenitors from phenylhydrazine chloride–treated mouse spleen were isolated as follows: CByD2F1/J mice (Jackson Laboratories, Bar Harbor, ME) were injected intraperitoneally on days 0 and 1 with a sterile solution of 6 mg/mL phenylhydrazine chloride to achieve a dose of 60 mg/kg body weight. On day 4, spleens were isolated and a single-cell suspension was generated using a 70-μM cell strainer. The cells were cultured in alpha-mem media (Invitrogen), 2% FCS in either the presence or absence of 10 U EPO/mL for the times indicated. For Figure 7C, the spleen cells were cultured in the absence of EPO for 3 hours prior to stimulation with EPO. For isolation of murine bone marrow cells, C57BL/6 mice (Jackson Laboratories) were euthanized at 6 to 8 weeks of age by CO₂ asphyxiation, and femurs were removed. Bone marrow was extracted in 5 mL IMDM, 10% FCS medium using a 23-gauge needle. Cells were enumerated with trypan blue and plated at the desired density.

To create the pTRE-JNK1dn plasmid, the human JNK1 gene containing alanine and phenylalanine substitutions at threonine (183) and tyrosine (185) was cloned into the pTRE2pur vector (BD Biosciences Clontech, Palo Alto, CA) that was converted to a Gateway destination vector using the Gateway Vector Conversion kit (Invitrogen) to create the destination vector pTRE2Pur-JNK1dn. This vector was transfected into HCD57^TET cells (described previously³³ ) using the Gene pulser electroporator (Bio-Rad, Hercules, CA). A 0.4-mM cuvette was used and the electroporation conditions were 250 V, 950 microfarads (μF), and 50 Ω resistance. Clones were isolated by limiting dilutions in IMDM, 30% FCS, 1 U EPO/mL, 2 μg/mL doxycycline (Sigma, St Louis, MO), 50 μg/mL geneticin, 0.5 μg/mL puromicin. However, the expression of the JNK1dn was leaky to the extent that all experiments were simply carried out in the absence of doxycycline.

Total RNA was isolated from HCD57^TET and HCD57-JNKdn cells using the RNEasy mini RNA isolation kit (Qiagen, Valencia, CA). Expression of the human JNK1dn was detected by Northern blot analysis as previously described³³  using the full-length human JNK1dn cDNA as a probe.

For each time point, 5 × 10⁶ HCD57 cells were used. For Western blot analysis, following treatment of the cells under the different conditions, HCD57, UT7-EPO, or TF-1 cells were lysed in 200 μL 1X sample buffer and sonicated for 10 seconds each to shear the genomic DNA. Equal volumes (20 μL) of sample were electrophoresed on an 8.5% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) gel and subjected to Western blot analysis with the phosphospecific antibodies JNK, ERK, and AKT as previously described.¹¹  The blot was stripped as previously described³⁴  and reprobed with an antibody to JNK-1 to insure equal loading of proteins. To obtain quantitative results for Western blots, the blots of the phosphorylated proteins were scanned and quantitated using the Scion Image for Windows program (Scion, Frederick, MD), and corrected against the total protein observed in the reprobed Western blot. For the in vitro kinase assays, 5 × 10⁶ cells per sample were incubated in EPO in the absence or presence of SP600125 or DMSO vehicle for the times indicated in the figure legends. Total cell extracts were immunoprecipitated with rabbit-anti-JNK-1 (C-17) and subjected to an in vitro kinase assay as previously described.³⁵ 

Proliferation of HCD57 cells was determined by MTT assay and by counting cells at 24-hour intervals as previously described.³⁵  For the MTT assay, 2.5 × 10⁴ cells in triplicate were washed 3 times as described above and then incubated for 24, 48, 72, and 96 hours with no additional growth factor or 1 U EPO/mL with or without SP600125 or DMSO vehicle as indicated in the figure. Proliferation is shown as an increase in the optical density at 540 nanometers (OD₅₄₀). Cell counts were determined by counting cells on a hemocytometer in the presence of 0.2% trypan blue. Proliferation is indicated as a percentage of the starting number of cells. For BrdU staining, HCD57 cells and HCD57-JNK1dn cells were plated at 1 × 10⁵ cells in 9 wells and cultured for 48 hours either with no inhibitor or in the presence of 6.25 μM or 12.5 μM SP600125 or 0.1% DMSO vehicle. BrdU was added to the cells to a final concentration of 10 μM, and the cells were incubated for 30 minutes at 37° C. Three samples of each cell treatment were collected immediately, and 6 samples were washed 2 times in DMEM and then cultured in fresh media for 6 and 9 hours either in the absence or in the presence of inhibitors. All samples were fixed in 100 μL 70% EtOH and stained with FITC–anti-BrdU and propidium iodide according to the BD Pharmingen specifications for the anti-BrdU antibody. Cell staining was analyzed using a FACSscan flow cytometer (Becton Dickinson, Franklin Lakes, NJ). When indicated, the P value is shown as the result of a 2-tailed, type 1 Student t test of 3 independent samples.

Apoptosis and cell-cycle determination of HCD57, TF-1 and UT7-EPO cells were detected using flow cytometry analysis of propidium iodide–stained cells as previously described.¹¹  Following staining, cells containing sub-G₀/G₁ DNA (indicating apoptosis), G₀/G₁, S phase, and G₂/M were gated and shown as a percentage of the total number of cells. Samples were plated in triplicate and error bars represent standard deviation.

Assessment of murine CFU-e and BFU-e colony formation was performed as follows: 6 × 10⁵ of total murine bone marrow cells were added to 3 mL methylcellulose media (Methocult M3334; Stem Cell Technologies, Vancouver, BC, Canada) containing 3 U EPO/mL but no other cytokines in the presence or absence of 0.1% DMSO vehicle or 6.25 μM, 12.5 μM, or 25 μM SP600125. A quantity of 1.1 mL of these cells was plated in duplicate onto 35-mM plates and cultured at 37° C in a moist CO₂ environment. To test whether the SP600125 could inhibit cytosine arabinoside (ara-C)–induced cell death, 6 × 10⁵ total bone marrow cells were cultured for 6 hours in IMDM, 25% FCS, 1 U EPO/mL, 10 ng SCF/mL, and 1% WEHI cell–conditioned media (a source of IL-3) in the presence of 12.5 μMor25 μM SP600125 (or DMSO vehicle) followed by the addition of ara-C to a final concentration of 10 μM (or no addition) for 1 hour. The cells were then washed twice in IMDM and plated onto the methylcellulose M3334 media containing EPO but no other growth factors or inhibitors as indicated above. CFU-e's and BFU-e's were counted 2 and 8 days after the start of the experiment, respectively. Data are presented as the number of BFU-e's or CFU-e's per 35-mm plate.

To test if JNK was necessary for proliferation and survival of HCD57 cells, these cells were treated with the recently described pharmacologic inhibitor of JNKs, SP600125.³⁶  SP600125 treatment resulted in a dose-dependent inhibition of JNK activity (Figure 1A). Approximately 90% of total JNK activity was inhibited by a concentration of 12.5 μM SP600125 (Figure 1B). By contrast, the phosphorylation of related MAP kinases ERK1/2 and p38 was unaffected by treatment with the SP600125 inhibitor (Figure 1C-D, respectively).

The biologic effects of the SP600125 inhibitor on the HCD57 cells were tested next. Proliferation was tested by a standard MTT dye reduction assay and counting the increase in number of cells visually using a hemacytometer over several days. Treatment with the SP600125 inhibitor resulted in a dose-dependent inhibition of proliferation of HCD57 cells irrespective of the method used (Figure 2A-C). Proliferation was inhibited by approximately 50% using 12.5 μM SP600125, the concentration of inhibitor required to significantly inhibit JNK activity and activation.

To confirm that the antiproliferative effects of the SP600125 inhibitor were due to delayed cell-cycle progression, pulse-chase labeling of cells with bromodeoxyuridine (BrdU) was performed. The rate at which BrdU⁺ cells progress into G₁ (G₁ entry) provides an indication of the rate of transit through S, G₂, and M. Flow cytometry analysis of cells stained with both propidium iodide and BrdU revealed that BrdU⁺ cells in the SP600125-treated cells entered G₁ more slowly than vehicle or untreated cells (Figure 2C). A 60% reduction in the number of BrdU⁺ cells that entered G₁ was observed in cells treated with 12.5 μM SP600125 (Figure 2Di). Similarly, an estimation of the relative time required to complete S phase is provided by the rate at which BrdU⁺ cells are depleted from the S-phase pool (S exit). Cells treated with SP600125 exhibited a delayed exit from S phase (Figure 2Dii). SP600125 treatment also resulted in an increase of the fraction of cells in G₂/M (Figure 2Diii).³⁶  Therefore, the inhibition of JNK activity correlated with a reduction of the rate of proliferation to approximately half that of control.

Apoptosis of HCD57 cells in the presence of SP600125 was assessed by the accumulation of apoptotic (sub-G₀/G₁) DNA of propidium iodide–stained cells using flow cytometry analysis. At 12.5 μM, the concentration necessary to inhibit JNK activity, 2% to 5% of HCD57 cells were apoptotic 24 hours after treatment with the SP600125 inhibitor (Figure 2E), whereas JNK activity was inhibited 6 hours after treatment with the inhibitor (Figure 1, lane D). Higher concentrations (≥ 25 μM) of SP600125 induced significant apoptosis immediately following addition of the inhibitor (Figure 2E, 25 μM SP600125): approximately 23% and 24% apoptosis was observed 24 hours and 48 hours, respectively, after addition of 25 μM inhibitor (Figure 2E, 25 μM, hatched column). The flow cytometry analysis of propidium iodide–stained HCD57 cells also revealed that treatment of these cells with either 12.5 μM or 25 μM SP600125 caused a decrease in the number of cells in G₀/G₁ and an increase in the percentage of cells in G₂ (Table 1).

The role of JNK in EPO-dependent proliferation and survival was further investigated in the human EPO-responsive cell lines TF-1 and UT7-EPO. EPO induced the activation of JNK1 (as assessed by phosphorylation of threonine 183 and tyrosine 185, which correlates with JNK1 activation¹⁷ ) in both UT7-EPO and TF-1 cells (Figure 3A, lanes 4 and 6). Whereas weak activation of JNK1 and JNK2 was apparent 5 minutes after EPO addition in HCD57 cells (Figure 3A, lane 2), only very weak activation of JNK2 was detected in UT7-EPO cells (Figure 3A, lane 6, top arrow), and only with an overexposure of this blot was this weak activation of JNK2 detected in TF-1 cells (data not shown). Treatment of both UT7-EPO and TF-1 cells with SP600125 inhibited EPO-induced proliferation by 60% at a concentration of 12.5 μM inhibitor (Figure 3B, left panels) with no significant increase in apoptosis (Figure 3B, right panels). Flow cytometry analysis of UT7-EPO and TF-1 cells treated with 12.5 μM SP600125 also revealed an increase in the percentage of cells in G₂/M (data not shown).

In order to independently verify that JNK activity is essential to EPO-induced proliferation, we stably transfected a dominant-negative form of the human JNK1 (JNK1dn) into HCD57 cells. Northern blot analysis revealed the high expression of the human JNK1dn gene (Figure 4A, lane 2). JNK1dn suppressed endogenous JNK activity as indicated by an in vitro kinase assay (Figure 4B, lane 2). Suppression of JNK1 activity resulted in an approximately 50% inhibition of proliferation by both MTT assay as compared with control cells (Figure 4C). BrdU pulse-chase labeling of HCD57 and HCD57-JNKdn cells revealed that the presence of the JNKdn caused a delay of S phase–labeled cells progressing into G₁ (Figure 4Di) and a delay in S-phase exit (Figure 4Dii). However, no increase in the fraction of cells in G₂/M was detected in the HCD57-JNK1dn cells (Figure 4Diii). Flow cytometry analysis of propidium iodide–stained cells revealed that 8% of the cells expressing JNK1dn underwent apoptosis, which is the same percent of apoptosis detected in the control (data not shown), suggesting that in the presence of EPO, JNK contributes to proliferation but not EPO-induced cell survival. Taken together, the above results suggest that inhibition of JNK activity inhibits EPO-induced proliferation of HCD57 cells.

The proliferative effects of EPO-induced JNK activity in the leukemic HCD57 cells led us to investigate the ability of EPO to induce JNK activity in primary erythroid cells. JNK activity was assessed in primary erythroid progenitors isolated from the spleens of mice injected with phenylhydrazine chloride.³⁷  Approximately 95% of the spleen cells isolated were erythroblasts as assessed by flow cytometry of transferrin receptor and TER-119 double-positive–staining cells (data not shown). An examination of JNK activity immediately following isolation of splenic primary erythroid progenitors isolated from phenylhydrazine-treated mice revealed an initial burst of JNK activation immediately following initiation of culture of the cells in IMDM media and fetal calf serum (Figure 5A, lanes 2 and 3) that was not suppressed by the addition of EPO (thus ruling out JNK activation due to withdrawal of EPO as the cause of this activation; Figure 5A, lanes 6 and 7). When the spleen cells were deprived of EPO for 3 hours, constitutive JNK1 and JNK2 activation was detected (Figure 5B, lane 1); however, there was no augmentation of JNK phosphorylation with the addition of EPO, whereas both EPO-dependent ERK and AKT phosphorylation were detected (Figure 5B, lanes 2-5). This experiment was repeated in the absence of serum or in the presence of 20% serum; the presence or absence of serum had no effect on the phosphorylation of JNK in these cells (Figure 5C). The induction of JNK activation during initiation of culture is therefore not a serum effect but may be the result of stress induced by manipulation of these cells. We observed similar JNK activity in primary murine erythroid progenitors infected with the anemia strain of the Friend spleen focus-forming virus (FVA cells), which undergo terminal differentiation within 48 to 72 hours after treatment with EPO; constitutive JNK activation was detected, but EPO did not augment this activation (data not shown). Whereas JNK activity was detected in both primary murine cells, this activity, at least under tissue-culture conditions, appeared to be EPO-independent.

To further test the function of JNK in primary erythroid cells, the effect of the SP600125 inhibitor on the development of primary murine erythroid progenitors was tested. Murine bone marrow cells were assayed for CFU-e's and BFU-e's in the absence or presence of increasing amounts of the SP600125 inhibitor in semisolid media. Morphologically, the CFU-e's were identical between the vehicle-treated and SP600125-treated cells in number, color, size, and shape (Figure 6B). Treatment of cultures of bone marrow cells with SP600125 resulted in markedly smaller BFU-e's (data not shown). The presence of the SP600125 inhibitor caused a dose-dependent reduction in the number of BFU-e's that developed; a 5-fold reduction in BFU-e's was observed at a concentration of 25 μM SP600125 (Figure 6A).

To test if the SP600125 treatment inhibited the proliferation of BFU-e's versus inducing apoptosis, we tested the ability of the SP600125 to protect BFU-e's and CFU-e's from apoptosis induced by the nucleoside analog ara-C. Because ara-C induces cell death by interfering with DNA synthesis and causing DNA damage, it is toxic to dividing cells, and conditions that inhibit the proliferation of cells protect them from apoptosis induced by ara-C (reviewed in Grant³⁸ ). Pretreatment of the bone marrow cells with 12.5 μM or 25 μM SP600125 significantly inhibited the ability of ara-C to induce apoptosis of the BFU-e's (Figure 6C), indicating that the number of BFU-e's in S phase of the cell cycle was reduced by the SP600125. ara-C also inhibited the formation of CFU-e's by about 50%, but the SP600125 had no effect on the ability of the ara-C to kill the cells (data not shown). Therefore, whereas the inhibition of JNK activity did not inhibit the development or cell-cycle progression of the more mature CFU-e's, the proliferation of the less mature erythroid progenitors was greatly reduced by the inhibition of JNK.

Although first identified as a stress-related kinase that was associated with the induction of apoptosis, JNK has recently been shown to be important in a wide variety of cellular processes, including protection from apoptosis,^39,40  growth factor–induced proliferation,²⁸  and cell-cycle progression.⁴¹  We have previously demonstrated that JNK was activated in the EPO-dependent HCD57 erythroid cell line following treatment with EPO for 4 to 24 hours.^11,35  We observed here that JNK could also be phosphorylated immediately following stimulation with EPO (Figure 4, lane 2), suggesting that in addition to the long-term phosphorylation of JNK by EPO-dependent synthesis of autocrine factors such as TNF-α,³⁵  the EPOR may directly activate the JNK pathway in HCD57 cells. The fact that the SP600125 inhibitor blocked EPO-induced proliferation of several cell lines in a dose-dependent manner suggests that JNK may play an important role in EPO-induced proliferation. The slowing of cell-cycle progression in both the SP600125-treated cells and the JNK1dn-expressing cells observed by BrdU pulse-chase experiments suggests that inhibiting JNK activity slowed the rate of proliferation of the cells. The 50% inhibition of proliferation resulting from either the SP600125 inhibitor or the JNK1dn mutant, however, suggests that signaling pathways other than JNK must also be active for maximum proliferation to occur. The lack of increase in the fraction of cells in G₂/M in the JNK1dn-treated cells raises the question of whether the SP600125 effect to block exit from G₂/M is a JNK-specific rather than a JNK-nonspecific effect. cJun and AP1 have long been demonstrated to be necessary for G₁ to S-phase progression,^33,42-44  and JNK has been shown to assist in this process by affecting the expression and activity of cyclin D1.²⁸  Recent evidence suggests, however, that JNK also has AP1-independent cell-cycle effects following G₁, including phosphorylation of JNK at the G₂/M phase of the cell cycle⁴¹  and association of JNK with the centrosomes of cycling cells.⁴⁵  The disruption of these JNK activities could explain the increase of cells in G₂/M observed in the HCD57, TF-1, and UT7-EPO cells treated with the SP600125 inhibitor. The SP600125 inhibitor has been shown to cause G₂ arrest and/or endoreduplication in a number of cell lines, including multiple myeloma cells, breast cancer cell lines MCF-7 and MDBA-MB-231, and prostate cancer cell lines LNCAP and PC-3, in addition to others.^46-49  However, in one of these reports, inhibition with antisense JNK oligonucleotides inhibited the proliferation of fibroblasts with no reported G₂ arrest.⁴⁶  We have observed the presence of multiple nuclei indicative of endoreduplication by staining SP600125-treated HCD57 cells with DAPI but did not observe this effect in the JNK1dn cells (data not shown). Therefore, it is possible that the endoreduplication and reduction in exit from G₂/M may be a non–JNK-related effect of the SP600125 inhibitor. Whereas the current study clearly shows that JNK activity is needed for the maximum rate of proliferation of 3 unrelated EPO-dependent erythroid cell lines, the role of JNK in survival is less clear; although JNK activity was not required for survival of the human leukemia cell lines studied, high concentrations of the JNK inhibitor resulted in apoptosis of approximately half of the HCD57 cells. The lack of increased apoptosis in cells expressing the JNK1dn suggests, however, that SP600125 may induce apoptosis at high concentrations through a JNK-independent mechanism. It is likely, therefore, that JNK does not play a significant role in EPO-induced survival of these erythroid cells.

This study suggests JNK activity may have a role in the proliferation of primary murine erythroid cells. The most compelling evidence that inhibition of JNK activity inhibits the proliferation of BFU-e's and does not simply cause massive cell death is the result that the SP600125 inhibitor protected the BFU-e's from cell death induced by ara-C. The lack of effect of SP600125 on the development of CFU-e's also rules out both a negative effect of the inhibitor on late differentiation events and a general toxicity of this compound toward erythroid cells. Therefore, it appears that JNK activity may play a role in the proliferation of immature erythroid progenitors but not CFU-e's and proerythroblasts.

Our experiments do not shed any light on what hormones or factors might trigger JNK activity in murine BFU-e's. Indeed, the data presented here do not show any ability of EPO to activate JNK in more mature primary murine erythroid cells but demonstrate that all these cells apparently express some level of constitutive JNK activity. In addition, the result that pretreatment of bone marrow cells with SP600125 for only 6 hours protected the BFU-e's from ara-C–induced death suggests that the cells that are affected by the SP600125 are early BFU-e's, which do not respond to EPO. Several questions remain, however, as to the role of JNK in hematopoiesis. For instance, homozygous deletion of either JNK1 or JNK2 does not have any apparent defects in hematopoiesis^30,50  and, although lethal at the embryonic stage, deletion of both JNK1 and JNK2 does not affect the development of fetal liver.³¹  Redundancy between the JNK isoforms may compensate for the loss of either JNK1 or JNK2. Differences in the proliferation of immature erythroid cells that depend on JNK1 and/or JNK2 may become more apparent during conditions of anemia in the JNK1^–/– or JNK2^–/– mice, as has been observed for STAT5a/b knockout mice^51,52 ; the effect of knockout of JNK1 or JNK2 on the development of BFU-e's has not been tested. The embryonic lethality of the JNK1^–/–/JNK2^–/– mice does not address whether one or the other may be required for adult erythropoiesis.

In summary, our data support a growing body of evidence that JNK is an important regulator of erythropoiesis in both primary and leukemia cells. In EPO-dependent leukemia cells, this JNK activity was manifested as EPO-induced proliferation, whereas primary erythroid progenitors may require JNK for the early proliferation of erythroid cells. Further studies on the function of JNK activity during normal erythropoiesis may shed light on the role of this important signal transduction pathway in erythropoiesis and in the development of the erythroleukemic phenotype.

The authors would like to thank William Steelman, Mousumi Sarkar, and Kwan-ho Roh at Virginia Commonwealth University for their technical assistance, Dr John Ryan at Virginia Commonwealth University for his assistance with the flow cytometry studies, and Dr Maurice Bondurant at Vanderbilt University for supplying extracts from FVA cells.