[Cell Cycle 2:5, 467-472; September/October 2003]; ©2003 Landes Bioscience
Report
An Intact NF-κB Pathway is Required for Histone Deacetylase InhibitorInduced G1 Arrest and Maturation in U937 Human Myeloid Leukemia Cells
Received 05/09/03; Accepted 06/15/03
is t
INTRODUCTION
sc
This work was supported by awards CA63753, CA93738, and CA83705 from the NIH, and award 6045-03 from the Leukemia and Lymphoma Society of America.
nc
leukemia, histone deacetylase inhibitor, NF-κB, differentiation, G1 arrest
ie
KEY WORDS
e.
N
Previously published online as a Cell Cycle E-publication at: http://www.landesbioscience.com/journals/cc/tocnew25.php?volume=2&issue=5
rd
*Correspondence to: Dr. Steven Grant; Division of Hematology/Oncology; Medical College of Virginia/Virginia Commonwealth University; MCV Station Box 230; Richmond, Virginia 23298 USA; Tel,: 804.828.5211; Fax: 804.828.8079; Email:
[email protected]
fo
of Biochemistry and 3Pharmacology; Virginia Commonwealth University; Medical College of Virginia; Richmond, Virginia USA
The role of NF-κB in regulating G1 arrest and maturation induced by the histone deacetylase inhibitor sodium butyrate (NaB) was examined in human myelomonocytic leukemia cells (U937). Cells stably transfected with an IκBα “super-repressor” lacking phosphorylation sites necessary for proteasomal degradation exhibited diminished IκBα phosphorylation and NF-κB DNA binding upon exposure to TNFα. When exposed to NaB (1 mM; 48 hr) or PMA (5 nM; 24 hr), IκBαM cells displayed a marked reduction in G1 arrest compared to Neo controls. In each case, this was accompanied by a significant reduction in the percentage of cells expressing the differentiation markers CD11a, CD11b, and CD18. The impairment in NaB-induced maturation in mutant cells was associated with a reciprocal increase in apoptosis. In contrast to impairment in NaB- or PMA-induced NF-κB DNA binding, stable expression of the IκBαM did not modify DNA binding of SP1 or AP2 transcription factors. IκBαM cells also displayed impairment in NaB- and PMA-mediated induction of p21CIP1 and phosphorylation (inactivation) of p34cdc2, as well as diminished levels of pRb-bound E2F1. Finally, the NF-κB inhibitor CAPE antagonized NaB- and PMA-related NF-κB DNA binding as well as induction of p21CIP1. Together, these findings suggest that NF-κB plays an important functional role in mediating NaB-induced p21CIP1 induction, G1 arrest, and maturation in human myelomonocytic leukemia cells, and that disruption of the NF-κB pathway causes cells to engage an alternative, apoptotic program.
ot
1Division of Hematology/Oncology; Department of Medicine, and the 2Departments
rib ut io n.
ABSTRACT
Yun Dai1 Mohamed Rahmani1 Steven Grant1,2,3
©
20 03
La
nd
es
B
io
The failure of cells to undergo an orderly program of differentiation represents a fundamental characteristic of neoplastic cells. Central to the process of maturation is cell cycle arrest, generally in the G1 phase of the cell cycle.1 In human leukemia cells, a variety of physiologic and non-physiologic compounds have been shown to trigger the differentiation program, indicating that in such cells, the capacity to undergo maturation is not irretrievably lost. Differentiation-inducing agents include retinoids (i.e., all-trans retinoic acid; ATRA),2 tumor-promoting phorboids (i.e., phorbol myristate acetate; PMA),3 and a class of compounds referred to as histone deacetylase inhibitors (HDIs).4 Histone deacetylases, in conjunction with histone acetyltransferases (HATs) reciprocally regulate the acetylation state of chromatin.5 In general, HDIs, by promoting histone acetylation, allow chromatin to assume a more relaxed state, thereby inducing expression of a variety of genes involved in diverse cellular processes, including differentiation.6 In this context, HDIs such as the short chain fatty acid sodium butyrate, have been shown to be potent inducers of cell cycle arrest and maturation in human leukemia cells.7 Induction of differentiation in leukemic cells by HDIs, as well as by phorbol esters such as PMA, involves the induction of endogenous cyclin-dependent kinase inhibitors (CDKIs) such as p21CIP1.8,9 Significantly, disruption of p21CIP1 induction has been shown to interfere with HDI-mediated leukemic cell maturation, suggesting an important functional role for this CDKI in growth arrest and differentiation.10 Recently, attention has focused on HDIs as anticancer agents, and clinical trials of butyrate derivatives have been initiated in patients with leukemia and other malignancies.11 The NF-κB family of transcription factors has been implicated in diverse cellular functions, including growth regulation, protection from oxidative stress, and cell survival.12-14 NF-κB is customarily sequestered in the cytoplasm due to binding to IκB proteins. Phosphorylation of IκB results in its proteasomal degradation, leading in turn to NF-κB nuclear translocation and transcription of diverse genes.15 Downstream targets of NF-κB involved in cell survival decisions include the anti-apoptotic proteins Bcl-xL and XIAP,
www.landesbioscience.com
Cell Cycle
467
AN INTACT NF-KB PATHWAY IS REQUIRED FOR HISTONE DEACETYLASE INHIBITOR-INDUCED G1 ARREST AND MATURATION IN HUMAN MYELOID LEUKEMIA CELLS U937
among others.16,17 NF-κB also appears to play a role in leukemic cell maturation. For example, evidence that activation of NF-κB is required for maturation of human myelomonocytic leukemia (U937) cells exposed to PMA has recently been presented.18 HDIs have been found to exert disparate effects on NF-κB activation, ranging from increases19 to decreases20 in activity. Currently, however, virtually no information exists concerning the functional role that NF-κB might play in mediating growth arrest and differentiation in human leukemia cells exposed to HDIs such sodium butyrate. To address this issue, we have employed U937 cells stably transfected with an IκBα “super-repressor” mutant which is resistant to proteasomal degradation, and which sequesters NF-κB in the cytoplasm, thereby inactivating it.21 Here we report that dysregulation of the NF-κB axis blocks sodium butyrate-induced G1 arrest and maturation in U937 cells, effects that are associated with a marked diminution in butyrate-related induction of p21CIP1. Collectively, these findings suggest that as in the case of phorbol esters, NF-κB activation plays a critical role in leukemic cell growth arrest and maturation in response to HDIs.
MATERIALS AND METHODS Cells and Reagents. U937 human histiocytic leukemia cells were obtained from ATCC and maintained in RPMI 1640 medium containing 10% FBS, 200 units/ml penicillin, 200 µg/ml streptomycin, minimal essential vitamins, sodium pyruvate, and glutamine as previously reported.22 U937 cells were stably transfected with Ser32/Ser36 mutant IκBα cDNA or an empty vector (pcDNA3.1), and clones selected with G418. All experiments were performed utilizing logarithmically growing cells (3–5 x 105 cells/ml). Histone deacetylase inhibitor sodium butyrate (NaB) was purchased from Biomol (Plymouth Meeting, PA). Phorbol 12-myristate 13-acetate (PMA) and caffeic acid phenylethyl ester (CAPE, NF-κB inhibitor) were purchased from Sigma (St. Louis, MO) and Alexis (San Diego, CA), respectively. These agents were dissolved in DMSO as a stock solution, stored under light-protected conditions at -20˚C. Recombinant human TNFα (Calbiochem, San Diego, CA) were rehydrated in PBS containing 0.5% BSA, aliquoted and stored at -80˚C. In all experiments, the final concentration of DMSO did not exceed 0.1%. Analysis of Apoptosis. The extent of apoptosis was evaluated by assessment of Wright-Giemsa stained cytospin preparation under light microscopy and scoring the number of cells exhibiting classic morphological features of apoptosis. For each condition, 5–10 randomly selected fields per condition were evaluated, encompassing at least 800 cells. Immunoprecipitation and Western Blot Analysis. Whole-cell pellets were lysed in SDS sample buffer and 30 µg protein for each condition was subjected to Western blot analysis following procedures previously described in detail.23 Where indicated, the blots were reprobed with antibodies against actin (Transduction Lab., Lexington, KY) or tubulin (Calbiochem) to ensure equal loading and transfer of proteins. The following antibodies were used as primary antibodies: anti-p21CIP/WAF1 (mouse monoclonal, Transduction Lab.), anti-p27KIP1 (mouse monoclonal, Pharmingen, San Diego, CA), phospho-cdc2 (Tyr15) antibody (rabbit polyclonal, Cell Signaling, Beverly, MA), IκBα antibody (rabbit polyclonal, Cell Signaling), phospho- IκBα (Ser32) antibody (rabbit polyclonal, Cell Signaling). Immunoprecipitation was performed to determine the extent of E2F1 bound to pRb. Briefly, 2 x 107 cells were lysed in RIPA buffer (1% NP-40, 0.5% Na deoxycholate, 1mM PMSF, 1mM Na vanadate, 5 µg/ml CLAPS, and 0.1% SDS in PBS) by syringing approximately 20 times with G23 needle. Protein samples were centrifuged at 12,800g for 30 min and quantified. 200 µg protein per condition were incubated under continuous shaking with 1 µg pRb antibody (mouse monoclonal, Pharmingen) overnight at 4˚C. 20 µl/condition of Dynabeads (Goat anti-mouse IgG, Dynal, Oslo, Norway) were added and incubated for an additional 4 hr. After washing three times with RIPA buffer, the bead-bound protein was eluted by vor-
468
texing and boiling in 20 µl 1X sample buffer. The samples were separated by 12% SDS-PAGE and subjected to Western blot analysis as described above. E2F1 antibody (mouse monoclonal, Upstate Biotech, Lake Placid, NY) was used as primary antibody. Electrophoretic Mobility Shift Assay (EMSA). Nuclear extracts were prepared as described previously.24 Double-stranded oligonucleotides corresponding to NF-κB binding site of Igκ promoter 5’- AGTTGAGGG GACTTTCCCAGGC-3’/3’-TCAACTCCCCTGAAAGGGTCCG-5’, SP1 binding site (5’-ATTCGATCGGGGCGGGGCGAGC-3’/3’-TAAGCTAGCCCCGC CCCGCTCG-5’), and AP2 binding site (5’-GATCGAACTGACCGCCCGCGGCCCGT-3’/3’-CTAGCTTGACTGGCGGGCGCCGGGCA-5’) were obtained from Promega (Madison, WI) and labeled with [γ-32P]ATP (3000 Ci/mmol, ICN Biomedicals, Irvine, CA) using T4 polynucleotide kinase (Promega) and purified using MicroSpin G-25 column (Amersham Pharmacia, Piscataway, NJ). Nuclear extracts (5 µg) were incubated at room temperature for 20 min with 105 cpm of labeled oligonucleotide probe in 15 µl of binding buffer (20 mM Hepes, pH 7.9, 5 mM MgCl2, 4 mM dithiothreitol, 20% glycerol, 0.1 mM phenylmethylsulfonyl fluoride, 5 mM benzamidine, 2 mM levamisol, 0.1 µg/ml aprotinin, 0.1 µg/ml bestatin, 2 µg poly (dI/dC). The reaction mixtures were then loaded onto 6 % native polyacrylamide gels in 0.5x TBE (pH 8.0), and electrophoresed for 1.5 hr at 150 V. The gels were dried at 80˚C and exposed to X-ray film for autoradiography. Cell Cycle Analysis. At the indicated intervals after treatment, 2 x 106 cells were pelleted by centrifugation at 4˚C, resuspended and fixed with 67% ethanol in PBS overnight at 4˚C. Then cells were incubated with a propidium iodide solution containing 3.8 mM Na citrate, 0.5 mg/ml RNase A (Sigma), and 0.01 mg/ml propidium iodide (Sigma) for 3 hr on ice. Cell cycle analysis was performed by flow cytometry using Verity Winlist software (Topsham, ME).25 Analysis of S Phase Content. S-phase content was evaluated by monitoring incorporation of bromodeoxyuridine (BrdU) into DNA during DNA synthesis. For each condition, 2 x 106 cells (cell density = 5 x 105/ml) were incubated with 10 µM BrdU for 30 min at 37˚C. After washing twice with 1% BSA/PBS, the cells were resuspended in 70% ethanol and fixed for 30 min on ice. The BrdU-labeled cells were denatured and nuclei released by incubation with 2N HCl/0.5% Triton X-100 for 30 min at room temperature. After centrifugation, the pellet was resuspended in 0.1M Na2B4O4 (pH 8.5) to neutralize the acid. 1 x 106 cells/100 µl in 0.5% Tween 20/1% BSA/PBS were incubated with FITC-conjugated anti-BrdU (1:10, mouse monoclonal, DAKO, Carpinteria, CA) for 30 min at 4˚C. After washing once with 0.5% Tween 20/1% BSA/PBS, the cells were resuspended in PBS containing 5 µg/ml propidium iodide and analyzed by flow cytometry. The content of S-phase cells was evaluated by determining percentage of BrdU FITC-positive part. Analysis of Cell Differentiation. At the indicated time points after treatment, 1x106 cells for each condition were resuspended in cold PBS and incubated with 1:10 diluted PE-CD11b, -CD11a and -CD18 antibody (Beckton Dockinson, San Jose, CA), and PE-IgG2a (monoclonal control) for 20 min on ice, respectively. The percentage of CD11b-, CD11a-, and CD18-positive cells was determined by flow cytometric analysis.26 Statistical Analysis. For analysis of cell cycle, cell differentiation and apoptosis, values represent the means ± SD for at least three separate experiments performed in triplicate. The significance of differences between experimental variables was determined using the student’s T test.
RESULTS U937 cell clones stably expressing an IκBα “super-repressor” were obtained by transfection with mutant IκBα cDNA (IκBα M; ser32, 36"ala). As shown by the EMSA analysis in Figure 1 (upper panel), exposure to TNFα resulted in a marked increase in NF-κB DNA binding in Neo control cells, but virtually none in the IκBα mutant. TNFα also induced phosphorylation of Neo IκBα, but not the IκBα mutant (Fig. 1, lower panel). Thus, NF-κB activation was impaired in the IκBα mutant cell line. Effects of expression of the IκBα mutant were then examined in relation
Cell Cycle
2003; Vol. 2 Issue 5
IκBαM
Ctrl
TNF
Neo
TNF
Ctrl
+ Cold Probe
+ N.S. Probe
AN INTACT NF-KB PATHWAY IS REQUIRED FOR HISTONE DEACETYLASE INHIBITOR-INDUCED G1 ARREST AND MATURATION IN HUMAN MYELOID LEUKEMIA CELLS U937
NF-κB
p-IκBα Figure 1. U937 cells were stably transfected with Ser32/Ser36 mutant IκBα cDNA or an empty vector (pcDNA3.1). Cells stably expressing IκBαmt32/36 (IκBαM) as well as empty vector controls (Neo) were treated with 5ng/ml TNFα for 5 min, after which nuclear extracts were prepared and subjected to electrophoretic mobility shift assay (EMSA) as described in Methods. Activity of NF-κB was reflected by the extent of binding to 32P-labeled oligonucleotides corresponding to the NF-κB binding site of the Igκ promoter (upper panel). Hela cells extract was used as positive control in lane 1 (preincubated with nonspecific oligo for 10 min) and 2 (preincubated with unlabeled specific oligo for 10 min). Alternatively, phosphorylation of IκBα was monitored by Western blot (lower panel). Results are representative of three separate experiments.
A
to NaB-mediated G1 arrest (Fig. 2A). Comparisons were also made to effects obtained with PMA. Whereas exposure of Neo control cells to 1 mM NaB for 48 hr resulted in a striking increase in the percentage of cells in G1 (e.g., from 50% to 85%), NaB-mediated G1 arrest was essentially absent in IκBαM mutants. A reciprocal reduction in the S-phase (e.g., 45% to 10%) was noted in Neo controls, but not in the mutant line. Essentially equivalent results were obtained in cells exposed to 5 nM PMA (24 hr), as previously described,18 and when a second IκBαM clone was employed (data not shown). The relative failure of IκBαM cells to undergo G1 arrest in response to NaB or PMA was confirmed by flow cytometric analysis of BrdU-FITClabeled cells (Fig. 2B). These results indicate that NF-κB activation is required for NaB-mediated G1 arrest. The impact of impaired NF-κB activation was then investigated in relation to NaB-mediated differentiation in U937 cells (Fig. 3, left panels). In Neo control cells, exposure to 1 mM NaB resulted in a marked increase in CD11b expression after 72 hr of exposure, and in CD11a and CD18 expression after 48 hr of treatment. In marked contrast, induction of each of these differentiation markers was essentially abrogated in IκBαM cells. Parallel reductions in PMA-induced maturation were also observed (Fig. 3, right panels), consistent with earlier reports.18 Thus, impairment in NaBmediated G1 arrest in IκBα mutant cells was accompanied by the failure of cells to undergo differentiation. Effects of disruption of the NF-κB axis were then examined in relation to NaB-mediated apoptosis in U937 cells. As shown in Figure 4A, exposure to NaB (24 or 48 hr) resulted in significantly more apoptosis in IκBαM cells than in Neo controls (P
