SEARCH

SEARCH BY CITATION

Keywords:

  • Retinoid-dependent signaling;
  • Cell cycle G1 exit;
  • Proliferation/differentiation transition;
  • Normal granulopoiesis

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Little is known about the mechanisms by which retinoic acid receptor α (RARα) mediates the effects of retinoic acid (RA) to coordinate granulocytic proliferation/differentiation (P/D) transition. Cyclin-dependent kinase-activating kinase (CAK) complex, whose activity in phosphorylation of RARα is determined by its targeting subunit ménage à trois 1 (MAT1), regulates G1 exit, a cell cycle stage when cells commonly commit to proliferation or to differentiation. We previously found that in myeloid leukemia cells, the lack of RA-induced RARα-CAK dissociation and MAT1 degradation suppresses cell differentiation by inhibiting CAK-dependent G1 exit and sustaining CAK hyperphosphorylation of RARα. This contrasts with our recent findings about the P/D transition in normal primitive hematopoietic cells, where MAT1 degradation proceeds intrinsically together with granulocytic development, in accord with dynamic expression of aldehyde dehydrogenases (ALDHs) 1A1 and 1B1, which catalyze RA synthesis. Blocking ALDH activity inhibits MAT1 degradation and granulocytic differentiation, whereas loss of RARα phosphorylation by CAK induces RA-target gene expression and granulocytic differentiation. These studies suggest that the subversion of RARα-CAK signaling during normal granulopoiesis is crucial to myeloid leukemogenesis and challenges the current paradigm that RA induces cell differentiation solely by transactivating target genes.

Disclosure of potential conflicts of interest is found at the end of this article.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Granulopoiesis refers to the stage-specific differentiation of hematopoietic stem cells (HSC) to common myeloid progenitors (CMP) to granulocyte/monocyte progenitors (GMP) and finally to mature granulocytes (supplemental online Fig. 1). Several genes influencing HSC self-renewal and differentiation have been identified, including Notch [1], bone morphogenetic protein [2], HOXB4 [3, 4], β-catenin [5], and retinoic acid receptor γ (RARγ) [6]. However, studies describing genes that reciprocally coordinate the restricted expansion and granulocytic development of HSC are lacking. The biological effects of retinoic acid (RA) on granulocytic differentiation of both normal and malignant myeloid cells are mediated by retinoid receptors, notably RARα, which transactivate RA-target genes [7, [8], [9]10]. The recent finding that RARα is a substrate for cyclin-dependent kinase-activating kinase (CAK) complex [11] led to the discovery that two RARα-CAK signaling events, ménage à trois 1 (MAT1) degradation and RARα hypophosphorylation in the presence of RA, coordinate cell cycle G1 exit and transition into differentiation in myeloid leukemia, squamous carcinoma, and neuroblastoma cells [12, [13], [14]15]. Determining the mechanisms of RARα-CAK signaling in mediating granulocytic development of HSC are crucial for the development of effective therapies that drive granulocytic differentiation of myeloid leukemia cells.

Differentiation is linked with cell cycle exit, which often occurs in G1 phase and may result from a dual function of cell cycle regulators that coordinate cell cycle exit and transition into differentiation [16, 17]. Human CAK complex is an enzyme consisting of cyclin-dependent kinase 7 (CDK7) [18], cyclin H [19], and MAT1 [20, 21]. It exists in cells either as free CAK or as part of the general transcription factor IIH (TFIIH) complex [22, 23]. Free CAK controls cell cycle progression by phosphorylation-activation of CDKs [22], whereas TFIIH-containing CAK regulates transcription by phosphorylation of the largest subunit of RNA polymerase II (RNA Pol II) [24, [25]26]. Accumulating evidence shows that CAK controls G1 exit, a stage when cells commonly commit to proliferation [27, [28], [29], [30]31] or to differentiation [12, 14, 15]. Adenoviral-MAT1 antisense mimics the effect of RA to inhibit CAK phosphorylation of the retinoblastoma tumor suppressor protein (pRb), which is associated with G1 arrest and cell differentiation [14, 15]. Loss of RARα phosphorylation by CAK due to RA-induced MAT1 ubiquitination [32] induces G1 arrest and granulocytic differentiation in leukemia cells [12, 14]. Hence, decreased CAK phosphorylation of pRb and RARα may be key to coordinating a reduction in cell division and a switch to granulocytic differentiation.

To date, all known functions of CAK are MAT1-dependent. MAT1 assembles CAK [20, 33] and determines CAK's substrate specificity [20, 25, 34, [35]36]. In the presence of MAT1, the substrate preference of cyclin H/CDK7 for CDK2 is diverted to RNA Pol II [25, 34, 37, 38], tumor suppressor p53 protein [35], and octamer transcription factors [36]. Anti-MAT1 antibody decreases TFIIH-mediated transcription [38], and mice lacking MAT1 are defective in RNA Pol II phosphorylation [39]. Retroviral-MAT1 antisense decreases CAK phosphorylation of pRb [31], leading to CAK-dependent G1 arrest [31, 40]. Adenovirus-MAT1 antisense mimics the effect of RA to induce a CAK-dependent proliferation/differentiation (P/D) transition, whereas this transition is blocked by MAT1 overexpression [15]. RA-induced ubiquitination-proteolysis of MAT1 decreases CAK phosphorylation of RARα, leading to granulocytic differentiation of myeloid leukemia cells [12, 14, 32]. These studies demonstrate that the intracellular regulation of MAT1 levels controls CAK activities during cell cycle, transcription, and the P/D transition, whereas decreased CAK phosphorylation of pRb and RARα due to RA-induced MAT1 degradation is crucial to coordinating G1 arrest and transition into cell differentiation.

RARα is a phosphoprotein [11, 41] and a transcription factor [7, 42]. Ser-77 of RARα (RARαS77) within the ligand-independent AF-1 domain is the main residue phosphorylated by CAK [11], whereas the ligand-dependent AF-2 activation domain core (AF-2ADc) transduces RA signaling [7, 42]. CAK hyperphosphorylates RARα in myeloid leukemia cells in the absence of RA; by contrast, RA-induced MAT1 degradation inhibits CAK phosphorylation of RARα [32], which correlates with granulocytic differentiation of leukemia cells [12, 14]. The lack of RA-induced MAT1 degradation and RARα hypophosphorylation is associated with the absence of leukemia cell differentiation [12]. These studies therefore indicate the presence of a RA-induced regulatory loop that couples post-translational modification to transcriptional control, in which RA:RARα binding induces MAT1 degradation to suppress CAK phosphorylation of RARα, leading to transcriptional expression of RA-target genes to mediate granulocytic differentiation of myeloid leukemia cells.

Growing evidence indicates that many pathways associated with cancer normally regulate stem cell development [43, [44]45]. RA has shown a potent effect on normal granulopoiesis [46, [47], [48]49], whereas RARα appears to be the most potent of the retinoid receptors in inducing these effects [6, 50, [51]52]. Because the RA-dependent RARα-CAK signaling events MAT1 degradation and RARα hypophosphorylation mediate granulocytic differentiation of leukemia cells [12, 14, 32], the availability of intrinsic RA may play a central role in initiation of RARα-CAK signaling to mediate normal granulocytic differentiation. Thus, we investigated whether RARα-CAK signaling coordinates CAK-dependent cell division and transition into differentiation in normal primitive hematopoietic cells. Our studies reveal that RA-induced RARα-CAK signaling events observed in leukemia cells appear to proceed intrinsically during normal granulopoiesis, suggesting that deregulation of CAK-dependent P/D transition is necessary to subvert the granulocytic development of HSC.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Isolation of Human CD34+, HSC, CMP, GMP, and CD66B+ Cells

Human umbilical cord blood (CB) was collected at Kaiser Permanente Hospital (Los Angeles, CA) according to guidelines under protocols approved by the Childrens Hospital Los Angeles Committee on Clinical Investigation. For isolation of HSC, CMP, GMP, and CD34+ cells, CB mononuclear cells were isolated from total CB using Ficoll-Hypaque (Amersham Biosciences Inc., Piscataway, NJ, http://www.amersham.com) density centrifugation. CB mononuclear cells were enriched for CD34+ cells using the MiniMacs CD34 Progenitor Isolation Kit (Miltenyi Biotec, Auburn, CA, http://www.miltenyibiotec.com) according to the manufacturer's instructions. Purified CD34+ cells (typical purity 95% or greater) were obtained by running enriched CD34+ cells through a second MiniMac separation column. By using four-color fluorescence-activated cell sorting (FACS) (FACSVantage; BD Biosciences, San Jose, CA, http://www.bd.com), HSC, CMP, and GMP were isolated from the purified CD34+ cells based on the defined surface immunophenotypes and established methods as described [53, [54]55]. CD66B+ mature granulocytes were separated from fresh CB cells (within 6 hours of delivery) as follows: following density centrifugation using Ficoll-Hypaque as described above, cells located below the buffy layer and the white cells at the very top of the red blood cell layer were collected, washed, and resuspended in phosphate-buffered saline (PBS). To remove remaining red cells, cell suspension was layered over Ficoll-Hypaque for a second 5-minute centrifugation at a speed that was 75% of that used for standard mononuclear cell isolation. Following a 5-minute centrifugation, fluid above the red blood cell layer was collected, washed, and stained with CD66B fluorescein isothiocyanate (FITC). Granulocytes were purified by positive selection (typical purity 95% or greater) using anti-FITC MicroBeads (Miltenyi Biotec) according to the manufacturer's instructions.

Cell Culture

HSC, CMP, GMP, or CD34+ cells were cultured in myeloid medium [56] containing 30% human stroma cell-conditioned medium (MM-HSCM) (myeloid medium: Iscove's modified Dulbecco's medium [Gibco-BRL, Gaithersburg, MD, http://www.gibcobrl.com]; 30% fetal calf serum [FCS]; 1% bovine serum albumin [Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com]; 2-mercaptoethanol [Sigma-Aldrich]; 10−6 mol/l hydrocortisone [Sigma-Aldrich]; glutamine; and a combination of 10 ng/ml interleukin-3 [IL-3], 50 U/ml IL-6, 50 ng/ml stem cell factor, and 50 ng/ml granulocyte-monocyte colony-stimulating factor). The MM-HSCM sustains granulopoiesis while blocking the growth of lymphoid and erythroid cells in the presence of hydrocortisone and in the absence of erythropoietin, respectively. The cells were split every 3 days with equal portions of fresh MM-HSCM. Human myeloid leukemia HL60, HL60R, and neuroblastoma CHP126 cells were cultured as described [12, 15]. Human brain tumor U251 cells were cultured in RPMI 1640 medium plus 10% FCS. All of the cells were grown in the medium supplemented with 100 units/ml penicillin and 100 μg/ml streptomycin at 37°C with 5% CO2. All-trans retinoic acid (ATRA) and disulfiram were purchased from Sigma-Aldrich. Either ATRA or disulfiram was dissolved in ethanol. ATRA (1 μM) or disulfiram (0.2 μM) was used in experiments.

Immunofluorescence Analysis

Rabbit polyclonal antibodies against MAT1 (FL-309), CDK7 (C-19), RARα (C-20), and pRb (C-15) were from Santa Cruz Biotechnology Inc. (Santa Cruz, CA, http://www.scbt.com). Secondary antibodies, including FITC-conjugated donkey anti-rabbit IgG and indocarbocyanine Cy3-conjugated donkey anti-goat IgG (Jackson Immunoresearch Laboratories, West Grove, PA, http://www.jacksonimmuno.com), were used at 1:100 dilutions. Immunofluorescence detection was performed variably depending on the antibodies used. In general, HSC, CMP, and GMP cells were spun onto microslides. Cells were fixed in 2% formaldehyde/PBS for 10 minutes at 4°C and then postfixed in 100% cold ethanol for 30 minutes at 4°C, followed by a permeabilization with 100 mM glycine and 0.5% Triton X-100 in PBS for 30 minutes at room temperature. Cells were then blocked with 5% nonimmune serum (the same host species of the labeled secondary antibody) in PBS for 30 minutes at 4°C. Primary and secondary antibodies were left on cells for 2 hours each. Cell nuclei were stained with 4,6-diamidino-2-phenylindole in aqueous mounting medium under a coverslip. For controls, we omitted either primary or both primary and secondary antibodies. Fluorescence intensity per cell abundance was determined through color threshold and integrated morphometry analysis by using MetaMorph software (Universal Imaging Corp., Downington, PA, http://www.moleculardevices.com) as described [15].

Immunologic Methods

Western blot and immunoprecipitation were performed as described before [31]. Rabbit anti-phosphoserine antibody was purchased from Invitrogen (Carlsbad, CA, http://www.invitrogen.com), and all other antibodies were purchased from Santa Cruz Biotechnology.

Vector Construction and Virus Production

The pCCL-c-MNDU3c-X2-PGK-enhanced green fluorescent protein (EGFP) (provided by Dr. Kohn, Childrens Hospital Los Angeles/University of Southern California) was generated from a third-generation lentiviral vector [57, 58]. The phosphorylation-defective human RARαS77A cDNA (provided by Dr. Crowe, University of Southern California) was cloned into the pCCL vector using our established techniques [31]. By using Lipofectamine 2000 (Invitrogen), vesicular stomatitis virus-pseudotyped vectors were produced as described [59] by cotransient transfection of 293FT cells (Invitrogen) with 10 μg of expression constructs (pCCL-RARαS77A or empty pCCL vectors), 10 μg of pRΔ8.9 packaging plasmids [60], and 2 μg of pMD.G envelope plasmids [61]. Sodium pyruvate (Invitrogen) induction was performed by following the manufacturer's instructions. After 72 hours of transfection, virus supernatants were harvested, centrifuged, and filtered with a 0.2-μm filter flask (Corning Costar, Lowell, MA, http://www.corning.com/lifesciences). Aliquots of supernatants were stored at −80°C.

Titer Determination of Vector Supernatants

The viral titer was determined in 293 cells (American Type Culture Collection, Manassas, VA, http://www.atcc.org) transduced with serial endpoint dilutions of different vector supernatants as described [57]. In brief, 293 cells were cultured in six-well plates (1 × 105 cells per well) for 24 hours and then transduced with 1-ml serial dilutions of vector supernatants. After 48 hours of transduction, expression levels of green fluorescent protein (GFP) were analyzed by FACS analyses, and titers were calculated by multiplying the cell number at the time of transduction with both the percentage of GFP-positive cells and the dilution factor to yield the number of infectious units (IU)/ml. Titers ranged between 2 × 106 and 1 × 107 IU/ml.

Lentiviral Transduction

Protein expression of pCCL-RARαS77A was determined by transducing HL60R cells (supplemental online Fig. 2A–2D). CD34+ cells were cultured in MM-HSCM for 24 hours and then plated onto non-tissue culture-treated 96-well plates coated with recombinant fibronectin (Takara, Otsu, Japan, http://www.takara.co.jp). After 12 hours of plating, CD34+ cells were transduced with pCCL-RARαS77A or vector (multiplicity of infection = 1:100) in the presence of MM-HSCM using established methods [62, 63]. A second cycle of transduction was performed 8 hours later with new virus supernatant in the presence of fresh MM-HSCM. Twenty-four hours after initial transduction, cells were washed and then cultured with MM-HSCM either in suspension or on a fibronectin-coated plate (for assessing colony-forming unit-granulocyte [CFU-G]).

Cell Proliferation, CFU-G Production, and Morphologic Differentiation Analysis

Cell proliferation was determined by cell count using a standard hemocytometer as described [40]. Depending on the available amounts of purified human primitive hematopoietic cells, equal numbers of cells were plated onto 12-well plates either in a single well (for HSC, CMP, and GMP) or in triplicate wells (for CD34+ cells). Twenty-four hours after plating, the cells were counted for up to 12 or 14 days. The proliferation pattern resulting from the single-well analysis was examined by three independent experiments. Using trypan blue exclusion methodology, the degree of cell death associated with cell proliferation in the cultures was simultaneously measured. Less than 5% of cell death observed in both control and experimental sample was considered normal and was not plotted into the graph. The total number of counted cells in the proliferation analysis is indicated by the y-axis scale in the graph. For CFU-G detection, RARαS77A- or vector-transduced cells were cultured in fibronectin-coated 96-well plates. Five days after initial transduction, CFU-G was viewed under a phase-contrast fluorescence microscope as described [15], and CFU-G production was calculated by randomly capturing at least 10 different areas of each well. For nuclear segmentation analysis, cell suspensions of HSC, CMP, and GMP were spun onto Micro slides. Cells were fixed in 4% paraformaldehyde and permeabilized in 0.1% Triton X-100/0.1% sodium citrate. Cell nuclei were stained with Wright-Giemsa stain (Sigma-Aldrich). Images were acquired as described [15].

Reverse Transcription-Polymerase Chain Reaction Analysis

Total RNA was extracted using the STAT-60 RNA isolation kit (Tel-Test, Inc., Friendswood, TX, http://www.tel-test.com). The iScript cDNA synthesis kit (Bio-Rad, Hercules, CA, http://www.bio-rad.com) was used for reverse transcription (RT). Specific primers designed to amplify the different target genes were synthesized by Operon Biotechnologies (Huntsville, AL, http://www.operon.com). Equal amounts of cDNA were used in polymerase chain reaction (PCR) using HotStarTaq Master Mix Kit (Qiagen, Hilden, Germany, http://www1.qiagen.com), and amplification of human β-actin was performed as control. PCR was performed on a Thermal ControllerPTC-100 (MJ Research, Waltham, MA, http://www.mjr.com), and PCR cycling conditions were varied based on primer-specific annealing and melting temperatures. PCR products were resolved by agarose gel, and gene amplification was visualized and determined by FluorChem-TM 8900 (Alpha Innotech, San Leandro, CA, http://www.alphainnotech.com).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

RARα-CAK Signaling Proteins Are Expressed in Primary HSC, CMP, and GMP

It is axiomatic in developmental biology that genes must be expressed in the same time and space if they are to interact functionally. Thus, we investigated whether RARα-CAK signaling proteins are expressed in primary HSC, CMP, and GMP using immunofluorescence analysis. We found that CDK7, MAT1, RARα, and pRb were expressed in HSC, CMP, and GMP and that MAT1 and RARα were coexpressed in each of these cell types (Fig. 1). These results indicate that coexpression of RARα-CAK signaling proteins in HSC and myeloid progenitors may be involved in modulating granulocytic development from HSC to CMP to GMP.

thumbnail image

Figure Figure 1.. RARα-CAK signaling proteins are expressed in primary HSC, CMP, and GMP. (A): Immunofluorescence analysis of the expression of CDK7, MAT1, RARα, and pRb. (B): Colocalization of RARα and MAT1. DAPI was used for nuclear staining. Controls included no primary AB (no 1st-AB) or no primary and secondary ABs (no-ABs). Abbreviations: AB, antibody; CB, cord blood; CDK7, cyclin-dependent kinase 7; CMP, common myeloid progenitors; DAPI, 4,6-diamidino-2-phenylindole; FITC, fluorescein isothiocyanate; GMP, granulocyte/monocyte progenitors; HSC, hematopoietic stem cells; MAT1, ménage à trois 1; pRb, retinoblastoma tumor suppressor protein; RARα, retinoic acid receptor α.

Download figure to PowerPoint

Granulocytic Development of HSC Is Enhanced by Exogenous ATRA and Correlates with Hypophosphorylation of pRb and RARα

Because RARα-CAK signaling is ATRA-dependent, we examined stage-specific normal granulopoiesis and its relationship to ATRA stimuli. HSC, CMP, and GMP were cultured in six-well plates with MM-HSCM in the presence of ATRA or control vehicle. Cell proliferation and morphology differentiation were examined as described [12, 40]. We found that without ATRA, HSC divided slowly, with limited replenishment and a low rate of granulocytic morphologic differentiation; CMP, by contrast, showed marked expansion, with an increased rate of granulocytic morphologic differentiation; and GMP had the slowest proliferation rate and the highest rate of granulocytic morphologic differentiation (Fig. 2A–2C). In parallel, ATRA inhibited proliferation while promoting differentiation in all three populations (Fig. 2A–2C). Since CD34+ cells contain a mixture of HSC, CMP, and GMP resulting from the differentiation of HSC to CMP to GMP stages, we also examined the consecutive process of granulocytic differentiation of CD34+ cells and its relationship to ATRA stimuli. Purified CD34+ cells were cultured with MM-HSCM in the presence of ATRA or control vehicle. The results from proliferation and differentiation analyses (Fig. 2D–2F) showed that granulocytic development of CD34+ cells correlated with reduced proliferation, whereas exogenous ATRA promoted this process. These data show that ATRA influences granulocytic differentiation of primitive hematopoietic cells.

thumbnail image

Figure Figure 2.. Granulocytic development of HSC is enhanced by exogenous ATRA and correlates with hypophosphorylation of pRb and RARα. (A): Proliferation analysis of HSC, CMP, and GMP by cell count. Data shown are representative of three independent experiments. (B): Morphologic differentiation of HSC, CMP, and GMP after 12 days of culture (data at days 3, 6, and 9 not shown). Nuclear segmentation (indicated by arrowheads) is characteristic of differentiated granulocytes. (C): Quantification of nuclear segmentation of cells in (B) (data at days 3 and 6 not shown). (D): Proliferation analysis of CD34+ cells by cell count. (E): Morphologic differentiation of CD34+ cells. (F): Quantification of nuclear segmentation of cells in (E). (G): Western analysis of pRb, RARα, and CDK7 in CMP after 12 days of culture. (H): Examination of phosphorylation status of pRb, RARα, and CDK7 of cells in G by rehybridizing the same blot using anti-Phos-Se antibody. (I): Densitometry quantification of Phos-Se levels of pRb, RARα, and CDK7 in (H). The levels of Phos-Se were normalized based on Actin expression. Abbreviations: ATRA, all-trans retinoic acid; CDK7, cyclin-dependent kinase 7; CMP, common myeloid progenitors; EtOH, ethanol; GMP, granulocyte/monocyte progenitors; HSC, hematopoietic stem cells; Phos-Se, phosphoserine; pRb, retinoblastoma tumor suppressor protein; RARα, retinoic acid receptor α; WB, Western blotting.

Download figure to PowerPoint

Previous studies showed that ATRA-induced granulocytic differentiation of leukemia cells accompanies decreased CAK phosphorylation of RARα and pRb [12, 14]. We therefore investigated whether ATRA-enhanced granulocytic differentiation (Fig. 2A–2F) correlates with decreased phosphorylation of RARα and pRb in CMP after 12 days of culture, a time point when we observed a reduction in proliferation (Fig. 2A, middle panel) and an induction in granulocytic differentiation (Fig. 2B, 2C). By using anti-phosphoserine antibody together with anti-pRb, anti-RARα, and anti-CDK7 antibodies in Western analysis, we examined the expression levels and phosphorylation status of pRb, RARα, and CDK7. We found that whereas the levels of pRb, RARα, and CDK7 remained relatively stable in the presence or absence of ATRA (Fig. 2G), ATRA inhibited phosphorylation of both pRb and RARα, which correlated with decreased CDK7 activity reflected by CDK7 autohypophosphorylation (Fig. 2H, 2I). Since RA inhibits CAK phosphorylation of both pRb and RARα during myeloid leukemia cell differentiation [12, 14], these data suggest that decreased CAK phosphorylation of pRb and RARα may drive granulocytic differentiation of primitive hematopoietic cells.

Stage-Specific MAT1 Degradation and RARα-CAK Dissociation Proceed Intrinsically During Granulocytic Development of HSC

MAT1 degradation occurs during ATRA-induced granulocytic differentiation of leukemia cells [12, 14]. Hence, we investigated whether normal granulopoiesis involves a spatiotemporal change in MAT1 levels during differentiation from the developmentally sequential HSC to CMP to GMP. Purified HSC, CMP, or GMP were cultured in six-well plates with MM-HSCM for 9 days in the presence of ATRA or control vehicle. Western analysis showed that MAT1 was essentially intact in HSC, cleaved in CMP, and extensively degraded in GMP (Fig. 3A, 3B). Because of the marked MAT1 degradation in GMP cultures, where more cells have completed granulocytic differentiation with reduced proliferation (Fig. 2A–2C), we examined the levels of MAT1 protein in mature granulocytes purified from CB. Western analysis showed that MAT1 was extensively degraded in primary granulocytes compared with HL60 cells (Fig. 3C). To further define this MAT1 degradation during normal granulopoiesis, we determined MAT1 expression in primary HSC, CMP, and GMP using immunofluorescence analysis. We found that progressive MAT1 degradation indeed proceeds from HSC to CMP to GMP (Fig. 3D, 3E) together with granulocytic differentiation (Fig. 2A–2C). Because MAT1 expression sustains proliferation whereas RA-induced MAT1 degradation induces differentiation by cross-regulating CAK-dependent cell division and the P/D transition in myeloid leukemia cells [12, 14], these results suggest that intrinsically programmed MAT1 expression and degradation coordinate the restricted expansion and granulocytic differentiation by mediating CAK activity during normal granulopoiesis.

thumbnail image

Figure Figure 3.. Stage-specific MAT1 degradation and RARα-CAK dissociation proceed intrinsically during granulocytic development of HSC. (A–C): Western analysis of MAT1. (D): Immunofluorescence analysis of MAT1 expression in primary HSC, CMP, and GMP. (E): Quantification of MAT1 levels of cells in (D). (F): Immunoprecipitation analysis of RARα-CAK interaction and MAT1-dependent CAK abundance. CHP126 cells were used as positive controls. Abbreviations: AB, antibody; ATRA, all-trans retinoic acid; CB, cord blood; CDK7, cyclin-dependent kinase 7; CMP, common myeloid progenitors; DAPI, 4,6-diamidino-2-phenylindole; FITC, fluorescein isothiocyanate; GMP, granulocyte/monocyte progenitors; HSC, hematopoietic stem cells; IP, immunoprecipitation; MAT1, ménage à trois 1; P-Gra, primary granulocytes; PI, preimmune IgG; RARα, retinoic acid receptor α; WB, Western blotting.

Download figure to PowerPoint

MAT1 assembles CAK [20, 21, 33], whereas RARα-CAK dissociation leads to MAT1 degradation and decreasing CAK abundance during RA-induced granulocytic differentiation of myeloid leukemia cells [12, 14, 32]. Therefore, we investigated the relationship among RARα-CAK interaction, MAT1 levels, and CAK abundance using cells that reflect an early stage (CMP after 6 days of culture) and a late stage (GMP after 9 days of culture) of normal granulocytic development. Immunoprecipitation analysis showed that exogenous ATRA induced RARα-CAK dissociation in day 6 CMP progeny, as evidenced by decreased RARα levels in CAK-complexed precipitates (Fig. 3F, lane 4 vs. lane 5). In day 9 GMP progeny, marked RARα-CAK dissociation was associated with both extensively decreased MAT1 levels and CAK abundance, as reflected by decreased levels of RARα, MAT1, and CDK7 in CAK-complexed precipitates (Fig. 3F, lanes 4 and 5 vs. lanes 6 and 7). RARα-CAK dissociation occurred earlier (in day 6 CMP progeny), whereas extensive MAT1 degradation occurred later (in day 9 GMP progeny). Such progressive RARα-CAK dissociation and MAT1 degradation correlated with the extent of granulocytic differentiation in CMP and GMP progeny (Fig. 2B, 2C). Thus, RARα-CAK dissociation and MAT1 degradation may mediate the granulocytic development of HSC.

RARαS77A-Mimicked Loss of RARα Phosphorylation by CAK Enhances Normal Granulopoiesis and Upregulates Cell Cycle Inhibitor p21 Expression

RARαS77 is the main RARα residue phosphorylated by CAK [11]. ATRA treatment decreases CAK phosphorylation of RARα, and RARα hypophosphorylation correlates with granulocytic differentiation of leukemia cells [12, 14]. Moreover, RARα hypophosphorylation mimicked by RARαS77A induces both granulocytic differentiation and ATRA-target gene expression in leukemia cells (supplemental online Fig. 2E–2I). Thus, we examined the effect of RARα hypophosphorylation on normal granulopoiesis and RA-target gene expression. CD34+ cells were plated on fibronectin-coated plates. After 6 hours of plating, cells were transduced with lentiviral-RARαS77A using established methods [62, 63]. The cells were then cultured with ATRA or control vehicle for an additional 4 days. We found that RARαS77A expression, as reflected by EGFP expression in transduced CD34+ cells (Fig. 4A), promoted CFU-G production (Fig. 4B) and granulocytic differentiation (Fig. 4C, 4D) with reduced proliferation (Fig. 4E). In RARαS77A-transduced cells, both the number of granulocytes and the degree of morphologic differentiation were profoundly enhanced by ATRA (Fig. 4C, panel III vs. panel IV; Fig. 4D). Such ATRA-enhanced effects of RARαS77A on driving granulocytic differentiation (Fig. 4C–4E) are through the intact AF-2ADc of RARαS77A, which transduces ATRA signaling [7]. Moreover, RARαS77A upregulates expression of p21 mRNA (Fig. 4F, 4G), the only known cell cycle inhibitor gene that is directly transactivated by liganded RARα [64] during granulocytic differentiation of leukemia cells [65]. These results show that RARα hypophosphorylation, a critical RARα-CAK signaling event, drives RA-target gene expression and granulocytic differentiation in CD34+ cells.

thumbnail image

Figure Figure 4.. RARα-mimicked loss of RARα phosphorylation by cyclin-dependent kinase-activating kinase enhances normal granulopoiesis and upregulates cell cycle inhibitor p21 expression. (A): Lentiviral-RARαS77A (S77A) expression reflected by enhanced green fluorescent protein expression. (B): CFU-G analysis. ***, Vector versus S77A (p < .001). (C): Morphologic differentiation of CD34+ cells. (D): Quantification of nuclear segmentation of cells in (C). ***, Vector versus S77A in the ethanol group (p < .001) and ATRA group (p < .001). (E): Proliferation analysis by cell count. *, Vector versus S77A in the ethanol group (p < .05) and ATRA group (p < .05). (F): RT-PCR analysis of p21 expression. (G): Densitometry quantification of p21 mRNA levels of cells in (F) based on Actin expression. Vector versus vector/ATRA (1.7-fold), S77A (1.4-fold), and S77A/ATRA (1.9-fold), respectively. Abbreviations: ATRA, all-trans retinoic acid; CFU-G, colony-forming unit-granulocyte; EtOH, ethanol; Fluo, fluorescence; RA, retinoic acid; RT-PCR, reverse transcription-polymerase chain reaction.

Download figure to PowerPoint

Dynamic Expression of Aldehyde Dehydrogenase 1A1 and Aldehyde Dehydrogenase 1B1 Correlates with Granulocytic Development of HSC

RARα-CAK signaling is RA-dependent, whereas two RARα-CAK signaling events, MAT1 degradation and RARα hypophosphorylation, are involved in the modulation of normal granulopoiesis (Figs. 2, Figure 3.4). These data suggest that the availability of intrinsic RA plays a central role in initiation of RARα-CAK signaling to mediate granulocytic differentiation. Aldehyde dehydrogenase (ALDH) catalyzes cellular RA synthesis [66, 67], and ALDH activity is retained in HSC and progenitor cells [68, [69], [70]71]. We therefore investigated which isoform(s) of ALDHs might be involved in mediating granulocytic development of HSC. Total RNAs were extracted from freshly isolated granulocytes and primary CD34+ cells as well as cultured CD34+ cells for RT-PCR analysis of ALDH1A1, 1A2, 1A3, and 1B1 expression. We found that ALDH1A1 levels peaked after 6 days of culture and then decreased rapidly, whereas ALDH1B1 levels progressively increased in CD34+ cells, eventually peaking in mature granulocytes (Fig. 5A, 5B). ALDH1A2 and 1A3, by contrast, were expressed in human primary hepatocytes and U251 cells but not in either CD34+ cells or granulocytes (Fig. 5C, 5D). To determine whether the expression of ALDH1A1 protein is consistent with the dynamic change in ALDH1A1 mRNA levels, we used anti-ALDH1A1 antibody (ALDH1B1 antibody is not commercially available) in Western analysis. We found that ALDH1A1 was indeed expressed in cultured CMP but not in cultured GMP or freshly isolated mature granulocytes (Fig. 5E, 5F). These results indicate that dynamic expression of ALDH1A1 and ALDH1B1 may be involved in modulating RARα-CAK signaling by governing RA availability during normal granulopoiesis.

thumbnail image

Figure Figure 5.. Dynamic expression of ALDH1A1 and ALDH1B1. (A): RT-PCR analysis of ALDH1A1 and ALDH1B1 expression. This similar expression pattern has been confirmed by three independent experiments. (B): Densitometry quantification of ALDH1A1 and ALDH1B1 mRNA levels of cells in (A). Samples were normalized based on Actin expression. (C, D): RT-PCR analysis of ALDH1A2(C) and ALDH1A3(D) expression. U251 cells and HepaC were used as positive controls. (E, F): Western analysis of ALDH1A1 in cultured CMP and GMP (E) and in freshly isolated mature granulocytes (F). Abbreviations: ALDH, aldehyde dehydrogenase; CMP, common myeloid progenitors; D, day; GMP, granulocyte/monocyte progenitors; HepaC, human primary hepatocytes; P-gra, primary granulocytes; RT-PCR, reverse transcription-polymerase chain reaction; WB, Western blotting.

Download figure to PowerPoint

Inhibition of ALDH Activity Impedes RA-Induced Granulocytic Differentiation, Suppresses Cell Cycle Inhibitor p21 Expression, and Inhibits MAT1 Degradation

ALDH activity can affect RA production to modulate differentiation of HSC and progenitor cells [68, 71]. These findings, coupled with the data in Figures 2, Figure 3., Figure 4.5, suggest that ALDH activity may govern the availability of RA during normal granulopoiesis. We tested this hypothesis by blocking ALDH activity in cultured CMP with disulfiram, an ALDH inhibitor [72] that blocks ALDH activity in hematopoietic cells [68]. CMP were cultured with MM-HSCM in the presence of ATRA or control vehicle for 12 days. Disulfiram was added to duplicate vehicle-treated cells for the last 48 hours of incubation. Morphologic analysis showed that, in contrast to ATRA, disulfiram inhibited granulocytic differentiation of CMP, as indicated by cell morphology (Fig. 6A, 6B). Western analysis showed that disulfiram inhibited the protein expression of p21 (Fig. 6C, 6D), the transcriptional target of liganded RARα that inhibits cancer cell proliferation. Furthermore, degradation of MAT1 into a 30-kDa fragment was enhanced by ATRA but inhibited by disulfiram (Fig. 6E). These results, together with the evidence of dynamic ALDH expression presented in Figure 5, suggest that ALDH1A1 and ALDH1B1 differentially mediate RA-dependent RARα-CAK signaling events to control granulocytic development of HSC by governing the availability of RA.

thumbnail image

Figure Figure 6.. Inhibition of aldehyde dehydrogenase activity by DS impedes RA-induced granulocytic differentiation, suppresses cell cycle inhibitor p21 expression, and inhibits MAT1 degradation. (A): Morphologic differentiation of cultured CMP. (B): Quantification of nuclear segmentation of cells in (A). (C): Western analysis of p21 expression. (D): Densitometry quantification of p21 levels of cells in (C) based on Actin expression. (E): Western analysis of MAT1 degradation in cultured CMP. Abbreviations: ATRA, all-trans retinoic acid; CMP, common myeloid progenitors; DS, disulfiram; EtOH, ethanol; MAT1, ménage à trois 1; WB, Western blotting.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

The previous finding that RARα is a substrate for CAK led to the discovery that two RARα-CAK signaling events, MAT1 degradation and RARα hypophosphorylation, coordinate RA-induced cell cycle G1 exit and transition into granulocytic differentiation in myeloid leukemia cells. Here we reveal that such RA-induced RARα-CAK signaling events appear to proceed intrinsically during granulocytic development of normal primitive hematopoietic cells and that ALDH-governed RA availability may mediate this process by initiating RARα-CAK signaling.

Intrinsically Programmed MAT1 Expression and Degradation: Involvement in Normal Granulopoiesis

MAT1 assembles CAK and determines CAK's substrate specificity. Extensive studies have shown that changes in MAT1 levels alter CAK activity in controlling transcription, cell cycle, and differentiation [25, 31, 34, [35]36, 39, 40]. In myeloid leukemia cells, RA treatment induces RARα-CAK dissociation [14] and MAT1 degradation [32], leading to decreased CAK phosphorylation of pRb and RARs, as well as the induction of P/D transition [12, 14]. By contrast, MAT1 degradation in normal hematopoietic cells proceeds intrinsically in accord with granulocytic development of HSC, during which MAT1 is essentially intact in HSC, cleaved in CMP, and markedly degraded in GMP (Fig. 3). This progressive MAT1 degradation (Fig. 3A–3E) is preceded by RARα-CAK dissociation (Fig. 3F) and is strongly correlated with the degree of granulocytic differentiation (Fig. 2A–2F). MAT1 remains intact in HSC, which divide slowly and differentiate only sporadically. CMP, which are less differentiated than GMP, cycle much more rapidly and demonstrate a higher rate of granulocytic differentiation. Thus, the initiation of MAT1 degradation may ensure the production of a large population of cells undergoing strictly controlled differentiation within the myeloid lineages during the CMP stage of granulopoiesis. GMP, representing a late developmental stage of granulocytes, show an extensive MAT1 degradation correlated with the lowest levels of proliferation and the highest rate of granulocytic differentiation. Hence, marked MAT1 degradation may promote terminal granulocytic differentiation during the GMP stage of granulopoiesis. These data suggest that by cross-regulating CAK-dependent cell division and P/D transition, MAT1 expression sustains the restricted expansion of HSC, whereas progressive MAT1 degradation drives this granulocytic development from HSC to CMP to GMP to mature granulocytes. Further determining the mechanisms of programmed MAT1 expression/degradation in mediating proliferation versus granulocytic differentiation choices of HSC, CMP, and GMP will advance our understanding of normal granulopoiesis and provide a basis for designing improved intervention strategies against myeloid malignancies.

RARα Hypophosphorylation Mediates Granulocytic Differentiation

RARα is a substrate for CAK [12, 14, 32, 41] and a transcription factor [7, 42]. RARαS77 is the main residue of RARα that is phosphorylated by CAK [11], and expression of phosphorylation-defective RARαS77A mimics the effects of RA by inhibiting cancer cell proliferation [13]. Whether or not the loss of CAK-dependent phosphorylation of RARα will influence the effect of liganded RARα on transcriptional control of granulocytic differentiation is unknown. We found that in myeloid leukemia cells, RA-induced MAT1 ubiquitination decreases CAK activity to induce a switch from RARα hyper- to hypophosphorylation [32], which is associated with granulocytic differentiation [12, 14]. In NB4 cells, promyelocytic leukemia/RARα hypophosphorylation occurs in RA-induced G1 arresting cells undergoing differentiation but not in synchronized G1 cells that do not differentiate [14]. The lack of such a switch to RARα hypophosphorylation correlates with the absence of granulocytic differentiation in HL60R cells [12]. Importantly, the loss of CAK-dependent RARα phosphorylation, as mimicked by RARαS77A, induces RA-target gene expression and granulocytic differentiation in both normal and malignant myeloid cells (Fig. 4; supplemental online Fig. 2). These data suggest that RA-induced RARα hypophosphorylation results in transactivation of RA-target genes to induce granulocytic differentiation. Extensive studies show that unliganded retinoid receptors bind to the retinoic acid-responsive element (RARE) of the target genes to repress transcription [8, 42] through the recruitment of corepressors [73, 74]. Ligand-induced conformational changes in receptors may cause the dissociation of corepressors and facilitate the positioning of coactivators at the promoter of target genes [42]. In this regard, hyperphosphorylated RARα, in the absence of RA, binds to the RARE of the promoter and represses transcription. A conformational change in RARα upon RA binding may release transcription-repression by dissociating itself from corepressors and then recruiting coactivators at the promoter of target genes. Indeed, RA treatment diminishes the binding of RARα or PML/RARα to RARE while stimulating transactivation [75, 76], and the lack of such diminished PML/RARα binding to RARE is associated with the inhibition of transactivation in RA-resistant cells [75]. Thus, diminished RARα binding to RARE may result from a conformational change in RARα due to the loss of CAK-dependent RARα phosphorylation. This allows hypophosphorylated RARα to dissociate from corepressors and then to recruit coactivators and/or transcriptional machinery to the promoter of target genes, leading to transcriptional expression of RA-target genes that mediate granulocytic differentiation. Defining this novel role and mechanism of RARα hypophosphorylation in transcriptional control of granulocytic differentiation in HSC and myeloid progenitors will introduce new concepts to the field of RA signaling and stem cell biology.

The Endogenous RA Action During Normal Granulopoiesis

Although little is known about the function of ALDH1B1, previous studies show that the primary role of ALDH1A1 is to catalyze retinol to RA [77]. ALDH1A1 is expressed in HSC and progenitor cells [78]. Inhibition of ALDH activity in RA production delays the differentiation of HSC [71], and such ALDH activity may be lost or diminished upon transformation of hematopoietic progenitors to leukemia cells [68]. These studies indicate that ALDH-dependent availability of RA action is important for normal hematopoiesis, although the specific isoforms of ALDHs and their relationship to RA actions during this developmental process are unknown. In exploring RA-activated RARα-CAK signaling during normal granulopoiesis, we found that ALDH1A1 in cultured CD34+ cells is expressed at earlier stages of granulocytic development but is absent in later stages of granulocytic differentiation and in mature granulocytes (Fig. 5A, 5B, 5E, 5F). On the other hand, the levels of ALDH1B1 mRNA progressively increase and peak in mature granulocytes (Fig. 5A, 5B) along with granulocytic development (Fig. 2). In contrast to the effects of ATRA that enhance granulocytic differentiation, induce p21 expression, and promote MAT1 degradation, blocking ALDH activity with disulfiram treatment inhibits granulocytic differentiation, suppresses p21 expression, and inhibits MAT1 degradation (Fig. 6). These studies suggest that ALDH1A1 and ALDH1B1 differentially mediate RARα-CAK signaling events by governing RA availability to control granulocytic development of HSC. It is possible that ALDH1A1 activity may be required to shift the balance toward more proliferation than differentiation at earlier stages of granulopoiesis (i.e., in HSC and CMP). In contrast, ALDH1B1 may be required to sustain commitment to the granulocytic fate during later stages of granulocytic development (i.e., in GMP) and in mature granulocytes. Determining the individual roles of ALDH1A1 and ALDH1B1 in mediating granulopoiesis will provide critical insights into the spatiotemporal restrictions on RA synthesis and into the mechanisms of RARα-CAK signaling during granulopoiesis.

In conclusion, contrary to exogenous RA-mediated granulocytic differentiation of myeloid leukemia cells that occurs through induction of MAT1 degradation and RARα hypophosphorylation, these RARα-CAK signaling events appear to proceed intrinsically during normal granulopoiesis. These are evidenced by (Fig. 7) (a) the correlation of stage-specific MAT1 degradation, progressive granulocytic development, and the dynamic expression of ALDH1A1 and ALDH1B1; and (b) the induction of RA-target gene expression and granulocytic differentiation resulting from loss of CAK-dependent RARα phosphorylation. Elucidating the mechanisms of RARα-CAK signaling in normal granulopoiesis and understanding how such mechanisms are subverted to promote HSC or myeloid progenitor transformation are crucial for the development of differentiation therapies for myeloid leukemia.

thumbnail image

Figure Figure 7.. RARα-cyclin-dependent kinase-activating kinase signaling events during dynamic expressions of ALDH1A1 and ALDH1B1 in normal granulopoiesis. Abbreviations: ALDH, aldehyde dehydrogenase; CMP, common myeloid progenitors; GMP, granulocyte/monocyte progenitors; HSC, hematopoietic stem cells; MAT1, ménage à trois 1; RA, retinoic acid; RARα, retinoic acid receptor α.

Download figure to PowerPoint

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

We thank Dr. D.L. Crowe for providing RARαS77A cDNA, Dr. D.B. Kohn for providing lentivirus pCCL plasmids, and Dr. G.M. Crooks for helpful suggestions. We acknowledge the FACS core facility of the Gene Immune and Stem Cell Therapy Program at Childrens Hospital Los Angeles for providing FACS services and L. W. Barsky and E. Zielinska for technical expertise in cell sorting. We thank the nurses at Kaiser Permanente Sunset, Los Angeles, CA, for umbilical cord blood collection. We thank Dr. W. Chen for some data collection. This work was supported by NIH Grants R21 CA111440 to L.W. and 5K01 DK066163 to K.J.P. P.L. and A.W. contributed equally to this work.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information
FilenameFormatSizeDescription
sc-07-0280-SupTable.pdf91KSupplemental Table

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.