| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
,25-dihydroxy-vitamin D3 in HL-60 phagocytes
* INRS-Institut Armand-Frappier, Laval, PQ, Canada; and
Department of Biochemistry, McGill University, Montréal, PQ, Canada
Correspondence: M. F. M. Cellier, INRS-Institut Armand-Frappier, 531 Bd des prairies, Laval, H7V 1B7, PQ, Canada. E-mail: mathieu.cellier{at}inrs-iaf.uquebec.ca
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
induced similar up-regulation. NRAMP1 protein expression paralleled the accumulation of mRNA and was localized in the phagosomal membrane after phagocytosis. A promoter construct extending 647 bp upstream of NRAMP1 ATG showed myeloid-specific transcription in transient transfection assays, which was up-regulated by VD in HL-60. In HL-60 clones stably transfected with this construct, transcription was apparently induced through indirect VD genomic effects, and there was accordance between the levels of reporter transcription and endogenous NRAMP1 mRNA in response to VD but not to IFN-. Thus, VD genomic effects stimulate NRAMP1 transcription and protein expression in maturing phagocytes.
Key Words: monocytic differentiation • interferon- • Northern blot • immunofluorescence • luciferase reporter gene
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Nramp1 was discovered in the laboratory mouse by positional cloning of the genetic factor responsible for innate susceptibility to several intravacuolar parasites such as Mycobacterium spp., S. typhimurium, and Leishmania donovani [6 ]. A recessive point mutation is responsible for inactivation of the encoded protein [7 ]. The human ortholog NRAMP1 was located in a conserved group of synteny at 2q35 [8 ]. NRAMP1 and Nramp1 genes are expressed specifically in the myeloid lineage (monocytes, macrophages, and polymorphonuclear neutrophils) [9 , 10 ], contrasting with their parologous genes NRAMP2 and Nramp2, which are expressed almost ubiquitously and required for iron homeostasis [11 ]. In vitro maturation of human monocytes (Mn) into macrophages is accompanied by NRAMP1 mRNA up-regulation, consistent with mRNA levels in explanted alveolar macrophages that are higher than those found in blood monocytes. The human leukemia cell line HL-60 was shown to be a useful model to study the regulation of NRAMP1 gene expression during experimentally induced granulocytic and monocytic differentiation [9 ].
NRAMP1 gene expression is induced in HL-60 cells treated with the active form of vitamin D [1
, 25-(OH)2D31,25-dihydroxyvitamin D3 (VD)], a seco-steroid hormone known to stimulate the production and terminal maturation of bactericidal macrophages. VD is viewed as a general regulator of cell growth, function, and differentiation, because virtually all cells can express the VD receptor (VDR), which acts as a transcriptional regulator in the cell nucleus after binding VD. In vivo, the VD endocrine system is linked to calcium and phosphate metabolism [12
]. VD is also recognized as a specific regulator of the differentiation and activation of the macrophage. VD influences myelopoiesis, favoring the development of the granulocyte-macrophage colony-forming precursor toward the mononuclear lineage [13
] and further stimulating the differentiation of monocytes toward the macrophage and osteoclast cell types [12
, 14
], but not toward monocyte-derived dendritic cells [15
]. VD also stimulates the phagocytosis and expression of natural defenses by macrophages [16
, 17
], which could suggest a basis for the positive effects on tuberculostasis shown for this hormone [18
].
Susceptibility to tuberculosis involves several predisposing genes inherited in a complex manner. Allelic polymorphisms in the VDR gene [19
] and others at the NRAMP1 locus [20
, 21
] were associated with mycobacterial diseases. VD deficiency was also correlated with increased risk of tuberculosis [22
]. In vitro studies indicate that VD pretreatment enhances MN control of the intracellular growth of mycobacteria, and VD action can be reinforced by a combination with interferon- (IFN-) and tumor necrosis factor (TNF-) [23
]. VD, alone or in combination with IFN- or TNF-, also induces the expression of the inducible nitric oxide synthase (iNOS) and suppresses intracellular growth of Mycobacterium tuberculosis in human monocytes [24
]. Because VD is an important modulator of macrophage differentiation and activation that could be relevant to tuberculosis resistance, we undertook a study of the kinetics and possible mechanisms controlling NRAMP1 protein expression during VD-induced differentiation of HL-60 cells.
VD mediates its biological activities primarily at the level of gene transcription. The mechanisms of action of VD seem to operate via two independent receptors, the VDR, a member of the family of steroid hormone receptors [25
], and a cell surface receptor identified as annexin II [26
, 27
]. The interaction of VD with a particular receptor is determined by the shape of the molecule of VD and conformationally restricted VD analogs enabled to discriminate between genomic (VDR-mediated) and (rapid) nongenomic effects of VD [28
]. Nongenomic and genomic effects of VD appear required for monocytic differentiation [29
, 30
]. Rapid (within minutes), nongenomic effects of VD include alterations of ion fluxes (e.g., Ca2+, Cl-) and activation of protein kinase C (PKC) and mitogen-activated protein kinases (MAPK) [12
, 31
]. In this study, we used the 6-s-cis conformer of VD, 1,25(OH)2-lumisterol3 (JN), which is unable to bind to VDR, as an agonist of VD nongenomic effects, and the analog 1ß,25(OH)2D3 (HL) as an antagonist of the VD nongenomic effects [27
, 31
, 32
]. VD genomic effects are mediated by the VDR in homo or heterodimers with other steroids or retinoid receptors and additional transcriptional coregulators [33
]. Several genes, including the cyclin-dependent kinase inhibitor p21waf1/cip1 and the homeobox transcription factor HOXA10, are targets for VDR-dependent up-regulation in myeloid cells at the onset of VD-induced differentiation [34
]. In this study, we have used the analogs 20-epi-22-oxa-24a,26a,27a-tri-homo-1,25(OH)2D3 (KH 1060) and [1(S), 3(R)-dihydroxy-20(R)-(5'-ethyl-5'-hydroxy-hepta-1'(E),3'(E)-dien-1'-yl)-9,10-secopregna-5(Z),7(E),10(19)-triene] (EB 1089) as potent agonists of VD genomic effects.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Differentiation agents
9-cis-Retinoic acid (RA) and all-trans-RA (ATRA; TRC Chemicals, Toronto, Canada) were dissolved in ethanol at 4 x 10-3 M and 2.5 x 10-3 M as stock solutions. Phorbol 12-myristate 13-acetate (PMA) was from Sigma Chemical Co. (St. Louis, MO). 1,25(OH2)D3 and analogs KH 1060 and EB 1089 were generously provided by Dr. Lise Binderup (Leo Pharmaceutical Products, Ballerup, Denmark). Stock solutions were in isopropanol at 4 x 10-3 M and were stored at -20°C protected from light. Analogs JN and HL were a kind gift of Dr. Anthony Norman from the University of California (Riverside). Stock solutions were in ethanol at 1 x 10-4 M and stored at -20°C protected from light. Actinomycin D obtained from Sigma Chemical Co. was dissolved in 10% ethanol. Recombinant human IFN- was obtained from Endogen (Woburn, MA).
Flow cytometric analysis and monoclonal antibodies (mAb)
Phenotypic analysis of cells was performed using mouse mAb detected by fluorescein isothiocyanate (FITC)-conjugated, affinity-purified donkey anti-mouse immunoglobulin G (IgG; Jackson ImmunoResearch, West Grove, PA). The following mAb were used: 3C10 (IgG2b, anti-CD14) and OKM1 (IgG2b, anti-CD11b) from ATCC. Cultured cells (5x105) were harvested, centrifuged, and resuspended in blocking solution [phosphate-buffered saline (PBS)-0.1% bovine serum albumin (BSA) supplemented with 5% donkey or goat normal serum depending on secondary Ab used] for 15 min. A 30-min incubation on ice followed with primary antibody or isotype-matched control (Cedarlane, Canada). This was followed by three washing steps with PBS, and finally with a 30-min incubation with secondary antibody. After washing three times, cells were resuspended in PBS-5% BSA. Ten thousand cells were analyzed considering a viable cell gate as delimited by forward and right-angle light scatter parameters using an Epics flow cytometer (Coulter, Hialeah, FL).
Phagocytosis
HL-60 cells were seeded in poly-D-lysine culture slides (Biocoat, BD-Falcon, Becton Dickinson, Franklin Lakes, NJ) and were exposed to differentiation inducers for 3 or 5 days. Culture medium was replaced by fresh medium containing heat-inactivated Candida albicans (Ca) at a cell:Ca ratio of 1:100. Incubation was at 37°C for 1 h. Following extensive washing steps, preparations were incubated again for 1 h to enable complete phagocytosis. Slides were washed and processed for light microscopy or epi-immunofluorescence as described below.
Immunofluorescence
Cells were allowed to differentiate on culture slides for phagocytosis experiments. Following 20 min fixation in Bouin’s solution, cells were washed extensively, permeabilized for 10 min in 0.1% Triton-X 100 in PBS, washed again in PBS and PBS-100 mM glycine, and blocked for 1 h at room temperature in PBS-1% BSA-20% heat-inactivated donkey or goat serum depending on secondary antibody. Primary antibody was incubated in blocking solution for 16 h at 4°C [rabbit polyclonal serum anti-NRAMP1 (see below, 1/50 dilution)] or for 1 h at room temperature (anti-c-Myc, 1/200 dilution, 9E10; Santa Cruz Biotechnology, Santa Cruz, CA). After washing with PBS-0.5% Tween 20, secondary antibodies were incubated for 1 h at room temperature (donkey R-phyco-erythrin-anti-rabbit, 1/200, Jackson ImmunoResearch; or goat FITC-anti-mouse, 1/400, Sigma Chemical Co.). After washing with PBS-0.5% Tween 20, culture slides were mounted with GelTol (Shandon, Pittsburgh, PA) and examined using a Leitz microscope with a 40x objective (400x magnification). Polyclonal- serum was generated against the N-terminus of human NRAMP1 protein, which was produced as a recombinant antigen fused to the glutathione-S-transferase as described previously [34
]. The oligodeoxynucleotides 5HF 5'-TCG GAT CCT CAA TGA CAG GTG AC-3' and 5HR 5'-TAG AAT TCG CAG GCT GAA GGT G-3' were used to polymerase chain reaction (PCR)-amplify a fragment of 156 bp using 0.2 unit Taq DNA polymerase (Gibco-BRL, Burlington, Ontario), 1 µM each oligonucleotide, 25 µM dXTP, and 1.25 mM MgCl2 in a 25 µL reaction of 1x reaction buffer supplied by the manufacturer. Plasmid C4 [8
] was used as template, and 25 cycles were performed (45 s at 94°C, 45 s at 42°C, 45 s at 72°C). The PCR fragment obtained was cloned after digestion with BamHI and EcoRI into the vector pGEX-2 [35
] and was sequenced with primers 5HF and 5HR. One clone encoding the linker peptide GSS followed by NRAMP1 sequence MTGDK. ... . TFSLR was used to immunize rabbits. The NRAMP1 insert was excised from the pGEX-2 vector and was subcloned into the bacterial expression vector pQE40 (Qiagen, Mississauga, Ontario) to produce a His6-dihydrofolate reductase fusion chimera, which was used to affinity-purify the anti-NRAMP1 polyclonal serum as described previously [11
]. The specificity of the purified antibody was established by Western blots of crude extracts and crude membrane preparations of transfected cells expressing a NRAMP1-c-Myc-tagged construct as described previously [11
].
RNA isolation and Northern hybridization
Total RNA was isolated from 107 cells by the guanidine isothiocyanate method (Tri-reagent, Sigma Chemical Co.). Following electrophoresis in 1.2% agarose/formaldehyde gel, RNA was blotted onto GeneScreen + membranes (NEN Life Science Products, Boston, MA) in 20x saline sodium citrate (SSC). After cross-linking with UV, blots were prehybridized overnight at 65°C in 1 M NaCl, 10% dextran sulfate, 1% sodium dodecyl sulfate (SDS), and 200 µg/mL heat-denatured herring sperm DNA. Hybridization probes for NRAMP1 and CD14 genes were as described [9
] using 2 x 106 cpm/mL hybridization mix. The same blots were probed for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) to assess equal loading per well, using 5 x 104 cpm/mL hybridization mix. Hybridization was performed overnight at 65°C in prehybridation solution containing 2 x 106 cpm/mL denatured [32P]-dCTP-labeled random-primed DNA probe. Membranes were washed one time in 1x SSC, two times with 1x SSC-0.1% SDS-0.02% lauryl sarcosyl, then three times with 0.1x SSC-0.1% SDS-0.02% lauryl sarcosyl. Filters were then exposed to Kodak XAR X-ray films at -70°C with intensifying screens or to a phosphorimaging system (Molecular Dynamics, Sunnyvale, CA).
Construction of human NRAMP1 promoter-Luciferase reporter
Restriction endonucleases were purchased from New England BioLabs (NEB; Mississauga, Ontario), except SacI, BglII, XbaI, and SmaI, which were from Amersham Pharmacia Biotech (APB; Baie d’Urfé, Quebec). Enzymatic reactions were performed following the supplier’s suggested conditions. A 290-bp fragment (base numbering starting at NRAMP1 ATG) was PCR-amplified using as template the 5' promoter region of NRAMP1 gene cloned in plasmid pKSB2-5 [8
] and the cloned Pfu DNA polymerase (Stratagene, La Jolla, CA) and primers ER1 5'-GGGAGGGGAACAAA GGTCCACT-3' (sense) and ER2 5'-CTGCCxATGGxAAATGCCGACTTCA-3' (antisense). The bases that were mutated to introduce an NcoI cleavage site surrounding the ATG are underlined. PCR conditions included five cycles (45 s at 94°C, 45 s at 55°C, 1 min at 72°C) followed by 25 cycles (45 s at 94°C, 45 s at 60°C, 1 min at 72°C). The PCR product of 290 bp was purified using a microspin S300 HR column (APB), digested with NcoI and ligated with the T4 DNA ligase (APB) to the pGL3 basic vector (Promega, Madison, WI) treated with NcoI and the calf intestine phosphatase (APB). The orientation of the insert was determined by digestion with EagI. One clone corresponding to the promoter fragment in sense orientation showing a sequence identical to the parental DNA was denominated pGL3 NR1S, and another corresponding to the antisense orientation was named pGL3 NR1AS.
A 645-bp promoter construct (pGL3 NR1L) was generated by ligation of three gel-purified fragments using the Geneclean II kit (Bio101, Vista, CA), including a 480-bp fragment produced by digestion of the plasmid pKSB2-5 with EagI and BstY and a 4276-bp EagI-BsrGI fragment derived from pGL3 NR1S and a 661-bp BsrGI-BglII fragment derived from the vector pGL3 basic. Two independent clones giving the expected pattern of restriction fragments after digestion with EagI, PstI, and NcoI were selected.
A shorter construct containing 584 bp of the promoter (pGL3 NR1M) was obtained by digestion of the plasmid pGL3 NR1L with the restriction endonucleases PstI and SacI and was treated with the T4 DNA polymerase (NEB) to produce blunt ends. The fragment of 5327 bp was gel-purified using the Geneclean II kit (Bio101) and self-ligated with T4 DNA ligase (APB). Two independent, resulting clones were verified with the restriction endonuclease BamHI and XhoI, giving the expected fragment of 5323 bp.
The construct pGL3 NR1dL containing a promoter region of 3629 bp was obtained using the 5350-bp NheI-NsiI DNA fragment derived from plasmid pGL3 NR1L and the 3032-bp XbaI-NsiI DNA fragment produced with the plasmid pB2-5. These restriction fragements were gel-purified using QIAquick gel extraction kit (Qiagen) and were ligated with T4 DNA ligase (NEB). The resulting clone was verified with the restriction endonucleases BamHI, BglII, and SmaI, giving the expected fragment of 8382 bp and with PstI, ScaI, and NcoI.
Transient tranfection assays
Large-scale plasmid purifications of all the constructions were performed with the EndoFree MaxiPrep kit (Qiagen) according to the manufacturer’s recommendations. The plasmids containing the luciferase reporter construction are described in the preceding section. The plasmid pXP2 with cytomegalovirus promoter driving the expression of the human growth hormone-coding sequence was a gift from Dr. D. G. Tenen (Harvard Institutes of Medicine, Boston, MA). Electroporations of the HL-60 cell line were performed as described [36
] with the following modifications: Day 2 (D2) differentiated cells (10-8 M VD and 10-8 M KH or the equivalent volume of solvent alone) were diluted to a concentration of 5 x 105 cells/mL. On D3, cells were electroporated with 19 µg reporter plasmid and 1 µg control pXP2 human growth hormone (hGH) plasmid at 300 V and 975 µF, incubated on ice for 15 min, and resuspended in 9 mL complete RPMI 1640. Cells were incubated at 37°C for 1 h before reincubating in the presence of inducer. Cells were incubated further at 37°C for 3 h before being harvested for luciferase extraction (Promega) and hGH quantification (Roche Diagnostics, Laval, Quebec). The tests were performed according to the manufacturer’s protocols. Jurkat cells were electroporated in conditions similar to those for HL-60 cells. 293-T cells were transfected with cationic lipids according the manufacturer indications (GenePORTER, Gene Therapy Systems, San Diego, CA).
Generation and analysis of stable transfectants
The promoter of the integrative expression vector Sr puromycin was removed by partial restriction of the plasmid with EcoRI and HindIII, and the fragment of 4 kb was purified using the QIAquick gel extraction kit (Qiagen). DNA overhangs were filled-in with the T4 DNA polymerase (NEB) in the presence of dXTP. Following ligation with the T4 DNA ligase (NEB), two independent clones verified by restriction with HindIII, SmaI, and BamHI were selected and named Sr-. This vector was digested with restriction endonucleases SmaI and BamHI, treated with the calf intestinal phosphatase (NEB), and gel-purified. It was ligated to a 2770-bp DNA fragment, which was gel-purified after digestion of plasmid pGL3 NR1L with endonucleases BamHI, NotI, and AseI. This fragment was ligated in the digested vector Sr- using the T4 DNA ligase (NEB). One clone (SrL) giving the expected length when digested with BamHI was sequenced using the primers ER2 and RVp3 (Promega). The plasmid SrL was produced with the EndoFree MaxiPrep kit (Qiagen). The day preceding the electroporation, the cells were washed and resuspended at a concentration of 5 x 105 cells/mL. Cells (1.4x107) were electroporated in the same conditions as for the transient transfection assays with 40 µg undigested DNA, incubated on ice for 15 min, resuspended in 10 mL complete RPMI medium, and incubated at 37°C for 48 h. Viable cells (
4x106) were seeded into 96-well trays at 4 x 104 cells/well in a total volume of 200 µL, and puromycin was added at a concentration of 1.5 µg/mL. After 2 weeks, half the medium was replaced with fresh medium containing puromycin. After 1–3 additional weeks, clones resistant to puromycin were detected in few wells (0.3%) and were amplified separately. Clones were tested by reading luciferase activity before and after differentiation with KH 10-8 M for 3 days, which was monitored by analysis of CD14 expression by flow cytometry and Northern blot. For luciferase assays, cells (2x105-1x106) were washed once with PBS, resuspended in 100 µL reporter lysis buffer (Promega), and then subjected to two freeze-and-thaw cycles. The lysate (20 µL) was used to quantify the luciferase as described in the transient transfection assays. Luciferase activity measured with a luminometer Lumat LB9507 (Berthold, Australia) was normalized with determination of total protein levels in each extract with the BCA-200 protein assay kit (Pierce, Rockford, IL) according to the manufacturer’s protocol. Seven clones were isolated out of five electroporations (frequency, 1x10-7). Of these, three were positive for luciferase expression and named clones HSRL1, HSRL4, and HSRL5.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
Cooperative effects of VD genomic agonists and IFN- stimulate NRAMP1 gene expression
VD and IFN- can exert additive effects on macrophage activation for bacteriostatic functions [42
] and on monocytic differentiation of HL-60 cells [43
]. Because IFN- is a key immunoregulatory cytokine required for tuberculostasis [23
] and up-regulates the autocrine production of VD by tissue macrophages [44
], we analyzed the effect of the combination of IFN- and VD compounds on NRAMP1 gene expression.
IFN- alone is a poor inducer of HL-60 differentiation and CD14 cell-surface expression when compared with VD (Fig. 2A
). However, the combination of IFN- with VD genomic agonist EB strongly stimulated cell-surface expression of CD14 (Fig. 2A)
. Figure 2B shows strong up-regulation of CD14 mRNA for the combination of EB and IFN- and little accumulation after stimulation with IFN- alone, thus indicating that CD14 cell-surface expression levels reflected mRNA accumulation levels. Determination of NRAMP1 mRNA levels in the same conditions revealed a very similar pattern of up-regulation, with high levels for the combination of the VD genomic agonist EB and IFN-, followed by the combination of VD compounds, then VD alone, and low NRAMP1 levels for cells induced only with IFN-. Hence, NRAMP1 and CD14 genes were similarly up-regulated in response to the combination of EB and IFN-.
|
Together, the results indicate that NRAMP1 and CD14 mRNA have different kinetics and turnover of expression. The combination of VD genomic agonist and IFN- had similar cooperative effects on NRAMP1 and CD14 gene expression, despite their different regulation. This could be a result of cross-talk between the corresponding signaling pathways acting at the level of gene transcription, because VD and IFN- can trigger rapid responses to initiate monocytic differentiation. Up-regulation of NRAMP1 gene expression was also observed after 3-day treatments using combinations of VD and granulocyte macrophage-colony stimulating factor (GM-CSF) or interleukin-6. The effect was dependent on the dose of cytokine used but remained modest in comparison with the combination of EB and IFN-. Up-regulation of NRAMP1 mRNA levels was even stronger in cells treated with VD, IFN-, and GM-CSF (unpublished results), a combination shown previously to exert synergistic effects on HL-60 differentiation [45
] and on activation for macrophage bacteriotoxic functions [46
, 47
]. Therefore, the strong up-regulation of the NRAMP1 gene in response to the combination of VD genomic agonist and IFN- probably reflects a cooperative effect of these compounds to induce macrophage differentiation and up-regulation of antibacterial functions.
Contribution of VD genomic and nongenomic effects to the induction of NRAMP1 gene expression
To gain further insight into the mechanism of action of VD on NRAMP1 gene expression, various treatments were assayed combining the native hormone VD and conformationally restricted VD analogs or retinoids (RA, ATRA). Combinations with other independent regulators of cellular transcriptional activity that stimulate macrophage maturation and activation, such as phorbol ester and IFN-, were also assayed. The effects of selected combinations on the expression of NRAMP1 and CD14 genes were determined after 3 days by Northern analysis (Fig. 3A
), phosphorimaging, and normalization of signal intensities to the level of GAPDH signal per sample (Fig. 3B)
. Combination treatments (20 nM VD compounds; e.g., 2VD, 20 nM VD; VD KH, 10 nM VD; and 10 nM KH) were performed to evaluate the relative influence of genomic versus rapid responses induced by VD on NRAMP1 gene expression.
|
Combinations of VD and retinoids such as the RA receptor (RAR) ligand, ATRA, and the RXR ligand, 9-cis RA, were also assayed because retinoids are known to potentialize HL-60 differentiation through the formation of VDR heterodimers [48 ]. Equimolar combinations with VD (Fig. 3A and 3B , lanes VD ATRA 10-8 M, and VD 9-cis RA 10-8 M) appeared similarly effective than VD alone to stimulate CD14 and NRAMP1 gene expression. Induction levels for both genes were inferior to those obtained with combinations of VD and genomic agonists (Fig. 3A and 3B , lanes KH VD and EB VD). 9-cis RA showed clear dose-dependent inhibition of NRAMP1 and CD14 expression (Fig. 3 , lane VD 9-cis RA 10-9 M), whereas ATRA inhibited minimally NRAMP1 and CD14 mRNA levels (Fig. 3 , lane VD ATRA 10-9 M and VD ATRA 10-7 M). These results suggest that NRAMP1 gene expression induced by VD can be modulated by cotreatment with 9-cis RA, further supporting a role for VDR in the stimulation by VD of NRAMP1 gene expression. The relatively weak effects observed with combinations of VD genomic agonists and retinoids contrast with the known abilities of these compounds to stimulate cellular differentiation strongly and could be a phenomenon of the HL-60 cell line [48 ]. These results support the view that the NRAMP1 gene is probably not a direct target for VDR and suggest that VDR-dependent up-regulation of NRAMP1 expression might be mediated via homodimers of this transcription factor [40 , 49 ].
To examine a possible contribution of the non-VDR signaling pathway, the effects of combinations of VD and the analogs JN and HL were tested. Equimolar combinations of VD and the agonist of VD nongenomic effects JN (Fig. 3A and 3B , lane VD JN 10-8 M) gave expression levels for CD14 or NRAMP1 mRNA that are similar to those obtained with VD (lane 2VD) and inferior to those obtained with VD and KH or EB (Fig. 3 , lanes KH VD or EB VD). In addition, dose-dependent stimulation of VD effects was observed with JN compound on mRNA levels of CD14 and NRAMP1 (Fig. 3 , compare lanes VD JN 10-7 M and VD JN 10-9 M). However, no stimulation was observed using JN as sole inducer (unpublished results). The combination of equimolar amounts of VD and the antagonist of nongenomic VD effects (HL) showed a strong inhibitory effect on the expression of NRAMP1 and CD14 genes induced by VD alone (Fig. 3A and 3B , lanes VD HL 10-8 M and 2VD). This effect was not clearly dose-dependent in the conditions tested (Fig. 3A and 3B , lanes VD HL 10-9 M and VD HL 10-7 M). Together, the results suggest that in addition to VDR-dependent signaling, the rapid, nongenomic responses induced by VD could contribute to up-regulate the NRAMP1 gene expression.
The JN agonist is thought to act in part through PKC stimulation [50 ]. Up-regulation of NRAMP1 mRNA observed in response to JN (Fig. 3A and 3B) suggested that PKC could stimulate NRAMP1 gene expression. The phorbol ester PMA is a pharmacological stimulator of PKC, whose isoform ß (PKC-ß) is required for macrophage differentiation of HL-60 cells [51 ], and a combination of VD and PMA increases macrophagic differentiation of HL-60 cells and iNOS-dependent production of NO [52 ]. In addition, previous studies showed that PMA induced accumulation of NRAMP1 mRNA in HL-60 cells, albeit less efficiently than VD [9 ]. Cotreatment of HL-60 cells with VD and PMA potentialized NRAMP1 gene expression, particularly in combination with low concentrations of PMA (Fig. 3A and 3B , lanes VD PMA 1–100 ng/mL). In contrast, CD14 gene expression was relatively inhibited. The latter result is consistent with the absence of CD14 mRNA accumulation in PMA-treated HL-60 cells [9 ]. Together, the results suggest that the VDR-dependent regulation of NRAMP1 mRNA may be modulated via PKC signaling, possibly involving nongenomic effects of VD.
The most potent differentiation treatment tested so far included KH, IFN-, and GM-CSF. It was about fivefold more efficient at inducing NRAMP1 gene expression than the equimolar combination of JN, IFN-, and GM-CSF (unpublished results). Thus VDR-mediated genomic effects of VD would be predominant in stimulating NRAMP1 expression, and nongenomic signals induced by VD could also contribute to this process. In addition, potent transcriptional activators that are important for macrophage differentiation and that signal through independent pathways exerted cooperative effects with VD to up-regulate NRAMP1 gene expression. Together, the data indicate a link between macrophage differentiation, up-regulation of antibacterial functions, and NRAMP1 gene expression.
IFN- and PMA cooperate with KH to induce NRAMP1 protein expression in HL-60 cells
NRAMP1 gene expression induced by VD seemed to correlate with terminal differentiation, and combinations of VD with PMA or IFN-, which potentialize the induction of phagocytic and bactericidal functions in HL-60 cells [43
, 52
], induced the highest levels of NRAMP1 accumulation (Fig. 3)
. Therefore, we hypothesized that HL-60 cells differentiated with combinations of the VD genomic agonist KH and PMA or IFN- should express NRAMP1 protein at detectable levels. A polyclonal antibody specific for NRAMP1 was generated. The purified polyclonal serum recognized NRAMP1 specifically in crude membrane preparations of stable U-937 transfectants (Fig. 4K
). The minor bands observed could be a result of differences in glycosylation of the protein or could indicate limited proteolysis during membrane preparations. This specific antibody was used to study the protein expression by indirect immunofluorescence. This technique enabled previous detection of the endogenous membrane protein in the mouse [53
] and suggested by analogy that NRAMP1 may be concentrated in intracellular vesicles of phagocytes. Therefore, an in situ immunofluorescence technique was chosen to favor detection in case the protein may be expressed at relatively low levels in HL-60 cells.
|
(Fig. 4A
4B
4C
4D
4E
4F
and 4J)
or 10 nM KH and 10 nM PMA (Fig. 4G 4H
4I)
. As expected from Northern studies (Figs. 2
and 3)
, NRAMP1 protein was expressed at high levels in HL-60 cells differentiated with KH and IFN- (Fig. 4A
and 4J)
and KH and PMA (Fig. 4H
and 4I) . Detection of NRAMP1 protein required cell permeabilization and revealed an intracellular punctate pattern in resting cells (Fig. 4A)
, which appeared similar to the intracellular distribution shown previously for Nramp1 in mouse macrophages [1
]. Nonpermeabilized cells (Fig. 4B)
and permeabilized cells analyzed with the preimmune serum (Fig. 4C)
did not show comparable signal (Fig. 4A
4B
4C)
. HL-60 cells differentiated with KH and IFN- were also analyzed with an anti-CD14 antibody, used as a control for cell-surface protein expression independent of cell permeabilization (Fig. 4D 4E
4F)
. Cell-surface fluorescence specific to CD14 was observed with cells that had been permeabilized (Fig. 4D)
or not (Fig. 4E) but was not observed when omitting anti-CD14 antibody (Fig. 4F) . Thus, high levels of NRAMP1 protein expression are correlated with the highest levels of NRAMP1 mRNA (Figs. 2
and 3)
, suggesting that NRAMP1 expression may be controlled primarily at the level of transcription and mRNA stabilization.
Redistribution of NRAMP1 protein localization was observed after phagocytosis of heat-killed C. albicans by HL-60 cells differentiated with KH and PMA (Fig. 4G
4H
4I)
. Figure 4I
shows NRAMP1-specific intracellular fluorescence in resting cells that had not been challenged with any particles. Phase-contrast and fluorescence visualizations of the same field are shown in Figure 4G
and 4H , respectively. Extracellular and intracellular Candida are identified in Figure 4G
by the small and large arrows, respectively. The fluorescent image of the same intracellular particle, marked in Figure 4H
by a replicate large, white arrow, shows decoration of the membrane surrounding the particle with anti-NRAMP antibody (Fig. 4H)
. Similar images were visualized with phagocytic HL-60 cells differentiated with KH and IFN- (Fig. 4J)
. The pattern observed was apparently similar in resting cells (Fig. 4I)
and in cells that did not ingest particles (Fig. 4H)
, suggesting that phagocytosis itself triggered the subcellular redistribution of NRAMP1 protein toward the phagosomal membrane. Additional in situ analyses of protein expression in freshly explanted cells showed relatively little NRAMP1 protein in adherent MN, whereas most of MN-derived Mac were positive for NRAMP1 protein expression, including in the presence of VD (unpublished results). Therefore, NRAMP1 protein expression seems to parallel mRNA levels detected in MN and MN-derived Mac [9
], suggesting that the expression of the antimicrobial NRAMP1 protein is a result of transcriptional up-regulation in terminally differentiated phagocytes [54
]. These results suggest a physiological significance for VD-dependent NRAMP1 expression in HL-60 cells, thus emphasizing the usefulness of this model for the study of the regulation of NRAMP1 gene expression.
The 5'-proximal region of the NRAMP1 gene is VD-responsive and contains myeloid-specific elements
NRAMP1 promoter activity was quantified by luciferase reporter assays using five constructs representing different lengths of the NRAMP1 5'-untranslated region. Luciferase activity was analyzed 4 h after electroporation of differentiated HL-60 cells, which had been treated for 3 days with a combination of VD and KH and reincubated with the inducers after transfection. A 263-bp NcoI fragment containing the NRAMP1 ATG mutated to an NcoI site was inserted in both orientations (S, sense; AS, antisense) in 5' upstream of the luciferase open-reading frame (ORF) from pGL3 basic plasmid. As expected, only the S construct gave significant reporter activity in nondifferentiated HL-60 cells [175 relative luciferase units (RLU/s); Fig. 5A
], suggesting that basic elements for transcription initiation were present in this short fragment. Luciferase activity driven by the S construct was not up-regulated in VD-differentiated HL-60 cells (Fig. 5A)
. Three other constructs extending 588 bp, 647 bp, and 3633 bp upstream of the ATG, respectively, M, L, and dL, were also active in HL-60. Constructs M and L exhibited similar luciferase activity (500 RLU/s; Fig. 5A
; P<0.05), representing approximately threefold the level measured for construct S, and which was further stimulated twofold after treament with VD and KH (1100 RLU/s; Fig. 5A
; P<0.05). This suggests that the 5' region of the NRAMP1 gene extending 588 bp upstream of the ATG contained cis elements conferring high-level expression in untreated HL-60 cells, as well as VD responsiveness in these cells. Finally, construct dL exhibited relatively lower activity in untreated HL-60 cells, suggesting that negative elements may be located further upstream of the proximal promoter (600 bp). VD responsiveness could still be detected with the dL construct (Fig. 5A)
, but the low levels of expression dampen the analysis. These results demonstrate that transcriptional up-regulation can contribute to NRAMP1 protein expression during VD-induced differentiation and further suggest that cis-active elements conferring VD responsiveness in HL-60 may be localized between 588 and 262 bp upstream of NRAMP1 ATG.
|
350 RLU/s; Fig. 5C
), whereas constructs M and L were also very efficient (300 RLU/s; Fig. 5C
), and construct dL showed slightly lower levels of reporter activity (300 RLU/s; Fig. 5C
). A similar picture emerged from transfection studies in the lymphoid cell line Jurkat (Fig. 5B)
. The AS construct again showed negligible activity, whereas the sense constructs S and L gave similar levels of reporter activity (400 RLU/s). Maximum and minimum levels of reporter activity were observed with constructs M and dL, respectively (550 RLU/s and 125 RLU/s). Lymphocytes constitute a target tissue for some immunosuppressive activities of VD that are mediated by the VDR [55
]. Therefore, we tested in conditions similar to the studies with HL-60 cells whether the NRAMP1 promoter showed similar VD responsiveness in Jurkat T cells. Importantly, we found that none of the constructs tested were responsive to VD induction in Jurkat cells. Together, these results demonstrate that the segment of 263 bp upstream NRAMP1 ATG contains minimal elements required for basal transcriptional activity, which is not restricted to myeloid cells. In contrast, possible myeloid-specific determinants are contained in the region spanning from 588 and 262 bp upstream of the ATG of the NRAMP1 gene. This region also contains elements that are required to mediate myeloid-specific VD responsiveness.
Transcriptional activation of NRAMP1 647-bp promoter is induced through VD genomic effects in stable transfectants
Sequence analyses of the 647-bp promoter region using the software MatInspector V2.2 [56
], the database Transfact, and the software SignalScan [57
] with the Transfact [58
] and Transcription Factor databases [59
] failed to detect any VD response element. The region was also inspected visually for any motif of sequence A/G G G/T T N N (3 or 6) A/G G G/T T N N, which represents the consensus VDR-binding site [60
, 61
]. Therefore, it appeared more likely that NRAMP1 promoter activity did not result from direct binding of VDR but rather from stimulation by other transcription factors that may be activated later in the monocytic differentiation program. As a first step to study the soluble factors responsible for the activation of the NRAMP1 promoter, stable transfectants were established in HL-60 cells to verify the transcriptional activity of the 647-bp construct in a chromosomal context.
The level of luciferase activity obtained after 3 days of differentiation induced with VD, KH, and EB was quantified in three independent clones (Fig. 6A ). Induction ratios in the range of 1.5–2.5 were obtained with VD and in the range of 2.5–4.5, with VD genomic analogs EB and KH. Kinetic studies of luciferase expression in clone HSRL5 confirmed the delayed expression of the NRAMP1 gene (Fig. 6B) . Luciferase expression was induced significantly by VD and the genomic agonist EB at day 2 of differentiation, reaching a peak of expression at day 3. These data are consistent with the levels and kinetics of expression that were measured with the same inducers by Northern blot analysis (Fig. 1) , confirming that the region starting 647 bp upstream of the ATG is sufficient to drive VD-dependent expression of NRAMP1.
|
because Northern analyses revealed that IFN- is a weak inducer of NRAMP1 expression, which stimulates EB- and VD-dependent expression strongly (Figs. 2
and 3
, respectively). However, we could not detect IFN--dependent stimulation of the NRAMP1-luciferase construct in transient transfection assays using VD-differentiated cells (unpublished results). Here, strong expression of the NRAMP1 construct was induced by IFN- alone, and merely additive effects were obtained by a combination of IFN- with VD or EB to stimulate luciferase expression (Fig. 6B)
. These results were unexpected from our previous Northern analyses (Fig. 2)
and indicated that unlike the regulation by VD genomic agonists, there is a clear difference in the response to IFN- between the transcriptional activity of the NRAMP1 647-bp promoter construct and the regulation of NRAMP1 mRNA levels. The clone HSRL5 was also used to verify that the transcriptional activity of the NRAMP1 647-bp construct in response to induction with the VD genomic agonists was dose-dependent (Fig. 6C) . KH analog was used to show it is between ten- and 100-fold more potent than VD to induce NRAMP1 transcription, which is consistent with the relative potency for induction of monocytic differentiation shown for this analog [62 ]. In contrast, the nongenomic agonist of VD, JN, showed no noticeable effect on the promoter region conferring VD responsiveness during monocytic differentiation. This result further suggests that the NRAMP1 647-bp promoter is the key determinant for the gene and protein expression in monocyte/macrophage in response to VD genomic effects. Because NRAMP1 mRNA accumulation and gene transcription are not significant before 48 h of differentiation and because no VDRE was found in the 647-bp region, it is logical to propose that transcription factors distinct from VDR, presumably involved in myeloid-specific expression, are the primary transacting factors for NRAMP1 gene expression.
The response of the NRAMP1 647-bp construct to IFN- was also dose-dependent, and combination with the VD-genomic agonist KH had only additive effects on luciferase activity (Fig. 6C)
, as observed in combinations of IFN- with VD and EB (Fig. 6B)
. To demonstrate that the stimulation of the NRAMP1 promoter construct by IFN- did not reflect the endogenous regulation accurately, we compared NRAMP1 mRNA levels and NRAMP1-dependent luciferase activity in the clone HSRL5. NRAMP1 and CD14 mRNA were stimulated very weakly in response to IFN- compared with KH, and IFN- had a potentializing effect on the accumulation of these mRNA induced by KH (Fig. 7A
and 7C
). These results are consistent with our previous analyses in HL-60 cells (Figs. 2
and 3)
, but this picture is clearly different from the results of luciferase activity obtained with clone HSRL5 (Fig. 7B)
, which shows the abnormally elevated response of the NRAMP1 promoter construct to IFN-. These results suggest that the transcriptional response to IFN- requires for proper regulation additional cis elements that are not present in the 647-bp promoter fragment of the NRAMP1 gene.
|
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
, by itself a poor inducer of NRAMP1 and CD14 genes, strongly stimulated their expression at the mRNA and protein levels. Immunofluorescence analyses of NRAMP1 protein expression in HL-60 cells differentiated with VD and IFN- showed a pattern typical of intracellular vesicles in resting cells. After phagocytosis, NRAMP1 protein was redistributed at the phagosomal membrane, where it is believed to mediate its antimicrobial function. Similar relocalization of NRAMP1 protein expression after phagocytosis was also observed after differentiation with VD and phorbol ester, a treatment that does not induce CD14 expression. Thus, HL-60 represents a valuable model to study the regulation of expression of NRAMP1 during the myelomonocytic differentiation induced by VD and more potent pharmacological VDR ligands. VDR ligands are known as potent immunomodulators that act directly on monocytes and their progeny (macrophages, dendritic cells) and on T-lymphocyte populations and that exert a dual role on the immune system. On one hand, their action on myeloid dendritic cells and T lymphocytes inhibits specific immune responses [63
, 64
], and on the other hand, their action on promyelocytes, monocytes, and macrophages stimulates innate, nonspecific immunity by the up-regulation of microbicidal functions such as iNOS [65
] and reduced nicotinamide adenine dinucleotide phosphate oxidase [56
, 66
]. It is interesting that the differentiation of monocytes toward dendritic cells is accompanied by down-regulation of NRAMP1 expression [67
] and is inhibited in the presence of VD [64
, 68
]. In this context, it is tempting to suggest that up-regulation of NRAMP1 expression at the level of the phagosome by VDR ligands in monocytes could contribute to enhance innate immunity. Thus, it will be important to characterize the molecular control of NRAMP1 gene expression by VDR ligands. The HL-60 model presented here will be valuable because we demonstrated that a DNA fragment extending 647 bp upstream of NRAMP1 ATG was sufficient to mediate transcriptional activation in response to VDR ligands in a dose- and time-dependent manner consistent with the induction of HL-60 differentiation. Of note, assays of NRAMP1 promoter activity in epithelial and lymphoid cell types gave patterns of transcriptional activation, which were similar between these two unrelated, nonmyeloid cellular backgrounds but clearly distinct from the activity observed in HL-60. This could suggest that NRAMP1 promoter activity in HL-60 cells is representative of myeloid cells, and further work will be required to demonstrate that this promoter operates in other myeloid cell lines. Contrarily to the regulation by VDR ligands, there was a clear discrepancy between NRAMP1 gene transcriptional activity and mRNA accumulation in response to IFN-. Thus, it is likely that additional cis elements outside the 647-bp DNA fragment contribute to regulate the expression of the NRAMP1 gene. Among the putative binding sites for transcription factors that may be important for VDR-dependent regulation is a motif located in the region spanning from -588 to -262 bp upstream of the ATG, which appears conserved with mouse and chicken promoter Nramp1 sequences [9
]. This site could bind the transcription factor PU.1 that is required for tissue-specific expression of various effector functions of professional phagocytes [9
, 69
]. Further studies are required to identify cis-acting elements contained in the basic promoter of the NRAMP1 gene and those involved in the transcriptional up-regulation of the NRAMP1 gene in response to VD in myeloid cells and to localize additional IFN--responsive regions. These studies will reveal some molecular determinants controlling the expression of a divalent metal ion transporter that is expressed in mature phagocytes and important for host resistance to infection.
|
ACKNOWLEDGEMENTS |
|---|
Received October 18, 2001; revised January 16, 2002; accepted January 22, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
