Background
Overexpression of the
Ecotropic
Virus
Integration site 1 (
EVI1) gene, which has been observed in subsets of patients with acute myeloid leukemia (AML) [
1-
4], myelodysplastic syndromes (MDS) [
5-
7], chronic myeloid leukemia (CML) [
8-
10], and certain solid tumors [
11-
14], is often associated with poor therapy response and shortened survival [
1-
4,
7,
9,
11,
12,
15,
16]. In mouse bone marrow transduction/transplantation models, experimental expression of
Evi1 led to development of an MDS-like disease [
17], or to AML-like disease when co-expressed with other oncogenes [
18,
19]. It also enhanced the growth of xenograft tumors in severe combined immunodeficient (SCID) mice [
20].
In vitro,
EVI1 stimulated cellular proliferation and inhibited differentiation and apoptosis in some experimental models [
14,
17,
20-
29], but evoked opposite responses in others [
17,
29-
37], indicating that the consequences of
EVI1 overexpression may be influenced by cell lineage, maturation stage, cooperating molecular events, and/or environmental stimuli. EVI1 is believed to exert its varied biological functions predominantly by regulating gene transcription, and recently large-scale approaches have been applied to identify its target genes in ovarian cancer and murine myeloid cell lines [
38,
39]. A limited number of genes were shown to be regulated by EVI1 in a direct manner and to contribute to some of its biological effects, e.g.,
Gata2 [
24],
Pbx1 [
40],
Pten [
41],
Gpr56 [
42],
miR-1-2 [
43],
miR-9 [
44],
miR-124 [
45,
46], and
miR-449A [
47]. In light of the multitude of cellular responses to EVI1, however, its target genes and mechanisms of action are still far from completely understood.
The membrane-spanning 4-domains subfamily A member 3 (
MS4A3) gene was expressed in specific subsets of hematopoietic cells, including myeloid precursors, basophilic granulocytes, and CD34-positive hematopoietic stem and progenitor cells induced to differentiate
in vitro by exposure to granulocyte colony stimulating factor (G-CSF) [
48-
50]. MS4A3 was present in a complex with cyclin-dependent kinase 2 (CDK2) and kinase-associated phosphatase (KAP), which inactivates CDK2 by dephosphorylation of Thr160 [
50]. MS4A3 stimulated the enzymatic activity of KAP, and caused cell cycle arrest when expressed in human myeloid U937 cells in a regulable manner [
50,
51].
In the present study, we found that MS4A3 was repressed by EVI1 in several experimental model systems. This repression was mediated by direct binding of EVI1 to a proximal region in the MS4A3 promoter, and was necessary for the tumor promoting effects of EVI1 in a murine xenograft model.
Discussion
EVI1 is an oncogene whose overexpression is associated with high aggressiveness of both hematological and solid tumors [
1-
4,
7,
9,
11,
12,
15,
16]. Even though this correlation is well established, and the molecular structure, nuclear localization, and DNA binding ability of EVI1 suggest that it acts as a transcription factor [
60], the target genes and molecular mechanisms through which it contributes to the emergence and therapy resistance of malignant diseases are still understood only to a limited extent. Recently, genome-wide large-scale approaches have been applied to identify genes regulated by EVI1 in murine hematopoietic cells and a human ovarian cancer cell line [
29,
38,
39]. In the present study, we used a complementary approach and searched for genes whose expression levels changed in response to inducible expression of EVI1 in a human myeloid cell line. Among 56 bona fide EVI1-regulated genes, the
MS4A3 gene, coding for a member of a family of four-transmembrane proteins, was repressed most strongly after induction of EVI1.
MS4A3 was also down-regulated in primary murine hematopoietic cells inducibly expressing
Evi1 [
29], and its mRNA levels changed in the expected direction after manipulation of
EVI1 expression in three additional human myeloid cell line based models (Figure
1 C-E). When CD34-positive primary human hematopoietic stem and progenitor cells were differentiated into the granulocytic lineage
in vitro,
EVI1 levels decreased [
36] while
MS4A3 levels increased [
50] (and KS, unpublished results). Furthermore,
EVI1 expression was negatively correlated with that of
MS4A3 in a panel of human myeloid cell lines and in primary samples from AML patients (Figure
1F, Table
1). Reporter gene assays and ChIP showed that EVI1 regulated
MS4A3 by directly binding to the proximal 268 bp of its promoter. ChIP-seq on a murine leukemic cell line also identified an EVI1 binding site near the
Ms4a3 gene [
39], yet at a greater distance from its transcriptional start site, and the functional significance of this site was not further investigated. Previous studies have defined a number of different consensus EVI1 binding sites [
38,
39,
61-
67], but interestingly, none of these sites was found in the 268 bp region delineated through the luciferase assays, suggesting that EVI1 has the ability to recognize DNA motifs in addition to those identified in these earlier studies.
To date, little is known about the biological functions of
MS4A3. Donato et al reported that inducible expression of this gene in U937 cells retarded their re-entry into the cell cycle after release from S-phase arrest [
50]. Using a constitutive overexpression approach in the same cell line, we did not observe any effect of
MS4A3 on the cell cycle distribution of asynchronously proliferating cells (Figure
3A), or on re-entry into the mitotic cycle of cells synchronized in the same manner as described by Donato et al (JE, unpublished results). Possible explanations for this divergence are the use of different expression systems and/or different U937 sublines between the Donato and our own studies. However, additional investigations will be required to resolve this discrepancy.
The reciprocal expression patterns of
EVI1 [
36] and
MS4A3 [
50] (and KS, unpublished results) during
in vitro differentiation of primary human CD34-positive cells into the granulocytic lineage raise the possibility that repression of
MS4A3 may contribute to the differentiation inhibiting effect of EVI1 [
17,
20,
29]. However, ectopic expression of
MS4A3 in U937_EVI1 or U937_vec cells did not affect their differentiation in response to 25-OH Vitamin D3 (Additional file
3: Figure S2), indicating either that induction of
MS4A3 is a consequence rather than a cause of myeloid maturation, or that other model systems are required to reveal a potential differentiation promoting effect of
MS4A3.
A gene expression signature characterizing leukemic stem and progenitor cells as opposed to the bulk leukemic population was associated with poor outcome in AML, and low expression of
MS4A3 constituted part of this signature [
68].
MS4A3 was also significantly down-regulated in a cyclophosphamide-resistant CML cell line as compared to the corresponding parental line (GEO data set GDS2729 [
69]). We therefore asked whether repression of
MS4A3 could play a role in
EVI1-mediated drug resistance of human myeloid leukemic cells [
20,
27,
70], yet re-expression of
MS4A3 in U937_EVI1 cells did not re-sensitize them to drugs used in the treatment of AML (JE and SK, unpublished results). Nevertheless, a role for down-regulation of
MS4A3 in
EVI1-induced disease aggressiveness was obtained in a murine xenograft model, in which tumors formed by U937_EVI1 cells grew significantly faster than U937_vec tumors, while re-expression of
MS4A3 abolished this effect. Interestingly,
MS4A3 did not slow the growth of
EVI1-negative U937_vec tumors, suggesting either that endogenous
MS4A3 was expressed at saturating levels in this cell line, or that
MS4A3 specifically interfered with tumor growth on the background of the gene expression pattern evoked by EVI1. The first possibility would predict that U937_EVI1_MS4A3 and U937_vec_MS4A3 grew at equal rates. The observation that in fact U937_EVI1_MS4A3 cells formed significantly smaller tumors than U937_vec_MS4A3 cells discredits the former explanation in favor of the latter. The
in vivo phenotypes of
EVI1 and
MS4A3 are also notable in light of the absence of an effect of either of these genes on cellular proliferation in suspension cultures
in vitro. This suggests that specific aspects of the growth condition
in vivo, e.g., interactions with the tumor microenvironment, are required for them to reveal their impact on cell and tumor growth.
Methods
Cell lines, retroviral transductions, immunofluorescence analysis, and gene knockdown
Cell lines U937T_EVI1-HA, represented by clones E10 and E14, and U937T_vec, represented by clone P2, have been described previously [
34]. They were cultured in RPMI 1640 (Life Technologies, Carlsbad, CA, USA) containing 10% fetal bovine serum (FBS; Life Technologies), 0.5 μg/ml puromycin (Sigma-Aldrich, St Louis, MO, USA), 500 μg/ml hygromycin (PAA, Pasching, Austria), and 1 μg/ml tetracycline (Sigma-Aldrich) in a humidified incubator at 37°C and 5% CO
2. To induce EVI1 expression, exponentially growing cells were washed 3 times with PBS (Life Technologies) and resuspended in growth media without tetracycline. Control cultures were washed in the same manner but were resuspended in media with tetracycline.
Cell lines U937_EVI1, U937_vec [
20], HL60_Evi1, and HL60_vec [
55] were grown in RPMI 1640 containing 10% FBS and 1% Penicillin/Streptomycin/Glutamine (PSG; Life Technologies). The coding sequence of the human
MS4A3 gene (transcript variant 1, NM_006138.4) was amplified using cDNA from U937_vec cells, the primers listed in Additional file
5: Table S2, and Phusion High Fidelity Polymerase (New England Biolabs, Ipswich, MA, USA). PCR products were cloned into the retroviral vector pMIA-II_IRES-Ametrine using the BamHI and XhoI sites to yield pMIA-II_MS4A3-IRES-Ametrine. DNA sequencing was performed to confirm the identity and accuracy of the insert. Retroviral particles were generated and U937_EVI1 and U937_vec cells were infected using standard procedures. After 3 days, cells were sorted for Ametrine positivity on a FACS Aria (Becton Dickinson, Franklin Lakes, NJ, USA). MS4A3 expression was confirmed by immunofluorescence analysis (IF). In brief, cells were transferred onto cover slips coated with Cell-Tak™ Cell and Tissue Adhesive (Corning Incorporated, Corning, NY) and fixed with ice-cold methanol (Roth, Karlsruhe, Germany). IF was performed using rabbit polyclonal MS4A3 antibody HPA019210 (Atlas Antibodies; dilution 1:30) and the Rhodamine (TRITC)-AffiniPure F(ab′)2 Fragment Goat Anti-Rabbit IgG (H + L) secondary antibody (Jackson ImmunoResearch, West Grove, PA, USA; dilution 1:200).
UCSD-AML1 cells [
56] were maintained in RPMI 1640 supplemented with 20% FBS, 1% PSG, and 10 ng/ml GM-CSF (PeproTech, Rocky Hill, NJ). 2.25 × 10
6 cells from a logarithmically growing culture were resuspended in 400 μl of PBS and electroporated either with a mix of EVI1 siRNAs (stealth siRNAs HSS103423 and HSS103424, Invitrogen) or with scrambled control siRNA (stealth siRNA 462001, Invitrogen) at final concentrations of 100 nM. Electroporation was carried out in a Gene Pulser Xcell Electroporation System (BioRad, Hercules, CA) at 300 V and 1000 μF using an exponential protocol. Electroporated cells were incubated under standard growth conditions for 24 h prior to RNA extraction.
Gene expression microarrays and statistical and bioinformatics analyses
For gene expression microarray analyses, U937T_EVI1-HA E10 and U937T_EVI1-HA E14 cells were washed and placed into media with or without tetracycline for 6, 12, 24, and 48 h. To control for potential effects of tetracycline removal in the absence of EVI1 induction, U937T_vec P2 and U937T cells incubated in the presence or absence of tetracycline for 48 h were also included in the experiment. Total RNA was extracted using the RNeasy kit (Qiagen, Hilden, Germany) as recommended by the manufacturer. RNA quality control, sample labelling and hybridization to Affymetrix HG-U133 plus 2.0 microarrays (Affymetrix, Santa Clara, CA, USA) were performed at the Center of Excellence for Fluorescent Bioanalytics (KFB; Regensburg, Germany). Robust Multi-array Analysis was used for background correction, quantile normalization and median polish summarization of probe levels. Only probe sets with a current gene annotation and with average log
2-intensities ≥3 at 24 and 48 h in E10 and E14, and at 48 h in P2 and U937T cells, were included in downstream analyses. Because we had previously observed background effects of tetracycline withdrawal in control cells [
53], probe sets were considered as regulated by EVI1 only if they were induced or repressed at least two-fold both at 24 and 48 h after tetracycline withdrawal and both in E10 and E14 cells, and in addition the effect of tetracycline removal at 48 h in E10 and E14 cells was at least 10^(fold-change expression/3) the effect in the control cell lines P2 and U937T. If more than one probe set for the same gene was found to be regulated in this manner, the probe set with the most pronounced regulation was included in the heatmap, which was generated using Genesis [
71]. All other computational analyses and filtering procedures were performed using R and custom PERL scripts. Microarray data were deposited in the GEO database (accession number GSE60100).
GO term enrichment was analysed using the term-for-term algorithm of Ontologizer [
72]. P-values were calculated using one-sided Fisher exact test, and adjusted for multiple hypothesis testing according to Benjamini and Hochberg [
73]. An adjusted p-value <0.1 was considered statistically significant.
GEO datasets GSE6891 [
57], GSE14471 [
58], and GSE35784 [
59], which contain gene expression data from primary AML samples, were probed for differences in
MS4A3 expression between samples with high or low levels of
EVI1 by bootstrap analysis. To determine cutoff values defining high versus low
EVI1 expression, the density distributions of the log
2 transformed EVI1 mRNA levels were estimated using a Kernel Density Estimator (KDE). The
EVI1 expression values at which the density distribution exhibited a minimum were used as cutoffs for the respective data set. In datasets where several local minima existed, the minimum closest to the EVI1
low distribution with <5% of the maximal density defined the cutoff. The respective groups of EVI1
high patients were compared to randomly sampled, equally sized groups of EVI1
low patients. 10.000 iterations of this setup were performed, and in each step the difference between the mean log
2 transformed
MS4A3 expression values in both groups was calculated. Finally, the mean value of the resulting distribution (log
2-fold change, M) and the two-sided P-value using the inverse standard normal cumulative distribution function were determined.
To predict potential binding sites of EVI1 in the
MS4A3 promoter, 19 different position frequency (weight) matrices (PWMs) were newly compiled or derived from experimentally verified binding sites [
38,
39,
61-
67] or from the Matbase matrix library 8.4 (Genomatix) and JASPAR [
74] databases. Potential EVI1 binding sites in the genomic region from -268 to -1 relative to the transcriptional start site of
MS4A3 were identified based on a PERL implementation of the MatInspector algorithm [
75] if the similarity score for a specific PWM was equal to or above a threshold that was defined by allowing one binding site per 10 kb of human coding sequences. Genomic sequences were derived from the UCSC genome browser [
76].
qRT-PCR
Total RNA for qRT-PCR was extracted using Trizol (Life Technologies) and reverse transcribed using random hexamer primers (Life Technologies) and M-MLV reverse transcriptase (Life Technologies) according to the manufacturer’s instructions. qRT-PCR was carried out in a Step One Plus Real Time PCR system (Applied Biosystems, Life Technologies) using standardized cycling conditions as recommended by the manufacturer. Levels of
EVI1,
MS4A3, and the housekeeping gene
Cyclophilin D were determined using the primers listed in Additional file
5: Table S2 and the Mesa Green qPCR MasterMix Plus (Eurogentec, Eraing, Belgium). All assays were performed in triplicate. Expression values for the gene of interest relative to the housekeeping gene and to a reference value were determined using the ΔΔC
T method [
77]. At least three biological replicates were analysed and mean fold changes in expression and standard errors of the mean (SEM) were calculated.
Reporter vectors and luciferase assays
All vectors used for luciferase assays were based on pGluc basic (New England Biolabs). The
MS4A3 5′ region (-3213/+11 relative to the transcription start site of NM_006138) was amplified from human genomic DNA using Phusion® High-Fidelity DNA Polymerase (New England Biolabs) and the primers listed in Additional file
5: Table S2. The resulting PCR product was ligated to the EcoRV digested vector, yielding pMS4A3(-3213/+11)/pGluc. A series of 5′ deletion constructs (-1992/-1, -1441/-1, -1118/-1, -668/-1, -268/-1 and -3213/-279) was generated by PCR amplification using the cloned promoter fragment MS4A3(-3213/+11) as a template, followed by subcloning using EcoRI and BamHI restriction enzymes (Fermentas Inc., Hanover, MD, USA).
pMS4A3(-3213/+11)/tk/pGluc, pMS4A3(-268/-1)/tk/pGluc, and pMS4A3(-3213/-279)/tk/pGluc were generated by subcloning the HSV tk promoter into the BamHI, or the KpnI and BamHI, sites of the respective pGluc basic based constructs.
For luciferase assays, 6 × 105 U937 cells/well were seeded into 12-well plates. Transient transfections of reporter constructs and either empty pcDNA3 (Life Technologies) or pcDNA3-EVI1 (containing a codon optimized version of the human EVI1 cDNA) were performed using 1 μg DNA (reporter:effector ratio = 1:3) and 4 μl of JetPEI cationic polymer transfection reagent (Polyplus, Illkirch, France) according to the manufacturer’s instructions. After 48 h, 50 μl of culture supernatant were mixed with 50 μl of Gluc assay solution from the BioLux® Gaussia Luciferase Flex Assay Kit (New England Biolabs). The bioluminescent reaction was measured immediately by detecting the emitted photons at 475 nm using a Tristar LB941 (Berthold Technologies, Bad Wildbad, Germany). The values represent means + SEMs of three independent experiments.
Chromatin immunoprecipitation (ChIP)
ChIP assays were performed using the chromatin immunoprecipitation assay kit (Upstate Biotechnology, Lake Placid, NY, USA) as reported previously [
78]. In brief, 5 × 10
6 U937_EVI1 or U937_vec cells were fixed by treatment with 1% formaldehyde for 10 min and then lysed. Chromatin was sheared to fragments of 200 - 1000 bp using Bioruptor (Diagenode, Liege, Belgium). Immunoprecipitation was performed using rabbit monoclonal EVI1 antibody C50E12 (Cell Signaling, Danvers, MA, USA; dilution 1:80) or rabbit polyclonal EVI1 antibody sc-8707X (Santa Cruz Biotechnology, Santa Cruz, CA, USA; 1:250). Nonspecific IgG (2729, Cell Signaling, 1:200) was used as a negative control. Immunoprecipitated DNA was extracted with phenol/chloroform (Sigma-Aldrich), precipitated with ethanol, and dissolved in 30 μl Tris-EDTA buffer (Sigma-Aldrich). 2 μl of recovered DNA were subjected to PCR analysis using the primers shown in Additional file
5: Table S2 and HotStarTaq DNA Polymerase (Qiagen). Cycling conditions were: 95°C for 12 min, followed by 32 cycles at 95°C for 30 s, 60°C for 40 s, and 72°C for 30 s, and a final incubation step at 72°C for 7 min. PCR products were separated on a 2% agarose gel stained with GelRed (Biotium, Hayward, CA, USA).
Analyses of myelomonocytic differentiation and of cell cycle distribution
To analyse myelomonocytic differentiation of U937 cells, logarithmically growing cells were seeded to a density of 2 × 105 cells/ml and incubated either with 100 nM 25-OH-Vitamin D3 (Calbiochem, La Jolla, CA) or with an equivalent amount of solvent (EtOH) for 5 days. Cells were diluted once during this period to avoid saturating densities, and fresh 25-OH-Vitamin D3 was added at the same time. After blocking of nonspecific epitopes with Human TruStain (Biolegend, San Diego, CA), cells were stained with monoclonal rat APC-Cy7 conjugated CD11b antibody (clone M1/70, Biolegend) or corresponding isotype control (clone RTK4530, Biolegend) using standard procedures. Flow cytometric analysis was performed on an LSRFortessa™ SORP (BD Biosciences, Bedford, MA, USA).
For cell cycle analyses, cells were adjusted to a density of 400 cells/μl. On the next day, cells were washed with PBS (Life Technologies) and incubated for 5 min in ice cold 0.5 M citrate/0.5% Tween-20. Cell membranes were disrupted mechanically before nuclei were pelleted and resuspended in PBS containing 100 μg/ml RNase A (Sigma-Aldrich) and 50 μg/ml propidium iodide (PI; Sigma-Aldrich). Nuclear DNA content was determined on a FACSCalibur™ (BD Biosciences) or a FACS LSRFortessa™ SORP using ModFit software (Verity Software House, Topsham, ME, USA) for data analysis.
Xenograft experiments and immunohistochemistry
Animal experiments were approved by the ethics committee of the Medical University of Vienna and the Bundesministerium für Wissenschaft und Forschung Ref. II/10b (Gentechnik und Tierversuche), application Nr. BMWF-66.009/0095-II/10b/1010, and were carried out according to the Austrian and FELASA guidelines for animal care and protection in order to minimize distress for the animals. Mice were sacrificed by cervical dislocation.
Six to eight week old male CB-17 scid/scid (SCID) mice were purchased from Harlan Laboratories (San Pietro al Natisone, Italy). The animals were kept in a pathogen-free environment and all procedures were performed in a laminar airflow cabinet. 5 × 106 U937_vec_vec, U937_vec_MS4A3, U937_EVI1_vec, or U937_EVI1_MS4A3 cells, resuspended in 50 μl of serum-free RPMI 1640 medium, were injected subcutaneously into the right flanks of 4 mice per cell line. Animals were controlled every day and tumor size was assessed regularly by caliper measurement. Tumor volume was calculated using the formula: (length × width2)/2. At experiment termination, mice were dissected and tumor tissue was processed for immunohistochemistry (IHC). For statistical analysis of tumor growth, two-way ANOVA and Bonferroni post-correction were applied.
IHC was performed using standard procedures. Briefly, 4 μm sections from xenograft tumor blocks were deparaffinized and rehydrated, heated for 10 min in 10 mM citrate buffer (pH 6.0) in a pressure cooker for epitope retrieval, and then incubated for 60 min at room temperature with rabbit monoclonal EVI1 (clone C50E12, Cell Signaling Technology; dilution 1:200) or rabbit polyclonal MS4A3 (HPA019210, Atlas Antibodies; dilution 1:50) antibodies, or for 30 min with mouse monoclonal Ki-67 antibody (MIB-1, Dako, Glostrup, Denmark; dilution 1:100). Antibody binding was detected by means of the UltraVision LP detection system (Lab Vision, Thermo Fisher Scientific, San Jose, CA, USA) according to the manufacturer’s recommendations. Color development was performed by 3-3′-diaminobenzidine (Dako) and counterstaining by hematoxylin (Merck, Vienna, Austria). Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) was carried out using the in situ cell death detection kit, TMR Red (Roche, Mannheim, Germany) according to the manufacturer’s instructions.
Images of stained tumor sections were acquired with TissueFAXS (TissueGnostics, Vienna, Austria). Percentages of TUNEL-positive cells were determined using TissueQuest software (TissueGnostics).
Acknowledgements
This work was supported by the Austrian Science Foundation (FWF), grants no P17896-B14, P20920-B12, and P21401-B12 to RW, and grant no F4709-B20 to SZM. The FWF did not have any role in the design of the study, the collection, analysis, and interpretation of data, the writing of the manuscript, or in the decision to submit it for publication.
pMIA-II_IRES-Ametrine was a kind gift from Dr. Dario Vignali of the St. Jude Children’s Research Hospital, Memphis, Tennessee, USA. Anita Brandstetter of the Medical University of Vienna is gratefully acknowledged for performing IHC analyses.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.
The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
GH designed and performed experiments and GO analyses, analyzed data, prepared figures and tables, and participated in writing of the manuscript. AR, BS, JE, KS, MF, PH, ET, SK, BZ, TT, and AS designed and performed experiments and analyzed data. HH performed statistical analyses of microarray results and bioinformatics analyses. SZM and WB interpreted data. RW designed, supervised, and coordinated research, interpreted data, and wrote the manuscript. All authors critically reviewed the final version of the manuscript and approved of its contents.