Background
The bone marrow (BM) microenvironment consists of different cell types, grouped in “niches”, defined according to the cellular composition and also the signals produced [
1,
2]. Detailed knowledge of the regulation and composition of the BM niches, including the osteoblastic and the vascular niches, is essential for our understanding of BM function and may also contribute towards the discovery of therapeutic targets to treat BM diseases. Emerging evidence suggests the “vascular niche”, and bone marrow endothelial cells (BMEC) in particular, conveys signals to hematopoietic progenitor and stem cells, promoting BM recovery via instructive “angiocrine” signals that tightly regulate the hematopoietic differentiation process [
3,
4]. The coordinated production and release (in such cases) of instructive signals is crucial for adequate BM recovery and function; hematopoietic differentiation and exit into peripheral organs is tightly regulated by the instructive signals from the BM vascular niche [
3,
5]. In particular, the communication between BMEC and hematopoietic elements, namely the hematopoietic stem and progenitor cells, is crucial for normal BM function and for maintaining BM integrity following stress. Whole body irradiation has been used to study BM turnover, since it rapidly induces BM cells apoptosis and allows detailed study of the mechanisms involved in BM cell recovery [
6,
7]. Whole body irradiation is also clinically relevant as a majority of cancer patients, most notably those suffering from hematological malignancies, receive some form of radiotherapy [
8,
9]. Little is known about the role of miRNAs in the BMEC that contribute to regulate the BM function and recovery. miRNAs are recognized post-transcriptional regulators of gene expression through mRNA targeting and/or translational repression thereby modulating biological homeostasis [
10]. In the present study, we discovered that miR-363-5p (previously termed miR-363*) is expressed by EC and is induced by irradiation
in vivo and
in vitro. miR-363-5p regulates the expression of angiocrine factors in EC, affects EC angiogenic properties and also modulates the interaction between EC and hematopoietic cells.
Discussion
Similar to solid tumors, neo-vessel formation (angiogenesis) has been associated with disease progression in hematological cancers including leukemias and lymphomas [
12‐
14]. This structural role of blood vessels in the BM is therefore linked to the needed increase in nutrients and oxygen to “feed” the expanding malignant cell clones. Nevertheless, recent evidence suggests BM vessels may play a more integrative role in the BM microenvironment, providing “instructive” or “angiocrine” cues to hematopoietic cells, this way maintaining BM homeostasis [
3]. This dynamic interaction between endothelial cells of BM vessels and hematopoietic elements (immature, undifferentiated precursors and differentiated progeny) should be tightly regulated, maintaining the balance between mature and immature hematopoietic cells and contributing towards BM recovery when needed. Nevertheless, the molecular signals that regulate angiocrine factor production and BM function are largely unknown and were the subject of the present study.
miRNAs are key regulators of gene expression at post-transcriptional level and are implicated in a wide range of biological functions including cell proliferation, differentiation, apoptosis, among many others [
15]. miRNAs deregulation is associated with several cancers [
16]. The miRNA expression profiles show that the vast majority of the miRNAs is down-regulated in many cancers [
17]. Interestingly, there has been a significant interest in the identification of miRNAs that selectively regulate EC function, namely during tumor angiogenesis. The term “angiomiRs” was coined a few years ago, to include the miRNAs that regulate particular EC functions [
18]. Nevertheless, to our knowledge, a specific miRNA regulating angiogenic and angiocrine properties on EC was not reported. We reasoned that the molecular profiling of BM EC exposed to stress might reveal the genes and target pathways involved in the homeostatic function of EC in BM microenvironment. For this, we used a well-established approach to induce BM stress (which consisted of whole body sub-lethal irradiation), isolated the EC from irradiated or control BM and discovered a set of miRNAs that are induced on BMEC following whole body irradiation. The miRNA profiling of irradiated BMEC revealed a large number of differentially expressed miRNAs from the non-irradiated control. Subsequent validation of miRNA induction by irradiation
in vitro allowed further mechanistic studies to be developed.
In detail, we identified one particular miRNA (miR-363-5p) that is induced by irradiation and selectively regulates EC properties, including the expression of angiocrine factors that are involved in the communication between BMEC and hematopoietic precursor cells. Interestingly, miR-363-5p previously named miR-363* is a miRNA generated from the upload of the miRNA* strand into RISC, which confirms the earlier observation that miRNA* are not always degraded and are active post-transcriptional gene regulators [
19]. miR-363-5p was consistently induced by irradiation and was found to modulate the expression of angiocrine factors. miR-363-5p regulates (directly and indirectly) the expression and availability of well known angiocrine and hematopoietic factors (although their altered expression could, in fact occur as a result of a feedback response to miR-363-5p modulation per se). We showed by luciferase reporter assays that miR-363-5p regulates the expression of TIMP1 and THBS3 at post-transcriptional level. Variations in the expression of these targets at the protein level will be addressed in future studies in our Laboratory, particularly since THBS3 biological functions are largely unknown, and thus the effect of its regulation by miR-363-5p and the relevance for bone marrow homeostasis remains to be investigated.
As there were no reports in the literature about the pathways or genes that miR-363-5p could regulate and as target identification can be complex, we followed a transcriptome analysis upon forced modulation of the miR-363-5p levels. This strategy was previously reported [
20] and was particularly useful to narrow-down the search of direct targets of a given miRNA. As miRNAs action depends on the miRNA-target stability, the strategy used, although allowed the identification of miRNA targets can result in the under-estimation of the targets identified, which can be dozens for a single miRNA [
21]. Importantly, along with the targets directly regulated by miR-363-5p, its function may be enhanced by indirect mechanisms, as shown with the indirect regulation of kit ligand (stem cell factor). Kit ligand is essential for normal BM function and for BM recovery following irradiation [
22]. In addition, TIMP1 is an inhibitor of matrix metalloproteinase function [
23]. Earlier studies had shown the activation of MMP9 in the BM microenvironment tightly regulates the cleavage of kit ligand from a membrane bound to a soluble form, promoting BM recovery through hematopoietic precursors mobilization, differentiation and proliferation [
24]. In the present report we show that EC with reduced miR-363-5p promote hematopoietic precursors adhesion and expansion, which is accompanied (and may be at least partially explained) by increased SCF production and release. The data presented here suggests that regulation of miR-363-5p expression on BMEC may regulate the
availability of SCF in the BM microenvironment, highlighting the relevance of this particular miRNA in BM homeostasis.
Methods
Mice irradiation and isolation of BMEC
All animal experiments were performed with the approval of the Instituto Gulbenkian de Ciência Animal Care Committee and Review Board. The mice were sub-lethally irradiated as previously published [
25,
26]; in the present study, FVB mice were used as subjects. miRNA profiling was performed in whole BM and in BMEC: the whole BM were extracted after irradiation as previously described and mononuclear cells were separated using Ficoll (Histopaque-1077). The BMEC were isolated using anti-CD31 fluorescein-conjugated Abs (1:100, Chemicon) and recovered using fluorescence-activated cell sorting (Modular Flow Cytometer, Beckman Coulter). Purity of recovered BMEC was ≥ 95%. Cells were centrifuged and the pellet was kept in TRIzol for further RNA extraction. Total BM from non-irradiated and irradiated mice was also used for RNA extraction as above.
RNA extraction, cDNA and qRT-PCR
Total RNA was extracted using TRIzol Reagent (Invitrogen™) according manufacturer instructions. mRNA and miRNAs were quantified according to the two protocols following described: for miRNA quantification, the cDNA was synthesized from 500 ng of total RNA using the NCode™ miRNA first-strand synthesis (Invitrogen™). miRNAs were quantified by quantitative RT-PCR (qRT-PCR) with SYBR Green (Invitrogen™) using a Universal qRT-PCR primer provided and primers to target specific miRNAs. Two μl of diluted cDNA (1:2) were used as template in 20 μl qRT-PCR reactions with 10 μM of each primer and 1x Platinum SYBR Green qRT-PCR Super Mix-UDG. The expression of U6 was used as endogenous control. miRNA levels were calculated using the comparative method 2^(−ΔΔCt) [
27]. To perform the quantification of coding genes, the cDNA was synthesized from 1 μg of total RNA and random hexamers and Superscript II (Invitrogen) were used according to manufacturers instructions. Two μl of diluted cDNA (1:2) were used as template in 20 μl qRT-PCR reactions with 10 μM of each primer and SYBR Green (Applied Biosystems). The relative mRNA levels were normalized against 18S rRNA expression and calculated using the comparative method 2^(−ΔΔCt) [
27]. All quantifications were performed with an ABI PRISM 7900HT Sequence Detection System (Applied Biosystems). Primers used in this study are available in Additional file
9. All reactions were run in triplicate.
miRNA profiling
Total RNA was isolated using TRIzol Reagent (Invitrogen). Quality and quantity of total RNA was analysed using the Agilent 2100 Bioanalyser (Agilent Technologies) and NanoDrop 2000. The miRNA profiling was performed using the miRCURY LNA (locked nucleic acid) V10.0 microarrays (Exiqon). The labeling was performed according to manufacturers recommendations using the miRCURY LNA microRNA Power labeling kit (Exiqon) from1 μg of total RNA (BMEC and whole BM). The microarrays used for miRNAs expression profiling comprised a total of 1154 probes from the miRBase Sequence Database version 8.0 (
http://microrna.sanger.ac.uk) [
28,
29]. The hybridized microarrays were washed, dried and scanned using a dual-laser Agilent Technologies scanner. Scanned images were analyzed using Feature Extraction Software (Agilent Technologies), which converts scanner-generated images into quantitative log2 ratios. Labelling efficiency was evaluated by the signals from the control spike-in capture probes. Background correction and normalization was performed using the Local Nearest Neighbour algorithm Lowess (locally weighted scatterplot smoothing regression algorithm), which uses multiple local backgrounds in the neighbourhood of a given spot to serve as background signal for that feature. Expression values were presented as log2 ratio of red signal/green signal. Log2 ratio errors and associated p-values, which determine the probability that a log ratio is significantly different form zero, were also calculated. The final expression values of each miRNA correspond to the average of the quadruplicates spots within the slide. The expression values were submitted to Nudge algorithm to identify differentially expressed miRNAs. TIGR Multiple Experiment Viewer software package (MeV version 4.1; [
30]) and Excel was used to perform data analysis and visualize the results.
Cell culture, transfection and functional analyses of miRNAs
Endothelial cells (Human Umbilical Vein Endothelial Cells - HUVEC) were cultured in EBM-2 complete medium supplemented with 5% foetal bovine serum (FBS). Passages older than 6 were not used in the transfection experiments. The anti-miR-363-5p, pre-miR-363-5p and scramble control (Ambion) were electroporated using HUVEC at 70% confluency at final concentration of 50 nM according to the manufacturers protocol (Nucleofector V kit VCA-1001; Amaxa). After 12 h, the media was replaced with fresh media. Experiments were performed 24 h post-electroporation.
The HUVEC (100,000 cells per well) transfected with anti-miR-363-5p or scramble control were seeded on 24-well plate coated with 200 μl of Growth Factor Reduced Matrigel (BD Biosciences). Branches were quantified after a 16 h incubation at 37°C. Three biological replicates were performed for each condition. Photographs were captured at 20x magnification using an Olympus Microscope.
Immunofluorescence
The HUVEC were fixed 48 h post-transfection with PFA (2%) for 10 min at room temperature. Cells were washed twice in PBS 1x and incubated with PBS/0.1%BSA for 30 min at room temperature. Cells were stained with phalloidin (1 μg/ml) for 30 min at room temperature, washed and mounted with Vectashield containing DAPI. Preparations were examined using a fluorescence microscope (Axioplan Microscope, Zeiss).
PCR array
Expression profiles of angiogenesis-related genes were generated using PCR Arrays (SABiosciences) in accordance with the manufacturers recommendations in an ABI PRISM 7900HT Sequence Detection System (Applied Biosystems). The RNA of HUVEC transfected with anti-miR-363-5p or scramble control was extracted as above and 1 μg was treated with DNase followed by cDNA synthesis using RT2 First Strand kit (SABiosciences). cDNA samples were mixed with RT2 qRT-PCR master mix and distributed across the PCR array 96-well. Fold-change of each gene from anti-miR-363-5p to scramble control HUVEC was reported as Log10(2^(−ΔΔCt)). Fold-changes greater or less than 0 are indicated as up- or down-regulation, respectively.
CD34+ isolation and adhesion experiment
Mononuclear cells from human cord bloods were separated over Ficoll and CD34+ progenitor cells were further isolated using magnetic beads (Miltenyi Biotec) following manufacturers instructions. Purity of the CD34+ was assessed by flow cytometry. CD34+ progenitor cells (1000 cells) were added to resting HUVEC monolayers that were previously (24 h before) transfected with anti-miR-363-5p, pre-miR-363-5p or scramble control on 24-well plates. As a rescue experiment, to test whether CD34+ cells adhesion to HUVEC transfected with anti-miR-363-5p, pre-miR-363-5p or scramble control could be blocked by increasing the levels of SCF, recombinant human stem cell factor (Life Technologies) was added (10 ng/mL).
Adherent CD34+ cells onto HUVEC were counted manually using a phase-contrast microscope. Triplicates were used for each experimental condition.
Cord blood-derived CD34+ cells isolated as described above were assessed for CFU frequency by culturing them in methylcellulose (R&D Technologies) according to manufacturers instructions. Cells were cultured in triplicate for seven days after which colonies were counted and morphologically analyzed.
Luciferase reporter assay
For the luciferase reporter experiments, the UTRs of putative miR-363-5p target genes were amplified and cloned in a luciferase reporter vector (pMIR-REPORT, Ambion) downstream the luciferase gene. Specifically, the wild-type UTRs having the miRNA binding(s) sites were amplified by PCR using primers having
Spe I and
Hind III restriction sites. The PCR products and the vector were then digested with
Spe I and
Hind III, cloned into pMIR-REPORT and transformed into
E. coli (One-Shot, Invitrogen). All constructs were verified by sequencing. Resulting plasmids were co-transfected (1.5 μg) into HUVEC with anti-miR-363-5p or pre-miR-363-5p or scramble (50 nM) and pRL-SV40 vector (0.5 μg) (Promega), which contains a Renilla Luciferase gene to normalize transfection rates. Mutation of seed sequence of the miRNA-binding site was performed using the Site-Directed Mutagenesis Kit (Promega) and the mutated plasmids were used for transfection as above. Primers used are listed in the Additional file
9. Luciferase activity was assayed after 48 h using the Dual-Luciferase Reporter System (Promega) and was normalized against Renilla activity. Results represent Luciferase/Renilla ratios of three independent experiments.
ELISA for Stem Cell Factor (SCF)
The SCF levels were quantified in the supernatants of HUVEC transfected with anti-miR-363-5p, pre-miR-363-5p or scramble control, 48 h post-transfection by ELISA (R&D Systems), according manufacturers guidelines. The assay was performed twice and error bars represent s.e.m. of three transfection experiments.
Transcriptomic studies
Gene expression profiles were performed using microarrays (Affymetrix GeneChip Human Gene 1.0 ST array) using total RNA extracted with TRIzol from HUVEC transfected with anti-miR-363-5p, pre-miR-363-5p or scramble control. The level of miR-363-5p (upon forced reduction or increase) was confirmed by qRT-PCR before performing the microarrays. GeneChip Hybridization and scanning were performed at Instituto Gulbenkian de Ciência (
http://www.igc.gulbenkian.pt).
Statistical analysis
Experiments were performed at least three times, unless indicated. Significance was performed using Student’s t test (p ≤ 0.05) or Anova, where indicated. Graphs show the standard error of the mean (s.e.m.) using Student’s t test. Single, double and triple asterisks indicate statistically significant differences: * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001.