Dual Role of miR-21 in CD4+ T-Cells: Activation-Induced miR-21 Supports Survival of Memory T-Cells and Regulates CCR7 Expression in Naive T-Cells

Immune cell-type specific miRNA expression patterns have been described but the detailed role of single miRNAs in the function of T-cells remains largely unknown. We investigated the role of miR-21 in the function of primary human CD4+ T-cells. MiR-21 is substantially expressed in T-cells with a memory phenotype, and is robustly upregulated upon αCD3/CD28 activation of both naive and memory T-cells. By inhibiting the endogenous miR-21 function in activated naive and memory T-cells, we showed that miR-21 regulates fundamentally different aspects of T-cell biology, depending on the differentiation status of the T-cell. Stable inhibition of miR-21 function in activated memory T-cells led to growth disadvantage and apoptosis, indicating that the survival of memory T-cells depends on miR-21 function. In contrast, stable inhibition of miR-21 function in activated naive T-cells did not result in growth disadvantage, but led to a significant induction of CCR7 protein expression. Direct interaction between CCR7 and miR-21 was confirmed in a dual luciferase reporter assay. Our data provide evidence for a dual role of miR-21 in CD4+ T cells; Regulation of T-cell survival is confined to activated memory T-cells, while modulation of potential homing properties, through downregulation of CCR7 protein expression, is observed in activated naive T-cells.


Introduction
Acquisition of effective, long-term immunity requires development of memory CD4+ T-cells. This process, induced by activation of a naive T-cell, involves extensive transcriptional activity. Consequently, naive and memory T-cells are characterized by distinct gene expression patterns [1,2]. However, not only the mere presence of transcripts, but also their regulation e.g. by microRNAs (miRNAs) is crucial for proper development and function of T-cells [3,4]. MiRNAs, a class of small, single-stranded RNA molecules, regulate gene expression at the post-transcriptional level. By binding to partly complementary target sequences in the 3' UTR of the mRNA, miRNAs induce degradation or translational inhibition of the targeted mRNA [5]. Like coding genes, expression of miRNAs is dynamically regulated during activation and differentiation of T-cells [6][7][8]. Consequently, various effector T-cell subsets are characterized by distinct miRNA expression profiles [9,10]. However, the contribution of single miRNAs in the function of individual T-cell subsets is still largely unknown.
Several miRNAs are highly expressed in freshly isolated human memory CD4+ T-cells [9]. Amongst these, miR-21 has anti-apoptotic properties which have been extensively studied in pathological conditions including cancer, cardiovascular disease and autoimmunity [11][12][13][14]. Indeed, miR-21 was recently shown to suppress apoptosis and induce proliferation of primary murine and human T-cells [15][16][17][18]. In accordance with the pronounced expression in memory T-cells, it has been shown that miR-21 can be induced upon activation of CD2+ Tcells [6]. However, the kinetics and degree of miR-21 upregulation, as well as the differential functional consequences thereof in naive and memory T-cells remain unknown.
In addition to apoptosis-related genes, bioinformatic analysis [19,20] of putative miR-21 targets relevant for T-cell biology revealed several immune-related genes, including CCchemokine receptor 7 (CCR7), which is substantially expressed on naive T-cells [21]. By binding with its ligands (CCL19 and CCL21) presented on the surface of high endothelial venules, CCR7 enables entry of T-cells into lymph nodes, and as such ensures (re)circulation of naive T-cells through the lymphatic system (reviewed in 22).
In this study we focused on the physiological role of high miR-21 expression in memory T-cells and the physiological consequences of activation-induced miR-21 expression in naive T-cells. By inhibiting endogenous miR-21 function during activation we showed that the survival of CD4+ T-cells, resulting from anti-apoptotic activity of miR-21, is confined mostly to the memory T-cell compartment. In contrast, activation-induced upregulation of miR-21 in naive T-cells posttranscriptionally regulates expression of the lymph node homing receptor CCR7. Thus, we provide evidence for a divergent role of miR-21 in two key aspects of T-cell biology, i.e. survival of memory T-cells and potential homing properties of naive T-cells.

MiR-21 expression is associated with memory T-cell phenotype and is induced upon αCD3/CD28 activation of CD4+ T-cells
MiR-21 expression was assessed in naive and memory CD4+ T-cells isolated from PBMC of healthy volunteers (as depicted in Figure S1A). The expression of miR-21 was detectable in both T-cell subsets, and in agreement with previous studies we observed that memory T-cells expressed higher levels of miR-21 than naive T-cells (median fold difference 5.1, p=0.016, Wilcoxon test) ( Figure 1A). No differences in miR-21 expression were observed between central and effector memory T-cells, indicating that high miR-21 expression is a general feature of the memory T-cell phenotype (Figure S1B-D). We next studied the kinetics of miR-21 induction upon T-cell activation. We assessed if and to what extent αCD3/CD28 activation can induce miR-21 expression in freshly isolated naive and memory T-cells. The expression of miR-21 was strongly induced in response to activation of both naive (median fold increase 13.8 at day 7, p<0.001, Friedman test) and memory T-cells (median fold increase 4.6 at day 7, p<0.001, Friedman test), indicating that miR-21 upregulation is a general feature of T-cell activation ( Figure 1B). Both T-cell subsets responded to the stimulation in a similar manner in regard to the increase in cell size (recorded by forward-and side-scatter) and expression of IL-2 receptor α (CD25). By day seven, all naive T-cells had obtained a memory phenotype as defined by acquisition of CD45RO expression ( Figure S1E and F).

MiR-21 mediates survival of activated memory CD4+ Tcells
To determine the functional role of activation-induced miR-21 expression we stably inhibited its function using a lentiviralbased, antisense miR-21 expression system (miR-21 inhibitor, Figure S2A) in activated naive and memory T-cells. To assess if miR-21 inhibition affects the growth of activated T-cells, we performed a competition assay in which the percentage of GFP + cells, harboring miR-21 or control inhibitor (a scrambled hairpin sequence), was monitored over 15 to 18 days in mixed GFP+ and GFP-cell cultures (as schematically depicted in Figure S2B). MiR-21 inhibition in activated naive T-cells led to a subtle, but consistent decrease in the percentage of GFP+ cells only at the beginning of the assay (day 8), and then remained stable until day 18 (Figure 2A, B). In contrast, miR-21 inhibition in activated memory T-cells led to a continuous depletion of GFP+ cells over time, indicating that survival of memory T-cells relies on the presence of miR-21 ( Figure 2C, D). To assess if the observed growth disadvantage results from increased apoptosis, we FACS isolated GFP+ cells harboring miR-21 or control inhibitor at day six following transduction, and assessed apoptosis after an additional 48h culture period. Analysis of the loss of mitochondrial transmembrane potential revealed an increased apoptosis rate in both naive ( Figure 2E) and memory ( Figure 2F) T-cells harboring miR-21 inhibitor. The increase was most pronounced in the memory T-cells, indicating that activation-induced miR-21 expression is an important anti-apoptotic factor for T-cells in general, but is especially necessary for the maintenance of activated memory T-cells.

MiR-21 regulates expression of CC-chemokine receptor 7 (CCR7)
Since miR-21 appeared to function differently in naive and memory T-cells, we next set out to study the functional role of miR-21 in activated naive T-cells. To this end, we examined the effect of miR-21 inhibition on specific aspects of the transition from a naive to a memory phenotype by analyzing the expression of CD45RA, CD45RO and CCR7, defining naive, central and effector memory T-cell stages [21], as well as the expression of CD25 on resting naive, and on activated GFP+ T-cells. Changes in the expression of CD45RA, CD45RO and CD25 resulting from the activation, were similar in T-cells harboring miR-21 and control inhibitor ( Figure S2C). In contrast, the expression of CCR7 was significantly increased on activated naive T-cells harboring miR-21, but not control inhibitor ( Figure 3A). The increase in the expression of CCR7 was evident at the level of CCR7 expression per cell (defined as geometric mean fluorescent intensity of the staining: MFI, Figure 3B), and in the amount of CCR7 positive cells in the culture ( Figure 3C). A similar result was observed at the CCR7 transcript level ( Figure 3D). The expression of CCR7 protein on memory T-cells harboring miR-21 inhibitor was also increased at day six after transduction, however to a significantly lesser extent than in naive T-cells (median fold increase 2.4 and 4.1 respectively) and with more variation between different donors ( Figure 3E-H). The miRNA target prediction programs TargetScan (www.targetscan.org/) [19] and MicroCosm Targets (www.ebi.ac.uk/enright-srv/microcosm/htdocs/targets/ v5/) [20] denote a conserved miR-21 binding site in the 3' UTR of the human CCR7 transcript suggesting a direct interaction between the CCR7 transcript and miR-21 ( Figure S2D). Cotransfection of a dual luciferase reporter construct harboring the 3' UTR of CCR7 together with miR-21 precursor molecules resulted in a significant decrease of the relative luciferase levels. This indicates that CCR7 is indeed a direct target of miR-21 ( Figure 4A). In agreement, freshly isolated naive and memory T-cells showed inversed miR-21 and CCR7 expression patterns ( Figure 4B). In addition, αCD3/CD28 activation of isolated naive T-cells led to a significant downregulation at the CCR7 protein level ( Figure 4C-E), in spite of increased CCR7 transcript levels ( Figure 4F), which also indicates a post-transcriptional regulation of CCR7 protein expression. Together these data show that activation-induced miR-21 expression ( Figure 1B) negatively regulates CCR7 protein levels, especially in activated naive T-cells.

Discussion
In this study we show that activation-induced miR-21 is an integral part of the T-cell activation process where it coregulates fundamentally different aspects of T-cell biology depending on the differentiation status of the T-cell. Specifically, we show that miR-21 regulates the potential migration capacity and transition towards the memory T-cell phenotype of activated naive T-cells, and supports survival of activated memory T-cells. Likely, the different gene expression profiles characterizing naive and memory T-cells [23,24] are fundamental to the observed dual functionality of miR-21. MiRNAs can theoretically regulate a multitude of diverse target genes, however, only a fraction of the predicted target genes is co-expressed together with the miRNA in a given cell type and, as such, can be subjected to miRNA-based regulation. Moreover, the miRNA-target gene interaction is also affected by the presence of other target genes, which are competing for the same miRNA [5,25]. Thus, the overall miRNA-target gene expression ratio, and the abundance of individual target genes are likely of central importance for the factual miRNA-based regulation.
MiR-21 levels are substantially higher in ex vivo isolated memory T-cells than in naive T-cells (this study and [9,10]). Sub division of memory T-cells into central and effector memory phenotypes did not reveal any differences in miR-21 expression, indicating that high miR-21 levels are a general feature of the memory T-cell phenotype. It would be interesting to study miR-21 expression also in terminally differentiated (i.e. CD45RO-CCR7-) T-cells, which are virtually absent in the CD4+ population but present in the CD8+ population.
MiR-21 expression is significantly induced upon T-cell activation in both naive and memory T-cells. However, we found that memory T-cells upregulate miR-21 more rapidly and achieve higher levels of expression, indicating a more prominent role of miR-21 in this population. Expression of the primary-miR-21 transcript can be induced by mitogenic stimuli, such as phorbol-12-myristate-13-acetate (PMA) in several cancer cell lines, and is under positive transcriptional control of were stimulated with 5µg/ml PHA and 100 U/mL IL-2, followed by transduction with lentiviral vectors harboring miR-21 or control inhibitor (scrambled hairpin sequence) and GFP. The percentage of GFP+ cells in culture over time was monitored by FACS. Data were normalized to the first measurement at day six. Each line represents a separate donor (n=5, two-way RM ANOVA with Bonferroni posttests, ns-not significant). B Data from A, depicted as median values with interquartile range. C Resting memory (CD3+ CD8-CD45RO-CD25-) T-cells were stimulated with 5µg/ml PHA and 100 U/mL IL-2, followed by transduction with lentiviral vectors harboring miR-21 or control inhibitor (scrambled hairpin sequence) and GFP. The percentage of GFP+ cells in culture over time was monitored by FACS. Data were normalized to the first measurement at day six. Each line represents a separate donor (n=5, two-way RM ANOVA with Bonferroni posttests). D Data from C, depicted as median with interquartile range. E Activated naive, and F activated memory GFP+ T-cells harboring miR-21 or control inhibitor were isolated by FACS from mixed cultures (A, and C respectively) at day six post lentiviral transduction. Isolated GFP+ cells, and activated not transduced cells (NT) were cultured in complete media supplemented with 100 U/mL of IL-2 for 48h (until day 8), and percentage of apoptotic cells was assessed by FACS-based measurement of mitochondrial trans-membrane potential loss, using DilC1 (5). Each line represents a separate donor (naive: n=4, memory n=5, RM ANOVA with Bonferroni posttests). *p<0.05, **p<0.01, ***p<0.0001. doi: 10.1371/journal.pone.0076217.g002 AP-1, STAT3 and NF-kB [26][27][28][29]. Two of these transcription factors, AP-1 and NF-κB, are directly downstream of the TcR/ CD28 and PKC signaling pathways [30,31], and as such may be responsible for the observed miR-21 induction following Tcell activation. Indeed, we observed an almost complete ablation of activation-induced miR-21 expression upon PKC inhibition (data not shown).
Previous studies in mice and man have shown the involvement of miR-21 in T-cell survival in general [15,18]. Our data, discriminating between naive and memory T-cells, indicate that the requirement for miR-21-based regulation of T-  [16]. Therefore it is possible, that the observed loss of memory T-cells lacking miR-21 is the result of enhanced apoptosis combined with a reduced proliferation rate. However, we did not observe a clear effect of miR-21 inhibition on the proliferative potential of either naive or memory T-cells (data not shown). In vivo studies are warranted to further study the details of the observed differential response of naive and memory T-cells towards the inhibition of miR-21 expression, especially in relation to defined aspects of long-term memory T-cell survival.

Figure 4. CCR7 is a direct target of miR-21 and correlates inversely with miR-21 expression in resting and activated Tcells. A
Ratio of Renilla luciferase (RL) to firefly luciferase (FL) signal, determined in lysates of Cos-7 cells co-transfected with psiCHECK-2 construct harboring the CCR7 3' UTR and a synthetic miR-21 or control precursor. Median values with range of data normalized to control precursor are depicted (n=4 independent experiments, paired t-test). B MiR-21 and CCR7 expression levels were assessed by qRT-PCR in freshly isolated naive (CD4+ CD45RO-) and memory (CD4+ CD45RO+) T-cells. Expression levels are shown relative to the RNU48 and U6 reference genes respectively. Each line represents a separate donor (n= 8, Wilcoxon signed ranked test). C Representative FACS staining plot depicting CCR7 expression on isolated naive (CD4CD45RO-) T-cells before (day 0), and after seven days of activation with plate-bound-anti-CD3/soluble-anti-CD28 mAbs. Isotype staining is depicted. D Quantification of CCR7 staining from C, depicted as MFI (geometric mean fluorescent intensity). Each line represents a separate donor (n=3, paired t-test, ns). E Percentage of CCR7 positive, GFP+ cells from C. Each line represents a separate donor (n=3, paired t-test). F CCR7 transcript expression was determined by qRT-PCR in naive (CD4+ CD45RO-) T-cells before (day 0), and after seven days of activation with plate-bound-anti-CD3/soluble-anti-CD28 mAbs. Each line represents a separate donor (n=5, paired t-test). Expression levels shown in the graph are normalized to the U6 reference gene. *p<0.05, **p<0.01. Stable inhibition of miR-21 in activated naive T-cells led to a significant upregulation of CCR7. To our knowledge this is the first report describing direct, miRNA-driven post-transcriptional regulation of CCR7 expression in human T-cells. It has been previously shown that the expression of CCR7 is downregulated following T-cell activation, and that activated Tcells within secondary lymphoid organs express lower levels of CCR7 protein [32,33]. Furthermore, decreased expression of CCR7, which is paralleled by increased expression of CXCR5, has been shown to guide activated T-cells out of the T-cell areas of the lymph node and direct them towards the B-cell follicles [34]. We found that activation-induced expression of miR-21 regulates CCR7 especially in naive T-cells. Therefore, we propose that fine-tuning of CCR7 expression by miR-21, following the activation of naive T-cells, enhances the egress of activated T-cells from the T-cell rich areas of the lymph nodes, and prevents consecutive recycling through other lymph nodes. The induction of miR-21 upon activation of naive T-cells is thus an inherent part of T-cell programming, which is actively associated with quelling of the naive T-cell state and induction of the memory T-cell phenotype. Regulation of CCR7 in activated (central) memory T-cells was also observed, however, to a much lesser extent than in naive T-cells, and with more variability between the donors. It is likely that abundant expression of other miR-21 target genes prevents regulation of CCR7 in central memory T-cells, and that the miR-21 targetome in different T-cell subsets varies based on the overall expression pool of potential miR-21 target genes. Of note, bioinformatics analysis revealed no miR-21 binding site in the 3' UTR of the murine CCR7 transcript, denoting possible fundamental differences between the regulation of CCR7 in mice and man.
Together, our results provide evidence that miR-21 differentially affects naive and memory T-cells. We propose that intrinsic, phenotype-specific differences between these two T-cell subsets, e.g. depending on the miRNA-mRNA target gene balance, dictate the outcome of miR-21 function.

Ethics statement
The study was approved by The Medical Ethical Committee (METC) of the University Medical Center Groningen (UMCG) (project number: 2009.118) and written informed consent was obtained from all donors.

PBMC isolation
Peripheral blood mononuclear cells (PBMC) from healthy donors were isolated immediately after blood withdrawal into heparin-containing vacutainer tubes (Becton Dickinson, Franklin Lakes, USA). Isolation was done by density gradient centrifugation of diluted blood (1:1 PBS) using Lymphoprep (Axis-shield, Oslo, Norway) according to the manufacturer instructions.

FACS analysis of cell surface markers
Expression of cell-surface markers on T-cell subsets before and after treatment with activating stimuli was assessed by FACS using mAbs against human CD45RO-PE (UCHL1), CD25-APC (BC96), CD45RA-eFluor605 (HI100, all from eBioscience), and CCR7-PE-Cy7 (3D12) or rat-IgG2a-PE-Cy7 isotype control (R35-9, both from BD Biosciences). To assess CCR7 expression, cells were incubated with mAb for 30min at RT. Cells were measured on BD LSR-II Flow Cytometer using Diva software (BD Biosciences) (Figure 3, Figure 2C) and on FACS Calibur flow cytometer using Cell Quest software (BD Biosciences) (Figure 4, Figure 1E, F). Data were analyzed using the Kaluza Flow Analysis Software (Beckman Coulter).  T-cells (1x10 6 cells/ml) were stimulated with 5µg/ml PHA (Thermo, Fisher Scientific Remel products) and 100U/mL recombinant human IL-2 (Peprotech) for 24h prior to transduction. Lentiviral transduction was carried out for 24h in the presence of 4µg/ml polybrene (Sigma-Aldrich, St. Louis, USA). After viral transduction, cells were washed three times with PBS and cultured in complete media supplemented with 100U/mL recombinant human IL-2 (Peprotech, London, England). The expression of green fluorescent protein (GFP) in mixed cultures was measured on a FACS Calibur flow cytometer (BD Biosciences), and monitored for up to 18 days. The percentage of GFP positive cells at the beginning of the assay (day 6) was set to one, and a fold difference per measurement was calculated. Transduced cells were sorted based on the GFP expression using MoFlo sorter (Beckman Coulter), or kept in mixed cultures for competition assay.

Apoptosis measurement
Activated naive and memory GFP+ T-cells harboring miR-21 or control inhibitor were isolated by FACS from mixed cultures at day six post lentiviral transduction, and seeded in 96-wells plates in complete media supplemented with 100 U/mL of IL-2. Control, not transduced cells cultured under the same conditions were seeded alongside. Percentage of apoptotic cells was assessed after 48h (day 8) by FACS-based measurement of mitochondrial transmembrane potential loss. Briefly, cells were stained for 20min at 37°C in cell culture medium containing 50nM DilC1(5) compound (Enzo Life Sciences, NY, USA), followed by washing with PBS. Cells were kept on ice and measured directly at the FACS Calibur flow cytometer using Cell Quest software (BD Biosciences). Data were analyzed using Kaluza Flow Analysis Software (Beckman Coulter).

Cloning of reporter construct, transient transfection and luciferase assays
The 3' UTR sequence harboring the putative miRNA binding site and a part of the open reading frame of the human CCR7 transcript was PCR-amplified from genomic DNA using primers harboring an XhoI (5') or NotI (3') restriction site: 5'-TTGCTCGAGGAGACCACCACCACCTTCTC-3' (forward) and 5'-TGGCGGCCGCCAGTGGAGCCAAGAGCTGAG-3' (reverse), and cloned into psiCHECK2 vector (Promega, Madison, USA), as described previously [35]. The insert was sequence verified (BaseClear, Leiden, The Netherlands). Cos-7 cells (1,2x10 4 ) were transfected with 125ng of the psiCHECK2 construct and 50nM Pre-miR-21 miRNA Precursor Molecule ID: PM10206 or Pre-miR miRNA Precursor Negative Control 1 (Life Technologies) using the Saint-MIX compound (Synvolux Therapeutics B.V., Groningen, The Netherlands) in 250µl serum-free medium. Four hours following the transfection, 500µl of medium supplemented with 10% FBS was added. Cells were lysed 24h after transfection and Renilla and Firefly luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega) according to manufacturer's instructions. For each transfection, luciferase activity was measured in duplicate with the Luminoskan Ascent Microplate Luminometer (Thermo Scientific). The Renilla over Firefly (RL/FF) luciferase ratio for miR-21 precursor was calculated. The RL/FF ratio of control precursor was set to one.

Statistical analysis
Statistical analyses were performed in the GraphPad Prism software (version 5.0). Comparisons of paired samples were performed using the Wilcoxon signed-rank test. Comparisons of paired samples with less than six replicates were performed using the paired t-test or with repeated measures (RM) ANOVA with Bonferroni posttests. Assessment of response to the stimuli over time was performed by the Friedman test with a Dunn's Multiple Comparison test. Comparisons of response to treatment over time between two groups were performed by the Two-way Repeated Measures ANOVA with Bonferroni posttests. All statistical analyses were two-sided, and the significance level used was p < 0.05.