NK cells were purified from peripheral blood mononuclear cells of healthy donors using the Human NK Cell Isolation kit (Miltenyi Biotec) (12). A total of 13 unrelated healthy donors were used for all the experiments described. Donors provided an informed consent according to the procedures approved by the Ethics Committee of Ospedale Policlinico San Martino (39/2012).

To obtain miRNA profiling, NK cells were incubated for 24 h in the presence of complete medium (RPMI 1640 with 10% of fetal bovine serum, 2 mM glutamine, 50 mg/mL penicillin and 50 mg/mL streptomycin) supplemented with TGF-β1 at the final concentration of 40 or 5 ng/mL. Control cells were represented by NK cells cultured in the presence of complete medium alone (12). For miRNA validation and mRNA expression analysis, NK cells were incubated in the presence of 40 or 5 ng/mL of TGF-β1 or complete medium for 12 or 24 h.

RNA containing the small RNA fraction was extracted using the miRCURY RNA Isolation Kit—Cell and Plant (Exiqon) according to the manufacturer guidelines. RNA amount was determined using the Quant-iT RiboGreen RNA Assay Kit (Invitrogen) following to the manufacturer instructions.

miRNA profiling was performed according to the TaqMan MicroRNA Array Workflow (Applied Biosystems). Briefly, 100 ng of RNA comprehensive of the small RNA fraction were reverse transcribed using the TaqMan MicroRNA Reverse Transcription kit with the Megaplex RT Primers Human Pool A + B set v3.0 (Applied Biosystems). A preamplification step was performed using the TaqMan PreAmp Master Mix with the Megaplex PreAmp Primers Human Pool A + B set v.3.0 (Applied Biosystems). Preamplified cDNAs were loaded in TaqMan Array Human MicroRNA Cards A and B set v3.0 (Applied Biosystems). Amplification reactions were performed using an Applied Biosystems ViiA 7 Real-Time PCR System. Data were analyzed using the miScript miRNA PCR Array Data Analysis Tool.¹ miRNAs expression was normalized to the U6 snRNA expression. Data from this miRNA profiling have been submitted to the NCBI Gene Expression Omnibus² under accession no. GSE98769. miRNAs that showed a fold change >2 or <0.5 between TGF-β1-treated and -untreated NK cells in three different donors were considered to be differentially expressed. Using the computational prediction on-line tools TargetScan³ (23), miRanda⁴ (24), and miRmap⁵ (25), selected miRNAs were investigated as possible regulators of CXCR4, CXCR3, and CX₃CR1 expression.

Expression of selected miRNAs was evaluated using TaqMan MicroRNA Assays as described by the manufacturer (Applied Biosystems). Briefly, 10 ng of RNA were reverse transcribed using the TaqMan MicroRNA Reverse Transcription kit primed with the specific RT primers. Real-time PCR was performed in quadruplicate using the specific primers. Expression of each miRNA was normalized to the RNU44 expression.

To evaluate the expression of CX₃CR1 mRNAs, 100 ng of RNA were reverse transcribed using the SuperScript VILO cDNA Synthesis Kit (Invitrogen). The cDNA was used for real-time PCR using the specific primers contained in the TaqMan Gene Expression Assay (Applied Biosystems). CX₃CR1 expression was normalized to the GAPDH expression. Experiments were performed in quadruplicate.

A 700 bp fragment of the CX₃CR1 3′ untraslated region (3′UTR) containing the putative target site for miR-27a-5p was amplified by PCR using primers CX3UTR2F and CX3UTR2R, containing the NheI and XhoI sites in their 5′ ends, respectively. The PCR product was purified using the MinElute PCR Purification Kit (Qiagen), NheI-XhoI digested and ligated to the NheI-XhoI digested pmirGLO vector (Promega) using the Rapid DNA Ligation kit (Roche, Basel, Switzerland). Ligation product was used for transformation of TOP10 Escherichia coli competent cells (Invitrogen). Colonies containing the recombinant plasmid were selected by PCR using primers CX3UTR2F/CX3UTR2R. Plasmid DNA was purified using the Illustra PlasmidPrep Mini Spin Kit (GE Healthcare). The inserted fragment and flanking sequences were sequenced using primers CX3UTR2F, CX3UTR2R, CX3UTR22R2, and CX3UTR22F2. Sequencing was performed using the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems) in a 3100 Genetic Analyzer (Applied Biosystems). The selected plasmid is reported as pmirCX3UTRWT. Primers, when not otherwise indicated, were designed by primerBLAST.⁶ Primer sequences are reported in Figure S1 in Supplementary Material.

A mutated version of pmirCX3UTRWT (pmirCX3UTRMT), containing a C>G mutation in the putative target site for miR-27a-5p, was prepared by site-directed mutagenesis using the Geneart Site-Directed Mutagenesis System (Invitrogen) according to the manufacturer guidelines. Oligonucleotides (CX3MUTF and CX3MUTR) were designed using the QuickChange Primer Design on-line tool (Agilent Technologies⁷). Presence of the mutation was confirmed by restriction analysis and by sequencing the mutagenized region using primers CX3UTR2FS and CX3UTR22R2.

Wild-type (pmirCX3UTRWT) and mutant (pmirCX3UTRMT) plasmids containing the 3′UTR of the CX₃CR1 mRNA, as well as the parental pmirGLO vector (Promega), were used in cotransfection experiments with the mirVana miRNA Mimic miR-27a-5p and the mirVana miRNA Mimic Negative Control #1 (Ambion). HEK293T cells were used for cotransfection experiments. The day before transfection 8 × 10⁴ cells per well were plated in 24-well plates in 500 µL of DMEM supplemented with 10% FCS. Cells were transfected with 100 ng of plasmid and 20 pmol of mimic miRNA using lipofectamine 2000 (Invitrogen) according to the manufacturer protocol. Twenty-four hours post transfection cells were harvested. Firefly and Renilla Luciferase activities were determined on cell lysates using the Dual Luciferase Reporter Assay System (Promega) according to the manufacturer guidelines. A single-tube DLReady validated luminometer TD-20/20 Turner Biosystems was used. Firefly luciferase activity was normalized to renilla luciferase activity. Normalized values were expressed as changes relative to the value of the negative control, which was set as 1. Three independent experiments were performed in quadruplicate.

The 1 × 10⁶ NK cells from healthy donors were cultured in the presence or in the absence of 5 ng/mL of TGF-β1. Real-time PCR was performed using the hsa-mir-23a TaqMan Pri-miRNA Assay (Applied Biosystems). Hsa-mir-23a pri-miRNA expression was normalized to the GAPDH expression. Experiments were performed in triplicate.

The 1 × 10⁶ NK cells from healthy donors were cultured in 24-well plates in 500 µL RPMI + 10% serum in the presence of TGF-β1 (5 ng/mL). Cells were transfected with 20 pmol of miR-27a-5p inhibitor (Applied Biosystems) or with a negative control (mirVana™ miRNA Inhibitor, Negative Control #1; Ambion) using Lipofectamine 3000 according to manufacturer instructions. Efficiency of transfection, determined with a fluorescently labeled miRNA (Cy3™ Dye-Labeled Pre-miR Negative Control #1; Ambion) was about 30% (data not shown). Cells were harvested after 72 h, RNA was extracted, and CX₃CR1 mRNA levels were determined as described previously.

Primer CX3ATGKZF, designed to contain a Kozak consensus sequence, and primer CX3UTR2R (Figure S1 in Supplementary Material) were used for the amplification of a genomic fragment (2,494 bp in length) containing the whole CX3CR1 gene open reading frame (ORF) and a portion of the 3′UTR in which the two putative target sites for miR-27a-5p are included. The product was cloned in a pcDNA 3.1/V5-His-TOPO vector (Invitrogen), excised by XbaI-BamHI digestion, and ligated to an XbaI-BamHI digested pCDH-CMV-MCS-EF1-GFP lentiviral vector (System Biosciences).

Ligation product was transformed in Stbl3 E. coli cells (Invitrogen), a positive colony was selected by PCR, endotoxin-free plasmid DNA was extracted using the QIAfilter Plasmid Midi Kit with the EndoFree Plasmid Buffer Set (Qiagen), the whole insert and the flanking vector regions were sequenced using 16 primers (whose sequences are reported in Figure S1 in Supplementary Material).

The lentiviral vector containing the CX₃CR1 gene was assembled in lentiviral particles using the pPACKH1 HIV Lentivector Packaging Kit (System Biosciences). Briefly, 3 × 10⁶ 293T cells plated in a 10 cm plate were tranfected with the packaging plasmids and the CX₃CR1 lentiviral vector with lipofectamine 2000 (Invitrogen). Medium was changed every 24 h and collected at 48 and 72 h. Virus containing supernatants were pooled and concentrated by the use of a PEG precipitation solution. Titration was performed by flow cytometric detection of GFP fluorescence emitted by infected cells.

HEK293T cells (2 × 10⁵) were plated in six-well plates and infected with viral preparation at a multiplicity of infection of 2 in 500 µL complete medium without antibiotics in the presence of 8 µg/mL of polybrene (Santa Cruz Biotechnology, Dallas, TX, USA). Infected cells were centrifuged for 1 h at room temperature at 2,000 rpm, then incubated at 37°C. 24 h after infection, medium was removed and replaced with 4 mL of fresh medium without polybrene. One week after infection, CX₃CR1 positive cells were selected using the anti-CX₃CR1 Microbead Kit (Miltenyi Biotec). Clones generated by limited dilution of the selected cells were analyzed for CX₃CR1 surface expression by flow cytometry. Clones showing a good and stable expression of CX₃CR1 were expanded and used for transfection with miR-27a-5p and with miR-Neg.

An HEK293T clone stably expressing CX₃CR1 was transfected with miR-27a-5p and with miR-Neg to evaluate the modulation of CX₃CR1 expression. The 10⁵ cells were plated in six-well plates in 2 mL complete medium, transfection was performed using 100 pmol of miRNA mimics and 3 μL of Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer guidelines. Cells were harvested 48 and 72 h after transfection. CX₃CR1 expression was analyzed by flow cytometry.

For coculture experiments in transwell condition, 2 × 10⁵ resting NK cells were cultured for 48 h with 3 × 10⁵ SH-SY5Y cells. NK cells and SH-SY5Y cells were placed in 24-well transwell (0.3-µm pore size; Corning Costar), upper and bottom chamber, respectively (12).

The statistical level of significance (p) is indicated using Student’s t-test, a parametric significance test. Graphic representation and statistical analysis were performed using GraphPad Prism 6 (GraphPad Software, La Jolla, CA, USA). Pearson’s correlation coefficient (two-tailed test) was used to evaluate the correlation between CX₃CR1 mRNA and miR-27a-5p expression.

To identify miRNAs possibly involved in the TGF-β1-mediated modulation of chemokine receptors expression, we investigated the miRNA expression profile variation in TGF-β1-treated NK cells purified from the peripheral blood of three unrelated healthy donors. The analysis, allowing the simultaneous profiling of 754 miRNAs (Figure 1), showed that miRNAs up- or downregulated by the treatment ranged from 127 in donor 2 to 234 in donor 3 (Figure S2 in Supplementary Material). A total of 14 miRNAs were modulated in NK cells from all donors analyzed upon treatment with either 5 or 40 ng/mL of TGF-β1. Among these 14 miRNAs, 11 were excluded from further analysis being expressed at low levels (threshold cycle <33), while 1 (miR-1201) was forsaken due to its sequence overlapping to an annotated small nucleolar RNAs (snoRNA, SNORD126), as reported in miRBase.⁸ The remaining two miRNAs (miR-302a and miR-27a-5p) were further investigated as putative regulators of CXCR4, CXCR3, or CX₃CR1 expression (Figure 1 and Figure S2 in Supplementary Material).

The differential expression of miR-302a and miR-27a-5p in untreated and TGF-β1-treated NK cells was checked for validation by using specific miRNA assays. The analysis revealed that miR-302a was virtually absent in both untreated and TGF-β1-treated NK cells (data not shown). On the contrary, in agreement with data obtained by miRNA profiling, TGF-β1-treated NK cells showed a significant increase of miR-27a-5p expression, which was higher at low TGF-β1 concentration (Figure 2). Thus, miR-27a-5p was further investigated as putative regulator of CXCR4, CXCR3, or CX₃CR1 expression using the computational prediction on-line tools TargetScan (see text footnote 3) (23), miRanda (see text footnote 4) (24), and miRmap (see text footnote 5) (25). The analysis indicated miR-27a-5p as a putative regulator of CX₃CR1 expression with relatively high scores. The three softwares predicted the same principal site of interaction in the CX₃CR1 3′UTR with the miR-27a-5p seed region (Figure S3A in Supplementary Material). An additional site in the CX₃CR1 3′UTR, with a weaker interaction with miR-27a-5p, was predicted by miRanda and miRmap tools only (Figure S3B in Supplementary Material).

Since miRNAs properties include the capability of inducing degradation of targeted mRNAs, expression profiles of miR-27a-5p and CX₃CR1 mRNAs were simultaneously analyzed over time. As shown in Figure 3A, at both concentrations used, TGF-β1 caused a significant increment of miR-27a-5p and decrease of CX₃CR1 mRNA expression. Importantly, expression of miR-27a-5p and CX₃CR1 mRNA was inversely correlated (Pearson’s correlation coefficient r = 0.766, p < 0.05; Figure 3B). While preliminary data showed that 1 ng/mL of TGF-β1 may be sufficient for induction of miR-27a-5p, concentrations of 5 and 40 ng/mL were chosen for a more systematic investigation. Next, to deepen the molecular mechanism responsible for the TGF-β1-induced increase of miR-27a-5p, we analyzed the expression of the miR-23a-27a-24-2 cluster, precursor of miRNAs, including miR-27a-5p. As shown in Figure 3C, TGF-β1 caused a significant increment of the primary miR-23a-27a-24-2 transcript, thus demonstrating that the increased amount of miR-27a-5p might be due, at least in part, to induction of its gene expression other than, for example, to miR-27a-5p egress from intracellular stores.

To unequivocally demonstrate that miR-27a-5p could interact with the 3′UTR of the CX₃CR1 mRNA, modulating its expression, we performed a luciferase reporter assay (Figure 4). We cloned a 700 bp fragment of the CX₃CR1 gene 3′UTR, containing all the putative sites of interaction with the seed region of miR-27a-5p, into a luciferase reporter vector, downstream of the firefly luciferase gene. Moreover, we also produced a mutated version of the construct, containing a C>G substitution in the main site of the CX₃CR1 3′UTR (Figure S3A in Supplementary Material). The two plasmids were separately cotransfected in HEK293T cells with a miR-27a-5p mimic or with a random sequence miRNA negative control. As shown in Figure 4A, a significant reduction of the normalized luciferase activity was detected in the presence of miR-27a-5p mimic compared to the miRNA negative control. Importantly, this effect was lost when the miR-27a-5p mimic was cotransfected with the construct having the mutated sequence of the CX₃CR1 3′UTR, thus confirming the interaction of miR-27a-5p with the CX₃CR1 3′UTR.

Next, to further confirm miR-27a-5p and CX₃CR1 mRNA interaction, before TGF-β1 treatment NK cells were transiently transfected with a specific miR-27a-5p inhibitor or with a scrambled miRNA as negative control (Figure 4B). As expected, we observed a significantly higher CX₃CR1 mRNA expression in TGF-β1-treated NK cells previously transfected with the miR-27a-5p inhibitor when compared with cells transfected with a scrambled miRNA.

To determine whether miR-27a-5p is able to downregulate the surface expression of CX₃CR1, we prepared a lentiviral vector containing the whole ORF of the CX₃CR1 gene and the portion of the 3′UTR region containing the two putative binding sites for miR-27a-5p. HEK293T cells where transduced with the vector and cultured under limiting dilution to obtain CX₃CR1 positive clones. A clone (#124) was selected that stably expressed at the cell surface optimal levels of the chemokine receptor. Clone #124 was transfected with a miR-27a-5p mimic or with a miRNA negative control and analyzed by flow cytometry for the chemokine receptor expression. As shown in Figure 5, miR-27a-5p mimic induced a significant downregulation of the percentage of CX₃CR1 positive cells. On the contrary, the expression of other molecules such as PVR (CD155) was unaffected (Figure S4 in Supplementary Material). These data further support the role of miR-27a-5p as modulator of the chemokine receptor.

To analyze the relevance of miR-27a-5p induction in a pathological context, we cocultured under transwell condition resting NK cells and the prototypic SH-SY5Y NB cell line, which we described to induce a TGF-β1-mediated downregulation of CX₃CR1 surface expression (12). Accordingly with our previous published data (12), NB conditioning resulted in a significant downregulation of CX₃CR1 expression at both protein and mRNA level. Importantly, this modulatory effect matched with a significant increase of miR-27a-5p as compared to unconditioned NK cells (Figure 6).