MicroRNA-150 modulates intracellular Ca2+ levels in naïve CD8+ T cells by targeting TMEM20

Regulation of intracellular Ca2+ signaling is a major determinant of CD8+ T cell responsiveness, but the mechanisms underlying this regulation of Ca2+ levels, especially in naïve CD8+ T cells, are not fully defined. Here, we showed that microRNA-150 (miR-150) controls intracellular Ca2+ levels in naïve CD8+ T cells required for activation by suppressing TMEM20, a negative regulator of Ca2+ extrusion. miR-150 deficiency increased TMEM20 expression, which resulted in increased intracellular Ca2+ levels in naïve CD8+ T cells. The subsequent increase in Ca2+ levels induced expression of anergy-inducing genes, such as Cbl-b, Egr2, and p27, through activation of NFAT1, as well as reduced cell proliferation, cytokine production, and the antitumor activity of CD8+ T cells upon antigenic stimulation. The anergy-promoting molecular milieu and function induced by miR-150 deficiency were rescued by reinstatement of miR-150. Additionally, knockdown of TMEM20 in miR-150-deficient naïve CD8+ T cells reduced intracellular Ca2+ levels. Our findings revealed that miR-150 play essential roles in controlling intracellular Ca2+ level and activation in naïve CD8+ T cells, which suggest a mechanism to overcome anergy induction by the regulation of intracellular Ca2+ levels.

and the regulation of intracellular Ca 2+ levels in naïve CD8 + T cells should be tightly regulated to avoid T cell anergy 5 . However, the regulating mechanism of intracellular Ca 2+ in naïve CD8 + T cells is not widely defined yet.
During CD8 + T cell activation, intracellular Ca 2+ concentrations are primarily regulated by the calcium release-activated Ca 2+ (CRAC) channel and plasma membrane Ca 2+ ATPase (PMCA) 6 . TCR stimulation induces Ca 2+ release from the endoplasmic reticulum (ER) through the inositol trisphosphate receptor, and the increased cytosolic Ca 2+ level, in association with calmodulin, activates PMCA to extrude Ca 2+ from the cell. The reduced Ca 2+ concentration in the ER lumen, referred to as store depletion, triggers stromal-interacting molecule 1 (STIM1) to open the CRAC channel via interactions with CRAC channel protein 1 (ORAI1) in the plasma membrane. The opened CRAC channel causes an influx of extracellular Ca 2+ into the cell 7 . Additionally, STIM1 inhibits PMCA-mediated Ca 2+ removal from the cell 8 . The Ca 2+ -regulating activity of STIM1 is mediated by the formation of a complex with transmembrane protein 20 (TMEM20) 9 , followed by the translocation of the complex to the adjacent side of the plasma membrane, eventually enabling STIM1 to bind PMCA. Based on this Ca 2+ -regulating mechanism, TMEM20 appears to play a crucial role in CD8 + T cell activation that must be precisely regulated; however, the specific mechanism by which this occurs and the role in naïve CD8 + T cells are largely unknown. microRNAs (miR) are noncoding RNAs about 22 nucleotides in length that mediate post-transcriptional gene regulation through translational repression and degradation of messenger RNA after binding to the 3′ untranslated region (UTR) of the target RNA 10 . Based on this regulatory activity, miRNAs, including miR-150, are widely associated with the development and biological function of immune cells. During T cell development in the thymus, miR-150 represses neurogenic locus notch-homolog protein 3, a factor involved in T cell differentiation and survival, suggesting that miR-150 is one of the controlling factors for CD8 + T cell development 11 . Recently, miR-150 was also shown to be required for CD8 + T cell activation 12 , and miR-150-deficient CD8 + T cells showed reduced proliferation, differentiation into terminal effector cells, and acquisition of effector function required for killing infected cells. Changes in the transcriptome, such as a decrease in proliferation/killing-associated RNAs and an increase in inhibitory RNAs, were suggested as a molecular mechanism associated with hypo-functional cells; however, the specific mechanism, including the direct target of miR-150 and the molecular events that followed antigenic stimulation in the absence of miR-150 in naïve CD8 + T cells, has not been elucidated.
In this study, we revealed that miR-150 is a critical factor in the prevention of an ectopic increase in intracellular Ca 2+ levels, which prevents anergy induction in naïve CD8 + T cells. Mechanistically, miR-150 suppresses increases in intracellular Ca 2+ levels by directly suppressing TMEM20 expression. This miR-150-induced reduction in TMEM20 allows naïve CD8 + T cells to remove intracellular Ca 2+ and eventually inhibits the expression of anergy-inducing genes, such as Casitas B lineage lymphoma b (Cbl-b), Egr2, and p27. These findings provide a specific molecular mechanism by which miR-150 regulates CD8 + T cell activation and affirms the central importance of intracellular Ca 2+ regulation in naïve CD8 + T cells.
To determine the molecular mechanism associated with reduced activation and functionality in mir-150 −/− CD8 + T cells, we first measured intracellular Ca 2+ levels, because a change in intracellular Ca 2+ level is one of the initial events during CD8 + T cell activation. Naïve mir-150 −/− CD8 + T cells cultured under physiological concentrations of Ca 2+ -containing media exhibited increased intracellular Ca 2+ levels relative to those in naïve mir-150 +/+ CD8 + T cells (Fig. 2a), and the increased intracellular Ca 2+ levels in naïve mir-150 −/− CD8 + T cells were sustained (Fig. S2) 14 . To determine the degree to which intracellular Ca 2+ levels increased in naïve mir-150 −/− CD8 + T cells, we compared the levels in mir-150 +/+ and mir-150 −/− CD8 + T cells before and after TCR stimulation. The basal levels of intracellular Ca 2+ in naïve mir-150 −/− CD8 + T cells were already similar to the increased levels achieved in mir-150 +/+ CD8 + T cells following TCR stimulation and were not further increased by TCR stimulation (Fig. 2b). These data suggested that miR-150 is required for regulation of intracellular Ca 2+ levels in naïve CD8 + T cells.
Elevation of intracellular Ca 2+ levels in naïve mir-150 −/− CD8 + T cells is associated with downregulated PMCA activity. Intracellular Ca 2+ concentrations are regulated primarily by CRAC and PMCA 6 .

Discussion
In CD8 + T cells, increasing the concentration of intracellular Ca 2+ after TCR stimulation is essential to inducing the expression of activation-associated genes 2, 25, 26 . Because controlling the Ca 2+ concentration is crucial for T cell activation, there has been an extensive search for molecules, such as channels, transporters, and other proteins,
Scientific RepoRts | 7: 2623 | DOI:10.1038/s41598-017-02697-x associated with Ca 2+ movement during T cell activation. The preconditions for increasing intracellular Ca 2+ levels during T cell activation are a relatively low level of intracellular Ca 2+ in naïve CD8 + T cells and an ectopic increase in intracellular Ca 2+ levels before T cell activation turns T cells into tolerant cells 5 . Although controlling intracellular Ca 2+ levels in naïve CD8 + T cells is also critical for T cell activation, regulatory molecules and associated mechanisms that determine the Ca 2+ level in naïve CD8 + T cells are largely unknown. In this study, we showed that miR-150 controls intracellular Ca 2+ levels by downregulating TMEM20 expression in naïve CD8 + T cells. The miR-150-induced suppression of TMEM20 expression prevented the expression of anergy-inducing genes and hypo-responsiveness of CD8 + T cells upon antigenic stimulation. Therefore, miR-150 is an essential regulator that creates an activation-favorable molecular milieu in naïve CD8 + T cells.
In this study, we showed that intracellular Ca 2+ levels are increased by miR-150 deficiency in naïve CD8 + T cells. In mir-150 −/− naïve CD8 + T cells, the expression level of TMEM20 was increased, while expression levels of other Ca 2+ regulating molecules did not changed. In addition, suppression of TMEM20 decreased intracellular Ca 2+ levels in mir-150 −/− naïve CD8 + T cells, indicating that the increased intracellular Ca 2+ levels in mir-150 −/− naïve CD8 + T cells are derived from the increased expression of TMEM20. Previous report showed that TMEM20 attenuates PMCA activity 9 . mir-150 −/− naïve CD8 + T cells showed high translocation of TMEM20 into PMCA in plasma membrane. In addition, Ca 2+ reducing rate also decreased in mir-150 −/− naïve CD8 + T cells. Treatment of PMCA inhibitor did not changed the intracellular Ca 2+ level in mir-150 −/− naïve CD8 + T cells whereas treatment of inhibitors for other Ca 2+ -regulating molecules changed the intracellular Ca 2+ levels in mir-150 −/− naïve CD8 + T cells, which also can be an indirect clue that PMCA activity is inactivated in mir-150 −/− naïve CD8 + T cells. Collectively, the high Ca 2+ levels in mir-150 −/− naïve CD8 + T cells are derived from increased expression of TMEM20, thereby inactivation of Ca 2+ extruding PMCA activity.
Several regulators such as PMCA, CRAC, SERCA, NCX, and MCU control intracellular Ca 2+ levels in TCR-stimulated CD8 + T cells 22,27 . In this study, we showed that PMCA, but not other regulators, mainly controls intracellular Ca 2+ levels in naïve CD8 + T cells. Although CRAC inhibition slowed the Ca 2+ uptake rate in naïve CD8 + T cells, overall intracellular Ca 2+ levels were not changed, indicating that the role of CRAC is limited to regulating intracellular Ca 2+ level in naïve CD8 + T cells. The unchanged intracellular Ca 2+ levels in naïve CD8 + T cells in SERCA-and NCX-inactivated conditions could be a result of Ca 2+ -regulating activity that is TCR-activation dependent. For MCU, their Ca 2+ -regulating function is associated with relocation of mitochondria. In TCR-stimulated T cells, mitochondria relocate to the immunological synapse, where the mitochondria act as Ca 2+ -controlling machinery through MCU 21 . Un-polarized mitochondria at the sites of Ca 2+ influx in naïve CD8 + T cells might be a reason that MCU is not associated with intracellular Ca 2+ levels in naïve CD8 + T cells.
In a previous study, TMEM20-mediated suppression of PMCA activity was shown to be dependent on calcium store depletion in the ER and co-localization with STIM1 in a Jurkat cell line model 9 . However, our co-localization data showed that TMEM20 does not co-localize with STIM1 either in naïve mir-150 +/+ or mir-150 −/− CD8 + T cells. In addition, the expression levels of STIM1 were similar between naïve mir-150 +/+ and mir-150 −/− CD8 + T cells. Considering the increased intracellular Ca 2+ levels in naive mir-150 −/− CD8 + T cells, STIM1 does not act as a regulator of intracellular Ca 2+ in naïve CD8 + T cells. Thus, PMCA may act in a STIM-independent manner in naïve CD8 + T cells. Given that mir-150 has many targets, expression of undefined factor(s) supporting TMEM20 function could be increased in naïve mir-150 −/− CD8 + T cells, and, thereby, the increase in TMEM20 with the associated factor(s) might suppress PMCA activity in naïve mir-150 −/− CD8 + T cells before calcium store-depletion in the ER. Another possible mechanism for calcium store depletion independent of TMEM20-mediated PMCA deactivation is TMEM20 expression in the plasma membrane, which might deactivate PMCA function itself, but its function could also be limited by insufficient expression to act in naïve mir-150 +/+ CD8 + T cells. In this case, increased TMEM20 expression in naïve mir-150 −/− CD8 + T cells may be sufficient to deactivate PMCA function, resulting in the increase in intracellular Ca 2+ levels in naïve mir-150 +/+ CD8 + T cells.
The expression levels of TMEM20 mRNA were similar between naïve mir-150 +/+ and mir-150 −/− CD8 + T cells. However, retroviral-mediated overexpression of miR-150 downregulated the expression of TMEM20 mRNA. Considering that miR-150 levels in retro-miR-150-infected mir-150 +/+ CD8 + T cells were more than 2-fold higher than those in the uninfected control mir-150 +/+ CD8 + T cells, retro-miR-150 infected mir-150 −/− CD8 + T cells may have higher levels of miR-150 than those in mir-150 +/+ CD8 + T cells. Thus, it is possible that a normal amount of miR-150, such as that in naïve mir-150 +/+ CD8 + T cells, might repress translation of TMEM20, whereas overexpression of miR-150, such as that in retro-miR-150 infected naïve mir-150 −/− CD8 + T cells, might suppress transcription of TMEM20. Given that miR-150 has many targets, miR-150 overexpression might induce suppression of TMEM20 as well as have additional or alternative effects, and the additional effects, not suppression of TMEM20 levels, could be the reason for the decreased intracellular Ca 2+ levels in retro-miR-150-infected naïve mir-150 −/− CD8 + T cells. However, we showed that suppression of TMEM20 expression decreased the intracellular Ca 2+ levels in naïve mir-150 −/− CD8 + T cells. Thus, TMEM20 expression levels, especially in the absence of miR-150, are critical to controlling intracellular Ca 2+ levels in naïve CD8 + T cells.
Whether simply overexpressing TMEM20 is sufficient to increase intracellular Ca 2+ levels in naïve mir-150 +/+ CD8 + T cells in questionable. In a previous report, TMEM20 overexpression did not change the intracellular Ca 2+ levels in HEK 293 cells expressing STIM1 and Orai1 in thapsigargin-treated conditions, but thapsigargin-untreated conditions were not tested 9 . Although we showed that miR-150 deficiency did not change the expression levels of Ca 2+ -regulating molecules, except for TMEM20, miR-150 deficiency may influence undefined factor(s) required for TMEM20 function, as described above. In this case, simply overexpressing TMEM20 might not increase intracellular Ca 2+ level in naïve mir-150 +/+ CD8 + T cell. To date, the mechanism of TMEM20 function and its associated/regulating molecules have not been widely investigated. Thus, studies on the mechanism of TMEM20 function and regulation should be conducted to better understand the detailed mechanism of Ca 2+ regulation in naïve CD8 + T cells before and after activation. miR-150 deficiency reduces CD8 + T cell activation, in terms of expansion, differentiation, and cytolytic function 12 . Although changes in the levels of several mRNAs in response to miR-150 deficiency were suggested in a previous study to be a mechanism for the observed reduction in CD8 + T cell activation, a direct relationship between altered mRNA levels and CD8 + T cell activation-associated molecular changes was not found. In this study, we showed that TMEM20 is a direct target for miR-150 in naïve CD8 + T cells and that the miR-150 deficiency-induced reduction in CD8 + T cell activation results from altered intracellular Ca 2+ levels. In the absence of miR-150, elevated expression of TMEM20, a Ca 2+ extruder, attenuates PMCA activity, resulting in increased intracellular Ca 2+ levels in naïve CD8 + T cells 8,9 . This increased intracellular Ca 2+ induced activation of NFAT1, which consequently induced expression of anergy-inducing genes such as Egr2, Cbl-b, and p27 16,17,28 . Because anergy-inducing genes such as p27 and Egr2 are reportedly miR-150 targets, increased expression of p27 and Egr2 in mir-150 −/− CD8 + T cells may be a result of the loss of translational suppression of these mRNAs by miR-150 in mir-150 −/− naïve CD8 + T cells 29,30 . However, following treatment with an NFAT1 inhibitor, mir-150 −/− naïve CD8 + T cells showed decreased levels of p27 and Egr2 mRNA, indicating that increased expression of anergy-inducing genes is primarily due to transcriptional regulation via the intracellular Ca 2+ /NFAT1 signaling pathway. Expression of these anergy-inducing genes related to miR-150 deficiency could explain the reduced proliferation, differentiation, and killing activity of naïve mir-150 −/− CD8 + T cells upon antigenic stimulation, as previously reported 12 and shown in our study. To avoid these miR-150-deficiency-induced molecular events, intracellular Ca 2+ levels need to be sufficiently low to prevent NFAT1 activation.
It was previously reported that in effector CD8 + T cells, miR-150 levels are lower than in naïve CD8 + T cells 13 . Given that miR-150 suppresses expression of CD25, an IL-2 receptor, reduced miR-150 levels in effector CD8 + T cells can favor survival and proliferation 31 . Relatively low levels of miR-150 in effector CD8 + T cells may also be explained by intracellular Ca 2+ -associated events. Because miR-150 deficiency elevated intracellular Ca 2+ levels, decreased levels of miR-150 in effector CD8 + T cells implies that effector CD8 + T cells may require higher intracellular Ca 2+ concentrations to promote activity such as that of cytotoxic T lymphocytes (CTLs). Previous reports showed that intracellular Ca 2+ is involved in granule reorientation toward the target cell contact region and exocytosis of granules [32][33][34] . Additionally, intracellular Ca 2+ mediates the production of effector cytokines and the expression of death receptor ligands [35][36][37] . To mediate these diverse functions associated with CTLs, high intracellular Ca 2+ concentrations may be required, which might be satisfied by miR-150 downregulation in effector CD8 + T cells.
miR-150 acts as a "fine tuner" of various immune cells. In B cells, miR-150 controls differentiation and receptor signaling [38][39][40] . In natural killer cells, miR-150 regulates development, as well as cytotoxicity 24,41 . T cell differentiation and the expression of cytokine receptors are also influenced by miR-150 11,31 . Here, we showed that miR-150 controlled Ca 2+ signaling and prevented the induction of anergy in naïve CD8 + T cells. Considering these various roles for miR-150 in immune cells and various target molecules, studies to determine additional miR-150 roles in the immune system will likely shed light on currently undefined mechanisms as well as possible therapeutic approaches using CD8 + T cells.
Given the impact of miR-150 on CD8 + T cell activation, regulation of miR-150 expression may be a therapeutic target in CD8 + T cell-associated diseases. Because we showed that mir-150 −/− naïve CD8 + T cells could not be activated, CD8 + T cell-specific suppression of miR-150 expression may be a novel approach to treating autoimmune diseases. In this context, our findings indicate a molecular mechanism that prevents the transition of CD8 + T cells into a hypo-responsive state, as well as a basis for regulation of CD8 + T cell activation.

Methods
Mice. TCR gag transgenic mice have been generated as previously described 42 . C57BL/6 (B6) mice and mir-150 −/− mice on a B6 background were purchased from The Jackson Laboratory and TCR gag :mir-150 +/+ and TCR gag :mir-150 −/− mice were generated in our animal facility. All mice were bred and maintained under Specific Pathogen Free conditions. All of the methods and experimental procedures were conducted according to the approved (approval ID: KRIBB-AEC-15088) guidelines and regulations by the animal ethics committee (IACUC) of KRIBB, South Korea.
Cell lines. The Friend virus-induced erythroleukemia of B6 origin, FBL, expresses the FMuLV-encoded gag epitope (peptide CCLCLTVFL purchased from Pi Proteomics), and was maintained in culture in RPMI1460 supplemented with 10% FBS and antibiotics. NIH3T3 cells for plasmid transfection were cultured in DMEM supplemented with 10% FBS and antibiotics.
Immunofluorescence. Glass coverslips-attached cells were stimulated with CaCl 2 , PMA/Ionomycin, and anti-CD3 for indicated time points, washed using PBS, fixed in 3.7% formaldehyde for 10 min at 37 °C, permeabilized with 0.2% Triton X-100 for 10 min at room temperature, washed, and then blocked in 1% BSA in PBS Scientific RepoRts | 7: 2623 | DOI:10.1038/s41598-017-02697-x for 30 min. To visualize NFAT1, the cells were stained with anti-NFAT1 (Abcam) for 1 h, washed and incubated with FITC-conjugated anti-mouse antibody (Santa Cruz Biotechnology) for 1 h. Images were captured with an Olympus DP30BW digital camera and processed using the Metamorph 7.1 program (Universal Imaging, Media, PA, USA). Nuclear translocation of NFAT1 was measured by percentage of NFAT1 localized in the nucleus.
Flow cytometry. All antibodies for flow cytometric analysis were purchased from Becton Dickinson or BD Pharmingen. CD8 + T cells were stained with the indicated antibodies in a staining buffer (PBS containing 1% FBS and 0.01% NaN3) for 20 minutes at 4 °C. After washing, flow cytometry was performed on a BD FACS Canto II and data analyzed with FlowJo software (Tree Star).
Reverse transcription and quantitative PCR. Total RNA extracted with TRIzol (Invitrogen) was reverse transcribed with Moloney murine leukemia virus reverse transcriptase and oligo-d(T). For real-time qPCR, cDNA was amplified with specific primers and SYBR Premix Ex Taq (Takara Bio) on a Dice TP800 Thermal Cycler (Takara Bio). The mRNA levels are presented relative to expression of GAPDH mRNA. Quantitative real-time PCR of mature miR was performed with a TaqMan MicroRNA Assay kit (Applied Biosystems) as previously described 23,24 . miR expressions are presented relative to the level of U6 small nuclear RNA. mRNA array and data analysis. To collect naïve, effector, and memory CD8 + T cells, mice infected with Lm-gag as described in "T cell proliferation in vivo", samples were isolated from the spleen at day 0, 5, and 50 after Lm-gag infection, respectively, and analyzed the phenotype of each samples (naïve: CD8 + CD44 lo CD62L hi , effector: CD8 + CD44 hi CD62L lo , and memory: CD8 + CD44 hi CD62L hi ). Labeling of target mRNAs and hybridization were performed using Agilent's mRNA Labeling Reagent and Hybridization Kit (Agilent Technology) as manufacturer's instruction. The hybridization images were analyzed with an Agilent DNA microarray Scanner and quantified by using Agilent Feature Extraction software. Normalization and selection of all data were performed by using GeneSpring GX 7.3 (Agilent Technology). For microarray data, genes were filtered with removing flag-out genes in each experiment. In the gene expression microarray, intensity-dependent normalization (LOWESS) was performed, where the ratio was reduced to the residual of the Lowess fit of the intensity vs. ratio curve. The averages of normalized ratios were calculated by dividing the average of normalized signal channel intensity by the average of normalized control channel intensity. Functional annotation of genes was performed according to Gene Ontology TM Consortium (http://www.geneontology.org/index.shtml) by GeneSpring GX 7.3. Gene classification was based on searches done by GeneCards (http://www.genecards.org/), miRanda (http:// www.microrna.org/), DAVID (http://david.abcc.ncifcrf.gov/), and Medline databases (http://www.ncbi.nlm.nih. gov/). The microarray data can be accessed from the Gene Expression Omnibus under the accession GSE62262.