Critical role for miR-181a/b-1 in agonist selection of invariant natural killer T cells

Abstract

T-cell receptor (TCR) signal strength determines selection and lineage fate at the CD4⁺CD8⁺ double-positive stage of intrathymic T-cell development. Members of the miR-181 family constitute the most abundantly expressed microRNA at this stage of T-cell development. Here we show that deletion of miR-181a/b-1 reduced the responsiveness of double-positive thymocytes to TCR signals and virtually abrogated early invariant natural killer T (iNKT) cell development, resulting in a dramatic reduction in iNKT cell numbers in thymus as well as in the periphery. Increased concentrations of agonist ligand rescued iNKT cell development in miR-181a/b-1^−/− mice. Our results define a critical role of miR-181a/b-1 in early iNKT cell development and show that miR-181a/b-1 sets a TCR signaling threshold for agonist selection.

Natural killer T (NKT) lymphocytes share features characteristic for NK cells as well as T cells, including the T-cell receptor (TCR). Upon TCR triggering they are able to rapidly release cytokines, such as IL-4 and IFN-γ, without prior priming. Thus, NKT cells are able to shape T helper cell differentiation and may, consequently, promote or suppress immune responses (1). NKT cells constitute various populations, the most extensively characterized of which comprises the invariant (i)NKT cells. These cells share a semiinvariant TCR that recognizes lipid antigen bound to the nonclassic MHC I molecule CD1d (2). It is composed of a Vα14Jα18 TCRα chain in mouse (Vα24Jα18 in human) and a limited pool of TCRβ chains, with a bias toward Vβ8, Vβ7, and Vβ2 (3). During intrathymic T-cell development the iNKT cell lineage diverges from conventional αβT cells at the CD4⁺CD8⁺ double-positive (DP) thymocyte stage and can be identified by its reactivity to CD1d-tetramers loaded with lipid antigen, such as α-galactosyl-ceramide (αGalCer) (4). Differentiation of iNKT cells proceeds through four phenotypically distinct precursor stages: CD24⁺DP^dim (stage 0), CD44^–NK1.1^– (stage 1), CD44⁺NK1.1^– (stage 2), and CD44⁺NK1.1⁺ (stage 3) (5–7). Stage 3 likely comprises a mixture of freshly generated as well as recirculating iNKT cells.

iNKT cells, as well as other nonconventional T cells, have been shown to be autoreactive to a certain degree (2). Consequently, iNKT cells have been proposed to be selected through strong TCR signals in a process termed “agonist selection.” They undergo massive intrathymic proliferation, and mature cells are CD44⁺, indicating an antigen-experienced phenotype. Furthermore, they express high levels of Nur77, which can be considered as a surrogate marker for TCR signal strength, immediately after positive selection (8). A further increase of TCR signal strength by addition of supraphysiological amounts of ligand or transgenic expression of CD1d provided some evidence for negative selection of iNKT cells (9, 10). Of note, the nature of positively selecting ligands remains largely elusive and is controversially discussed (1). In addition to strong TCR signals, development of iNKT cells depends on costimulatory signals. These are mediated through homotypic interactions of signaling lymphocytic-activation molecule (SLAM) family members (11). Consequently, mice deficient in the SLAM-associated protein (SAP) and its downstream kinase Fyn have severe defects in iNKT cell development at the stage 0 to stage 1 transition (11–15).

microRNAs (miRNAs) are short noncoding RNAs that modulate a large number of biological processes, mostly by down-regulating expression of target genes via mRNA degradation, mRNA destabilization, or interference with translation. miR-181 comprises a family of six miRNAs, which are organized in three clusters (miR-181a/b-1, miR-181a/b-2, miR-181c/d). miR-181a constitutes the most prominently expressed miRNA species in DP thymocytes (16, 17) and has been associated with modulating TCR signal strength via targeting serine/threonine as well as tyrosine phosphatases (18). Consequently, elevated expression of miR-181a results in reduced phosphatase activity and increased TCR signal strength. Recently it has been shown that miR-181a expression prevents the generation of αβ T cells that are strongly reactive toward positively selecting peptides (19).

To date, the effect of aberrant expression of miR-181a on TCR signaling has only been analyzed using short-term assays and in vitro organ cultures. Here we studied the consequences of deletion of miR-181a/b-1 on T-cell development in vivo in the steady state. We found that miR-181a/b-1–deficient mice displayed an almost complete block in early iNKT cell development, resulting in dramatically reduced numbers of iNKT cells in thymus as well as in the periphery. DP thymocytes from miR-181a/b-1–deficient mice displayed diminished signaling upon TCR triggering, leading to an altered TCRβ repertoire in iNKT cells and reduced cytokine production in the periphery. In turn, increasing the availability of agonist ligand overcame the early block in iNKT cell development in these mice.

Taken together, we identified miR-181a/b-1 as a regulator of iNKT cell development and provided evidence for the critical importance of fine-tuned TCR signal strength for agonist-selected T cells.

Results and Discussion

Development of αβ T Cells in Mice Lacking miR-181a/b-1.

Among all miRNAs, miR-181a/b is most prominently expressed in DP thymocytes, in which it constitutes up to 40% of all miRNAs (16, 17). We generated mice carrying a targeted deletion in miR-181a/b-1 (miR-181a/b-1^−/− mice) (Fig. S1). Deletion of miR-181a/b-1 was verified by Northern blot (Fig. 1A). Quantitative RT-PCR with primers recognizing both miR-181a-1 and miR-181a-2 showed that deletion of miR-181a/b-1 resulted in a reduction of all miR-181a species by 98%. This confirmed that miR-181a/b-1 but not miR-181a/b-2 is predominantly expressed in thymus (Fig. 1B). Thymus cellularity of miR-181a/b-1^−/− mice was indistinguishable from WT and heterozygous littermate controls both at 2 wk and 8 wk of age, and gross composition as assessed by staining for CD4 or CD8 was essentially normal (Fig. 1 C–E). In addition, we did not detect major alterations in early thymocyte subsets (Fig. 1F). Despite reduced expression of miR-181a-1 and miR-181b-1 in heterozygous miR-181a/b-1^+/− mice, T-cell development in these mice was comparable to that in WT littermates. Therefore, miR-181a/b-1^+/− mice were used as controls throughout this study. We detected a mild effect of homozygous deletion of miR-181a/b-1 in development of conventional αβ T cells, reflected by slight alterations in early T lineage progenitor numbers compared with controls, an ∼4% decrease in the frequency of DP thymocytes, and a concomitant increase in the frequency of CD4 single positive (SP) thymocytes. These findings are consistent with minor alterations in thymocyte subsets in a different mouse model of deletion of miR-181a/b-1 (20). Taken together, these data indicate that deletion of miR-181a/b-1 does not result in substantial defects in early T-cell development.

Fig. 1.

Development of αβ T cells in mice lacking miR-181a/b-1. (A) Northern blot analysis of miR-181a-1 and miR-181b-1 expression in total thymocytes of miR-181a/b-1^+/+, miR-181a/b-1^+/−, and miR-181a/b-1^−/− mice. U6 snRNA expression was assessed as loading control. (B) Quantitative RT-PCR for mature miR-181a expression in total thymocytes of miR-181a/b-1^+/+, miR-181a/b-1^+/−, and miR-181a/b-1^−/− mice. n = 3 for each genotype. (C) FACS analysis of DN, DP, and SP thymocytes from miR-181a/b-1^+/− and miR-181a/b-1^−/− mice. Representative plots of multiple experiments are depicted. Quantitative plot for DN, DP, and SP thymocytes, n = 5. Percentage within of total thymocytes is shown. (D) Frequency of major thymic subsets and (E) thymus cellularity in miR-181a/b-1^+/− and miR-181a/b-1^−/− mice at various ages (mean + SEM, n = 5–7). (F) FACS analysis of DN thymocyte subsets; ETP, early T lineage progenitor (lineage^–CD44⁺CD117⁺CD25^–); DN2 (lineage^–CD44⁺CD117⁺CD25⁺); DN3 (lineage^–CD44^loCD117^–CD25⁺); DN4 (lineage^–CD44^loCD117^–CD25^–). Mean + SEM, n = 3–4.

Development of iNKT Cells Depends on miR-181a/b-1.

Next we analyzed whether deletion of miR-181a/b-1 resulted in altered generation of iNKT cells. Thymi from miR-181a/b-1^−/− mice showed a 15-fold reduced frequency of iNKT cells, as assessed by surface staining for TCRβ⁺, NK1.1⁺ in conjunction with αGalCer CD1d tetramers (CD1d-tet), compared with controls (Fig. 2 A and B). Absolute numbers of iNKT cells were reduced to a similar extent (Fig. 2B). CD1d-tet^–TCRβ⁺NK1.1⁺ variant (v)NKT cell frequencies and numbers, which are present at much lower frequencies in WT mice than iNKT cells, were also reduced in miR-181a/b-1^−/− mice, albeit less dramatically, compared with iNKT cells (Fig. 2C). To test whether the paucity of thymic iNKT cells penetrated into the periphery, we assessed frequencies and numbers of splenic and liver iNKT cells. In both organs we found ninefold and 11-fold reduced numbers, respectively, of iNKT cells in miR-181a/b-1^−/− mice compared with controls (Fig. 2B). Analysis of vNKT cells in spleen and liver yielded comparable results as analysis of thymus, although the observed differences might be underestimated because low frequencies of vNKT cells and lack of specific staining reagents rendered analysis of this subset difficult (Fig. 2C). Thus, we conclude that deletion of miR-181a/b-1 results in impaired generation of iNKT cells, which cannot be compensated for in the periphery. Mice lacking miR-150 display a defect in the development of iNKT cells, resulting in a modest reduction of iNKT cell numbers in the thymus, which does not penetrate into the periphery (21, 22). Nevertheless, our data imply that the previously reported massive defect in iNKT cell development in mice with conditional deletion of the RNA-processing enzyme Dicer, which is essential for the generation of miRNAs, in hematopoietic cells or DP thymocytes, can be attributed to a large extent to a lack of miR-181a/b-1 (23–25). Of note, other unconventional T-cell populations that have been suggested to depend on agonist selection, such as some γδ T cells and intestinal intraepithelial lymphocytes (iIELs), remained essentially unaffected by deletion of miR-181a/b-1 (Fig. S2). γδ T cells do not pass the DP stage during development, and it is a matter of debate whether iIELs are derived from DP thymocytes (26, 27). Relative expression of miR-181a/b-1 is highest in DP cells, whereas it is expressed at lower levels in double-negative (DN) cells (17). Therefore, it is conceivable that miR-181a/b-1 selectively controls unconventional T-cell subsets that originate from DP thymocytes, such as iNKT cells.

Fig. 2.

Development of iNKT cells depends on miR-181a/b-1. (A) FACS analysis of CD1d-tet⁺ iNKT cells from thymus, spleen, and liver. Plots are representative from two to four individual experiments with three to five mice each. (B) Percentage and absolute numbers of iNKT cells (CD1d-tet⁺TCRβ⁺) in thymus and spleen. Percentage of iNKT cells (CD1d-tet⁺TCRβ⁺) in liver. Mean + SEM, n = 3–10 for each genotype. (C) Percentage and absolute numbers of vNKT cells (CD1d-tet^–NK1.1⁺TCRβ⁺) in thymus and spleen. Percentage of vNKT cells (CD1d-tet^–NK1.1⁺TCRβ⁺) in liver. Mean + SEM, n = 3–10 for each genotype. (D) iNKT cell reconstitution in thymi of competitive BM chimeras. Recipients were CD45.1, competitor donor BM was CD45.1/2, and miR-181a/b-1^+/− or miR-181a/b-1^−/− test BM was CD45.2. Competitor and test BM were administered at a 1:1 ratio. Twelve weeks after BM transplantation, cells were analyzed by FACS. Plots are representative for six mice from two individual experiments. (E) Analysis of indicated thymocyte subsets from competitive BM chimeras. DN, double negative thymocytes. x axis labels denote genotypes of test populations. Each dot represents an individual mouse. (F) iNKT cell reconstitution in spleens of competitive BM chimeras as described in D. Each dot represents an individual mouse.

Deletion of miR-181a/b-1 might prevent the generation of iNKT cells because of failure of extrinsic signals. To directly assess whether deficiency in miR-181a/b-1 inhibited generation of iNKT cells in a cell-intrinsic manner we generated mixed bone marrow (BM) chimeras. To this end, BM cells from congenic competitors (CD45.1/CD45.2) were mixed at a 1:1 ratio with BM cells from miR-181a/b-1^−/− mice or heterozygous controls (CD45.2) and transferred into lethally irradiated hosts (CD45.1). Analysis of mixed chimeras after 12 wk revealed that only few iNKT cells derived from miR-181a/b-1–deficient donor cells were found in thymus and spleen, indicating that lack of miR-181a/b-1 results in a cell-intrinsic defect to generate iNKT cells (Fig. 2 D–F). Of note, heterozygous controls were not significantly outcompeted by WT competitors, further arguing for the absence of a gene dosage effect of miR-181a/b-1 in iNKT cell development (Fig. 2 D–F). Although we noted a slight competitive disadvantage of miR-181a/b-1^−/− DP thymocytes compared with their DN precursors, loss of cells derived from miR-181a/b-1^−/− donors was much more prominent in the iNKT cell fraction compared with DP thymocytes (Fig. 2E). This suggests that miR-181a/b-1 acts at or after the DP stage of development.

miR-181a/b-1 Controls Development of iNKT Cells at the Transition from Stage 0 to Stage 1.

Development of iNKT cells originates from DP thymocytes and then proceeds through CD1d-tet⁺ stages distinguishable by differential expression of CD24, CD44, and NK1.1. To define the developmental stage at which miR-181a/b-1 is required for generation of iNKT cells, we first assessed the frequency of preselection (CD5^–TCRβ^lo) DP thymocytes carrying Vα14Jα18 TCRα rearrangements. Vα14Jα18 rearrangements occur late during the life of a DP thymocyte and, thus, impaired thymocyte survival might affect development of iNKT cells as has been previously reported for mice deficient in the transcription factor c-Myb (28). Semiquantitative RT-PCR analysis revealed that preselection DP thymocytes displayed a similar frequency of Vα14Jα18 rearrangements compared with heterozygous controls, indicating that miR-181a/b-1 is not required to control iNKT cell precursor frequency before selection (Fig. 3A).

Fig. 3.

miR-181-a/b-1 controls development of iNKT cell at the transition from stage 0 to stage 1. (A) Semiquantitative RT-PCR analysis of Vα14Jα18 rearrangements from sorted preselection DP thymocytes (CD5^–TCRβ^lo). Wedges indicate threefold serial dilutions. Expression of the TCRα constant region (Cα) was assessed as positive control. Data are representative for three mice from each genotype. (B) Flow cytometric analysis of iNKT cell development in CD1d-tet⁺ enriched thymocytes from 13-d-old mice. Plots are representative for 12 mice from three individual experiments. Numbers within gates denote developmental stages. Numbers in quadrants and adjacent to gates indicate frequencies of cells within the corresponding gates. (C) Percentage and (D) absolute numbers of CD1d-tet⁺TCRβ⁺ cells in each stage of development gated as depicted in A as mean + SEM; n = 12 for each genotype. (E) BrdU incorporation in DP thymocytes and iNKT cells from 13-d-old mice at different developmental stages after a 2-h pulse. Numbers in plots indicate frequency of cells within gates ± SEM; n = 6 for each genotype.

To faithfully quantify rare progenitor populations, we enriched CD1d-tet⁺ cells from young mice using magnetic beads (Fig. 3B). Frequencies of stage 2 and stage 3 iNKT cells were similar in thymi of miR-181a/b-1^−/− and miR-181a/b-1^+/− mice (Fig. 3 B–D). However, whereas most CD44^–NK1.1^– precursors in miR-181a/b-1^+/− mice were at stage 1 (CD24^–), the vast majority of CD44^–NK1.1^– precursors had retained CD24 expression, indicating that miR-181a/b-1 is required to promote the transition from stage 0 to stage 1 (Fig. 3 B and C). Analysis of absolute numbers of iNKT cell precursors revealed a minor reduction of stage 0 precursors in thymi of miR-181a/b-1^−/− mice compared with heterozygous controls (Fig. 3D). However, in contrast to controls even fewer stage 1 than stage 0 precursors were detectable in thymi of miR-181a/b-1^−/− mice, substantiating the conclusion that the stage 0 to stage 1 transition of iNKT cell development is dependent on miR-181a/b-1. Absolute numbers of stage 2 and stage 3 iNKT cells were also reduced, indicating that these cells do not undergo sufficient compensatory proliferation. To directly test whether the observed developmental block is due to a failure of proliferation, we administered a 2-h BrdU pulse to 13-d-old miR-181a/b-1^−/− mice and heterozygous controls before isolation of iNKT cell precursors. Consistent with our previous observations, a massively reduced number of stage 0 miR-181a/b-1–deficient iNKT cell precursors had incorporated BrdU, whereas all other subsets showed similar levels of BrdU incorporation (Fig. 3E). Taken together, we conclude that miR-181a/b-1 is critical for proliferative expansion of stage 0 iNKT cell precursors. Similarly, failure to undergo proliferative expansion toward developmental stage 1 has been demonstrated in c-Myc–deficient as well as in Pdk1-deficient mice (29, 30). However, this developmental block may also reflect a direct consequence of aberrant TCR signaling during selection as shown for mice deficient in the transcription factors Egr1 and Egr2 (31). Both factors are rapidly induced upon TCR triggering. Analysis of expression of the transcriptional regulators of iNKT cell development promyelocytic leukemia zinc finger (PLZF), Egr1, Egr2, and c-Myc at stage 0 and 1 revealed no clear differences between cells from miR-181a/b-1^+/− and miR-181a/b-1^−/− mice (Fig. S3). These data suggest that either miR-181a/b-1 acts independently of these transcription factors or, alternatively, that miR-181a/b-1–deficient stage 1 iNKT cells represent cells that have escaped developmental arrest. The latter explanation is supported by normal rates of BrdU incorporation at developmental stages 1–3.

Impaired Agonist Selection of iNKT Cells in the Absence of miR-181a/b-1.

To assess whether impaired development of iNKT cells in miR-181a/b-1^−/− mice is a consequence of altered TCR signal strength, we first analyzed expression levels of CD69 and CD5 as indicators of TCR signal strength. Stage 0 cells from miR-181a/b-1^−/− mice displayed a consistent, albeit mild, reduction of expression levels of CD69 and CD5 compared with heterozygous controls (Fig. 4A). Lower levels of CD69 were not apparent in preselection DP cells or at later stages of development, whereas CD5 expression remained lower until stage 3. Next we assessed whether miR-181a/b-1 directly modulated the TCR response of DP thymocytes. To this end, thymocytes from miR-181a/b-1^−/− mice and controls were stimulated with anti-CD3 antibody and assessed for Ca²⁺-flux. In contrast to controls, miR-181a/b-1–deficient DP thymocytes generated only little Ca²⁺ signal upon TCR triggering, directly demonstrating that TCR signaling is impaired in the absence of miR-181a/b-1 (Fig. 4B). Targets of miR-181a comprise several phosphatases capable of negatively regulating TCR signaling (18). Thus, we assessed whether these targets displayed altered expression levels in DP thymocytes from miR-181a/b-1^−/− mice. We found that levels of mRNA coding for Ptpn22, Shp-2 (encoded by Ptpn11), and Dusp6 were increased between 1.3- and 1.8-fold (Fig. 4C). This differential expression was comparable to alterations of phosphatase expression induced by targeting of miR-181a in vitro (18). Thus, permanent ablation of miR-181a/b-1 does not result in compensatory restoration of phosphatase levels. These data are consistent with a previous report implicating miR-181a as a positive regulator of TCR signaling (18) and support the hypothesis that increased expression of miR-181a/b-1–dependent negative regulators results in impaired TCR signaling in thymocytes from miR-181a/b-1^−/− mice.

Fig. 4.

Impaired agonist selection of iNKT cells in the absence of miR-181a/b-1. (A) Surface expression of CD5 and CD69 on iNKT cells at different developmental stages of iNKT cells at 13 d of age. Data are shown as mean fluorescence intensity (mean +SEM). n = 6 for each genotype. (B) Total thymocytes were stimulated with anti-CD3, and Ca²⁺-flux was recorded flow cytometrically over time. Plots for electronically gated DP thymocytes are representative for five mice from two individual experiments. “stim” indicates time point of stimulation. Bar graph shows analysis of peak Ca²⁺-flux over background from five mice per group from two independent experiments. (C) Analysis of expression of Ptpn22, Ptpn11 (encoding Shp-2), and Dusp6 in DP thymocytes by quantitative RT-PCR. Expression levels were normalized to Actb. Data are shown as mean relative expression + SEM from four mice per group out of two individual experiments. (D) TCR Vβ repertoire of splenic CD1d-tet⁺TCRβ⁺ iNKT cells was analyzed by flow cytometry. Pie charts represent frequencies of individual Vβ chains within these iNKT cells. Cumulative data from 13 to 14 mice per genotype from four independent experiments. Table represents quantification of data presented in pie charts. Statistical analysis was performed using unpaired Student t test. (E) Absolute numbers of splenic CD1d-tet⁺TCRβ⁺ iNKT cells expressing different Vβ chains. (F) Cytokine production by splenic iNKT cells from miR-181a/b-1^+/− or miR-181a/b-1^−/− mice after stimulation with αGalCer-loaded splenocytes for the indicated periods of time. IFNγ, IL-4, and IL-13 concentrations were determined using cytometric bead assays. Data are shown as mean + SEM from five mice in two independent experiments. Statistical analysis was performed using unpaired Student t test.

It has been reported that the restricted Vβ chain use in WT iNKT cells is a consequence of antigen recognition and that changes in exposure of iNKT cell precursors to selecting self-ligands resulted in alterations of the TCRβ chain repertoire (3, 32). Furthermore, it has been shown that Vβ7⁺ and Vβ8.2⁺ iNKT cells displayed different affinities for αGalCer, suggesting that different TCRβ chains transmit signals with different signal strength (33). Development of iNKT cells in miR-181a/b-1^−/− mice was not completely abrogated, thus allowing us to test the prediction that altered TCR signaling results in changes of the TCR repertoire of successfully selected iNKT cells. Analysis of Vβ chain use by splenic iNKT cells revealed that a reduced proportion of splenic iNKT cells from miR-181a/b-1^−/− mice expressed Vβ8 compared with splenic iNKT cells from miR-181a/b-1^+/− littermates (Fig. 4D). In contrast, frequencies of Vβ7⁺ and Vβ2⁺ were only marginally affected by homozygous deletion of miR-181a/b-1. Instead, we noted increased relative use of Vβ4, Vβ5, and Vβ11 chains, all of which are barely detectable on WT iNKT cells. Of note, absolute quantification of iNKT cells expressing different Vβ chains revealed that changes in relative Vβ use were largely due to a selective loss of iNKT cells expressing Vβ8, Vβ7, and Vβ2, which constitute the majority of iNKT cells in WT mice, but also Vβ9 and Vβ10. In contrast, absolute numbers of Vβ4⁺, Vβ5.1/5.2⁺, Vβ6⁺, and Vβ11⁺ iNKT cells remained essentially unaffected by deletion of miR-181a/b-1 (Fig. 4E). Given that the biased Vβ chain use in WT iNKT cells is a consequence of antigen recognition rather than preferential pairing of TCR chains (3), these findings are consistent with the hypothesis that miR-181a/b-1 controls iNKT cell differentiation via setting a TCR signaling threshold. A selective loss of iNKT cell subsets expressing Vβ8, Vβ7, and Vβ2 is also consistent with our data showing that key transcriptional regulators of iNKT cell development were not aberrantly expressed in iNKT cell precursors that could still be recovered from miR-181a/b-1^−/− mice (Fig. S3). To assess possible functional consequences of an altered TCR repertoire of iNKT cells from miR-181a/b-1^−/− mice, we analyzed their capacity to produce cytokines in response to stimulation. To this end, miR-181a/b-1–deficient and sufficient splenic NKT cells were cocultured with αGalCer-pulsed splenic dendritic cells, and cytokine production was assessed after 16 and 72 h. At both time points splenic NKT cells from miR-181a/b-1^−/− mice produced considerably less cytokines, in particular IL-4, compared with controls (Fig. 4F), suggesting that altered selection of iNKT cells affects their responsiveness, at least to stimulation with a model antigen, in the periphery. Thus, deletion of miR-181a/b-1 results in a reduced responsiveness toward TCR stimulation and, consequently, in an altered repertoire and reduced peripheral function of iNKT cells.

Exogenous Ligand Rescues iNKT Cell Development Beyond Stage 0 in miR-181a/b-1^−/− Mice.

If reduced responsiveness to TCR signals is causative for defective development of iNKT cells in miR-181a/b-1^−/− mice, administration of supraphysiological concentrations of agonist ligand should be able to compensate for cell-intrinsic reduction of TCR signal strength. To test this prediction, we injected αGalCer i.p. into 13 d-old miR-181a/b-1^−/− mice and heterozygous controls and assessed its effect on the frequency of the iNKT cell precursor stages after 16 h. Administration of αGalCer did not result in significant changes in frequency of stage 2 and stage 3 cells in either miR-181a/b-1–deficient or control mice after this short course of treatment with αGalCer (Fig. 5A). In contrast, we observed a reversal of frequencies of stage 0 and stage 1 precursors in miR-181a/b-1–deficient mice after αGalCer injection, whereas the frequency of stage 1 precursors in control mice was only mildly increased. Notably, αGalCer treatment of miR-181a/b-1–deficient mice resulted in an increase of absolute numbers of stage 1 cells, whereas numbers of stage 0 cells remained almost constant (Fig. 5B). We conclude that increased availability of agonist ligand, and thus increased capacity for TCR signaling, is able to rescue both the developmental block as well as the failure of proliferative expansion at the stage 0 to stage 1 transition in miR-181a/b-1^−/− mice.

Fig. 5.

Exogenous ligand rescues iNKT cell development beyond stage 0 in miR-181a/b-1^−/− mice. (A) Flow cytometric analysis of thymic iNKT cells from 13-d-old mice 16 h after i.p. injection of αGalCer or vehicle control. Plots are representative for 10–13 mice from each genotype from two individual experiments. (B) Numbers of stage 0 and stage 1 iNKT cell precursors after i.p. injection of αGalCer relative to vehicle control. n = 10–13 for each genotype.

Taken together, our data indicate that deficiency in miR-181a/b-1 results in reduced TCR signaling and, consequently, in an altered TCR repertoire and peripheral responsiveness of iNKT cells. Conversely, rescue by supraphysiological concentrations of agonist ligand supports the hypothesis that miR-181a/b-1 controls iNKT cell development by modulating agonist selection.

Modest alterations of target gene expression, such as increased expression of Ptpn22, Shp-2, and Dusp6 in miR-181a/b-1^−/− mice, seem to be the rule rather than an exception for gene regulation by miRNA. Consequently, it has been suggested that, rather than controlling major biological switches, miRNAs contribute to managing noise and/or setting regulatory thresholds (34). Thus, virtually complete control of development of a cell lineage by a single miRNA as described in this study is to our knowledge unparalleled in the immune system. miRNAs have been suggested to frequently target multiple genes within a particular pathway, thus generating synergistic effects (35). Here, deletion of miR-181a/b-1 resulted in increased expression of multiple negative regulators of TCR signaling, and it has been proposed that their coordinated, but not individual, regulation by miR-181a modulates TCR signaling thresholds (18). iNKT cells seem to be particularly sensitive to alterations in TCR signal strength. For example, reduction of TCR signal strength by reduction of immunoreceptor tyrosine-based activation motif motifs in the CD3ζ chain resulted in a preferential defect in iNKT cell development (36). In addition, generation of iNKT cells is dependent on costimulation via SLAM-SAP-Fyn signaling (11). Fyn activity is also negatively regulated by target phosphatases of miR-181a/b-1, suggesting that its deletion results in defective integration of TCR stimulation and costimulation, thus providing an additional explanation why iNKT cell development is selectively controlled by miR-181a/b-1.

In conclusion, we have identified a single miRNA species, miR-181a/b-1, that controls differentiation of iNKT cells by setting a threshold for agonist selection.

Materials and Methods

Mice.

C57BL/6J mice (CD45.2) were purchased from Charles River. B6.SJL-Ptprc^aPepc^b/BoyJ mice (termed “B6 CD45.1” throughout this article) and (C57BL/6J × B6 CD45.1) F1 mice (CD45.1/CD45.2 heterozygous) were bred at the animal facility of Hannover Medical School. Animals were maintained under specific-pathogen-free conditions. All animal experiments were conducted in accordance with local and institutional guidelines (Permit: 33.9-42502-04-12/0869, 07/1393, 08/1480). Generation of miR-181a/b-1 knockout mice is described in detail in SI Materials and Methods.

Flow Cytometry and Cell Sorting.

Phycoerythrin (PE)- and allophycocyanin (APC)-conjugated CD1d/αGalCer tetramer was provided by R. Hurwitz (Max Planck Institute for Infection Biology, Berlin, Germany), PE conjugated CD1d/αGalCer tetramer was purchased from ProImmune. APC-conjugated CD1d/PBS-57 (αGalCer analog) loaded and unloaded tetramer were provided by the National Instiutes of Health Tetramer Facility at Emory University (Atlanta, GA). Monoclonal antibodies used in this study are described in SI Materials and Methods.

Enrichment and Analysis of iNKT Development.

For analysis of iNKT cells, thymocyte single-cell suspensions obtained from 13-d-old mice were stained for 15 min at room temperature with APC-labeled CD1d/PBS-57 tetramers. CD1d-tet⁺ cells were then enriched with anti-APC magnetic microbeads by using a magnetic-activated cell sorting cell separator (Miltenyi Biotec) and subjected to further staining by flow cytometry.

Application of αGalCer.

Thirteen-day-old miR-181a/b-1^+/− or miR-181a/b-1^−/− mice were injected i.p. with αGalCer (Alexis Biochemicals) at 1 µg per mouse per 50 µL of PBS, or PBS/DMSO. Sixteen hours later, mice were killed, and intrathymic development of iNKT cells was analyzed as described above.

Additional Methods.

Northern Blot analysis, PCR, procedures for enrichment of DN thymocytes, generation of competitive BM chimeras, analysis of TCR Vβ chain repertoire, measurement of intracellular Ca²⁺-flux in DP thymocytes, and BrdU incorporation are described in SI Materials and Methods.

Statistical Analysis.

All analysis was performed using GraphPad Prism software. Data are represented as mean + SEM. Analysis of significance between two groups of mice was performed using unpaired Student t tests.