Dicer-Dependent MicroRNA Pathway Controls Invariant NKTCell Development1

Maya Fedeli,* Anna Napolitano,* Molly Pui Man Wong,† Antoine Marcais,‡ Claudia de Lalla,*Francesco Colucci,† Matthias Merkenschlager,‡ Paolo Dellabona,2* and Giulia Casorati2*

Invariant NK T (iNKT) cells are a separate lineage of T lymphocytes with innate effector functions. They express an invariantTCR specific for lipids presented by CD1d and their development and effector differentiation rely on a unique gene expres-sion program. We asked whether this program includes microRNAs, small noncoding RNAs that regulate gene expressionposttranscriptionally and play a key role in the control of cellular differentiation programs. To this aim, we investigatediNKT cell development in mice in which Dicer, the RNase III enzyme that generates functional microRNAs, is deleted incortical thymocytes. We find that Dicer deletion results in a substantial reduction of iNKT cells in thymus and theirdisappearance from the periphery, unlike mainstream T cells. Without Dicer, iNKT cells do not complete their innate effectordifferentiation and display a defective homeostasis due to increased cell death. Differentiation and homeostasis of iNKT cellsrequire Dicer in a cell-autonomous fashion. Furthermore, we identify a miRNA profile specific for iNKT cells, which exhibitsfeatures of activated/effector T lymphocytes, consistent with the idea that iNKT cells undergo agonist thymic selection.Together, these results define a critical role of the Dicer-dependent miRNA pathway in the physiology of iNKT cells. TheJournal of Immunology, 2009, 183: 2506 –2512.

I nvariant natural killer T cells are a separate lineage of innate-like T lymphocytes that in mice express a TCR made of theinvariant V�14–J�18 rearrangement paired with V�8.2, �7,

or �2 (1). Invariant NK T (iNKT)3 cells recognize both self andexogenous lipids presented by the MHC class I-like molecule CD1dand play a role in infection, autoimmunity, allergy, and cancer (1).

iNKT cell developmental pathway is distinct from that of main-stream T lymphocytes (1, 2). CD4CD8 double positive (DP) pre-cursors that have randomly rearranged the canonical invariantTCR are positively selected into the iNKT cell lineage upon rec-ognition of endogenous agonist lipids, presented by CD1d on DPthymocytes (1, 2). Postselection iNKT cells undergo an orderedphenotypic maturation characterized by four stages: the first,detectable immediately after positive selection, is the HSAhigh

CD69� stage 0, which becomes HSAlowCD44lowNK1.1� stage 1,followed by CD44highNK1.1� stage 2, and finally by the mature

CD44highNK1.1� stage 3, which occurs both in the thymus andperiphery (3–5). Maturation is accompanied by a massive cellularexpansion between stage 1 and 2 (4). Furthermore, developingiNKT cells acquire also effector cytokine expression in the thymusindependently of foreign Ag encounter (4, 5), resulting in the rapiddisplay of effector functions in thymic iNKT cell emigrants.

The iNKT cell development is selectively controlled by theSLAM/Fyn/SAP/PKC-� signal transduction pathway, the NF�B,T-bet, PLZF, and Egr2 transcription factors, and the cytokineIL-15 (1, 2, 6–8). Deletion of one of these molecules impairsiNKT cell ontogeny, whereas it has little or no effects on the de-velopment of mainstream T lymphocytes.

Collectively, this evidence argues for a unique genetic programcontrolling iNKT cell development.

Small microRNAs (miRNAs) are short (�22 nt) noncoding RNAsthat control gene expression at the posttranscriptional level by bindingtarget protein-encoding mRNAs, inducing their translational repres-sion or degradation, depending on the degree of complementarity (9).Functional mature miRNAs are generated in the cytoplasm by a com-plex containing the RNase III enzyme Dicer and TRBP (HIV 1–trans-activating response RNA-binding protein) (9).

miRNAs regulate basic functions such as proliferation, apopto-sis, lineage commitment, or differentiation of cells that constitutedifferent tissues and organs, including the immune system (9, 10).Conditional deletion of Dicer provided insight into the role ofmiRNAs in controlling the development of B and T lymphocytes(11–16). Dicer deletion early in T cell development (DN stage)caused a 10-fold decrease in DP and single-positive (SP) thymo-cyte numbers, though the CD4/CD8 lineage choice was unaffected(13). Dicer deletion at later stage of T cell development (DP stage)did not alter number and composition of mainstream thymocytes,though it resulted in a substantial reduction in thymic CD4�CD25�

Foxp3� natural regulatory T (Treg) cells (14, 17). In the peripheralcompartment, these mice showed a modest reduction in CD4� andCD8� T cells, Th1 polarization, and increased apoptosis in CD4�

T cells (12).

*Experimental Immunology Unit, Division of Immunology, Transplantation and In-fectious Diseases, H. San Raffaele Scientific Institute, Milan, Italy; †Laboratory ofLymphocyte Signaling and Development, Babraham Institute, Cambridge, UnitedKingdom; and ‡Lymphocyte Development Group, Medical Research Council ClinicalSciences Centre, Imperial College London, London, U.K.

Received for publication April 30, 2009. Accepted for publication June 20, 2009.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 This study was supported by Associazione Italiana per la Ricerca sul Cancro, Fonda-zione Cariplo, Italian Ministry of Health (to P.D. and G.C.) and Biotechnology andBiological Sciences Research Council (to F.C.). M.F. is supported by fellowshipsfrom the PhD Program in Molecular Medicine, University of Vita-Salute San Raffaele(Milano, Italy).2 Address correspondence and reprint requests to Dr. Paolo Dellabona or Giulia Ca-sorati, Experimental Immunology Unit, Division of Immunology, Transplantation andInfectious Diseases, H. San Raffaele Scientific Institute, Via Olgettina 58, Milan,Italy. E-mail addresses: dellabona.paolo@hsr.it and casorati.giulia@hsr.it3 Abbreviations used in this paper: iNKT, invariant NK T cell; DP, double positive;miRNA, microRNA; Treg, regulatory T cell; wt, wild type; SP, single positive; qRT-PCR, quantitative RT-PCR.

Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00

The Journal of Immunology

www.jimmunol.org/cgi/doi/10.4049/jimmunol.0901361

Foxp3-driven deletion of Dicer selectively in Treg cells im-paired their peripheral homeostasis (16) and suppressor functions(15, 16), supporting the critical requirement for Dicer-controlledmiRNA pathway in this T cell lineage.

In the light of these considerations, we asked whether Dicer-dependent miRNAs play any role in controlling the unique iNKTcell developmental program. To this aim, we investigated theiNKT cell development in a set of mice in which Dicer has beendeleted from cortical thymocytes.

Materials and MethodsMice

Dicerlox/lox, cd4Dicer�/�, and lckDicer�/� transgenic mice were described(13, 14). Dicerlox/lox were crossed with both hCD2Cre and R26R-EYFPtransgenic mice (18) to delete one (cd2Dicer�/wt) or both (cd2Dicer�/�)Dicerlox alleles expressing EYFP upon deletion of a floxed transcriptionalstop cassette inserted into the rosa26 locus. All mice were housed in apathogen-free environment. Procedures involving animals were approvedby the Institutional Animal Care and Use Committee at San Raffaele Sci-entific Institute or performed according to the Animals (ScientificProcedures) Act.

Flow cytometry and cell sorting

Cells from thymus, spleen, and liver were purified, stained, and sorted asdescribed (19) using the following mAbs: anti-TCR�-allophycocyanin,CD3-PECy5, anti-NK1.1-PerCP, HSA-FITC, CD4-PerCP or allophyco-cyanin, CD8�-PE, CD44-allophycocyanin, CD19-FITC, CD1d-PE,streptavidin-allophycocyanin, SLAMF3-bio (BD Bioscience), SLAMF1-APC, SLAMF5-bio (Biolegend), SLAMF6-bio (eBioscience). Enrichmentof HSAlow mature thymocytes by anti-HSA mAb and rabbit complementwas performed as described (19). CD1d-IgG Dimerix (BD Bioscience) ormCD1d tetramers (Proimmune) were loaded with �GalCer (Alexis) as de-scribed (19). BrdU and 7-aminoactinomycin D stainings were performedusing the BrdU Flow Kit (BD Biosciences). To investigate the iNKT cellstage 0 of development, 2 � 106 thymocytes were stained for 1 h on icewith 0.5 �g DimerX-CD1d followed by 0.5 �g rat anti-mouse IgG1 PE(clone A85-1, BD). DimerX-positive cells were enriched with anti-PE im-munomagnetic beads (Miltenyi Biotec) following the manufacturer’s in-struction and stained with anti-TCR�-allophycocyanin and anti-HSA-FITCmAbs. As control, empty DimerX-CD1d did not stain thymocytes abovethe background level. Cells were analyzed on FACSCanto or LSRII andsorted on MoFlo (Beckman Coulter), always excluding dead cells by theuse of the DNA specific dye DAPI (Santa Cruz Biotechnology) and dou-blets by comparing side scatter width to forward scatter area.

Mixed radiation bone marrow chimeras

Mixed radiation bone marrow chimeras were performed as described (20).

MicroRNA array profiling

Total RNA was extracted from total thymocytes, and from sorted maturethymic iNKT cells or T cells. Three independent samples for each cell typewere obtained from 18 pooled thymi of 4-wk-old C57BL6 mice. Its qualitywas verified by an Agilent 2100 Bioanalyzer profile. Ninety nanograms oftotal RNA from samples (iNKT or T cells) and reference (total thymocytes)was labeled with Hy3 and Hy5 fluorescent label, respectively, using themiRCURY LNA Array power labeling kit (Exiqon). The Hy3-labeled sam-ples and a Hy5-labeled reference RNA sample were mixed pair-wise andhybridized to the miRCURY LNA array version 11.0 (Exiqon), whichcontains capture probes targeting all miRNAs for human, mouse, or ratregistered in the miRBASE version 12.0 at the Sanger Institute. The hy-bridization was performed according to the miRCURY LNA array manualusing a Tecan HS4800 hybridization station (Tecan). After hybridization,the microarray slides were scanned and stored in an ozone-free environ-ment (ozone level below 2.0 ppb) to prevent potential bleaching of thefluorescent dyes. Scanning was performed with the Agilent G2565BA Mi-croarray Scanner System (Agilent Technologies) and the image analysiswas conducted using the ImaGene 8.0 software (BioDiscovery). The quan-tified signals were background corrected (Normexp with offset value 10)and normalized using the global Lowess (LOcally WEighted ScatterplotSmoothing) regression algorithm.

The microarray data are deposited in Geo record: http://www.ncbi.nlm.nih.gov/projects/geo GSE15925 - Mature thymic NKT and T cells.

Quantitative real-time RT-PCR

To quantify miRNA expression, gene-specific reverse transcription wasperformed using TaqMan MicroRNA Reverse Transcription kit (AppliedBiosystems). RT-PCR was performed using TaqMan MicroRNA AssayMix containing PCR primers and TaqMan probes (Applied Biosystems)according to manufacturer instructions. Expression values were normalizedto snor420 and reconfirmed using snor202.

Statistical analysis

The results were analyzed by a two-tailed Student t test. A value of p �0.05 was considered statistically significant.

ResultsAbsence of Dicer in cortical thymocytes results in a substantialiNKT cell reduction

To investigate the role for Dicer-controlled miRNAs in the reg-ulation of the iNKT cell developmental program, we first de-termined the frequency and number of these cells in the thymusand peripheral organs of mice carrying a conditional mutationof dicer-1 gene (Dicerlox/lox), activated by Cre recombinase un-der the control of the cd4 enhancer/promoter/silencer (CD4Cre)(14). CD4Cre causes 90% deletion of Dicer in DP thymocytesthat, however, still express abundant mature miRNAs (12, 14).Dicerlox/lox deletion induced by CD4Cre is essentially completein SP thymocytes and mature miRNAs are reduced �10-foldin naive T cells (12, 14). As shown in Fig. 1, Dicer deletion inDicerlox/lox mice induced by CD4Cre (cd4Dicer�/� mice) in-deed resulted in �10-fold reduction in thymic iNKT cells com-pared with control thymi from Dicerlox/lox littermate mice.iNKT cell depletion in cd4Dicer�/� mice was even more pro-found in the peripheral compartment. iNKT cells could behardly detected by CD1d-tetramer staining in the liver andspleen of cd4Dicer�/� mice (Fig. 1), approaching backgroundvalues obtained in CD1d�/� mice that completely lack iNKTcells. In lymph nodes, iNKT cells from cd4Dicer�/� mice werealso markedly reduced compared with Dicerlox/lox littermates,although this difference did not reach statistical significance(Fig. 1).

By contrast, total thymocyte numbers, and both frequency andnumber of mainstream DP and SP thymocytes, were normal incd4Dicer�/� thymi, while peripheral CD4� and CD8� T cells weremoderately reduced (2-fold and 4-fold, respectively) in cd4Dicer�/�

mice (supplemental Fig. 1)4 compared with Dicerlox/lox control mice,and consistent with published data (12, 14).

Thus, Dicer deletion in DP thymocytes resulted in a dramaticreduction of iNKT but not of T cells, in both thymic and peripheralcompartments.

We also investigated the effects of Dicerlox/lox depletion oniNKT cell development by Cre recombinase driven by hCD2 pro-moter (cd2Dicer�/�) or proximal pLck (lckDicer�/�) promoters,which delete at the DN3 stage (18, 21) significantly earlier in T celldevelopment than CD4Cre, resulting in the depletion of Dicer andof mature miRNAs already at the DP stage (13) (and supplementalFig. 2A). Dicerlox/lox deletion by pCD2cre resulted in a 14-foldreduction in the total number thymocytes compared with heterozy-gous deleted control mice (8.4 � 6 � 106 in cd2Dicer�/� vs 120 �45.1 � 106 in cd2Dicer�/wt mice), which essentially depended ona dramatic drop in the number of DP (�100-fold), CD4 SP (�20-fold) and, to a lesser extent, CD8 SP (6-fold) thymocyte (supple-mental Fig. 2B), while DN thymocytes were even slightly in-creased in cd2Dicer�/� compared with cd2Dicer�/wt thymi (6.3 �5 � 106 and 2.6 � 1.4 � 106, respectively). In the peripheral

4 The online version of this article contains supplemental material.

2507The Journal of Immunology

compartment, mainstream T cells from cd2Dicer�/� mice werereduced only three times compared with T cells from heterozygouscd2Dicer�/wt controls, due to the expansion of T cells that failed todelete Dicer in thymus (supplemental Fig. 2C). By contrast,analysis of iNKT cells demonstrated a complete loss of �Gal-Cer-CD1d tetramer� cells from thymus, spleen and liver ofcd2Dicer�/� mice (supplemental Fig. 2D). The reduction ofiNKT cells in cd2Dicer�/� mice was not proportional to thatof mainstream T cells, suggesting again that the development of thissubset was critically dependent on Dicer-dependent miRNAs.

Comparable depletion of iNKT cells and mainstream T cellswere observed in thymus and periphery of lckDicer�/� mice (datanot shown).

Together, these results showed that, unlike mainstream T cellsthat require Dicer during the early developmental stage (DN) butnot after their lineage commitment (DP stage), iNKT cells con-tinue to require Dicer-dependent miRNAs also after this stage,suggesting a critical role for miRNAs in controlling the iNKT celldifferentiation program.

Dicer deletion in cortical thymocytes impairs iNKT celldifferentiation program

We next investigated whether Dicer-dependent miRNAs playedany role in the phenotypic maturation followed by iNKT cells inthymus. We concentrated on cd4Dicer�/� mice because theyhad detectable iNKT cells in the thymus. As shown in Fig. 2A,the differentiation program of thymic iNKT cells in 4-wk-oldcd4Dicer�/� mice was arrested at the stage 2. In Dicerlox/lox

littermate controls, iNKT cells maturation progressed to stage3, although the percentage of NK1.1 expressing cells was lowbecause of the young age. The defective maturation of iNKTcells in cd4Dicer�/� mice did not improve with age (8-wk-oldmice) (Fig. 2B), when the majority of thymic iNKT cells fromDicerlox/lox littermate were NK1.1�.

Interestingly, the great majority of thymic iNKT cells fromcd4Dicer�/� mice expressed CD4 (Fig. 2C), unlike iNKT cellsfrom Dicerlox/lox thymi, which developed into both CD4� andCD4� subsets, a bifurcation that is suggested to occur betweenstage 1 and 2 (22). Thus, without Dicer-controlled miRNAs, iNKT

cells may not reach the maturation stage at which the CD4� subsetemerges. Alternatively, but not mutually exclusive, CD4 shut offand the emergence of the CD4� iNKT cell subset could be directlycontrolled by miRNAs.

Dicer deletion in DP thymocytes by CD4Cre, however, did notaffect the very early stages of iNKT cell development (stage 0),immediately following positive selection (Fig. 2D), in line with thenormal miRNA content displayed by DP thymocytes in this model(12, 14). Comparable quantities of canonical invariant V�14-J�18transcripts were also detected by quantitative RT-PCR (qRT-PCR)in sorted DP thymocytes from cd4Dicer�/� mice and Dicerlox/lox

littermate controls, supporting the flow cytometry data on stage 0iNKT cells.

Collectively, these findings strongly suggested that the iNKTcell differentiation program is critically controlled by Dicer-depen-dent miRNAs.

Altered homeostasis of the iNKT cells in the absence of Dicer

The reduced number of thymic iNKT cells caused by the lack ofDicer could result either from an impaired expansion of imma-ture iNKT cells or from an increased cell death, or both. Asdetermined by BrdU incorporation in vivo, depletion of Dicer incd4Dicer�/� mice did not modify the proliferation of thymiciNKT cell precursors compared with control mice (Fig. 3A).Coupling BrdU incorporation in vivo with in vitro staining forthe DNA content revealed that the frequency of immature iNKTcells at the G2/M phase was significantly higher in cd4Dicer�/�

thymi than Dicerlox/lox control thymi (8 and 1%, respectively,Fig. 3B), suggesting a possible mitotic defect in iNKT cells thatlack Dicer.

In contrast to the rate of cell division, cell death in thymic iNKTcells from cd4Dicer�/� mice exceeded by 10-fold that of iNKTcells from Dicerlox/lox controls (Fig. 3C). Cell death in Dicer-de-ficient iNKT cells occurred mainly at stage 2 (CD44highNK1.1�),which is normally characterized by marked proliferation (4), sug-gesting a possible link between cell division and death of iNKTcells in the absence of miRNAs.

FIGURE 1. Dicer deletion in corti-cal thymocytes results in a substantialiNKT cell reduction. Quantification ofiNKT cells from thymus and peripheryof 8-wk-old cd4Dicer�/� mice. A,Cells were stained with �GalCer-loaded CD1d tetramers (CD1d tet),TCR�, and HSA (thymus) or CD19(periphery) specific mAbs. Shown arethe percentages of iNKT cells in thy-mus (HSAlowCD1d�TCR��) andperiphery (CD19�CD1d�TCR�). B,Frequency and number of iNKTcells in thymus and periphery ofcd4Dicer�/� mice and Dicerlox/lox

controls. Data are representative offive pairs of mice. (t test: �, p �

0.05; ��, p � 0.01).

2508 miRNAs CONTROL iNKT CELLS

Unlike iNKT cells, both cell division and death of mainstreamT cells were comparable in cd4Dicer�/� and Dicerlox/lox litter-mates, indicating that the homeostasis of thymic T cells was notaffected by the deletion of Dicer at the DP stage (data not shown).

Collectively, these results suggested that the profound depletionof iNKT cells observed in the absence of Dicer resulted from analtered homeostasis of this lymphocyte subset due to increased celldeath.

The iNKT cell developmental defect caused by Dicer deficiencyis cell-autonomous

To determine the mechanisms by which Dicer deletion impairediNKT cell development, we first investigated whether DPthymocytes from cd4Dicer�/� mice exhibited defective CD1dexpression, CD1d-dependent lipid Ag presentation or expres-sion of the SLAMF molecules shown to play a role in iNKT

cell development (23). DP thymocytes from cd4Dicer�/� andDicerlox/lox littermate control mice expressed similar levels ofCD1d and presented with comparable efficiency the exogenousglycolipid �GalCer to an iNKT cell hybridoma (supplemental Fig.3). Furthermore, SLAMF1, F3, F5, and F6 molecules were also ex-pressed at comparable levels in DN, DP, CD4 SP, and CD8 SPthymocyte subsets from cd4Dicer�/� and Dicerlox/lox control mice(supplemental Fig. 3). These results ruled out that the lack of Dicer-dependent miRNAs was affecting key molecular interactions neces-sary for iNKT cell development.

To determine whether the iNKT cell developmental defectcaused by the lack of Dicer was cell autonomous, we verifiedwhether the development of Dicer-deficient iNKT cells could berescued by Dicer-sufficient thymocytes in mixed BM chimeras.Lethally irradiated cd4Dicer�/� mice were reconstituted with anequal mixture of BM cells derived from CD45.1 wild type (wt)

FIGURE 2. Dicer deletion in cor-tical thymocytes impairs iNKT celldifferentiation program. Phenotype ofthymic iNKT cells from cd4Dicer�/�

mice. Total thymocytes were stainedwith CD1d tet or CD1d dimerX,HSA, CD3, NK1.1, and either CD44or CD4 specific mAbs. Phenotype,frequency and number of iNKT cells(HSAlowCD3�CD1d tet�) from 4-(A) and 8-wk-old (B) cd4Dicer�/�

mice and controls. Histograms dataare expressed as the mean � SD.Three to four mice per group wereindependently analyzed in each ex-periment. C, CD4 and NK1.1 ex-pression on iNKT cells from 4-wk-old cd4Dicer�/� mice and controls.D, Phenotype of iNKT cells enrichedfrom total thymocytes of 2-wk-oldmice by CD1d dimerX staining andimmunomagnetic sorting. Thymo-cytes pooled from five mice per groupwere analyzed. One experiment rep-resentative of two is shown.

2509The Journal of Immunology

mice and CD45.2 cd4Dicer�/� mice. As shown in Fig. 4, the ma-jority of iNKT cells present in the thymi of the BM chimeras weremature NK1.1� cells derived from the CD45.1 wt BM. Only a

minority of iNKT cells were derived from the CD45.2 cd4Dicer�/�

BM cells and displayed an immature NK1.1� phenotype. UnlikeiNKT cells, T cells developed normally from both wt andcd4Dicer�/� mice (Fig. 4).

Hence, the impaired iNKT cell development caused by the de-letion of Dicer could not be rescued by wt thymocytes, arguingstrongly for an iNKT cell-autonomous defect.

iNKT cells display a distinct miRNA profile

Given the selective role played by Dicer-dependent miRNAs inthe control of iNKT cell development, we sought to determinethe miRNA profile expressed by iNKT cells in comparison withmature thymic T cells. The miRNAs obtained from sortedHSAlow-enriched iNKT and T cells, in triplicate samples eachoriginating from a pool of six C57BL/6 mice, were profiled withLNA (locked nucleotide acid)-based miRNA microarrray. Theanalysis identified 70 miRNAs expressed in thymic iNKT and Tcells, 17 of which were differentially expressed between the twocell subsets at a statistically significant level (Fig. 5A). Quan-titative RT-PCR confirmed that miR-21 was overexpressed,while 13 miRNAs were underexpressed in iNKT cells comparedwith T cells (Fig. 5B). The overexpression of miR-290, miR-483* and miR-720 in iNKT cells was not confirmed byqRT-PCR.

FIGURE 3. Altered homeostasis ofthe iNKT cell compartment in the ab-sence of Dicer. BrdU incorporation andapoptosis in thymic iNKT cells fromcd4Dicer�/� mice. A, Two-wk-oldmice received BrdU iv and were sacri-ficed after 1 h. Enriched HSAlow thy-mocytes from were stained with CD1dtet, TCR�, NK1.1, and BrdU-specificAbs. Depicted are BrdU� iNKT cellsgated in CD1d tet�TCR�� cells. B,BrdU-labeled mature thymocyteswere obtained and stained as indi-cated in A plus 7-aminoactinomycinD. Shown is the cell cycle distribu-tion of iNKT cells. The frequenciesof iNKT cells in G2/M phase fromDicer-deficient and sufficient mice aresignificantly different (p � 0.01). C, En-riched HSAlow thymocytes from4-wk-old mice were incubated withfluorescent CaspAce for 20 h at 37°C,and stained with CD1d tet, TCR�,NK1.1, CD44-specific mAbs. Pheno-type and frequency of apoptotic iNKTcells gated in TCR��CD1d tet� isshown. D, Fold change in apoptoticmature T (TCR��CaspAce�) andiNKT (TCR��CD1dtet�CaspAce�)cells from thymi of cd4Dicer�/� com-pared with controls (indicated by thedashed line). Data are expressed as themean � SD calculated from three in-dependent experiments. A and B showone representative experiment.

FIGURE 4. Dicer deficiency causes a cell-autonomous iNKT cell de-velopmental defect. Generation of mixed BM chimeras. Lethally irradiatedcd4Dicer�/� mice were reconstituted with equal numbers of BM cells fromCD45.1� C57BL/6 and CD45.2� cd4Dicer�/� mice. After 8 wk, thymocyteswere stained with CD1d tet, CD3, NK1.1, and HAS-specific mAbs. Shown arephenotype and frequency of HSAlow thymocytes and iNKT cells derived fromwt or cd4Dicer�/� BM cells. Data are representative of five BM chimerasanalyzed. The experiment was repeated twice with comparable results.

2510 miRNAs CONTROL iNKT CELLS

Thus, this analysis revealed an iNKT lineage-specific miRNAprofile, substantially different from that displayed by maturethymocytes, in line with the selective role of Dicer-controlledmiRNAs in the control of the iNKT lineage-specific geneticprogram.

DiscussionCD1d-dependent iNKT cells are a separate lineage of T lympho-cyte that undergoes a distinct developmental pathway controlledby a unique gene expression program. In this study, we show thatthis program includes microRNAs, small noncoding RNAs thatregulate gene expression posttranscriptionally and play a key rolein the control of cellular differentiation programs. We provide a setof compelling evidences suggesting that the differentiation and ho-meostasis of iNKT cells require Dicer, the RNase III enzyme thatgenerates functional miRNAs, in a cell-autonomous fashion. Re-markably, the development of mainstream T cells is largely unaf-fected by Dicer deletion, underscoring the uniqueness of the iNKTcell program.

Furthermore, we have identified an iNKT cell-specific miRNAprofile, different from that of mature mainstream thymocytes,which exhibits features of activated/effector T cells.

These findings reveal the critical role of the Dicer-dependentmiRNA pathway in the physiology of iNKT cells. Furthermore,they provide new targets to investigate the molecular pathwayscontrolling iNKT cell development in future work.

Several studies have shown that conditional Dicer ablation isassociated with an increased cell death in the targeted cells (24,25), suggesting that a Dicer-dependent miRNA pathway plays animportant role in controlling cell survival and tissue homeostatis(26). Dicer deletion in B cells resulted in the block of B cell de-velopment that was partially due to increased apoptosis at thepre-B stage, linked to an aberrantly high expression of the pro-apoptotic gene Bim (11). Without Dicer, cells seem particularlyvulnerable to cell death during cell division, possibly because ofmitotic defects due to centromer dysfunctions and premature sisterchromatid separation (24). The extensive cell division occurring inthymocytes between the DN and DP stages has also been impli-cated in the marked apoptosis of DP thymocytes observed in lck-Dicer�/� mice, in which Dicer is deleted at the DN stage (13).Furthermore, Muljo et al. (12) showed that mature peripheral Tcells from cd4Dicer�/� undergo increased cell death compared

with controls upon TCR-dependent activation, which induces apotent proliferative stimulus.

Interestingly, the miRNA profile found in thymic iNKT cellsshares several similarities with that described for effector T cells(14, 27), obtained upon in vitro activation of peripheral naiveCD8� or CD4� T cells. Indeed, miR-15b, miR-16, miR-30c, miR-150, and Let-7 family were down-regulated in thymic iNKT cellsas well as in activated CD8� and CD4� T cells; while, among thefive miRNAs that are up-regulated in iNKT cells, miR-21 wasfound also preferentially expressed in the peripheral T cells acti-vated in vitro (14, 27). miR-21 was shown to inhibit apoptosis intumor cells (28) and to down-regulate the mRNA encoding theproapoptotic PTEN (phosphatase and tensin homolog deleted onchromosome TEN) molecule (29). The fact that Dicer deletionfrom DP thymocytes results in a selective cell death in immatureiNKT cells could be therefore compatible with the lack ofmiRNA-21 expression in these cells. Nevertheless, it is likely thatalso other miRNAs take part in the coordinate regulation of themolecular pathways involved in iNKT cell development.

Cobb et al. (14) have also shown that Treg cells express a char-acteristic set of miRNAs that is distinct from that of naive CD4�

T cells, and more similar to that acquired by effector CD4� T cellsupon activation in vitro. Accordingly, iNKT and Treg cells seem toshare part of their miRNA profile: miR-21 is preferentially ex-pressed in iNKT and Treg cells compared with naive T cells, whilemiR-30c, miR-106a, miR-106b, miR-150, and Let-7 family aredown-regulated in both iNKT and Treg cells. The “effector-like”miRNA profiles exhibited by iNKT cells and Tregs would be con-sistent with the evidence suggesting that both iNKT and Treg cellsundergo an agonist selection process in the thymus, resulting in theearly activation and acquisition of the effector/memory phenotype(1, 30, 31). The miRNAs shared by iNKT, Treg, and activated Tcells might be implicated in the regulation of common molecularpathway(s) that lead to the acquisition of the effector phenotype.

The quantitative variation of miRNAs that are differentially ex-pressed between iNKT cells and T cells should result in the mod-ulation of expression levels of their target mRNAs. As suggestedby the data showing a dose-dependent regulation of c-Myb expres-sion by graded concentration of miR-150 (10), this would have amajor impact on the regulation of protein synthesis between thetwo T cell subsets, resulting in a lineage-specific expression levelof particular proteins in iNKT and in T cells. A finely regulated

FIGURE 5. iNKT cells display adistinct miRNA profile. Total RNA wasextracted from sorted HSAlow enrichedthymic iNKT (TCR��CD1dtet�) andmainstream T (HSAlowTCR�high) cells.The miRNA expression in iNKT and Tcells was normalized against a referenceconsisting of miRNAs from total thy-mocytes. Shown are the miRNAs ex-pressed at statistically different levelsbetween iNKT and T cells (p � 0.05).A, Heat map: overexpression (red), un-derexpression (blue), and identical ex-pression (white) of miRNA in iNKT orT cells relative to the reference. ThemiRNA clustering (left) and sampleclustering trees (top) are shown. B,Fold change expression of miRNAs iniNKT cells compared with mature thy-mic T cells, validated by qRT-PCR.

2511The Journal of Immunology

level of expression would be critical for proteins that function overa narrow range of concentration, such as PTEN and Bim (10),which control lymphocyte proliferation and survival and are tar-geted by miR-21 and miR-17, differentially expressed in iNKT andT cells.

Nevertheless, we cannot discount the possibility that miRNAsthat are equally expressed by both iNKT and T cells may also playa lineage-specific role in the development of the two subsets, be-cause the target mRNAs may code for proteins involved in path-ways critical for the program of one but not the other lineage, assuggested for other biological processes (10).

Harnessing miRNA function (32) and identification of the tran-scripts that are targets of iNKT cell-expressed miRNAs might pro-vide further insights into the molecular pathways that specificallycontrol the development of this subset, and also shed light on mo-lecular cues common to Treg and effector T cells.

In conclusion, our results suggest that differentiation and ho-meostasis of iNKT cells depend critically on Dicer-dependentmiRNAs, consistent with the idea that miRNAs control the ca-nalization of cell differentiation by buffering stress-related vari-ations in the expression of genes specifically involved in theprograms (33).

AcknowledgmentsWe thank Drs. Massimiliano Pagani, Riccardo Rossi, Grazisa Rossetti (Mi-lano), Aurore Saudemont, Klaus Okkenhaug, and Elena Vigorito (Babra-ham), as well as the Babraham Small Animal Facility, for help anddiscussions.

DisclosuresThe authors have no financial conflict of interest.

References1. Bendelac, A., P. B. Savage, and L. Teyton. 2007. The biology of NKT cells.

Annu. Rev. Immunol. 25: 297–336.2. Godfrey, D. I., and S. P. Berzins. 2007. Control points in NKT-cell development.

Nat. Rev. Immunol. 7: 505–518.3. Benlagha, K., D. G. Wei, J. Veiga, L. Teyton, and A. Bendelac. 2005. Charac-

terization of the early stages of thymic NKT cell development. J. Exp. Med. 202:485–492.

4. Benlagha, K., T. Kyin, A. Beavis, L. Teyton, and A. Bendelac. 2002. A thymicprecursor to the NK T cell lineage. Science 296: 553–555.

5. Pellicci, D. G., K. J. Hammond, A. P. Uldrich, A. G. Baxter, M. J. Smyth, andD. I. Godfrey. 2002. A natural killer T (NKT) cell developmental pathway in-volving a thymus-dependent NK1.1�CD4� CD1d-dependent precursor stage.J. Exp. Med. 195: 835–844.

6. Kovalovsky, D., O. U. Uche, S. Eladad, R. M. Hobbs, W. Yi, E. Alonzo, K. Chua,M. Eidson, H. J. Kim, J. S. Im, et al. 2008. The BTB-zinc finger transcriptionalregulator PLZF controls the development of invariant natural killer T cell effectorfunctions. Nat. Immunol. 9: 1055–1064.

7. Savage, A. K., M. G. Constantinides, J. Han, D. Picard, E. Martin, B. Li,O. Lantz, and A. Bendelac. 2008. The transcription factor PLZF directs the ef-fector program of the NKT cell lineage. Immunity 29: 391–403.

8. Lazarevic, V., A. J. Zullo, M. N. Schweitzer, T. L. Staton, E. M. Gallo,G. R. Crabtree, and L. H. Glimcher. 2009. The gene encoding early growthresponse 2, a target of the transcription factor NFAT, is required for the devel-opment and maturation of natural killer T cells. Nat. Immunol. 10: 306–313.

9. Bartel, D. P. 2004. MicroRNAs: genomics, biogenesis, mechanism, and function.Cell 116: 281–297.

10. Xiao, C., and K. Rajewsky. 2009. MicroRNA control in the immune system:basic principles. Cell 136: 26–36.

11. Koralov, S. B., S. A. Muljo, G. R. Galler, A. Krek, T. Chakraborty,C. Kanellopoulou, K. Jensen, B. S. Cobb, M. Merkenschlager, N. Rajewsky, andK. Rajewsky. 2008. Dicer ablation affects antibody diversity and cell survival inthe B lymphocyte lineage. Cell 132: 860–874.

12. Muljo, S. A., K. M. Ansel, C. Kanellopoulou, D. M. Livingston, A. Rao, andK. Rajewsky. 2005. Aberrant T cell differentiation in the absence of Dicer. J. Exp.Med. 202: 261–269.

13. Cobb, B. S., T. B. Nesterova, E. Thompson, A. Hertweck, E. O’Connor,J. Godwin, C. B. Wilson, N. Brockdorff, A. G. Fisher, S. T. Smale, andM. Merkenschlager. 2005. T cell lineage choice and differentiation in the absenceof the RNase III enzyme Dicer. J. Exp. Med. 201: 1367–1373.

14. Cobb, B. S., A. Hertweck, J. Smith, E. O’Connor, D. Graf, T. Cook, S. T. Smale,S. Sakaguchi, F. J. Livesey, A. G. Fisher, and M. Merkenschlager. 2006. A rolefor Dicer in immune regulation. J. Exp. Med. 203: 2519–2527.

15. Zhou, X., L. T. Jeker, B. T. Fife, S. Zhu, M. S. Anderson, M. T. McManus, andJ. A. Bluestone. 2008. Selective miRNA disruption in T reg cells leads to un-controlled autoimmunity. J. Exp. Med. 205: 1983–1991.

16. Liston, A., L. F. Lu, D. O’Carroll, A. Tarakhovsky, and A. Y. Rudensky. 2008.Dicer-dependent microRNA pathway safeguards regulatory T cell function.J. Exp. Med. 205: 1993–2004.

17. Chong, M. M., J. P. Rasmussen, A. Y. Rudensky, and D. R. Littman. 2008. TheRNaseIII enzyme Drosha is critical in T cells for preventing lethal inflammatorydisease. J. Exp. Med. 205: 2005–2017.

18. de Boer, J., A. Williams, G. Skavdis, N. Harker, M. Coles, M. Tolaini, T. Norton,K. Williams, K. Roderick, A. J. Potocnik, and D. Kioussis. 2003. Transgenicmice with hematopoietic and lymphoid specific expression of Cre. Eur. J. Im-munol. 33: 314–325.

19. Schumann, J., P. Pittoni, E. Tonti, H. R. Macdonald, P. Dellabona, andG. Casorati. 2005. Targeted expression of human CD1d in transgenic mice re-veals independent roles for thymocytes and thymic APCs in positive and negativeselection of V�14i NKT cells. J. Immunol. 175: 7303–7310.

20. Tonti, E., G. Galli, C. Malzone, S. Abrignani, G. Casorati, and P. Dellabona.2009. NKT-cell help to B lymphocytes can occur independently of cognate in-teraction. Blood 113: 370–376.

21. Lee, P. P., D. R. Fitzpatrick, C. Beard, H. K. Jessup, S. Lehar, K. W. Makar,M. Perez-Melgosa, M. T. Sweetser, M. S. Schlissel, S. Nguyen, et al. 2001. Acritical role for Dnmt1 and DNA methylation in T cell development, function,and survival. Immunity 15: 763–774.

22. Coquet, J. M., S. Chakravarti, K. Kyparissoudis, F. W. McNab, L. A. Pitt,B. S. McKenzie, S. P. Berzins, M. J. Smyth, and D. I. Godfrey. 2008. Diversecytokine production by NKT cell subsets and identification of an IL-17-producingCD4�NK1.1� NKT cell population. Proc. Natl. Acad. Sci. USA 105:11287–11292.

23. Griewank, K., C. Borowski, S. Rietdijk, N. Wang, A. Julien, D. G. Wei,A. A. Mamchak, C. Terhorst, and A. Bendelac. 2007. Homotypic interactionsmediated by Slamf1 and Slamf6 receptors control NKT cell lineage development.Immunity 27: 751–762.

24. Fukagawa, T., M. Nogami, M. Yoshikawa, M. Ikeno, T. Okazaki, Y. Takami,T. Nakayama, and M. Oshimura. 2004. Dicer is essential for formation of theheterochromatin structure in vertebrate cells. Nat. Cell Biol. 6: 784–791.

25. Harfe, B. D., M. T. McManus, J. H. Mansfield, E. Hornstein, and C. J. Tabin.2005. The RNaseIII enzyme Dicer is required for morphogenesis but not pat-terning of the vertebrate limb. Proc. Natl. Acad. Sci. USA 102: 10898–10903.

26. Xu, P., M. Guo, and B. A. Hay. 2004. MicroRNAs and the regulation of celldeath. Trends Genet. 20: 617–624.

27. Wu, H., J. R. Neilson, P. Kumar, M. Manocha, P. Shankar, P. A. Sharp, andN. Manjunath. 2007. miRNA profiling of naive, effector and memory CD8 Tcells. PLoS ONE 2: e1020.

28. Chan, J. A., A. M. Krichevsky, and K. S. Kosik. 2005. MicroRNA-21 is anantiapoptotic factor in human glioblastoma cells. Cancer Res. 65: 6029–6033.

29. Meng, F., R. Henson, H. Wehbe-Janek, K. Ghoshal, S. T. Jacob, and T. Patel.2007. MicroRNA-21 regulates expression of the PTEN tumor suppressor gene inhuman hepatocellular cancer. Gastroenterology 133: 647–658.

30. Jordan, M. S., A. Boesteanu, A. J. Reed, A. L. Petrone, A. E. Holenbeck,M. A. Lerman, A. Naji, and A. J. Caton. 2001. Thymic selection of CD4�CD25�

regulatory T cells induced by an agonist self-peptide. Nat. Immunol. 2: 301–306.31. Apostolou, I., A. Sarukhan, L. Klein, and H. von Boehmer. 2002. Origin of

regulatory T cells with known specificity for antigen. Nat. Immunol. 3: 756–763.32. Krutzfeldt, J., N. Rajewsky, R. Braich, K. G. Rajeev, T. Tuschl, M. Manoharan,

and M. Stoffel. 2005. Silencing of microRNAs in vivo with “antagomirs.” Nature438: 685–689.

33. Hornstein, E., and N. Shomron. 2006. Canalization of development by microR-NAs. Nat. Genet. 38 Suppl: S20–S24.

2512 miRNAs CONTROL iNKT CELLS