of April 13, 2018.This information is current as

Expression of Rictor and mTORRegulatory T Cells by Regulating the MicroRNA-15b/16 Enhances the Induction of

S. CobbYogesh Singh, Oliver A. Garden, Florian Lang and Bradley

http://www.jimmunol.org/content/195/12/5667doi: 10.4049/jimmunol.1401875November 2015;

2015; 195:5667-5677; Prepublished online 4J Immunol 

MaterialSupplementary

5.DCSupplementalhttp://www.jimmunol.org/content/suppl/2015/11/04/jimmunol.140187

Referenceshttp://www.jimmunol.org/content/195/12/5667.full#ref-list-1

, 20 of which you can access for free at: cites 47 articlesThis article

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists, Inc. All rights reserved.Copyright © 2015 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on April 13, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

by guest on April 13, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

The Journal of Immunology

MicroRNA-15b/16 Enhances the Induction of RegulatoryT Cells by Regulating the Expression of Rictor and mTOR

Yogesh Singh,*,† Oliver A. Garden,‡ Florian Lang,† and Bradley S. Cobb*

CD4+ regulatory T cells (Tregs) are essential for controlling immune responses and preventing autoimmunity. Their development

requires regulation of gene expression by microRNAs (miRNAs). To understand miRNA function in Treg development, we

searched for important miRNAs and their relevant target genes. Of the more abundantly expressed miRNAs in Tregs, only

miR-15b/16, miR-24, and miR-29a impacted the production of in vitro–induced Tregs (iTregs) in overexpression and blocking

experiments. miRNA mimics for these significantly enhanced the induction of iTregs in Dicer2/2 CD4+ T cells. Furthermore, the

overexpression of miR-15b/16 in conventional CD4+ T cells adoptively transferred into Rag22/2 mice increased the in vivo

development of peripheral Tregs and diminished the severity of autoimmune colitis. In searching for targets of miR-15b/16, we

observed that the mammalian target of rapamycin (mTOR) signaling pathway was enhanced in Dicer2/2 CD4+ T cells, and its

pharmacological inhibition restored induction of iTregs. Suppression of mTOR signaling is essential for induction of iTregs from

naive CD4+ T cells, and the mTORC2 component, Rictor, contained a functional target site for miR-15b/16. Rictor was more

abundantly expressed in Dicer2/2 T cells as was mTOR, and their expression was downregulated by the overexpression of miR-

15b/16. This led to a reduction in mTOR signaling, as measured by phosphorylation of the downstream target, ribosomal protein

S6. Finally, knockdown of Rictor by small interfering RNAs enhanced Treg induction in Dicer2/2 CD4+ T cells. Therefore, an

important mechanism of miRNA regulation of Treg development is through regulation of the mTOR signaling pathway. The

Journal of Immunology, 2015, 195: 5667–5677.

Regulation of the immune response and the prevention ofautoimmunity require a class of Th cells called regulatoryT cells (Tregs). Because of their essential function in

maintaining peripheral tolerance, they play important roles inmany immune-related disorders. Therefore, a significant interestexists in understanding their development and function to developnew therapeutic strategies for diseases involving the immuneresponse (1).The majority of circulating Tregs are produced in the thymus at

the double-positive stage and are designated thymus-derived Tregs(tTregs). Current thought is that they enter the Treg lineage be-cause of an intermediate affinity of their TCR to self-antigens,which is above that normally required for positive selection butnot sufficient for negative selection. This is hypothesized to inducea suppressor phenotype that protects the body against any sub-sequent immune response that might develop against self-antigens

(1). Tregs can also develop in the periphery from naive T CD4+

T cells when activated under the influence of specific factors(e.g., TGF-b and retinoic acid) produced by APCs and other celltypes present at the site of Ag challenge. These are designatedperipherally-derived Tregs (pTregs). Both tTregs and pTregs areimportant in preventing autoimmune disease (2, 3). Each is de-fined by the expression of the Treg-specific transcription factorFoxp3, which plays an essential role in regulating the gene ex-pression profile required for Treg development and function. Lossof Foxp3 results in the absence of Tregs and subsequent lethalautoimmunity (4).MicroRNAs (miRNAs) are critical regulators of Treg devel-

opment and function. These are small double-stranded RNAs ∼22nt in length that negatively regulate gene expression at a post-transcriptional stage (5). miRNAs are encoded in the genomewithin pol II transcripts. Their critical feature is a stem loopstructure made up of the complementary strands of the miRNA,which is recognized and cleaved by protein complexes containingthe RNases Drosha and Dicer to give the mature miRNA. This isincorporated into an effector complex called the RNA-inducedsilencing complex. One strand is degraded, and the other cantarget the complex to messages through imperfect base pairing,which typically occurs in the 39-untranslated region (UTR) of thegene. This results in inhibition of translation and the subsequentdestabilization of the message (6). The loss of miRNAs at thedouble-positive stage of T cell development through a conditionalknockout of Drosha or Dicer significantly inhibits the develop-ment of Tregs and ultimately leads to autoimmunity (7, 8). Inaddition, a deletion of either RNase specifically in Tregs disruptstheir suppressor function and results in an acute autoimmuneresponse that is as severe as the loss of Foxp3 (8–10). Therefore,miRNAs are important in the steps leading up to commitmenttoward the Treg lineage and also downstream in mediating theirsuppressor function. Several studies examined the roles of miR-NAs in Treg function and identified individual miRNAs that

*Department of Comparative Biomedical Sciences, The Royal Veterinary College,London NW1 0TU, United Kingdom; †Institute of Physiology I, Eberhard KarlsUniversity of T€ubingen, D-72076 T€ubingen, Germany; and ‡Department of ClinicalSciences and Services, The Royal Veterinary College, London NW1 0TU, UnitedKingdom

ORCIDs: 0000-0001-6919-6273 (Y.S.); 0000-0001-6783-0509 (B.S.C.).

Received for publication July 25, 2014. Accepted for publication October 9, 2015.

This work was supported by Biotechnology and Biological Sciences Research Coun-cil Grant BB/H018573/1 and a BD Biosciences grant.

Address correspondence and reprint requests to Dr. Bradley S. Cobb, Department ofComparative Biomedical Sciences, The Royal Veterinary College, Royal CollegeStreet, London NW1 0TU, U.K. E-mail address: bcobb@rvc.ac.uk

The online version of this article contains supplemental material.

Abbreviations used in this article: iTreg, in vitro–induced Treg; miRNA, microRNA;mTOR, mammalian target of rapamycin; mTORC, mTOR complex; PTEN, phos-phatase and tensin homolog; pTreg, peripherally-derived Treg; qPCR, quantitativereal-time PCR; qRT-PCR, quantitative real time RT-PCR; SGK, serum glucocorticoidkinase; siRNA, small interfering RNA; Treg, regulatory T cell; tTreg, thymus-derivedTreg; UTR, untranslated region.

Copyright� 2015 by The American Association of Immunologists, Inc. 0022-1767/15/$25.00

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1401875

by guest on April 13, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

regulate key genes required for the function and stability of Tregs(11–17). However, it is still unknown how miRNAs regulate theexpression of genes involved in the developmental steps leadingto the Treg lineage.We set about to examine miRNA function in Treg development

by first determining functional miRNAs and then identifyingcritical genes that they regulate. We found three miRNAs (miR-15b/16, miR-24, and miR-29a) that regulated the induction ofTregs from naive CD4+ T cells, with miR-15b/16 having thegreatest effect in overexpression and blocking experiments. Im-portant genes regulated by miR-15b/16 were Rictor and Mtor,which encode components of the mammalian target of rapamycin(mTOR) signaling pathway. Downregulation of the mTOR sig-naling pathway is important for the induction of Tregs (18–20).Therefore, miR-15b/16 plays an important role in regulating themTOR signaling pathway and controlling the development ofTregs.

Materials and MethodsMouse strains, Abs, and plasmids

C57BL/6 (Charles River, Harlow, U.K.), Rag22/2 (kindly donated byDr. Jian-Guo Chai, Imperial College London, London, U.K.), and CD4-CreDicerlox/lox (7) mice were used for the experiments and kept in a con-ventional specific pathogen–free facility. All animal work was performedaccording to the Animals (Scientific Procedures) Act 1996, UK underanimal Project License 70/6965.

Abs for flow cytometry experiments were all directed against mouseAgs. They included CD4-FITC/PE/PerCP/allophycocyanin (clone GK1.5),CD8a-PerCP (clone 53-6.7), CD25-PE (clone PC61.5), Foxp3-allophycocyanin(clone FJK-16s), IFN-g–FITC or IFN-g–eFluor 450 (clone XMG1.2),IL-17–PE (all from eBioscience), and IL-17a–PE (clone TC11-18H10)(BD Biosciences). Western blot Abs (all rabbit mAbs) were p-SMAD2(S465/467), total SMAD2/3, phosphatase and tensin homolog (PTEN;clone D4.3) XP, PDK-1, p–PDK-1 (S241), p-AKT (S473) (clone D9E) XP,p-AKT (T308) (clone C31E5E), total AKT (pan) (clone C67E7), mTOR(clone 7C10), Rictor (clone 53A2), p-FoxO1 (S256), total FoxO1a (cloneC29H4), p-FoxO3a (S253), total FoxO3a (clone 75D8), p-S6 ribosomalprotein (Ser235/236) (clone D57.2.2E) XP, GAPDH (clone D16H11) XP, andb-actin (clone 8H10D10) (all from Cell Signaling).

miRNA expression vectors were derived by cloning the genomicregion encoding specific miRNAs into the retroviral pMIG vector. Ex-pression of miRNAs was confirmed by miRNA–quantitative real timeRT-PCR (qRT-PCR). miRNA decoy or sponge transcripts were designedas described (21, 22). Decoy sequences were cloned into the pSIFlentiviral vector for PolIII expression. Sponge sequences were clonedinto pMIG. miRNA reporter constructs contained the 39UTR of Rictorinserted downstream of luciferase in pGL3. Details about the vectors areavailable upon request.

Isolation of naive T cells and activation into Th subsets

Naive CD4+ CD62Lhigh CD252 T cells were isolated from spleen andlymph nodes (inguinal, axillary, brachial, mediastinal, superficial cervical,mesenteric) of 6–8-wk-old female C57BL/6 mice using magnetic beads forCD4+ selection (Dynabeads Untouched Mouse CD4 Cells Kit; Invitrogen),followed by CD25 and CD62L selection using biotinylated anti-CD25(7D4 clone) or biotinylated anti-CD62L (clone MEL-14; both from BDBiosciences) and Streptavidin MicroBeads with MACS separation col-umns (Miltenyi Biotec).

Naive CD4+ CD62Lhigh CD252 T cells were cultured in completemedium (RPMI 1640, penicillin/streptomycin, L-glutamine, HEPES, and2-ME; all from Sigma) and 10% FBS. They were activated in the pres-ence of plate-bound anti-CD3/anti-CD28 Abs (eBioscience) at a 1:2 ratio(1 mg/ml anti-CD3/2 mg/ml anti-CD28) for Th1, Th2, and in vitro–induced Tregs (iTregs) and a 1:10 ratio (1 mg/ml anti-CD3/10 mg/mlanti-CD28) for Th17. Activated T cells were differentiated into Th1using 20 ng/ml rIL-12 (eBioscience) and anti–IL-4 (5 mg/ml; BD Bio-sciences); into Th2 using 20 ng/ml rIL-4 (BD Biosciences) and anti–IFN-g (5 mg/ml; BD Biosciences); into Th17 using 2.5 ng/ml rTGF-b,50 ng/ml rIL-6 (eBioscience), anti–IFN-g (5 mg/ml), anti–IL-4 (5 mg/ml),and anti–IL-2 (5 mg/ml; BD Biosciences); and into iTregs using 2.5 ng/mlrTGF-b and 5 ng/ml rIL-2 (eBioscience). Analysis was performed on cellsafter 3–4 d of culture.

Retroviral transduction and transfection of miRNA mimics andsmall interfering RNAs into T cells

Retroviruses were produced by calcium phosphate transfection of HEK293T cells with the retroviral expression and the pcL-Eco helper virusvectors. Culture supernatants were harvested and used to spin-infect naiveCD4+ T cells that were activated overnight by plate-bound anti-CD3/anti-CD28 Abs. Cells were differentiated into iTregs as described above. Tocompare the effects of different miRNAs, an average Z score was calcu-lated from individual experiments, as described (23): Z = (x – y)/SD. Inthis calculation, x = iTreg index, which is the percentage of Foxp3+ cellsinduced in the control or with the expression of an individual miRNA; y =the mean of all the iTreg indices in a given experiment; and SD = the SD ofall these iTreg indices.

miRNA mimics or small interfering RNAs (siRNAs) were transfectedinto naive CD4+ T cells using DharmaFECT 3 miRNA mimic transfectionreagent or siRNA Accell delivery medium, respectively, per the manu-facturer’s guidelines (Dharmacon). For the miRNA mimic transfections,the negative control oligonucleotide, consisting of a random sequence,contained a DY547 fluorescent tag for determining transfection efficiency,which was ∼50% in each experiment.

Immunoassays

Immunostaining for flow cytometric analysis was performed using therespective Abs. For intracellular staining of cytokines, cells were treatedwith 1 mg/ml PMA, 1 mg/ml ionomycin, and 1 mg/ml brefeldin A for 2 hprior to staining. For Foxp3 staining, cells were fixed with Foxp3Fixation/Permeabilization Buffer (eBioscience) for 30 min prior tostaining; if GFP was also detected, cells were fixed with 2% para-formaldehyde for 5 min prior to fixation/permeabilization. Data wereacquired using a FACSCanto II (BD) and analyzed using FlowJo software(TreeStar).

Cytokine in culture supernatants from differentiated T cells was de-tected using the FlowCytomix multiple analysis detection kit for Th1/Th2cytokines (eBioscience).

Western blot analysis was performed from total cell extracts preparedfrom naive CD4+ T cells that were either left unactivated or activated for30 min with plate-bound anti-CD3/anti-CD28 (1:2), with or without theaddition of 2.5 ng/ml TGF-b. Proteins were detected using the indicatedAbs and a HRP-conjugated secondary Ab and then visualized using theECL detection system. Quantitation was performed by densitometric anal-ysis of exposed films.

miRNA and mRNA qRT-PCR

RNA was isolated from CD4+ T cells using miRNeasy (for miRNAs) ormRNeasy (for mRNAs) isolation kits (QIAGEN). For miRNA detection,cDNA synthesis and subsequent quantitative real-time PCR (qPCR) withlocked nucleic acid primers for specific miRNAs was performed usingmiRCURY LNA Universal RT microRNA cDNA synthesis and qPCR kit(EXIQON). For mRNA detection, cDNA was prepared using the miScriptcDNA synthesis kit (QIAGEN). qPCR was performed using IQ Bio-RadSYBR Green Master Mix (Bio-Rad) with the following primers for Rictor(forward: 59-ACCGGGCTTCTGACCATTAAA-39 and reverse: 59-TTGT-ATGAACCGCCGACACT-39). Measurements were performed using Chromo4 Bio-Rad PCR machines. The relative expression level of miRNAs wasnormalized to that of 5S rRNA and U6 small nuclear RNA, whereas therelative expression level of mRNAs was normalized to that of b-actin, asdescribed previously (24).

Adoptive T cell–transfer colitis model

C57BL/6 Rag22/2 mice were injected into the peritoneal cavity with106 GFP FACS-sorted naive CD4+ CD62Lhigh CD252 T cells that weretransduced with either control or miR-15b/16–expressing retroviruses.Weight was monitored every week, and mice were sacrificed after 8 wk(or when weight was reduced by 20%) for analysis. Histopathologicalreview was performed as previously described (12, 25). The large in-testine (from the ileocecocolic junction to anorectal junction) was re-moved and processed. Sections were stained with H&E, coded, andassigned a score in a blinded fashion. The severity of colitis was gradedsemiquantitatively on an overall scale of 0 to 5, assessing the degree ofepithelial hyperplasia and goblet cell depletion, leukocyte infiltration inthe lamina propria, area of tissue affected, and the presence of markers ofsevere inflammation, such as crypt abscesses, submucosal inflammation,and ulcers. The total colonic score was calculated as the median of theindividual scores of sections of proximal colon, midcolon, and distalcolon.

5668 MicroRNA REGULATION OF Treg DEVELOPMENT

by guest on April 13, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

39UTR luciferase reporter assay

The 39UTR of Rictor was cloned downstream of the firefly luciferase genein the pGL-3 vector. HEK 293 T cells were transfected with this vector (orthe mutated miR-15/16 target site vector), internal control pRL Renillaluciferase vector, and the miRNA overexpression vector. After 48 h, lu-ciferase activity was measured using the Dual-Luciferase reporter assaysystem (Promega). Luciferase activity was measured with a Wallac 1420VICTOR2 multilabel plate reader (Perkin Elmer).

Statistical analysis

Statistical analyses, using the Student t test, were performed with Prismsoftware (GraphPad). Figures were created using Excel and GraphPadPrism software. The p values #0.05 were considered significant.

ResultsIdentification of important miRNAs

To analyze miRNA function in Treg development, a model systemwas needed in which individual miRNAs could be manipulatedand tested. Therefore, we used an in vitro model of Treg inductioninvolving the activation of naive CD4+ T cells in the presence ofTGF-b plus IL-2 and measured the differentiation into the Treglineage by the expression of Foxp3. This model represents thedevelopment of pTregs, which, along with tTregs, are critical insuppressing the immune response and preventing autoimmunedisease (2, 3). Furthermore, the induction of iTregs is significantlyreduced in Dicer2/2 CD4+ T cells just like the in vivo develop-ment of Tregs in the mutant mice (7). For this model, we hy-pothesized that important miRNAs would be abundant and morehighly expressed in iTregs compared with conventionally acti-vated CD4+ T cells or those polarized to other Th subsets. Our

previous miRNA-profiling experiments (7) identified multiplemiRNAs that are more highly expressed in ex vivo–derivedTregs compared with activated conventional CD4+ T cells. ThesemiRNAs were felt to be the most likely candidates to be involvedin the induction of iTregs, so their relative level was determinedby qPCR in iTregs and compared with other Th subsets. Theseincluded resting naive CD4+ T cells and those activated undernonpolarizing or polarizing conditions toward Th1, Th2, or Th17subsets. miRNAs that were more abundantly expressed in ex vivoTregs (miR-15b, 16, 21, 24, 29a, 92b, 142 59, 142 39 146a, 150,and 223) were also the most abundant in iTregs compared withT cells cultured in other conditions; conversely, those that werenot abundantly expressed in ex vivo Tregs (miR-23a, 30c, 99b,125a, 191, and 326) were not abundantly expressed in iTregs

(Fig. 1A). Therefore, we focused on the function of the miRNAsthat were most abundantly expressed in iTregs.The function of individual miRNAs was examined by measuring

the effect of their overexpression or blocking in iTreg induction.Overexpression was achieved through the use of retroviruses thatexpressed genomic sequences encoding the miRNAs. Blockingwas achieved through the use of lentiviruses or retroviruses thatexpressed miRNA decoy or “sponge” target sequences, respec-

tively. These viruses were transduced into CD4+ T cells, andiTregs were induced by the addition of TGF-b plus IL-2. Becausethe viruses also encoded GFP to track their expression, GFP+ cellswere analyzed for the induction of iTregs by the expression ofFoxp3. The miRNAs that had the greatest effect in these experi-ments were miR-15b and miR-16, which are encoded within thesame primary transcript and are closely related, such that theytarget the same sequences in mRNAs. Overexpression of miR-15b/16 significantly increased the induction of iTregs comparedwith cells transduced with a control retrovirus lacking any miRNAsequences (Fig. 1B). Likewise, expression of decoys for eithermiR-15b or miR-16 inhibited the induction of iTregs (Fig. 1C).

These effects were dependent on the level of expression of themiRNAs or decoys, as deduced from GFP expression. Cell pop-ulations within gates of increasing GFP expression were moreaffected in iTreg induction (Fig. 1D).Comparing the effects of all of the miRNAs tested using a

Z-score analysis (as described in Materials and Methods) revealedthat miR-24 and miR-29a were the only other miRNAs to have asignificant positive effect on the induction of iTregs (Fig. 1E). Allof these positive-acting miRNAs could also partially reverse thedefect in iTreg induction in Dicer2/2 CD4+ T cells. Transfectionof miRNA mimics for miR-15b, miR-24, or miR-29a enhancedthe induction of iTregs in Dicer2/2 CD4+ T cells, whereas miR-21(which did not have an effect in overexpression and blockingexperiments) failed to do so. In addition, transfection of a com-bination of miR-15b, miR-16, and miR-29a into Dicer2/2 CD4+

T cells brought the level of iTreg induction nearly back to thatseen in Dicer+/2 CD4+ T cells (Fig. 1F). Therefore, these miRNAsappear to be most important in regulating the induction of iTregs.It is interesting to note that, in contrast to the importance of miR-15b/16 in the overexpression and blocking studies, miR-29a wasmost effective in the complementation studies. However, thebiological significance of this observation remained unclear.The difference could have been due to the different experimentalsystems or to differences in expression of the miRNAs betweentransfection and retroviral transduction.To test whether overexpression of miR-15b/16 could also

enhance the development of pTregs in vivo, Rag22/2 mice werereconstituted with conventional Th cells (CD4+ CD252) trans-duced with control or miR-15b/16–expressing retroviruses. Theautoimmune reaction observed with this reconstitution was sig-nificantly inhibited by the overexpression of miR-15b/16, asmeasured by weight loss and histopathological review of thecolon (Fig. 2A–C). In addition, the total numbers of mononu-clear cells populating the spleen were significantly reduced whenmiR-15b/16 was overexpressed, as was the number of CD4+

T cells (Fig. 2D). Consistent with the reduction in the totalnumber of T cells, there were fewer IFN-g–producing cells in thespleen when miR-15b/16 was overexpressed. Likewise, thereappeared to be fewer IL-17–producing cells, but this differencedid not reach statistical significance. Interestingly, the percent-ages of IFN-g and IL-17–producing cells were not statisticallydifferent, so the overexpression of miR-15b/16 did not appear tospecifically impact the development of Th1 and Th17 cells overother Th subsets. Therefore, the reduction in the inflammatoryresponse was most likely due to the overall reduction in T cellnumbers. This could have been due to multiple effects ofmiR-15b/16 overexpression, but most importantly, there was asignificant increase in the percentage of pTregs at the sites ofinflammation (spleen, lungs, and mesenteric lymph nodes) but notaway from the site of inflammation within peripheral lymph nodes(Fig. 2F). Therefore, similar to the in vitro experiments, theoverexpression of miR-15b/16 in CD4+ T cells enhanced the pe-ripheral development of Tregs, and this was consistent with thereduced numbers of T cells and the attenuation of the autoimmuneresponse. Alternative explanations for the decrease in autoim-munity might include suppression of T cell proliferation or acti-vation of cell death because miR-15/16 can inhibit cell growth(26) and induce apoptosis (27). However, overexpression of miR-15b/16 did not affect T cell proliferation or apparent viabilityin vitro (Supplemental Fig. 1A), so it is uncertain whether theseissues were important in vivo. Additionally, miR-15b/16 could haveimpacted the suppressive function of Tregs, but our attempts toexamine the suppressive function of Tregs was hampered by the factthat ex vivo–isolated Tregs transduced to overexpress miR-15b/16

The Journal of Immunology 5669

by guest on April 13, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

maintained Foxp3 expression much more stably when culturedin vitro (Supplemental Fig. 1B). Therefore, any effects on sup-

pressive function would be difficult to interpret. Finally, to deter-

mine whether the effect of miR-15b/16 on pTreg development was

cell intrinsic or due to extrinsic factors created in the immune en-

vironment where T cells overexpressed miR-15b/16, unsorted miR-

15b/16–transduced cells were adoptively transferred into Rag22/2

mice, and the percentage of pTregs was determined in both the

GFP+ and GFP2 populations of CD4+ T cells. Cells overexpressing

miR-15b/16 had significantly more pTregs in mesenteric lymph

nodes and spleen (Supplemental Fig. 2), suggesting that the effect

of miR-15b/16 was intrinsic to the T cells overexpressing these

miRNAs.

Identification of relevant genes targeted by miR-15b/16

Because miR-15b/16 had the greatest effect on iTreg induction inoverexpression and blocking studies, we focused on finding rele-vant target genes for this miRNA family. As a first step, targetprediction algorithms (28, 29) identified nearly a thousand geneswith potential target sites for miR-15b/16. To narrow down thislist, we decided to examine signal transduction pathways known tobe important for Treg development and determine whether anywere altered in Dicer2/2 CD4+ T cells. Candidate target geneswithin affected pathways could then be focused upon for analysis.

miRNA regulation of IFN-g is not important for iTreg induction.IFN-g inhibits the induction of iTregs (30, 31), and Dicer2/2 CD4+

T cells inherently produce IFN-g when activated by costimulation

FIGURE 1. miR-15b/16, miR-24, and miR-29 are important for the in vitro induction of Tregs. (A) Relative expression levels of miRNAs (as determined

by qPCR) in naive Th cells (CD4+ CD62L+ CD252 CD442) from lymph nodes and spleen of C57BL/6 mice that were left unactivated or activated under

the indicated polarizing conditions. (B) The effect of overexpression of miR-15b/16 on the induction of iTregs. Naive CD4+ T cells were transduced with

GFP-expressing retroviruses that did not express any miRNA (control) or expressed miR-15b/16. iTreg induction was measured in GFP+ cells by the

expression of Foxp3. Representative experiment (left panels) and the mean + SD of six independent experiments (right panel), which demonstrated a

significant difference between control and miR-15b/16–overexpressing cells. **p = 0.005. (C) The effect of blocking miR-15b/16 on iTreg induction.

Similar to above, naive CD4+ cells were transduced with control or miR-15b or miR-16 decoy-expressing lentiviruses, and iTreg induction was measured by

the expression of Foxp3. Representative experiment (left panels) and mean + SD from six independent experiments showing a significant difference

between control and decoy-expressing cells (miR-15b, *p = 0.045; miR-16, *p = 0.02). (D) The effect of the level of expression of miR-15b/16 or its decoys

on iTreg induction. The level of expression of the miRNAs or their decoys was deduced from GFP expression. iTreg induction (measured by Foxp3

expression) was determined in populations of cells within gates of no, low, medium, or high GFP expression (insets), and statistically significant differences

were found when comparing no-GFP versus high-GFP expression (overexpression, **p = 0.005; decoy expression, ***p = 0.001). (E) Comparison of the

effects of individual miRNAs in overexpression and blocking experiments using a Z-score analysis described in Materials and Methods. Only miR-15b/16,

miR-24, and miR-29a were found to impact iTreg induction. Data are derived from six individual experiments for each miRNA. (F) Restoration of iTreg

induction in Dicer2/2 CD4+ T cells by miR-15b/16, miR-24, and miR-29a. miRNA mimics for the indicated miRNAs were transfected into naive Dicer+/2

or Dicer2/2 CD4+ T cells, and the effect on iTreg induction was determined by Foxp3 expression (miR-all indicates a combination of miR-15b, miR-24,

and miR-29). Representative experiment (left panels) and mean and SD from three independent experiments (right panel), which demonstrated a significant

difference in Dicer2/2 CD4+ T cells between the positive acting miRNAs (but not the negative acting miR-21) and the negative control miRNA mimic

(miR-15b, *p = 0.02; miR-24, *p = 0.05; miR-29, **p = 0.002; all three miRNAs together, ***p = 0.0005).

5670 MicroRNA REGULATION OF Treg DEVELOPMENT

by guest on April 13, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

of the TCR and CD28 (32). Therefore, the importance of miRNAsin iTreg induction could be through their inhibition of IFN-g ex-pression. Previous studies indicated that miR-29 regulates the ex-pression of IFN-g through its regulation of the transcription factors

T-bet and Eomes, which activate the expression of IFN-g (23, 33,34). The gene encoding T-bet (Tbx21) contains a predicted targetsite for miR-15b/16 in its 39UTR, in addition to the site targeted bymiR-29 (29, 35). Therefore, the importance of these miRNAs in

iTreg induction could be through their targeting of Tbx21, therebyregulating IFN-g expression. To test the importance of miRNAs inthe regulation of IFN-g expression, IFN-g production was deter-mined in Dicer+/2 and Dicer2/2 CD4+ T cells that were activated

under different Th-polarizing conditions. Production of IFN-g wassignificantly higher in Dicer2/2 CD4+ T cells polarized toward Th0,Th1, Th2, and Th17 subtypes, but it was virtually absent in iTregsfrom Dicer+/2 or Dicer2/2 CD4+ T cells (Fig. 3A) (note: the pro-

duction of IFN-g in Th2 conditions in Dicer+/2 cells was due to thelack of the addition of neutralizing Ab for IFN-g, which is requiredfor full polarization to the Th2 subset but would prevent the mea-surement of IFN-g in the culture supernatants). Therefore, IFN-g

production does not appear to be responsible for the reduction inthe induction of iTregs in Dicer2/2 CD4+ T cells. Furthermore, theaddition of a neutralizing Ab to the low amount of IFN-g usedin these studies did not significantly enhance the induction ofiTregs in Dicer+/2 or Dicer2/2 CD4+ T cells (Fig. 3B) (in fact, in-duction was slightly inhibited in Dicer2/2 CD4+ T cells). In contrast,the addition of IFN-g inhibited iTreg induction in Dicer+/2 andDicer2/2 CD4+ T cells. Therefore, miRNA regulation of IFN-gexpression does not appear to be important for iTreg induction.

Proximal TGF-b signaling events are not regulated by miRNAs.Because TGF-b signaling during T cell activation results in iTreginduction (36), miRNAs could potentially regulate the expressionof key repressors of this pathway and provide a means to fine-tuneits signaling. Therefore, miRNA regulation of proximal events inthis pathway were analyzed by examining the phosphorylation ofthe downstream transcription factor, SMAD2, which is activatedthrough the TGF-bR complex by phosphorylation on serines 465and 467. As shown in Fig. 4, Dicer+/2 and Dicer2/2 CD4+ T cellshad no difference in phosphorylation of SMAD2 when activatedby anti-CD3/CD28 and the addition of TGF-b. Therefore, miRNAs

FIGURE 2. Overexpression of miR-15b/16 enhances the production of pTregs and decreases the autoimmune response in Rag22/2 reconstitution ex-

periments. Naive CD4+ T cells were transduced with the control or miR-15b/16-expressing retroviruses. GFP+ cells were sorted and adoptively transferred

into Rag22/2 mice, and the effect of miR-15b/16 overexpression was examined in the following assays. (A) Loss of body weight was followed over 55 d

after reconstitution, and it was significantly less. ***p = 0.0003, **p = 0.01, *p = 0.05. (B) Thickness of colons (*p = 0.01) and colitis score (*p = 0.03) at

the end point of analysis were measured using a scale of 1–5, and these were significantly less. (C) Representative histological sections of the large intestine

of mice demonstrated a large decrease in the inflammation. Scale bars, 100 mm. (D) Total numbers of mononuclear cells were significantly reduced (*p =

0.05; left panel), as were CD4+ T cells (GFP+ cells, *p = 0.02; right panel). (E) Analysis of inflammatory IFN-g– and IL-17–producing cells. Representative

experiments (left panels) and bar graphs showing mean + SD (right panels). There was a significant reduction in IFN-g–producing cells (*p = 0.03) but not

in IL-17–producing cells. Interestingly, the percentage of CD4+ T cells producing IFN-g or IL-17 was not significantly different. (F) The development of

pTregs was significantly increased. Representative experiment from spleen (left panels) and bar graphs (mean + SD) showing significant differences in the

percentages of Tregs in the spleen (***p = 0.0001), lungs (**p = 0.007), and mesenteric lymph nodes (mLNs; **p = 0.002) but not away from the site of

inflammation at the peripheral lymph nodes (pLNs) (right panels). Data were derived from six control and five miR-15b/16 mice. Two independent ex-

periments were performed with similar results.

The Journal of Immunology 5671

by guest on April 13, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

are not important for regulating TGF-b signaling up to the point ofSMAD2 phosphorylation. However, it is possible that miRNAsregulate TGF-b signaling downstream of this point. Target pre-diction algorithms found target sites of miR-15b/16 in SMAD7,SMURF1/2, and WWP1, which are negative regulators of TGF-bsignaling, but little is known about the importance of these factorsin TGF-b signaling leading to iTreg induction, so their analysiswas not pursued further.

The mTOR signaling pathway is regulated by miRNAs. Inductionof iTregs requires suppression of signaling through the mTORpathway (18–20). A highly simplified diagram of the pathway isshown in Fig. 5A. Briefly, stimulation of the TCR and coreceptorsleads to the activation of PI3K. These phosphorylate inositol lipidsat the cell membrane (countered PTEN), which then attract andactivate PDK1. PDK1 phosphorylates protein kinase B/AKT (atthreonine 308), leading to its activation, which ultimately resultsin the activation of the mTOR complex (mTORC). mTORC reg-ulates multiple facets of cell metabolism and ultimately controlsmany steps in the determination of developmental fates (37). Onekey mediator is ribosomal protein S6 kinase, which is activatedthrough its phosphorylation by mTOR and then phosphorylatesmany cellular proteins involved in cell growth, one being ribo-

somal protein S6. mTORC contains the mTOR kinase and otherassociated proteins. Two main complexes exist that differ pri-marily by the inclusion of Raptor (mTORC1) or Rictor (mTORC2).

mTORC1 lies downstream and is activated via events mediatedby AKT. However, mTORC2 is required for full activation ofAKT through its phosphorylation of serine 473. In addition toactivating mTORC1, AKT phosphorylates FoxO1a and FoxO3a

proteins, which are important transcription factors required forthe expression of Foxp3 and the development of Tregs (38, 39).Phosphorylation of FoxO proteins by AKT leads to their exclusionfrom the nucleus and subsequent degradation, which is one im-

portant mechanism through which mTOR signaling inhibits Tregdevelopment (40).Pharmacological inhibitors of this pathway enhance the in-

duction of iTregs. These include LY294002, rapamycin, andAKT1/2, which inhibit PI3K, mTORC1, and AKT, respectively(20). To determine whether miRNAs are important for suppressingthe mTOR signaling pathway, the above inhibitors were tested for

their effect on iTreg induction in Dicer2/2 CD4+ T cells. Asshown in Fig. 5B and 5C, LY294002 enhanced the induction ofiTregs in Dicer+/2 CD4+ T cells, but it had no significant effect onDicer2/2 CD4+ T cells. In contrast, both rapamycin and iAKT1/2

FIGURE 3. miRNA regulation of IFN-g expression is not important for induction of iTregs. (A) IFN-g expression in CD4+ T cells polarized toward the

indicated Th cell subsets. IFN-g levels were measured by a FlowCytomix-based assay, and the mean + SD from three independent experiments are shown.

Expression was significantly higher in Dicer2/2 CD4+ T cells in Th0 (*p = 0.03), Th1 (*p = 0.05), Th2 (*p = 0.04), and Th17 (**p = 0.003) cell subsets.

However, in iTregs, IFN-g production was virtually absent in both Dicer+/2 and Dicer2/2 CD4+ T cells. (B) iTreg induction in the presence of a neu-

tralizing IFN-g Ab or exogenously added IFN-g (from culture supernatants of cells polarized under Th1 conditions). Representative experiment (left

panels). Bar graph (mean + SD) showing data from three independent experiments (right panel); the addition of an IFN-g–neutralizing Ab did not enhance

iTreg induction in Dicer+/2 or Dicer2/2 CD4+ T cells; whereas exogenously added IFN-g inhibited induction in both Dicer+/2 and Dicer2/2 CD4+ T cells.

5672 MicroRNA REGULATION OF Treg DEVELOPMENT

by guest on April 13, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

significantly increased the induction of iTregs in Dicer+/2 andDicer2/2 CD4+ T cells. Therefore, miRNAs appear to be impor-tant for regulating the mTOR signaling pathway downstream ofPI3K. Examining the levels of key proteins in this pathway(Fig. 5D–F), PTEN appeared to be slightly upregulated in Dicer2/2

CD4+ T cells; however, upon further experiments, this differencewas not apparent and revealed no statistically significant changes.Looking downstream, the level of PDK displayed a slight increase(10%) in Dicer2/2 CD4+ T cells that was consistent in furtherexperiments, yielding statistically significant results. In contrast, thelevel of AKT was not affected by miRNAs. However, its phos-phorylation at Thr308 and Ser473 was significantly enhanced inDicer2/2 CD4+ T cells, indicating an upregulation of AKT activity.This correlated with an increase in the levels of mTOR and Rictor,which are part of mTORC2 that activates AKT through phosphor-ylation of Ser473. Consistent with an increase in AKT activity wasthe downstream activation of mTORC1, as measured by phos-phorylation of ribosomal protein S6. Likewise, the level of FoxO1awas significantly downregulated in Dicer2/2 CD4+ T cells, and itsrelative phosphorylation was greatly enhanced. Because FoxOproteins are important regulators of Treg development throughtheir activation of Foxp3 transcription (38, 39), we examinedwhether overexpression of FoxO3a could enhance iTreg inductionin Dicer2/2 CD4+ T cells. Induction was enhanced by FoxO3aoverexpression, and this was further enhanced by overexpressionof a phosphorylation mutant of FoxO3a (A3A) that contains ala-nine residues at the three sites phosphorylated by AKT, thusmaking it resistant to degradation by AKT phosphorylation(Supplemental Fig. 3) (41). A similar increase in induction wasobserved with overexpression of these FoxO3a proteins in Dicer+/2

CD4+ T cells. However, the level of iTreg induction did not cor-relate with the absolute level of FoxO3a expression. Therefore,FoxO3a protein expression cannot, by itself, explain the effects oniTreg induction in Dicer2/2 CD4+ T cells, and its relative impor-tance is unclear.

Rictor and mTOR are relevant targets of miR-15b/16 in iTreginduction. Target prediction algorithms identified potential targetsites of miR-15b/16 in the 39UTRs of several components of themTOR signaling pathway (29, 35). These included Akt3, Sgk1,and Rictor. Of these, Akt3 expression was greater in Dicer2/2

CD4+ T cells, as measured by RNA (Y. Singh and B. Cobb, un-

published observations), but this was not thought to be relevantbecause, as shown above, there was no increase in the totalamount of AKT in Dicer2/2 CD4+ T cells in which the AKT Abused for the Western blot recognizes all three isoforms (AKT1,AKT2, and AKT3). Therefore, any changes in the level of AKT3did not affect the overall level of AKT. Sgk1 encodes serumglucocorticoid kinase (SGK)1, which is related to AKT and canphosphorylate FoxO proteins at the same sites as AKT andsimilarly regulate their activity (42). SGK1 is important for Th17development (43, 44), as well as for Th1 and Th2 development(45). However, its role in Treg development is less clear. SGK1levels (as measured by RNA) did not change significantly be-tween conventionally activated T cells and iTregs, and over-expression of SGK1 had minimal effect on iTreg induction(Y. Singh and B. Cobb, unpublished observations). Therefore,SGK1 was not thought to be a relevant target of miR-15b/16 inthe induction of iTregs.In contrast, Rictor did appear to be a relevant target of miR-

15b/16. Its expression was reduced at both the protein and mRNAlevels when miR-15b/16 was overexpressed (Fig. 6A, 6B). Theincrease in protein level in Dicer2/2 CD4+ T cells and the de-crease in miR-15b/16–overexpressing cells was also observed atthe mRNA level (Fig. 6B). A luciferase reporter construct con-taining the 39UTR of Rictor was suppressed by overexpression ofmiR-15b/16, and suppression was significantly reduced when thepredicted target sequence was mutated (Fig. 6C). Furthermore, themiR-15b miRNA mimic reduced the level of Rictor expression inDicer2/2 CD4+ T cells (Fig. 6D). The relevance of the increasedRictor expression impacting iTreg induction in Dicer2/2 CD4+

T cells was tested by siRNA knockdown of Rictor, which enhancedthe induction of iTregs in Dicer2/2 CD4+ T cells by ∼1.5-fold(Fig. 6E). This did not bring the level of induction back to that inDicer+/2 CD4+ T cells transfected with a control oligonucleotide,but the knockdown of Rictor in Dicer2/2 CD4+ T cells did notreduce its level to what was observed in Dicer+/2 CD4+ T cells.Therefore, the differences in iTreg induction between Rictor siRNA-transfected Dicer2/2 CD4+ T cells and control-transfected Dicer+/2

CD4+ T cells were consistent with the prevailing level of Rictorexpression. Interestingly, there was not a significant knockdown ofRictor expression in Dicer+/2 cells, but this could be due to itsalready low level of expression. Despite this caveat, the transfection

FIGURE 4. Proximal events in TGF-b signaling are not affected in Dicer2/2 CD4+ T cells. TGF-b signaling was examined in Western blots of

p-SMAD2 at serines 465 and 467. (A) Representative experiment. (B) Mean + SD of the relative expression or phosphorylation levels in anti-CD3/CD28 +

TGF-b–stimulated cells from three separate experiments. Owing to the variability in the relative signal strength of the indicated protein compared with

b-actin, we normalized the results in each experiment to b-actin, determined the ratio of Dicer2/2/Dicer+/2 CD4+ T cells, and calculated the mean + SD of

this value for all experiments. No significant difference was observed in the levels of SMAD2/3 protein or the phosphorylation of SMAD2.

The Journal of Immunology 5673

by guest on April 13, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

of Rictor siRNAs still resulted in an ∼1.3-fold increase in Foxp3+

cells in these experiments. Therefore, small changes in Rictor ex-pression around its normal level could result in significant differ-ences in mTOR signaling and iTreg induction. Overexpression ofmiR-15b/16 significantly reduced the level of Rictor in wild-typecells, increasing the percentage of Foxp3, on average, by ∼1.5-fold.However, these experiments were performed at different times, andthe basal level of Foxp3-expressing cells was higher with the cellsused in these experiments. This and the problems inherent incomparing retroviral-transduced and siRNA-transfected experi-ments make it difficult to compare the effects of siRNA knockdownand miR-15b/16 expression on iTreg induction. Nevertheless,changes in Rictor levels over a broad range were important foriTreg induction, and most notably, the upregulation of Rictor inDicer2/2 CD4+ T cells appeared to play an important role in thereduced levels of iTreg induction.To further address the relevance of miR-15b/16 regulation of

Rictor on iTreg induction, we found that the siRNA knockdownof Rictor expression in T cells reversed the inhibition of iTreginduction when miR-15b/16 were inhibited by the miR-15 decoy(Fig. 6G). However, quantifying the importance of miR-15b/16

regulation of Rictor expression, as opposed to other targets,will require the gene mutation of the target site in the Rictor39UTR. Undoubtedly, other targets of miR-15b/16 will also playa role in iTreg induction; in fact, within the mTOR signalingpathway, miR-15b/16 additionally regulated the expression ofmTOR. Interestingly, target prediction algorithms did not iden-tify sites for miR-15b/16 in its 39UTR (29, 35); however, likeRictor, its expression was reduced by overexpression of miR-15b/16 (Fig. 6A), and an miRNA mimic for miR-15b de-creased its expression in Dicer2/2 CD4+ T cells (Fig. 6D). Therefore,mTOR could have target sites for miR-15b/16 not recognizedby the target prediction algorithms, or it may be regulatedindirectly.Changes in mTOR and Rictor levels were consistent with a

decrease in AKT activation, as measured by phosphorylation ofSer473, but this was not reflected by any significant changes inFoxO1a or FoxO3a expression (Fig. 6A). In contrast, mTORC1activity was inhibited, as measured by phosphorylation of ribo-somal protein S6. Therefore, miR-15b/16 appears to affect mTORsignaling downstream of mTORC1 but not by regulating the ex-pression or phosphorylation of FoxO proteins.

FIGURE 5. The mTOR signaling pathway is enhanced in Dicer2/2 CD4+ T cells. (A) A simplified diagram of the mTOR signaling pathway showing the

targets of specific pharmacological inhibitors. See text for details. Representative experiment (B) and mean + SD (C) for three independent experiments

examining the effect of mTOR signaling pathway inhibitors on iTreg induction in Dicer2/2 and Dicer+/2 CD4+ T cells. The PI3K inhibitor LY294002 had

minimal effect, whereas the mTORC1 inhibitor rapamycin (*p = 0.05) and the AKT inhibitor iAKT1/2 (*p = 0.02) significantly enhanced iTreg induction in

Dicer2/2 CD4+ T cells. (D) Levels of the indicated proteins in a representative Western blot. Mean + SD for the relative level of expression (E) or p-protein

(F) of the indicated protein between Dicer2/2 and Dicer+/2 CD4+ T cells. Ratios were calculated as in Fig. 4 and were derived from three separate ex-

periments. In Dicer2/2 CD4+ T cells there was no significant change in PTEN or AKT expression, a slight increase in the level of PDK-1 (*p = 0.02), a

significant decrease in FoxO1a (**p = 0.004), and a significant increase in mTOR (**p = 0.005) and Rictor (**p = 0.007). In addition, in Dicer2/2 CD4+

T cells, there was an increase in p-AKT at both Thr308 (*p = 0.03) and Ser473 (**p = 0.01) and an increase in p-FoxO1a (**p = 0.002) and p-S6 (*p = 0.03),

indicating an enhanced response in the mTOR signaling pathway.

5674 MicroRNA REGULATION OF Treg DEVELOPMENT

by guest on April 13, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

DiscussionmiRNAs are critical regulators of Treg development and func-tion. This work identified three individual miRNAs (miR-15b/16,miR-24, and miR-29a) that are important for the in vitro inductionof iTregs, and it also demonstrated the importance of miR-15b/16in the in vivo generation of pTregs. These miRNAs were identifiedthrough overexpression and blocking experiments, but they couldalso reverse the defect in the induction of iTregs in Dicer2/2

CD4+ T cells. Further support for the importance of miR-29came from the T cell–specific deletion of the locus encoding it,which results in a significant reduction in Tregs in vivo. However,little difference in the induction of iTregs was found under theconditions used, which included the addition of retinoic acid (34).We found that the influence of miRNAs is greatest under subop-timal conditions for iTreg induction. Therefore, it would be in-teresting to determine whether the deletion of miR-29a affects

the induction of iTregs in the conditions used in this study. NoT cell–specific deletions of miR-15/16 or miR-24 have beencharacterized, to the best of our knowledge. However, suchexperiments would be complicated because both miR-15/16and miR-24 are encoded on two separate loci, and it is notknown whether one locus is preferentially expressed over theother in T cells or whether the deletion of one locus wouldimpact the expression of the other. Furthermore, miR-15/16 ispart of a family of related miRNAs that are capable of targetingthe same sequences in messages. These include miR-195, miR-322, miR-424, miR-497, and miR-1907, so it could be difficultto ensure lack of expression of all of these miRNAs, whichwould make the interpretation of the effects on gene expressionchallenging. Therefore, the overexpression and blocking ap-proach offers the advantage of characterizing the effects of thisentire family.

FIGURE 6. miR-15b/16 regulates the expression of Rictor and mTOR in iTregs. (A) Expression levels and phosphorylation of indicated proteins in

control or miR-15b/16-expressing CD4+ T cells, as measured by Western blot. Representative experiment (left panel) and mean + SD for the relative level

of expression or p-protein of the indicated protein between control or miR-15b/16–expressing CD4+ T cells (right panels). In miR-15b/16–overexpressing

T cells, there was a reduction in the level of Rictor (*p = 0.03) and mTOR (**p = 0.002), as well as the phosphorylation of AKT Ser473 (*p = 0.02) and S6

(*p = 0.02). Values are derived from between three and six experiments. (B) The level of Rictor mRNA was similarly affected by miR-15b/16 over-

expression, and it was increased in Dicer2/2 CD4+ T cells. (C) Luciferase reporter constructs containing the 39UTR of Rictor were regulated by miR-15b/

16 overexpression in HEK 293T cells, and this was significantly reduced when the predicted miR-15b/16 target site was mutated. ***p = 0.0009. (D) An

miRNA mimic for miR-15b decreased the expression of Rictor and mTOR in Dicer2/2 CD4+ T cells, as measured by Western blot. (E) siRNA knockdown

of Rictor enhanced iTreg induction in Dicer2/2 and Dicer+/2 CD4+ T cells. Representative experiment (left panel) and mean + SD from two independent

experiments, each performed in triplicate (Dicer2/2, ***p = 0.0002; Dicer+/2, **p = 0.005) (right panel). (F) Expression levels of Rictor in one of the two

siRNA-knockdown experiments, with the relative level compared with GAPDH listed below each band. Similar results were obtained in each experiment.

(G) The reduction in iTreg induction with the inhibition of miR-15b/16 is counteracted by the knockdown of Rictor expression. CD4+ T cells were

transduced with a control (Cont LV) or miR-16 decoy expressing lentivirus (16 decoy LV). After 16–18 h, cells were transfected with a control or Rictor

siRNA; cells were harvested 48 h later and analyzed for Foxp3 expression. Representative experiment (left panels) and bar graph showing mean + SD from

four independent experiments (right panel). Under these conditions, the miR-15b decoy significantly inhibited iTreg induction. *****p = 0.000006, Cont

LV + Cont siRNA versus 16 decoy + Cont siRNA. Likewise, the Rictor siRNA enhanced iTreg induction. *p = 0.012, Cont LV + Cont siRNA versus Cont

LV + Rictor siRNA. Most importantly, the Rictor siRNA counteracted the inhibition of iTreg induction by the miR-16 decoy. *p = 0.011, 16 decoy LV +

Cont siRNA versus 16 decoy LV 2 Rictor siRNA.

The Journal of Immunology 5675

by guest on April 13, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Because miR-15b/16 had the greatest effect in overexpressionand blocking experiments, we focused on identifying relevanttargets regulated by this miRNA family. Signaling through themTOR pathway was enhanced in Dicer2/2 CD4+ T cells, andexpression of the mTORC2 components Rictor and mTOR wasfound to be regulated by miR-15b/16. It has been well establishedthat a reduction in signaling through the mTOR pathway is re-quired for Treg development. A T cell–specific deletion of mTORresults in the spontaneous induction of iTregs without the additionof TGF-b (18). The expression of a constitutively active AKTinhibits iTreg induction (19), and pharmacological inhibition ofthe mTOR signaling pathway enhances iTreg induction (20).Therefore, regulation of this pathway by miRNAs is an obviousmechanism for fine-tuning its function in T cell development.The regulation of mTORC2 by miR-15b/16 affected mTOR

signaling by reducing AKT activity and subsequent downstreamevents. This disruption in mTOR signaling is consistent with theT cell–specific deletion of Rictor, which diminishes AKT activationand enhances iTreg induction (46). It is also consistent with theeffects observed in the T cell–specific deletion of mTOR (18).However, in contrast to the T cell–specific deletion of mTOR, iTreginduction in miR-15b/16–overexpressing cells still required theaddition of TGF-b. Also, the proximal events of TGF-b signalingwere not affected inDicer2/2 CD4+ T cells as they were inMtor2/2

CD4+ T cells. Therefore, the effects on signaling by the reduction inmTOR levels do not entirely follow the complete loss of mTORexpression, suggesting that other signaling mechanisms involvingthe mTOR pathway exist that are only sensitive to large changes inmTORC activity. Interestingly, the reduced AKT activity in T cellsoverexpressing miR-15b/16 had no effect on FoxO protein levels orphosphorylation, suggesting that the change in AKT activation wasnot sufficient to regulate this part of the pathway. In contrast, it wassufficient to regulate mTORC1 activity, leading us to speculate thatthe decrease in both mTORC1 and mTORC2 levels from reducedmTOR expression amplified the effect of the reduced AKT activity.Alternatively, regulation of mTORC1 in T cells might be moresensitive to small changes in AKT activity.The regulation of the mTOR signaling pathway is a major

mechanism controlling Treg development, and miR-15b/16 affectsthis pathway by regulating the expression of Rictor and mTOR.However, there will undoubtedly be other important targets for thismiRNA, as well as targets for miR-24 and miR-29a. Furthermore,the regulation of expression of these miRNAs will need to bedifferentially controlled in iTreg induction to enhance their levelsover what is found in other Th subtypes. T cell activation resultsin a global change in miRNA levels through their turnover due tothe destabilization of Argonaute proteins; however, transcriptionalregulation of miRNA transcripts and their posttranscriptionalprocessing are also important (47). Therefore, it will be interestingto dissect the mechanisms controlling the expression of these im-portant miRNAs.

AcknowledgmentsWe thank Robert Unsell, Poorna Sharma, and Claudia Gale for technical

assistance. We also thank Anna Morgunowicz, Grainne Mcgeever, and

Wendy Balderson for technical help with animal experiments.

DisclosuresThe authors have no financial conflicts of interest.

References1. Josefowicz, S. Z., L.-F. Lu, and A. Y. Rudensky. 2012. Regulatory T cells:

mechanisms of differentiation and function. Annu. Rev. Immunol. 30: 531–564.

2. Josefowicz, S. Z., R. E. Niec, H. Y. Kim, P. Treuting, T. Chinen, Y. Zheng,D. T. Umetsu, and A. Y. Rudensky. 2012. Extrathymically generated regulatoryT cells control mucosal TH2 inflammation. Nature 482: 395–399.

3. Samstein, R. M., S. Z. Josefowicz, A. Arvey, P. M. Treuting, and A. Y. Rudensky.2012. Extrathymic generation of regulatory T cells in placental mammals mit-igates maternal-fetal conflict. Cell 150: 29–38.

4. Fontenot, J. D., M. A. Gavin, and A. Y. Rudensky. 2003. Foxp3 programs thedevelopment and function of CD4+CD25+ regulatory T cells. Nat. Immunol. 4:330–336.

5. He, L., and G. J. Hannon. 2004. MicroRNAs: small RNAs with a big role in generegulation. Nat. Rev. Genet. 5: 522–531.

6. Meijer, H. A., Y. W. Kong, W. T. Lu, A. Wilczynska, R. V. Spriggs,S. W. Robinson, J. D. Godfrey, A. E. Willis, and M. Bushell. 2013. Translationalrepression and eIF4A2 activity are critical for microRNA-mediated gene regu-lation. Science 340: 82–85.

7. 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.

8. 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.

9. 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.

10. 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.

11. Lu, L.-F., T.-H. Thai, D. P. Calado, A. Chaudhry, M. Kubo, K. Tanaka,G. B. Loeb, H. Lee, A. Yoshimura, K. Rajewsky, and A. Y. Rudensky. 2009.Foxp3-dependent microRNA155 confers competitive fitness to regulatory T cellsby targeting SOCS1 protein. Immunity 30: 80–91.

12. Kohlhaas, S., O. A. Garden, C. Scudamore, M. Turner, K. Okkenhaug, andE. Vigorito. 2009. Cutting edge: the Foxp3 target miR-155 contributes to thedevelopment of regulatory T cells. J. Immunol. 182: 2578–2582.

13. Lu, L.-F., M. P. Boldin, A. Chaudhry, L.-L. Lin, K. D. Taganov, T. Hanada,A. Yoshimura, D. Baltimore, and A. Y. Rudensky. 2010. Function of miR-146a incontrolling Treg cell-mediated regulation of Th1 responses. Cell 142: 914–929.

14. Beyer, M., Y. Thabet, R.-U. M€uller, T. Sadlon, S. Classen, K. Lahl, S. Basu,X. Zhou, S. L. Bailey-Bucktrout, W. Krebs, et al. 2011. Repression of the ge-nome organizer SATB1 in regulatory T cells is required for suppressive functionand inhibition of effector differentiation. Nat. Immunol. 12: 898–907.

15. de Kouchkovsky, D., J. H. Esensten, W. L. Rosenthal, M. M. Morar,J. A. Bluestone, and L. T. Jeker. 2013. microRNA-17-92 regulates IL-10 pro-duction by regulatory T cells and control of experimental autoimmune en-cephalomyelitis. J. Immunol. 191: 1594–1605.

16. Jiang, S., C. Li, V. Olive, E. Lykken, F. Feng, J. Sevilla, Y. Wan, L. He, andQ.-J. Li. 2011. Molecular dissection of the miR-17-92 cluster’s critical dualroles in promoting Th1 responses and preventing inducible Treg differen-tiation. Blood 118: 5487–5497.

17. Jeker, L. T., X. Zhou, K. Gershberg, D. de Kouchkovsky, M. M. Morar,G. Stadthagen, A. H. Lund, and J. A. Bluestone. 2012. MicroRNA 10a marksregulatory T cells. PLoS One 7: e36684.

18. Delgoffe, G. M., T. P. Kole, Y. Zheng, P. E. Zarek, K. L. Matthews, B. Xiao,P. F. Worley, S. C. Kozma, and J. D. Powell. 2009. The mTOR kinase differ-entially regulates effector and regulatory T cell lineage commitment. Immunity30: 832–844.

19. Haxhinasto, S., D. Mathis, and C. Benoist. 2008. The AKT-mTOR axis regulatesde novo differentiation of CD4+Foxp3+ cells. J. Exp. Med. 205: 565–574.

20. Sauer, S., L. Bruno, A. Hertweck, D. Finlay, M. Leleu, M. Spivakov,Z. A. Knight, B. S. Cobb, D. Cantrell, E. O’Connor, et al. 2008. T cell receptorsignaling controls Foxp3 expression via PI3K, Akt, and mTOR. Proc. Natl.Acad. Sci. USA 105: 7797–7802.

21. Haraguchi, T., Y. Ozaki, and H. Iba. 2009. Vectors expressing efficient RNAdecoys achieve the long-term suppression of specific microRNA activity inmammalian cells. Nucleic Acids Res. 37: e43.

22. Ebert, M. S., J. R. Neilson, and P. A. Sharp. 2007. MicroRNA sponges: com-petitive inhibitors of small RNAs in mammalian cells. Nat. Methods 4: 721–726.

23. Steiner, D. F., M. F. Thomas, J. K. Hu, Z. Yang, J. E. Babiarz, C. D. Allen,M. Matloubian, R. Blelloch, and K. M. Ansel. 2011. MicroRNA-29 regulatesT-box transcription factors and interferon-g production in helper T cells. Immunity35: 169–181.

24. Singh, Y., V. Kaul, A. Mehra, S. Chatterjee, S. Tousif, V. P. Dwivedi, M. Suar,L. Van Kaer, W. R. Bishai, and G. Das. 2013. Mycobacterium tuberculosiscontrols microRNA-99b (miR-99b) expression in infected murine dendritic cellsto modulate host immunity. J. Biol. Chem. 288: 5056–5061.

25. Patton, D. T., O. A. Garden, W. P. Pearce, L. E. Clough, C. R. Monk, E. Leung,W. C. Rowan, S. Sancho, L. S. Walker, B. Vanhaesebroeck, and K. Okkenhaug.2006. Cutting edge: the phosphoinositide 3-kinase p110 delta is critical for thefunction of CD4+CD25+Foxp3+ regulatory T cells. J. Immunol. 177: 6598–6602.

26. Porrello, E. R., A. I. Mahmoud, E. Simpson, B. A. Johnson, D. Grinsfelder,D. Canseco, P. P. Mammen, B. A. Rothermel, E. N. Olson, and H. A. Sadek.2013. Regulation of neonatal and adult mammalian heart regeneration by themiR-15 family. Proc. Natl. Acad. Sci. USA 110: 187–192.

27. Cimmino, A., G. A. Calin, M. Fabbri, M. V. Iorio, M. Ferracin, M. Shimizu,S. E. Wojcik, R. I. Aqeilan, S. Zupo, M. Dono, et al. 2005. miR-15 and miR-16induce apoptosis by targeting BCL2. [Published erratum appears in 2006 Proc.Natl. Acad. Sci. USA 103: 2464.] Proc. Natl. Acad. Sci. USA 102: 13944–13949.

5676 MicroRNA REGULATION OF Treg DEVELOPMENT

by guest on April 13, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

28. Lewis, M. A., E. Quint, A. M. Glazier, H. Fuchs, M. H. De Angelis, C. Langford,S. van Dongen, C. Abreu-Goodger, M. Piipari, N. Redshaw, et al. 2009. AnENU-induced mutation of miR-96 associated with progressive hearing loss inmice. Nat. Genet. 41: 614–618.

29. Betel, D., M. Wilson, A. Gabow, D. S. Marks, and C. Sander. 2008. The micro-RNA.org resource: targets and expression. Nucleic Acids Res. 36: D149–D153.

30. Wei, J., O. Duramad, O. A. Perng, S. L. Reiner, Y.-J. Liu, and F. X. Qin. 2007.Antagonistic nature of T helper 1/2 developmental programs in opposing pe-ripheral induction of Foxp3+ regulatory T cells. Proc. Natl. Acad. Sci. USA 104:18169–18174.

31. Chang, J.-H., Y.-J. Kim, S.-H. Han, and C.-Y. Kang. 2009. IFN-gamma-STAT1signal regulates the differentiation of inducible Treg: potential role for ROS-mediated apoptosis. Eur. J. Immunol. 39: 1241–1251.

32. 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.

33. Ma, F., S. Xu, X. Liu, Q. Zhang, X. Xu, M. Liu, M. Hua, N. Li, H. Yao, andX. Cao. 2011. The microRNA miR-29 controls innate and adaptive immuneresponses to intracellular bacterial infection by targeting interferon-g. Nat.Immunol. 12: 861–869.

34. Smith, K. M., M. Guerau-de-Arellano, S. Costinean, J. L. Williams, A. Bottoni,G. Mavrikis Cox, A. R. Satoskar, C. M. Croce, M. K. Racke, A. E. Lovett-Racke,and C. C. Whitacre. 2012. miR-29ab1 deficiency identifies a negative feedbackloop controlling Th1 bias that is dysregulated in multiple sclerosis. J. Immunol.189: 1567–1576.

35. Lewis, B. P., C. B. Burge, and D. P. Bartel. 2005. Conserved seed pairing, oftenflanked by adenosines, indicates that thousands of human genes are microRNAtargets. Cell 120: 15–20.

36. Chen, W., W. Jin, N. Hardegen, K.-J. Lei, L. Li, N. Marinos, G. McGrady, andS. M. Wahl. 2003. Conversion of peripheral CD4+CD252 naive T cells to CD4+

CD25+ regulatory T cells by TGF-beta induction of transcription factor Foxp3. J.Exp. Med. 198: 1875–1886.

37. Laplante, M., and D. M. Sabatini. 2012. mTOR signaling in growth control anddisease. Cell 149: 274–293.

38. Ouyang, W., O. Beckett, Q. Ma, J.-H. Paik, R. A. DePinho, and M. O. Li. 2010.Foxo proteins cooperatively control the differentiation of Foxp3+ regulatoryT cells. Nat. Immunol. 11: 618–627.

39. Kerdiles, Y. M., E. L. Stone, D. R. Beisner, M. A. McGargill, I. L. Ch’en,C. Stockmann, C. D. Katayama, and S. M. Hedrick. 2010. Foxo transcriptionfactors control regulatory T cell development and function. Immunity 33: 890–904.

40. Chi, H. 2012. Regulation and function of mTOR signalling in T cell fate deci-sions. Nat. Rev. Immunol. 12: 325–338.

41. Hedrick, S. M., R. Hess Michelini, A. L. Doedens, A. W. Goldrath, andE. L. Stone. 2012. FOXO transcription factors throughout T cell biology. Nat.Rev. Immunol. 12: 649–661.

42. Brunet, A., J. Park, H. Tran, L. S. Hu, B. A. Hemmings, and M. E. Greenberg.2001. Protein kinase SGK mediates survival signals by phosphorylating theforkhead transcription factor FKHRL1 (FOXO3a). Mol. Cell. Biol. 21: 952–965.

43. Kleinewietfeld, M., A. Manzel, J. Titze, H. Kvakan, N. Yosef, R. A. Linker,D. N. Muller, and D. A. Hafler. 2013. Sodium chloride drives autoimmunedisease by the induction of pathogenic TH17 cells. Nature 496: 518–522.

44. Wu, C., N. Yosef, T. Thalhamer, C. Zhu, S. Xiao, Y. Kishi, A. Regev, andV. K. Kuchroo. 2013. Induction of pathogenic TH17 cells by inducible salt-sensing kinase SGK1. Nature 496: 513–517.

45. Heikamp, E. B., C. H. Patel, S. Collins, A. Waickman, M.-H. Oh, I.-H. Sun,P. Illei, A. Sharma, A. Naray-Fejes-Toth, G. Fejes-Toth, et al. 2014. The AGCkinase SGK1 regulates TH1 and TH2 differentiation downstream of themTORC2 complex. Nat. Immunol. 15: 457–464.

46. Lee, K., P. Gudapati, S. Dragovic, C. Spencer, S. Joyce, N. Killeen,M. A. Magnuson, and M. Boothby. 2010. Mammalian target of rapamycinprotein complex 2 regulates differentiation of Th1 and Th2 cell subsets viadistinct signaling pathways. Immunity 32: 743–753.

47. Bronevetsky, Y., A. V. Villarino, C. J. Eisley, R. Barbeau, A. J. Barczak,G. A. Heinz, E. Kremmer, V. Heissmeyer, M. T. McManus, D. J. Erle, et al.2013. T cell activation induces proteasomal degradation of Argonaute and rapidremodeling of the microRNA repertoire. J. Exp. Med. 210: 417–432.

The Journal of Immunology 5677

by guest on April 13, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from