Essential Functions for ID at Multiple Checkpoints in Invariant NKT Cell Development

This information is current as Mihalis Verykokakis, Veena Krishnamoorthy, Antonio of September 27, 2021. Iavarone, Anna Lasorella, Mikael Sigvardsson and Barbara L. Kee J Immunol 2013; 191:5973-5983; Prepublished online 15 November 2013;

doi: 10.4049/jimmunol.1301521 Downloaded from http://www.jimmunol.org/content/191/12/5973

Supplementary http://www.jimmunol.org/content/suppl/2013/11/18/jimmunol.130152 Material 1.DC1 http://www.jimmunol.org/ References This article cites 52 articles, 22 of which you can access for free at: http://www.jimmunol.org/content/191/12/5973.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision by guest on September 27, 2021

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2013 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Essential Functions for ID Proteins at Multiple Checkpoints in Invariant NKT Cell Development

Mihalis Verykokakis,*,† Veena Krishnamoorthy,* Antonio Iavarone,‡,x,{ Anna Lasorella,‡,x,{ Mikael Sigvardsson,‖ and Barbara L. Kee*,†

Invariant NKT (iNKT) cells display characteristics of both adaptive and innate lymphoid cells (ILCs). Like other ILCs, iNKT cells constitutively express ID proteins, which antagonize the E transcription factors that are essential for adaptive lymphocyte development. However, unlike ILCs, ID2 is not essential for thymic iNKT cell development. In this study, we demonstrated that ID2 and ID3 redundantly promoted iNKT cell lineage specification involving the induction of the signature PLZF and that ID3 was critical for development of TBET-dependent NKT1 cells. In contrast, both ID2 and ID3 limited iNKT cell numbers by enforcing the postselection checkpoint in conventional thymocytes. Therefore, iNKT cells show both adaptive and innate-like requirements for ID proteins at distinct checkpoints during iNKT cell development. The Journal of Immunology, Downloaded from 2013, 191: 5973–5983.

atural killer T cells are T lymphocytes that display the Tcra locus and are recombined through secondary rearrange- characteristics of innate immune cells, including the use ments that occur late in the life of CD4+CD8+ (double-positive N of invariant receptors to recognize pathogen and rapid [DP]) thymocytes (5). Mice harboring mutations that decrease DP activation without prior Ag exposure. NKT cells diverge from the thymocyte survival and mice with limited Tcra recombination http://www.jimmunol.org/ conventional program during positive selection. Following lack iNKT cells (6–8). Positive selection of iNKT cells requires selection, NKT cell maturation is coupled with the ability to rapidly lipid Ag presentation by the nonclassical MHC class I protein CD1d secrete cytokines when challenged (1). As a consequence of their expressed on DP thymocytes, along with signals from the signaling “poised” effector state and ability to produce numerous cytokines, lymphocyte activation molecule (SLAM) family receptors (9, 10). NKT cells can act as both positive and negative regulators of an This selection pathway results in the TCR-dependent induction of immune response. They promote pathogen and tumor clearance, the lineage-specifying transcription factor promyelocytic leukemia but their activity can contribute to diseases such as autoimmunity, zinc finger (PLZF), which is essential for iNKT development and atherosclerosis, and asthma (2, 3). To harness the therapeutic potential confers innate properties to conventional CD4 T cells when ectop- of NKT cells, a comprehensive understanding of the mechanisms ically expressed (11–14). by guest on September 27, 2021 controlling NKT cell selection, maturation, and effector function is iNKT cell maturation is divided into four stages, based on the required. surface expression of CD24, CD44, and NK1.1 (15, 16). Stage 2 2 Invariant NKT (iNKT) cells, characterized by a Va14-Ja18 0 (CD24+CD44 NK1.1 ) represents rare iNKT cell precursors TCRa-chain paired with TCR Vb7, Vb8, and Vb2 chains (4), are among postselection (PS) DP thymocytes. Stage 1 cells down- the most abundant and well-characterized NKT cell population in regulate CD8 and CD24 and express low levels of the memory mice. The Va14 and Ja18 segments are located far apart in marker CD44. Stage 2 cells have increased CD44 and can progress tostage3inthethymus,wheretheyexpressseveralNKcell receptors including NK1.1, or they can exit the thymus and mature *Department of Pathology, University of Chicago, Chicago, IL 60637; †Committee on further in the peripheral tissues. Stage 2 iNKT cells have been ‡ Immunology, University of Chicago, Chicago, IL 60637; Institute for Cancer Genet- considered an immature stage, although these cells can robustly ics, Columbia University Medical Center, New York, NY 10032; xDepartment of Neurology, Columbia University Medical Center, New York, NY 10032; {Department produce both Th1 and Th2 cytokines. However, a subset of stage of Pathology, Columbia University Medical Center, New York, NY 10032; and ‖Ex- 2 iNKT cells are terminally differentiated cells that express the perimental Hematopoiesis Unit, Department of Clinical and Experimental Medicine, transcription factor GATA3, and these have been classified recently Faculty for Health Sciences, Linko¨ping University, 58183 Linko¨ping, Sweden as NKT2 cells. Stage 3 iNKT cells preferentially produce the Th1 Received for publication June 12, 2013. Accepted for publication October 9, 2013. cytokine IFN-g with lower amounts of Th2 cytokines and have been This work was supported by National Institutes of Health Grant CA099978 (to B.L.K.). B.L.K. is a scholar of the Leukemia and Lymphoma Society. classified as NKT1 (17). TBET is critical for the maturation, sur- vival, and Th1-like characteristics of NKT1 (18, 19). Therefore, The sequence data presented in this article have been submitted to the Gene Expres- sion Omnibus (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc) under accession acquisition of an iNKT cell TCR and induction of PLZF and TBET number GSE50933. define three critical checkpoints during iNKT cell development that Address correspondence and reprint requests to Dr. Barbara L. Kee and Dr. Mihalis control their abundance and functional competence. Verykokakis, Department of Pathology, University of Chicago, 924 East 57th Street, The E protein transcription factors are important regulators of JFK, Room R318, Chicago, IL 60637. E-mail addresses: [email protected] (B.L.K.) and [email protected] (M.V.) conventional T cell development and selection, and they control the The online version of this article contains supplemental material. life span and gene signature of DP thymocytes (20, 21). E protein Abbreviations used in this article: BM, bone marrow; DP, double-positive; ILC, function can be modulated through antagonistic interactions with innate lymphoid cell; iNKT, invariant NKT; LMC, littermate control; PLZF, promye- any of the four members of the ID family (ID1–4) (22). TCR- locytic leukemia zinc finger; PS, postselection; qPCR, quantitative PCR; SLAM, dependent induction of ID3, and the consequent decrease in E signaling lymphocyte activation molecule; Tet, tetramer; WT, wild-type. protein activity, is critical for positive selection of conventional Copyright Ó 2013 by The American Association of Immunologists, Inc. 0022-1767/13/$16.00 CD4 and CD8 T lymphocytes (23–25). However, a role for ID3 in www.jimmunol.org/cgi/doi/10.4049/jimmunol.1301521 5974 ID PROTEINS CONTROL iNKT CELL DEVELOPMENT the TCR-dependent selection of iNKT cells has not been dem- Cell culture, IFN-g/IL-4 production, and intracellular staining onstrated. Moreover, although development of all non-T cell line- Thymocytes were treated with 50 ng PMA, 1 mg ionomycin, and brefeldin age innate lymphoid cells (ILC) depends on the ID2 protein (26), A for 5 h, before harvesting and staining intracellularly for IFN-g and IL-4 thymic development of iNKT cells appears to be independent of using the BD Biosciences Cytofix/Cytoperm kit. ID2 (27). Why iNKT cells differ from other ILCs in their re- RNA analysis and real-time quantitative PCR quirement for ID2 remains to be determined. We and others (28–31) recently demonstrated that ID3 restricts Total RNA was extracted and DNAase-treated from sorted thymocytes us- the development of ab and gd NKT-like cells. In this study, we ing the RNAeasy Mini Kit (Qiagen) and was reverse transcribed using Superscript III (Invitrogen). Quantitative PCR (qPCR) was performed with showed that ID proteins were central regulators of the three major gene-specific primers in an iCycler (Bio-Rad), using the iQ SYBR Green thymic iNKT cell developmental checkpoints. ID proteins limited Supermix (Bio-Rad). Hypoxanthine phosphoribosyltransferase was used to selection into the iNKT cell pathway, at least in part, by extin- normalize expression levels of other transcripts. guishing the potential for secondary Tcra rearrangements in PS Microarray analysis DP thymocytes. Once an appropriate TCR was acquired, ID2 and 2 2 2 ID3 functioned redundantly in stage 0 iNKT cells to promote Ten thousand stage 1 (CD24+CD44 NK1.1 ) and stage 2 (CD24 CD44+ 2 2/2 expression of PLZF and entry into the NKT cell lineage. Fol- NK1.1 ) WT and Id3 iNKT cells from three independent experiments were sorted and subjected to microarray procedures as described previ- lowing iNKT cell specification, ID3 was required for the emer- + ously (34). Normalized mRNA expression data from WT stage 1, 2, and 3 gence of TBET NKT1 cells in the thymus. We have demonstrated iNKT cells, obtained from the Immunological Genome Consortium (35), critical functions for ID proteins in the temporal regulation of were analyzed to identify that are progressively upregulated 1.5- fold iNKT cell differentiation and in the control of their effector fates. from stage 1 to stage 2 and 1.5-fold from stage 2 to stage 3. The gene list Downloaded from was then subjected to dChIP analysis to identify differentially regulated Moreover, we revealed that ID proteins are universally required 2 2 genes between WT and Id3 / stage 2 iNKT cells. The data can be found for development of innate lymphoid cells. at http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc using the accession number GSE50933. Materials and Methods Statistics Mice

A standard Mann–Whitney unpaired test was used (unless otherwise http://www.jimmunol.org/ Mice were housed at the University of Chicago Animal Resource Center, stated) to determine the statistical probability of the differences observed and experiments were performed according to the guidelines of the Uni- between two populations of cells, using the GraphPad Prism software. versity of Chicago Institutional Animal Care and Use Committee. Id32/2, Id2f/f,andVa14Tg mice have been described previously (9, 32, 33). Tbx212/2 mice were purchased from The Jackson Laboratory, and CD4Cre Results mice were purchased from Taconic. Sh2d1a2/2 mice were a gift from ID3 limited the development of PLZF+ PS DP thymocytes C. Terhorst (Beth Israel Deaconess Medical Center, Boston, MA). All mice were on a C57BL/6 background. We and others previously showed that ID3 limits the number of PLZF+ innate ab and gd T lymphocytes in the thymus (28, 30, 31) Flow cytometry, cell sorting, and Abs In this study, we tested directly whether ID3 functions in the se- by guest on September 27, 2021 Thymocyte suspensions from 3- to 10-wk-old mice were incubated with lection of NKT cells at the DP stage of T cell development. gd anti-FcgR prior to staining with fluorochrome-conjugated Abs. Cells were T cells were electronically gated out from our analysis to avoid acquired in an LSRII Fortessa or sorted in a FACSAria and analyzed with potential cross-contamination of ab T cells. We found an accu- FlowJo. Enrichment of thymocytes for iNKT cells was performed by mulation of Zbtb16 (encoding PLZF) mRNA in Id32/2 PS DP staining total thymocytes with allophycocyanin-conjugated CD1dPBS57 tetramers (Tet), followed by anti-allophycocyanin microbeads and subjected thymocytes as compared with WT cells, where Zbtb16 mRNA to auto-MACS–based magnetic cell separation. Propidium iodide was in- was below detection (Fig. 1A). There was a small but reproducible 2 2 cluded in all samples to exclude dead cells from the analysis. Abs specific population of PLZF+ cells among Id3 / PS DP thymocytes, for the following Ags were purchased from BD Biosciences, eBioscience, whereas no PLZF+ cells were observed in the WT population (Fig. or BioLegend: CD4 (GK1.5), CD8a (53-6.7), TCRb (H57-597), CD122 1B, 1C). Therefore, an increased frequency of Id32/2 DP cells (5H4), CD44 (IM7), CD24 (M1/69), IFN-g (XMG1.2), IL-4 (11B11), CD69 (H1.2F3), NK1.1 (PK136), NKG2D (CX5), CD45.1 (A20), CD45.2 underwent specification to the NKT cell lineages, as measured by (104), RORgt (AFKJS-9), T-BET (4B10), and CXCR3 (CXCR3-173). The PLZF expression. anti-PLZF Ab was purchased from Santa Cruz Biotechnology (clone D9, NKT cells require SAP for their development (36). We found that catalog number sc-28319) and conjugated with Pacific Blue using the Pa- the emergence of PLZF+ cells among Id32/2 PS DP thymocytes cific Blue mAb Labeling Kit (P30013; Invitrogen). Intracellular staining for PLZF, RORgt, and T-BET was performed using the Foxp3/Transcription required SAP, indicating that PLZF expression was a consequence Factor Staining Buffer Set (00-5523-00; eBioscience), and Fixable Viabil- of selection into the innate T lymphocyte lineage rather than being ity Dye eF506 (65-0866; eBioscience) was added to exclude dead cells. a consequence of deregulated PLZF expression (Supplemental Fig. Allophycocyanin- or PE-conjugated Cd1d Tet, loaded with PBS57, were obtained 1A, 1B). The majority of Id32/2PLZF+ PS DP thymocytes were from the National Institutes of Health Tetramer Facility at Emory University. detected with Cd1dPBS57-loaded Tet, indicating that these cells were In vivo BrdU incorporation iNKT cell progenitors (Fig. 1D, 1E). Therefore, ID3 limited the number of DP thymocytes recognizing CD1d and undergoing SAP- Mice were injected i.p. with BrdU (1 mg/6g) 12 h prior to euthanasia. For dependent selection into the NKT cell pathway. pulse-chase experiments, BrdU was added to the drinking water for 4 d (0.8 mg/ml). Water was changed daily. Subsequently, BrdU was removed for the ID3 limited Va14Ja18 recombination in PS DP thymocytes indicated time, and thymocytes were stained for surface markers and BrdU, according to the manufacturer’s instructions (BD Biosciences). Previous studies showed that E proteins maintain the DP gene “identity” (20, 37) and regulate genes involved in V(D)J recom- Adoptive transfers of bone marrow bination (Rag1 and Rag2) and cell survival (Rorc). Given that Id3 2 2 Bone marrow (BM) single-cell suspensions from Id3 / CD45.2 mice is induced by TCR signaling during positive selection (24), we were prepared from femur and tibiae and were mixed in 1:1 ratio with hypothesized that ID3-deficient PS DP thymocytes maintained competitor wild-type (WT) CD45.1 BM cells. Recipient mice were le- thally irradiated (1000 rad) 5 h prior to i.v. injection of total 5 3 106 expression of these E protein target genes and therefore continued cells. Thymocytes were stained for flow cytometry 6–8 wk posttran- to rearrange their Ag receptors. To test this hypothesis, we ex- 2 2 splantation. amined Rorc, Rag1, and Rag2 mRNA in WT and Id3 / PS DP The Journal of Immunology 5975

FIGURE 1. ID3 limited the number of PLZF+ and Tet+ cells among PS DP thymocytes. (A) qPCR analysis for Zbtb16 (PLZF) mRNA in sorted WT and Id32/2 TCRgd2CD4+CD8+TCRb+CD69+ thymocytes. The y-axis indicates Zbtb16 mRNA relative to Hprt mRNA analyzed in the same samples. The graphs show the mean 6 SD of five independent experiments. n =5.*p , 0.05, two-tailed unpaired t test. (B) Histograms showing expression of PLZF in gated TCRgd2CD4+CD8+TCRb+CD69+ (black line, post-DP) or TCRgd2CD4+CD8+TCRb2CD692 cells (gray line, pre-DP) thymocytes from the in- dicated mouse strains. The number above the line shows percentage of PLZF+ cells among PS DP cells. (C) Average number and percentage of PLZF+ Downloaded from among TCRgd2CD4+CD8+TCRb+CD69+ thymocytes. The graphs show mean 6 SD of eight independent experiments. n =8.*p , 0.05, **p , 0.01. (D) Flow cytometry plots showing PLZF and Tet staining on TCRgd2CD4+CD8+TCRb+CD69+ thymocytes from mice of the indicated genotype. (E) Graphs show the average number and percentage of Tet+ PS DP cells among thymocytes in WT and Id32/2 mice. The graphs show mean 6 SD of 10 independent experiments. n = 10. *p , 0.05, ***p , 0.001, two-tailed paired t test.

thymocytes. Despite apparent TCR signaling, Rorc, Rag1,andRag2 ID3 limited the number of iNKT cells http://www.jimmunol.org/ 2/2 mRNAs were expressed more abundantly in Id3 than in WT Our data indicated that Id32/2 PS DP thymocytes continued to cells (Fig. 2A). At the single-cell level, RORgt was detected in a rearrange the Ja locus and thereby produced more cells with a subset of Id32/2 PS DP thymocytes (Fig. 2B, 2C). Unlike PLZF, + TCR capable of selection into the NKT cell lineages. Consistent however, RORgt cells persisted in the absence of SAP, indicating with this conclusion, there was a 3-fold increase in the frequency that Rorc was not induced as a consequence of NKT cell selection and 1.8-fold increase in the number of iNKT cells in Id32/2 mice (Supplemental Fig. 1C). Our results indicated that a subset of 2/2 compared with WT littermates (Fig. 3A, 3B). The increased number Id3 PS DP thymocytes maintained expression of the E protein of iNKT cells was an intrinsic consequence of ID3 deficiency as target genes Rorc, Rag1, and Rag2. revealed in mixed BM chimeras (Fig. 3C, 3D). Relative to DP We hypothesized that sustained RAG1 and RAG2 expression in 2/2 ∼ by guest on September 27, 2021 2/2 thymocytes, Id3 iNKT cells were increased by 1.6-fold, Id3 PS DP thymocytes increased the probability of secondary whereas WT iNKT cells decreased by 40% in the same chimera rearrangements, including the Va14-Ja18 rearrangement used in (Fig. 3E, Supplemental Fig. 1F). As expected (23), Id32/2 CD4 the iNKT cell TCR. Consistent with this hypothesis, there was 2/2 and CD8 thymocytes decreased by 35 and 55%, respectively, with a 4.5-fold increase in Va14Ja18 mRNA in Id3 PS DP thy- a concomitant increase in WT CD4 and CD8 thymocytes (Fig. 3F, mocytes compared with WT littermates (Fig. 2D). In addition, 2/2 2/2 Supplemental Fig. 1F). Therefore, Id3 DP thymocytes pro- Id3 PS DP cells had a 2.5-fold increase in the frequency of duced more iNKT cells and fewer conventional thymocytes than cells with a Va14-Ja18 rearrangement (Fig. 2E). We detected their WT counterparts in mixed BM chimeras. The increased fre- increased usage of distal Ja transcripts and decreased usage of quency of Id32/2 iNKT cells was not a consequence of altered proximal Ja transcripts, suggesting that ID3 limited the V(D)J expression of the selection-associated proteins CD1d, SLAM, recombination process (Fig. 2F). ID3 deficiency could have pro- LY108, or SAP (Fig. 3F, 3G). Therefore, ID3 deficiency led to an longed the survival of DP thymocytes, resulting in increased distal intrinsic enhancement of iNKT cell numbers. rearrangements. However, we did not detect an increase in Va14- Ja18 rearrangement in preselection DP cells, and Id32/2 prese- lection and PS DP cells did not show prolonged survival in vivo, ID3 was required for normal development of stage 3 iNKT as measured by BrdU dilution (Supplemental Fig. 1D, 1E). More- (NKT1) cells over, mRNA encoding prosurvival proteins such as Bclxl or Bcl2, Our data indicated that ID3 limited iNKT cell selection; however, which are targets of RORgt, were similar in WT and Id32/2 pre- we questioned whether ID3 might also have functioned in iNKT selection and PS DP thymocytes (data not shown). cell maturation or effector fate determination. When stimulated To assess whether the increased frequency of cells with a Va14- with PMA and ionomycin, a subset of WT iNKTs produced IFN-g Ja18 rearrangement could contribute to the increased number of or IFN-g plus IL-4, whereas only a few cells produced IL-4 without iNKT cells, we introduced a CD4 promoter–driven, prerearranged IFN-g. In contrast, a lower frequency of Id32/2 iNKT cells pro- Va14Ja18 TCR (Va14 transgene (9)) into the ID3-deficient back- duced IFN-g alone, and an increased frequency of cells produced ground. By removing the need for recombination of the TCRa- only IL-4 (Fig. 4A, 4B). Consistent with the decline in IFN-g– chain, we should have overcome any bias in iNKT cell develop- only producing cells, there was a decrease in stage 3 iNKT cell ment provided by altered Va14-Ja18 rearrangement. We found number and frequency, as measured by multiple surface markers, that Id3+/+ and Id32/2 Va14 transgenic mice had a similar fre- which are the major IFN-g–only producing iNKT cells in WT mice quency and absolute number of iNKT cells in the thymus (Fig. 2G, (Fig. 4C, 4D, Supplemental Fig. 2A). By contrast, the frequency 2H). These data support the hypothesis that the increased number and number of Id32/2 stage 1 and stage 2 cells increased (Fig. 4C, of iNKT cells in Id32/2 mice was a consequence, at least in part, 4D). The reduction of stage 3 iNKT cells and the accumulation of extended Va14-Ja18 recombination. of stage 1 and 2 iNKT cells also were apparent in mixed BM 5976 ID PROTEINS CONTROL iNKT CELL DEVELOPMENT Downloaded from http://www.jimmunol.org/

FIGURE 2. Id32/2 PS DP thymocytes had more Va14Ja18 recombination events. (A) qPCR analysis of Rorc, Rag1, and Rag2 mRNA in sorted CD4+ CD8+TCRb+CD69+ thymocytes from WT and Id32/2 mice. The graphs show mean 6 SD of five independent experiments. n =5.*p , 0.05, two-tailed 2 2 unpaired t test. (B)RORgt in gated PS (black line) or preselection (gray line) DP thymocytes from WT and Id3 / mice. The shaded histogram shows by guest on September 27, 2021 RORgt expression in CD8+TCRb+ cells, as a negative control. The number above the line shows percent of RORgt+ cells among PS DP cells. (C) Graph shows the average percentage of RORgt+ cells among PS DP cells in WT and Id32/2 mice. n = 8. ***p , 0.001. qPCR analysis of Va14Ja18 mRNA (D) and Va14Ja18 DNA (E) rearrangements in sorted CD4+CD8+TCRb+CD69+ thymocytes from WT and Id32/2 mice. Numbers show change in detection relative to WT. The graphs show the mean 6 SD from five independent experiments. n =5.*p , 0.05, two-tailed unpaired t test. (F) qPCR analysis of Ja transcripts containing Ja gene segments that are proximal and distal to Va in TCRgd2CD4+CD8+TCRb+CD69+ thymocytes sorted from WT and Id32/2 mice. Numbers show change in detection relative to WT for each Ja transcript. The graphs show the mean 6 SD from three independent experiments. n =3. *p , 0.05, **p , 0.01, two-tailed unpaired t test. (G) Flow cytometry plots of total thymocytes from Id3+/+;Va14Tg and Id32/2;Va14Tg mice showing the percentage of Tet+TCRb+ iNKT cells. (H) Average number and percentage of iNKT cells in the indicated mouse strains. The graphs show the mean 6 SD from 10 independent experiments; n = 10. chimeras (Fig. 4E, 4F). Id32/2 stage 3 cells did not stain with cells. However, WT and Id32/2 stage 1 and 2 iNKT cells were annexin V, and they expressed the CXCR3, labeled similarly with BrdU during a 12-h pulse, suggesting that which functions in thymic retention (Supplemental Fig. 2B, 2C) these cells have an equivalent rate of proliferation (Supplemental (38), indicating that these cells are not apoptotic and they should Fig. 2E, 2F). We also considered the possibility that arrested matu- be retained in the thymus. Conversely, stage 2 iNKT cells did not ration contributed to the increased number of stage 1 and stage 2 iNKT express CXCR3, which could have led to their thymic retention. cells. However, Tbx212/2 iNKT cells, which also fail to generate stage Indeed, Id32/2 stage 2 iNKT cells were capable of exiting the 3 iNKT cells (18, 19), did not have more total iNKT cells (Supple- thymus because iNKT cells were present in the periphery of Id32/2 mental Fig. 2G, 2H), arguing that arrested maturation per se does not mice, although their numbers were decreased (Fig. 5A, 5B). More- augment iNKT cell numbers. Therefore, in addition to preventing over, there were few peripheral NK1.1+ stage 3 cells, indicating PS recombination of the iNKT TCR, ID3 promoted iNKT cell that maturation also may be impaired in the periphery (Fig. 5C, maturation to stage 3 but had little impact on NKT cell proliferation.

5D). We conclude that ID3 was required intrinsically for the de- + velopment of stage 3 iNKT cells. We note that the decreased fre- ID3 supported the progenitors of TBET stage 3 iNKT (NKT1) quency of IFN-g–only producing iNKT cells is less severe than the cells decreased frequency of stage 3 iNKT cells, which could be a con- PLZF and TBET are critical regulators of iNKT cell specification sequence of increased expression of EOMES in stage 2 iNKT cells and maturation (11, 12, 39). Intracellular staining and flow cy- (Supplemental Fig. 2D). tometry revealed higher PLZF (Fig. 6A, left panel)andlower Proliferation of iNKT cells occurs primarily at the immature TBET expression in Id32/2 iNKT cells as compared with WT stages. Therefore, the increased iNKT cell frequency or the decline in (Fig. 6B). However, PLZF was expressed similarly in WT and stage 3 iNKT cells could be because of increased proliferation of these Id32/2 stage 1 and stage 2 iNKT cells (Fig. 6A). Therefore, the The Journal of Immunology 5977

FIGURE 3. Id32/2 mice had increased iNKT cell numbers. (A) Flow cytometry plots of total thymocytes from WT and Id32/2 mice showing the percentage of iNKT cells indentified by CD1dPBS57-Tet and TCRb staining. (B) Average percentage and number of total iNKT thymocytes in the indicated mouse strains. n = 20 for each genotype. The graphs show the average of .15 independent experiments. Error bars indicate SD. *p , 0.05, ***p , 0.001. (C) Flow cytometry plots showing the percentage of Tet+TCRb+ iNKT cells among CD45.1+ (WT) and CD45.2+ (Id32/2) cells in competitive BM chimeras. (D) Average percentage of Tet+TCRb+ iNKT cells among CD45.1+ (WT) and CD45.2+ (Id32/2) cells. n = 8 chimeric mice. Error bars indicate SD. *p , 0.05, two-tailed paired t test. (E) Analysis of DP, iNKT, CD4, and CD8 cells in competitive BM chimeras shown as the percentage of WT or Id32/2 DP cells. n = 8. Error bars indicate SD. ***p , 0.001, two-tailed paired t test. (F) Histograms showing expression of CD150, LY108, and CD1d in Downloaded from DP thymocytes from WT and Id32/2 mice. Solid gray histogram indicates isotype control. (G) RT-QPCR analysis of Sh2d1a mRNA (encoding for the adaptor protein SAP) in CD42CD82 and CD4+CD8+ sorted thymocytes from WT and Id32/2 mice. http://www.jimmunol.org/ increased PLZF was a consequence of the loss of stage 3 iNKT absence of ID3 (Fig. 6B). Overall, we detected a 3.5-fold decrease cells, which typically express lower PLZF (12). In contrast to in the number of TBET+ cells (Fig. 6C), consistent with the re- PLZF, TBET expression at stage 2 was severely reduced in the duction of stage 3 cells. Importantly, Tbx21 (encoding TBET) by guest on September 27, 2021

FIGURE 4. ID3 was required for iNKT cell maturation. (A) Intracellular staining for IFN-g and IL-4 in gated Tet+CD242 cells after in vitro culture with (lower panel)orwithout(upper panel) PMA and ionomycin for 5 h. Numbers indicate percentage of cells in each quadrant. (B) Average percentage and number of iNKT thymocytes expressing the indicated cytokines from WT and Id32/2 mice. The graphs show the mean 6 SD from four independent experiments. n =4 for each genotype. *p , 0.05, **p , 0.01. (C) WT and Id32/2 thymocytes were analyzed for Tet+CD242 iNKT cells (upper panel) and CD44 and NK1.1 expression is shown in gated Tet+CD242 iNKT cells (lower panel). The numbers in plots indicate the percentage of cell subsets in the respective gates. (D) Average percentage and number of stage 1, stage 2, and stage 3 iNKT cells in WT and Id32/2 thymi. The graphs show the mean 6 SD from .15 independent experiments. n = 19. ***p , 0.001. (E) Chimeric mice shown in Fig. 1 were analyzed for Tet+CD242 thymocytes (upper panel) and surface expression of CD44 and NK1.1 (lower panel). (F) Average percentage of stage 1, stage 2, and stage 3 iNKT cells in CD45.1+ (WT)andCD45.2+ (Id32/2) cells from the competitive BM chimeras. The graphs show the mean 6 SD from eight independent experiments. n =8.*p , 0.05, ***p , 0.001, two-tailed paired t test. 5978 ID PROTEINS CONTROL iNKT CELL DEVELOPMENT

FIGURE 5. Id32/2 peripheral iNKT cells failed to mature. (A) Flow cytometry plots showing the per- centage of Tet+TCRb+ iNKT cells in the spleen and liver from WT and Id32/2 mice. (B) The graphs show the average number and percentage of iNKT cells in the spleen and liver of the indicated mouse strains. The graphs show mean 6 SD from 8 to 10 independent experiments. *p , 0.05, **p , 0.01. (C) Histograms showing the expression of NK1.1 on total iNKT cells in the spleen and liver of WT and Id32/2 mice. The shaded histogram is the isotype control. The numbers indicate percentage of NK1.1+ cells among iNKT cells. (D) The graphs show the average percentage of NK1.1+ cells among the iNKT population in the spleen and liver of the indicated mouse strains. The graphs Downloaded from show mean 6 SD from 8 to 10 independent experi- ments. *p , 0.05, **p , 0.01. http://www.jimmunol.org/ mRNA could be detected in WT stage 1 iNKTs and was increased (Fig. 6D), suggesting that Tbx21 expression is perturbed at the at stage 2. In contrast, there was a 2- and a 4-fold reduction of earliest stages of iNKT maturation. Tbx21 is part of a broader gene Tbx21 mRNA in ID3-deficient stage 1 and 2 iNKTs, respectively expression program that is implemented over the course of the by guest on September 27, 2021

FIGURE 6. Id32/2 iNKT cells failed to induce the stage 3–associated transcrip- tional program. Histograms showing PLZF (A)orTBET(B)intotal(left), stage 1 (middle),andstage2(right) iNKT cells from WT (gray line) and Id32/2 (black line) mice. The shaded histogram is the expres- sion of the respective protein in conventional CD4+ thymocytes, which are a negative control. (C) Average number of TBET+ iNKT cells in the thymus. The graph shows the mean 6 SD from 12 independent experiments. n = 12. ***p , 0.001. (D) qPCR analysis of Tbx21 mRNA in sorted stage 1 and 2 iNKT thymocytes from WT and Id32/2 mice. The y-axis indicates change relative to expression in WT stage 1. The graph shows the mean 6 SD from three independent experiments. Two mice for each mouse strain were pooled for each experiment. n =3.**p , 0.01, ***p = 0.001, two-tailed unpaired t test. (E) A set of 70 genes that are progressively upregu- lated during the transition from WT stage 1 through stage 3 was identified by analyzing normalized data from the Immunological Genome Consortium. The expression of each gene in WT and Id32/2 stage 1 and stage 2 iNKT cells is shown as a heat map. The Journal of Immunology 5979 stage 1 to 3 transition (40). Microarray analysis revealed that Va14Ja18 mRNA transcripts in Id2D/D;Id32/2 PS DP thymocytes, .70% of the genes that are progressively upregulated from stage compared with the LMC control mice (Fig. 7F). Our results demon- 1 to stage 3 failed to initiate expression during the stage 1 to 2 strated that the failure to express ID2 and ID3 during positive selection transition in Id32/2 iNKT cells (Fig. 6E). Given that Id3 is altered the TCR repertoire, allowing for accumulation of iNKT cells. expressed more highly in stage 1 and 2 iNKT cells than in stage 3 Redundant functions for ID2 and ID3 during iNKT cell lineage cells (Supplemental Fig. 3A), we concluded that iNKT cells rely specification on ID3 to support the emergence of the stage 3–associated program at the earliest stages of iNKT cell maturation. We next examined the consequences of Id2 and Id3 deletion on iNKT cell maturation. Remarkably, whereas Id32/2 iNKT cells ID2 and ID3 cooperated to limit iNKT cell selection were arrested at stage 2, Id2D/D;Id32/2 iNKT cells lacked CD44hi ID3 was required for the initiation of the stage 3 gene program early stage 2 and stage 3 cells (Fig. 8A). The total number and frequency during iNKT cell maturation. However, Id2 and Id3 mRNA were of stage 1 iNKT cells was increased by 85- and 10-fold, respec- coexpressed in stage 1 and stage 2 iNKT cells (Supplemental Fig. tively, in Id2D/D;Id32/2 compared with LMC mice (Fig. 8B). In vitro 3A). Although ID2 deficiency has no effect on thymic iNKT cell stimulation of Id2D/D;Id32/2 iNKT cells revealed impaired IFN-g, development (Supplemental Fig. 3B, 3C) (27), we hypothesized but normal IL-4 production and the cells failed to coproduce IFN-g that ID2 may have compensated for some ID3 functions. In ad- and IL-4 (Fig. 8C), a trait of iNKT cells that is conferred by PLZF. dition, ID3 may have compensated for the loss of ID2 thereby Consistent with this, PLZF protein expression was reduced at least masking the requirements for ID proteins in iNKT cells. To ad- 2-fold in Id2D/D;Id32/2 stage 0 and stage 1 iNKT cells relative to f/f 2/2 D/D D/D dress this possibility, we created CD4-Cre;Id2 ;Id3 (Id2 ; LMC (Fig. 8D, 8E), whereas it was expressed in stage 1 Id2 and Downloaded from Id32/2) mice that lacked both Id2 and Id3 starting in DP thy- Id32/2 iNKT cells (Fig. 6, Supplemental Fig. 4A). Importantly, mocytes. Analysis of Id2D/D;Id32/2 mice revealed a reduced Id2D/D;Id32/2 stage 0 and stage 1 iNKT cells showed a 5- and 2.5- number of CD4 and CD8 T cells, consistent with published results fold decrease, respectively, in Zbtb16 mRNA compared with their (21). However, .30% of Id2D/D;Id32/2 CD4 T cells were iNKT LMC controls (Fig. 8D, 8E). Therefore, ID proteins are required to cells, compared with only 2% of their littermate controls (LMC) support expression of Zbtb16 in iNKT cell progenitors and to allow

(Supplemental Fig. 3D). The frequency and absolute number of their specification to the NKT cell fate (Fig. 8E). These data are in http://www.jimmunol.org/ iNKT cells among total thymocytes was increased by an average accordance with previous studies showing that haploinsufficiency of of 7- and 5-fold, respectively, in Id2D/D;Id32/2 mice compared PLZF leads to a naive-like iNKT phenotype (41). Importantly, Tet+ with LMC (Fig. 7A, 7B). This increase was significantly larger cells failed to accumulate in the spleen and the liver of Id2D/D;Id32/2 than the increase observed in Id32/2 mice (Fig. 7C). mice, and the few cells that were present completely lacked NK1.1 In Id2D/D;Id32/2 thymi, there was a substantial increase in the (Fig. 8F, 8G, Supplemental Fig. 4B). In contrast, Tet+ cells were frequency of PS DP cells (Fig. 7D). Moreover, there was a 6-fold detected in higher frequency in the lymph nodes of the compound increase in the percentage and a 4-fold increase in the number of mutant mice (Supplemental Fig. 4C, 4D); however, the majority Tet+ PS DP thymocytes in Id2D/D;Id32/2 mice (Fig. 7E). Consistent showed a CD44loCD62L+ phenotype, indicative of a naive state. with these results, we observed a 9-fold increase in the canonical The failure to induce PLZF appeared to be independent of EGR2, by guest on September 27, 2021

FIGURE 7. Increased iNKT cell development in the absence of both Id2 and Id3.(A) Flow cytometry plots of total thymocytes from Id2f/f and CD4Cre; Id2f/f;Id32/2 (Id2D/D;Id32/2) mice showing the percentage of Tet+TCRb+ iNKTs. (B) Average percentage and number of total iNKT thymocytes in the indicated mouse strains. (C) Fold change in iNKT percentage (left) and cell numbers (right)ofId32/2 and Id2D/D;Id32/2 relative to LMC. The graphs show the mean 6 SD of 10 independent experiments. n = 10. *p , 0.05, **p , 0.01. (D) TCRb+CD69+ (PS) thymocytes from the indicated mouse strains were analyzed for CD4 and CD8 expression (upper panel), and DP cells were then analyzed for CD1dPBS57-Tet staining. The numbers in the plots indicate the percentage of DP cells among PS thymocytes (upper panel), and the percentage of Tet+ cells among PS DP cells (lower panel). (E) The average number of CD4+CD8+TCRb+CD69+Tet+ (upper graph) and percentage of Tet+ thymocytes among PS DP cells (lower graph). The graphs show the mean 6 SD of 10 independent experiments. n = 10. *p , 0.05, **p , 0.01. (F) qPCR analysis of Va14Ja18 mRNA (in sorted CD4+CD8+TCRb+CD69+ thymocytes from LMC and Id2D/D;Id32/2 mice. The graphs show the mean 6 SD of three independent experiments. n =3.*p , 0.05, two-tailed unpaired t test. 5980 ID PROTEINS CONTROL iNKT CELL DEVELOPMENT Downloaded from http://www.jimmunol.org/

FIGURE 8. Id2 and Id3 regulate iNKT cell specification. (A) Id2f/f and Id2D/D;Id32/2 thymocytes were analyzed for Tet+CD242 iNKT cells (upper panel). The percentage of Tet+CD242 cells is indicated. CD44 and NK1.1 expression is shown in gated Tet+CD242 iNKT cells (lower panel). The numbers indicate the percentage of cells in the respective quadrants. (B) The average cell number and percentage of stage 1 iNKT cells in the indicated mouse strains. The graphs show the mean 6 SD of eight independent experiments. n = 8. ***p , 0.001. (C) Total thymocytes from Id2f/f and Id2D/D;Id32/2 mice + 2 were cultured in vitro with (lower panel) or without (upper panel) PMA and ionomycin for 5 h. Intracellular staining of IFN-g and IL-4 in gated Tet CD24 by guest on September 27, 2021 cells is shown. The numbers indicate the percentage of cells in each quadrant. (D) The histogram (left panel) shows PLZF expression in stage 1 iNKT cells from Id2f/f (gray line) and Id2D/D;Id32/2 (black line) mice. Solid gray, PLZF expression from conventional CD4 thymocytes from littermate control mice. The graph (middle panel) shows the average mean fluorescence intensity (MFI) from 10 independent experiments. n = 12. The graph (right panel) shows qPCR analysis of Zbtb16 mRNA in sorted stage 1 (Tet+CD242CD442NK1.12) iNKT thymocytes from LMC and Id2D/D;Id32/2 mice. The numbers on the y-axis indicate fold change relative to LMC stage 1 expression. The graphs show the mean 6 SD from three independent experiments. n =3.Middle panel, *p , 0.05, right panel,**p , 0.01, two-tailed unpaired t test. (E) The histogram (left panel) shows PLZF expression in stage 0 (Tet+CD24+CD69+CD442) iNKT cells from Id2f/f (gray line) and Id2D/D;Id32/2 (black line) mice. Solid gray, PLZF expression from conventional CD4 thymocytes from littermate control mice. The histogram is representative of two independent experiments. The graph (right panel) shows qPCR analysis of Zbtb16 mRNA in sorted stage 0 (Tet+CD24+CD69+CD442) iNKT thymocytes from LMC and Id2D/D;Id32/2 mice. The numbers on the y-axis indicate fold change relative to LMC stage 1 expression. The graphs show the mean 6 SD from three independent experiments. n =3.**p , 0.01, two-tailed unpaired t test. Flow cytometry plots showing lymphocytes from the spleen (F) and liver (G) of the indicated mouse strains analyzed for Tet+TCRb+ cells (upper panel). The percentage of Tet+TCRb+ cells is indicated. Histograms showing the percent of NK1.1+ iNKTs (lower panel). Solid gray indicates the isotype control. a direct transcriptional regulator of Zbtb16, because Egr2 mRNA limited selection into the iNKT cell pathway and total iNKT cell and EGR2 protein were increased in Id2D/D;Id32/2 iNKT cells numbers by extinguishing the potential for secondary rearrange- (Supplemental Fig. 4E, 4F). Moreover, Id2D/D;Id32/2 stage 1 iNKT ments at Tcra during conventional T cell–positive selection. In cells incorporated less BrdU during a 12-h pulse than their LMC contrast, postpositive selection, ID proteins were essential for (Supplemental Fig. 4G), indicating that these cells were not in- iNKT cell lineage specification and maturation in the thymus. ID2 creased as a consequence of increased proliferation. These data and ID3 redundantly promoted expression of PLZF and entry of revealed that deletion of Id2 and Id3 in DP thymocytes resulted in Tet+ DP cells into the iNKT cell lineage. After iNKT cell lineage more PS DP thymocytes with a CD1dPBS57-Tet binding TCR; specification, ID3 was required for the emergence of TBET+ stage however, these cells failed to induce sufficient PLZF to progress 3 iNKT cells. Our data demonstrated that ID protein dependency beyond stage 1 or to acquire innate-like characteristics. Moreover, is a common requirement for all innate lymphoid lineages and these cells resembled conventional CD4 T cells in that they ac- revealed that alteration of ID protein dose modulates the effector cumulated in lymph nodes rather than the liver and had a naive subset composition within the iNKT cell lineage. phenotype. Therefore, ID2 and ID3 functioned redundantly to specify The development of iNKT cells depends on the lifespan of DP the iNKT cell fate. thymocytes. RORgt- or HEB deficiency, which leads to decreased Rorc expression and DP survival, results in a loss of iNKT cells Discussion because of a failure to rearrange Va14 and Ja18, a phenotype We have shown that ID proteins are major regulators of at least that is rescued by transgenic expression of the survival proteins three critical checkpoints in iNKT cell development. ID proteins BCLxL or BCL2 or a rearranged Va14Ja18 TCR (6, 7). We ob- The Journal of Immunology 5981 served that RORgt was maintained in a subset of Id32/2 PS DP ical regulator of iNKT cell specification, EGR2 directly regulates thymocytes. Nonetheless, Bclxl and Bcl2 were not expressed, and the transcription of both Id2 and Zbtb16 in developing iNKT cells we found no evidence for increased survival of DP thymocytes in (48). In Id2D/D;Id32/2 Tet+ cells, EGR2 protein was present even Id32/2 mice. Therefore, we propose that the increased number of though Zbtb16 failed to be induced, indicating that ID proteins iNKT cells in Id32/2 and Id2D/D;Id32/2 mice was largely a conse- were required for additional components of the transcriptional quence of the failure of PS DP thymocytes to extinguish expression program leading to Zbtb16 induction. Furthermore, Id2D/D;Id32/2 of Rag1 and Rag2. These cells continued to rearrange Tcra until Tet+ cells were present in the lymph nodes rather than the liver and aTCRa-chain was generated, such as Va14Ja18, that allowed spleen, a phenotype consistent with the loss of PLZF and the selection into the iNKT cell lineage. Consistent with this hypoth- innate T cell identity. Collectively, our results indicate that ID esis, Va14Ja18 TCR rearrangements and Tet+ cells were increased proteins control the transcriptional network required for entry into among Id32/2 and Id2D/D;Id32/2 PS DP thymocytes. Moreover, the iNKT cell lineage. expression of a Va14Ja18 TCR extinguished the advantage of After specification to the iNKT cell lineage Id3 mRNA declines, Id32/2 T cells to generate iNKT cells as compared with Va14Ja18 whereas Id2 mRNA increases. Given that baseline Id3 mRNA transgenic Id3+/+ T cells. The accumulation of iNKT cells in Id32/2 expression is a gauge for E protein activity, we propose that E or Id2D/D;Id32/2 mice is unlikely to be a consequence of arrested protein activity also declines during iNKT cell maturation. Sur- maturation because both PLZF-deficient and TBET-deficient mice, prisingly, despite the decline of Id3 mRNA in stage 3 iNKT cells, which show altered maturation similar to Id2D/D;Id32/2 and Id32/2 ID3 was required for the emergence of this population. The lack of iNKT cells, respectively, do not have more thymic iNKT cells. Id3 mRNA expression in these cells indicates that ID3 does not

Although we cannot exclude additional roles for ID proteins during directly support the survival of these cells or the maintenance of Downloaded from the NKT lineage selection, our results are consistent with the their transcriptional program. It seems unlikely that the increase in conclusion that ID proteins limited the ability of DP thymocytes stage 2 iNKTs in the thymus of Id32/2 mice was due to defective with a functional TCR to undergo secondary recombination events thymic egress because iNKT cells were readily detected in the that increased the number of iNKT cells. periphery, and CXCR3 was expressed on the appropriate iNKT One implication of our findings is that ID3 is not required for cell subsets. Nonetheless, Id32/2 iNKT cells failed to initiate ex-

positive selection of cells with an iNKT cell receptor. Consistent pression of a cohort of genes associated with stage 3 including http://www.jimmunol.org/ with this, Id3 mRNA showed significantly less induction in stage multiple NK cell receptors. Moreover, the transcription factor TBET, 0 iNKT cells than in conventional PS DP thymocytes (Supple- which controls the transcriptional program of stage 3 iNKT cells mental Fig. 3A). ID3 is a target of the ITK/RAS/ERK kinase (18), was severely reduced in stage 1 and stage 2 Id32/2 iNKT cells. signaling pathway, which is activated during positive selection Therefore, ID3 may support the development of a precursor to stage into both the conventional and the iNKT cell pathway (42). Why 3 or it may promote a critical regulator of the stage 3 gene program. the signals initiated during iNKT cell selection fail to strongly Surprisingly, the requirement for ID3 is not absolute in pe- induce Id3 mRNA remains to be determined. Although Itk2/2 and ripheral iNKT cells, where there was only a modest reduction of Id32/2 mice have similar alterations in iNKT cell maturation, NK1.1+ cells. However, Id2 expression is higher in splenic and Itk2/2 mice do not have more iNKT cells (43), suggesting that liver iNKT cells than in thymic iNKT, whereas Id3 expression is by guest on September 27, 2021 ITK is not required for ID3 induction during positive selection. low in peripheral iNKT cells. These findings raise the possibility However, whether attenuation of ITK activation limits Id3 mRNA that in the periphery ID2 performs the functions attributed to ID3 induction during iNKT cell specification remains to be deter- in the thymus. Consistent with this idea, ID2 has been implicated mined. An alternative possibility is that the SLAM/SAP signaling in the survival of hepatic iNKT cells (27). Therefore, we propose pathway attenuates Id3 expression in the face of strong TCR sig- that the requirements for ID2 and ID3 differ in the thymus and the naling during iNKT cell lineage specification. b-Catenin, which periphery due to their pattern of expression but that ID proteins are is activated by TCR signaling and can inhibit expression of both necessary for NKT cell specification and effector fate determi- Id2 and Id3 mRNA in thymocytes (44), may play a role in Id3 nation. Indeed, in the absence of both ID proteins, thymic iNKT cells attenuation during iNKT cell lineage specification because mice lose their innate properties and the peripheral iNKT cell pool is se- expressing an activated form of b-catenin in T lymphocytes verely diminished and functionally compromised. phenocopy Id32/2 mice (45). However, it remains to be deter- We demonstrated that ID2 and ID3 play a critical role in con- mined whether the increased number of iNKT cells in these mice trolling the number of iNKT cells and their effector repertoire. Our is a consequence of decreased ID3 expression in PS DP thymocytes. results imply that E protein function needs to be tightly regulated Regardless, ID proteins may serve as integrators of multiple sig- during iNKT cell development because deletion of both Id2 and naling pathways to regulate the output of thymic selection. Id3 relaxes the constraints on E protein activity. Whether E pro- Specification of DP thymocytes into the iNKT cell lineage is teins are required for any functions in iNKT cells remains to be regulated by PLZF (11, 12). We showed in this study that in- determined; however, it seems likely that ID proteins restrain but duction of PLZF in iNKT cells requires either ID2 or ID3. ID2 is do not totally ablate E protein function. Therefore, an increase in tightly associated with innate lymphoid lineages and is absolutely E protein function could augment expression of genes that are required for development of all non-T lineage ILC subsets (26). normally regulated by E protein homodimers in NKT cells. Alter- Nonetheless, ID2 deficiency had little effect on thymic iNKT cell natively, an increase in E protein concentration could deregulate development (27), suggesting that iNKT cells differ significantly genes with low-affinity E boxes. For example, E protein homo- from other ILC populations in their requirement for ID2. In con- dimers could antagonize expression of genes that are regulated by trast to this view, we propose that ID2 is a critical driver of the class II basic helix-loop-helix transcription factors because these iNKT cell fate but that ID3 compensates for the loss of ID2 during proteins have subtly different consensus binding sequences and iNKT cell lineage specification. Previous studies demonstrated therefore may have distinct gene target specificity when present at that Id3 mRNA is induced by E proteins and is increased in many low doses (22). E proteins have been proposed to interact with the instances of ID2 deficiency (46, 47), raising the possibility that helix-loop-helix protein BHLHE40, a transcriptional repressor ID3 is critical for iNKT cell specification only when ID2 fails to that is part of the core effector program of iNKT cells and there- be expressed. Consistent with the hypothesis that ID2 is the crit- fore increased E protein concentration may augment or compete 5982 ID PROTEINS CONTROL iNKT CELL DEVELOPMENT for BHLHE40 functions (40, 49). Regardless of the mechanism, 19. Townsend, M. J., A. S. Weinmann, J. L. Matsuda, R. Salomon, P. J. Farnham, C. A. Biron, L. Gapin, and L. H. Glimcher. 2004. T-bet regulates the terminal our results demonstrate that tight regulation of E protein dose is maturation and homeostasis of NK and Va14i NKT cells. Immunity 20: 477–494. critical for iNKT cell development and effector differentiation. 20. Jones, M. E., and Y. Zhuang. 2007. Acquisition of a functional T cell receptor We and others have shown previously that cytokine production during T lymphocyte development is enforced by HEB and E2A transcription hi factors. Immunity 27: 860–870. by PLZF NKT cells can influence the fate of bystander CD8 21. Jones-Mason, M. E., X. Zhao, D. Kappes, A. Lasorella, A. Iavarone, and T cells and the production of IgE (31, 50, 51, 52). Therefore, Y. Zhuang. 2012. E protein transcription factors are required for the development + a minor difference in the number and effector composition of of CD4 lineage T cells. Immunity 36: 348–361. 22. Kee, B. L. 2009. E and ID proteins branch out. Nat. Rev. Immunol. 9: 175–184. NKT cells can have widespread effects on the immune response. 23. Rivera, R. R., C. P. Johns, J. Quan, R. S. Johnson, and C. Murre. 2000. Thy- Our results show that ID proteins have multiple roles in deter- mocyte selection is regulated by the helix-loop-helix inhibitor protein, Id3. Immunity 12: 17–26. mining iNKT cell number and effector fate and further reveal 24. Bain, G., C. B. Cravatt, C. Loomans, J. Alberola-Ila, S. M. Hedrick, and a global role for the ID family of proteins in shaping the com- C. Murre. 2001. Regulation of the helix-loop-helix proteins, E2A and Id3, by the position of the innate immune system. Ras-ERK MAPK cascade. Nat. Immunol. 2: 165–171. 25. Engel, I., C. Johns, G. Bain, R. R. Rivera, and C. Murre. 2001. Early thymocyte development is regulated by modulation of E2A protein activity. J. Exp. Med. Acknowledgments 194: 733–745. We thank Dr. Albert Bendelac for helpful discussions, comments on the 26. Walker, J. A., J. L. Barlow, and A. N. McKenzie. 2013. Innate lymphoid cells— manuscript, and numerous reagents including the Va14 transgenic mice; how did we miss them? Nat. Rev. Immunol. 13: 75–87. 27. Monticelli, L. A., Y. Yang, J. Knell, L. M. D’Cruz, M. A. Cannarile, I. Engel, Liselotte Lenner for performing the microarray; and members of the Kee M. Kronenberg, and A. W. Goldrath. 2009. Transcriptional regulator Id2 con- laboratory for helpful discussions. This work benefited from data assem- trols survival of hepatic NKT cells. Proc. Natl. Acad. Sci. USA 106: 19461–19466. bled by the Immunological Genome Consortium. 28. Alonzo, E. S., R. A. Gottschalk, J. Das, T. Egawa, R. M. Hobbs, P. P. Pandolfi, P. Pereira, K. E. Nichols, G. A. Koretzky, M. S. Jordan, and D. B. Sant’Angelo. Downloaded from 2010. Development of promyelocytic zinc finger and ThPOK-expressing innate Disclosures gd T cells is controlled by strength of TCR signaling and Id3. J. Immunol. 184: The authors have no financial conflicts of interest. 1268–1279. 29. Ueda-Hayakawa, I., J. Mahlios, and Y. Zhuang. 2009. Id3 restricts the devel- opmental potential of gd lineage during thymopoiesis. J. Immunol. 182: 5306– 5316. References 30. Verykokakis, M., M. D. Boos, A. Bendelac, E. J. Adams, P. Pereira, and

1. Bendelac, A., P. B. Savage, and L. Teyton. 2007. The biology of NKT cells. B. L. Kee. 2010. Inhibitor of DNA binding 3 limits development of murine slam- http://www.jimmunol.org/ Annu. Rev. Immunol. 25: 297–336. associated adaptor protein-dependent “innate” gd T cells. PLoS One 5: e9303. 2. Van Kaer, L., V. V. Parekh, and L. Wu. 2013. Invariant natural killer T cells as 31. Verykokakis, M., M. D. Boos, A. Bendelac, and B. L. Kee. 2010. SAP protein- sensors and managers of inflammation. Trends Immunol. 34: 50–58. dependent natural killer T-like cells regulate the development of CD8+ T cells 3. Berzins, S. P., M. J. Smyth, and A. G. Baxter. 2011. Presumed guilty: natural with innate lymphocyte characteristics. Immunity 33: 203–215. killer T cell defects and human disease. Nat. Rev. Immunol. 11: 131–142. 32. Niola, F., X. Zhao, D. Singh, A. Castano, R. Sullivan, M. Lauria, H. S. Nam, 4. Park, S. H., A. Weiss, K. Benlagha, T. Kyin, L. Teyton, and A. Bendelac. 2001. Y. Zhuang, R. Benezra, D. Di Bernardo, et al. 2012. Id proteins synchronize The mouse CD1d-restricted repertoire is dominated by a few autoreactive T cell stemness and anchorage to the niche of neural stem cells. Nat. Cell Biol. 14: receptor families. J. Exp. Med. 193: 893–904. 477–487. 5. Engel, I., and M. Kronenberg. 2012. Making memory at birth: understanding the 33. Pan, L., S. Sato, J. P. Frederick, X. H. Sun, and Y. Zhuang. 1999. Impaired differentiation of natural killer T cells. Curr. Opin. Immunol. 24: 184–190. immune responses and B-cell proliferation in mice lacking the Id3 gene. Mol. 6. Egawa, T., G. Eberl, I. Taniuchi, K. Benlagha, F. Geissmann, L. Hennighausen, Cell. Biol. 19: 5969–5980.

A. Bendelac, and D. R. Littman. 2005. Genetic evidence supporting selection of 34. Dias, S., R. Ma˚nsson, S. Gurbuxani, M. Sigvardsson, and B. L. Kee. 2008. E2A by guest on September 27, 2021 the Va14i NKT cell lineage from double-positive thymocyte precursors. Im- proteins promote development of lymphoid-primed multipotent progenitors. munity 22: 705–716. Immunity 29: 217–227. 7. D’Cruz, L. M., J. Knell, J. K. Fujimoto, and A. W. Goldrath. 2010. An essential 35. Heng, T. S., M. W. Painter; Immunological Genome Project Consortium. 2008. role for the transcription factor HEB in thymocyte survival, Tcra rearrangement The Immunological Genome Project: networks of gene expression in immune and the development of natural killer T cells. Nat. Immunol. 11: 240–249. cells. Nat. Immunol. 9: 1091–1094. 8. Callen, E., R. B. Faryabi, M. Luckey, B. Hao, J. A. Daniel, W. Yang, H. W. Sun, 36. Nichols, K. E., J. Hom, S. Y. Gong, A. Ganguly, C. S. Ma, J. L. Cannons, G. Dressler, W. Peng, H. Chi, et al. 2012. The DNA damage- and transcription- S. G. Tangye, P. L. Schwartzberg, G. A. Koretzky, and P. L. Stein. 2005. Regulation of associated protein paxip1 controls thymocyte development and emigration. Im- NKT cell development by SAP, the protein defective in XLP. Nat. Med. 11: 340–345. munity 37: 971–985. 37. Bain, G., M. W. Quong, R. S. Soloff, S. M. Hedrick, and C. Murre. 1999. 9. Griewank, K., C. Borowski, S. Rietdijk, N. Wang, A. Julien, D. G. Wei, Thymocyte maturation is regulated by the activity of the helix-loop-helix pro- A. A. Mamchak, C. Terhorst, and A. Bendelac. 2007. Homotypic interactions tein, E47. J. Exp. Med. 190: 1605–1616. mediated by Slamf1 and Slamf6 receptors control NKT cell lineage develop- 38. Drennan, M. B., A. S. Franki, P. Dewint, K. Van Beneden, S. Seeuws, S. A. van ment. Immunity 27: 751–762. de Pavert, E. C. Reilly, G. Verbruggen, T. E. Lane, R. E. Mebius, et al. 2009. 10. Wei, D. G., H. Lee, S. H. Park, L. Beaudoin, L. Teyton, A. Lehuen, and Cutting edge: the chemokine receptor CXCR3 retains invariant NK T cells in the A. Bendelac. 2005. Expansion and long-range differentiation of the NKT cell thymus. J. Immunol. 183: 2213–2216. lineage in mice expressing CD1d exclusively on cortical thymocytes. J. Exp. 39. Raberger, J., A. Schebesta, S. Sakaguchi, N. Boucheron, K. E. Blomberg, Med. 202: 239–248. A. Berglo¨f, T. Kolbe, C. I. Smith, T. Ru¨licke, and W. Ellmeier. 2008. The 11. Kovalovsky, D., O. U. Uche, S. Eladad, R. M. Hobbs, W. Yi, E. Alonzo, K. Chua, transcriptional regulator PLZF induces the development of CD44 high memory M. Eidson, H. J. Kim, J. S. Im, et al. 2008. The BTB-zinc finger transcriptional phenotype T cells. Proc. Natl. Acad. Sci. USA 105: 17919–17924. regulator PLZF controls the development of invariant natural killer T cell ef- 40. Cohen, N. R., P. J. Brennan, T. Shay, G. F. Watts, M. Brigl, J. Kang, fector functions. Nat. Immunol. 9: 1055–1064. M. B. Brenner, et al. 2013. Shared and distinct transcriptional programs underlie 12. Savage, A. K., M. G. Constantinides, J. Han, D. Picard, E. Martin, B. Li, the hybrid nature of iNKT cells. Nat. Immunol. 14: 90–99. O. Lantz, and A. Bendelac. 2008. The transcription factor PLZF directs the 41. Constantinides, M. G., D. Picard, A. K. Savage, and A. Bendelac. 2011. A naive- effector program of the NKT cell lineage. Immunity 29: 391–403. like population of human CD1d-restricted T cells expressing intermediate levels 13. Savage, A. K., M. G. Constantinides, and A. Bendelac. 2011. Promyelocytic of promyelocytic leukemia zinc finger. J. Immunol. 187: 309–315. leukemia zinc finger turns on the effector T cell program without requirement for 42. Qi, Q., A. K. Kannan, and A. August. 2011. Tec family kinases: Itk signaling and the agonist TCR signaling. J. Immunol. 186: 5801–5806. development of NKT ab and gd Tcells.FEBS J. 278: 1970–1979. 14. Kovalovsky, D., E. S. Alonzo, O. U. Uche, M. Eidson, K. E. Nichols, and 43. Felices, M., and L. J. Berg. 2008. The Tec kinases Itk and Rlk regulate NKT cell D. B. Sant’Angelo. 2010. PLZF induces the spontaneous acquisition of memory/ maturation, cytokine production, and survival. J. Immunol. 180: 3007–3018. effector functions in T cells independently of NKT cell-related signals. J. 44. Guo, Z., M. Dose, D. Kovalovsky, R. Chang, J. O’Neil, A. T. Look, H. von Immunol. 184: 6746–6755. Boehmer, K. Khazaie, and F. Gounari. 2007. b-Catenin stabilization stalls the 15. Benlagha, K., T. Kyin, A. Beavis, L. Teyton, and A. Bendelac. 2002. A thymic transition from double-positive to single-positive stage and predisposes thymo- precursor to the NK T cell lineage. Science 296: 553–555. cytes to malignant transformation. Blood 109: 5463–5472. 16. Benlagha, K., D. G. Wei, J. Veiga, L. Teyton, and A. Bendelac. 2005. Charac- 45. Sharma, A., Q. Chen, T. Nguyen, Q. Yu, and J. M. Sen. 2012. T cell factor-1 and terization of the early stages of thymic NKT cell development. J. Exp. Med. 202: b-catenin control the development of memory-like CD8 thymocytes. J. Immunol. 485–492. 188: 3859–3868. 17. Constantinides, M. G., and A. Bendelac. 2013. Transcriptional regulation of the 46. Boos, M. D., Y. Yokota, G. Eberl, and B. L. Kee. 2007. Mature natural killer cell NKT cell lineage. Curr. Opin. Immunol. 25: 161–167. and lymphoid tissue-inducing cell development requires Id2-mediated suppres- 18. Matsuda, J. L., Q. Zhang, R. Ndonye, S. K. Richardson, A. R. Howell, and sion of E protein activity. J. Exp. Med. 204: 1119–1130. L. Gapin. 2006. T-bet concomitantly controls migration, survival, and effector 47. Yang, C. Y., J. A. Best, J. Knell, E. Yang, A. D. Sheridan, A. K. Jesionek, functions during the development of Va14i NKT cells. Blood 107: 2797–2805. H. S. Li, R. R. Rivera, K. C. Lind, L. M. D’Cruz, et al. 2011. The transcriptional The Journal of Immunology 5983

regulators Id2 and Id3 control the formation of distinct memory CD8+ T cell 50. Lai, D., J. Zhu, T. Wang, J. Hu-Li, M. Terabe, J. A. Berzofsky, C. Clayberger, subsets. Nat. Immunol. 12: 1221–1229. and A. M. Krensky. 2011. KLF13 sustains thymic memory-like CD8+ T cells in 48. Seiler, M. P., R. Mathew, M. K. Liszewski, C. J. Spooner, K. Barr, BALB/c mice by regulating IL-4-generating invariant natural killer T cells. J. F. Meng, H. Singh, and A. Bendelac. 2012. Elevated and sustained ex- Exp. Med. 208: 1093–1103. pression of the transcription factors Egr1 and Egr2 controls NKT line- 51. Weinreich, M. A., O. A. Odumade, S. C. Jameson, and K. A. Hogquist. 2010. age differentiation in response to TCR signaling. Nat. Immunol. 13: 264– T cells expressing the transcription factor PLZF regulate the development of 271. memory-like CD8+ T cells. Nat. Immunol. 11: 709–716. 49. Dear, T. N., T. Hainzl, M. Follo, M. Nehls, H. Wilmore, K. Matena, and 52. Felices, M., C. C. Yin, Y. Kosaka, J. Kang, and L. J. Berg. 2009. Tec kinase Itk in T. Boehm. 1997. Identification of interaction partners for the basic-helix-loop- gdT cells is pivotal for controlling IgE production in vivo. Proc. Natl. Acad. Sci. helix protein E47. Oncogene 14: 891–898. USA 106: 8308–8313. Downloaded from http://www.jimmunol.org/ by guest on September 27, 2021