Cutting Edge: Ly9 (CD229), a SLAM Family Receptor, Negatively Regulates the Development of Thymic Innate Memory-like CD8+ T and Invariant NKT Cells Jordi Sintes,*,†,1 Marta Cuenca,*,†,1 Xavier Romero,*,†,1 Ricardo Bastos,*,† Cox Terhorst,‡ Ana Angulo,*,† and Pablo Engel*,† Signaling lymphocytic activation molecule family re- among these T cell types with innate features. The latter ceptors and the specific adapter signaling lymphocytic population represents a subset of thymic single-positive (SP) activation molecule–associated protein modulate the CD8 cells characterized by the constitutive expression of high development of innate-like lymphocytes. In this study, levels of activation markers, which are indicative of a mem- we show that the thymus of Ly9-deficient mice contains ory-like phenotype (1). Disruption of either the IL4 or the an expanded population of CD8 single-positive cells Cd1d genes in BALB/c mice leads to a reduction in thymic with the characteristic phenotype of innate memory- CD8 SP T cells, demonstrating that the cytokine IL-4 se- like CD8+ T cells. Moreover, the proportion of these creted by iNKT cells is needed for their development (2, 3). innate CD8+ T cells increased dramatically postinfec- Consistently, recent studies have shown that IL-4 produced by promyelocytic leukemia zinc finger (PLZF)+ iNKT cells tion with mouse CMV. Gene expression profiling of + Ly9-deficient mice thymi showed a significant upregu- drives CD8 T cells to acquire an innate-like phenotype (re- viewed in Ref. 4). lation of IL-4 and promyelocytic leukemia zinc finger. 2/2 2/2 The homotypic interaction between the SLAM family Analyses of Ly9 IL4ra double-deficient mice (SLAMF) receptors (SLAMF1 and SLAMF6), expressed on revealed that IL-4 was needed to generate the thymic + cortical DP thymocytes and iNKT cell precursors, and the innate CD8 T cell subset. Furthermore, increased downstream signaling SLAM-associated protein (SAP) adaptor numbers of invariant NKT cells were detected in Ly9- are essential to the positive selection required for iNKT cell deficient thymi. In wild-type mice, IL-4 levels induced development (5–7). Recently, it has been demonstrated that by a-galactosylceramide injection could be inhibited by the innate CD8+ thymocytes driven by iNKT cells need SAP a mAb against Ly9. Thus, Ly9 plays a unique role as an to properly develop (8). However, nothing is known about by guest on October 1, 2021. Copyright 2012 Pageant Media Ltd. inhibitory cell surface receptor regulating the size of the specific contribution of other SLAM family members to + the thymic innate CD8 Tcellpoolandthedevelopment the development of innate-like T lymphocytes. of invariant NKT cells. The Journal of Immunology, Ly9 (SLAMF3, CD229) is a homophilic cell surface receptor 2013, 190: 21–25. present on all thymocytes and is highly expressed on innate- like lymphocytes such as iNKT cells (9, 10). Our findings show that in contrast to SLAMF1 and SLAMF6, Ly9 plays a nega- nnate-like T cells are derived from double-positive (DP) tive rather than a positive role in the signaling pathways re- thymocytes and are selected by nonclassical MHC class I quired for innate-like lymphocyte development in the thymus. molecules. These cells have several characteristics that http://classic.jimmunol.org I distinguish them from conventional T cells, including a Materials and Methods highly restricted TCR repertoire and the rapid generation of Mice effector functions after TCR stimulation without previous 2/2 3 exposure to Ag. CD1d-resticted invariant NKT (iNKT) cells, Ly9 mice (129 B6), provided by Dr. D.J. McKean (11), were back- crossed for 12 generations to BALB/c or C57BL/6 backgrounds. IL-4R–de- 2 2 MHC class Ib–restricted CD8 T cells (H2–M3-restricted), ficient mice (BALB/c-Il4ratm1Sz/J, IL4ra / ) obtained from The Jackson + 2 2 and MHC class Ia–restricted innate-like CD8 T cells are Laboratory were crossed with Ly9 / mice to generate Ly9 3 IL-4Ra double- Downloaded from

*Immunology Unit, Department of Cell Biology, Immunology, and Neurosciences, versity of Barcelona, C/Casanova 143, Barcelona E-08036, Spain. E-mail address: Medical School, University of Barcelona, 08036 Barcelona, Spain; †August Pi i Sunyer [email protected] Biomedical Research Institute, 08036 Barcelona, Spain; and ‡Division of Immunology, The online version of this article contains supplemental material. Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA 02215 Abbreviations used in this article: DP, double-positive; Eomes, eomesodermin; aGalCer, 1J.S., M.C., and X.R. contributed equally to this work. a-galactosylceramide;Id3,inhibitorofDNAbinding3;iNKT,invariantNKT; Received for publication September 11, 2012. Accepted for publication November 5, MCMV,murineCMV;PLZF,promyelocyticleukemiazincfinger;SAP,signaling 2012. lymphocytic activation molecule–associated protein; SLAM, signaling lymphocytic acti- vation molecule; SLAMF, signaling lymphocytic activation molecule family; SP, single- This work was supported by Ministerio de Educacio´n y Ciencia Grants SAF2009-7071 positive; wt, wild-type. (to P.E.) and SAF2011-25155 (to A.A.) and by National Institutes of Health Grant 2P01AI65687-06A1 (to P.E. and C.T.). X.R. was supported by Beatriu de Pinos Fel- Ó lowship 2010 BP-B. Copyright 2012 by The American Association of Immunologists, Inc. 0022-1767/12/$16.00 Address correspondence and reprint requests to Dr. Pablo Engel, Immunology Unit, Department of Cell Biology, Immunology, and Neurosciences, Medical School, Uni-

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1202435 22 CUTTING EDGE: Ly9 SHAPING THYMIC INNATE T CELL RESPONSE

2 2 2 2 deficient mice (Ly9 / IL4ra / ). All mice were maintained in a pathogen- free facility. Experiments were conducted on animals 8–12 wk old and in compliance with institutional guidelines as well as with national laws and policies. Flow cytometry Cells were stained with fluorochrome-labeled Abs using standard methods. R-PE–labeled CD1d/a-galactosylceramide (aGalCer) tetramer (Proimmune) was used to detect iNKT cells. For intracellular staining with anti-PLZF, eomesodermin (Eomes) or IFN-g cells were made permeable with an intra- cellular staining buffer (eBioscience). Data were acquired using a FACSCanto II (BD Biosciences) flow cytometer. Anti-mouse mAbs were obtained from BD Pharmingen, ImmunoTools, and eBioscience. The mouse anti-mouse Ly9 mAb (clone Ly9.7.144, IgG1 isotype) was generated in our laboratory. Gene expression analysis RNA was reverse-transcribed into cDNA using the High Capacity RNA-to- cDNA Master Mix (Applied Biosystems), whereas PCR reactions were as- sembled with TaqMan Universal PCR Master Mix (Applied Biosystems). Samples were loaded onto TaqMan low-density arrays and run in duplicate on a 7900HT Fast real-time PCR system (Applied Biosystems). Relative quantification was based on the comparative cycle threshold method using GAPDH as an endogenous control. Lymphocyte and iNKT cell activation g IFN- intracellular levels were measured by FACS in CD8 SP thymocytes + treated ex vivo with PMA (100 ng/ml), ionomycin (500 ng/ml; Sigma- FIGURE 1. Ly9 gene regulates the size of innate-like CD8 Tcellpoolin +/+ Aldrich), and GolgiStop (BD Biosciences) for 4 h. Isolated thymocytes and thymus. (A) CD4 and CD8 expression in thymocytes from wt (Ly9 )and 2 2 2 2 splenocytes (105/well) were also stimulated with different doses of the Ly9-deficient (Ly9 / ) BALB/c mice. (B)PercentageofCD8 CD4 (double- 2 2 iNKT cell ligand aGalCer (KRN7000; Enzo Life Sciences) for 48 h. Mice negative), CD8+CD4+ (DP), and CD4 and CD8 (SP) cells in Ly9+/+ and Ly9 / were i.p. injected with 250 mg Ly9.7.144 mAb or mouse IgG1 control 48 h mice. (C) Cell numbers of total thymocytes and CD8 SP cells (n =6).(D) before administrating 6 mg aGalCer. IL-4, IFN-g (BioLegend), and IL-17 Representative data of CD124 and CD122 expression in CD8 SP thymocytes (R&D Systems) production was detected by ELISA. (left panels) and frequencies of CD8+CD124+CD122+ cells from Ly9+/+ and 2/2 E Murine CMV infections Ly9 mice (n =6)(right panel). ( ) Expression of Eomes in CD8 SP thy- mocytes (left panels), cumulative data of percentage of cells (middle panel), and 6 2 2 Female mice were i.p. injected with 1–2 3 10 PFU tissue culture–propagated absolute numbers (right panel)ofEomes+ CD8 SP cells from Ly9+/+ and Ly9 / murine CMV (MCMV; Smith strain) in 500 ml DMEM. IFN-g levels were mice (n = 12) are shown. Data are representative of at least three independent measured by ELISA (BioLegend). Mice were sacrificed at day 4 postinfection experiments. Experiments were conducted on animals 8–12 wk old. and thymic populations were analyzed by FACS. by guest on October 1, 2021. Copyright 2012 Pageant Media Ltd. Statistical analysis expression levels were detected on splenic CD8+ T cells int Comparisons between groups were performed with an unpaired two-tailed (Supplemental Fig. 1B). Further analysis of the CD3 CD8 Student t test with a p value ,0.05 as the cutoff for statistical significance. SP thymocytes showed that most cells display the phenotype CD122highCD124highCD24lowCD62LhighCD69low, provid- Results and Discussion ing further evidence that these cells correspond to innate-like + Ly9 shapes the size of the innate CD8 T pool CD8+ T cells (Supplemental Fig. 1C). Additionally, we ob- 2 2 Ly9-deficient (Ly9 / ) mice on a BALB/c background de- served that the expression levels of CD122 and CD124 were veloped normally but showed a striking difference in the higher in the CD3intCD8 SP thymocytes from the Ly9-defi- http://classic.jimmunol.org distribution of thymocyte subsets in comparison with wild- cient mice compared with those of the wt mice (Supplemental type (wt) mice (Fig. 1A). An expanded population of CD8 SP Fig. 1C). The observation that CD8+ innate-like T cells thymocytes having no apparent defect in the development of expressed intermediate amounts of CD3 has not been noted the remaining thymic subsets was detected (Fig. 1B). This in previous studies. Consistent with this phenotype, the Ly9- increased percentage of CD8 SP T cells, also apparent in cell deficient thymocyte population contained a .3-fold higher numbers (Fig. 1C), was specific to the thymus, as it could not number of CD8 SP cells expressing Eomes, a transcription Downloaded from be observed in the blood or spleen (data not shown). A similar factor characteristic of innate-like CD8+ Tcells(Fig.1E). expanded population of thymic CD8+ T cells has also been This cell subset was further characterized by their negative shown to be present in several gene-deficient mice, including expression of the transcription factor T-bet (Supplemental the Kruppel-like factor 2–, IL-2–inducible T cell kinase–, Fig. 1D). Notably, the expansion of the innate-like CD8+ CREB-binding protein–, inhibitor of DNA binding 3 (Id3)–, T cells in Ly9-deficient mice could not be found in age- and NF-kB1–deficient mice (reviewed in Ref. 1). Analysis of matched Ly9-deficient mice on a C57BL/6 background 2 2 the CD8 SP T cell compartment in Ly9 / mice confirmed (Supplemental Fig. 1E). Taken together, the data show that that the increased CD8 SP population displayed the pheno- the innate CD8+ T pool in the thymus is regulated by the typic characteristics of CD8+ memory-like T cells, as judged homophilic receptor Ly9. by expression of CD122 and CD124 (Fig. 1D). Interestingly, MCMV infection induces a predominant pool of thymic innate-like these CD8 SP cells corresponded to a subset that expressed + intermediate amounts of CD3, not only readily apparent in CD8 T cells in vivo 2 2 the Ly9 / mice but also present in wt BALB/c mice (Sup- Although thymic innate memory CD8+ T cells are unlikely to plemental Fig. 1A, 1B). In contrast, no differences in CD3 play a significant role in secondary infections, they may do so The Journal of Immunology 23

during early phases of primary infections. The presence of controlling viral replication, in particular during the neonatal 2 2 elevated numbers of innate-like CD8+ T cells in the Ly9 / and early life stages before conventional memory networks are mice gave us the opportunity to test the behavior of these cells established (1). To our knowledge, this is the first report to in response to a viral infection. Although iNKT cells have show that viruses are capable of inducing an important in- been found to participate in the initial response to MCMV crease in the proportion of thymic innate CD8+ cells and infection, the role of memory-like CD8+ T cells during viral demonstrates that these cells may function as viral sensors in infections has not been examined (12). Thus, we determined an innate cell-like manner. + the kinetics of memory-like CD8 T cells in the course of + a primary MCMV infection. The infection induced a rapid and Ly9 deficiency causes the expansion of PLZF iNKT cells large increase in the proportion of CD3intCD124highCD122high To gain insight into the mechanism of the perturbed T cell 2 2 CD8 SP thymocytes, which was significantly higher in Ly9 / development observed in the Ly9-deficient mice, we per- 2 2 versus Ly9+/+ mice (Fig. 2A, 2B). Moreover, in Ly9 / mice formed gene expression profile analysis. The expression of 44 these cells also expressed higher levels of CD124 compared genes was compared between the thymi from Ly9-deficient with those in wt mice (Fig. 2B). An atrophy of the thymus and wt BALB/c mice (Supplemental Table I). We observed a induced by the MCMV infection, also reported by others, was significant upregulation of a subset of genes associated with observed (13). Further experiments are required to test the effector/memory CD8+ T cells and iNKT cells, such as PLZF differential maturation and migration of thymic subsets in- and IL4 (Fig. 3A). These two genes are specifically expressed duced by MCMV infection. by thymic iNKT cells and both are essential to their devel- 2 2 A time course analysis revealed that Ly9 / mice responded opment and effector functions (4). Remarkably, the lack of with higher levels of IFN-g compared with wt mice (Fig. 2C). Ly9 induced the upregulation of several genes encoding the Analysis of the viral load in different organs 96 h postinfec- chemokine receptors Cxcr3 and Cxcr5, which are known 2 2 tion showed slightly lower virus in the Ly9 / mice compared markers of effector/memory CD8+ T cells. The most signif- with their wt counterparts. However, these differences were icant increase found was that of Cxcr3, which has similarly not statistically significant (data not shown). Further studies been reported to increase in the thymic innate-like CD8+ cells will be needed to analyze the specific impact of these cells on of Kruppel-like factor 2–deficient mice (14). Interestingly, by guest on October 1, 2021. Copyright 2012 Pageant Media Ltd. http://classic.jimmunol.org Downloaded from

FIGURE 3. Ly9 is critical for thymic iNKT cell numbers, a major source 2 2 of IL-4 required for innate-like CD8+ T cell expansion in Ly9 / mice. (A) 2 2 FIGURE 2. MCMV infection augments the ratio of the innate-like CD8+ Genes differentially expressed between Ly9 / and Ly9+/+ thymocytes. (B) T cells in the thymus. (A) Frequencies of CD3intCD8 SP cells among thy- Contour plots of CD4 and CD8 expression in gated CD3 intermediate 2 2 2 2 2 2 2 2 mocytes from Ly9+/+ and Ly9 / mice at the indicated time points after thymocytes from Ly9+/+, Ly9 / , and Ly9 / IL4ra / .(C) Absolute infection. (B) Contour plot of CD124 and CD122 expression in gated CD3intCD8 SP cell numbers and (D) expression of Eomes and IFN-g in 2 2 CD3intCD8 SP cells (left panels), cumulative data of percentage of CD8 SP thymocytes (black bar/line, Ly9+/+; red bar/line, Ly9 / ; and blue 2 2 2 2 CD3intCD8+CD124+CD122+ cells (middle panel), and mean fluorescence bar/line, Ly9 / IL4ra / ). (E) Representative data of PLZF and CD3 ex- intensity of CD124 expression in gated CD3intCD8 SP cells from Ly9+/+ and pression (left panels) and cumulative data of percentage of PLZF+ cells from 2 2 2 2 Ly9 / mice 96 h after infection (right panel). (C) Analysis of IFN-g plasma thymocytes of Ly9+/+ and Ly9 / mice (right panel). (F) Representative iNKT 2 2 levels from Ly9+/+ and Ly9 / mice (n = 6/group) at various time points after staining (left panel), frequency (middle panel), and absolute iNKT cell num- 2 2 viral infection. Data are representative of at least three independent experi- bers (right panel) from Ly9+/+ and Ly9 / thymi. Data are representative of ments. two (A), three (B–D), or at least four (E, F) independent experiments. 24 CUTTING EDGE: Ly9 SHAPING THYMIC INNATE T CELL RESPONSE

thymic iNKT cells express Cxcr3 that retains these cells in the thymus (15). Moreover, an increase in Cxcr5 has also been re- ported in the CD8 SP T cells of Id3-deficient thymocytes (16). The expansion of a population of innate CD8 SP T cells in several deficient mice (Klf2, Itk,andId3) has been attributed to an IL-4–dependent mechanism. These studies demonstrate that the acquisition of the innate T CD8+ phenotype was not an intrinsic defect in these cells, but rather stemmed from an expanded population of IL-4–producing iNKT cells. Thus, to further understand the requirements for the expansion of 2 2 2 2 innate-like CD8+ T cells in Ly9 / mice, we crossed Ly9 / mice with others deficient in the IL-4 receptor. This com- pletely prevented the expansion of CD3intCD8 SP cells (Fig. 3B, 3C). Moreover, the increased levels of Eomes found in 2 2 the Ly9-deficient mice were undetectable in the Ly9 / IL- 2 2 4ra / mice (Fig. 3D). Consistent with this result, the per- centage of CD8 SP T cells producing IFN-g after ex vivo activation with PMA and ionomycin was dramatically re- duced in the double-deficient cells (Fig. 3D). All of these data clearly demonstrate that in a Ly9 deficiency setting IL-4 is an essential requirement for the generation of an expanded memory-like or innate-like CD8+ population. We also ob- served a 3- to 6-fold increase in the percentage of thymic + + FIGURE 4. Ly9 deletion enhances cytokine production in iNKT stimu- PLZF and CD1d tetramer iNKT cells and a 6-fold increase A +/+ 2/2 2/2 lated cells. ( )IL-4inplasmafromLy9 and Ly9 mice (n = 11) i.p. in the numbers of iNKT cells in Ly9 mice (Fig. 3E, 3F), injected with aGalCer (6 mg). Blood samples were collected at the indicated indicating that expansion of the CD8 SP cells was unlikely time points. (B) IL-4 levels in the culture supernatants of thymocytes from 2 2 due to an increased functional capacity of the iNKT cells, but Ly9+/+ and Ly9 / (n = 7) activated with aGalCer for 48 h. (C) IL-4 (left rather resulted from an increase in the number of these cells. panel), IL-17 (middle panel), and IFN-g (right panel) measured in plasma +/+ In contrast, we only observed a slight increase in the number from Ly9 mice injected with anti-Ly9 or a matching isotype (each group 2 2 of iNKT cells in the Ly9 / C57BL/6 mice, which may be treatment n = 12) 48 h before administration of aGalCer. Blood was obtained 4 h after aGalCer stimulation. (D) iNKT cell numbers or (E) changes of unable to support an expansion of thymic innate CD8+ + Τ PLZF mRNA expression levels after anti-Ly9 or a matching isotype treatment T cells (Supplemental Fig. 1F). The number of PLZF gd for 48 h previous to aGalCer administration. Data are representative of at

by guest on October 1, 2021. Copyright 2012 Pageant Media Ltd. cells, a subset also involved in thymic IL-4 secretion, was not least two independent experiments. significantly altered (data not shown). Additionally, no sig- nificant alteration was found in the expression of SAP, SLAMF1, or SLAMF6 in Ly9-deficient mice (Supplemental anti-Ly9 mAb. The injection of wt mice with anti-Ly9 mAb Table I and data not shown). Thus, in contrast to the positive (Ly9.7.144) was able to induce a significant decrease in the regulators SLAMF1 and SLAMF6, Ly9 (SLAMF3) is a nega- production of IL-4 but not in IL-17 or IFN-g after aGalCer tive regulator of iNKT cell development in the thymus. administration (Fig. 4C). No effect of this mAb was observed in Ly9-deficient mice, supporting its specificity. This decrease Ly9 deletion enhances cytokine production by stimulated iNKT cells in IL-4 production was most likely due to the agonistic effect http://classic.jimmunol.org Given the important roles played by iNKT cells and IL-4 in the of this mAb. This noncytotoxic mAb did not significantly + development of innate-like CD8 T cells, we next sought to affect the numbers of iNKT cells or the levels of PLZF in determine the response of Ly9-deficient mice to aGalCer, thymus (Fig. 4D, 4E). We have previously shown that mAbs a specific iNKT cell agonist. Compared with wt mice, Ly9- against human Ly9 (CD229) negatively regulate TCR sig- deficient mice presented an enhanced production of IL-4 2 h naling, thereby inhibiting ERK phosphorylation (17). after in vivo injection with aGalCer (Fig. 4A). Consistent 2 2 Collectively, our data suggest that a Ly9 deficiency may alter Downloaded from with our in vivo results, the ex vivo activation of Ly9 / the threshold for thymic selection, resulting in a strong pos- thymocytes with aGalCer also induced significantly higher itive selection of those iNKT cells that would normally have levels of IL-4 as compared with thymocytes of wt mice (Fig. been negatively selected, thus indirectly resulting in an aug- 4B). In contrast, no significant difference in the in vitro mented commitment of innate-like CD8 SP thymocytes via production of IL-4 after aGalCer activation could be ob- the production of high levels of IL-4. This proposed role of served in splenocytes, where only a moderate increase in the Ly9 is consistent with our observation that immature thymic iNKT cell numbers was detected in the Ly9-deficient mice iNKT cells (stage 0) expressed very high levels of Ly9 com- and no alteration in activation markers was observed (Sup- pared with mature cells (Supplemental Fig. 1H). plemental Fig. 1G). Thus, these results suggest that the in- In conclusion, the Ly9 cell surface receptor emerges as creased levels of IL-4 largely resulted from elevated numbers a uniquely important element for the modulation of innate of iNKT cells in Ly9-deficient mice and were likely unrelated T cell function, acting as an inhibitory molecule that regulates to the hyperreactivity of these cells. iNKT development and innate-like CD8 T cell expansion. Importantly, we demonstrate in this study that IL-4 pro- Further studies will be required to establish the molecular duction could be modulated by the in vivo treatment with an mechanisms by which Ly9 regulates innate lymphocyte de- The Journal of Immunology 25

velopment and effector functions. Nevertheless, this work 7. Kageyama, R., J. L. Cannons, F. Zhao, I. Yusuf, C. Lao, M. Locci, P. L. Schwartzberg, and S. Crotty. 2012. The receptor Ly108 functions as a SAP demonstrates that Ly9 is a potential target for therapeutic in- adaptor-dependent on-off switch for T cell help to B cells and NKT cell develop- tervention in diseases where iNKT cell numbers play a relevant ment. Immunity 36: 986–1002. 8. Verykokakis, M., M. D. Boos, A. Bendelac, and B. L. Kee. 2010. SAP protein- role, including cancer, autoimmune and inflammatory dis- dependent natural killer T-like cells regulate the development of CD8+ T cells with eases, and infections. innate lymphocyte characteristics. Immunity 33: 203–215. 9. Romero, X., N. Zapater, M. Calvo, S. G. Kalko, M. A. de la Fuente, V. Tovar, C. Ockeloen, P. Pizcueta, and P. Engel. 2005. CD229 (Ly9) lymphocyte cell surface receptor interacts homophilically through its N-terminal domain and relocalizes to Acknowledgments the immunological synapse. J. Immunol. 174: 7033–7042. We thank A. La´zaro and J. de Salort for technical assistance. 10. Sintes, J., M. Vidal-Laliena, X. Romero, V. Tovar, and P. Engel. 2007. Charac- terization of mouse CD229 (Ly9), a leukocyte cell surface molecule of the CD150 (SLAM) family. Tissue Antigens 70: 355–362. Disclosures 11. Graham, D. B., M. P. Bell, M. M. McCausland, C. J. Huntoon, J. van Deursen, The authors have no financial conflicts of interest. W. A. Faubion, S. Crotty, and D. J. McKean. 2006. Ly9 (CD229)-deficient mice exhibit T cell defects yet do not share several phenotypic characteristics associated with SLAM- and SAP-deficient mice. J. Immunol. 176: 291–300. 12. Tyznik, A. J., E. Tupin, N. A. Nagarajan, M. J. Her, C. A. Benedict, and References M. Kronenberg. 2008. Cutting edge: the mechanism of invariant NKT cell 1. Lee, Y. J., S. C. Jameson, and K. A. Hogquist. 2011. Alternative memory in the responses to viral danger signals. J. Immunol. 181: 4452–4456. CD8 T cell lineage. Trends Immunol. 32: 50–56. 13. Price, P., S. D. Olver, A. E. Gibbons, H. K. Teo, and G. R. Shellam. 1993. 2. Weinreich, M. A., O. A. Odumade, S. C. Jameson, and K. A. Hogquist. 2010. Characterization of thymic involution induced by murine cytomegalovirus infec- T cells expressing the transcription factor PLZF regulate the development of tion. Immunol. Cell Biol. 71: 155–165. memory-like CD8+ T cells. Nat. Immunol. 11: 709–716. 14. Weinreich, M. A., K. Takada, C. Skon, S. L. Reiner, S. C. Jameson, and 3. Lai, D., J. Zhu, T. Wang, J. Hu-Li, M. Terabe, J. A. Berzofsky, C. Clayberger, and K. A. Hogquist. 2009. KLF2 transcription-factor deficiency in T cells results in A. M. Krensky. 2011. KLF13 sustains thymic memory-like CD8+ T cells in BALB/c unrestrained cytokine production and upregulation of bystander chemokine mice by regulating IL-4-generating invariant natural killer T cells. J. Exp. Med. 208: receptors. Immunity 31: 122–130. 1093–1103. 15. Drennan, M. B., A. S. Franki, P. Dewint, K. Van Beneden, S. Seeuws, S. A. van de 4. Alonzo, E. S., and D. B. Sant’Angelo. 2011. Development of PLZF-expressing Pavert, E. C. Reilly, G. Verbruggen, T. E. Lane, R. E. Mebius, et al. 2009. Cutting innate T cells. Curr. Opin. Immunol. 23: 220–227. edge: the chemokine receptor CXCR3 retains invariant NK T cells in the thymus. J. 5. Griewank, K., C. Borowski, S. Rietdijk, N. Wang, A. Julien, D. G. Wei, Immunol. 183: 2213–2216. A. A. Mamchak, C. Terhorst, and A. Bendelac. 2007. Homotypic interactions 16. Miyazaki, M., R. R. Rivera, K. Miyazaki, Y. C. Lin, Y. Agata, and C. Murre. 2011. mediated by Slamf1 and Slamf6 receptors control NKT cell lineage development. The opposing roles of the transcription factor E2A and its antagonist Id3 that or- Immunity 27: 751–762. chestrate and enforce the naive fate of T cells. Nat. Immunol. 12: 992–1001. 6. Horai, R., K. L. Mueller, R. A. Handon, J. L. Cannons, S. M. Anderson, 17. Martı´n, M., J. M. Del Valle, I. Saborit, and P. Engel. 2005. Identification of Grb2 M. R. Kirby, and P. L. Schwartzberg. 2007. Requirements for selection of con- as a novel binding partner of the signaling lymphocytic activation molecule- ventional and innate T lymphocyte lineages. Immunity 27: 775–785. associated protein binding receptor CD229. J. Immunol. 174: 5977–5986. by guest on October 1, 2021. Copyright 2012 Pageant Media Ltd. http://classic.jimmunol.org Downloaded from Supplemental material.

Supplemental figure 1. Characterization of Ly9-deficient thymic-innate-like

CD8+ T and iNKT cell subsets.

(A-C) CD3 intermediate expression characterizes innate-like CD8+ T cell subset. (A) Flow cytometry of thymocytes stained with CD4, CD8, CD124,

CD122 and CD3. Expression of CD3 in gated CD124+ CD122+ SP CD8+ T cells compared with total CD8 SP cells. (B) Comparison of CD3 expression between single positive CD8 thymocytes and splenic cells from Ly9+/+ and Ly9-/- mice.

Data are representative of at least 5 independent experiments (A and B). (C)

Characterization of the Ly9, CD122 (IL2Rβ), CD124 (IL4R), CD24, CD62L and

CD69 expression in CD3 intermediate (CD3int) and CD3 high expressing thymocytes from Ly9+/+ and Ly9-/- mice. Data are representative of at least 2 independent experiments with at least 2 mice per group.

(D) Representative contour histogram of Eomes and T-bet expression in CD8

SP thymocytes from Ly9+/+ and Ly9-/- BALB/c mice.

(E) CD4 and CD8 expression in thymocytes (left panel) and Eomes expression in CD8 SP thymocytes from wt (Ly9+/+) and Ly9-deficient (Ly9-/-) C57BL/6 mice

(middle panel). Percentage of CD8-CD4- (DN), CD8+CD4+ (DP), CD4 and CD8

(SP) cells in Ly9+/+ and Ly9-/- C57BL/6 thymi (right panel). Cumulative data of at least 2 experiments are shown.

(F) Representative iNKT staining (left panel), frequency (middle panel) and absolute iNKT cell numbers (right panel) from Ly9+/+ and Ly9-/- C57BL/6 thymi.

Cumulative data of at least 2 experiments are shown. (G) IL-4 production measured by ELISA in 48h αGalCer in vitro activated splenocytes from Ly9+/+ and Ly9-/- BALB/c mice (n=8). The data are means and

SEM (upper left panel). Frequency of iNKT cells (αGalCer-CD1d tetramer+

B220- CD3+) among splenocytes from Ly9+/+ and Ly9-/- (n=12 per group) mice

(upper right panel) and iNKT cell numbers from Ly9+/+ and Ly9-/- (n=6 per group)

(lower left panel). Representative histogram of CD122 and CD69 expression in splenic iNKT cells (lower right panel). Each circle represents an individual mouse. Small horizontal bars indicate the mean

(H) Wild-type CD1d tet+ cell sorted iNKT cells from pooled thymocytes and then labelled with anti-CD24 and anti-Ly9. Sorted iNKT cells were further classified base on their CD24 expression (left panel). Representative histogram of Ly9 expression in iNKT CD24+ (Stage 0; red line) and CD24- (Stages 1, 2 and 3; blue line) subsets from Ly9+/+ thymus are shown. Tinted lines are controls using Ly9 expression on Ly9-/- thymocytes (middle panel). A bar diagram of cumulative data of three independent cell sorting experiments are shown in right panel.

Supplemental Table I. Gene expression levels from Ly9-/- and Ly9+/+ thymocytes.

Data represent levels of mRNA expression of 44 genes from a customized low density array after statistical analysis of four independent gene expression profiles from Ly9+/+ and Ly9-/- mice. Columns show gene symbol, entrez gene number, identifier from Applied Biosystems, P-values (NS means not significant) and fold change difference between Ly9-/- vs Ly9+/+ thymocytes on a logarithmic scale (base 2). Fold differences; positive values Ly9-/- > Ly9+/+, negative values Ly9-/- < Ly9+/+. Supplemental figure 1

A BALB/c thymus CD8 SP CD122 CD124 CD8 SP Ly9/

16% 4% CD4 CD124

CD8 CD122 Events (% of max) CD3

B Thymic CD8 SP Splenic CD8 SP Ly9/ 100 Ly9/ 80 Cd229 60

40

20 Events (% of max) CD3 C Ly9/ Thymic CD8 SP Ly9/

CD3int

CD3high

Events (% of max) Ly9 CD122 CD124 CD24 CD62L CD69 (IL2R) (IL-4R) D BALB/c thymus: CD8 SP Ly9/ Ly9/ 50.5 0.080 72.7 1.14

Eomes 48.1 0.565 25.8 0.383 T-bet E C57BL/6 Thymus CD8 SP / /   Ly9 82.4% / 7.88%   9.9% Ly9 / 90 60 30 15 3.7% 4%   / / 85.4% 10.2% 10 8.3% 5 Percentage (%) 0 Eomes CD4 3% 3.3% DN DP CD4 CD8 CD8 SP SP

F C57BL/6 thymus / 0.368% p= .0179 p= .0165 1.5 ) 1.0 6 0.8 1.0 0.6 / 0.572% 0.4 NKT cells

i 0.5 0.2 % NKT cells (x10 i

CD1d tet 0.0 0.0 Ly9/ Ly9/ Ly9/ Ly9/ CD3   G Ly9 / BALB/c Spleen   Ly9 / 250 3 p= .0158 200 2 150 NKT i 100 %

IL-4 (pg/ml) 1 50 ND 0 0     Medium 5 50 100 Ly9 / Ly9 / GalCer (ng/ml)

/ 3 Ly9 )

6   100 Ly9 / 2 80 60 1 40 20 NKT cells (x10 i

0 Events (% of max) CD122 CD69 H BALB/c Thymus MFI ST1/2/3; 1745 MFI ST0; 3885 5 )

ST0; 0.123 % 3 4 3 2 1

ST1/2/3; Ly9 MFI (x10 99.7 % CD24 0 Events (% of max) CD1d Ly9 ST0 ST1/2/3 Gene symbol Entrez gene Identifier P -value Fold change

Cytokines/chemokines Ifng 15978 Ifng-Mm01168134_m1 0.069 (NS) -0.74 Tnf 21926 Tnf-Mm00443258_m1 0.060 (NS) 0.39 Tnfrsf13b 57916 Tnfrsf13b-Mm00840182_m1 0.151 (NS) 0.66 Il4 16189 Il4-Mm00445260_m1 0.001 1.13

Cytokine/chemokine receptors Il2ra 16184 Il2ra-Mm01340213_m1 0.983 (NS) 0.00 Cxcr4 12767 Cxcr4-Mm01292123_m1 0.597 (NS) 0.13 Ccr7 12775 Ccr7-Mm00432608_m1 0.150 (NS) 0.57 Ccr6 12458 Ccr6-Mm01323931_m1 0.045 0.75 Il2rb 16185 Il2rb-Mm00434264_m1 0.009 0.99 Ccr1 12768 Ccr1-Mm01216147_m1 0.003 1.05 Cxcr5 12145 Cxcr5-Mm00432086_m1 0.004 1.07 Cxcr3 12766 Cxcr3-Mm00438259_m1 0.001 1.66

Costimulatory molecules Cd40lg 21947 Cd40lg-Mm00441911_m1 0.007 0.52 Icos 54167 Icos-Mm00497600_m1 0.003 0.75 Cd40 21939 Cd40-Mm00441891_m1 0.002 1.10

SLAMF members and SAP protein Ly9 17085 Ly9-Mm00500512_m1 <0,0001 -2.50 Slamf1 27218 Slamf1-Mm00443316_m1 0.100 (NS) -0.21 Sh2d1a 20400 Sh2d1a-Mm00448717_m1 0.589 (NS) 0.12 Slamf6 30925 Slamf6-Mm00472836_m1 0.169 (NS) 0.24 Adaptors and signal molecules Csk 12988 Csk-Mm00432751_m1 0.650 (NS) -0.07 Fasl 14103 Fasl-Mm00438864_m1 0.882 (NS) -0.03 Fas 14102 Fas-Mm00433237_m1 0.630 (NS) 0.07 S1pr1 13609 S1pr1-Mm00514644_m1 0.321 (NS) 0.38 Traf1 22029 Traf1-Mm00493827_m1 0.074 (NS) 0.48

Transcription factors Myb 17863 Myb-Mm00501741_m1 0.004 -0.53 Gata3 14462 Gata3-Mm00484683_m1 0.250 (NS) -0.15 Irf2 16363 Irf2-Mm00515204_m1 0.391 (NS) -0.13 Foxp3 20371 Foxp3-Mm00475162_m1 0.887 (NS) 0.06 Ets1 23871 Ets1-Mm00468970_m1 0.710 (NS) 0.10 Nfatc1 18018 Nfatc1-Mm00479445_m1 0.322 (NS) 0.15 Tbx21 57765 Tbx21-Mm00450960_m1 0.311 (NS) 0.24 Notch2 18129 Notch2-Mm00803077_m1 0.035 0.27 Irf8 15900 Irf8-Mm00492567_m1 0.227 (NS) 0.28 Zbtb7b (Thpok) 22724 Zbtb7b-Mm00784709_s1 0.426 (NS) 0.30 Pou2f2 18987 Pou2f2-Mm01149405_m1 0.181 (NS) 0.40 Stat1 20846 Stat1-Mm00439531_m1 0.005 0.66 Irf9 16391 Irf9-Mm00492679_m1 0.004 0.70 Zbtb16 (PLZF) 235320 Zbtb16-Mm01176865_m1 <0,0001 1.04

Others Cd24a 12484 Cd24a-Mm00782538_sH 0.038 -0.34 Cd8a 12525 Cd8a-Mm01182108_m1 0.404 (NS) -0.14 Mki67 17345 Mki67-Mm01278616_m1 0.710 (NS) 0.06 Cd44 12505 Cd44-Mm01277164_m1 0.503 (NS) 0.09 Cd69 12515 Cd69-Mm01183378_m1 0.049 0.37 Cd1d1 12479 Cd1d1-Mm00783541_s1 0.029 0.46