Histone H3K27 Demethylase Negatively Controls the Memory Formation of Antigen-Stimulated CD8+ T Cells

This information is current as Takeshi Yamada, Shogo Nabe, Koji Toriyama, Junpei of September 28, 2021. Suzuki, Kazuki Inoue, Yuuki Imai, Atsushi Shiraishi, Katsuto Takenaka, Masaki Yasukawa and Masakatsu Yamashita J Immunol 2019; 202:1088-1098; Prepublished online 9

January 2019; Downloaded from doi: 10.4049/jimmunol.1801083 http://www.jimmunol.org/content/202/4/1088

Supplementary http://www.jimmunol.org/content/suppl/2019/01/08/jimmunol.180108 http://www.jimmunol.org/ Material 3.DCSupplemental References This article cites 43 articles, 9 of which you can access for free at: http://www.jimmunol.org/content/202/4/1088.full#ref-list-1

Why The JI? Submit online. by guest on September 28, 2021 • Rapid Reviews! 30 days* from submission to initial decision

• 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 © 2019 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Histone H3K27 Demethylase Negatively Controls the Memory Formation of Antigen-Stimulated CD8+ T Cells

Takeshi Yamada,*,† Shogo Nabe,‡ Koji Toriyama,x Junpei Suzuki,{,‖ Kazuki Inoue,# Yuuki Imai,# Atsushi Shiraishi,x Katsuto Takenaka,‡ Masaki Yasukawa,‡,** and Masakatsu Yamashita{,‖,**

Although the methylation status of histone H3K27 plays a critical role in CD4+ differentiation and its function, the role of Utx histone H3K27 demethylase in the CD8+ T cell–dependent immune response remains unclear. We therefore generated T cell– specific Utxflox/flox Cd4-Cre Tg (Utx KO) mice to determine the role of Utx in CD8+ T cells. Wild-type (WT) and Utx KO mice were infected with Listeria monocytogenes expressing OVA to analyze the immune response of Ag-specific CD8+ T cells. There was no significant difference in the number of Ag-specific CD8+ T cells upon primary infection between WT and Utx KO mice. However, Utx deficiency resulted in more Ag-specific CD8+ T cells upon secondary infection. Adoptive transfer of Utx KO CD8+ T cells Downloaded from resulted in a larger number of memory cells in the primary response than in WT. We observed a decreased expression of effector-associated transcription factors, including Prdm1 encoding Blimp1, in Utx KO CD8+ T cells. We confirmed that the trimethylation level of histone H3K27 in the Prdm1 gene loci in the Utx KO cells was higher than in the WT cells. The treatment of CD8+ T cells with Utx-cofactor a-ketoglutarate hampered the memory formation, whereas Utx inhibitor GSK-J4 enhanced the memory formation in WT CD8+ T cells. These data suggest that Utx negatively controls the memory formation of Ag-stimulated CD8+ T cells by epigenetically regulating the . Based on these findings, we identified a critical link between Utx and http://www.jimmunol.org/ the differentiation of Ag-stimulated CD8+ T cells. The Journal of Immunology, 2019, 202: 1088–1098.

ntigen-stimulated CD8+ T cells play a critical role in the localization (6–9). However, the control mechanisms of memory elimination of cells infected with intracellular patho- T cell differentiation are largely unknown. The role of various A gens, such as viruses and bacteria (1, 2). Activated CD8+ cytokines and transcription factors (TFs) as well as TCR stim- T cells are supposed to differentiate into short-lived terminal ef- ulation and costimulation have been implicated in the differ- fectors and long-lived memory precursors. Immunological mem- entiation of CD8+ T cells. Furthermore, it is evident that the ory T cells protect the host through the rapid recall response to immunological memory formation is influenced by epigenetic by guest on September 28, 2021 pathogens. Self-renewing memory CD8+ T cells can persist for changes. Therefore, a clear epigenetic understanding of the life after Ag-stimulation, inducing rapid responses and a robust differentiation into memory T cells can aid in the development effector function upon secondary infection, thereby providing of successful vaccination protocols. long-term immunological protection (3–5). Thus, a better under- The gene expression is affected by histone modification, in- standing of memory differentiation is required to develop effective cluding methylation, acetylation, and ubiquitination, as well as vaccination protocols. DNA methylation (10). It has been reported that the trimethylation Memory T cells can be categorized into central memory T of lysine 27 in the histone H3 tail is involved in X (Tcm) cells, effector memory cells, and tissue-resident memory T inactivation and tissue development via gene silencing (11–13). (Trm) cells, distinguished by differences in their expression of cell Thus, histone H3K27 modification is critical for global cell dif- surface markers, their proliferative capacity, and their anatomical ferentiation from stem cells by epigenetic gene regulation (14),

*Department of Medical Technology, Ehime Prefectural University of Health Sci- Address correspondence and reprint requests to Dr. Takeshi Yamada or Dr. Masakatsu ences, Tobe, Ehime 791-2101, Japan; †Department of Infection and Host Defenses, Yamashita, Department of Medical Technology, Ehime Prefectural University of Graduate School of Medicine, Ehime University, Shitsukawa, Toon, Ehime 791- Health Sciences, Tobe, Ehime 791-2101, Japan (T.Y.) or Department of Immunol- 0295, Japan; ‡Department of Hematology, Clinical Immunology and Infectious ogy, Graduate School of Medicine, Ehime University, Shitsukawa, Toon, Ehime Diseases, Graduate School of Medicine, Ehime University, Shitsukawa, Toon, Ehime 791-0295, Japan (M. Yamashita). E-mail addresses: [email protected] (T.Y.) 791-0295, Japan; xDepartment of Ophthalmology, Graduate School of Medicine, or [email protected] (M. Yamashita) Ehime University, Shitsukawa, Toon, Ehime 791-0295, Japan; {Department of The online version of this article contains supplemental material. Immunology, Graduate School of Medicine, Ehime University, Shitsukawa, Toon, Ehime 791-0295, Japan; ‖Department of Translational Immunology, Translational Abbreviations used in this article: ChIP, chromatin immunoprecipitation; DM-aKG, Research Center, Ehime University Hospital, Shitsukawa, Toon, Ehime 791-0295, dimethyl a-ketoglutarate; GSK-J4, ethyl 3-((6-(4,5-dihydro-1H-benzo[d]azepin-3 Japan; #Division of Integrative Pathophysiology, Department of Proteo-Inovation, (2H)-yl)-2-(pyridin-2-yl)pyrimidin-4-yl) amino) propanoate; GzmB, granzyme B; Proteo-Science Center, Graduate School of Medicine, Ehime University, Shitsukawa, histone H3K27me3, trimethylated histone H3K27; iLN, inguinal lymph node; KO, Toon, Ehime 791-0295, Japan; and **Division of Immune Regulation, Department knockout; Lm-OVA, Listeria monocytogenes expressing OVA; mLN, mesenteric of Proteo-Innovation, Proteo-Science Center, Graduate School of Medicine, Ehime lymph node; Pent+, pentamer-positive; qPCR, quantitative PCR; qRT-PCR, quanti- University, Toon, Ehime 791-0295, Japan tative RT-PCR; Tcm, central memory T; tet, translocation; TF, ; Tg, transgenic; Trm, tissue-resident memory T; TSS, transcription start site; Utx KO, ORCIDs: 0000-0003-2384-9659 (T.Y.); 0000-0002-6119-0034 (Y.I.); 0000-0003- Utxflox/flox Cd4-Cre Tg; WT, wild-type. 2138-5555 (A.S.); 0000-0002-1538-433X (M. Yasukawa). Received for publication August 3, 2018. Accepted for publication December 8, Copyright Ó 2019 by The American Association of Immunologists, Inc. 0022-1767/19/$37.50 2018. This work was supported by the Takeda Science Foundation, JSPS KAKENHI Grant 17K08887, BioLegend/TOMY Digital Biology, The Naito Foundation, and the Waksman Foundation. www.jimmunol.org/cgi/doi/10.4049/jimmunol.1801083 The Journal of Immunology 1089 suggesting that the status of trimethylated histone H3K27 (histone and 1% penicillin–streptomycin (catalog no. 26253-84; Nacalai Tesque) in H3K27me3) is a signature of gene loci associated with gene tran- the absence or presence of 2 mM DM-aKG, 1 mM GSK-J4, or a combi- scription and cellular differentiation. In the thymus, the demethy- nation of 2 mM DM-aKG and 1 mM GSK-J4 for 4 d. For in vitro pro- liferation assay, purified naive CD8+ T cells were labeled with 5 mM Cell lation of histone H3K27me3 has been observed at the promoter Proliferation Dye eFluor 670 (catalog no. 65-0840; eBioscience) before region of the gene encoding Thpok for CD4+ T cell development stimulation, according to the manufacturer’s protocol. (15). The degree of trimethylation of the histone H3K27 is highest in Tcm and naive cells, relative to effectors (16). Therefore, histone Chromatin immunoprecipitation assay H3K27demethylasemayplayanimportant role in the differentiation The Magna ChIP kit was used according to the manufacturer’s protocol of CD8+ T cells from naive cells into effector and immunological (EMD Millipore, Darmstadt, Germany). The Abs used for the chromatin memory cells. immunoprecipitation (ChIP) assay were as follows: anti-Utx mAb (catalog no. ab36938; Abcam), anti-histone H3K27me3 pAb (catalog no. 39157; The methylation of histone H3K27 is controlled by histone Active Motif), anti-histone H3K4me3 mAb (catalog no. 61379; Active H3K27–specific methyltransferase and demethylase, and the Motif), and anti-histone H3K27ac mAb (catalog no. 39133; Active Motif). removal of the trimethyl mark is associated with the induction The degree of trimethylation of histone H3K27 and H3K4 and degree of of gene expression (17, 18). Enhancer of zeste homolog 2 is a acetylation of histone H3K27 at the transcription start sites (TSSs) of the histone methyltransferase toward histone H3K27 and a catalytic Prdm1 and Tbx21 gene loci were assessed by quantitative PCR (qPCR) using the StepOnePlus Real-Time PCR System (Thermo Fisher Scientific). component of polycomb repressive complex 2 (19), which si- The specific primers and Roche Universal Probes (Roche, Basel, Switzerland) lences gene expression and regulates differentiation (20). In used were as follows: 59UTR (Prdm1),59-GGGTAAACCGTGTTAACT- contrast, Utx and Jmjd3 were discovered as histone H3K27 GCTG-39(forward) and 59-AGTGTAGTCATCTCAGGGTCTGG-39(reverse), demethylases that promote HOX gene expression for tissue probe 94; exon 1 (Prdm1),59-CTCTCTCATGTCCACCCAGTC-39(forward) Downloaded from and 59-GCGGCCGTAGAAAAGGAG-39(reverse), probe 88; intron 1 (Prdm1), development (21). Jmjd3 ablation has been reported to promote + 59-GAGCTCCCATGGACCACA-39(forward) and 59-CAAAACTGCTG- the differentiation of CD4 T cells into Th2 and Th17 cells in GGTGTTTCC-39(reverse), probe 60; exon 1 (Tbx21),59-TGCAGAGA- the small intestine and inhibit its differentiation into Th1 cells AAGCCCAG-39(forward) and 59-CGTCCGAAGACCAATG-39(reverse), (22). Utx, in turn, plays a critical role in embryonic tissue de- probe 33; intron 1 (Tbx21),59-CATCCCAGACCTAGCC-39(forward) and velopment (23–26); however, its function in T cells is largely 59-GAGGACCAGGGACAG-39(reverse), probe 26. unknown. Recently, another group reported that Utx is a critical Flow cytometry http://www.jimmunol.org/ regulator of Tfh differentiation in CD4+ Tcells(27).We therefore hypothesized that Utx also plays an important role The cell suspensions were prepared by manual disruption of the spleens and + lymph nodes with frosted glass slides, followed by lysis of erythrocytes with in CD8 T cell differentiation during infection via epigenetic an ammonium chloride/potassium solution. The livers and lungs were regulation. We investigated how Utx is involved in the immune perfused with ice-cold PBS as previously described (28, 29). Both sets of response of CD8+ T cells that are activated by Listeria infection. tissues were homogenized and incubated in PBS containing Collagenase In this article, we demonstrate the role of Utx in the memory III (400 U/ml, catalog no. CLS3; Funakoshi, Tokyo, Japan) at 37˚C for 30 min. + The digested tissues were then applied to a Percoll gradient (catalog formation of Ag-stimulated CD8 T cells upon infection. Our no. 17089101; GE Healthcare Life Sciences) to collect the lymphocytes, findings may provide important information to improve both according to the manufacturer’s protocol. The OVA-specific CD8+ T cells vaccination and immunotherapy. were detected using MHC class I Pentamer H-2Kb SIINFEKL (catalog no. by guest on September 28, 2021 F093-0A-G; ProImmune) and Fluorotag R-PE labeled (catalog no. K2A; ProImmune) according to the manufacturer’s protocol. The following Abs Materials and Methods were used for cell surface staining and intracellular staining: anti-Thy1.1 Mice FITC (HIS51; eBioscience), anti-Thy1.1 Alexa Fluor 647 (OX-7; BioLegend), anti-Thy1.2 allophycocyanin-Cy7 (30-H12; BioLegend), anti-Thy1.2 PE + + C57BL/6 Thy1.2 mice, C57BL/6 Thy1.1 mice, Cd4-Cre transgenic (Tg) (53-2.1; BD Biosciences), anti-CD3 violetFluor450 (17A2; Tonbo Bio- mice, and OT-1 Tg mice, were purchased from The Jackson Laboratory. sciences), anti-CD8 violetFluor450 (2.43; Tonbo Biosciences), anti-CD8 flox/flox Utx mice were established by Drs. K. Inoue and Y. Imai (Ehime Alexa Fluor 488 (53-6.7; BioLegend), anti-CD4 eFluor 780 (RM4-5; flox/flox University). Utx mice were crossed with Cd4-Cre Tg mice to gen- eBioscience), anti-CD62L FITC (MEL-14; BD Biosciences), anti-CD62L flox/flox erate Utx Cd4-Cre Tg (Utx KO) mice. Next, we crossed wild-type allophycocyanin (MEL-14; Tonbo Biosciences), anti-CD44 PE (IM7; (WT) mice or Utx KO mice with OT-1 Tg mice to generate WT OT-1 Tg Tonbo Biosciences), anti-CD27 PE (LG.3A10; BD Biosciences), anti- mice or Utx KO OT-1 Tg mice. The mice were genotyped by PCR using CD127 Alexa Fluor 488 (SB/199; BioLegend), anti-CD69 PE (H1.2F3; genomic DNA isolated from tails. Female mice were used at 6–12 wk of BD Biosciences), anti-CD103 FITC (2E7; eBioscience), anti-KLRG1 PE age in the experiments. (2FI/KLRG1; BD Biosciences), anti–granzyme B Alexa Fluor 647 (GB11; All experiments using mice were performed with the approval of BD Biosciences), and anti–IFN-g FITC (XMG1.2; BD Biosciences). For the Ehime University Administrative Panel for Animal Care. All animal cytokines and granzyme B (GzmB) staining, intracellular staining was care was conducted in accordance with the guidelines of Ehime University. performed as previously described (30). In brief, the splenocytes were isolated and stimulated with 1 mg/ml of H-2Kb OVA peptide SIINFEKL Reagents (catalog no. TS-5001-P; MBL, Nagoya, Japan) in the presence of monensin Dimethyl a-ketoglutarate (DM-aKG) was obtained from Tokyo Chem- (2 mM) in a 96-well culture plate for 6 h and then stained with anti-Thy1.2, ical Industry (catalog no. K0013; Tokyo, Japan). Ethyl 3-((6-(4,5-dihydro- anti-Thy1.1, and anti-CD8a and fixed and permeabilized, followed by 1H-benzo[d]azepin-3(2H)-yl)-2-(pyridin-2-yl)pyrimidin-4-yl) amino) intracellular staining for IFN-g or GzmB. Flow cytometry was performed propanoate (GSK-J4) was purchased from Abcam (catalog no. ab144395; using a Gallios instrument (Beckman Coulter), and the data were analyzed Cambridge, U.K.). with the FlowJo software program (Tree Star). CD8+ T cell stimulation and cell culture Adoptive transfer of CD8+ T cells and Listeria infection Naive CD8+ T cells were purified from the spleens of WT OT-1 Tg Naive (CD44loCD62Lhi) CD8+ T cells were purified from the spleens of (Thy1.1+ or Thy1.2+)orUtx KO OT-1 Tg (Thy1.2+) mice and then stimu- WT OT-1 Tg (Thy1.1+ or Thy1.2+)orUtx KO OT-1 Tg (Thy1.2+) mice lated with immobilized anti-CD3ε mAb (10 mg/ml, 2C11) and anti-CD28 using a Naive CD8a+ T Cell Isolation Kit and an autoMACS Pro Separator mAb (1 mg/ml, 37.5; BioLegend) and cultured in RPMI 1640 (catalog no. (Miltenyi Biotec) or MojoSort Isolation Kit (BioLegend), and their purities 189-02025; Wako Chemicals) supplemented with 10% FBS, IL-2 (10 ng/ml, were checked by flow cytometry (.95% purity). The purified CD8+ T cells + + catalog no. 212-12; PeproTech), 2 mM L-glutamine (catalog no. 16948-04; were mixed at a 1:1 ratio (Thy1.1 cells/Thy1.2 cells) and adoptively Nacalai Tesque, Kyoto, Japan), 1 mM sodium pyruvate (catalog no. 06977- transferred to double congenic (Thy1.1+Thy1.2+) mice (1 3 104 cells per 34; Nacalai Tesque), 1% MEM nonessential amino acids (catalog no. 06344- mouse, i.v.). After 18–24 h, the mice were infected with a Listeria mono- 56; Nacalai Tesque), 10 mM HEPES (catalog no. 15630-080; Thermo Fisher cytogenes expressing OVA (Lm-OVA) at 5 3 103 CFU i.v., as previously Scientific), 55 mM 2-ME (catalog no. 21985-023; Thermo Fisher Scientific), described (28). For the analysis or purification of donor cells, the splenocytes 1090 Utx INHIBITS MEMORY CD8+ T CELL DIFFERENTIATION Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 1. A lack of Utx increases the number of Ag-specific CD8+ T cells upon Lm-OVA secondary infection. The immune response of Ag-specific CD8+ T cells was analyzed at different time points after primary and secondary Lm-OVA infection by flow cytometry with an OVA-specific pentamer. (A)Numbers indicate the percentage of Pent+ cells among CD8+ T cells in the spleen at 7 d after primary infection and at 5 d after secondary infection (upper panels) (day 7 [d7] and day 72 [d72], respectively). The absolute numbers of Pent+ CD8+ T cells in the spleen were compared between WT and Utx KO mice (lower panels) (mean 6 SD, n =4‒5 per genotype). (B) The liver and lung were analyzed on days 7 and 72. Numbers indicate the percentage of Pent+ cells among CD8+ T cells (upper panels). The absolute numbers of Pent+ CD8+ T cells in the liver and lung were compared between WT and Utx KO mice (Figure legend continues) The Journal of Immunology 1091 Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 2. Utx deficiency increases memory CD8+ T cells in a cell-intrinsic manner. For the cell-intrinsic analysis, a 1:1 mixture of WT OT-1 Tg CD8+ T (Thy1.1+)/WT or Utx KO OT-1 Tg CD8+ T(Thy1.2+) cells was adoptively transferred into WT congenic (Thy1.1+Thy1.2+) mice, which were then infected with Lm-OVA to activate the donor cells. (A) A schematic outline of a competitive assay of the T cell immune response by adoptive transfer of naive OT-1 Tg CD44loCD8+ T cells into congenic mice. (B) The donor cells were collected from the spleen, liver, and lung on day 7 postinfection and analyzed by flow cytometry (upper panels), as shown in (A1). The absolute number of donor cells was calculated in each tissue (lower panels) (mean 6 SD, n = 5 per genotype). (C) The infected mice were reinfected with Lm-OVA on day 74, as shown in (A2). The donor cells were collected from the spleen, liver, and lung on day 5 after secondary infection and analyzed by flow cytometry (upper panels). The absolute number of donor cells was calculated in each tissue (lower panels) (mean 6 SD, n = 5 per genotype). (D) The donor memory cells were analyzed in different tissues on day 74 postinfection by flow cytometry (upper panels). The absolute number of donor cells was calculated in each tissue after Lm-OVA infection (lower panels) (mean 6 SD, n = 3 per genotype). The data for iLNs and mLNs are the number of Pent+ cells per 1 3 102 total cells isolated from tissues, and the data for bone marrow and blood are the number of Pent+ cells per 1 3 103 total cells isolated from tissues. Each symbol represents an individual mouse. Data are representative of at least two independent experiments. *p , 0.05, **p , 0.01, two-tailed Student t test. isolated from recipient mice were stained with anti-CD8, anti-CD4, anti- CD42CD8+Thy1.12Thy1.2+ (Utx KO) cells. Purified donor cells were Thy1.1, and anti-Thy1.2 Abs. The donor cells were then analyzed or pu- mixed at a 1:1 ratio (Thy1.1+/Thy1.2+ cells) and adoptively transferred to rified using a Gallios instrument or a BD FACSAria II cell sorter (BD double congenic (Thy1.1+Thy1.2+) mice (2 3 103 cells per mouse, i.v.). Biosciences) by gating on CD42CD8+Thy1.1+Thy1.22 (WT) cells or After 24 h, the mice were infected with the Lm-OVA strain at 5 3 104 CFU.

(lower panels) (mean 6 SD, n =4‒5 per genotype). (C) Kinetics of the absolute numbers of Pent+ cells among CD8+ T cells in the spleen after primary (1˚) and secondary (2˚) Lm-OVA infections (mean 6 SD, n =3‒5 per genotype). Each symbol represents an individual mouse. Data are representative of three independent experiments. *p , 0.05, **p , 0.01, two-tailed Student t test. 1092 Utx INHIBITS MEMORY CD8+ T CELL DIFFERENTIATION Downloaded from

FIGURE 3. Utx deficiency has no effect on the memory function in CD8+ T cells. For the analysis of the memory function, a 1:1 mixture of purified + A memory CD8 T cells was adoptively transferred into congenic mice and infected with Lm-OVA. ( ) A schematic outline of a competitive assay of the http://www.jimmunol.org/ memory CD8+ T cell response. (B) The secondary expansion of donor cells was compared in the spleen, liver, and lung on day 5 postinfection and analyzed by flow cytometry (upper panels). The absolute number of donor cells was calculated in each tissue after Lm-OVA infection (lower panels) (mean 6 SD, n = 5 per genotype). (C) The production of effector-function molecules in the donor cells was assessed by intracellular staining. The splenocytes were collected from recipient mice on day 60 and restimulated with OVA peptide for 6 h. Percentages of GzmB+ cells (left panels) and IFN-g+ cells (right panels) were compared in WT and Utx KO donor cells. The respective results of statistical analyses are shown (lower panels) (mean 6 SD, n = 3 per genotype). Each symbol represents an individual mouse. Data are representative of three independent experiments.

All experiments using Lm-OVAwere performed according to the protocols The cells were then transferred to a new plate and further cultured in the approved by the Ehime University Institutional Biosafety Committee. presence of IL-2 (10 ng/ml) for 3 d. The isolated total RNA was analyzed by guest on September 28, 2021 using the Agilent SurePrint G3 Mouse Gene Expression 8 3 60K v2 Array RNA isolation and quantitative RT-PCR (Takara Bio). The raw data were subjected to log2 transformation and Total RNA was isolated using TRIzol Reagent (catalog no. 15596018; analyzed using the GeneSpringGX12. Thermo Fisher Scientific) or NucleoSpin RNA XS (catalog no. 740902.10; Takara Bio, Shiga, Japan) according to the manufacturer’s protocols. From Results this process, cDNA was synthesized using the Superscript VILO cDNA + Synthesis Kit (Life Technologies). Quantitative RT-PCR (qRT-PCR) was A lack of Utx increases the number of Ag-specific CD8 T cells performed using the StepOnePlus Real-Time PCR System (Life Tech- upon Lm-OVA secondary infection nologies). The specific primers and Roche Universal Probes (Roche) used were as follows: CD3«,59-CCAGCCTCAAATAAAAACACG-39 (for- We generated T cell–specific Utx-deleted mice by crossing Utx- ward) and 59-GATGATTATGGCTACTGCTGTCA-39 (reverse), probe 10; floxed mice with Cd4 promoter–derived Cre-recombinase Tg mice Utx,59-CGGGTTCGTGAGGTTTCAT-39(forward) and 59-GAGATTCG- to determine the role of Utx in the T cell differentiation and TAGCAGCGAACA-39 (reverse), probe 3; Prdm1,59-TGCGGAGAGAG- function. The gene expression of Utx was markedly reduced in the GCTCCACTA-39 (forward) and 59-TGGGTTGCTTTCCGTTTG-39 (reverse), splenic CD8+ T cells from Utx KO mice according to qRT-PCR probe 80; Tbx21,59-AAACATCCTGTAATGGCTTGTG-39 (forward) and 59-TCAACCAGCACCAGACAGAG-39 (reverse), probe 19; Bcl6,59-CTGCA- (Supplemental Fig. 1A). T cells in the thymuses of WT and Utx GATGGAGCATGTTGT-39 (forward) and 59-GCCATTTCTGCTTCACTCG-39 KO mice were analyzed. There were no significant differences in (reverse), probe 4; Serpinb1a,59-TGGATTTTCTGCATGCCTCT-39 (forward) the percentage or absolute number of CD8+ T cells in the thymus and 59-GACAACAGTTCTGGGATTTTCC-39 (reverse), probe 12; Pmaip1 between WT and Utx KO mice, suggesting no effect of Utx de- (Noxa), 59-CAGATGCCTGGGAAGTCG-39 (forward) and 59-TGAGCA- letion during the double-positive phase of thymic development of CACTCGTCCTTCAA-39 (reverse), probe 15; Bim,59-GGAGACGAGTTCA- + ACGAAACTT-39 (forward) and 59-AACAGTTGTAAGATAACCATTTGAGG- CD8 T cells (Supplemental Fig. 1B), as previously reported (31). 39 (reverse), probe 41; Bmf,59-ACCCCAGAGACTCTTTTACGG-39 (forward) Before infection studies, the cell number was assessed and an and 59-GAGCCTGCAGGGAAACTG-39 (reverse), probe 55; Bcl2,59-GTAC- immunophenotypic analysis of CD8+ T cells was performed at the CTGAACCGGCATCTG-39 (forward) and 59-GGGGCCATATAGTTCCACAA- steady state. The Utx KO mice showed no significant differences 39 (reverse), probe 75; and BclxL,59-TGACCACCTAGAGCCTTGGA-39 + (forward) and 59-GCTGCATTGTTCCCGTAGA-39 (reverse), probe 2. Specific in the number of CD8 T cells in the spleen, blood, inguinal reagents for Eomes, Tcf7,andLef1 were purchased from Applied Biosystems lymph nodes (iLNs), mesenteric lymph nodes (mLNs), liver, or (catalog no. Mm01351985_ml, Mm00493445_ml, and Mm00550265_ml, re- lung from WT mice (Supplemental Fig. 1C). The immunophe- spectively). The gene expressions were calculated as the relative expression, notypic analysis based on the CD44 and CD62L expression normalized to the expression of CD3ε. showed a similar result between WT and Utx KO mice in different Gene expression profiling tissues although the loss of Utx subsequently reduced the CD44 hi + 3 6 expression on CD62L cells (Supplemental Fig. 1D). Naive CD8 T cells (1.5 10 ) were stimulated with immobilized anti– + TCR-b mAb (3 mg/ml, H57-597; BioLegend) and anti-CD28 mAb (1 mg/ml, Next, naive CD8 T cells from WT and Utx KO mice were 37.5; BioLegend) for 2 d in the presence of IL-2 (10 ng/ml; PeproTech). purified and activated in vitro using anti-CD3 and anti-CD28 Abs, The Journal of Immunology 1093 Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 4. Utx deficiency enhances the memory formation of activated CD8+ T cells via the reduced demethylation of histone H3K27me3 in the Prdm1 gene loci. An immunophenotypic analysis of Ag-specific CD8+ T cells was performed by flow cytometry after Lm-OVA infection. Purified naive OT-1 Tg CD8+ T cells adoptively transferred into WT congenic mice were analyzed on different days postinfection. (A) KLRG1hiCD127lo and KLRG1loCD127hi populations were compared between WT and Utx KO donor cells. Numbers in quadrants indicate the percentage of each population. (B) The absolute number of KLRG1loCD127hi cells was analyzed on different days postinfection (mean 6 SD, n = 5 per genotype). (C) The gene expression of differ- entiation-associated TFs in donor cells sorted on days 5 and 7 was analyzed by qRT-PCR and compared between WT and Utx KO (mean 6 SD, n = 4 per genotype). (D) An immunophenotypic analysis of CD8+ T cells was performed on different days after in vitro activation to compare differentiation between WT and Utx KO by staining with anti-CD62L and anti-CD27. Numbers in quadrants indicate the percentage of each (Figure legend continues) 1094 Utx INHIBITS MEMORY CD8+ T CELL DIFFERENTIATION and the cell proliferation was compared. There were no significant and lungs by staining with anti-CD44, CD62L, CD69, and CD103 differences with respect to cell division between the WT and Utx Abs. There were no significant differences in the percentage of KO cells (Supplemental Fig. 1E). CD44hiCD62Lhi Tcm cells in the spleen or CD69+CD103+ Trm We then infected the mice with an intracellular pathogen, Lm-OVA, cells in the liver and lung among donor memory cells between WT to determine the impact of Utx deficiency in the immune response and Utx KO, although the absolute number of each memory pop- of CD8+ T cells, as previously described (28, 32). We investigated ulation was significantly higher in the Utx KO cells than in the the primary and secondary immune response of endogenous CD8+ WT cells (Supplemental Fig. 2). These results suggest that Utx T cells to Lm-OVA infection by flow cytometry using a pentamer deficiency does not affect specific memory differentiation. for the detection of OVA-specific CD8+ T cells. Although there Utx deficiency has no effect on the memory function in were no significant differences in the primary expansion between CD8+ T cells WT and Utx KO mice on day 7 after Lm-OVA infection, Utx KO CD8+ T cells showed a higher frequency and absolute number of Next, we examined the memory function by a cell-based analysis + OVA-specific pentamer-positive (Pent+) cells in the spleen on day [Fig. 3A]. We purified naive CD8 T cells from the spleens of WT + + 72, at the peak of the secondary immune response, than did the OT-1 Tg (Thy1.1 ) and Utx KO OT-1 Tg (Thy1.2 ) mice and then WT CD8+ T cells (Fig. 1A). Other tissues, namely the liver and adoptively transferred a 1:1 mixture of WT and Utx KO OT-1 Tg + + lung, also showed a significantly higher frequency and absolute cells into the double congenic (Thy1.1 Thy1.2 ) recipient mice number of Pent+ cells among Utx KO CD8+ T cells at secondary for a competitive assay of the immune response. Memory cells infection, indicating that the increased level of Ag-specific CD8+ were purified by cell sorting on day 54 after primary infection,

T cells in Utx KO mice was not caused by tissue distribution and a 1:1 mixture of donor cells was then adoptively transferred Downloaded from (Fig. 1B). into double congenic mice again. The recipient mice were then in- The kinetics showed a significantly higher absolute number of fected with Lm-OVA to analyze the secondary response of memory Pent+ cells in Utx KO cells after day 9 of primary infection than cells. Intriguingly, the secondary expansion of donor cells showed no WT cells (Fig. 1C). significant differences between WT and Utx KO on day 5 postin- fection (Fig. 3B). Furthermore, restimulated donor cells showed no Utx deficiency increases the memory CD8+ T cell number in a

significant differences in the production of GzmB and IFN-g be- http://www.jimmunol.org/ cell-intrinsic manner tween WT and Utx KO mice (Fig. 3C). These results indicate that In a T cell–specific gene deletion system with Cd4-Cre Tg mice, Utx deficiency has no impact on the secondary expansion or Utx is absent in CD4+ T cells as well as CD8+ T cells. This raises acquisition of the effector function in memory cells. the possibility that the immune response of Utx-deleted CD8+ T cells might have been improved because of functional changes Utx deficiency enhances memory precursor formation via in Utx-deleted CD4+ T cells (i.e., in a cell-extrinsic manner). reduced demethylation of histone H3K27me3 in Prdm1 To exclude this possibility, we used an adoptive transfer approach gene loci by transferring donor cells from OT-1 Tg mice into congenic re- Next, we analyzed the differentiation from naive cells into terminal cipient mice in a competitive assay to compare WT and Utx KO effectors or memory precursors by comparing immunophenotypes by guest on September 28, 2021 CD8+ T cells in the same environment. We confirmed the presence using flow cytometry. We examined Ag-specific CD8+ T cells in of .90% Pent+ cells among naive CD8+ T cells from both WT the spleen on days 3–121 after Lm-OVA infection to analyze the OT-1 Tg and Utx KO OT-1 Tg mice by flow cytometry (data not differentiation by adoptive transfer of OT-1 Tg cells. The pop- shown). ulations of activated CD8+ T cells can be divided into short-lived We purified naive CD8+ cells from the spleens of WT OT-1 Tg terminal effectors and long-lived memory precursors, according to (Thy1.1+) and Utx KO OT-1 Tg (Thy1.2+) mice and then adop- the expression of KLRG1 and CD127 (IL-7 a) (5). The tively transferred a 1:1 mixture of WT and Utx KO OT-1 Tg cells time course study showed that the differentiation of donor Utx KO into the double congenic (Thy1.1+Thy1.2+) recipient mice to distin- CD8+ T cells into CD127hiKLRG1lo memory precursors during guish cells by flow cytometry with cell surface staining and analyze expansion, after the peak of expansion on day 7, was enhanced the primary [Fig. 2A (1)] and secondary [Fig. 2A (2)] immune re- compared with WT cells (Fig. 4A). An analysis of the absolute sponses. Consistent with the results of the endogenous immune re- number clearly showed more Utx KO memory precursors than sponse, donor Utx KO cells showed expansion similar to WT cells in WT precursors in the spleen after day 7 of infection (Fig. 4B). the spleen, liver, and lung on day 7 after Lm-OVA infection, at Therefore, to identify the molecular programs associated with the peak of the primary immune response (Fig. 2B), and showed a the functional spectrum of CD8+ T cells, we compared the gene higher frequency and absolute number of expanded cells on day expression of the TFs associated with T cell differentiation by 79 at the peak of the secondary immune response in the spleen, qRT-PCR between WT and Utx KO cells. liver, and lung than did WT cells (Fig. 2C). We purified naive CD8+ T cells from the spleens of WT OT-1 Tg Next, we analyzed memory CD8+ T cells on day 74 [Fig. 2A (Thy1.1+) and Utx KO OT-1 Tg (Thy1.2+) mice and then adop- (2)] without secondary infection. As expected, donor Utx KO cells tively transferred a 1:1 mixture of WT and Utx KO OT-1 Tg cells showed a higher frequency and absolute number of memory CD8+ into the double congenic (Thy1.1+Thy1.2+) recipient mice. Ag- T cells in the spleen, liver, lung, iLN, mLN, bone marrow, and stimulated donor cells were then isolated by cell sorting on days 5 blood than WT cells (Fig. 2D). We then analyzed the distinct and 7 after Lm-OVA infection. Utx deficiency reduced the gene immunophenotypes of donor memory T cells in the spleen, liver, expression of Prdm1 encoding Blimp1 and Tbx21 encoding Tbet,

population. (E) Purified naive CD8+ T cells were stimulated with anti-CD3ε and anti-CD28 mAbs and cultured for 7 d. The Utx binding at the TSSs of the Prdm1 gene loci was determined using WT and Utx KO CD8+ T cells by a ChIP-qPCR assay. (F) The degree of trimethylation of histone H3K27 (left panel) and histone H3K4 (middle panel) and the degree of acetylation of histone H3K27 (right panel) at the TSSs of the Prdm1 gene loci were compared between WT and Utx KO CD8+ T cells by a ChIP-qPCR assay. The results are presented relative to the input DNA (mean 6 SD, n = 3 per genotype). Each symbol represents an individual mouse. Data are representative of at least two independent experiments. *p , 0.05, **p , 0.01, two-tailed Student t test. The Journal of Immunology 1095 Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 5. Memory formation is controlled by the modulation of histone H3K27 demethylase. (A) A schematic outline of a competitive assay of the T cell immune response by adoptive transfer of in vitro–activated CD8+ T cells into congenic mice. Purified naive OT-1 Tg CD8+ T (Thy1.2+) cells were stimulated with anti-CD3 and anti-CD28 mAbs and cultured in the absence (control [Ctrl]) or presence of cofactor (aKG), inhibitor (GSK-J4), or com- bination thereof (aKG + GSK-J4) for 4 d. The treated cells were then mixed with Ctrl (Thy1.1+) cells at a 1:1 ratio. The mixture of cultured CD8+ T cells was adoptively transferred into WT congenic (Thy1.1+Thy1.2+) mice. The memory formation of donor cells was analyzed in the spleen 51 d after adoptive transfer (day 55 [d55]), and the recall response of donor cells was analyzed in the spleen 5 d after Lm-OVA infection (Figure legend continues) 1096 Utx INHIBITS MEMORY CD8+ T CELL DIFFERENTIATION which are associated with effector differentiation. In contrast, Utx percentage of CD44hiCD62Lhi cells among Utx KO donor cells deficiency increased the gene expression of Eomes, Tcf7, Lef1, and did not change, regardless of the treatments (Supplemental Fig. 4). Bcl6, which are associated with memory differentiation (Fig. 4C). We compared the donor memory cells in the spleen on day 51 These results are mostly consistent with the results of gene pro- after adoptive transfer (day 55). Surprisingly, the cell frequency filing we conducted (https://www.ncbi.nlm.nih.gov/geo/query/acc. and absolute number of WT donor cells was dramatically reduced cgi?acc=GSE90163) and indicate that Utx KO cells lost the effector by the treatment with aKG and increased by the treatment with signature and acquired the memory signature. GSK-J4 in correlation with the percentage of CD44hiCD62Lhi We then compared T cell differentiation between WT and Utx cells. In addition, treatment with both GSK-J4 and aKG rescued KO using CD8+ T cells stimulated with anti-CD3 and anti-CD28 the reduction of WT donor cells treated with aKG. In contrast, Utx mAbs and cultured for 7 d. The expression of CD27 and CD62L KO donor cells remained more numerous than WT donor cells, re- was analyzed on different days by flow cytometry. Utx KO cells gardless of the treatments (Fig. 5B). Consistent with the results from exhibited a lower frequency of CD27hiCD62Llo terminal effector memory cells, the recall response showed a reduced number of cells than did WT cells (Fig. 4D). donor cells following the treatment with aKG and more WT donor The Blimp1 expression is known to be required for CD8+ T cell cells following the treatment with GSK-J4. In this example as differentiation into terminal effectors upon Ag stimulation. We well, treatment with both GSK-J4 and aKG rescued the reduction therefore determined whether Utx directly targets the Prdm1 gene of donor cells treated with aKG. In contrast, the Utx KO donor by a ChIP assay of Utx at the Prdm1 gene loci. As expected, cells outcompeted the WT donor cells, regardless of the treatments the results showed Utx binding at the Prdm1 gene loci (Fig. 4E). (Fig. 5C).

We then analyzed the trimethylation level of histone H3K27 and Next, we compared the gene expression of TFs in in vitro– Downloaded from histone H3K4 and the acetylation level of histone H3K27 at the activated CD8+ T cells just before adoptive transfer. The results TSSs of the Prdm1 gene loci in WT and Utx KO CD8+ T cells on showed that aKG treatment increased the expression of Prdm1 day 7 after TCR stimulation by a ChIP-qPCR assay. The results and Tbx21 in WT cells, whereas GSK-J4 treatment decreased the showed a higher level of trimethylation of histone H3K27 in Utx expression of Prdm1 and increased the expression of Bcl6, Tcf7, KO cells than in WT cells and no significant differences in the and Lef1 in WT cells (Fig. 5D). Regardless of the treatments, Utx

level of trimethylation of histone H3K4 or acetylation of histone KO cells showed no changes in the gene expression of TFs. These http://www.jimmunol.org/ H3K27, indicating demethylation specific to histone H3K27me3 results suggest that the memory formation of activated CD8+ at the Prdm1 gene loci (Fig. 4F). These results suggest that Utx de- T cells is epigenetically regulated via the Utx activity. ficiency enhances the memory formation by reducing the demethy- lation of histone H3K27me3, thereby regulating gene expression, Discussion including that of Prdm1. Upon infection, Ag-specific CD8+ T cells expand and differentiate We also compared the rates of apoptosis in an attempt to explain into long-lived memory T cells that can respond more efficiently the greater number of Utx KO cells than WT cells after Lm-OVA to subsequent infections to eliminate pathogens than naive cells infection. As expected, Utx KO cells showed a lower frequency of (33). Thus, understanding the mechanisms underlying memory annexin V+ populations, a lower expression of proapoptotic , formation can help improve vaccination protocols. Epigenetic by guest on September 28, 2021 and a higher expression of antiapoptotic genes than WT cells factors have been implicated in the regulation of T cell differen- (Supplemental Fig. 3A, 3B). tiation; however, the molecular mechanisms underlying memory CD8+ T cell differentiation remain poorly understood. In the Memory formation is controlled by the modulation of histone current study, we demonstrated using T cell–specific Utx KO mice H3K27 demethylase that the histone H3K27 demethylase Utx plays a role in memory To clarify the relationship between the Utx activity and CD8+ T cell formation. Our adoptive transfer experiments revealed that WT differentiation, we altered the Utx activity for immunomodulation CD8+ T cells were outcompeted by Utx KO cells upon secondary using an activator and inhibitor of Utx. aKG is a cofactor of catalytic infection in a cell-intrinsic manner. The purified WT and Utx KO Jumonji C domain in Utx as well as a cofactor of 10–11 translocation memory cells showed no significant differences in secondary ex- (tet) enzymes for DNA demethylation, whereas GSK-J4 functions as pansion or the expression of effector molecules, as determined by a competitive inhibitor of aKG by binding the catalytic pockets a cell-based analysis, suggesting that Utx deficiency has no effect of Jumonji enzyme. We thus regulated the Utx activity using aKG on the memory function. Furthermore, there was no significant and GSK-J4 in the immune response of Ag-specific CD8+ T cells difference in the frequency of CD44hiCD62Lhi memory cells in in an adoptive transfer model of in vitro–activated CD8+ T cells the spleen or in CD103+CD69+ memory cells in the liver and lung in mice. between WT and Utx KO mice, indicating that Utx deficiency does We cultured CD8+ T cells in the presence or absence of aKG and not dominate specific memory cells, such as Tcm and Trm cells. GSK-J4 for 4 d before adoptive transfer (Fig. 5A). An immunophe- The T cell fate is known to be determined by the expression of notypic analysis revealed that the percentage of CD44hiCD62Lhi cells TFs associated with differentiation, and transcriptional regulation among WT cells was decreased by treatment with aKG and, is known to be achieved via distinct epigenetic modifications (34). conversely, increased by treatment with GSK-J4. In contrast, the The crucial TFs have been identified as cell-fate determinants of

(day 60 [d60]). (B) Memory formation was assessed by comparing the percentage of donor cells (upper panels) and the absolute number of cells (lower panel). Numbers indicate the percentage of donor cells among CD8+ T cells. The absolute number of donor cells was calculated in the spleen (mean 6 SD, n = 5 per genotype). (C) Recall responses were analyzed by comparing the percentage of donor cells (upper panels) and absolute number of cells (lower panel). Numbers indicate the percentage of donor cells among CD8+ T cells. The absolute number of donor cells was calculated in the spleen (mean 6 SD, n = 5 per genotype). (D) The gene expression of differentiation-associated TFs was analyzed by qRT-PCR and compared between WT and Utx KO using CD8+ T cells activated in vitro and cultured for 4 d (mean 6 SD, n = 4 per genotype). Each symbol represents an individual mouse. Data are representative of at least two independent experiments. *p , 0.05, **p , 0.01, two-tailed Student t test. GSK or GSK-J4, ethyl 3-((6-(4,5-dihydro-1H-benzo[d]azepin- 3(2H)-yl)-2-(pyridin-2-yl)pyrimidin-4-yl) amino) propanoate. The Journal of Immunology 1097

CD8+ T cells based on their differential gene expression (35–37). 4. Kaech, S. M., and W. Cui. 2012. Transcriptional control of effector and memory CD8+ T cell differentiation. Nat. Rev. Immunol. 12: 749–761. Consistently, our experiments also revealed that Utx deficiency 5. Kaech, S. M., J. T. Tan, E. J. Wherry, B. T. Konieczny, C. D. Surh, and reduces the gene expression of effector-associated TFs, such as R. Ahmed. 2003. Selective expression of the interleukin 7 receptor identifies Prdm1 and Tbx21, and increased the gene expression of memory- effector CD8 T cells that give rise to long-lived memory cells. Nat. Immunol. 4: 1191–1198. associated TFs, such as Eomes, Tcf7, Lef1, and Bcl6, suggesting 6. Gebhardt, T., L. M. Wakim, L. Eidsmo, P. C. Reading, W. R. Heath, and that Utx deficiency enhances the memory formation at the cost of F. R. Carbone. 2009. Memory T cells in nonlymphoid tissue that provide effector differentiation. Other groups have reported that the gene enhanced local immunity during infection with herpes simplex virus. Nat. Immunol. 10: 524–530. expression of effector molecules, such as GzmB, is regulated by 7. Jiang, X., R. A. Clark, L. Liu, A. J. Wagers, R. C. Fuhlbrigge, and T. S. Kupper. the histone H3K27 methylation status (38, 39), which supports 2012. Skin infection generates non-migratory memory CD8+ T(RM) cells pro- our finding that Utx epigenetically regulates T cell differentia- viding global skin immunity. Nature 483: 227–231. 8. Park, C. O., and T. S. Kupper. 2015. The emerging role of resident memory tion. We considered Prdm1, which induces terminal effector T cells in protective immunity and inflammatory disease. Nat. Med. 21: differentiation, a direct target of Utx, because a ChIP-qPCR 688–697. analysis showed that the trimethylation status of histone 9. Sallusto, F., D. Lenig, R. Fo¨rster, M. Lipp, and A. Lanzavecchia. 1999. Two subsets of memory T lymphocytes with distinct homing potentials and effector H3K27 at the TSS of Prdm1 genes was higher in Utx KO cells functions. Nature 401: 708–712. than in WT cells. Interestingly, the results of treatment with 10. Kouzarides, T. 2007. Chromatin modifications and their function. Cell 128: 693–705. aKG and GSK-J4 showed that the Utx activity regulates T cell 11. Hu¨bner, M. R., and D. L. Spector. 2010. Role of H3K27 demethylases Jmjd3 and differentiation, although aKG may induce differentiation via tet UTX in transcriptional regulation. Cold Spring Harb. Symp. Quant. Biol. 75: enzyme activity because DNA demethylation is known to be 43–49. 12. Schuettengruber, B., and G. Cavalli. 2009. Recruitment of polycomb group related to T cell differentiation (40, 41). We will determine the complexes and their role in the dynamic regulation of cell fate choice. Downloaded from effects of aKG on T cell differentiation via the tet enzyme in a Development 136: 3531–3542. future study. We found that aKG treatment reduces memory 13. Plath, K., J. Fang, S. K. Mlynarczyk-Evans, R. Cao, K. A. Worringer, H. Wang, C. C. de la Cruz, A. P. Otte, B. Panning, and Y. Zhang. 2003. Role of histone H3 formationincorrelationwithanincreasedexpressionofeffector- lysine 27 methylation in X inactivation. Science 300: 131–135. associated genes, whereas GSK-J4 treatment inhibits effector 14. Bannister, A. J., and T. Kouzarides. 2011. Regulation of chromatin by histone modifications. Cell Res. 21: 381–395. differentiation and enhances memory formation in correlation 15. Wang, L., and R. Bosselut. 2009. CD4-CD8 lineage differentiation: Thpok-ing with a decreased expression of effector-associated genes. into the nucleus. J. Immunol. 183: 2903–2910. http://www.jimmunol.org/ Regarding intracellular energy metabolism, metabolic switching 16. Tyrakis, P. A., A. Palazon, D. Macias, K. L. Lee, A. T. Phan, P. Velic¸a, J. You, G. S. Chia, J. Sim, A. Doedens, et al. 2016. S-2-hydroxyglutarate regulates CD8+ is known to be associated with memory differentiation in Ag- T-lymphocyte fate. Nature 540: 236–241. stimulated CD8+ T cells (42). Naive T cells undergo metabolic 17. Wei, G., L. Wei, J. Zhu, C. Zang, J. Hu-Li, Z. Yao, K. Cui, Y. Kanno, T. Y. Roh, reprogramming upon activation and differentiate into terminal W. T. Watford, et al. 2009. Global mapping of H3K4me3 and H3K27me3 reveals specificity and plasticity in lineage fate determination of differentiating CD4+ effectors and memory precursors, suggesting that there is a link T cells. Immunity 30: 155–167. between intracellular metabolism and epigenetic changes. We 18. Zhang, J. A., A. Mortazavi, B. A. Williams, B. J. Wold, and E. V. Rothenberg. 2012. Dynamic transformations of genome-wide epigenetic marking and tran- observed that the inhibition of glutamine metabolism enhanced scriptional control establish T cell identity. Cell 149: 467–482. + the memory formation of activated CD8 T cells (data not shown). 19. Chou, R. H., L. Chiu, Y. L. Yu, and W. C. Shyu. 2015. The potential roles of The Utx activity is able to regulate the downstream elements of EZH2 in regenerative medicine. Cell Transplant. 24: 313–317. by guest on September 28, 2021 20. Margueron, R., and D. Reinberg. 2011. The polycomb complex PRC2 and its metabolic pathways because cofactor aKG is produced as a me- mark in life. Nature 469: 343–349. tabolite of glucose and glutamine (43). Given these findings, we 21. Agger, K., P. A. Cloos, J. Christensen, D. Pasini, S. Rose, J. Rappsilber, conclude that Ag-stimulation activates energy metabolism and I. Issaeva, E. Canaani, A. E. Salcini, and K. Helin. 2007. UTX and JMJD3 are histone H3K27 demethylases involved in HOX gene regulation and develop- increases the production of aKG, which activates Utx in turn. Utx ment. Nature 449: 731–734. then targets the gene expression of TFs, including Prdm1, for en- 22. Li, Q., J. Zou, M. Wang, X. Ding, I. Chepelev, X. Zhou, W. Zhao, G. Wei, J. Cui, K. Zhao, et al. 2014. Critical role of histone demethylase Jmjd3 in the regulation hancement of terminal effector differentiation and inhibition of of CD4+ T-cell differentiation. Nat. Commun. 5: 5780. memory precursor differentiation. 23. Morales Torres, C., A. Laugesen, and K. Helin. 2013. Utx is required for proper In the current study, we identified Utx as an epigenetic regulator induction of ectoderm and mesoderm during differentiation of embryonic stem + cells. PLoS One 8: e60020. involved in the cell fate of activated CD8 T cells via histone H3K27 24. Shpargel, K. B., T. Sengoku, S. Yokoyama, and T. Magnuson. 2012. UTX and demethylation, regulating the gene expression of differentiation- UTY demonstrate histone demethylase-independent function in mouse embry- associated TFs. Furthermore, our data suggest that epigenetic changes onic development. PLoS Genet. 8: e1002964. + 25. Wang, R., and D. R. Green. 2012. Metabolic reprogramming and metabolic during CD8 T cell activation can be controlled by treatment with the dependency in T cells. Immunol. Rev. 249: 14–26. Utx inhibitor GSK-J4. Our findings may prove useful for developing 26. Welstead, G. G., M. P. Creyghton, S. Bilodeau, A. W. Cheng, S. Markoulaki, vaccination strategies against infectious pathogens and immuno- R. A. Young, and R. Jaenisch. 2012. X-linked H3K27me3 demethylase Utx is required for embryonic development in a sex-specific manner. Proc. Natl. Acad. therapies against cancer. Sci. USA 109: 13004–13009. 27. Cook, K. D., K. B. Shpargel, J. Starmer, F. Whitfield-Larry, B. Conley, D. E. Allard, J. E. Rager, R. C. Fry, M. L. Davenport, T. Magnuson, et al. 2015. Acknowledgments T follicular helper cell-dependent clearance of a persistent virus infection re- We thank Hao Shen and Shiki Takamura for providing Lm-OVA. We also quires T cell expression of the histone demethylase UTX. Immunity 43: 703–714. thank Aya Tamai for performing the maintenance of mice and Kenji 28. Yamada, T., M. Kanoh, S. Nabe, T. Yasuoka, J. Suzuki, A. Matsumoto, Kameda for cell sorting assistance. M. Kuwahara, S. Maruyama, T. Fujimoto, R. Sakisuka, et al. 2016. Menin plays a critical role in the regulation of the antigen-specific CD8+ T cell response upon Listeria infection. J. Immunol. 197: 4079–4089. Disclosures 29. Yamada, T., C. S. Park, M. Mamonkin, and H. D. Lacorazza. 2009. Transcription factor ELF4 controls the proliferation and homing of CD8+ T cells via the The authors have no financial conflicts of interest. Kru¨ppel-like factors and KLF2. Nat. Immunol. 10: 618–626. 30. Yamashita, M., K. Hashimoto, M. Kimura, M. Kubo, T. Tada, and T. Nakayama. 1998. Requirement for p56(lck) tyrosine kinase activation in Th subset differ- References entiation. Int. Immunol. 10: 577–591. 1. Butterfield, L. H. 2015. Cancer vaccines. BMJ 350: h988. 31. Northrup, D., R. Yagi, K. Cui, W. R. Proctor, C. Wang, K. Placek, L. R. Pohl, 2. Comber, J. D., and R. Philip. 2014. MHC class I antigen presentation and im- R. Wang, K. Ge, J. Zhu, and K. Zhao. 2017. Histone demethylases UTX and plications for developing a new generation of therapeutic vaccines. Ther. Adv. JMJD3 are required for NKT cell development in mice. Cell Biosci. 7: 25. Vaccines 2: 77–89. 32. Khan, S. H., and V. P. Badovinac. 2015. Listeria monocytogenes: a model 3. Arens, R., and S. P. Schoenberger. 2010. Plasticity in programming of effector pathogen to study antigen-specific memory CD8 T cell responses. Semin. and memory CD8 T-cell formation. Immunol. Rev. 235: 190–205. Immunopathol. 37: 301–310. 1098 Utx INHIBITS MEMORY CD8+ T CELL DIFFERENTIATION

33. Veiga-Fernandes, H., U. Walter, C. Bourgeois, A. McLean, and B. Rocha. 2000. 38. Zediak, V. P., J. B. Johnnidis, E. J. Wherry, and S. L. Berger. 2011. Cutting edge: Response of naı¨ve and memory CD8+ T cells to antigen stimulation in vivo. Nat. persistently open chromatin at effector gene loci in resting memory CD8+ T cells Immunol. 1: 47–53. independent of transcriptional status. J. Immunol. 186: 2705–2709. 34. Russ, B. E., M. Olshanksy, H. S. Smallwood, J. Li, A. E. Denton, J. E. Prier, 39. Zediak, V. P., E. J. Wherry, and S. L. Berger. 2011. The contribution of epige- A. T. Stock, H. A. Croom, J. G. Cullen, M. L. Nguyen, et al. 2014. Distinct netic memory to immunologic memory. Curr. Opin. Genet. Dev. 21: 154–159. epigenetic signatures delineate transcriptional programs during virus-specific 40. Nestor, C. E., A. Lentini, C. Ha¨gg Nilsson, D. R. Gawel, M. Gustafsson, CD8(+) T cell differentiation. [Published erratum appears in 2014 Immunity L. Mattson, H. Wang, O. Rundquist, R. R. Meehan, B. Klocke, et al. 2016. 41: 1064.] Immunity 41: 853–865. 5-hydroxymethylcytosine remodeling precedes lineage specification during dif- 35. Best, J. A., D. A. Blair, J. Knell, E. Yang, V. Mayya, A. Doedens, M. L. Dustin, ferentiation of human CD4(+) T cells. Cell Reports 16: 559–570. and A. W. Goldrath, Immunological Genome Project Consortium. 2013. Tran- 41. Yue, X., S. Trifari, T. A¨ ijo¨, A. Tsagaratou, W. A. Pastor, J. A. Zepeda-Martı´nez, scriptional insights into the CD8+ T cell response to infection and memory T cell C. W. Lio, X. Li, Y. Huang, P. Vijayanand, et al. 2016. Control of Foxp3 stability formation. Nat. Immunol. 14: 404–412. through modulation of TET activity. J. Exp. Med. 213: 377–397. 36. Doering, T. A., A. Crawford, J. M. Angelosanto, M. A. Paley, C. G. Ziegler, and 42. van der Windt, G. J., B. Everts, C. H. Chang, J. D. Curtis, T. C. Freitas, E. Amiel, E. J. Wherry. 2012. Network analysis reveals centrally connected genes and E. J. Pearce, and E. L. Pearce. 2012. Mitochondrial respiratory capacity is a pathways involved in CD8+ T cell exhaustion versus memory. Immunity 37: critical regulator of CD8+ T cell memory development. Immunity 36: 68–78. 1130–1144. 43. Carey, B. W., L. W. Finley, J. R. Cross, C. D. Allis, and C. B. Thompson. 2015. 37. Hu, G., and J. Chen. 2013. A genome-wide regulatory network identifies key tran- Intracellular a-ketoglutarate maintains the pluripotency of embryonic stem cells. scription factors for memory CD8+ T-cell development. Nat. Commun. 4: 2830. Nature 518: 413–416. Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021