The Interferon Regulatory Factor 4 is required for the generation of protective effector + CD8 T cells

Friederike Raczkowskia,1, Josephine Ritterb,1, Kira Heescha, Valéa Schumachera, Anna Guralnikb, Lena Höckerb, Hartmann Raiferb, Matthias Kleinc, Tobias Boppc, Hani Harbd, Dörthe A. Kesperd, Petra I. Pfefferled, Melanie Grusdate, Philipp A. Lange, Hans-Willi Mittrückera,2,3, and Magdalena Huberb,2,3

aInstitute for Immunology, University Medical Center Hamburg–Eppendorf, 20246 Hamburg, Germany; bInstitute for Medical Microbiology and Hospital Hygiene, University of Marburg, 35033 Marburg, Germany; cInstitute for Immunology, University Medical Center of the Johannes Gutenberg-University Mainz, 55131 Mainz, Germany; dInstitute for Laboratory Medicine and Pathobiochemistry, University of Marburg, 35043 Marburg, Germany; and eDepartment of Gastroenterology, Hepatology, and Infectious Diseases, Heinrich Heine University, 40225 Düsseldorf, Germany

Edited by Tak W. Mak, The Campbell Family Institute for Breast Cancer Research, Ontario Cancer Institute at Princess Margaret Hospital, University Health Network, Toronto, ON, Canada, and approved July 29, 2013 (received for review May 20, 2013)

+ Robust cytotoxic CD8 T-cell response is important for immunity to The transcription factor IFN regulatory factor 4 (IRF4) controls intracellular pathogens. Here, we show that the transcription fac- class-switch recombination, germinal center B-cell formation, and + tor IFN Regulatory Factor 4 (IRF4) is crucial for the protective CD8+ plasma cell development (12). In CD4 T cells, IRF4 is crucial for T-cell response to the intracellular bacterium Listeria monocyto- the differentiation into T helper (Th) subsets such as Th2, Th9, .IRF4-deficient (Irf4−/−) mice could not clear L. monocytogenes Th17, and Tfh cells (13–18). Mechanistically, IRF4 controls B-cell infection and generated decreased numbers of L. monocytogenes- and dendritic cell differentiation by cooperative DNA binding + specificCD8 T cells with impaired effector phenotype and function. with TFs of the Ets family on Ets-IRF composite elements (EICE) + − − Transfer of wild-type CD8 T cells into Irf4 / mice improved bacte- as well as by cooperation with basic transcription rial clearance, suggesting an intrinsic defect of CD8+ T cells in Irf4−/− factor ATF-like (BATF)-JUN heterodimers in binding to AP-1- − − + – / IRF4 composite elements (AICE) (19 23). In contrast, differen- IMMUNOLOGY mice. Following transfer into wild-type recipients, Irf4 CD8 + T cells became activated and showed initial proliferation upon tiation of CD4 T cells relies mainly on IRF4 binding to AICE el- L. monocytogenes infection. However, these cells could not sus- ements (19, 21, 22). Moreover, there is evidence for cooperation of IRF4 with other TFs, including members of the NFAT, STAT, or tain proliferation, produced reduced amounts of IFN-γ and TNF-α, families (12). and failed to acquire cytotoxic function. Forced IRF4 expression in −/− + There is only limited information on the function of IRF4 in Irf4 CD8 T cells rescued the defect. During acute infection, + fi −/− −/− + CD8 T cells. IRF4-de cient (Irf4 ) mice are impaired in their Irf4 CD8 T cells demonstrated diminished expression of B lym- response to lymphocytic choriomeningitis virus infection (24) phocyte-induced maturation protein-1 (Blimp-1), inhibitor of DNA and IRF4 appears to control expression of Eomes in these cells + binding (Id)2, and T-box expressed in T cells (T-bet), transcription (25, 26). Here, we investigate the role of IRF4 in CD8 T cells factors programming effector-cell generation. IRF4 was essential for during an immune response against the intracellular bacterium expression of Blimp-1, suggesting that altered regulation of Blimp-1 Listeria monocytogenes and demonstrate an intrinsic role for contributes to the defects of Irf4−/− CD8+ T cells. Despite increased − − IRF4 in the differentiation of peripheral cytotoxic T lymphocytes. levels of B-cell lymphoma 6 (BCL-6), , and Id3, Irf4 / + CD8 T cells showed impaired memory-cell formation, indicating Results additional functions for IRF4 in this process. As IRF4 governs B-cell IRF4 Is Essential for Clearance of L. monocytogenes. Infection of + + and CD4 T-cell differentiation, the identification of its decisive mice with L. monocytogenes induces a robust effector CD8 + role in peripheral CD8 T-cell differentiation, suggests a common T-cell response, which is crucial for clearance of bacteria (27). + regulatory function for IRF4 in adaptive lymphocytes fate decision. To elucidate the role of IRF4 in generation of protective CD8 − − T cells, Irf4 / and WT mice were infected with L. monocytogenes. + −/− ollowing infection with intracellular pathogens, specific CD8 Compared with WT mice, Irf4 mice were compromised in the eradication of L. monocytogenes (Fig. 1A and Fig. S1A). Fur- FT cells become activated, proliferate, and differentiate into + cytotoxic T cells, which are critical for the clearance of infection. thermore, the expansion of the CD8 T-cell population and the fl acquisition of the CD62LloCD44hiKLRG1hi effector phenotype Upon antigen encounter, these effector cells produce in am- + − − matory and have the capability to kill infected cells. by CD8 T cells were greatly impaired in Irf4 / mice (Fig. S1 B– After resolution of infection, the bulk of effector cells dies; F). However, we observed an increase in the proportion of CD44hi − − + however, a small fraction remains as long-lived memory T cells cells, indicating activation of Irf4 / CD8 T cells during infec- that respond with rapid conversion into effector cells upon tion (Fig. S1D). To evaluate whether the impaired clearance of reexposure to the cognate pathogen (1). Phenotypic and functional markers allow distinction between + short-lived effector CD8 T cells and cells that give rise to long- Author contributions: F.R., J.R., T.B., P.A.L., H.-W.M., and M.H. designed research; F.R., J.R., lived memory cells already at early stages of the response. Ef- K.H., V.S., A.G., L.H., H.R., M.K., H.H., D.A.K., P.I.P., M.G., H.-W.M., and M.H. performed hi lo hi αlo research; F.R., J.R., H.-W.M., and M.H. analyzed data; and F.R., H.-W.M., and M.H. wrote fector cells display a CD44 CD62L KLRG1 IL7-R phenotype the paper. and memory precursor cells can be defined as CD44hiKLRG1loIL7- + The authors declare no conflict of interest. Rαhi cells (1). Differentiation of CD8 T cells into effector and This article is a PNAS Direct Submission. memory cells is regulated by balanced expression of several tran- 1F.R. and J.R. contributed equally to this work and are listed in alphabetical order. scription factors (TF). Whereas BCL-6 (2, 3), Eomesodermin 2H.-W.M. and M.H. contributed equally to this work. (Eomes) (4), Id3 (5, 6), and TCF-1 (7) are associated with memory 3To whom correspondence may be addressed. E-mail: [email protected] or cell differentiation and longevity of cells, T-bet (encoded by Tbx21) [email protected]. (4, 8), Id2 (9), and Blimp-1 (encoded by Prdm1) (10, 11) promote This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. effector cell development. 1073/pnas.1309378110/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1309378110 PNAS Early Edition | 1of6 Downloaded by guest on September 28, 2021 A 5 * * ** * B 4 ** ** CD44, CXCR3, and CD25, suggesting that they reacted to 4 LmOVA, although to a lesser extent than WT cells. Furthermore, 3 they failed to down-regulate CD62L and CD27 and to up-regulate 3 + 2 KLRG1, again confirming that phenotypic alterations of CD8 2 −/− CFU/liver [log10]

CFU/liver [log10] T cells observed in Irf4 mice were due to an intrinsic defect of 1 1 WT Irf4-/- WT Irf4-/- WT Irf4-/- WT Irf4-/- WT Irf4-/- Irf4-/- these cells (Fig. 3C and Fig. S3 D and G). Consistently, IRF4 d 4 d 12 d 25 d 40 + WT CD8+ was rapidly induced by polyclonal or antigen-specific stimulation WT Ly5.1+ CD8+ Irf4-/- Ly5.1 - and during L. monocytogenes infection, its induction correlated C + 27.5 39.9 49.9 16.4 0.2 44.5 with the acquisition of the effector phenotype by CD8 Tcells CD62L CD44 KLRG1 (Fig. S5 A–F). − − + To elucidate whether reduced accumulation of Irf4 / CD8 23.9 8.6 1.5 32.1 48.8 6.5 Ly5.1 T cells was caused by restricted proliferation, we measured car- boxyfluorescein succinimidyl ester (CFSE) dilution. Both WT and − − Fig. 1. Impaired control of L. monocytogenes infection by IRF4-deficient Irf4 / OT-I cells had proliferated extensively as measured at day + −/− − − CD8 T cells. (A) Irf4 and WT mice were infected with L. monocytogenes 5 posttransfer (Fig. 3D). Because impaired accumulation of Irf4 / and colony forming units (CFU) in livers were determined at indicated days − − + + OT-I cells was evident at day 5 but not at day 3 posttransfer and / − − + p.i. (B) Irf4 mice were reconstituted with Ly5.1 congenic WT CD8 T cells infection, we speculated that Irf4 / CD8 T cells might not and infected with L. monocytogenes. CFU in livers were determined at day + maintain an initial proliferation. Indeed, BrdU incorporation by 12 p.i. (C) CD44, CD62L, and KLRG1 expression by splenic CD8 T cells (day 12 − − − − − / fi p.i.) in reconstituted Irf4 / (Ly5.1 ) mice. (A and B) CFU for individual mice Irf4 cells between days 4 and 5 was signi cantly lower, com- and the median of combined results from two independent experiments are pared with that by WT cells (Fig. 3E), suggesting reduced pro- −/− shown (n = 12–14). The dashed line gives the detection limit. (C) Numbers liferation of Irf4 cells at this stage of infection. The − − give percentage of positive cells. Experiments were repeated twice with proliferative defect of Irf4 / cells was also detectable in vitro consistent results. and could not be rescued by the addition of high amounts of IL-2 (250 units) (Fig. S6 A and B). The rate of apoptosis was reduced −/− + in Irf4 OT-I cells at day 5 p.i. as determined by fluorescent L. monocytogenes was caused by defective function of CD8 T + − − labeled inhibitor of caspases (FLICA) and Annexin V staining, / −/− cells, we transferred WT CD8 TcellsintoIrf4 mice. At day 12 and Irf4 cells expressed higher levels of the prosurvival factor postinfection (p.i.), significantly lower bacterial numbers were − − + Bcl2 (Fig. 3F and Fig. S3 E and F), reflecting again their defect found in Irf4 / mice after transfer of WT CD8 T cells and in − − + to mature into effector cells that are more prone to apoptosis contrast to Irf4 / CD8 T cells, a substantial proportion of trans- lo hi (28). Furthermore, the expression of the exhaustion markers ferred WT cells displayed a CD62L KLRG1 effector phenotype −/− + CD244, LAG-3, and PD-1 (29) was not enhanced in Irf4 OT-I (Fig. 1 B and C and Fig. S1G). Thus, WT CD8 T cells acquired cells compared with WT cells at day 5 of infection (Fig. S3G). effector properties in an IRF4-deficient environment, indicating −/− + + Thus, reduced numbers of Irf4 CD8 T cells during acute that intrinsic defects in CD8 T cells were at least in part responsible −/− L. monocytogenes infection were most likely caused by an intrinsic for the impaired clearance of L. monocytogenes by Irf4 mice. failure to maintain proliferation and not due to increased apoptosis + − − or exhaustion. Impaired Pathogen-Specific Effector CD8 T-Cell Response in Irf4 / Mice. To characterize the function of IRF4 in an antigen-specific − − setting, WT and Irf4 / mice were infected with an L. mono- cytogenes strain recombinant for chicken ovalbumin (LmOVA). −/− A B 6 * * * Irf4 mice also failed to clear the LmOVA infection (Fig. S2A). naive WT Irf4 -/- + +

257-264 0.05 3.5 0.9 4 We noted a significant reduction in OVA-specific CD8 T cells - Dex

− − [log10]

/ + /Ova b in tissues of Irf4 mice and these cells failed to acquire a 257-264 2

lo hi – fi CD8 Ova CD62L KLRG1 phenotype (Fig. 2 A C). OVA-speci c cy- H-2K − − / 0 tokine production was also greatly impaired in Irf4 mice SPL LV BM CD8 compared with WT mice (Fig. 2D). Infection with an L. mono- WT Irf4 -/- WT Irf4 -/- cytogenes strain recombinant for gp33 from LCMV revealed C 1.1 13.3 9.1 86.1 52.4 6.8 0.8 2.9 a comparable defect (Fig. S2 B–E). Furthermore, we failed to KLRG1 detect substantial responses to H2-M3–restricted formyl-methionin CD62L (f-met) peptides of L. monocytogenes (Fig. S2 B and C). Thus, 0.7 84.9 0.0 4.8 15.6 25.2 36.0 60.3 −/− + CD44 IL-7Rα Irf4 mice fail to mount a regular CD8 effector response to -/- WT Irf4 WT Irf4 -/- 4 several immunodominant peptides presented by different MHC D 1.0 0.1 0.5 0.1 * + 3 molecules during infection with L. monocytogenes. α * 2 IL-2 TNF-

−/− + % of CD8 1 Irf4 CD8 T Cells Display Altered Proliferative Behavior. The anal- 1.0 0.2 ysis of L. monocytogenes clearance suggested an intrinsic de- IFN- γ CD8 0 IFN-γ+ IFN-γ+ + − − + fect of CD8 T cells in Irf4 / mice (Fig. 1 B and C). To fully TNF-α fi + isolate the IRF4 de ciency to CD8 T cells, we conducted Fig. 2. IRF4 deficiency impairs generation of pathogen-specific effector −/− + − − competitive transfers of small numbers of both WT and Irf4 CD8 T cells. (A–D) Irf4 / and WT mice were infected with LmOVA and + b + OVA-specificOT-ICD8 T cells into congenic WT recipients, analyzed at day 12 p.i. (A) H-2K OVA257–264-dextramer staining of CD8 b + followed by LmOVA infection. At day 3 after transfer and in- T cells isolated from spleens. (B) Numbers of H-2K OVA257–264-dextramer − − + fection, we found similar numbers of WT and Irf4 / OT-I cells CD8 T cells in spleens (SPL), livers (LV), and bone marrow (BM) (mean ± SEM, b n = 4). (C) CD44 and CD62L or KLRG1 and IL-7Rα expression on H-2K OVA257–264- in spleens of recipient mice; however, at day 5 the ratio changed + + − − dextramer CD8 T cells isolated from spleens of WT and Irf4 / mice. (D) to approximately 10:1 (Fig. 3 A and B and Fig. S3 A and B). −/− −/− Spleen cells from Irf4 and WT were stimulated with OVA257–264 peptide + Reduced accumulation of Irf4 OT-I cells was also found in and analyzed on a CD8 gate for intracellular IFN-γ, TNF-α, and IL-2. Bar + + + + other tissues (Fig. S3C), and noncompetitive T-cell transfers graphs display percentage values of IFN-γ and IFN-γ TNF-α CD8 T cells gave similar results (Fig. S4 A and B). Phenotypically, transferred (mean ± SEM, n = 4). (A, C, and D) Numbers give percentage of positive cells. −/− Irf4 OT-I cells displayed less pronounced up-regulation of Experiments were repeated twice with consistent results.

2of6 | www.pnas.org/cgi/doi/10.1073/pnas.1309378110 Raczkowski et al. Downloaded by guest on September 28, 2021 Fig. 3. IRF4 intrinsically regulates the phenotype d5 + fi ABd3 5 *** C and proliferation of effector CD8 T cells. Puri ed 1849 20009 939 OT-I WT + −/− − + 13897 8753 242 OT-I Irf4-/- CD8 T cells from Irf4 (CD90.1 CD90.2 ) OT-I mice 4 + + and WT OT-I mice (CD90.1 CD90.2 ) were mixed in 52.7 % of Max. OT-I WT × 3 OT-I Irf4-/- a ratio of 1:1 and labeled with CFSE. A total of 4 47.3 OT-I [log10] 104 OT-I cells were injected into LmOVA-infected 2 + − d3 d5 CD62L CD44 KLRG1 CD90.1 CD90.2 congenic mice. Transferred cells d5 D from spleens of recipient mice were analyzed at CFSE d0 d5 E BrdU d4 d5 F d5 6 15 30 92.4 210 4367 6847 *** indicated days p.i. (A) CD90.1 and CD90.2 staining 5 339 1493 *** 1288 of transferred OT-I cells at days 3 and 5 p.i. Numbers CD90.1 4 10 20 − − 7.2 ns. % of Max. % of Max. give the percentage values for Irf4 / and WT OT-I 3 % of Max. 10 % FLICA+ − − % BrdU+ 5 cells. (B) Total numbers of transferred Irf4 / and 2 1

CD90.2 MFI CFSE [log10] WT OT-I cells at indicated days p.i. (C) Expression of 0 0 0 − − CFSE d0 d5 BrdU d5 FLICA d0 d5 CD62L, CD44, and KLRG1 on WT and Irf4 / OT-I − − cells at day 5 p.i. (D) CFSE profiles of Irf4 / and WT OT-I cells at day 5 p.i. [dotted line and shaded bar: CFSE staining of mixed cells before transfer (day 0)]. (E) − − Mice received BrdU at day 4 p.i. and were analyzed at day 5 p.i. Plots show BrdU incorporation by transferred cells. (F) FLICA binding by transferred Irf4 / and WT cells before transfer (day 0) and at day 5 p.i. Numbers in all histograms give the mean fluorescence intensity (MFI). Bars give the mean ± SEM, n = 3–5. Experiments were repeated twice with consistent results.

+ − − + IRF4 Regulates CD8 T-Cell Effector Development in a Cell-Intrinsic Eomes, and Id3 was increased in Irf4 / CD8 T cells. This Manner. Acquisition of effector functions such as cytotoxicity and feature combined with the CD44hiCD62Lhi memory-like phe- inflammatory production is central for the protective −/− + + + notype of activated Irf4 CD8 T cells (Fig. S1D and Fig. 3C) capacity of CD8 T cells. To evaluate the role of IRF4 in CD8 suggested that absence of IRF4 might promote the formation of + T cells in this process, we again used competitive transfer of WT memory CD8 T cells. However, in two experimental approaches −/− + − − and Irf4 OT-I cells. Although we detected similar accumula- in which either CD8 T cells were directly analyzed in Irf4 / −/− − − tion of Irf4 and WT OT-I cells at day 3 posttransfer and in- mice or Irf4 / OT-I cells were analyzed after transfer into WT −/− + fection, Irf4 cells already displayed impaired IFN-γ, TNF-α, recipients, we noted decreased numbers of OVA-specific CD8 and GzmB production at this time point and the defect in GzmB T cells 40 d after LmOVA infection and these cells were pro- expression was even more pronounced at day 5 (Fig. 4 A, B, and γ α IMMUNOLOGY −/− foundly impaired in IFN- and TNF- production after stimu- D and Fig. S7 A and B). Irf4 cells also failed to produce IL-2 lation (Fig. S8 A–H). Thus, the formation as well as the function −/− γ + (Fig. 4C). The defect of Irf4 cells in production of IFN- and of long-lived memory CD8 T cells was markedly impaired in the α TNF- was also detectable in vitro and could not be rescued by absence of IRF4. addition of high amounts of IL-2 (Fig. S6C). The mRNA analysis −/− + of sorted WT and Irf4 OT-I cells isolated from acute infection IRF4 Binds Directly to Regulatory Elements of the Prdm1 in CD8 + revealed decreased levels for the cytotoxic molecules GzmB, T Cells. Blimp-1–deficient CD8 T cells display impaired cyto- Granzyme K, and Perforin 1 (Fig. 4E). Consistent with this re- −/− toxicity and express diminished levels of KLRG1 and Tbx21, sult, we detected impaired cytotoxicity of Irf4 OT-I cells in an whereas the expression of Eomes and Bcl6 is increased in these in vivo kill assay after transfer of LmOVA-activated WT and fi − − cells. Therefore, Blimp-1 has been de ned as a central TF for Irf4 / OT-I cells (Fig. 4F). Importantly, diminished cytotoxicity + − − terminal effector CD8 T-cell differentiation (10, 11). Because / − − + was not caused by loss of Irf4 OT-I cells because we detected of similarities in the phenotype of Irf4 / CD8 T cells and that similar numbers of transferred cells in recipients of WT and – fi + −/− described for Blimp-1 de cient CD8 T cells and strong re- Irf4 cells (Fig. 4F). In summary, these results demonstrate an −/− + −/− + duction of Prdm1 expression in Irf4 CD8 T cells during acute intrinsic defect of Irf4 CD8 T cells to acquire functions of L. monocytogenes infection, we hypothesized that IRF4 regulates terminal effector cells during L. monocytogenes infection. + −/− − − + Prdm1. To test this, CD8 T cells from WT and Irf4 mice were To exclude developmental defects of Irf4 / CD8 T cells, we −/− activated in vitro and then cultured with cytokines contributing retrovirally overexpressed IRF4 in LmOVA-primed WT or Irf4 + to CD8 T-cell activation (Fig. 6A). Without addition of cyto- OT-I cells (7, 8). Cells were transferred into congenic mice, which − − + kines, Irf4 / CD8 T cells showed reduced expression of Prdm1 were infected with LmOVA (Fig. 4 G and H). WT cells transduced compared with WT cells. IL-2 even in high concentrations and with control virus (control-RV) or IRF4-expressing virus (IRF4- − − IL-12 did not change the expression level of Prdm1 in both RV) displayed similar production of IFN-γ and TNF-α. Irf4 / populations (Fig. 6A and Fig. S6D). Consistent with published OT-I cells transduced with control-RV barely produced cytokines. + −/− data for CD4 T cells (30), IL-21 strongly induced Prdm1 and In contrast, Irf4 cells transduced with IRF4-RV produced IFN-γ + − − + and TNF-α. Thus, forced expression of IRF4 in Irf4 / CD8 Blimp-1 protein in WT CD8 T cells, which corresponded to enhanced IRF4 levels. In agreement with our ex vivo data, Prdm1 T cells rescued at least partially the cytokine production, corrob- − − + / orating the crucial role of IRF4 for CD8 effector differentiation expression was markedly lower in Irf4 cells and Blimp-1 was undetectable (Fig. 6 A–C). Reduced expression was not due to and excluding developmental defects. − − + a failure of Irf4 / CD8 T cells to react to IL-21, because WT and − − + Irf4 / cells displayed similar phosphorylation of STAT3 upon IL- IRF4 Controls Expression of Transcription Factors Regulating CD8 − − + 21 treatment (Fig. S9A). Furthermore, the transduction of Irf4 / T-Cell Fate Decision. Differentiation of effector CD8 T cells is + controlled by coordinated expression of several TFs (2–11). CD8 T cells with IRF4-expressing retrovirus induced enhanced − − + Therefore, WT and Irf4 / OT-I CD8 T cells were sorted from expression of Prdm1 compared with transduction with control recipient mice during acute L. monocytogenes infection and retrovirus, suggesting direct regulation of Blimp-1 by IRF4 (Fig. mRNA levels for different TFs were determined by quantitative 6D and Fig. S9B). Previous studies (30) have identified an IL-21 RT-PCR analysis (Fig. 5). Consistent with severely impaired response element downstream of Prdm1 that binds IRF4 in B cells − − + + effector differentiation of Irf4 / CD8 T cells, we found di- and CD4 T cells and is required for optimal Prdm1 expression. + minished expression of TFs important for CD8 effector-cell Our chromatin immunoprecipitation (ChIP) analysis revealed development such as Prdm1 (encoding Blimp-1), Id2, and Tbx21 strong binding of IRF4 to this element after treatment of WT + (encoding T-bet) in these cells. Notably, the expression of TFs CD8 T cells with IL-21. Computational analysis of the 5′ region associated with memory T-cell differentiation such as BCL-6, of the Prdm1 gene revealed further putative IRF4-binding

Raczkowski et al. PNAS Early Edition | 3of6 Downloaded by guest on September 28, 2021 + -/- WT Irf4 d3 d3 Characterization of CD8 T cells in transfer assays revealed that A B − − + 4593 80 ** WT Irf4 / CD8 T cells responded to infection with initial pro- 3684 60 Irf4-/- of OT-I + 40 liferation. Activated cells showed phenotypic changes and pro- −/− + 20 duced cytokines. However, Irf4 CD8 T cells could not sustain 83.6 52.3

% GzmB 0 “ ” d5 d5 proliferation and retained a precursor-like state, as indicated 60957 80 *** WT by diminished cytokine response, low expression of cytotoxic 16051 60

of OT-I -/-

+ Irf4 fi 40 , and defective cytotoxicity. This de ciency in maturation % of Max TNF- α + 20 into fully functional effector cells was CD8 T-cell intrinsic and 83.2 53.1 − − + 0 / % GzmB could be rescued by retroviral IRF4 expression in Irf4 CD8 IFN- γ GzmB T cells. The retroviral expression studies also exclude develop- C WT Irf4 -/- d5 D 100 mental defects, such as altered thymic maturation, as a main cause 1.0 4.5 0.2 6.0 ** *** 5 *** + −/− 80 4 of impaired CD8 T-cell responses in Irf4 mice. This result is of OT-I

+ 60 3 γ IL-2 40 2 consistent with our results and a recent publication (25), which −/− 20 1 78.8 52.6 % IL-2 of OT-I demonstrate normal thymic T-cell maturation in Irf4 mice. % IFN- 0 0 −/− + IFN- γ d3 d5 d5 The impaired maturation of Irf4 CD8 T cells during acute

E F ns infection was accompanied by decreased expression levels of TFs 100 Gzmb 16 Gzmk 10 Prf1 30 ** 5 central to the formation of effector cells such as Blimp-1, Id2, 75 12 7.5 4 20 and T-bet (1), suggesting that IRF4 controls their expression. 50 8 5 3 −/− + – fi + 10 Because Irf4 CD8 T cells resemble Blimp-1 de cient CD8 25 4 2.5 2 % cytotoxicity T cells with regard to impaired cytotoxicity and diminished 0 0 0 0 [log10] OT-I/spleen 0 Relative expression KLRG1 expression (10, 11), we hypothesized that IRF4 func- GHcontrol-RV IRF4-RV control-RV IRF4-RV tions upstream of Blimp-1 in regulating effector-cell maturation. 0.9 ± 0.4 9.3 ± 3.6 Indeed, we could demonstrate that Blimp-1 expression was re- −/− +

TNF- α Irf4

TNF- α duced in CD8 T cells. Conversely, the overexpression of −/− + 4.1 ± 1.6 IRF4 in Irf4 CD8 T cells enhanced Prdm1 expression. Re- 8.9 ± 5.3 −/− + IFN- γ IFN- γ duced Prdm1 level was also evident in Irf4 CD8 T cells after

+ addition of high amounts of IL-2. These results point to a direct Fig. 4. IRF4 regulates effector CD8 T-cell differentiation. (A–D) Purified + − − control of Prdm1 by IRF4. Accordingly, IRF4 specifically bound CD8 T cells from Irf4 / and WT OT-I mice were mixed in a ratio of 1:1 and 4 × 104 cells were injected into LmOVA-infected congenic mice. Transferred to the IL-21 responsive element and other regulatory regions of the Prdm1 gene. Recently, IRF4 was shown to be crucial in IL- cells from spleens of recipient mice were analyzed at day 3 and day 5 p.i. Bars + give the mean ± SEM, n = 4. (A, C, and D) Cytokine production is shown after 21–induced Blimp-1 expression in B cells and CD4 T cells (30) stimulation with OVA – peptide. Numbers in plots and bars show per- as well as in Treg cells (31). Our data extend these observations 257 264 + centages of positive cells. (B) Numbers in histograms and bars give the MFI of to CD8 T cells and suggest the IRF4–Blimp-1 axis as a common − − GzmB. (E) Irf4 / and WT OT-1 cells were sorted on day 6 after transfer and regulatory mechanism in lymphocytes. LmOVA infection and mRNA expression levels were determined by qRT-PCR. −/− Because Blimp-1 controls terminal effector differentiation and Relative expression was calculated by setting the expression levels in Irf4 the expression of other TFs involved in lymphocyte differentia- OT-I cells to 1 (mean ± SD of duplicate PCR samples). (F) After activation in − − LmOVA-infected recipient mice, equal numbers of purified Irf4 / and WT tion such as T-bet, Eomes, BCL-6, and Id3 (10, 11), impaired OT-I cells were retransferred into congenic WT mice that received target and induction of Blimp-1 in the absence of IRF4 could be largely −/− control cells. Graphs show percentage of killing and total numbers of OT-I responsible for the defective effector differentiation of Irf4 + cells recovered from spleens after 4 h (mean ± SEM, n = 3). (G and H) Acti- CD8 T cells. To prove this concept, we tried to overexpress − − − − + vated WT (G)orIrf4 / (H) OT-I cells were transduced with retrovirus Blimp-1 in Irf4 / CD8 T cells. However, due to the large size of + encoding IRF4-GFP (IRF4-RV) or a control retrovirus encoding GFP (control- Blimp-1, we were not able to reliably express Blimp-1 in CD8 RV) and transferred into LmOVA-infected congenic mice. At day 8 p.i., spleen – fi + + T cells. Furthermore, Blimp-1 de cient CD8 T cells differ from cells were stimulated with OVA257–264 peptide and GFP OT-I cells were an- −/− + ± = γ+ γ+ α+ Irf4 CD8 T cells in IL-2 production and memory cell forma- alyzed. Numbers give mean SEM, n 4, of IFN- and IFN- TNF- cells. fl + Experiments were repeated twice (A–F) or four times (G and H) with con- tion (10, 11); thus, it is likely that IRF4 in uences CD8 T-cell differentiation by additional Blimp-1–independent mechanisms. sistent results. + It is likely that similarly to that in CD4 T cells, cooperative binding of IRF4 with BATF-JUN heterodimers also occurs in + sites, for two of which we found specific binding (Fig. 6E). Pre- CD8 T cells. BATF-deficient mice show impaired effector + cipitation with anti-IRF4 Ab resulted in significant enrichment of CD8 T-cell differentiation, characterized by reduced expression the analyzed regulatory elements of Prdm1 compared with pre- of IFN-γ, Perforin, and T-bet, which was caused by changes in cipitation with control IgG and there was no significant binding of IRF4 to control sequences from the 5′ region of the Rpl32 gene, which does not contain canonical IRF motifs (Fig. S9C). These Prdm1 Tbx21 Id2 Bcl6 Eomes Id3 + 24 8 12 12 4.5 results demonstrate that in CD8 T cells, IRF4 is essential for 12 18 6 Blimp-1 expression and binds specifically to regulatory regions of 8 8 3.0 12 4 8 the Prdm1 gene and suggest that the regulation of Blimp-1 by 4 4 1.5 −/− expression 6 2 4 IRF4 contributes to the impaired effector differentiation of Irf4 Relative mRNA + CD8 T cells. WT Irf4-/-

Discussion Fig. 5. IRF4 balances the expression of transcription factors controlling + + + 4 −/− Despite the established role for IRF4 in B-cell (12) and CD4 CD8 T-cell differentiation. Purified CD8 T cells (4 × 10 ) from Irf4 and WT + T-cell differentiation (13–19, 21, 22), its function in CD8 T cells OT-I mice were individually transferred into LmOVA-infected congenic WT recipients. After 6 d, OT-I cells were sorted and analyzed by qRT-PCR. mRNA during infection has not been closely evaluated. Here, we show −/− levels were normalized to Hprt1 and relative expression was calculated by that Irf4 mice mostly failed to eradicate the intracellular setting of the lowest experimental value to 1. Bars give the mean (± SD) of bacterium L. monocytogenes. This defect was caused largely by duplicate PCR samples. The experiment was repeated twice with consis- + a failure in the generation of a protective CD8 T-cell response. tent results.

4of6 | www.pnas.org/cgi/doi/10.1073/pnas.1309378110 Raczkowski et al. Downloaded by guest on September 28, 2021 A C already at early differentiation steps that are common to both WT Irf4-/- 8 Prdm1 fates. Along with this model, IRF4 may function as a permissive IL-21 + - + - factor by rendering precursor-like cells responsive to fate-shap- Ctrl IRF4 6 IL-21 ing environmental cues such as cytokines, costimulatory signals, IL-2 or different antigen loads that are encountered in the course of β-Actin 4 IL-12 infection. Consequently, loss of IRF4 causes a disbalanced tran- scriptional program that immediately affects effector-cell gener- 2 D ation during acute infection but also prevents the formation of + 2.5 Prdm1 functional memory CD8 T cells. Relative mRNA expression 0 2.0 In conclusion, our results identify IRF4 as a central regulator 6h 24h 6h 24h + 1.5 of peripheral CD8 T-cell differentiation. These data parallel ob- WT Irf4-/- 1.0 servations in late stages of B-cell development (12) and in gener- B WT Irf4-/- ation of Th-cell subsets (13–18) and thus support the idea of IL-21 + - + - 0.5 conserved transcriptional modules regulating peripheral differen- Blimp-1 Relative expression 0.0 Ctrl-RV IRF4-RV tiation of adaptive lymphocytes (33). β-Actin Methods E Mice and L. monocytogenes Infection. C57BL/6, congenic CD45.1 (Ly5.1) (6.SJL- Prdm1, Prdm1, Prdm1, 3`region 5`region 5`region Ptprca Pep3b/BoyJ; The Jackson Laboratory), congenic CD90.1 (B6.PL-Thy1a/ IL-21 responsive element -35 to +183 -1631 to -1430 × −/− −/− 0.15 0.3 0.3 CyJ; The Jackson Laboratory), OT-I, (OT-I CD90.1)F1, Irf4 , and Irf4 OT-I mice were bred at the animal facilities of the University Medical Center – 0.10 0.2 0.2 Hamburg Eppendorf or of the Biomedical Research Center, University of Marburg. Experiments were conducted according to the German animal protection law. Mice were infected i.p. with 5 × 104 colony-forming units 0.05 0.1 0.1 (cfu) of the WT L. monocytogenes strain EGD (Lm) or with 105 cfu of nd. nd. nd. nd. nd. nd. nd. nd. nd. L. monocytogenes strains recombinant for ovalbumin (LmOVA) or for gp33 0.00 0.0 0.0 IL-21 - + - + IL-21 - + - + IL-21 - + - + of LCMV (LmGP33) (34). Bacterial titers were quantified as previously de- Specific pulldown (% of input) WT Irf4-/- WT Irf4-/- WT Irf4-/- scribed (35). IMMUNOLOGY + + + − − Fig. 6. IRF4 controls Prdm1 expression in CD8 T cells. (A) Purified CD8 T Cell Transfer Experiments. CD8 T cells from spleens of WT or Irf4 / OT-I mice −/− cells from Irf4 and WT OT-I mice were preactivated with immobilized anti- were purified by negative magnetic activated cell sorting (MACS) selection CD3 mAb plus soluble anti-CD28 mAb and IL-2 for 3 d and then rested for (Miltenyi Biotec) according to the manufacturer’s protocol. For noncom- − − + 24 h. Prdm1 mRNA was determined by qRT-PCR at 6 and 24 h after treat- petitive transfers, 4 × 104 WT or Irf4 / OT-I CD8 T cells were injected i.v. ± + + ment as indicated. Bars give the mean ( SD) of duplicate PCR samples. into Ly5.1 or CD90.1 congenic mice. For competitive transfers, a total of −/− + + + − − − + (B and C) Preactivated Irf4 and WT CD8 T cells remained untreated or 4 × 104 WT OT-I (CD90.1 CD90.2 )andIrf4 / OT-I (CD90.1 CD90.2 )cellsin β + − were treated with IL-21 for 24 h and Blimp-1 (B), IRF4 (C), or -Actin levels ratio 1:1 were injected i.v. into congenic mice (CD90.1 CD90.2 ). In some fi + −/− were determined by immunoblotting. (D) Puri ed CD8 T cells from Irf4 experiments, cells were stained with 2 μM of CFSE (Molecular Probes). Re- OT-I mice were preactivated as in A and transduced with retrovirus encoding cipient mice were infected with 105 cfu LmOVA. For reconstitution ex- IRF4 (IRF4-RV) or control retrovirus (Ctrl-RV). After 3 d, cells were rested for periments, MACS-purified WT Ly5.1+ CD8+ T cells were injected into Irf4−/− 24 h and then stimulated with IL-21. After 24 h, Prdm1 mRNA was determined recipients (5 × 106 per mouse) 1 d before L. monocytogenes infection. by qRT-PCR. Bars give the mean (± SD) of duplicate PCR samples. (A–D)The experiments were repeated two times with consistent results. (E) ChIP of IRF4 −/− + Isolation of Cells and Flow Cytometry. Lymphocytes were isolated from dif- was performed with preactivated Irf4 and WT CD8 T cells without stim- ferent tissues as previously described (36). For extracellular staining, cells ulation or treated with IL-21 for 1 h. Input DNA and precipitated DNA were were incubated with rat serum and anti-CD16/CD32 mAb and then stained quantified by qRT-PCR with primer pairs specific for regulatory regions of + with specific mAb as indicated. Ovalbumin-specific CD8 T cells were Prdm1; the same chromatin was used for control ChIP experiments with b detected with H-2K Ova257–264 dextramers (Immudex). The cells were ana- control IgG. Precipitated DNA is presented relative to input (% of input). lyzed by flow cytometry, using a FACS Canto II, a FACSCalibur, or an Aria III Values for nonspecific binding (as determined by using control IgG) were (Becton-Dickinson). Results were analyzed with the DIVA software (Becton- subtracted; nd, not detectable. Shown is mean ± SEM of combined results Dickinson) or the FlowJo Software (Tree Star). Fluorochrome-conjugated from three independent experiments (n = 9); ns, not significant. mAbs to CD3 (eBio500A2 or 145-2C11), CD4 (RM4-5), CD8α (53-6.7), CD25 (PC61.5), CD44 (IM7), Ly5.1 (A20), CD62L (MEL-14), CD90.1 (His51), CD90.2 (53-2.1), CD127/IL-7Rα (A7R34), CD244 (2B4), TCR-Vα2 (B20.1), CXCR3 epigenetic remodeling and energy metabolism (32). It is possible + (CXCR3-173), KLRG1 (2F1), LAG-3 (C9B7W), PD-1 (J43), Granzyme B (GB12), that BATF and IRF4 cooperatively affect CD8 T-cell effector IRF4 (3E4), Bcl2 (10C4), TNF-α (MP6-XT22), IFN-γ (XMG1.2), and IL-2 (JES6- differentiation either by regulation of effector proteins and TFs 5H4) were purchased from BioLegend, eBioscience, or BD Pharmingen. Flow or by promoting changes in metabolic pathways. In B cells and cytometric sorting was performed on an Aria III. + CD4 T cells, IRF4 additionally cooperates with several other TFs, including members of the ETS family, E47, NFATc2, Flow Cytometry-Based Assays. For the characterization of cytokine production, + −6 STAT3, and STAT6 (12). In CD8 T cells IRF4 might interact cells were incubated with 10 M of the peptides OVA257–264 (SIINFEKL) or LCMV gp – (KAVYNFATM) or of a mixture of the L. monocytogenes f-met with some of these molecules as well to influence differentiation. 33 41 −/− + peptides fMIGWII, fMIVIL, and fMIVTLF (all JPT) in the presence of brefeldin A Analysis of TFs in Irf4 CD8 T cells revealed up-regulation (Sigma Aldrich) for 4–5 h. Cells were stained extracellularly, fixed, and stained of BCL-6, Eomes, and Id3. We and others have recently de- intracellularly. Intracellular staining for GzmB and IRF4 was done without − − + scribed high Eomes expression in Irf4 / CD8 T cells (25, 26). prior stimulation. High expression of BCL-6, Eomes and Id3 and low expression of Anti-BrdU mAbs (BD Pharmingen) were used according to the manu- Blimp-1, Id2, and T-bet are associated with the development facturer’s protocols. Mice received 1 mg BrdU i.p. 1 d before analysis. Apo- + of memory CD8 T cells (1). However, only marginal numbers of ptosis was determined with the FLICA Kit (Immunochemistry Technologies) + − − ’ fi LmOVA-specific CD8 T cells were detectable in Irf4 / mice according to the manufacturer s protocol. Dead cells were identi ed by DAPI staining. and they mostly failed to produce cytokines upon stimulation. Thus, IRF4 controls maturation processes essential for the × 4 −/− + + Quantitative Real-Time PCR. A total of 4 10 WT or Irf4 OT-I CD8 T cells generation of both effector and memory CD8 T cells. The rapid were transferred into congenic WT mice, which were infected with LmOVA. induction of IRF4 after T-cell activation suggests that IRF4 acts Six days p.i., OT-I cells were FACS-sorted from spleens and total RNA was

Raczkowski et al. PNAS Early Edition | 5of6 Downloaded by guest on September 28, 2021 + prepared using the RNeasy Micro Kit (Qiagen). cDNA synthesis and PCR were In Vivo Cytotoxicity Assay. Single transfers of CD8 T cells from WT OT-I mice − − performed as described previously (37). mRNA expression levels were nor- or Irf4 / OT-I mice into congenic mice and LmOVA infection were done. On − − + malized to hypoxanthine–guanine phosphoribosyl transferase (Hprt1)expres- day 5 p.i., WT or Irf4 / OT-I CD8 T cells were recovered from spleens of sion and relative fold differences were calculated. The lowest experimental recipient mice by MACS purification and equalized numbers of cells (1 × 106) value was set to 1. The primer sets have been described previously (37). The were transferred into naive mice. One day later, target and control cells were primer pair for Bcl6 detection was forward, 5′-CCTGTGAAATCTGTGGCACTCG- −6 prepared by loading of spleen cells with 10 M of the peptides OVA257–264 3′, and reverse, 5′-CGCAGTTGGCTTTTGTGACG-3′. (target cells) or LCMV gp33–41 (control cells), which were then stained with 2 μMor0.2μM CFSE, respectively. Target and control cells were then mixed fi + Immunoblotting. Whole-cell lysates were prepared from puri ed CD8 T cells in a 1:1 ratio and a total of 6 × 106 cells were injected into the recipients of without stimulation or after in vitro stimulation. Immunoblotting was per- activated OT-I cells. Four hours later, spleen cells were isolated and killing fl + formed, as described previously (26). Brie y, proteins were fractionated by SDS/ as well as OT-I CD8 T-cell numbers were determined. Killing was calculated PAGE, transferred to nitrocellulose membrane, immunoblotted with pSTAT3 as follows: Tyr705 (9131; Cell Signaling Technology), IRF4 (M-17; sc6059; Santa Cruz) or . Blimp-1 (Novus) antibodies, and then reprobed with antibodies to total STAT3 % Killing = 100 − CFSEhisample=CFSElosample (124H6; 9139; Cell Signaling Technology) or β-Actin (Sigma-Aldrich). CFSEhicontrol=CFSElocontrol × 100 : + Chromatin Immunoprecipitation Assays. For ChIP experiments, CD8 T cells enriched by negative MACS selection were preactivated with plate-bound anti-CD3 mAb (2 μg/mL) and soluble anti-CD28 mAb (1 μg/mL) in the pres- Statistics. All described experiments were performed at least two times with ence of rhIL-2 (50 units/mL) for 3 d, rested overnight, and stimulated with similar results. For statistical analysis of frequencies and cell numbers, we 6 IL-21 (100 ng/mL) for 1 h. A total of 2 × 10 cells were fixed with 1% formal- applied an unpaired two-tailed Student’s t test. Bacterial titers were compared – dehyde for 10 min at room temperature to preserve the protein DNA using the Mann–Whitney test. P values are indicated as follows: *P < 0.05; interactions. Subsequently, ChIP was performed as described previously (38) **P < 0.01; ***P < 0.001. Additional methods can be found in SI Material with antibodies against IRF4 (M-17; Santa Cruz). Quantitative RT-PCR with and Methods. the precipitated chromatin was performed to calculate the percentage of input. Primer sequences are provided in Table S1. All amplifications were ACKNOWLEDGMENTS. The authors thank Dr. Hao Shen for providing re- performed in triplicate with SYBR Green PCR Master Mix (Qiagen). Control combinant L. monocytogenes strains and Dr. Timo Lischke and Bärbel ChIP was performed with a respective isotype control antibody to ensure Camara for technical advice and assistance. This work was supported by specificity. After normalization of the data according to the isotype control, the Deutsche Forschungsgemeinschaft through grants to M.H. (HU 1824/2- the specific pulldown (percentage of input chromatin) was calculated. 1) and H.-W.M. (MI 476/3, SFB841, and KFO228) as well as to P.I.P. (SFB/TR22).

1. Kaech SM, Cui W (2012) Transcriptional control of effector and memory CD8+ Tcell 20. Tussiwand R, et al. (2012) Compensatory dendritic cell development mediated by differentiation. Nat Rev Immunol 12(11):749–761. BATF-IRF interactions. Nature 490(7421):502–507. 2. Ichii H, et al. (2002) Role for Bcl-6 in the generation and maintenance of memory CD8+ 21. Li P, et al. (2012) BATF-JUN is critical for IRF4-mediated transcription in T cells. Nature Tcells.Nat Immunol 3(6):558–563. 490(7421):543–546. 3. Cui W, Liu Y, Weinstein JS, Craft J, Kaech SM (2011) An interleukin-21-interleukin-10- 22. Ciofani M, et al. (2012) A validated regulatory network for Th17 cell specification. Cell + STAT3 pathway is critical for functional maturation of memory CD8 T cells. Immunity 151(2):289–303. – 35(5):792 805. 23. Brass AL, Kehrli E, Eisenbeis CF, Storb U, Singh H (1996) Pip, a lymphoid-restricted IRF, 4. Intlekofer AM, et al. (2005) Effector and memory CD8+ T cell fate coupled by T-bet contains a regulatory domain that is important for autoinhibition and ternary com- and eomesodermin. Nat Immunol 6(12):1236–1244. plex formation with the Ets factor PU.1. Genes Dev 10(18):2335–2347. 5. Ji Y, et al. (2011) Repression of the DNA-binding inhibitor Id3 by Blimp-1 limits the 24. Mittrücker HW, et al. (1997) Requirement for the transcription factor LSIRF/IRF4 for formation of memory CD8+ T cells. Nat Immunol 12(12):1230–1237. mature B and T lymphocyte function. Science 275(5299):540–543. 6. Yang CY, et al. (2011) The transcriptional regulators Id2 and Id3 control the formation 25. Nayar R, et al. (2012) TCR signaling via Tec kinase ITK and interferon regulatory factor 4 of distinct memory CD8+ T cell subsets. Nat Immunol 12(12):1221–1229. + – 7. Zhou X, et al. (2010) Differentiation and persistence of memory CD8(+) T cells depend (IRF4) regulates CD8 T-cell differentiation. Proc Natl Acad Sci USA 109(41):E2794 E2802. + on T cell factor 1. Immunity 33(2):229–240. 26. Huber M, et al. (2013) IL-17A secretion by CD8 T cells supports Th17-mediated au- 8. Joshi NS, et al. (2007) Inflammation directs memory precursor and short-lived effector toimmune encephalomyelitis. J Clin Invest 123(1):247–260. CD8(+) T cell fates via the graded expression of T-bet transcription factor. Immunity 27. Pamer EG (2004) Immune responses to Listeria monocytogenes. Nat Rev Immunol 27(2):281–295. 4(10):812–823. 9. Cannarile MA, et al. (2006) Transcriptional regulator Id2 mediates CD8+ T cell im- 28. Kalia V, et al. (2010) Prolonged interleukin-2Ralpha expression on virus-specific CD8+ munity. Nat Immunol 7(12):1317–1325. T cells favors terminal-effector differentiation in vivo. Immunity 32(1):91–103. 10. Kallies A, Xin A, Belz GT, Nutt SL (2009) Blimp-1 transcription factor is required for the 29. Wherry EJ (2011) T cell exhaustion. Nat Immunol 12(6):492–499. differentiation of effector CD8(+) T cells and memory responses. Immunity 31(2):283–295. 30. Kwon H, et al. (2009) Analysis of interleukin-21-induced Prdm1 gene regulation reveals 11. Rutishauser RL, et al. (2009) Transcriptional repressor Blimp-1 promotes CD8(+) T cell functional cooperation of STAT3 and IRF4 transcription factors. Immunity 31(6):941–952. terminal differentiation and represses the acquisition of central memory T cell 31. Cretney E, et al. (2011) The transcription factors Blimp-1 and IRF4 jointly control the properties. Immunity 31(2):296–308. differentiation and function of effector regulatory T cells. Nat Immunol 12(4):304–311. 12. De Silva NS, Simonetti G, Heise N, Klein U (2012) The diverse roles of IRF4 in late 32. Kuroda S, et al. (2011) Basic leucine zipper transcription factor, ATF-like (BATF) reg- – germinal center B-cell differentiation. Immunol Rev 247(1):73 92. ulates epigenetically and energetically effector CD8 T-cell differentiation via Sirt1 13. Brüstle A, et al. (2007) The development of inflammatory T(H)-17 cells requires in- expression. Proc Natl Acad Sci USA 108(36):14885–14889. terferon-regulatory factor 4. Nat Immunol 8(9):958–966. 33. Crotty S, Johnston RJ, Schoenberger SP (2010) Effectors and memories: Bcl-6 and 14. Lohoff M, et al. (2002) Dysregulated T helper cell differentiation in the absence of Blimp-1 in T and B lymphocyte differentiation. Nat Immunol 11(2):114–120. interferon regulatory factor 4. Proc Natl Acad Sci USA 99(18):11808–11812. 34. Foulds KE, et al. (2002) Cutting edge: CD4 and CD8 T cells are intrinsically different in 15. Rengarajan J, et al. (2002) Interferon regulatory factor 4 (IRF4) interacts with NFATc2 their proliferative responses. J Immunol 168(4):1528–1532. to modulate interleukin 4 . J Exp Med 195(8):1003–1012. 35. Kursar M, et al. (2005) Differential requirements for the chemokine CCR7 in T 16. Staudt V, et al. (2010) Interferon-regulatory factor 4 is essential for the developmental – program of T helper 9 cells. Immunity 33(2):192–202. cell activation during Listeria monocytogenes infection. J Exp Med 201(9):1447 1457. 17. Chen Q, et al. (2008) IRF-4-binding protein inhibits interleukin-17 and interleukin-21 pro- 36. Mittrücker HW, et al. (2007) Poor correlation between BCG vaccination-induced T cell duction by controlling the activity of IRF-4 transcription factor. Immunity 29(6):899–911. responses and protection against tuberculosis. Proc Natl Acad Sci USA 104(30):12434–12439. 18. Bollig N, et al. (2012) Transcription factor IRF4 determines germinal center formation 37. Huber M, et al. (2009) A Th17-like developmental process leads to CD8(+) Tc17 cells through follicular T-helper cell differentiation. Proc Natl Acad Sci USA 109(22):8664–8669. with reduced cytotoxic activity. Eur J Immunol 39(7):1716–1725. 19. Glasmacher E, et al. (2012) A genomic regulatory element that directs assembly and 38. Brand S, et al. (2011) Epigenetic regulation in murine offspring as a novel mechanism for function of immune-specific AP-1-IRF complexes. Science 338(6109):975–980. transmaternal asthma protection induced by microbes. J Allergy Clin Immunol 128:618–625.

6of6 | www.pnas.org/cgi/doi/10.1073/pnas.1309378110 Raczkowski et al. Downloaded by guest on September 28, 2021