1. Immunology and Inflammation
Download icon

c-Maf restrains T-bet-driven programming of CCR6-negative group 3 innate lymphoid cells

  1. Caroline Tizian
  2. Annette Lahmann
  3. Oliver Hölsken
  4. Catalina Cosovanu
  5. Michael Kofoed-Branzk
  6. Frederik Heinrich
  7. Mir-Farzin Mashreghi
  8. Andrey Kruglov
  9. Andreas Diefenbach
  10. Christian Neumann  Is a corresponding author
  1. Laboratory of Innate Immunity, Department of Microbiology, Infectious Diseases and Immunology, Charité-Universitätsmedizin Berlin, Germany
  2. Mucosal and Developmental Immunology, Deutsches Rheuma-Forschungszentrum, Germany
  3. Chronic Immune Reactions, Deutsches Rheuma-Forschungszentrum, Germany
  4. Therapeutic Gene Regulation, Deutsches Rheuma-Forschungszentrum, Germany
  5. Chronic Inflammation, Deutsches Rheuma-Forschungszentrum, Germany
  6. Belozersky Institute of Physico-Chemical Biology and Biological Faculty, M.V. Lomonosov Moscow State University, Russian Federation
Short Report
  • Cited 2
  • Views 1,087
  • Annotations
Cite this article as: eLife 2020;9:e52549 doi: 10.7554/eLife.52549

Abstract

RORγt+ group 3 innate lymphoid cells (ILC3s) maintain intestinal homeostasis through secretion of type 3 cytokines such as interleukin (IL)−17 and IL-22. However, CCR6- ILC3s additionally co-express T-bet allowing for the acquisition of type 1 effector functions. While T-bet controls the type 1 programming of ILC3s, the molecular mechanisms governing T-bet are undefined. Here, we identify c-Maf as a crucial negative regulator of murine T-bet+ CCR6- ILC3s. Phenotypic and transcriptomic profiling of c-Maf-deficient CCR6- ILC3s revealed a hyper type 1 differentiation status, characterized by overexpression of ILC1/NK cell-related genes and downregulation of type 3 signature genes. On the molecular level, c-Maf directly restrained T-bet expression. Conversely, c-Maf expression was dependent on T-bet and regulated by IL-1β, IL-18 and Notch signals. Thus, we define c-Maf as a crucial cell-intrinsic brake in the type 1 effector acquisition which forms a negative feedback loop with T-bet to preserve the identity of CCR6- ILC3s.

Introduction

Tissue-resident, RORγt-dependent group 3 innate lymphoid cells (ILC3s) protect mucosal surfaces against infections and maintain the integrity of the epithelial barrier by secretion of cytokines such as IL-17 and IL-22 (Sonnenberg et al., 2012; Gronke et al., 2019; Hernández et al., 2015; Zheng et al., 2008; Gladiator et al., 2013; Satoh-Takayama et al., 2008; Zenewicz et al., 2008). In mice, ILC3s consist of two major subsets, CCR6+ ILC3s that include lymphoid tissue inducer (LTi) cells and CCR6- ILC3s that can differentiate to cells expressing type 1 effector molecules, such as IFN-γ and the natural cytotoxic receptor NKp46 (Sawa et al., 2010; Klose et al., 2013). Functionally, fetal LTi and adult LTi-like CCR6+ ILC3s are essential for lymphoid organ development (Sun et al., 2000; Eberl et al., 2004), while NKp46+ CCR6- ILC3s are implicated in type 1 inflammatory immune responses and thus may also have pathogenic functions during intestinal inflammation (Klose et al., 2013; Powell et al., 2012; Buonocore et al., 2010; Bernink et al., 2013; Vonarbourg et al., 2010; Rankin et al., 2016; Song et al., 2015).

The transcription factor (TF) RORγt is strictly required for the development of all ILC3s, as mice deficient for RORγt lack all ILC3 subsets (Eberl et al., 2004; Sanos et al., 2009). RORγt also controls the functionality of ILC3s by regulating the production of effector cytokines, such as IL-17 and IL-22 (Ivanov et al., 2006; Rutz et al., 2013). Interestingly, CCR6- ILC3 co-express RORγt and the master regulator of type 1 immunity, T-bet. T-bet is key to the differentiation of NKp46+ CCR6- ILC3s, as those cells fail to develop in mice lacking the gene encoding T-bet, Tbx21 (Klose et al., 2013; Rankin et al., 2013). Importantly, T-bet not only contributes to NKp46+ CCR6- ILC3 development, but an increasing T-bet gradient enables functional plasticity of NKp46+ CCR6- ILC3s by instructing a type 1 effector program in ILC3s (Klose et al., 2013; Sciumé et al., 2012; Klose et al., 2014; Cella et al., 2019). Tunable T-bet expression in NKp46+ CCR6- ILC3s serves as a dynamic molecular switch from a type 3 to a type 1 phenotype (Klose et al., 2013). Once T-bet expression reaches a sufficient level, it can also act as a repressor of RORγt, resulting eventually in a full conversion of ILC3s to ILC1-like cells (referred to as ILC3-to-1 plasticity) (Vonarbourg et al., 2010; Cella et al., 2019; Bernink et al., 2015). Thus, the balance between RORγt versus T-bet expression dictates the fate and function of CCR6- ILC3s (Fang and Zhu, 2017).

Importantly, the molecular mechanisms controlling the dynamic and quantitative co-expression of RORγt and T-bet in CCR6- ILC3s are largely undefined. Several extrinsic signals were shown to promote or restrain T-bet-dependent plasticity, most prominently cues from the microbiota, IL-23, IL-7 and Notch signaling (Klose et al., 2013; Sanos et al., 2009; Rankin et al., 2013; Viant et al., 2016; Chea et al., 2016). Moreover, exposure to pro-inflammatory cytokines, such as IL-12, IL-15 and IL-18, was reported to further support transdifferentiation to an ILC1-like fate (Vonarbourg et al., 2010; Bernink et al., 2015; Satoh-Takayama et al., 2010). However, despite this, the intrinsic molecular mediators governing ILC3 plasticity have not been discovered yet.

In the past, our group and others could identify the AP-1 TF c-Maf as a central regulator of RORγt+ CD4+ T cells, including RORγt+ Foxp3+ Treg cells (Neumann et al., 2019; Xu et al., 2018; Wheaton et al., 2017), RORγt+ Th17 cells (Ciofani et al., 2012; Aschenbrenner et al., 2018; Tanaka et al., 2014) and RORγt+ γδ T cells (Zuberbuehler et al., 2019), both in mouse and human. Specifically, c-Maf was shown to directly bind and regulate key genes of RORγt+ T cells, including IL-22 and RORγt itself (Tanaka et al., 2014; Zuberbuehler et al., 2019; Rutz et al., 2011). Recently, a broad transcriptional network analysis also identified c-Maf as an important regulator of the ILC3-ILC1 balance, although the precise underlying molecular mechanisms have remained unclear (Pokrovskii et al., 2019).

Here, we demonstrate that c-Maf was essential for CCR6- ILC3s to establish a physiological equilibrium between type 1 and type 3 effector states. c-Maf directly restrained T-bet expression, thereby preventing CCR6- ILC3s from acquiring excessive type 1 effector functions. c-Maf expression itself was dependent on T-bet and tightly correlated with its expression level. Upstream, we identified IL-1ß- and IL-18-mediated NF-κB, as well as Notch signals, as potent extrinsic enhancers of c-Maf expression in CCR6- ILC3s. Thus, our data define c-Maf as an integral regulator within the type 3-to-1 conversion program that acts as a cell-intrinsic gatekeeper of T-bet expression to maintain the function and lineage-stability of CCR6- ILC3s.

Results and discussion

c-Maf specifically preserves the type 3 identity of CCR6- ILC3s

Given the pivotal role of c-Maf in CD4+ T cells, we aimed to define its function in ILCs, which share a similar transcriptional program with T cells (Vivier et al., 2018). We first investigated the expression pattern of c-Maf in different ILC subsets of the small intestinal lamina propria (siLP) by staining for c-Maf. This analysis showed that ILC3s expressed higher levels of c-Maf when compared to ILC1s or ILC2s (Figure 1A, gating strategy see Figure 1—figure supplement 1). Among the ILC3 subsets, c-Maf was particularly highly expressed by NKp46+ CCR6- ILC3s at levels comparable to RORγt+ CD4+ T cells (Figure 1B). Collectively, these data suggested a potential function of c-Maf in these cells.

Figure 1 with 1 supplement see all
c-Maf is required to maintain the type 3 phenotype of CCR6- ILC3s.

(A) Protein expression of c-Maf by ILC1s, ILC2s and ILC3s isolated from the siLP of naïve C57BL/6N wild-type mice as measured by flow cytometry (pregated on viable Lin- CD127+ cells). Representative histograms show c-Maf expression. (B) Representative histograms showing expression of c-Maf by different siLP ILC3 subsets and RORγt+ T cells as comparison. (C) Frequencies and total numbers of ILC1s, ILC2s and ILC3s from siLP of RorcCre Maffl/fl and control mice as measured by flow cytometry (pregated on viable Lin- CD127+ cells). Representative flow cytometric profiles of RORγt vs. T-bet and GATA3 expression (left); right, quantification (n = 5–6, mean ± SEM, *p<0.05, **p<0.01). (D) Total numbers of different siLP ILC3 subsets of RorcCre Maffl/fl and control mice as measured by flow cytometry. Representative flow cytometric profiles of NKp46 vs. CCR6 expression (left); right, quantification (n = 5–6, mean ± SEM, *p<0.05). (E) Expression of RORγt by different siLP ILC3 subsets from RorcCre Maffl/fl and control mice. Representative histograms (upper panel) show RORγt expression. Graphs below show quantification of RORγt gMFI (n = 5–6, mean ± SEM, **p<0.01, ***p<0.001). (F, G) Frequencies of IL-17A and IL-22 positive cells among different siLP ILC3 subsets of RorcCreMaffl/fl and control mice after ex vivo restimulation with PMA/ionomycin and IL-23. Representative flow cytometric profiles of IL-17A or IL-22 vs. CCR6 are shown on the left; quantification on the right (n = 4, mean ± SEM, *p<0.05, **p<0.01). Data are representative of three independent experiments. Statistical differences were tested using an unpaired Students’ t-test (two-tailed).

In order to directly study the role of c-Maf in ILC3s, we crossed mice carrying floxed Maf alleles (Maffl/fl) to mice expressing Cre recombinase driven by the regulatory elements of the Rorc(γt) gene locus (Rorc-CreTg), thereby generating mice with a specific deletion of c-Maf in RORγt+ ILC3s and T cells (RorcCre Maffl/fl). Phenotypic analysis of siLP ILCs revealed decreased frequencies of ILC3s in the absence of c-Maf, whereas frequencies and total numbers of ILC1s and ILC2s were increased in RorcCre Maffl/fl mice when compared to littermate controls (Figure 1C). Among ILC3s, we detected a selective loss of NKp46- CCR6- ILC3s in RorcCre Maffl/fl mice, while total numbers of NKp46+ CCR6- ILC3s and CCR6+ ILC3s were not significantly changed (Figure 1D).

Thus, c-Maf was selectively required for the maintenance of intestinal NKp46- CCR6- ILC3s. Its absence in ILC3s resulted in significant changes in the proportions of individual ILC subsets in the gut. Importantly, flow cytometric intracellular staining also revealed a significant downregulation of RORγt protein levels in c-Maf-deficient NKp46+ and NKp46- CCR6- ILC3s, which was not observed in CCR6+ ILC3s (Figure 1E).

Next, we tested the functionality of c-Maf-deficient ILC3s by assessing their capacity to produce the type 3 signature cytokines IL-17A and IL-22. In line with the decrease in RORγt expression, NKp46+ and NKp46- CCR6- ILC3s from RorcCre Maffl/fl mice exhibited significantly reduced frequencies of IL-17A and IL-22 producers after ex vivo restimulation as compared to control cells (Figure 1F and G). Again, no differences were detected in CCR6+ ILC3s, corroborating the selective role of c-Maf for the homeostasis and function of CCR6- ILC3s. Notably, c-Maf did not act as a repressor of IL-22 production in ILC3s, as it was shown for Th17 cells (Rutz et al., 2011), indicating essential differences in c-Maf function between ILCs and T cells.

In summary, these findings demonstrated a crucial requirement of c-Maf in maintaining the type 3 identity of CCR6- ILC3s, including their expression of RORγt, IL-17A and IL-22.

c-Maf suppresses the acquisition of type 1 properties by CCR6- ILC3s

RORγt+ CCR6- ILC3s have the capacity to acquire type 1 effector characteristics, such as IFN-γ and NKp46 expression (Vonarbourg et al., 2010). This differentiation is facilitated by graded co-expression of T-bet, together with RORγt, enabling functional flexibility during inflammatory immune responses (Klose et al., 2013; Powell et al., 2012; Bernink et al., 2013; Rankin et al., 2013; Sciumé et al., 2012). However, the molecular mechanisms governing the T-bet-dependent type 1 programming of CCR6- ILC3s are undefined.

Interestingly, c-Maf expression strongly correlated with T-bet expression in both NKp46+ and NKp46- CCR6- ILC3s (Figure 2A). More importantly, c-Maf deficiency resulted in strong upregulation of T-bet and NKp46 expression, both on the population (frequencies) and at single cell level (gMFI), in CCR6- ILC3s (Figure 2B). These data, together with the selective loss of NKp46- CCR6- ILC3s, which are considered to contain precursors of NKp46+ CCR6- ILC3s, in RorcCreMaffl/fl mice (Figure 1D), suggested an amplified type 1 conversion of CCR6- ILC3s in the absence of c-Maf.

Figure 2 with 2 supplements see all
CCR6- ILC3s acquire a hyper type 1 phenotype in the absence of c-Maf.

(A) Representative flow cytometric profiles of T-bet vs. c-Maf expression by siLP NKp46+ and NKp46- CCR6- ILC3s. (B) T-bet and NKp46 expression (frequency and gMFI) by siLP CCR6- ILC3s of RorcCreMaffl/fl and control mice. Representative flow cytometric profiles are shown on the left; quantification on the right (n = 4, mean ± SEM, **p<0.01, ***p<0.001). (C) Frequency of IFN-γ positive cells among siLP CCR6- ILC3s of RorcCreMaffl/fl and control mice after ex vivo restimulation with PMA/ionomycin and IL-23. Representative flow cytometric profiles are shown on the left; quantification on the right (n = 4, mean ± SEM, ***p<0.001). (D) Analysis of mixed bone marrow chimeras of CD45.1+ wild-type and CD45.2+ RorcCreMaffl/f (Maf KO) bone marrow cells. Pregated on viable Lin- CD127+ CD90.2+ RORγt+ cells. Recipient mice were CD90.1 positive. Representative flow cytometric profiles of T-bet vs. CCR6 expression are shown left; quantification on the right (n = 7, mean ± SEM, **p<0.01). All statistical differences were tested using an unpaired Students’ t-test (two-tailed). (E) NKp46+ CCR6- ILC3s were sorted from siLP of RorcCreMaffl/fl and control mice and subjected to RNA sequencing. Vulcano plot showing comparison of gene expression between c-Maf-deficient and control NKp46+ CCR6- ILC3s. Data represent the combined analysis of three biologically independent samples. Genes considered significant (FC > 1.5, FDR < 0.05) fall into the grey background, while selected genes are highlighted in red. (F) Gene set enrichment plots showing enrichment of ILC1 and NK cell signature genes in c-Maf-deficient vs. control NKp46+ CCR6- ILC3s (FDR < 0.01). Normalized enrichment score (NES).

In accordance with the increase in T-bet expression, we also detected increased frequencies of IFN-γ producing cells within c-Maf-deficient CCR6- ILC3s as compared to c-Maf-competent control cells (Figure 2C). Of note, Ki67 staining of c-Maf-deficient CCR6- ILC3s was not altered, ruling out that proliferative differences facilitated the skewing towards a type 1 phenotype in the absence of c-Maf (Figure 2—figure supplement 1A).

In addition to the lack of c-Maf expression in ILC3s, RorcCre Maffl/fl mice also harbour a c-Maf-deficient T cell compartment, due to the expression of RORγt during T cell development (Sun et al., 2000). Moreover, conditional deletion of c-Maf in T cells (Cd4Cre Maffl/fl or Foxp3Cre Maffl/fl) was reported to cause disturbances in intestinal homeostasis (Neumann et al., 2019; Imbratta et al., 2019), raising the possibility that changes in the gut microenvironment contributed to the ‘hyper type 1’ phenotype of CCR6- ILC3s in RorcCre Maffl/fl mice. To interrogate this scenario, we analysed intestinal CCR6- ILC3s from Cd4Cre Maffl/fl mice. The expression of T-bet by CCR6- ILC3s was not altered when c-Maf was selectively deleted in T cells, excluding that T cell-dependent alterations affected the type 1 conversion of ILC3s in RorcCre Maffl/fl mice (Figure 2—figure supplement 1B).

Interestingly, a small subset of CCR6- ILC3s also co-expresses CD4 (ca. 2% of CCR6- ILC3s, Figure 2—figure supplement 1C). Indeed, c-Maf staining of CD4+ CCR6- ILC3s from Cd4Cre Maffl/fl mice confirmed c-Maf deletion in a substantial fraction of cells (ca. 50%), allowing us to compare c-Maf-competent and c-Maf-deficient CCR6- ILC3s within the same mouse (Figure 2—figure supplement 1C). We detected higher expression of T-bet and NKp46 in cells lacking c-Maf as compared to their c-Maf-expressing counterparts, indicating that the type 1 deviation of c-Maf-deficient CCR6- ILC3s was a cell-intrinsic phenomenon (Figure 2—figure supplement 1C). Importantly, mixed bone marrow chimeras of wild-type and c-Maf-deficient cells confirmed the cell-intrinsic function of c-Maf in restraining T-bet expression (Figure 2D).

Together, these data identified c-Maf as a potent repressor of the T-bet-dependent type 1 conversion of CCR6- ILC3s. Notably, c-Maf has also been shown to attenuate the differentiation of Th1 cells by a yet-to-be-defined mechanism (Ho et al., 1998), suggesting a broader role of c-Maf in restraining type 1 immune effector cells.

c-Maf globally restrains type 1 and NK cell specific gene programs in CCR6- ILC3s

To better define c-Maf-dependent gene regulation, we performed RNA sequencing (RNA-seq) of siLP ILC3s of RorcCreMaffl/fl and control mice. We isolated NKp46+ CCR6- ILC3s and as a comparison CCR6+ ILC3s for our analysis (sorting strategy see Figure 2—figure supplement 2A). Principal component analysis showed that c-Maf-deficient ILC3s clustered separately from their wild-type counterparts, demonstrating a unique role of c-Maf in shaping the transcriptome of ILC3s (Figure 2—figure supplement 2B).

We identified 941 genes as differentially expressed (FC > 1.5, p-value < 0.05) between c-Maf-deficient and control NKp46+ CCR6- ILC3s (Figure 2E, Supplementary file 1). Interestingly, only ca. 30% of these genes (294 genes) were also found to be differentially expressed in CCR6+ ILC3s (Figure 2—figure supplement 2C, Supplementary file 2), indicating a distinct function of c-Maf in NKp46+ CCR6- ILC3s. Consistent with our flow cytometry data, we detected a striking enrichment of type 1 and NK cell-related genes (e.g. Tbx21, Ifng, Nos2, Klri2, Klrc1, Klrc2, Gzmb, Gzmd, Fasl) in c-Maf-deficient NKp46+ CCR6- ILC3s (Figure 2E). In line with this, Bcl6 and Irf8, both recently described as crucial transcription factors promoting ILC1/NK cell gene programs (Pokrovskii et al., 2019; Adams et al., 2018), were also significantly upregulated in the absence of c-Maf in NKp46+ CCR6- ILC3s (Figure 2E, Figure 2—figure supplement 2D and E). Conversely, several type 3 effector genes (e.g. Il17a, Il17f, Il17re, Ccl20) were downregulated in NKp46+ CCR6- ILC3s upon c-Maf-deficiency (Figure 2E).

To better understand the global c-Maf-dependent changes in gene expression programs, we performed gene set enrichment analysis (GSEA). Given the strong upregulation of type 1 and NK cell features in c-Maf-deficient CCR6- ILC3s, we made use of published RNA-seq data comparing ILC1s and NK cells with CCR6- ILC3s (Pokrovskii et al., 2019). In detail, we created sets of genes that were most highly overexpressed in ILC1s or NK cells, thereby defining ILC1 and NK cell gene signatures that distinctly separated those lineages from CCR6- ILC3s. Importantly, when applied to GSEA, both gene signatures were significantly enriched in c-Maf-deficient NKp46+ CCR6- ILC3s (Figure 2F), indicating a global shift in gene expression towards an ILC1/NK cell phenotype.

In summary, these data demonstrated that c-Maf was essential to globally balance type 1 and type 3 gene programs within CCR6- ILC3s. This function of c-Maf was specific to NKp46+ CCR6- ILC3s as opposed to CCR6+ ILC3s, most likely due to the particularly high c-Maf expression and the selective accessible of type 1 gene loci in these cells (Pokrovskii et al., 2019; Shih et al., 2016). In the absence of c-Maf, NKp46+ CCR6- ILC3s downregulated type 3 effector genes, while overexpressing numerous genes encoding for type 1 and cytotoxic effector molecules. The latter finding is particularly interesting, since NKp46+ CCR6- ILC3s are largely considered to be non-toxic cells (Melo-Gonzalez and Hepworth, 2017). Nevertheless, NKp46+ CCR6- ILC3s share a considerable transcriptional overlap with ILC1s, which also exhibit a degree of cytotoxic capacity (Robinette et al., 2015; Cortez and Colonna, 2016). Thus, c-Maf-deficiency may result in marked cytotoxicity of NKp46+ CCR6- ILC3s. More work is needed to precisely define the role of c-Maf for ILC3 functionality during homeostasis and in the context of intestinal inflammation.

c-Maf restrains T-bet expression by directly repressing the Tbx21 promoter

The strong upregulation of T-bet expression in c-Maf-deficient CCR6- ILC3s raised the possibility that c-Maf acted as a direct repressor of T-bet, thereby restraining the type 1 differentiation program. Indeed, in silico analysis identified Maf response elements (MARE) within the Tbx21 promoter, as well as in a conserved distant Tbx21 enhancer (Kataoka et al., 1994; Yang et al., 2007), both regions accessible in NKp46+ CCR6- ILC3s as evidenced by ATAC-sequencing (Shih et al., 2016; Figure 3A and B).

Figure 3 with 3 supplements see all
c-Maf directly represses T-bet as part of a negative feedback loop.

(A) Representative ATAC sequencing tracks across the Tbx21 locus of NKp46+ ILC3s (Shih et al., 2016). The Tbx21 promoter and enhancer regions are highlighted in red (Yang et al., 2007). ATAC sequencing tracks were visualized using the WashU browser from the Cistrome project (Mei et al., 2017). (B) Positioning of putative Maf response elements (MARE) within the Tbx21 promoter and enhancer (highlighted in blue). Depicted sequences represent selected regions of the Tbx21 promoter and enhancer. Mutated sequences are shown in lower lines in red. (C) Schematic representation of plasmids containing the Tbx21 promoter/enhancer linked to the firefly luciferase reporter gene (Luc). Blue dots indicate MARE sites within the Tbx21 promoter and enhancer region. (D) Relative luciferase activity (RLU) of different reporter constructs driven by the Tbx21 promoter alone or in combination with an enhancer sequence compared to a promoterless (w/o) control vector (pGL3 basic) (n = 3, mean ± SEM, *p<0.05, ***p<0.001). (E) Analysis of c-Maf-dependent suppression of luciferase activity. The MARE sites within the Tbx21 promoter or the Tbx21 promoter and enhancer were mutated in the Tbx21 enhancer/promoter construct. Comparison of RLU between unmutated and mutated Tbx21 enhancer/promoter constructs upon c-Maf overexpression (n = 3, mean ± SEM, ***p<0.001). All reporter assay data are pooled from three independent experiments. (F) Expression of c-Maf by siLP CCR6- ILC3s from T-bet-deficient (T-bet-/-) and -sufficient (T-bet+/+) mice. Representative flow cytometric profiles are shown on the left; graph on the right shows quantification of c-Maf gMFI (n = 4, mean ± SEM, ***p<0.001). (G–H) Sort-purified siLP NKp46+ CCR6- ILC3s from RorcCre R26EYFP mice were cultured in vitro for 36 hr in the presence of IL-7/SCF (w/o) or IL-7/SCF plus indicated cytokines. Subsequently, Maf expression was measured by qPCR (n = 4, mean ± SEM, *p<0.05, **p<0.01, ***p<0.001). In one condition the NF-κb inhibitor BMS-345541 (BMS) was added at 1 μM to the culture. Data are pooled from two independent experiments each with two replicate wells. (I) Sort-purified siLP NKp46+ CCR6- ILC3s from RorcCre R26EYFP mice were cultured in the presence of IL-7/SCF on OP9 or OP9-DLL1 stromal cells as indicated. After 12 days, cells were analysed by flow cytometry for the cell surface expression of NKp46. Representative contour plots are shown on the left (pregated on CCR6- ILC3s); graph on the right shows quantification of the frequency of NKp46+ cell among CCR6- ILC3s (n = 10, mean ± SEM, ***p<0.001). Data are pooled from two independent experiments with 4 to 6 replicate wells. (J) Expression of c-Maf by NKp46+ CCR6- ILC3s cultured on OP9 or OP9-DLL1 cells. Representative histogram is shown left; graph on the right shows quantification of c-Maf gMFI (n = 10, mean ± SEM, ***p<0.001). All statistical differences were tested using an unpaired Students’ t-test (two-tailed).

In order to study the transcriptional activity of c-Maf at these sites, we cloned the Tbx21 promoter alone or in conjunction with the Tbx21 enhancer upstream of a luciferase reporter (Figure 3C). Indeed, the Tbx21 promoter showed strong transcriptional activity when compared to a promoterless control vector (Figure 3D). The Tbx21 enhancer further increased this activity when cloned upstream of the Tbx21 promoter (Figure 3D). Importantly, upon exogenous overexpression of c-Maf, we detected a strong de-repression of the reporter signal when we mutated the MARE sites within the Tbx21 promoter (Figure 3B and E). Notably, mutating the Tbx21 enhancer did not result in further increase of reporter activity, indicating that c-Maf facilitated its suppressive function mainly by acting on the Tbx21 promoter (Figure 3B and E).

Thus, these data identified the Tbx21 promoter as a c-Maf-responsive region through which c-Maf directly controls T-bet expression.

c-Maf expression in CCR6- ILC3s is dependent on T-bet

Despite the fact that c-Maf acted as a direct repressor of T-bet, we observed strong correlation of c-Maf expression with T-bet expression in CCR6- ILC3s (Figure 2A). This finding let us hypothesize that c-Maf expression was co-regulated with T-bet expression as part of the type one conversion program. To explore this hypothesis, we studied c-Maf expression in CCR6- ILC3s from T-bet-deficient mice. Indeed, in the absence of T-bet, c-Maf expression was strongly reduced when compared to T-bet-sufficient cells (Figure 3F), indicating that T-bet positively regulated its own repressor to establish an equilibrated state within the ILC3-to-ILC1 continuum.

In silico analysis of the Maf locus using ATAC-Seq (Shih et al., 2016) and T-bet ChIP-Seq (Gökmen et al., 2013) data further identified a striking overlap of T-bet binding peaks with open chromatin regions in NKp46+ CCR6- ILC3s within two conserved non-coding sequences (CNS-0.5 and CNS-1) upstream of Maf, suggesting that T-bet directly controlled c-Maf expression in CCR6- ILC3s (Figure 3—figure supplement 1).

IL-1β- and IL-18-mediated NF-κb and Notch signalling promote c-Maf expression in CCR6- ILC3s

Next, we aimed to identify signals that control c-Maf expression in CCR6- ILC3s. Several cytokines were shown to regulate the differentiation and function of T-bet+ CCR6- ILC3s, including IL-1ß, IL-12, IL-15, IL-18 and IL-23 (Klose et al., 2013; Vonarbourg et al., 2010; Bernink et al., 2015; Satoh-Takayama et al., 2010). Thus, we short-term stimulated sort-purified NKp46+ CCR6- ILC3s with these cytokines in vitro and subsequently assessed their Maf expression by qPCR. Among all tested conditions, the pro-inflammatory IL-1 family cytokines IL-1ß and IL-18 stood out as potent inducers of Maf expression (Figure 3G). IL-1ß/IL-18 stimulation also induced Ifng and Il22 expression (Figure 3—figure supplement 2A). In fact, c-Maf-deficient NKp46+ CCR6- ILC3s expressed more Ifng in response to IL-1ß/IL-18 as compared to controls (Figure 3—figure supplement 2A), functionally connecting c-Maf expression downstream of these stimuli with the control of the type 1 effector response.

Since both IL-1ß and IL-18 signal via NF-κb and a conserved NF-κb binding site is present within CNS-0.5 upstream of Maf (Figure 3—figure supplement 1C and D), we hypothesized that NF-κb is involved in regulation of c-Maf expression downstream of IL-1ß/IL-18. Indeed, pharmacological inhibition of NF-κb signalling completely abrogated the cytokine-mediated induction of c-Maf expression (Figure 3H). In addition to IL-1ß/IL-18, we found that IL-12 suppressed Maf expression (Figure 3G). However, IL-12 did not interfere with the IL-1ß/IL-18-mediated c-Maf induction nor was c-Maf expression altered in NKp46+ CCR6- ILC3s from Il12a-/- mice as compared to controls (Figure 3—figure supplement 2B and C), questioning a dominant role of IL-12 in c-Maf regulation.

Of note, the cytokine-mediated induction of Maf expression was accompanied by a concomitant decrease in Tbx21 expression (Figure 3—figure supplement 2D). Yet, this suppression was independent of NF-κb and c-Maf, since Tbx21 expression was similarly suppressed by IL-1ß/IL-18 upon NF-κb inhibiton and in c-Maf-deficient NKp46+ CCR6- ILC3s (Figure 3—figure supplement 2E and F). Taken together, these data demonstrate that, besides their reciprocal regulation, c-Maf and T-bet expression level are further critically controlled by cytokine signals.

In addition to cytokines, we also tested Notch signals as a potential extrinsic cue controlling both type 1 conversion of and c-Maf expression by ILC3s. Indeed, Notch was shown to be necessary for the induction and maintenance of T-bet and NKp46 expression in CCR6- ILC3s (Rankin et al., 2013; Viant et al., 2016). Similarly, we could recently identify Notch signalling as a potent inducer of c-Maf expression in T cells (Neumann et al., 2019; Neumann et al., 2014). To test the role of Notch we cultured purified NKp46+ CCR6- ILC3s on OP9 or OP9-DLL1 stromal cells, the latter ectopically express the Notch ligand Delta-like 1 (DLL1). As reported earlier, Notch signals were essential to drive type 1 conversion of NKp46+ CCR6- ILC3s, as evidenced by reduction and loss of NKp46 expression in the absence of Notch (Figure 3I). Importantly, c-Maf showed a similar expression pattern, being significantly reduced in NKp46+ CCR6- ILC3s cultured on OP9 cells as compared to cells cultured on OP9-DLL1 cells (Figure 3J). This downregulation of NKp46 and c-Maf expression in the absence of continuous Notch signalling was independent of potential survival promoting effects of Notch (Figure 3—figure supplement 3), establishing a molecular link between Notch signalling and type 1 conversion of NKp46+ CCR6- ILC3s.

Concluding remarks

Collectively, our study adds c-Maf as a novel key factor to the complex transcriptional network that governs the differentiation and function of ILC3s. In line with the emerging concept that co-expression and cross-regulation of multiple master regulators determines the fate and function of ILCs (Fang and Zhu, 2017), we have uncovered an essential negative feedback loop between c-Maf and T-bet, which restrains the type 1 conversion of ILC3s. Given the antagonism between T-bet and RORγt, the c-Maf-dependent suppression of T-bet also indirectly stabilizes RORγt expression, thus preserving the type 3 identity of ILC3s. In addition, c-Maf may also directly contribute to RORγt expression in ILC3s, as it has been reported for T cells (Tanaka et al., 2014; Zuberbuehler et al., 2019). Our data supports and extends the findings of a very recent report that was released after completion of this manuscript (Parker et al., 2020).

Materials and methods

Animals

To generate conditional c-Maf-deficient mice, RorcCre mice or Cd4Cre mice were crossed to Mafflox mice (provided by C. Birchmaier, MDC, Berlin, Germany) (Wende et al., 2012). Cd4Cre (Lee et al., 2001), RorcCre (Eberl and Littman, 2004), R26EYFP (Srinivas et al., 2001) and T-bet-deficient mice (Szabo et al., 2002) were described before. Il12a-/- mice were kindly provided by U. Schleicher, Erlangen. All mice were on a C57BL/6 background and bred and maintained under specific pathogen-free conditions at our animal facilities (FEM Charité Berlin, Germany). All animal experiments were in accordance with the ethical standards of the institution or practice at which the studies were conducted and were reviewed and approved by the responsible ethics committees of Germany (LAGeSo Berlin, I C 113 – G0172/14) and Russia.

Antibodies

A list of antibodies used in this study is provided in Supplementary file 3.

Cell isolation from small intestine and flow cytometry

Request a detailed protocol

Small intestinal tissue was treated with HBSS buffer (without calcium and magnesium) containing 5 mM EDTA and 10 mM HEPES (pH 7.5) at 37°C for 30 min to remove epithelial cells, minced and digested in HBSS buffer (with calcium and magnesium) containing 10 mM HEPES, 4% FCS, 0.5 mg/ml Collagenase D, 0.5 mg/ml DNaseI (Sigma), 0.5 U/ml Dispase (BD) with constantly stirring at 37°C for 30 min. The supernatant was filtered and the remaining tissue was mashed through a 70 μm mesh. siLP cells were separated using a 40%/80% step-gradient (Percoll solution, GE Healthcare). Flow cytometry was performed according to previously defined guidelines (Cossarizza et al., 2019). In detail, single-cell suspensions were stained with different antibodies (Supplementary file 3). For cytokine analysis, cells were restimulated with PMA (Sigma, 10 ng/ml), ionomycin (Sigma, 1 μg/ml) and IL-23 (50 ng/ml) for 5 hr in TexMACS medium (Miltenyi Biotec) containing 10% FCS. After 1 hr of stimulation, Brefeldin A (Sigma, 5 μg/ml) was added to block cytokine secretion. For intracellular staining of cytokines and transcription factors, cells were first stained for surface markers and dead cells were labeled with Fixable Viability Dye eFluor780 (eBioscience). After that, cells were fixed in Fix/Perm buffer (eBioscience) at 4°C for 1 hr, followed by permeabilization (eBioscience) at 4°C for 2 hr in the presence of antibodies. Cells were acquired with a BD LSRFortessa X-20 and analysis was performed with FlowJo (Tree Star) software.

RNA-seq analysis

Request a detailed protocol

NKp46+ CCR6- ILC3s and CCR6+ ILC3s were sorted from the siLP of 8–12 weeks old RorcCre Maffl/fl or littermate Maffl/fl control mice using a BD FACSAria sorter (sorting strategy Figure 2—figure supplement 2A). RNA was isolated with the RNeasy Micro kit from Qiagen according to the manufacturer`s protocol. RNA libraries were prepared using the Smart-Seq V4 Ultra low Input RNA kit (Takara Clontech). Sequencing was performed on an Illumina Nextseq 500 generating 75 bp paired-end reads. Three biological replicates of each subset were sequenced. Raw sequence reads were mapped to the mouse GRCm38/mm10 genome with TopHat2 (Kim et al., 2013) in very-sensitive settings for Bowtie2 (Langmead and Salzberg, 2012). Gene expression was quantified either by HTSeq (Anders et al., 2015) for total RNA or featureCounts (Liao et al., 2014) for mRNA and analyzed using DESeq2 (Love et al., 2014). A cut-off (FC > 1.5 and p-value < 0.05) was applied for calling differentially expressed genes. Furthermore, differentially expressed genes were filtered for ‘gene_type = protein_coding’ before further analysis.

Gene set enrichment analysis (GSEA)

Request a detailed protocol

GSEA was performed using the GSEA tool from the Broad Institute. Gene sets used in this study were generated by taking the top upregulated genes (log2FC > 2) from published differential gene expression analysis of RNA-seq data comparing ILC1s or NK cells with CCR6- ILC3s (Pokrovskii et al., 2019).

Luciferase reporter assay

Request a detailed protocol

HEK293T cells were transfected with the pGL3 basic luciferase plasmid (Promega) containing the T-bet promoter alone or the T-bet promoter in combination with an upstream enhancer region (Yang et al., 2007), or the empty pGL3 basic in combination with an internal control pRL-TK Renilla plasmid (Promega). The T-bet enhancer/promoter plasmid was described before (Hosokawa et al., 2013) and kindly provided by H. Hosokawa (Tokai University, Japan). In order to assess gene regulation by c-Maf, putative Maf responsive elements in the promoter and enhancer were mutated using the Q5 Site-Directed Mutagenesis Kit (New England Biolabs). In addition to mutated reporter plasmids, cells were co-transfected with c-Maf coding sequence in pMSCV. Luciferase activity was measured on a SpectraMax i33 microplate reader (Molecular Devices) after 24 hr using dual luciferase assay system (Promega). Luciferase activity was determined relative to Renilla.

In vitro stimulation of NKp46+ CCR6- ILC3s with cytokines

Request a detailed protocol

CD45+ Lineage- (Lineage: anti-CD19, anti-Gr-1, anti-CD3, anti-CD5) RORγtfm+ CD127+ NKp46+ CCR6- cells were sort-purified from the siLP of 11–14 week old RorcCre R26EYFP mice. Sorted cells were transferred in complete RPMI medium to a 96 U bottom well plate at a density of 15.000 cells/well. Subsequently cells were cultured in the presence of IL-7 (20 ng/ml) and SCF (20 ng/ml) plus different cytokines (IL-1ß, IL-12, IL-15, IL-18, IL-23; each at 20 ng/ml) for 36 hr before mRNA analysis. For blocking NF-κb activation, BMS-345541, a selective inhibitor of IκB kinase (Burke et al., 2003), was added at a concentration of 1 μM to the culture.

In vitro culture of NKp46+ CCR6- ILC3s on OP9-DLL1 cells

Request a detailed protocol

CD45+ Lineage- (Lineage: anti-CD19, anti-Gr-1, anti-CD3, anti-CD5) RORγtfm+ CD127+ NKp46+ CCR6- cells were sort-purified from the siLP of 11–14 week old RorcCre R26EYFP mice. Sorted cells were transferred in complete RPMI medium to OP9 or OP9-DLL1 cells at a density of 10.000 cells/well and cultured in the presence of IL-7 (20 ng/ml) and SCF (20 ng/ml) for 12 days before flow cytometric analysis. OP9 cells are murine stromal cells derived from OP/OP mice used as feeder cells in lymphocyte differentiation assays. OP9-DLL1 cells are transfected with Notch ligand delta-like-1 (Schmitt and Zúñiga-Pflücker, 2002). Prior to adding isolated lymphocytes, confluent feeder cells were treated with 5 μg/ml Mitomycin C (Sigma) for 3 hr at 37°C and subsequently seeded on a 96 flat bottom well plate at a density of 50.000 cells/well.

Bone marrow chimeras

Request a detailed protocol

Bone marrow cells from wild-type CD45.1+ CD90.2+ C57BL/6 and CD45.2+ CD90.2+ RorcCre Maffl/fl mice were mixed in a 1:1 ratio and intravenously injected into sub-lethally irradiated CD90.1+ wild-type recipient mice. Small and colonic lamina propria of reconstituted mice were analysed 6 weeks after cell transfer.

qPCR

Request a detailed protocol

mRNA for real-time qPCR was isolated with the RNeasy Plus Micro Kit according to the manual of the manufacturer (QIAGEN). Reverse transcription was done with the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) as it is described in the manufacturer’s protocol. qPCR was performed using a Quant Studio five system (Applied Biosystems) and the SYBR Green PCR Master Mix Kit (Applied Biosystems). The mRNA expression is presented relative to the expression of the housekeeping gene hypoxanthine-guanine phosphoribosyl-transferase (HPRT). Real-time qPCR primer can be found in Supplementary file 4.

Statistical analysis

Request a detailed protocol

Data are the mean with SEM and summarize or are representative of independent experiments as specified in the text. Statistical analyses were performed using Prism software (GraphPad) with two-tailed unpaired Student's t test (except RNA-seq data).

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
    Guidelines for the use of flow cytometry and cell sorting in immunological studies (second edition)
    1. A Cossarizza
    2. HD Chang
    3. A Radbruch
    4. A Acs
    5. D Adam
    6. S Adam-Klages
    7. WW Agace
    8. N Aghaeepour
    9. M Akdis
    10. M Allez
    11. LN Almeida
    12. G Alvisi
    13. G Anderson
    14. I Andrä
    15. F Annunziato
    16. A Anselmo
    17. P Bacher
    18. CT Baldari
    19. S Bari
    20. V Barnaba
    21. J Barros-Martins
    22. L Battistini
    23. W Bauer
    24. S Baumgart
    25. N Baumgarth
    26. D Baumjohann
    27. B Baying
    28. M Bebawy
    29. B Becher
    30. W Beisker
    31. V Benes
    32. R Beyaert
    33. A Blanco
    34. DA Boardman
    35. C Bogdan
    36. JG Borger
    37. G Borsellino
    38. PE Boulais
    39. JA Bradford
    40. D Brenner
    41. RR Brinkman
    42. AES Brooks
    43. DH Busch
    44. M Büscher
    45. TP Bushnell
    46. F Calzetti
    47. G Cameron
    48. I Cammarata
    49. X Cao
    50. SL Cardell
    51. S Casola
    52. MA Cassatella
    53. A Cavani
    54. A Celada
    55. L Chatenoud
    56. PK Chattopadhyay
    57. S Chow
    58. E Christakou
    59. L Čičin-Šain
    60. M Clerici
    61. FS Colombo
    62. L Cook
    63. A Cooke
    64. AM Cooper
    65. AJ Corbett
    66. A Cosma
    67. L Cosmi
    68. PG Coulie
    69. A Cumano
    70. L Cvetkovic
    71. VD Dang
    72. C Dang-Heine
    73. MS Davey
    74. D Davies
    75. S De Biasi
    76. G Del Zotto
    77. GV Dela Cruz
    78. M Delacher
    79. S Della Bella
    80. P Dellabona
    81. G Deniz
    82. M Dessing
    83. JP Di Santo
    84. A Diefenbach
    85. F Dieli
    86. A Dolf
    87. T Dörner
    88. RJ Dress
    89. D Dudziak
    90. M Dustin
    91. CA Dutertre
    92. F Ebner
    93. SBG Eckle
    94. M Edinger
    95. P Eede
    96. GRA Ehrhardt
    97. M Eich
    98. P Engel
    99. B Engelhardt
    100. A Erdei
    101. C Esser
    102. B Everts
    103. M Evrard
    104. CS Falk
    105. TA Fehniger
    106. M Felipo-Benavent
    107. H Ferry
    108. M Feuerer
    109. A Filby
    110. K Filkor
    111. S Fillatreau
    112. M Follo
    113. I Förster
    114. J Foster
    115. GA Foulds
    116. B Frehse
    117. PS Frenette
    118. S Frischbutter
    119. W Fritzsche
    120. DW Galbraith
    121. A Gangaev
    122. N Garbi
    123. B Gaudilliere
    124. RT Gazzinelli
    125. J Geginat
    126. W Gerner
    127. NA Gherardin
    128. K Ghoreschi
    129. L Gibellini
    130. F Ginhoux
    131. K Goda
    132. DI Godfrey
    133. C Goettlinger
    134. JM González-Navajas
    135. CS Goodyear
    136. A Gori
    137. JL Grogan
    138. D Grummitt
    139. A Grützkau
    140. C Haftmann
    141. J Hahn
    142. H Hammad
    143. G Hämmerling
    144. L Hansmann
    145. G Hansson
    146. CM Harpur
    147. S Hartmann
    148. A Hauser
    149. AE Hauser
    150. DL Haviland
    151. D Hedley
    152. DC Hernández
    153. G Herrera
    154. M Herrmann
    155. C Hess
    156. T Höfer
    157. P Hoffmann
    158. K Hogquist
    159. T Holland
    160. T Höllt
    161. R Holmdahl
    162. P Hombrink
    163. JP Houston
    164. BF Hoyer
    165. B Huang
    166. FP Huang
    167. JE Huber
    168. J Huehn
    169. M Hundemer
    170. CA Hunter
    171. WYK Hwang
    172. A Iannone
    173. F Ingelfinger
    174. SM Ivison
    175. HM Jäck
    176. PK Jani
    177. B Jávega
    178. S Jonjic
    179. T Kaiser
    180. T Kalina
    181. T Kamradt
    182. SHE Kaufmann
    183. B Keller
    184. SLC Ketelaars
    185. A Khalilnezhad
    186. S Khan
    187. J Kisielow
    188. P Klenerman
    189. J Knopf
    190. HF Koay
    191. K Kobow
    192. JK Kolls
    193. WT Kong
    194. M Kopf
    195. T Korn
    196. K Kriegsmann
    197. H Kristyanto
    198. T Kroneis
    199. A Krueger
    200. J Kühne
    201. C Kukat
    202. D Kunkel
    203. H Kunze-Schumacher
    204. T Kurosaki
    205. C Kurts
    206. P Kvistborg
    207. I Kwok
    208. J Landry
    209. O Lantz
    210. P Lanuti
    211. F LaRosa
    212. A Lehuen
    213. S LeibundGut-Landmann
    214. MD Leipold
    215. LYT Leung
    216. MK Levings
    217. AC Lino
    218. F Liotta
    219. V Litwin
    220. Y Liu
    221. HG Ljunggren
    222. M Lohoff
    223. G Lombardi
    224. L Lopez
    225. M López-Botet
    226. AE Lovett-Racke
    227. E Lubberts
    228. H Luche
    229. B Ludewig
    230. E Lugli
    231. S Lunemann
    232. HT Maecker
    233. L Maggi
    234. O Maguire
    235. F Mair
    236. KH Mair
    237. A Mantovani
    238. RA Manz
    239. AJ Marshall
    240. A Martínez-Romero
    241. G Martrus
    242. I Marventano
    243. W Maslinski
    244. G Matarese
    245. AV Mattioli
    246. C Maueröder
    247. A Mazzoni
    248. J McCluskey
    249. M McGrath
    250. HM McGuire
    251. IB McInnes
    252. HE Mei
    253. F Melchers
    254. S Melzer
    255. D Mielenz
    256. SD Miller
    257. KHG Mills
    258. H Minderman
    259. J Mjösberg
    260. J Moore
    261. B Moran
    262. L Moretta
    263. TR Mosmann
    264. S Müller
    265. G Multhoff
    266. LE Muñoz
    267. C Münz
    268. T Nakayama
    269. M Nasi
    270. K Neumann
    271. LG Ng
    272. A Niedobitek
    273. S Nourshargh
    274. G Núñez
    275. JE O'Connor
    276. A Ochel
    277. A Oja
    278. D Ordonez
    279. A Orfao
    280. E Orlowski-Oliver
    281. W Ouyang
    282. A Oxenius
    283. R Palankar
    284. I Panse
    285. K Pattanapanyasat
    286. M Paulsen
    287. D Pavlinic
    288. L Penter
    289. P Peterson
    290. C Peth
    291. J Petriz
    292. F Piancone
    293. WF Pickl
    294. S Piconese
    295. M Pinti
    296. AG Pockley
    297. MJ Podolska
    298. Z Poon
    299. K Pracht
    300. I Prinz
    301. CEM Pucillo
    302. SA Quataert
    303. L Quatrini
    304. KM Quinn
    305. H Radbruch
    306. T Radstake
    307. S Rahmig
    308. HP Rahn
    309. B Rajwa
    310. G Ravichandran
    311. Y Raz
    312. JA Rebhahn
    313. D Recktenwald
    314. D Reimer
    315. C Reis E Sousa
    316. EBM Remmerswaal
    317. L Richter
    318. LG Rico
    319. A Riddell
    320. AM Rieger
    321. JP Robinson
    322. C Romagnani
    323. A Rubartelli
    324. J Ruland
    325. A Saalmüller
    326. Y Saeys
    327. T Saito
    328. S Sakaguchi
    329. F Sala-de-Oyanguren
    330. Y Samstag
    331. S Sanderson
    332. I Sandrock
    333. A Santoni
    334. RB Sanz
    335. M Saresella
    336. C Sautes-Fridman
    337. B Sawitzki
    338. L Schadt
    339. A Scheffold
    340. HU Scherer
    341. M Schiemann
    342. FA Schildberg
    343. E Schimisky
    344. A Schlitzer
    345. J Schlosser
    346. S Schmid
    347. S Schmitt
    348. K Schober
    349. D Schraivogel
    350. W Schuh
    351. T Schüler
    352. R Schulte
    353. AR Schulz
    354. SR Schulz
    355. C Scottá
    356. D Scott-Algara
    357. DP Sester
    358. TV Shankey
    359. B Silva-Santos
    360. AK Simon
    361. KM Sitnik
    362. S Sozzani
    363. DE Speiser
    364. J Spidlen
    365. A Stahlberg
    366. AM Stall
    367. N Stanley
    368. R Stark
    369. C Stehle
    370. T Steinmetz
    371. H Stockinger
    372. Y Takahama
    373. K Takeda
    374. L Tan
    375. A Tárnok
    376. G Tiegs
    377. G Toldi
    378. J Tornack
    379. E Traggiai
    380. M Trebak
    381. TIM Tree
    382. J Trotter
    383. J Trowsdale
    384. M Tsoumakidou
    385. H Ulrich
    386. S Urbanczyk
    387. W van de Veen
    388. M van den Broek
    389. E van der Pol
    390. S Van Gassen
    391. G Van Isterdael
    392. RAW van Lier
    393. M Veldhoen
    394. S Vento-Asturias
    395. P Vieira
    396. D Voehringer
    397. HD Volk
    398. A von Borstel
    399. K von Volkmann
    400. A Waisman
    401. RV Walker
    402. PK Wallace
    403. SA Wang
    404. XM Wang
    405. MD Ward
    406. KA Ward-Hartstonge
    407. K Warnatz
    408. G Warnes
    409. S Warth
    410. C Waskow
    411. JV Watson
    412. C Watzl
    413. L Wegener
    414. T Weisenburger
    415. A Wiedemann
    416. J Wienands
    417. A Wilharm
    418. RJ Wilkinson
    419. G Willimsky
    420. JB Wing
    421. R Winkelmann
    422. TH Winkler
    423. OF Wirz
    424. A Wong
    425. P Wurst
    426. JHM Yang
    427. J Yang
    428. M Yazdanbakhsh
    429. L Yu
    430. A Yue
    431. H Zhang
    432. Y Zhao
    433. SM Ziegler
    434. C Zielinski
    435. J Zimmermann
    436. A Zychlinsky
    (2019)
    European Journal of Immunology 49:1457–1973.
    https://doi.org/10.1002/eji.201970107
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21
  22. 22
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
  35. 35
  36. 36
  37. 37
    c-Maf regulates the plasticity of group 3 innate lymphoid cells by restraining the type 1 program
    1. ME Parker
    2. A Barrera
    3. JD Wheaton
    4. MK Zuberbuehler
    5. DSJ Allan
    6. JR Carlyle
    7. TE Reddy
    8. M Ciofani
    (2020)
    The Journal of Experimental Medicine, 217, 10.1084/jem.20191030, 31570496.
  38. 38
  39. 39
  40. 40
  41. 41
  42. 42
  43. 43
  44. 44
  45. 45
  46. 46
  47. 47
  48. 48
  49. 49
  50. 50
  51. 51
  52. 52
  53. 53
  54. 54
  55. 55
  56. 56
  57. 57
  58. 58
  59. 59
  60. 60
  61. 61
  62. 62
  63. 63
  64. 64
  65. 65
  66. 66
  67. 67

Decision letter

  1. Satyajit Rath
    Senior and Reviewing Editor; Indian Institute of Science Education and Research (IISER), India
  2. David Withers
    Reviewer

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Acceptance summary:

This paper describes that cMAF regulates the balance between ILC1 and ILC3, by controlling T-bet expression. This occurs in the CCR6- ILC3 population. cMAF expression regulates the dynamic acquisition of type 1 effector functions as deletion of this factor results in upregulation of NKp46, which was previously been found to be regulated by T-bet. While it is not novel that cMAF is involved in regulating the balance of ILC3 and ILC1, the present paper examines mechanisms involved. The authors provide convincing data that c-Maf negatively regulates the transition of ILC3 to ILC1 like cells. How ILC plasticity is regulated is an important in topical area of research and the identification of the role of c-Maf significantly adds to our understanding. The data provided are convincing, and support, with more insight in certain areas, a very recent paper (Parker et al., 2019), a publication that was released after completion of the present manuscript.

Decision letter after peer review:

Thank you for submitting your article "c-Maf restrains T-bet-driven programming of CCR6-negative group 3 innate lymphoid cells" for consideration by eLife. Your article has been reviewed by two peer reviewers, and the evaluation has been overseen by Satyajit Rath as the Senior and Reviewing Editor. The following individual involved in review of your submission has agreed to reveal their identity: David Withers (Reviewer #2).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Essential revisions:

1) The claim that c-Maf is required for maintenance of the total CCR6- ILC3 is not really supported by the data. There is a reduction but this reduction is less than 30-40%. The strongest reduction is seen in NKp46-CCR6- cells (Figure 2). A more likely conclusion would be that cMAF is required for maintenance of the NKp46-CCR6-ILC3s. This should be clarified.

2) Whereas the authors show that IL-1b and IL-18 can upregulate cMAF, the effects of IL-12 do not get much attention. IL-12 is a powerful inducer of ILC3 to ILC1 conversion and IL-12 receptor deficient mice show a strong increase in NKp46+RORgt+ ILC3 (Vonarbourg et al., 2010). One would expect a strong effect of IL-12 on c-Maf expression. Indeed, inspection of Figure 3E reveals a significant reduction of cMAF by IL-12. This should be addressed.

3) In Figure 1, the authors should show total numbers of ILC populations rather than only percentages. Total numbers would clarify whether other populations were really unaffected.

4) Again, in Figure 1, it would be useful to compare CCR6-NKp46- and CCR6-NKp46+ cell populations for further clarity. It is not necessary to group both CCR6- populations together. Since the CCR6-NKp46- cells are considered to be the precursors, expression of c-Maf here is of interest.

5) Any evidence for T-bet binding in the Maf locus to regulate its expression, building on the observation in Figure 3D, would be very useful to add.

https://doi.org/10.7554/eLife.52549.sa1

Author response

Essential revisions:

1) The claim that c-Maf is required for maintenance of the total CCR6- ILC3 is not really supported by the data. There is a reduction but this reduction is less than 30-40%. The strongest reduction is seen in NKp46-CCR6- cells (Figure 2). A more likely conclusion would be that cMAF is required for maintenance of the NKp46-CCR6-ILC3s. This should be clarified.

We agree with the reviewer that our analysis of c-Maf-deficient ILC3s in Figure 1 was unprecise because we divided ILC3s solely into the CCR6- and CCR6+ ILC3 subsets, neglecting potential differences between NKp46+ vs. NKp46- CCR6- ILC3s. Therefore, we revisited our data and now present a comprehensive characterization of c-Maf-deficient ILC3s in Figure 1, by directly comparing NKp46+ CCR6- ILC3s, NKp46- CCR6- ILC3s and NKp46- CCR6+ ILC3s (new Figure 1B, D-G).

This analysis indeed revealed that the NKp46- CCR6- ILC3 subset was selectively reduced in RorcCreMaffl/fl mice, whereas total numbers of NKp46+ CCR6- ILC3s and NKp46- CCR6+ ILC3s were normal or even slightly increased as compared to controls (new Figure 1D, subsection “c-Maf specifically preserves the type 3 identity of CCR6- ILC3s”, second paragraph). Thus, we corrected our claim, now concluding that c-Maf is specifically required for the maintenance of intestinal NKp46- CCR6- ILC3 (see the third paragraph of the aforementioned subsection). Given the fact that NKp46- CCR6- ILC3 are considered to contain precursors of NKp46+ CCR6- ILC3s, the selective loss of NKp46- CCR6- ILC3s in RorcCreMaffl/fl mice suggested an enhanced differentiation from NKp46- to NKp46+ CCR6- ILC3s in the absence of c-Maf, which is in line with our concept that c-Maf-deficiency results in an amplified type 1 conversion of CCR6- ILC3s (subsection “c-Maf suppresses the acquisition of type 1 properties by CCR6- ILC3s”, second paragraph).

2) Whereas the authors show that IL-1b and IL-18 can upregulate cMAF, the effects of IL-12 do not get much attention. IL-12 is a powerful inducer of ILC3 to ILC1 conversion and IL-12 receptor deficient mice show a strong increase in NKp46+RORgt+ ILC3 (Vonarbourg et al., 2010). One would expect a strong effect of IL-12 on c-Maf expression. Indeed, inspection of Figure 3E reveals a significant reduction of cMAF by IL-12. This should be addressed.

We followed the reviewers suggestion and further explored the role of IL-12 in c-Maf regulation. First, we comparatively assessed c-Maf expression in siLP NKp46+ ILC3s from Il12a-/- and wild-type control mice. However, we could not detect any alteration in c-Maf expression in the absence of IL-12 (new Figure 3—figure supplement 2C, subsection “IL-1β- and IL-18-mediated NF-κb and Notch signalling promote c-Maf expression in CCR6- ILC3s”, second paragraph), suggesting that c-Maf expression is largely independent of IL-12 at least during homeostatic conditions.

Second, we tested whether IL-12 was capable to interfere with the IL-1ß/IL-18-mediated induction of c-Maf expression. Yet, in vitro stimulation of NKp46+ CCR6- ILC3s with IL-1ß/IL-18 plus IL-12, as compared to IL-1ß/IL-18 stimulation alone, did not result in a reduction of c-Maf expression (new Figure 3—figure supplement 2B, see the aforementioned paragraph), questioning a dominant role of IL-12 in regulation of c-Maf expression.

We have included the above mentioned data as supplementary figures (new Figure 3—figure supplement 2B, C) in the revised version of our manuscript.

3) In Figure 1, the authors should show total numbers of ILC populations rather than only percentages. Total numbers would clarify whether other populations were really unaffected.

As suggested by the reviewer we now show in Figure 1 total numbers of the different siLP ILC populations (ILC1s, ILC2s, ILC3s) (new Figure 1C, subsection “c-Maf specifically preserves the type 3 identity of CCR6- ILC3s”, second paragraph) as well as of the different siLP ILC3 subsets (NKp46+ CCR6- ILC3s, NKp46- CCR6- ILC3s and NKp46- CCR6+ ILC3s) (new Figure 1D, see the aforementioned paragraph).

This analysis demonstrated that frequencies and total numbers of ILC1s and ILC2s were significantly increased in RorcCreMaffl/fl mice as compared to controls (new Figure 1C, subsection “c-Maf specifically preserves the type 3 identity of CCR6- ILC3s, second paragraph). For ILC3s, we detected reduced relative proportions among Lin- CD127+ cells, while total numbers were not changed in RorcCreMaffl/fl mice (new Figure 1C, see the aforementioned paragraph). However, among ILC3s, we found a selective loss of NKp46- CCR6- ILC3s (new Figure 1D, see the aforementioned paragraph). In contrast, total numbers of NKp46+ CCR6- ILC3s and NKp46- CCR6+ ILC3s were not significantly changed (new Figure 1D, see the aforementioned paragraph).

4) Again, in Figure 1, it would be useful to compare CCR6-NKp46- and CCR6-NKp46+ cell populations for further clarity. It is not necessary to group both CCR6- populations together. Since the CCR6-NKp46- cells are considered to be the precursors, expression of c-Maf here is of interest.

As mentioned already in response to point 1, we now directly compare NKp46+ CCR6- ILC3s, NKp46- CCR6- ILC3s and NKp46- CCR6+ ILC3s in Figure 1 (new Figure 1B, D-G) to better dissect the phenotype of CCR6- ILC3s based on NKp46 expression.

The analysis of c-Maf expression (new Figure 1B, subsection “c-Maf specifically preserves the type 3 identity of CCR6- ILC3s”, first paragraph) showed that c-Maf was particularly highly expressed by NKp46+ CCR6- ILC3s at levels comparable to RORγt+ CD4+ T cells. This is in line with our findings, that c-Maf expression (like NKp46 expression itself) in CCR6- ILC3s is dependent on T-bet and thus tightly correlates with T-bet expression (new Figure 2A, subsection “c-Maf suppresses the acquisition of type 1 properties by CCR6- ILC3s”, second paragraph). In contrast, c-Maf expression level of NKp46- CCR6- ILC3s were more heterogenous (new Figure 1B). However, also in NKp46- CCR6- ILC3s we observed tight correlation of c-Maf and T-bet expression (new Figure 2A, see the aforementioned paragraph). Together with the selective loss of NKp46- CCR6- ILC3s in RorcCreMaffl/fl mice, these data suggest an enhanced T-bet-dependent differentiation from the NKp46- CCR6- to the NKp46+ CCR6- ILC3 compartment in the absence of c-Maf (see the aforementioned paragraph).

In addition to the analysis of c-Maf expression, we also revisited our data to directly compare RORγt (new Figure 1E) and IL-17A/IL-22 expression (new Figure 1F and 1G) between NKp46+ CCR6- ILC3s, NKp46- CCR6- ILC3s and NKp46- CCR6+ ILC3s in RorcCreMaffl/fl mice and control mice. This analysis showed that RORγt protein levels (subsection “c-Maf specifically preserves the type 3 identity of CCR6- ILC3s”, third paragraph) as well as frequencies of IL-17A and IL-22 producers (see the fourth paragraph of the aforementioned subsection) were significantly reduced in both NKp46+ and NKp46- CCR6- ILC3s in the absence of c-Maf, whereas NKp46- CCR6+ ILC3s were unaffected.

5) Any evidence for T-bet binding in the Maf locus to regulate its expression, building on the observation in Figure 3D, would be very useful to add.

We thank the reviewer for bringing up this very interesting point. In order to gain evidence for a direct binding of T-bet to Maf, we performed a detailed in silicoanalysis of the Maf locus using published ATAC-Seq data from NKp46+ CCR6- ILC3s [1] and T-bet ChIP-seq data from Th1 cells [2]. We also performed motif analysis using HOMER to identify potential T-bet binding sites across the Maf locus. We present the results of these analyses in the new Figure 3—figure supplement 1 and in the text (subsection “c-Maf expression in CCR6- ILC3s is dependent on T-bet”, second paragraph).

As a result, based on the ATAC-Seq data, we find that the Maf locus in NKp46+ ILC3s is only accessible in a very specific and narrow region (stretching ca. 1kb) upstream of the TSS (new Figure 3—figure supplement 1A). Interestingly, this open chromatin region overlaps with two highly conserved non-coding sequences (CNS-0.5 and CNS-1) upstream of Maf, suggesting an important role of these regions in regulation of Maf in NKp46+ CCR6- ILC3s (new Figure 3—figure supplement 1A).

Importantly, analysis of T-bet ChIP-seq data from Th1 cells revealed a striking overlap of T-bet binding peaks with CNS-0.5 and CNS-1 (new Figure 3—figure supplement 1B). Furthermore, T-bet motif analysis identified several T-bet binding sites within CNS-0.5 and CNS-1 (new Figure 3—figure supplement 1C and D).

Taken together, these data strongly suggest that T-bet directly binds to Maf in NKp46+ CCR6- ILC3s thereby promoting c-Maf expression.

https://doi.org/10.7554/eLife.52549.sa2

Article and author information

Author details

  1. Caroline Tizian

    1. Laboratory of Innate Immunity, Department of Microbiology, Infectious Diseases and Immunology, Charité-Universitätsmedizin Berlin, Berlin, Germany
    2. Mucosal and Developmental Immunology, Deutsches Rheuma-Forschungszentrum, Berlin, Germany
    Contribution
    Data curation, Formal analysis, Validation, Investigation, Methodology, Designed and performed experiments, Analysed data
    Contributed equally with
    Annette Lahmann
    Competing interests
    No competing interests declared
  2. Annette Lahmann

    Chronic Immune Reactions, Deutsches Rheuma-Forschungszentrum, Berlin, Germany
    Contribution
    Data curation, Formal analysis, Validation, Investigation, Methodology, Designed and performed experiments, Analysed data
    Contributed equally with
    Caroline Tizian
    Competing interests
    No competing interests declared
  3. Oliver Hölsken

    1. Laboratory of Innate Immunity, Department of Microbiology, Infectious Diseases and Immunology, Charité-Universitätsmedizin Berlin, Berlin, Germany
    2. Mucosal and Developmental Immunology, Deutsches Rheuma-Forschungszentrum, Berlin, Germany
    Contribution
    Formal analysis, Validation, Investigation, Methodology, Designed and performed experiments, Analysed data
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-6086-9275
  4. Catalina Cosovanu

    1. Laboratory of Innate Immunity, Department of Microbiology, Infectious Diseases and Immunology, Charité-Universitätsmedizin Berlin, Berlin, Germany
    2. Mucosal and Developmental Immunology, Deutsches Rheuma-Forschungszentrum, Berlin, Germany
    Contribution
    Formal analysis, Methodology, Performed experiments, Analysed data
    Competing interests
    No competing interests declared
  5. Michael Kofoed-Branzk

    1. Laboratory of Innate Immunity, Department of Microbiology, Infectious Diseases and Immunology, Charité-Universitätsmedizin Berlin, Berlin, Germany
    2. Mucosal and Developmental Immunology, Deutsches Rheuma-Forschungszentrum, Berlin, Germany
    Contribution
    Methodology, Provided discussion, Proofread the manuscript
    Competing interests
    No competing interests declared
  6. Frederik Heinrich

    Therapeutic Gene Regulation, Deutsches Rheuma-Forschungszentrum, Berlin, Germany
    Contribution
    Formal analysis, Methodology, Performed computational analyses of 16S rRNA-seq data
    Competing interests
    No competing interests declared
  7. Mir-Farzin Mashreghi

    Therapeutic Gene Regulation, Deutsches Rheuma-Forschungszentrum, Berlin, Germany
    Contribution
    Resources, Supervision, Supervised sequencing experiments, Provided reagents and equipment for their execution
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-8015-6907
  8. Andrey Kruglov

    1. Chronic Inflammation, Deutsches Rheuma-Forschungszentrum, Berlin, Germany
    2. Belozersky Institute of Physico-Chemical Biology and Biological Faculty, M.V. Lomonosov Moscow State University, Moscow, Russian Federation
    Contribution
    Resources, Methodology, Assisted in chimera experiments
    Competing interests
    No competing interests declared
  9. Andreas Diefenbach

    1. Laboratory of Innate Immunity, Department of Microbiology, Infectious Diseases and Immunology, Charité-Universitätsmedizin Berlin, Berlin, Germany
    2. Mucosal and Developmental Immunology, Deutsches Rheuma-Forschungszentrum, Berlin, Germany
    Contribution
    Conceptualization, Resources, Supervision, Funding acquisition, Provided crucial reagents, Contributed to the design of the study, Proofread the manuscript
    Competing interests
    No competing interests declared
  10. Christian Neumann

    1. Laboratory of Innate Immunity, Department of Microbiology, Infectious Diseases and Immunology, Charité-Universitätsmedizin Berlin, Berlin, Germany
    2. Mucosal and Developmental Immunology, Deutsches Rheuma-Forschungszentrum, Berlin, Germany
    Contribution
    Conceptualization, Resources, Data curation, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Visualization, Methodology, Project administration, Conceived the project, Designed and performed most experiments, Analysed the data, Generated the figures and wrote the manuscript with the input of all co-authors
    For correspondence
    c.neumann@charite.de
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-2202-1876

Funding

Deutsche Forschungsgemeinschaft (Priority Program 1937 "Innate Lymphoid Cells")

  • Andreas Diefenbach
  • Christian Neumann

European Regional Development Fund (ERDF 2014-2020)

  • Frederik Heinrich
  • Mir-Farzin Mashreghi

European Regional Development Fund (EFRE 1.8/11)

  • Mir-Farzin Mashreghi

Russian Science Foundation (17-74-20059)

  • Andrey Kruglov

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

This research was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) Priority Program 1937 ‘Innate Lymphoid Cells’ (to CN project number: 428192857 and to AD project number: DI764/9-2). FH and MFM were supported by the state of Berlin and the ‘European Regional Development Fund’ (ERDF 2014–2020, EFRE 1.8/11, Deutsches Rheuma‐Forschungszentrum). We acknowledge support from the DFG and the Open Access Publication Funds of Charité – Universitätsmedizin Berlin. We thank Irene Mattiola, Christoph S N Klose, Fabian Guendel-Rojas, Mario Witkowski, Pawel Durek and Katrin Lehmann for discussion, technical and experimental help and proofreading of the manuscript.

Ethics

All animal experiments were in accordance with the ethical standards of the institution or practice at which the studies were conducted and were reviewed and approved by the responsible ethics committees of Germany (LAGeSo Berlin, I C 113 – G0172/14) and Russia.

Senior and Reviewing Editor

  1. Satyajit Rath, Indian Institute of Science Education and Research (IISER), India

Reviewer

  1. David Withers

Publication history

  1. Received: October 8, 2019
  2. Accepted: February 10, 2020
  3. Accepted Manuscript published: February 10, 2020 (version 1)
  4. Version of Record published: February 17, 2020 (version 2)
  5. Version of Record updated: March 10, 2020 (version 3)

Copyright

© 2020, Tizian et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 1,087
    Page views
  • 192
    Downloads
  • 2
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Immunology and Inflammation
    Matthew D Lauver et al.
    Research Article

    JCPyV polyomavirus, a member of the human virome, causes Progressive Multifocal Leukoencephalopathy (PML), an oft-fatal demyelinating brain disease in individuals receiving immunomodulatory therapies. Mutations in the major viral capsid protein, VP1, are common in JCPyV from PML patients (JCPyV-PML) but whether they confer neurovirulence or escape from virus-neutralizing antibody (nAb) in vivo is unknown. A mouse polyomavirus (MuPyV) with a sequence-equivalent JCPyV-PML VP1 mutation replicated poorly in the kidney, a major reservoir for JCPyV persistence, but retained the CNS infectivity, cell tropism, and neuropathology of the parental virus. This mutation rendered MuPyV resistant to a monoclonal Ab (mAb), whose specificity overlapped the endogenous anti-VP1 response. Using cryo EM and a custom sub-particle refinement approach, we resolved an MuPyV:Fab complex map to 3.2 Å resolution. The structure revealed the mechanism of mAb evasion. Our findings demonstrate convergence between nAb evasion and CNS neurovirulence in vivo by a frequent JCPyV-PML VP1 mutation.