The AP-1 transcription factors c-Jun and JunB are essential for CD8α conventional dendritic cell identity

Dendritic cell (DC) development is orchestrated by lineage-determining transcription factors (TFs). Although, members of the activator-protein-1 (AP-1) family, including Batf3, have been implicated in conventional (c)DC specification, the role of Jun proteins is poorly understood. Here, we identified c-Jun and JunB as essential for cDC1 fate specification and function. In mice, Jun proteins regulate extrinsic and intrinsic pathways, which control CD8α cDC1 diversification, whereas CD103 cDC1 development is unaffected. The loss of c-Jun and JunB in DC progenitors diminishes the CD8α cDC1 pool and thus confers resistance to Listeria monocytogenes infection. Their absence in CD8α cDC1 results in impaired TLR triggering and antigen cross-presentation. Both TFs are required for the maintenance of the CD8α cDC1 subset and suppression of cDC2 identity on a transcriptional and phenotypic level. Taken together, these results demonstrate the essential role of c-Jun and JunB in CD8α cDC1 diversification, function, and maintenance of their identity.


Introduction
Dendritic cells (DCs) are innate immune cells essential for the initiation of antigen-specific immune responses and maintenance of tolerance [1]. DCs originate from multipotent, hematopoietic precursors in the bone marrow (BM) and can be categorized into two distinct branches, conventional (cDCs) and plasmacytoid DCs (pDCs), characterized by their unique transcriptional, functional, and cell surface phenotypes [2][3][4]. cDCs can be subgrouped in cDC1, specialized in cross-presentation and cDC2, critical for the induction of T h 2 and T h 17 immunity [5]. cDC1 can be identified by the expression of X-C motif chemokine receptor 1 (XCR1) and comprise a lymphoidresident CD8α + and a non-lymphoid CD103 + population [6,7]. Over the last decade studies have shown that the differentiation into cDC1 is controlled by a network of specific, lineage-determining transcription factors (TF) [8,9]. TFs like the interferon regulatory factor 8 (IRF8), inhibitor of DNA binding 2 (ID2), and nuclear factor interleukin 3-regulated (NFIL3) are required for the diversification of the common DC progenitor (CDP) to cDC1 [10][11][12]. Furthermore, the basic leucine zipper transcription factor, ATF-like 3 (Batf3), a member of the dimeric activator protein-1 (AP-1) TF family, has been shown to maintain IRF8 expression in pre-cDCs, the direct clonogenic precursor to cDCs, to safeguard cDC1 differentiation [13,14].
AP-1 TFs comprise a group of homo-or heterodimers of basic leucine zipper proteins belonging to the Jun, Fos, MAF or ATF family with essential functions in development, cancer and immunity [15][16][17]. c-Jun and JunB are AP-1 family members with prominent roles in T cell development [18] and T h 17 identity [19], respectively. Moreover, JunB inactivation in stem cells results in a myeloproliferative disorder characterized by increased numbers of granulocyte/macrophage progenitors (GMP) [20,21]. Functionally, c-Jun and JunB can be antagonistic but also compensatory, emphasizing that expression of some target genes is regulated equally by c-Jun and JunB. The inducible epidermal deletion of both c-Jun and JunB, for example, causes a psoriasis-like skin phenotype in adult mice, whereas single deletion results in no phenotype [22][23][24][25].
The interaction between Batf and Jun was originally considered to inhibit AP-1 transcriptional activity [26,27]. However, a transcriptional complex between Jun/Batf and IRF4 or IRF8 enables binding to AP-1-IRF composite elements (AICEs), which is critical for the development of cDC1s [28] and the differentiation of T h 17 cells [29,30]. Although deletion of Batf3 in a genetically engineered mouse model (GEMM) demonstrated its role in CD8α/ CD103 development, the importance of Jun proteins, is poorly defined. In this study, we show that combined, immune cell-specific deletion of c-Jun and JunB leads to the selective loss of lymphoid CD8α cDC1, whereas nonlymphoid CD103 cDC1 differentiate normally. Importantly, c-Jun and JunB are required for cDC1-dependent immune responses, including activation upon TLR triggering and antigen cross-presentation. Mechanistically, our data reveal for the first time that c-Jun and JunB maintain cDC1 identity and suppress expression of cDC2 TF and surface markers.

Methods
Animal studies c-Jun fl/fl and JunB fl/fl mice were crossed with CD11c-Cre, Mx-Cre or Tamoxifen (Tx) inducible K5-Cre ER mice, [31][32][33][34]. Mx-Cre mice received one or two (indicated in Fig. legends) intraperitoneal injections of poly I: C (200 µg; VWR) and K5-Cre ER mice five consecutive injections of Tx (1 mg; Sigma Aldrich) for deletion. A 5% cream formulation of Imiquimod (Aldara, Meda Pharma) was applied topically on shaved and waxed (Veet, Heidelberg, Germany) back skin for up to 7 days [35]. C57Bl/6 CD45.1, CD45.2 or CD45.1/CD45.2 were bred in house. Female and male mice of a mixed background (129 Sv × C57BL/6) and 8-16 weeks of age were used for studies. Mice were kept in the animal facility of the Medical University of Vienna in accordance with institutional policies and federal guidelines. All mice had access to food and water ad libitum.

Flow cytometry
Single cell suspensions of lymphoid organs and BM were prepared as follows. BM was flushed, spleen and lymph nodes were digested for 30 min (37°C) with Liberase™ (50 µg/mL; Roche) and DNAse I (100 µg/mL; Sigma Aldrich), filtered through a 70 µm filter and washed. Red blood cells were lysed with ACK buffer. Single cell suspensions were blocked with anti-CD16/32 antibody (BioLegend) before addition of fluorescently labeled antibodies. Antibodies (see Supplementary Table 1) were diluted according to the manufacturer´s recommendation and samples incubated on 4°C for 30 min. Absolute cell number was assessed with a CASY cell counter (Beckman Coulter) or by addition of 123count eBeads (Thermo Fisher Scientific). Cells were acquired on a LSR Fortessa cell analyzer (BD Biosciences) and analyzed with FlowJo software (Treestar).

Cell culture
BM cells were isolated from femur/tibia and red blood cells were lysed. 3 × 10 6 cells/mL were cultured in RPMI supplemented with 10% FCS, 1% Pen/Strep, sodium pyruvate, nonessential amino acids and 0.1% βmercaptoethanol. Recombinant mFLT3L (80 ng/mL, Peprotech) was added to induce cDC and pDC differentiation. For CD103 + cDC in vitro differentiation BM cells were depleted of lineage-positive cells using the lineage cell depletion kit (Miltenyi Biotec). Enriched cells were cultured in complete medium supplemented with recombinant mFLT3L (80 ng/mL, Peprotech) and recombinant mGM-CSF (10 ng/mL, Peprotech) for 12 days. On day 6 fresh culture media supplemented with recombinant mFLT3L and mGM-CSF was added.

Generation of bone marrow chimeric mice
BM chimeric mice were generated by intravenous (i.v.) injection of CD3ε/CD90.1.-MACS-depleted BM cells (4 × 10 6 ) into lethally irradiated mice . For mixed BM chimeras, BM cells of two different donors (CD45.1/2 wild-type and CD45.2 c-Jun/JunB deficient) were mixed in a ratio 1:1 before injection. Experiments were performed after a reconstitution period of at least 8 weeks. Poly I: C was injected into chimeric mice twice, 5 days apart, and mice were analyzed 14 days after Cre induction.
In vivo α-G-CSF antibody treatment Mice were treated with 50 μg of α-GCSF IgG (clone 67604, Sigma Aldrich), or matched isotype control (rat IgG, BioXCell) by intraperitoneal injection 2 and 4 days after poly I: C induced deletion of c-Jun/JunB in the Mx-Cre mouse model.

Listeria infection model
An overnight culture of L. monocytogenes strain LO28 was re-cultured in BHI medium to late log phase, pelleted and diluted in PBS. The concentration of bacteria was quantified by optical density measurements at 600 nm and confirmed by plating serial dilutions on BHI agar plates and colony counting. 1 × 10 6 cfu were injected into the peritoneum of 8-to 15-week-old mice. Animals were euthanized after 24 and 72 h, spleens and livers were isolated, weighed and homogenized, and serial dilutions were plated on BHI agar plates. Colonies were counted after~30 h culture at 37°C.
In vitro cross-presentation c-Jun/JunB cDC FL were obtained by negative selection of B220 + pDC FL by MACS on day 7 of FLT3L-supplemented BM cultures. cDC FL were then loaded with Ovalbumin (250 µg/mL) or SIINFEKL (20 µg/mL) for 6 h, washed and re-plated at 2 × 10 4 with CFSE labelled (1 µM, 10 min) OT-I T cells in a 1:100 cDC FL to OT-I T cell ratio. OT I-T cells were enriched from lymphoid organs of OT-I mice by negative depletion of B (CD19 + , B220 + ), NK (NK 1.1 + ), DC (CD11c + ) and myeloid (CD11b + ) cells. After 3 days of co-culture, proliferation of OT-I T cells was analyzed by CFSE dilution in TCRβ + CD8α + cells.

Quantitative PCR
Total RNA was isolated with TRIzol reagent (Invitrogen) and cDNA synthesis was done with SuperScript IV Reverse Transcriptase (Thermo Fisher Scientific). Real-time PCR was performed with SYBR Green Master Mix (Applied Biosystems) on a CFX96 Touch System (BioRad). Differential expression is shown as fold change and was calculated with the ΔΔCt method. Expression was normalized to the house keeping gene Tbp (TATA-binding protein). Primers used for quantitative PCR are listed in Supplementary Table 2. RNA-sequencing FLT3L-derived cDC1 FL (CD24 + CD11c + B220 − CD115 − CD172a − ) cells were sorted (FACSAria Fusion cell sorter; BD Biosciences) into TRIzol LS reagent (Invitrogen) and RNA was isolated with the miRNeasy Micro Cleanup Kit (Qiagen). Sequencing libraries were prepared using the NEBNext Poly (A) mRNA Magnetic Isolation Module and the NEBNext Ultra ™ II Directional RNA Library Prep Kit for Illumina according to manufacturer's protocols (New England Biolabs). Libraries were QC-checked on a Bioanalyzer 2100 (Agilent) using a High Sensitivity DNA Kit for correct insert size and quantified using Qubit dsDNA HS Assay (Invitrogen). Pooled libraries were sequenced on a NextSeq500 (Illumina) in 2 × 75 bp paired-end sequencing mode. Reads in fastq format were aligned to the murine reference genome version mm10 (https://www.ensembl.org) with Gencode vM19 annotations (https://www.gencodegenes.org/) using STAR aligner [36] version 2.6.1a in 2-pass mode. Reads per gene were counted by STAR, and differential gene expression was calculated using DESeq2 [37] version 1.20.0. Principal component analysis (PCA) and Euclidean distance plots were generated using R functions prcomp and dist, respectively. GO term enrichment analysis was performed with GOrilla [38] using as an input selections of differentially expressed genes (DEGs) and all mouse Ensemble gene ids as a background. Gene set enrichment analysis (GSEA) was performed using GSEA [39] version 3.0 using as an input regularized logtransformed count data from DESeq2 and selected gene sets from MSigDB [40] version 6.2 as well as custom gene sets.

Venn diagram
To generate Venn diagrams we first made a list of probe sets derived from the published microarray data with the Transcriptome Analysis Software (Thermo Fisher Scientific). We then used the GSEA software (http://www. gsea-msigdb.org/gsea/index.jsp) to collapse the values of the probe sets to a single gene value by gene using the median probe mode. We considered only genes that were annotated on both, the microarray and RNA-Seq platform. We used the R function euler to generate a Venn diagram.

Statistics
GraphPad Prism 8 was used for statistical analysis. Significance was analyzed on pooled data from independent experiments. Animals were randomly assigned to experimental groups. Data were analyzed by unpaired two-tailed Students t test or one-way ANOVA with Tukey post-test, or multiple t-tests with the Holm-Sidak method for grouped data and are shown as mean ± SEM. t-test with Welch´s correction was performed, if variance was significantly different between two sets of data. Outliers were identified by Grubb´s or ROUT method. A P value below 0.05 was considered statistically significant (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001).

Results
Combined deletion of c-Jun and JunB results in loss of resident CD8α cDC1, while migratory CD103 cDC1 remain unaffected To study the role of Jun/AP-1 proteins during DC diversification, we first conducted an unbiased gene expression analysis of c-Jun and JunB across the immune cell compartment using IMMGEN datasets. We found prominent expression of c-Jun and JunB in cDCs, but not in pDCs ( Supplementary Fig. 1a).
To confirm these results, FLT3L-derived cDCs (cDCs FL ) were generated from BM to analyze the expression of c-Jun and JunB during DC development (Fig. 1a). RNA expression from sorted cell subpopulations revealed an increase in JunB expression from progenitors to cDCs FL , whereas levels remained low in pDCs FL . In contrast, c-Jun mRNA levels were highest in pre-cDCs FL and cDC2 FL , intermediate in cDC1 FL and low in pDC FL (Fig. 1b). Analysis of c-Jun and JunB expression in cDCs, pre-cDCs and CDPs in vivo, using previously published microarray data [14], confirmed these results ( Supplementary Fig. 1b).
To determine, whether these Jun/AP-1 factors are functionally involved in DC development, c-Jun fl/fl and JunB fl/fl mice were crossed with Mx-Cre mice allowing induction of Cre-recombinase in a diverse set of cells, including the hematopoietic stem cell compartment, by injection of poly I: C [31,32,34]. No differences in splenic cDC1 and cDC2 subsets, identified by CD8α and CD11b marker expression among cDCs, were observed in c-Jun Δ/Δ Mx-Cre and JunB Δ/Δ Mx-Cre mice (Supplementary Fig. 1c-f). Thus, although c-Jun and JunB are induced during cDC development, these Jun/AP-1 proteins individually are dispensable for cDC diversification.
Closely related AP-1 proteins can functionally substitute for the absence of a family member during development and proliferation [24]. We speculated that c-Jun and JunB might compensate for each other, so that only combined deletion of both AP-1 factors would affect DC diversification.
Loss of c-Jun/JunB in c-Jun/JunB Δ/Δ Mx-Cre mice (Fig. 1c) induced a dramatic reduction in the frequency and total cell number of splenic CD8α cDC1, whereas numbers of CD11b cDC2 were unchanged (Fig. 1d, e). DEC-205 expression correlates with CD8α on cDC1s [41]. The DEC-205 + CD11c + population in c-Jun/JunB Δ/Δ Mx-Cre mice was reduced, confirming that the observed phenotype in cDC1 cells was not caused by down-regulation of CD8α on cDCs (Fig. 1f). Moreover, changes in the splenic immune cell compartment, in particular myeloid cells (CD11b + Ly-6C/G + ), could be observed. The frequency of splenic B and T cells was altered, the absolute number, however, was not affected (Supplementary Fig. 2a, b). Interestingly, these observed changes in the immune cell composition preceded a more severe phenotype that developed within 2-3 weeks of Cre recombinase induction and included an inflamed skin ( Supplementary  Fig. 2c), weight loss and increased mortality (data not shown).
In c-Jun/JunB Δ/Δ Mx-Cre mice deletion of c-Jun/JunB occurs already in hematopoietic stem cells and might affect the commitment of multipotent progenitors to a single fate. We detected a significant expansion of granulocytemacrophage progenitors (GMPs) in the BM of c-Jun/JunB Δ/ Δ Mx-Cre mice, whereas progenitors, precursors of cDCs and lymphocytes were unchanged (Supplementary Fig. 2d-f).
Taken together, these data imply that c-Jun and JunB have an overlapping function and are dispensable for early DC development, but crucial for the terminal differentiation to CD8α cDC1.
Intrinsic and extrinsic factors determine the fate of CD8α cDC1 in c-Jun/JunB Δ/Δ Mx-Cre mice To determine, if the diminished CD8α cDC1 population in c-Jun/JunB Δ/Δ Mx-Cre mice was due to cell-intrinsic effects of c-Jun/JunB in DCs or a consequence of extrinsic c-Jun/ JunB dependent signals BM chimeras were generated. Loss of c-Jun/JunB only in hematopoietic cells partially rescued the splenic cDC1 phenotype after Cre recombinase activation, when compared to control mice (Fig. 2a, b). Thus, extrinsic and intrinsic factors are critical for CD8α cDC1 specification. In contrast, the increase in myeloid cells was rescued, if c-Jun/JunB was deleted only in BM cells (Fig. 2c).
To specify, if extrinsic signals from hematopoietic cells determine CD8α cDC1 fate competitive, mixed BM chimeras were generated by giving wild-type (CD45.1/2) and c-Jun/JunB fl/fl or c-Jun/JunB fl/fl Mx-Cre (CD45.2) BM cells in a ratio of 1:1 into c-Jun/JunB fl/fl mice. After recombination by poly I: C, the overall frequency of CD8α cDC1 was significantly reduced in mice harboring wild-type and c-Jun/JunB Δ/Δ Mx-Cre BM (Fig. 2d, e, left panel). The few remaining CD8α cDC1 were equally derived from both, wild-type and c-Jun/JunB-deficient BM (Fig. 2e,  right panel).
Among other organs Mx-Cre activation also results in genomic recombination in skin cells. To investigate, whether extrinsic factors influence CD8α cDC1 differentiation we used a mouse model with deletion of c-Jun/JunB restricted to keratinocytes (c-Jun/JunB Δ/Δ K5-Cre ER ) that results in a psoriasis-like skin inflammation [25,43]. Splenic CD8α cDC1 were diminished in c-Jun/JunB Δ/Δ K5-Cre ER mice, although not to the extent as in c-Jun/JunB Δ/Δ Mx-Cre mice ( Supplementary Fig. 3a-c). In contrast, we did not observe a reduction in the Imiquimod induced psoriasis-like skin inflammation model (Supplementary Fig. 3d, e).
G-CSF, a JunB target gene [44], has been shown to influence cDC1 development [45]. To test the role of G-CSF signaling in our observed phenotype, we treated c-Jun/JunB Δ/Δ Mx-Cre with anti-G-CSF neutralizing antibodies in vivo (Fig. 2f). We found that suppression of G-CSF signaling partially restored splenic cDC1 in c-Jun/JunB Δ/Δ Mx-Cre mice and normalized the frequency of monocytes and neutrophils (Fig. 2g).
In summary, these data show that extrinsic and intrinsic mechanisms controlled by c-Jun/JunB have synergistic effects and together cause a severe reduction of cDC1 in c-Jun/JunB Δ/Δ Mx-Cre mice.  Fig. 4a). However, loss of c-Jun/JunB at the pre-cDC stage induced a significant reduction of CD8α cDC1 (Fig. 3a, b) implying a phenotype similar to c-Jun/JunB Δ/Δ Mx-Cre mice. The overall reduction in CD8α cDC1 was less severe than in c-Jun/JunB Δ/ Δ Mx-Cre mice and more comparable to WT + Δ/Δ BM chimeras (Fig. 2b). Other markers (XCR1, CD24 and DEC-205) [6] that identify the splenic cDC1 subset were diminished in the spleen, while the frequency of CD11b cDC2 was not altered (Fig. 3a, b, Supplementary Fig. 4b). Consistent to our data in the Mx-Cre mouse model, skin-draining LNs of c-Jun/JunB Δ/Δ Itgax-Cre mice showed a reduced number of CD8α rDC, whereas the CD103 mDC subset was not affected (Fig. 3c, d). Moreover, c-Jun/JunB Δ/Δ Itgax-Cre mice showed normal development of cDC-and myeloid progenitors in the BM, including pre-cDC1 and pre-cDC2 subpopulations ( Supplementary Fig. 4c-e). These results further support our hypothesis that c-Jun and JunB together promote the differentiation of pre-cDCs to cDC1s under homeostatic conditions.
In vitro, significantly less cDC1 FL developed from c-Jun/ JunB Δ/Δ Itgax-Cre BM. However, a prominent CD24 int population was seen in c-Jun/JunB Δ/Δ Itgax-Cre FLT3L BM cultures that lacked expression of typical DC markers, like XCR1 or MHC-II, resembling an immature cDC1 FL population (Fig. 3e-f). In contrast, normal differentiation of  , i.p.). Anti-G-CSF IgG antibodies (50 µg, i.p.) were injected 2 and 4 days after deletion of c-Jun/JunB. Isotype was injected in controls. Analysis of splenic cDC1 and myeloid cells was done on day 6 post poly I: C injection. g Splenic CD8α cDC1 (CD11c + MHCII + XCR1 + CD8α + ), monocytes (Ly6-C + CD11b + ) and neutrophils (Ly6G + CD11b + ) were analyzed in mice of the indicated genotype treated as described in (f). Data are representative of 1-3 independent experiments. Flow cytometry plots shown are pre-gated on single, live cells. Dots indicate number of individual mice per experimental group. Error bars represent mean ± SEM. Statistical significance was determined by unpaired twotailed Student's t test (b, c and e) and one-way ANOVA with Tukey post-test (g). **P < 0.01, *P < 0.05 and ns > 0.05. cDC2 FL , pDC FL and CDP FL populations was observed, whereas pre-cDCs FL were reduced after one week of culture (Fig. 3e, g and Supplementary Fig. 4f, g). Thus, c-Jun/JunB are required cell-intrinsically to orchestrate the final step of cDC1 development, the transition of pre-cDCs to cDC1, both in vitro and in vivo.

Global transcriptional changes in c-Jun/JunB deficient cDC1
Next, we performed RNA sequencing to characterize the transcriptional profile in c-Jun/JunB deficient cDC1 in an unbiased manner. We compared in vitro generated cDC1s FL from c-Jun/JunB Δ/Δ Itgax-Cre to c-Jun/JunB fl/fl mice and found 432 genes to be differentially expressed (Supplementary Table 3). Gene ontology (GO) term analysis of the DEGs revealed down-regulation of genes linked to antiviral responses, like double-stranded RNA binding and 2′-5′oligoadenylatesynthetase in c-Jun/JunB-deficient cDC1s FL , whereas genes linked to cell surface expression, enabling carbohydrate binding or detection of external stimuli, were up-regulated (Fig. 5a). Principal component analysis (PCA) further confirmed a clear segregation of c-Jun/JunB-deficient cDC1s FL from control cDCs (Fig. 5b) and GSEA showed significant enrichment of cDC1 specific gene sets in the control only (Fig. 5c). Moreover, previously identified cDC1-specific genes (e.g., Ifi205, Rab7b, Snx22, Tlr3, Wdfy4 and Xcr1) [49,50] were all reduced in c-Jun/JunBdeficient cDC1s FL (Fig. 5d). Among the top-ranked DEGs we found genes that are involved in functional processes, like the prostaglandin inactivating enzyme Hpgd, or a protease, Ctse, involved in antigen processing, a gene related to glycolysis (Eno1b), to interferon-signaling (Ifi44l) and an uncharacterized zinc finger transcription factor (Zfp990) (Fig. 5e).
To define genes that are common or unique to c-Jun/ JunB, we compared our RNA-Seq data to published microarray data of Batf3 −/− cDC1 (Batf3 genes) and Irf4 −/− Irf8 −/− cDC1 FL with restored IRF8 expression (IRF8 genes) [28,49]. Among others, transcripts for the cDC2associated TFs Irf4, Klf4 and Ehf [51] were increased in both c-Jun/JunB and Batf3 deficient cDC1, whereas transcripts associated with cDC1 (Snx22, Tlr3, Xcr1) or cellular signaling (Plce1, Plpp1, Spp1, Fzd1) were reduced ( Fig. 5f and Supplementary Table 4). Similar overlaps were observed comparing c-Jun/JunB to IRF8-dependent transcripts ( Fig. 5g and Supplementary Table 5). These results show that c-Jun/JunB play an essential role to promote the gene expression in cDC1. cDC1 lacking c-Jun/JunB are functionally impaired in immune defense mechanisms Our RNA-Seq analysis highlighted a role for c-Jun/JunB in the expression of genes (Tlr3, Wdfy4) and pathways (ds-RNA binding) associated with cDC1 function. Hence, we next investigated, whether these transcriptional changes impaired typical immune-functions of cDC1 in vivo and in vitro.
Importantly, we found that splenic CD8α cDC1 and in vitro generated cDC1 FL lacking c-Jun/JunB expressed significantly less IL-12p40 and failed to up-regulate the co-stimulatory molecules CD80 and CD86 after stimulation with the TLR3 ligand poly I: C (Fig. 6a-c, Supplementary Fig. 5c-d). Moreover, antigen crosspresentation was impaired in the absence of c-Jun/JunB in cDCs FL and a significantly reduced antigen-specific (Ovalbumin) CD8α OT-I T cell proliferation was induced, when compared to control (Fig. 6d, e). CD8α cDC1 have been shown to be a cellular entry point for infection with Listeria monocytogenes [52]. We observed a significantly reduced pathogen burden in the spleen and liver of c-Jun/JunB Δ/Δ Itgax-Cre mice infected with Error bars represent mean ± SEM. Statistical significance was determined by unpaired two-tailed Student's t test (b, d and g). ***P < 0.001, **P < 0.01, *P < 0.05 and ns > 0.05.
L. monocytogenes (Fig. 6f). Hence, these data suggest that c-Jun/JunB are important regulators for the functional properties of cDC1 in immune responses.

c-Jun and JunB maintain cDC1 identity
Next, we investigated whether the functional impairment of the remaining CD8α splenic cDC1 was caused by an altered cDC1 lineage specification. Interestingly, the expression of TFs essential for cDC2 development and function, namely Irf4 and Klf4, were significantly increased in c-Jun/JunBdeficient cDC1, although expression levels were lower compared to cDC2 (Fig. 7a). Despite comparable levels of Batf3 and TFs involved in pDC development (Tcf4 and Spi-B), Irf8 was significantly reduced in cDC1 lacking c-Jun/ JunB (Fig. 7a). Similarly, c-Jun/JunB-deficient cDC1 FL showed upregulated Irf4 and Klf4 transcript levels, which remained high during the course of differentiation compared and LCs (XCR1 − CD11b + CD24 + cDCs). Representative plots are shown and graphs are pregated on single, live cell populations. b Frequency and total numbers of cutaneous CD103 − and CD103 + cDC1, cDC2 and LCs as defined in (a) are shown. c Representative flow cytometry plots of cDCs (B220 − Gr-1 − CD11c + MHC-II + ) generated from FLT3L (cDC FL ) or FLT3L + GM-CSF (cDC FL + G ) supplemented BM cell cultures of the indicated genotypes. Gating for CD103 − and CD103 + cDC1 (CD24 + CD172a − cDCs) and cDC2 (CD24 − CD172a + cDCs) in vitro is shown. d Frequency and total numbers of in vitro generated CD103 − or CD103 + cDC1 FL or cDC1 FL+ G described in (c) are shown. Data are representative of 2-3 independent experiments. Dots indicate number of individual mice per experimental group. Error bars represent mean ± SEM. Statistical significance was determined by unpaired two-tailed Student's t test (b) and multiple t-test with the Holm-Šídák method (d). **P < 0.01, *P < 0.05 and ns > 0.05. to control cells, while pDC-and cDC1-defining TFs were unchanged, except for Irf8 in pre-cDCs ( Supplementary  Fig. 6a-c). Consistently, splenic cDC1 lacking c-Jun/JunB showed significantly elevated IRF4 and reduced IRF8 protein levels (Fig. 7b-c).
Besides TFs, cellular identity is defined by exclusive expression of specific cell surface markers. Our RNA-Seq data showed an elevated expression of surface marker transcripts on c-Jun/JunB deficient cDC1 FL that are normally found on cDC2 or pDCs (Supplementary Fig. 7a). UMAP analysis, revealed a reduction of cDC1 and a shift to a cDC2 phenotype in c-Jun/JunB Δ/Δ Itgax-Cre mice (Fig. 7d). Analysis of the cDC1 population, defined in the UMAP space, showed an elevated expression of cDC2 specific markers (CD4 and CD11b) on c-Jun/JunB deficient cDC1 (Fig. 7e). Correct population clustering by UMAP and increased surface expression of CD4 and CD11b was confirmed by conventional gating (Supplementary Fig. 7b-d). Next, we sorted CD8α rDC and CD103 mDC from sd-LN to better characterize the TF profile in these two population. Similar to splenic CD8α cDC1 resident as well as migratory cDC1 showed a shift to a cDC2 phenotype with reduced Irf8, but increased Irf4 and Klf4 mRNA expression, whereas the pDC TFs Tcf4 and Spi-B were unchanged (Fig. 7f). Analysis of IRF4 and IRF8 protein by flow cytometry confirmed this altered cDC2-like phenotype (Fig. 7g, h). Lastly, deletion of c-Jun and JunB in sorted cDC1 and cDC2 was confirmed by quantitative RT-PCR ( Supplementary Fig. 7e-f).
In conclusion, in this study we discovered a novel TF pair, c-Jun/JunB, required for CD8α cDC1 development, function, and maintenance of their identity.
Discussion cDC1 differentiation and identity is controlled by a complex network of linage-determining TFs. So far IRF8, ID2, NFIL3 and Batf3 have been reported as critical regulators at different stages of CD8α and CD103 cDC1 development [7]. In this study, we identify the AP-1 family member's c-Jun and JunB as two novel TFs essential for the development and function of the CD8α cDC1 lineage. cDC1 develop from progenitor cells residing in the BM (CDPs), and differentiate in the tissue from a transitory, immediate progenitor population, the pre-cDC1 [7]. Our results demonstrate that the combined deletion of c-Jun and JunB in CD11c + cells induces a severe defect in the specification of pre-cDCs to CD8α cDC1. Importantly, other immune cell subsets developed normally in this mouse model, emphasizing c-Jun/JunB as specific-lineage determining TFs for CD8α cDC1 commitment.
In single c-Jun or JunB knock-out mice no effect on the CD8α cDC1 pool was observed. Thus, c-Jun and JunB seem to either compensate for each other or cooperate by a currently unclear mechanism. Our data demonstrate for the first time such a non-redundant function for c-Jun and JunB in immune cells, but similar observations have been made in keratinocytes and hair follicle stem cells [25,53].
Pathogenic infections or transplantations can restore the cDC1 pool independently of Batf3 [28,56]. Batf3 can be replaced by its family members Batf or Batf2 during infections with intracellular pathogens [28]. Batf has been described to hetero-dimerize with c-Jun, JunB or JunD in vitro, whereas for Batf2 and Batf3 only heterodimerization with c-Jun has been tested [26,57,58]. Our findings suggest that both c-Jun and JunB are nonredundant partners for Batf3 in the lineage decision of pre-cDCs. Additional studies are required to test, if c-Jun and JunB are also essential interaction partners for Batf or Batf2 to mediate compensatory DC development. In Batf3 −/− mice, also T-cell function is affected [59]. Therefore, c-Jun/JunB Δ/Δ Itgax-Cre mice could be used as an alternative mouse model to study CD8α cDC1 biology under homeostatic and inflammatory conditions.
The classic DNA-binding element bound by AP-1 proteins is the TPA-responsive element (TRE) [16], but cooperation with IRF proteins expands the repertoire to AICE elements [28]. Although a proportion of the DEGs in c-Jun/JunB deficient cDC1 was overlapping with Batf3, the majority of genes were exclusive and associated with cell function, signaling and metabolism.  Future studies could elucidate the role of classical elements (TRE) bound by AP-1 dimers compared to AICE elements recognized by AP-1-IRF hetero-trimers for cDC1 biology.
Importantly, our findings emphasize that not only deletion of c-Jun/JunB in hematopoietic cells, but also in other cell types affects cDC1 development. In a recent study, G-CSF, which is a direct transcriptional target of JunB in keratinocytes [44], has been shown to interrupt IRF8 dependent cDC1 development in the BM [45]. Here, we show that neutralization of G-CSF in c-Jun/JunB Δ/Δ Mx-Cre mice partially rescued cDC1 numbers in the spleen. Future work should examine this novel extrinsic mechanism of cDC1 differentiation in more detail. c-Jun/JunB deficient cDC1 have reduced Irf8 expression, despite normal levels of Batf3, and show a shift towards a cDC2 phenotype. Similarly, Batf3 −/− -deficient pre-CD8 DC cannot maintain IRF8 expression and divert to cDC2 [14].
Irf8 expression was also reduced in c-Jun/JunB deficient CD103 mDC, but their frequency was normal. Given the dependence of this DC subset on GM-CSF [47], further studies are needed to clarify its role in promoting CD103 mDC development/survival in the absence of c-Jun/JunB. Collectively, our findings identify a key role for c-Jun/ JunB in cDC1 development, identity and function through extrinsic and intrinsic mechanisms.

Data availability
The RNA-Seq dataset has been deposited to GEO under the accession number GSE156484.
Author contributions PN and BD contributed equally to this work. BD and PN performed and analyzed most of the experiments. MH participated in experimental design, interpretation of data, in the antigenpresentation experiment, cell sorting and writing of the paper. CDSF performed some of the cell culture in vitro experiments and assisted in flow cytometry experiments. RT performed the Listeria monocytogenes infections. SD performed analysis of data generated by RNA-Sequencing, TD and EFW provided technical support. MS supervised the project and provided the requested funding.

Compliance with ethical standards
Conflict of interest The authors declare no competing interests.
Ethical approval All animal experiments conducted were compliant with federal laws and guidelines of the Medical University of Vienna.
Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons. org/licenses/by/4.0/. Fig. 7 c-Jun and JunB maintain cDC1 identity. a Splenic CD8α cDC1 (CD3 -CD19 -BST-2 -Siglec-H -CD11c + I-A/I-E + XCR1 + CD8α + ) from c-Jun/JunB Δ/Δ Itgax-Cre and c-Jun/JunB fl/fl mice were sorted and RT-qPCR analysis of Irf8, Batf3, Irf4, Klf4, Tcf4 and Spi-B mRNA was performed (n = 11-17 for cDC1 and 4-9 for cDC2 subgroups). b Splenic cDC (CD11c + I-A/I-E + ) subsets cDC1 (XCR1 + cDC) and cDC2 (CD172a + cDC) were analyzed for the protein expression of the transcription factors IRF4 and IRF8 by intracellular flow cytometry. Shown are representative histograms of IRF4 and IRF8 expression in cDC1 and cDC2 from indicated mice. c Graph depicts the mean fluorescence intensity (MFI) of IRF4 and IRF8 in splenic cDC subsets defined in (b). d UMAP analysis of splenic cDCs (CD11c + I-A/I-E + ) acquired by flow cytometry, including the surface markers CD4, CD8α, CD11b, CD172a, ESAM, and XCR1 in c-Jun/JunB Δ/Δ Itgax-Cre and c-Jun/JunB fl/fl mice. Samples were concatenated (n = 4 per genotype) to perform the analysis. e Expression of CD4 and CD11b on cells within the cDC1 population, as defined by UMAP in (d), was analyzed. Flow cytometry plots (left) and histograms (right) are shown. f Lymph-node resident (CD3 -CD19 -NK1.1 -TCRβ -CD11c + I-A/ I-E int XCR1 + CD8α + ) and migratory cDC1 (CD3 -CD19 -NK1.1 -TCRβ -CD11c + I-A/I-E high XCR1 + CD103 + ) from c-Jun/JunB Δ/Δ Itgax-Cre and c-Jun/JunB fl/fl mice were sorted and RT-qPCR analysis of Irf8, Batf3, Irf4, Klf4, Tcf4 and Spi-B mRNA was performed (n = 6-7). g Lymphnode CD8α rDC (CD11c + I-A/I-E int XCR1 + CD11b -), CD103 mDC (CD11c + I-A/I-E high XCR1 + CD11b -CD103 + ) and CD11b rDC (CD11c + I-A/I-E int XCR1 -CD11b + ) were analyzed by intracellular flow cytometry for IRF4 and IRF8 protein expression in indicated mice. Shown is the log 2 Fold change of the MFI for IRF4 and IRF8 relative to the c-Jun/JunB fl/fl CD103 mDC subset, which is set to 0. (n = 7 c-Jun/JunB fl/fl and n = 8 c-Jun/JunB Δ/Δ Itgax-Cre). h Graph depicts the MFI of IRF4 and IRF8 in skin-draining lymph node cDC subsets as defined in (g). Data are representative of 2-4 independent experiments. Dots indicate number of individual mice per experimental group. Error bars represent mean ± SEM. Statistical significance was determined by multiple t-test with the Holm-Šídák method (a, c, f, h). ****P < 0.0001, ***P < 0.001, **P < 0.01, * P < 0.05, ns > 0.05.