Classical dendritic cells (cDC) can be classified into two major subsets: Irf8-dependent cDC1 and Irf4-expressing cDC2. Although these subsets play distinct roles in intestinal immune homeostasis, their functions in T cell–driven colitis remain unknown. To assess the role of IRF4 expression in cDC2 in T cell–driven colitis, CD11c-Cre.Irf4fl/fl and Irf4fl/fl mice were backcrossed onto a Rag-1−/− background and used as recipients of CD45RBhiCD4+ T cells. Colitis score and innate immune cell influx were reduced in Cre+ mice 4 wk posttransfer, and these changes were associated with reduced CD4+ T cell counts in both the mesenteric lymph nodes and colon. By 7 wk, colitis score and colon CD4+ T cell numbers were similar in Cre+ and Cre− mice despite a selective reduction in Th17 cells in the colon of Cre+ mice and a continued reduction in CD4+ T cell numbers in mesenteric lymph nodes. Cotransfer of CD25+CD45RBlo CD4+ T cells prevented CD45RBhiCD4+ T cell–driven colitis in both Cre+ and Cre− recipients, demonstrating that IRF4 expression by cDC is not required for CD4+ regulatory T cell–mediated control of colitis. Collectively these results suggest a role for IRF4 expression in cDC2 in the generation of colitogenic CD4+ T cells, which becomes redundant as colitis progresses.
The intestinal lining represents the largest surface area of the body that interfaces with the external environment and is continually exposed to foreign material derived from our diet and the complex microbiota found in the intestinal lumen (1, 2). The intestinal microbiota is essential for health, including immune system development and function (3, 4), and continual cross-talk between the microbiota and mucosal immune system is required for tissue homeostasis (5–7). Alterations in this dialogue, for example, through changes in the microbiome, challenges to the environment, or unusual host genetic susceptibility can result in dysregulated tissue responses that are thought to contribute to the development and maintenance of inflammatory bowel disease (IBD) (8–10). However, our understanding of the intestinal immune cell populations that contribute to the initiation and pathologic condition of IBD remains incomplete.
Classical dendritic cells (cDCs) are the major APCs of the immune system and additionally sense and relay environmental signals that determine the resulting adaptive immune response. In this way, cDC are the central drivers and regulators of adaptive immunity. As in all tissues, intestinal cDC can be divided into two major subsets, cDC1 and cDC2 (for review, see Ref. 11). cDC1, broadly identified by their expression of X-C motif chemokine receptor 1 (XCR1) and lack of expression of signal-regulatory protein α (SIRPα) or CD11b, are developmentally dependent on the transcription factors Batf3, Irf8, and Id2 (12–14). In contrast, cDC2, identified as SIRPα+ XCR1−, are phenotypically heterogeneous in the intestine, comprising CD103+CD11b+ and CD103−CD11b+ subsets that develop independently of Batf3, Irf8, and Id2, but are, in part, dependent on Notch2 and Irf4 (15, 16). We and others have demonstrated that intestinal cDC1 and cDC2 play nonredundant roles in intestinal immune homeostasis (16–19). Whereas cDC2 participate in intestinal Th17 and Th2 responses (15–17, 20–22), cDC1 are required for cross-presentation of epithelial derived self-antigen to CD8+ T cells, and for intestinal Th1 and intraepithelial lymphocyte homeostasis (18, 19, 23).
Recent studies have begun to address the function of intestinal cDC subsets in colitis. cDC1 deficiency in XCR1–diphtheria toxin A (DTA) mice, or depletion of cDC1 by administration of diphtheria toxin to CLEC9A–diphtheria toxin receptor (DTR) mice, leads to enhanced susceptibility to dextran sodium sulfate (DSS)–induced colitis (19, 24). In contrast, deficiency of CD103+CD11b+ cDC2 in human langerin-DTA mice (25) or depletion of cDC2 by diphtheria toxin administration to CLEC4a-DTR mice (24) has no effect on DSS-induced colitis (24). Importantly, the acute DSS-induced colitis model used in these studies is driven by innate immune cells that are recruited into the intestine following DSS-induced epithelial destruction and is T cell independent (26), and the role of intestinal cDC subsets in T cell–dependent colitis thus remains unstudied.
We have shown previously that CD11c-cre.Irf4fl/fl mice display a selective loss of CD103+CD11b+ cDC2 in the small intestinal lamina propria and in particular in intestinal draining mesenteric lymph nodes (MLN). cDC2 that remain in CD11c-cre.Irf4fl/fl mice upregulate CCR7 as efficiently as wild-type (WT) cDC2 but die rapidly in vitro, indicating a key role for IRF4 in intestinal cDC2 survival but not migration (16). In this study, we explored the role of IRF4 expression by cDC in T cell–dependent colitis by transferring CD45RBhi CD4+ naive T cells into CD11c-cre.Irf4fl/fl mice backcrossed onto an Rag-1−/− background and have further examined if CD45RBlo CD4+ regulatory T cells (Tregs) can prevent colitis in this setting.
Materials and Methods
Rag-1−/−, Rag-1−/−.Irf4fl/fl, and Rag-1−/−.CD11c-Cre.Irf4fl/fl mice on a C57BL/6 background were bred and maintained at the Lund University Biomedical Center. Irf4fl/fl mice were originally obtained from U. Klein [Columbia University (27)] and CD11c-Cre mice from B. Reizis [Columbia University Medical Center (28)]. Female C57BL/6 mice were purchased from Taconic Biosciences (Germantown, NY) or Janvier Labs (Le Genest-Saint-Isle, France). Littermate controls were used for all experiments. All animal experiments were performed in accordance with the Lund/Malmö Animal Ethics Committee.
PBS, HBSS, RPMI 1640, FCS, HEPES buffer, EDTA, sodium pyruvate, 2-ME, penicillin, streptomycin, and gentamicin were obtained from Invitrogen. FCS, brefeldin A, PMA, collagenase IV, and DTT were obtained from Sigma-Aldrich. Ionomycin was from Calbiochem, whereas DNase I and Liberase were obtained from Roche.
Cells were isolated from the colon lamina propria (cLP) by enzymatic digestion after removal of epithelial cells, as described previously (29). In short, the colon was cut longitudinally and cut into 5-mm pieces. Epithelial cells and mucus were removed by incubating tissue in HBSS supplemented with FCS (2%), DTT (0.015%), and EDTA (2 mM) for 15 min at 37°C with continual shaking at 450 rpm; the samples were then filtered, and EDTA treatments were repeated twice without DTT supplementation. Remaining tissue pieces were incubated in RPMI 1640 supplemented with FCS (10%), HEPES (10 mM), sodium pyruvate (1 mM), penicillin (100 U/ml), and streptomycin (100 mg/ml) (R10 medium) containing DNase I (12.5 μg/ml) and Liberase (0.3 U/ml) for 45 min at 37°C. Resulting cell suspensions were filtered consecutively through 100- and 40-μm cell strainers (BD Biosciences) prior to further analysis. For isolation of MLN cells, MLN were cut into small pieces and digested with R10 medium containing collagenase IV (0.5 mg/ml) and DNase I (12.5 μg/ml) for 30 min while shaking at 37°C, followed by filtering through a 70-μm filter as described previously (29).
Flow cytometry and Abs
Flow cytometry was performed according to standard procedures. Dead cells identified as fluorescent reactive dye+ by Fixable Viability Dye eFluor 450 (eBioscience) or by LIVE/DEAD Fixable Near-IR or Red Dead Cell Stain Kit (Life Technologies), and cell aggregates were excluded from analyses using forward scatter-area-A versus forward scatter-height scatterplots. For intracellular staining, cells were fixed and permeabilized using the Foxp3 Fixation/Permeabilization Kit (eBioscience). Cells were analyzed on an LSR II (BD Biosciences) and flow cytometry data analyzed using FlowJo software (Tree Star). The following Abs were from BioLegend: anti-CD45RB (C363.16A), CD45.2 (104), CD45.1 (A20), CD19 (6D5), CD25 (M1/69), Ly-6–C (HK1.4), Ly-6–G (1A8), TER-119 (TER-119), CD172a/SIRPα (P84), NK1.1 (PK136), MHC class II (MHCII; M5/114.15.2), B220 (RA3-6B2), XCR1 (ZET), Helios (22F6), IFN-y (XMG1.2), and IL-17 (TC11-18H10.1). Anti-CD45 (30-F11), CD64 (x45-5/7.1), Siglec-F (E50-2440), CD103 (M290), and BV605 streptavidin were from BD Biosciences. Anti-CD4 (RM4-5), CD11c (N418), CD11b (M1/70), TCRβ (H57-597), Foxp3 (FJK-16s), RORγt (B2D), and IRF4 (3E4) were from eBioscience.
T cell transfers
T cell transfers were performed as described previously (30). Briefly, CD4+ T cells were enriched from spleen cell suspensions using anti-CD4–conjugated MACS beads (Miltenyi Biotec). After surface staining, CD4+CD25−CD45RBhi T cells were purified by cell sorting using an FACSAria II (BD Biosciences), and 3 × 105 cells were injected i.p. into 6–8-wk-old Rag-1−/−.CD11c-cre.Irf4fl/fl mice or Rag-1−/−.Irf4fl/fl littermates. When indicated, CD4+CD25−CD45RBhi and CD4+CD25+CD45RBlo T cells were coinjected (1.5 × 105 cells of each population/mouse) into recipient mice as described above.
Samples of proximal, mid, and distal colon were prepared and inflammation scored in a blinded manner as described previously (31, 32). Briefly, colon sections were scored from 0 to 3 on each of four criteria: 1) crypt length, degree of epithelial hyperplasia, and goblet cell depletion; 2) leukocyte infiltration in the lamina propria; 3) area of tissue affected; and 4) presence of markers of severe inflammation, such as crypt abscesses, submucosal inflammation, and ulcers. Total scores from the proximal, mid, and distal colon were averaged to give an overall inflammation score of 0–12. Slides were scanned with an Aperio Cs2 (Leica Biosystems) and analyzed with Aperio ImageScope software (Leica Biosystems).
Serum cytokine assays
Lipocalin-2 levels were assessed by ELISA (R&D Systems). Serum IL-6, TNF-α, and IFN-γ levels were determined using CBA mouse cytokine flex sets in combination with CBA mouse soluble protein Master Buffer Kit (BD Biosciences). Samples were analyzed using an Accuri C6 (BD Biosciences) and the FCAP Array Software (BD Biosciences).
Gene expression analysis
Total RNA was isolated from colonic tissue using the RNeasy Micro Kit (QIAGEN). RNA was reverse transcribed to complementary DNA using the SuperScript III First-Strand Synthesis System (Invitrogen). Gene expression was analyzed by quantitative reverse transcription PCR using the SYBR Green Master Mix (Bio-Rad Laboratories) and the following primers (Eurofins Genomics): IL-10 forward 5′-GGCCAGTACAGCCGGGAAGA-3′, IL-10 reverse 5′-AAACTGGATCATTTCCGATAAGGC-3′, IL-23p19 forward 5′-AGCGGGACATATGAATCTACTAAGAGA-3′, and IL-23p19 reverse 5′-GTCCTAGTAGGGAGGTGTGAAGTTG-3′. Gene expression was analyzed using GAPDH as an endogenous control, and the mean relative gene expression was determined, and differences were calculated with the 2−ΔC(t) method (33).
Statistical significance was determined using GraphPad Prism software (GraphPad), using the nonparametric Mann–Whitney U test.
Analysis of colon and MLN cDC subsets in Rag-1−/− mice
Because intestinal cDC have not been assessed in detail in Rag-1−/− mice that were to be used as recipients of CD45RBhi CD4+ T cells, we first compared the intestinal cDC subset composition of Rag-1−/− with age- and sex-matched C57BL/6 mice (Fig. 1). XCR1+ cDC1 and SIRPα+ cDC2 were readily detected in the colon of Rag-1−/− mice (Fig. 1A). Despite the expected reduction in CD45+ cells in the colon of Rag-1−/− compared with WT mice (Fig. 1B), cDC1 and cDC2 numbers were increased in the colon of Rag-1−/− mice (Fig. 1C, Supplemental Fig. 1A). The ratio of cDC1 and cDC2 in the colon of Rag-1−/− mice was similar to that of WT mice (Fig. 1D), but the proportion of colon cDC2 that expressed CD103 was higher in Rag-1−/− mice (Fig. 1E). IRF4 expression, as assessed by intracellular staining, was restricted to cDC2 in both Rag-1−/− and C57BL/6 colon, and similar levels were found in these mice (Fig. 1F).
Both MHCIIhi migratory and a minor population of MHCIIint lymph node–resident cDC (34, 35) were present in the MLN of Rag-1−/− mice (Fig. 1G), but their numbers and those of total CD45+ cells were reduced compared with WT MLN (Fig. 1H, 1I). The proportion of cDC1 and cDC2 within MHCIIhi MLN cDC was the same in Rag-1−/− and C57BL/6 mice (Fig. 1J), but the levels of XCR1 on MHCIIhi cDC1 and SIRPα on MHCIIhi cDC2 were reduced in Rag-1−/− mice (Supplemental Fig. 1B). Despite their reduced SIRPα expression, MHCIIhi cDC2 in Rag-1−/− mice expressed comparable levels of IRF4 (Fig. 1K) and CD103 to those found in C57BL/6 mice (Fig. 1L). Thus, IRF4-expressing cDC2 are present in the colon and MLN of Rag-1−/− mice and at similar proportions as in WT mice.
CD11c-Cre.Irf4fl/fl.Rag-1−/− mice have reduced numbers of intestinal cDC2 in the colon and MLN
To assess the role of IRF4 expression by cDC2 in the T cell transfer model of colitis, Rag-1−/− mice lacking IRF4 in their cDC compartment were generated by crossing Rag-1−/− with CD11c-Cre.Irf4fl/fl mice (16). cDC2 were present in the colon of CD11c-Cre.Irf4fl/fl.Rag-1−/− (Cre+) mice (Fig. 2A) despite efficient deletion of Irf4 (Fig. 2B). Although CD45+ numbers did not differ between the colon of Cre+ and Cre− mice (Fig. 2C), cDC numbers were significantly reduced in Cre+ mice (Fig. 2D). This involved a marked reduction in cDC2 and a minor, yet statistically significant, reduction in cDC1 (Fig. 2E), resulting in a significant increase in the cDC1/cDC2 ratio in the colon of Cre+ mice (Supplemental Fig. 2A). Both CD103+ and CD103− cDC2 were reduced in the colon of Cre+ mice (Supplemental Fig. 2B), with no difference in the proportion of CD103-expressing cDC2 (Fig. 2F). As in immunocompetent CD11c-Cre.Irf4fl/fl mice (17), the number of colonic CD11c+CD64+ macrophages did not differ between Cre+ and Cre− mice (Supplemental Fig. 2C, 2D), consistent with their lack of expression of IRF4 (17).
In MLN, the numbers of CD45+ cells and of MHCIIint and MHCIIhi cDC were reduced in Cre+ mice (Fig. 2G–J). As in the colon, and as reported in the MLN of lymphocyte replete CD11c-Cre.Irf4fl/fl mice (16), the total number of MHCIIhi cDC2, but not of cDC1, was dramatically reduced in the MLN of Cre+ mice (Fig. 2K), resulting in a marked increase in the cDC1/cDC2 ratio (Supplemental Fig. 2E). CD103+ and CD103− cDC2 were reduced to similar extents in the MLN of Cre+ mice (Fig. 2L, Supplemental Fig. 2F). These results demonstrate efficient deletion of Irf4 in cDC2 from Cre+ mice, which resulted in a significant reduction in cDC2 numbers in the colon and their almost complete absence in MLN.
CD11c-Cre.Irf4fl/fl.Rag-1−/− mice have delayed colitis onset after transfer of CD45RBhiCD4+ T cells
To investigate the role of IRF4 expression by cDC in T cell–mediated colitis, Cre+ and Cre− littermates were injected i.v. with CD45RBhi CD4+ T cells and sacrificed 4 or 7 wk later. Although Cre+ and Cre− mice lost weight to similar extents following T cell transfer (Supplemental Fig. 3A), Cre+ mice had reduced pathology in the colon at 4 wk after T cell transfer compared with Cre− mice, an effect that was not seen at 7 wk (Fig. 3A, 3B). We next assessed inflammatory cell infiltrates in the colon of Cre+ and Cre− mice by flow cytometric analysis. Cell numbers were normalized to the average value of Cre− mice within each experiment because of variation in immune cell numbers between experiments. Consistent with the pathology scores, total CD45+ cell counts were lower in the colon of Cre+ mice at 4, but not 7, wk posttransfer compared with Cre− mice (Fig. 3C). In parallel CD64-expressing, monocyte-derived cells, eosinophils and neutrophils were also all reduced in Cre+ mice at 4, but not 7, wk (Fig. 3D, 3E). In contrast, the total number of CD45+ cells in the MLN were reduced at both 4 and 7 wk after T cell transfer in Cre+ mice (Fig. 3C), including numbers of CD64-expressing, monocyte-derived cells, eosinophils, and neutrophils (Supplemental Fig. 3B). The levels of serum lipocalin-2 were measured as a systemic indicator of inflammation (36) and these did not differ significantly between Cre+ and Cre− mice at 4 or 7 wk after T cell transfer, although there was a trend to reduced levels in Cre+ mice at 4 wk (Fig. 3F). Collectively, these results suggest that deletion of IRF4 in cDC delays the onset of local disease in T cell transfer colitis.
CD4+ T cell numbers and function are altered in CD11c-Cre.Irf4fl/fl.Rag-1−/− mice following T cell transfer
Given the delay in colitis development in Cre+ mice, we next assessed the numbers and function of CD4+ T cells in the MLN and colon of colitic mice (Fig. 4). As above, for statistical analysis of pooled experiments, CD4+ T cell numbers were normalized to the average value of Cre− mice within each experiment (Fig. 4). CD4+ T cell numbers were significantly reduced in the MLN of Cre+ mice at both 4 and 7 wk after T cell transfer (Fig. 4A), affecting IL-17+IFN-γ−, IL-17−IFN-γ+, and IL-17+IFN-γ+ CD4+ T cells (Fig. 4B, 4C). In the colon, total CD4+ T cell numbers were reduced in Cre+ mice at 4, but not 7, wk after T cell transfer (Fig. 4D), with both IL-17+IFN-γ− and IL-17−IFN-γ+ CD4+ T cells being affected at 4 wk (Fig. 4E, 4F), whereas IL-17+IFN-γ− CD4+ T cells remained significantly reduced after 7 wk (Fig. 4F). Results from one experiment with total cell numbers can be seen in Supplemental Fig. 4A–D. Finally, the proportion of transferred CD4+ T cells that converted to Foxp3+ Treg was low in Cre+ and Cre− mice (Fig. 4G, 4H), consistent with previous results in CD45RBhi CD4+ T cell transfer colitis (37, 38); however, as with effector CD4+ T cell subsets, Cre+ mice had lower total numbers of these cells in the MLN at 4 and 7 wk posttransfer (Fig. 4I). Cre+ mice also displayed a slight, yet significant, reduction in Foxp3+ Treg in the cLP at 7, but not 4, wk after T cell transfer (Fig. 4I).
Levels of IL-23 mRNA, which is essential for colitis development in this model (31), were reduced, whereas levels of IL-10 mRNA were increased in the colon of Cre+ compared with Cre− recipients at 4, but not 7, wk after T cell transfer (Supplemental Fig. 4E), consistent with reduced effector CD4+ T cell numbers and colonic inflammation in Cre+ mice early during colitis development. Further serum IFN-γ levels were reduced in Cre+ compared with Cre− recipients at 4, but not 7, wk (Supplemental Fig. 4F), whereas serum levels of TNF-α and IL-6 showed a trend toward reduced levels at 4, but not 7, wk after T cell transfer (Supplemental Fig. 4G, 4H).
Together, these results suggest that IRF4 expression in cDC is required for the optimal generation of colitic CD4+ T cells, but as colitis progresses, CD4+ effector T cell expansion in the colon is independent of IRF4 expression in cDC.
IRF4 expression by cDC is not required for CD25+CD45RBlo CD4+ T cells to inhibit colitis
Cotransfer of CD25+CD45RBlo CD4+ Tregs prevents CD45RBhi CD4+ T cell–mediated colitis in Rag-1−/− mice, and this has been suggested to require recipient-derived cDC (39, 40). However, it remains unclear how and which cDC contribute to the proliferation and regulatory function of these cells in vivo. To assess whether IRF4 expression by cDC is required for the ability of CD25+CD45RBlo CD4+ T cells to prevent colitis, Cre+ and Cre− mice were injected with CD25+CD45RBloCD4+ (CD45.1+) and CD45RBhiCD4+ (CD45.2+) T cells and colitis development assessed 7 wk later (Fig. 5). CD25+CD45RBloCD4+ T cells prevented CD45RBhiCD4+ T cell–mediated colitis in both Cre+ and Cre− mice (Fig. 5A, 5B), and this was associated with similar reductions in myeloid, eosinophil, and neutrophil infiltration into the colon and similarly reduced serum levels of IFN-γ, TNF-α, and IL-6 (Fig. 5C, 5D). Consistent with these findings, the proportion of CD45.1+ CD4+ T cells that expressed Foxp3 7 wk posttransfer was similar between Cre+ and Cre− mice (Fig. 5E, 5F). Thus, IRF4 expression by cDC is not required for the ability of CD25+CD45RBloCD4+ T cells to prevent CD45RBhi CD4+ T cell–driven colitis.
In the current study, we explored the role of IRF4 expression by cDC in T cell–driven colitis. In Rag-1−/− mice, deletion of IRF4 in CD11c+ cells resulted in a reduction in colonic cDC2 and an almost complete loss of MHCIIhi cDC2 in MLN, including both CD103+ and CD103− cDC2. The absence of these cells correlated with reduced colon pathology 4 wk after CD45RBhiCD4+ T cell transfer in Cre+ mice compared with Cre− mice. This was associated with reduced CD4+ T cell and innate immune cell numbers in the colon and MLN. By 7 wk after T cell transfer, Cre+ and Cre− mice had similar colon pathology and colonic innate and CD4+ T cell counts, although Th17 cell remained reduced in the colon of Cre+ mice. In contrast, CD4+ T cell and innate immune cell numbers remained reduced in the MLN of Cre+ mice 7 wk after T cell transfer. Finally, coadministration of CD45RBloCD4+ T cells prevented colitis development in both Cre+ and Cre− mice. Collectively, these results suggest a role for IRF4-expressing cDC2 in the initial priming of colitogenic T cells that becomes redundant as colitis develops.
As Rag-1−/− Cre+ and Cre− mice were used as the recipients for T cell transfer colitis, we initially compared the cDC subset composition of Rag-1−/− with that of WT mice. Consistent with recent results (40), the colon of Rag-1−/− mice contained both cDC1 and cDC2. However, the colon of Rag-1−/− mice contained more cDC compared with C57BL/6 mice, whereas the MLN of Rag-1−/− mice had reduced numbers of MHCIIhi cDC and few MHCIIint cDC. The reason for altered DC homeostasis in Rag-1−/− mice remains unclear; however, given the observed phenotype, we speculate that direct or indirect signals from adaptive immune cells in the colon may be required for driving optimal steady-state migration of cDC from the colon to the MLN. Furthermore, adaptive immune cell–derived TNF superfamily members have been implicated in promoting cDC survival within lymph nodes (41, 42). In addition to alterations in cDC numbers, we also noted alterations in the phenotype of cDC1 and cDC2 in Rag-1−/−mice. First, a greater proportion of cDC2 in Rag-1−/− colon expressed CD103. As CD103 induction on intestinal cDC2 has been linked to TGF-β (40), we hypothesize that the absence of adaptive immune cells in Rag-1−/− mice may result in increased availability of bioactive TGF-β in the colon. Second, the levels of SIRPα on cDC2 and XCR1 on cDC1 in the MLN were significantly reduced, indicating a potential role for adaptive immune cells in regulating expression of these receptors. These findings highlight a potential caveat to our current approach that uses recipients on a Rag-1−/− background, and it will be important in the future to confirm our findings in T cell–dependent colitis models in which control mice have DC numbers and a DC phenotype that reflect that of WT mice.
CD11c-driven deletion of Irf4 in Rag-1−/− mice led to reduced numbers of colon cDC2, a slight reduction in cDC1, and a general skewing of the colonic cDC ratio toward cDC1. As we found previously in lymphocyte replete CD11c-Cre.Irf4fl/fl mice (16), the impact of Irf4 deletion on cDC2 numbers was even more dramatic in MLN, a phenomenon that appears to reflect reduced cDC2 survival in the absence of Irf4 (16). Importantly, cDC2 numbers in MLN also remained selectively reduced in Cre+ compared with Cre− Rag-1−/− mice following CD4+ T cell transfer. In contrast to lymphocyte-replete CD11c-Cre.Irf4fl/fl mice, in which CD103-expressing cDC2 were preferentially reduced in the MLN (16), both CD103+ and CD103− cDC2 were reduced in the MLN of Cre+ Rag-1−/− mice.
A major finding of the current study was that Cre+ mice displayed reduced colon pathology 4 wk after CD45RBhiCD4+ T cell transfer. These results contrast with previous findings in Clec4a-DTR and human Langerin-DTA mice demonstrating that cDC2 do not play a role in DSS-induced acute colitis (24, 25). Nevertheless, in contrast to the CD45RBhiCD4+ T cell transfer model used in the current study, pathology in acute DSS colitis is primarily driven by innate immune cells that enter the mucosa following DSS-induced epithelial destruction with little involvement of adaptive immune cells (26). Reduced pathology at 4 wk correlated with markedly reduced CD4+ T cell numbers in both the MLN and colon of Cre+ mice, indicating an early, nonredundant role for cDC2 in the generation of colitogenic CD4+ T cells. Such results extend previous studies demonstrating reduced CD4+ T cell priming in lymphocyte-replete CD11c-Cre.Irf4fl/fl mice (16, 43) and indicate cDC2 as a possible therapeutic cellular target in IBD. In this regard, it was interesting to note that Cre+ and Cre− mice had similar symptoms of wasting disease (weight loss) after CD45RBhiCD4+ T cell transfer, possibly reflecting a role for IRF4 expression by cDC in local, but not systemic, inflammation (44). Although the precise mechanisms underlying reduced CD4+ T cell priming in Cre+ mice remain to be determined, it likely reflects the lower numbers of cDC2 in Cre+ mice, together with reduced MHCII Ag-presenting capacity of remaining Irf4-deficient cDC2 (43).
Despite the observed differences between Cre+ and Cre− mice 4 wk after T cell transfer, colitis scores and colonic CD4+ T cell numbers were similar in Cre+ and Cre− mice by 7 wk whereas CD4+ T cell numbers in the MLN of Cre+ mice remained reduced. Thus, IRF4 expression by cDC plays a nonredundant role in driving CD4+ T cell responses in MLN during colitis development but is not required for the sustained development of T cell–dependent colitis or for the maintenance and expansion of colitogenic CD4+ T cells subsequent to their entry into the colon. Interestingly, Foxp3+ Treg numbers were lower in the colon of Cre+ compared with Cre− mice at 7 wk posttransfer, potentially contributing to the expansion of pathogenic T cells in the colon of these mice. As monocyte-derived cells have been implicated in promoting CD4+ T cell responses during colitis (45, 46), we speculate that such cells may support colitogenic CD4+ T cell expansion in the colon once colitis is established. Alternatively, although cDC1 are generally less able to prime CD4+ T cells than cDC2 (47), they may contribute to colitogenic T cell expansion in the colon. Finally, Ag-independent signals, such as microbiome-induced cytokines, may assist in the expansion of pathogenic CD4+ T cells subsequent to their entry into the colon (48, 49).
We, and others, have shown that cDC2 play a nonredundant role in differentiation and subsequent homeostasis of intestinal IL-17+ CD4+ T cells (16, 17, 50). In this study, we found that although the numbers of IL-17+, IFN-γ+, and IL-17+FNγ+CD4+ T cells were significantly reduced in the MLN of Cre+ mice 7 wk posttransfer, this reduction was most significant for IL-17+CD4+ T cells. Furthermore, Cre+ mice had a selective reduction in IL-17+CD4+ T cells in the colon 7 wk posttransfer. Thus, IRF4 expression in cDC, in addition to promoting colitogenic CD4+ T cell responses, appears to have a nonredundant role in promoting Th17 responses during T cell–dependent colitis. However, similar to our observations under steady-state conditions (16), IL-17+CD4+ T cells were not completely absent in colitic Cre+ mice. This may reflect a role of the residual cDC2 in Irf4-deficient mice, or of the infiltrating myeloid cells, which are a documented source of the Th17-promoting cytokines IL-1β, IL-6, and IL-23 during colitis (45, 49, 51, 52).
Finally, although cotransfer of CD25+CD45RBlo CD4+ T cells prevents CD45RBhi CD4+ T cell–driven colitis, the role of cDC subsets in this process remains unclear (39, 40). Our results demonstrate that IRF4 expression by cDC2 is not required for CD25+CD45RBloCD4+ T cell–mediated prevention of colitis. Whether additional APC subsets play nonredundant roles in this process or whether cDC2 have a redundant role controlling the in vivo function of regulatory CD25+CD45RBlo CD4+ T cells remains to be determined.
In summary, our findings indicate a role for IRF4 expression in cDC2 for the initial priming of colitogenic T cells in Th17 induction and in early disease onset in T cell–dependent colitis. Further, we show that IRF4 expression by cDC2 is not required for the in vivo function of CD25+CD45RBloCD4+ T cells to prevent colitis. Further studies assessing the contribution of additional APC subsets during T cell–dependent colitis should help unravel potential unique functions and pathways these cells play in colitis development.
We thank Dr. K. Kotarsky for technical support, A.-C. Selberg for animal care and genotyping, and Dr. A. Mowat for valuable feedback during manuscript preparation.
This work was supported by a Sapere Aude III senior researcher grant from the Danish Research Council, the Swedish Medical Research Council (2017-02072), Cancerfonden (16 0370 and 18 0598), the Kocks, Österlund, Prof. Nanna Svartz, and Richard and Ruth Julins Foundations, the IngaBritt and Arne Lundbergs Foundations, the Royal Physiographic Society, and a clinical grant from the Swedish national health service.
The online version of this article contains supplemental material.
Abbreviations used in this article:
classical dendritic cell
colon lamina propria
dextran sodium sulfate
diphtheria toxin A
diphtheria toxin receptor
inflammatory bowel disease
MHC class II
mesenteric lymph node
signal-regulatory protein α
regulatory T cell
X-C motif chemokine receptor 1.
The authors have no financial conflicts of interest.