|Home | About | Journals | Submit | Contact Us | Français|
Although several subsets of intestinal antigen presenting cells (APCs) have been described, to date is no systematic evaluation of their phenotypes, functions and geographical localization. Here we used 10-color flow cytometry, to define the major APC subsets in the small and large intestine LP. LP APCs could be subdivided into CD11c+CD11b−, CD11c+CD11b+, and CD11cdullCD11b+ subsets. CD11c+CD11b− cells were largely CD103+F4/80− DCs, while the CD11c+CD11b+ subset comprised of CD11c+CD11b+CD103+F4/80− DCs versus CD11c+CD11b+CD103-F4/80+ macrophage-like cells. The majority of CD11cdullCD11b+ cells were CD103-F4/80+ macrophages. Although macrophages were more efficient at inducing Foxp3+ Treg cells than DCs, at higher ratios of T:APC cells all DC subsets efficiently induced Foxp3+ Treg cells. In contrast, only CD11c+CD11b+CD103+ DCs efficiently induced TH-17 cells. Consistent with this, the geographical distribution of CD11c+CD11b+CD103+ DCs correlated with that of TH-17 cells, with duodenum>jejunum>ileum>colon. Conversely, CD11c+CD11b−CD103+ DCs, macrophages and Foxp3+ Treg cells were most abundant in the colon and scarce in the duodenum. Importantly however, the ability of DC and macrophage subsets to induce Foxp3+ Tregs versus TH17 cells was strikingly dependent on the source of the mouse strain. Thus, DCs from C57BL/6 mice from Charles River, (that have segmented filamentous bacteria, which induce robust levels of TH-17 cells in situ (1, 2)), were more efficient at inducing TH-17 cells, and less efficient at inducing FoxP3+ Treg cells than DCs from B6 mice from Jackson Laboratories. Thus the functional specializations of APC subsets in the intestine is dependent on the T:APC ratios, regional localization, and the source of the mouse strain.
Sensing of microbes by innate immune cells is critical for the establishment of effective adaptive immune responses and host defense (3). Antigen-presenting cells (APCs) including dendritic cells and macrophages are endowed with the ability to detect microbial components via several evolutionarily conserved receptor/signaling pathways including toll-like receptors (TLRs), Nod-like receptors (NLRs), and c-type lectin receptors (CLRs) among others (4-6). The integration of microbial-derived signals via these pathways and local microenvironmental cues ultimately dictates the balance between ensuing tolerogenic and proinflammatory responses (7, 8). Additionally, specific APC subsets, expressing various levels of cell surface markers such as CD11c, CD11b, F4/80, CD8α, and CD4 differentially respond to microbial encounter (9, 10).
DCs and macrophages are situated in various lymphoid and nonlymphoid tissues and under homeostatic conditions they encounter the majority of microbial antigens at mucosal surfaces (11). In particular, DCs and macrophages in the intestinal Peyer’s patches (PP) and LP (LP) are situated just beneath a single layer of epithelial cells that separate them from a vast commensal microflora. The major portals of entry for many microbes are the PP of the small intestine (12). Within the PP exist several subsets of DCs that differentially regulate immune responses dependent upon whether they are CD11b+, CD8α+ or CD11b−CD8α− (13). CD11b+ PP DCs can be stimulated to secrete IL-10 and induce the differentiation of Tr1-like cells and may promote IgA B cells responses (14). Alternatively, CD8α+ and CD11b−CD8α− PP DCs can produce IL-12 and generate Th1-like cells (15). In the LP, CD8α+ and CD11b+ DCs have been identified and the functions of these subsets have yet to be clearly defined. CD11b+ LP DCs appear to express tight junction proteins and can extend dendrites that interdigitate between neighboring intestinal epithelial cells (IECs) to sample luminal microbes (16). The extension of dendrites is controlled by expression of CX3CR1 by DCs and such CX3CR1-dependent processes may control intestinal bacterial clearance (17). Additionally, CD11b+ LP DCs can promote the induction of TH-17 differentiation either constitutively (18) or following stimulation by bacterial flagellin in a TLR5-dependent manner (19). Intestinal DCs have also been reported to specifically induce the gut homing molecules α4β7 and CCR9 on responding T and B cells (20-22) and FoxP3+ regulatory T cells (Treg) from naïve CD4+ T cells in vitro in a retinoic acid dependent fashion (23-25), and recent work from our lab has demonstrated a potent role for β-catenin signaling in programming such DCs to a tolerogenic state (26). The ability of LP DCs to induce FoxP3+ Treg conversion appears closely linked with the commensal microbiota as such conversion can be inhibited by bacterial CpG and TLR9-deficient mice have increases in FoxP3+ Tregs in the LP (27). The role of CD8α+ LP DCs are less clear but may also be to induce differentiation of regulatory-type T cells (28).
In addition to DCs a major macrophage network exists in the intestinal LP. LP macrophages are avidly phagocytic, hyporesponsive to various inflammatory stimuli, spontaneously secrete IL-10, and can efficiently promote FoxP3+ Treg conversion in vitro in the presence of exogenous TGF-β (18, 29-33). Steady-state LP macrophages may also dampen immune responses during homeostasis and intestinal inflammation or injury (18, 31, 34-36). For example, depletion of intestinal macrophages using clodronate liposomes led to exacerbated DSS-induced experimental colitis (37). Alternatively, certain macrophage subsets may contribute to the pathogenesis of intestinal inflammation in mice and humans (31, 38). Recently, TREM-1+ as well as CD14+ LP macrophages were shown to be associated with the pathogenesis of human IBD (39, 40). Thus, steady-state LP macrophages may play important roles in maintaining tolerance while inflammatory macrophages may contribute to the pathogenesis of intestinal inflammation (38).
Thus, unique APCs subsets in different regions of the intestine may contribute to maintaining the balance between tolerogenic and pro-inflammatory immune responses. Here we have defined major DC and macrophage subsets in the small and large intestine LP and demonstrate that the functional specializations of LP DC and macrophage subsets in the intestine is dependent on several parameters, including the T:APC ratio, regional localization, and the source of the mouse strain. Thus, F4/80+ macrophage subsets are the major IL-10 producers in the LP, and concomitantly express TGF-β, retinoic acid generating enzymes and promote differentiation of inducible FoxP3+ Treg. Furthermore, all DC subsets could induce FoxP3+ Treg cells at particular T:APC ratios. Alternatively, CD11b+CD103+ LP DCs in C57BL/6 mice from Charles River, uniquely express IL-6 and TGF-β mRNA as well as retinoic acid-generating enzymes and efficiently induce the differentiation of TH-17 cells. CD11b+CD103+ LP DCs are enriched in the duodenum and rare in the colon during the steady state and accumulate in the colonic LP along with TH-17 cells during intestinal inflammation. Strikingly, CD11c+CD11b+ LP DCs from C57BL/6 mice purchased from Charles River, (that have segmented filamentous bacteria, which induce robust levels of TH17 cells in situ (1, 2)), were more efficient at inducing TH-17 cells, and less efficient at inducing FoxP3+ Treg cells than their counterparts from B6 mice from Jackson Laboratories. Therefore, these results demonstrate that the functions of intestinal APC subsets is dependent on several parameters, including the ratio of T:APC, the source of the mouse strain, and regional localization.
C57BL/6 male mice 6-12 weeks of age obtained from Charles River Laboratory were used in all studies. In specific experiments mice obtained from Jackson Laboratory were also used. OT-II (Rag1/2+/+), originally obtained from Jackson Laboratories, were bred on-site. IL-10–IRES-EGFP reporter mice (Vert-X mice; (41)) by insertion of a floxed neomycin–IRES-EGFP cassette between the endogenous stop site and the polyadenosine site of Il10. The neomycin resistance marker was excised by breeding the mice with Zp3-Cre mice, and successful Cre-mediated deletion was confirmed by Southern blot analysis (41). Il10−/− mice (B6.129P2-Il10tm1Cgn/J) were purchased from Jackson Laboratory. Mice were maintained under specific pathogen-free conditions in the Emory Vaccine Center or Emory Whitehead Biomedical Research Building vivariums. All animal protocols were reviewed and approved by the Institute Animal Care and Use Committee of Emory University.
Small/Large Intestine: Briefly, intestines were removed and carefully cleaned of their mesentery, Peyer’s patches were excised (small intestine), and intestines were opened longitudinally and washed of fecal contents. Intestines were then cut into 0.5-cm pieces, transferred into 50-ml conical tubes, and shaken at 250 rpm for 20 min at 37°C in HBSS medium (Life Technologies) supplemented with 5% FBS (CellGro), containing 2mM EDTA. This process was repeated two additional times. The cell suspensions were passed through a strainer and remaining intestinal tissue was washed and then minced, transferred to a 50-ml conical tube, and shaken for 20 min at 37°C in HBSS+5% FBS containing Type VIII collagenase at 1.5 mg/ml (Sigma). The cell suspension was collected and passed through a strainer, and pelleted by centrifugation at 1200 rpm. CD11b+ or CD11c+ cells were enriched by positive selection with CD11b or CD11c microbeads (Miltenyi Biotec), respectively. For FACS-sorting experiments intestinal CD11b/c-enriched cells were stained with PE-conjugated anti-CD103 (clone M290; Pharmingen), PerCP-conjugated anti-CD45 (clone 30-F11; Pharmingen), PE-Cy7 conjugated anti-F4/80 (clone BM8), APC conjugated anti-CD11c (clone HL3), Alexa Fluor 700 conjugated anti-IAb (clone M5/114), and Pacific Blue conjugated anti-CD11b (clone M1/70) all from eBioscience unless otherwise noted. Stained cells were sorted to purify indicated populations on a FACS Aria at The Emory Vaccine Center Flow Cytometry Core Facility.
Isolated splenocytes or small intestinal LP cells were resuspended in PBS containing 5% FBS. After incubation for 15 min at 4°C with the blocking 2.4G2 anti-Fc©RIII/I, the cells were stained at 4°C for 30 min with labeled antibodies. Samples were then washed two times in PBS containing 5% FBS. The samples were immediately analyzed at this point, or they were fixed in PBS containing 2% paraformaldehyde and stored at 4°C. Antibodies used for analysis were from eBioscience unless otherwise noted: FITC-labeled anti mouse B220, PE-labeled anti mouse CD103 (BD Pharmingen), PE-Texas Red-labeled anti mouse CD8α (Caltag), PerCP-labeled anti mouse CD45 (BD Pharmingen), PE-Cy7-labeled anti mouse F4/80, APC-labeled anti mouse CD11c, Alexa Fluor 700-labeled anti mouse IAb, APC-Alexa Fluor 750-labeled anti mouse Gr-1, Pacific Blue-labeled anti mouse CD11b, and Alexa Fluor 430 (live/dead stain; Invitrogen). Intracellular IL-17 or Foxp3 staining was performed using PE-labeled anti-mouse IL-17 or Foxp3 (eBioscience). ALDH activity was determined using the ALDEFLOUR staining kit (STEMCELL Technologies Inc.) according to the manufacturer’s protocol. In brief, cells were suspended at 106 cells/ml in ALDEFLUOR assay buffer containing activated ALDEFLUOR substrate (final concentration 1.5 μM) with or without the ALDH inhibitor DEAB (final concentration 15 μM) and incubated at 37°C for 30 min. Flow cytometric analysis was performed on a Becton Dickinson FACSCaliber or LSR II flow cytometer at the Emory Vaccine Center.
To analyze cell morphology, sorted LP antigen presenting cells were spun onto glass slides using a CytoSpin centrifuge, allowed to air dry for 1 hour, fixed in 100% methanol and then stained using Wright-Giemsa stain.
Total RNA was isolated from FACS-sorted LP antigen presenting cells or enriched, splenic DCs (CD11c+ cells) or macrophages (CD11b+ cells) using the Qiagen RNeasy Mini Kit, according to the manufacturer’s protocol with on-column DNase digestion using the RNase-Free DNase set. cDNA was generated using the Superscript First-Strand Synthesis System for RT-PCR and oligo dT primers (Invitrogen), according to the manufacturer’s protocol. cDNA was used as a template for quantitative real-time PCR using SYBR Green Master Mix (BioRad). PCR and analysis was performed using a MyiQ ICycler (BioRad). Gene expression was calculated relative to Gapdh. Stimulation of lymphocytes. In vitro: FACS-sorted lamina propria APCs (1×105 unless otherwise noted) were co-cultured with naïve CFSE labeled or unlabeled CD4+CD62L+ OT-II CD4+ T cells (1×105) and OVA (ISQVHAAHAEINEAGR; 10μg/ml) in 200 μl RPMI complete medium in 96-well round bottom plates. For CFSE experiments cells were harvested after 90 hours. As a positive control for CFSE culture stimulation αCD3/28 coated beads (Dynal) were used at a ratio of 1 bead per cell. For restimulation, cells were harvested after 90 h of primary culture, before being restimulated with PMA (50ng/ml) and ionomycin (500ng/ml) for 6 h (intracellular cytokine detection) in the presence of GolgiPlug. Small/large intestine LP lymphocytes were stimulated with PMA and ionomycin for 6 h in the presence of GolgiPlug. For FoxP3+ Treg induction assays 1ng/ml rhTGF-β (Peprotech) was added to cultures.
For intracellular cytokine analysis, cells were stimulated with PMA and ionomycin 6 h in the presence of GolgiPlug (PharMingen) and subsequently incubated with blocking 2.4G2 anti-FcγRIII/II (PharMingen) and stained with PerCP-conjugated anti-CD4 and APC-labeled anti mouse TCR-b (BD Pharmingen). Cells were permeabilized by using Cytofix/Cytoperm and stained by using fluorochrome-conjugated antibodies according to the manufacturer’s protocol. IL-10 ELISAs (eBioscience) were performed according to the manufacturer’s protocol.
Naïve OT-II CD4+ T cells (1×105) were cultured with CD11c-LP macrophages (1×105) in the presence of 1ng/ml rhTGF-β for 90 hours, then harvested, washed, and added at a ratio of 1:1 to cultures of CFSE labeled naïve polyclonal CD4+CD45.1+ splenic T cells stimulated with CD11c+ splenic DCs and soluble anti-CD3ε (2μg/ml; BD PharMingen). Cultures were incubated for 48 hours and analyzed for CFSE dilution after gating on CD4+CD45.1+ cells.
Chronic colitis was induced by multiple-cycle administration of dextran sodium sulfate (DSS) drinking water (42, 43). Mice received either regular drinking water (control) or 3% (wt/vol) DSS drinking water (molecular weight: 40,000 – 50,000) (MP Biomedicals, Irvine, CA) on days 1–5, 8–12, 15–19, and 22–26 and mice were sacrificed on day 29. Histology was used to evaluate inflammation and colonic shortening was analyzed. This model of colitis is referred to as “chronic” since it involved 4 cycles of DSS treatment over the course of nearly one month and is distinct from the commonly used “acute” model of DSS-induced colitis where mice are typically treated for 5-7 days.
Statistical significance of differences in means ± SEM values from different groups were calculated using the unpaired t test (two-tailed) using GraphPad Prism software. P values less than 0.05 were considered statistically significant.
A clear definition of intestinal LP DCs and macrophages has been plagued by a lack of specific markers that clearly delineate each population and subsets existing within. Therefore we performed 10-color flow cytometry on isolated small intestinal LP cells combining the use of 9 markers (CD8α, CD11b, CD11c, CD45, CD103, IAb, F4/80, Ly6C/G, B220) and a vital dye. After gating on live, CD45+IAb+ cells three major cell populations expressing either CD11b and/or CD11c were identified (Fig. 1a). Two populations of CD11c+ cells were distinguished by differential expression of CD11b with the CD11b+CD11c+ cells (4.64% ± 9.2%) approximately 2.5-fold more abundant than the CD11bdull/−CD11c+ subset (1.67% ± 0.22%). This difference was statistically significant (p<0.05). Notably, CD11bdull/−CD11c+ cells were a discrete subset while CD11b+CD11c+ cells displayed variable levels of CD11c that formed a continuum. The gate for CD11b+CD11c+ cells, with regards to CD11c intensity, was identified based upon the same level of CD11c expression as the clearly defined CD11bdull/−CD11c+ subset gate. The other abundant population of cells identified was CD11b+CD11cdull/− cells (22.75% ± 1.4%), which were 5.4-fold more abundant than the CD11b+CD11c+ subset and 13.8-fold more abundant than the CD11bdull/− CD11c+ subset. These differences were statistically significant (p<0.05). All of the above mentioned populations and subsets did not express B220 or Ly6C/G (data not shown). Additionally, cells that did not stain for either CD11b or CD11c included B cells as identified by B220 or CD19 staining and other undefined cells (data not shown).
To further characterize each of the three major populations expressing either CD11b and/or CD11c they were analyzed for expression of CD103 and F4/80. The large majority (78.9% ± 2.2%) of CD11bdull/−CD11c+ LP cells expressed high levels of CD103 and were negative for F4/80 (Fig. 1b). Interestingly, CD11b+CD11c+ cells included approximately equal frequencies of CD103+F4/80− and CD103−F4/80+ subsets (39% ± 2.5% and 42.7% ± 1.3%, respectively). The most abundant CD11b+CD11cdull/− population was almost entirely CD103− F4/80+ (79.2% ± 2%). In order to visualize the cell morphology of each of these populations Geimsa staining was performed on FACS sorted cells. CD11bdull/−CD11c+ cells, which were largely CD103+F4/80−, displayed classical DC morphology with numerous dendrites extending from the cell surface and few detectable phagocytic vacuoles (Fig. 1c). Similarly, among the CD11b+CD11c+ population the CD103+F4/80− subset was characterized by classical DC morphology while the CD103−F4/80+ subset displayed typical macrophage morphology characterized by abundant phagocytic vacuoles and a relative lack of dendrites. As previously reported by our group CD11b+CD11cdull/− cells also displayed typical macrophage morphology (18). These results collectively demonstrate that IAb+ DCs in the small intestine LP are largely defined by expression of CD103 with variable expression of CD11b (hereafter referred to as CD11b+CD103+ and CD11b−CD103+ LP DCs for simplicity), while IAb+ macrophages are defined by expression of CD11b and F4/80 with variable expression of CD11c (hereafter referred to as CD11c−F4/80+ and CD11c+F4/80+ LP macrophages). Additionally, we calculated absolute cell numbers for each of these populations. As shown in Fig. 1d both LP DC subsets as well as CD11c+F4/80+ LP macrophages were present in similar numbers, while CD11c−F4/80+ LP macrophages were approximately 10-fold more abundant (p < 0.05). Of note, our strategy of pre-gating on CD45+IAb+ cells for these analyses intentionally did not include what may be considered SSChiIAb− macrophages (44) or eosinophils (19).
Since LP DCs and macrophages have been reported to generate Foxp3+ Treg cells via the production of retinoic acid (RA) and IL-10 in the presence of TGF-β1, we analyzed the expression of the retinoic acid generating enzymes retinaldehyde dehydrogenase 1 and 2 (aldh1a1, aldh1a2), il10 and tgfb1 by quantitative real-time PCR. For comparative purposes, splenic DCs (CD11c+ cells) and macrophages (CD11b+ cells) were included in the analyses. LP DCs and macrophages further displayed a clear difference in expression of retinoic acid generating enzymes. CD11c−F4/80+ LP macrophages expressed high levels of aldh1a1, CD11c+F4/80+ LP macrophages and CD11b+CD103+ LP DCs expressed intermediate levels, and CD11b−CD103+ LP DCs expressed very low levels that were comparable to splenic DCs and macrophages (Fig. 2a). Alternatively, CD11b+CD103+ and CD11b−CD103+ LP DCs expressed high levels of aldh1a2 mRNA and both LP macrophage populations displayed significantly lower levels of this RA-generating enzyme (p < 0.05). Thus, while both LP DCs (24) and macrophages (18) are capable of generating retinoic acid and inducing Treg in a retinoic acid-dependant fashion, they may differentially employ aldh1a1 and aldh1a2 for this purpose. To assess whether such differences correlated with differences in aldehyde dehydrogenase (ALDH) activity, CD11b+CD103+ and CD11b−CD103+ LP DCs, and CD11c−F4/80+ and CD11c+F4/80+ LP macrophages were incubated with the fluorescent ALDH substrate ALDEFLUOR, and ALDEFLUOR expression was assessed by flow cytometry. CD11b+CD103+ and CD11b−CD103+ LP DCs exhibited significantly (p < 0.05) higher ALDEFLUOR expression than did CD11c− F4/80+ and CD11c+F4/80+ LP macrophages, in the small intestine, as measured by mean fluorescence intensity (MFI) for ALDEFLUOR expression. The change in MFI (ΔMFI) was calculated by subtracting the MFI for DEAB treated samples from the MFI for the corresponding non-DEAB (diethylaminobenzaldehyde) treated samples (Fig. 2b). In all conditions ALDH activity was blocked with the ALDH inhibitor DEAB. Interestingly, lower ALDH activity was measured in the large intestine, especially for CD11b+CD103+ and CD11b−CD103+ LP DC subsets. Although ALDH expression was significantly lower among LP macrophage subsets when compared to LP DC subsets in the small intestine, all subsets possessed some, specific ALDH expression. Given that LP macrophages 1) express both aldh1a1 and aldh1a2 (albeit at significantly lower levels than LP DCs) mRNA, 2) are noticeably positive using the ALDEFLUOR assay (albeit at significantly lower levels than LP DCs) and 3) that addition of LE540 to block RA signaling reduces the ability of LP macrophages to induce Foxp3+ Tregs in vitro (similar to LP DCs; (18)) strongly suggest that these cells do indeed possess the ability to generate RA.
While all populations of LP DCs and macrophages expressed tgfb1, both macrophage subsets expressed significantly more mRNA than did either DC subset (Fig. 2c). We next probed the expression of il6 mRNA among LP DCs and macrophages since it has been shown to inhibit Foxp3+ Treg differentiation and can combine with TGF-β to generate IL-17 producing T cells. Interestingly, CD11b+CD103+ LP DCs expressed significantly higher levels of il6 mRNA that the other APC subsets (Fig. 2c), however this cytokine was undetectable by ELISA and intracellular cytokine staining suggesting low-level protein expression. mRNAs for IL-12p35, IL-12p40 and IL-23p19 were not detectable in any of the groups that we analyzed (data not shown). Interestingly, CD11c−F4/80+ and CD11c+F4/80+ LP macrophages both expressed significantly higher levels of il10 mRNA than CD11b+CD103+ and CD11b−CD103+ LP DCs (p < 0.05). The statistically (p < 0.05) different expression of il10 was confirmed at the protein level using ELISA (Fig. 2d) and IL-10–IRES-EGFP reporter mice (Vert-X mice; Fig. 2e). IL-10-GFP was clearly detected in CD11c−F4/80+ LP macrophages and to a lesser extent in CD11c+F4/80+ LP macrophages, but was undetectable in CD11b+CD103+ and CD11b−CD103+ LP DC (Fig. 2e). Collectively these data provide evidence that LP macrophages are unique in their expression of high levels of il10, tgfb1, and aldh1a1, while both LP DC populations express high levels of aldh1a2 and CD11b+CD103+ LP DCs specifically express high levels of il6. These data for each LP DC and macrophage subset are summarized in Table I.
In order to address the differential role of LP DC and macrophage populations in inducing the differentiation of TH-17 or Foxp3-expressing Treg cells in vitro we setup an antigen presentation assay as previously described (18) by co-culturing freshly isolated LP DCs or macrophages with naive OT-II CD4+CD62L+ T cells and ovalbumin (OVA) in the absence (TH-17 assay) or presence (Treg assay) of exogenous TGF-β1. Following 90 hours of co-culture, both CD11b− CD103+ and CD11b+CD103+ LP DCs as well as CD11c−F4/80+ and CD11c+F4/80+ LP macrophages all induced similar levels of CFSE dilution that were of the same magnitude as when aCD3/28 beads were used to stimulation OT-II T cells (Fig. 3a). Interestingly, although LP macrophages subsets induced similar levels of CFSE dilution as LP DC subsets, LP macrophages induced less T cell accumulation in culture (approximately 3-fold less than LP DC subsets), and less T cell clustering and IL-2 production (data not shown).
We next observed that among the major LP DC and macrophage populations, CD11b+CD103+ LP DCs preferentially induced significant IL-17-expressing T cell differentiation (6.1%; p < 0.05) as detected by intracellular cytokine staining (Fig. 3b). These data are summarized in Fig. 3b (lower panel) which displays means from 3 independent experiments +/− SEM. Importantly, this ability of CD11b+CD103+ LP DCs to induce IL-17-expressing T cells was highly dependant on IL-6 production (Fig. 3c; p < 0.05) since addition of neutralizing antibody to IL-6 (αIL-6), but not isotype control antibody (Rat IgG), significantly reduced IL-17 expressing T cells.
Using the same antigen presentation assay described above, but in the presence of exogenous TGF-β1, various LP DCs and macrophages populations were probed for their ability to promote the induction of Foxp3 in differentiating, FACS-sorted naïve Foxp3−CD4+CD62L+ cells. As shown in Fig. 4a CD11c+F4/80+ and CD11c−F4/80+ LP macrophages, but not CD11b+CD103+ or CD11b−CD103+ LP DCs can induce Foxp3-expressing Treg cell differentiation at an T:APC cell ratio of 1:1. Since it has been reported that CD103+ and CD103− LP DCs can induce Foxp3-expressing Treg cell differentiation at higher T:APC ratios (24), we performed a titration of LP DCs or macrophages while keeping the OT-II T cells number constant to increase the T:APC ratio. Changes in the T:APC ratio from 1:1 up to 40:1 did not dramatically alter the high frequency of Foxp3+ Treg cells induced by either CD11c+F4/80+ or CD11c−F4/80+ LP macrophages. Alternatively, LP DCs which were poor at generating Foxp3+ Treg cells at a ratio of 1:1, were capable of inducing Foxp3 in responding T cells at T:APC ratios of 5:1 and higher. These reproducible data confirm previous observations that LP DCs and macrophages differentially Foxp3 expressing Treg cells and highlight the dynamics of T:APC ratios in the induction of Foxp3+ Treg by LP DCs in vitro. As has been reported for LP DCs (23, 24), Foxp3+ Treg cells induced by CD11c−F4/80+ LP macrophages were capable of suppressing naïve CD4+ responder cell proliferation as analyzed by CFSE-dilution (Fig. 4b). The majority of naive CD4+CD45.1+ responder cells proliferated 4-5 times in the absence of Foxp3+ Treg cells, while the addition of CD11c−F4/80+ LP macrophage induced putative Treg cells (approximately 55% FoxP3+) prevented most responder cells from dividing at all (66.6 % suppression in Expt 1 and 71% suppression in Expt 2; Fig. 4c).
Recent work has demonstrated striking differences in microbiota and intestinal TH-17 cells in the genetically identical strains of mice from different vendors (1). Thus, C57BL/6 mice from Charles River have segmented filamentous bacteria that induce high levels of intestinal TH-17 cells in contrast to the same strain of mice from Jackson Laboratories (2). We thus determined whether the intestinal DC subsets in C57BL/6 mice from the different vendors had intrinsic differences in their capacity to induce TH-17 versus Foxp3+ Tregs. Both CD11c+CD11b− and CD11c+CD11b+ LP cells isolated from Jackson Laboratories mice were more efficient at inducing Foxp3+ Tregs (Fig. 5a) and the preferential induction of Foxp3+ Tregs by CD11c+CD11b+ LP cells was significant (p < 0.05). Alternatively, both CD11c+CD11b− and CD11c+CD11b+ LP cells from Charles River mice were significantly (p < 0.05) more efficient at inducing TH-17 cells than those from Jackson Laboratories mice (Fig. 5b), while the Charles River CD11c+CD11b+ LP cells were more efficient than Charles River CD11c+CD11b− LP cells at inducing TH-17 cells, confirming previous results (Fig. 3b). In light of this data using Charles River CD11c+CD11b+ LP cells, which contain a mixture of CD103+ LP DCs and F4/80+ macrophages, we further confirmed the preferential and significant (p < 0.05) TH-17 inducing capacity of the more specifically defined CD11c+CD11b+CD103+ LP DC subset that was sorted from Charles River mice (Fig. 5c). Importantly, differences in the capacity of Charles River and Jackson Laboratories LP cells to induce TH-17 versus Foxp3+ Tregs was not due to differences in the abundance of DC and/or macrophages subsets, since phenotypic characterization of these subsets revealed no significant changes (Fig. 5d). These data reveal that the source of the mouse strain is a critical determinant of intestinal APC function, most likely due to differences in the microbiota.
To further probe the relationship between specific LP DC populations and TH-17 and Treg cells in situ we performed detailed analysis of the duodenum, jejunum, ileum (defined as equal thirds by length), and colon (not including cecum). After gating on live, CD45+IAb+ cells the frequencies (Fig. 6a) and absolute cell numbers (Fig. 6b) of CD11b+CD103+ and CD11b−CD103+ LP DCs and CD11c+F4/80+ and CD11c−F4/80+ LP macrophages were determined. The frequencies of CD11b−CD11c+ LP DCs (R1) were significantly (p < 0.05) enriched in the colon, while the frequencies of CD11b+CD11c+ cells (R2) were significantly (p < 0.05) enriched in the duodenum, as compared to other regions. Further, the frequencies of CD11b+CD11c− cells (R3) were least abundant in the duodenum (p < 0.05) and showed similar accumulation in the jejunum, ileum, and colon (Fig. 6a). Among CD11b+CD11c+ LP cells, the duodenum also displayed the highest frequencies of CD103+F4/80− LP DCs at 59.8% ± 2.5% (p < 0.05), while the jejunum and ileum had similar frequencies of these cells (33.5% % ± 2.6% and 33.8% ± 3.3%, respectively). The colon had the lowest frequency of CD103+F4/80− LP DCs among CD11b+CD11c+ LP cells at 23% ± 2.9% (p < 0.05). The absolute numbers of these cells are shown in Fig. 6b and statistically significant differences are indicated. It is important to note that differences in trends between frequency and absolute numbers of cells are the result of differences in cellularity between the intestinal regions with the total cellularity as follows: duodenum>jejunum>ileum>colon. Alternatively, the frequency of CD11c+F4/80+ LP macrophages within the CD11b+CD11c+ population increased from 26.1% ± 2.5% % in the duodenum to 37.3% ± 5.6% in the jejunum and 38.5% ± 3.9% in the ileum and nearly 46.8% ± 4.8% in the colon. The duodenum contained the lowest frequencies of CD11c+F4/80+ LP macrophages within the CD11b+CD11c+ population (p < 0.05) compared to all other regions, while the colon contained the highest (p < 0.05). This increase in frequency was not directly consistent with changes in absolute cell number due to decreased total cellularity in equal length pieces of intestine obtained from descending intestinal regions.
Another important means of comparing relative changes in the two major LP DC subsets along the intestinal tract is to pre-gate on live, CD45+IAb+CD11c+CD103+ DCs and to then analyze expression of CD11b to discern the relative abundance of CD11b+CD103+ and CD11b− CD103+ LP DC subsets. In doing so a clear gradient from high levels of CD11b+CD103+ LP DCs in the duodenum (69.7% ± 2.4%) to intermediate levels in the jejunum (54.9% ± 3.1%) and ileum (47.3% ± 3.2%) to low levels in the colon (26.1% ± 1.3%) was observed (Fig. 6c), consistent with what was observed with the previous gating strategy (Fig. 6a). These data are summarized in Fig. 6c, which displays means from all experiments ± SEM. All regions are significantly different from each other (p < 0.05) except for the jejunum and ileum.
These data demonstrate that when descending the intestinal tract the presence of CD11b+CD103+ LP DCs decreases while CD11b−CD103+ DCs increase in abundance. The region specific localization and T cells differentiation data for each LP DC and macrophage subset are summarized in Table II.
We next evaluated the presence of IL-17 producing CD4+ T cells and Foxp3-expressing Treg in the 3 regions of the small intestine and large intestine. Remarkably, the presence of IL-17 producing CD4+ T cells was highly enriched in the upper small intestine with 25.5% ± 1.8% of CD4+ T cells isolated from the duodenum expressing IL-17 upon ex vivo restimulation with PMA and ionomycin. CD4+ T cells isolated from the jejunum and ileum expressed 16.6% ± 0.9% and 12.3% ± 0.8% IL-17 producing CD4+ T cells, respectively, and colonic CD4+ T cells displayed the lowest frequency of IL-17+ T cells at 8.3% ± 1.1% (Fig. 7a). All regions were significantly different from each other (p < 0.05) both in frequency and absolute cell number thus highlighting a clear and statistically significant gradient for IL-17 producing CD4+ T cell along the intestinal tract (Fig. 7b). Of note, the majority of IL-17 expressing CD4+ T cells did not co-express IFN-γ and the majority of IFN-γ expressing CD4+ T cells were detected in the colon.
In contrast to the decreased representation of TH-17 cells in descending regions of the intestinal tract, an opposite trend was observed for Foxp3-expressing Treg in the LP. The presence of Foxp3-expressing Treg cells was lowest in the upper small intestine at 15.7% ± 0.9% CD4+ T cells. CD4+ T cells isolated from the jejunum and ileum expressed 17.3% ± 0.9% and 16.3% ± 1.8% CD4+Foxp3+ T cells, respectively, and colonic CD4+ T cells displayed the highest frequency of CD4+Foxp3+ T cells at 33.7% ± 2% that was approximately 2-fold higher than corresponding frequencies in small intestinal regions (Fig. 7a). The enrichment of CD4+Foxp3+ T cells in the colon was statistically significant (p < 0.05) compared to all other regions (which were not statistically different from one another) (Fig. 7b). The opposite trends of TH-17 and Foxp3+ Treg cell representation along the length of the intestine, both in frequency and absolute cell number (Fig. 7b), suggest that the milieu of the upper small intestine favors TH-17 differentiation and that the large intestine microenvironment is conducive to Foxp3+ Treg differentiation. These gradients for TH-17 and Foxp3+ Treg appear to be lost during dextran sodium sulfate (DSS)-induced intestinal inflammation with no statistically significant difference in either TH-17 or Foxp3+ Treg cells observed in the duodenum, jejunum, and ileum (Fig. 7c). Of note, the frequencies of IL-17 producing CD4+ T cells were higher in the inflamed small intestine compared with normal small intestine.
During the analysis of intestinal region specific presence of DC and macrophage populations and TH-17 and Foxp3+ Treg cells several strong correlations were noted. In particular, the trend of TH-17 cells in specific regions of the intestine correlated with the abundance of the CD11b+CD103+ LP DC subset (r=0.92; Fig. 7d). Alternatively, the abundance of Foxp3+ Treg cells correlated with the abundance of the CD11b−CD11c+ LP DC subset (r=0.76) and also with CD11c−F4/80+ LP macrophages (r=0.67) and to a lesser extent with CD11c−F4/80+ LP macrophages (r=0.63; Fig. 6c).
Having noted a paucity of IL-17 producing CD4+ T cells in the colon, we next examined if these cells were increased during colonic inflammation and whether any such increases correlated with accumulation of CD11b+CD11c+ LP DCs. Thus, we induced chronic colitis in mice using 4 (1-week each) cycles of dextran-sodium sulfate (DSS) administration in the drinking water. Colonic shortening (data not shown) as well as prominent mononuclear cell infiltration (Fig. 8a) were observed and this treatment led to the development of chronic colitis associated with TH1 and TH-17 responses (42, 43). When LP cells were isolated and restimulated, CD4+ T cells from colitic DSS-treated mice expressed high levels of IL-17 (29%) compared to control mice (7.6%). Data from all mice analyzed were averaged and summarized in Fig. 8b (right panel) and statistically significant differences are indicated. Unlike the pattern observed in control mice, the majority of the IL-17 producing cells in the colons of the DSS-treated group of mice co-expressed IFN-γ (Fig. 8b). Associated with the approximate 5-fold increase in IL-17 producing CD4+ T cells in the LP of DSS treated mice (Fig. 8b), an specific and statistically significant (p < 0.05) accumulation in the frequency of CD45+IAb+ cells, CD11b+CD11c+ cells (Fig. 8c) and CD11b+CD103+ LP DCs (Fig. 8d) was noted in comparison to cells isolated from control mice. The increased frequency of CD11b+CD103+ LP DCs in DSS-treated mice compared to control mice was especially reflected in the absolute number of cells (Fig. 8e; (p < 0.05) since the increased cellularity in the inflamed colon led to increases in total cell counts among all LP DC and macrophage subsets analyzed (p < 0.05). Collectively, these data suggest that the abundance of CD11b+CD103+ LP DCs positively influences the differentiation of TH-17 cells both in the upper small intestine of normal mice and in the large intestine during colitis, however it deserves mention that CD11b+CD103+ LP DCs in the inflamed intestine did not correlate as strongly with TH-17 cells as they did in healthy intestine.
Here we have shown that the functional specializations of DC and macrophage subsets in the intestine is a complex function of several variables, including the regional localization of the subsets, the ratio of T:APCs and the source of the mouse strains. In order to systematically characterize the DC and macrophage subsets in the small and large intestine LP, we established a 10-color FACS panel to phenotypically identify 5 major subsets of APCs: CD11c+CD11b− CD103+F4/80− and CD11c+CD11b+CD103+F4/80− LP DCs, CD11c+CD11b+CD103−F4/80+ and CD11c−CD11b+CD103−F4/80+ LP macrophages, and CD11c−CD11b− cells (that may include B cells and other non-DC, non-macrophage APCs), as well as several rare subsets including CD11c+CD11b+CD103−F4/80−, CD11c+CD11b−CD103−F4/80− and CD11c−CD11b+CD103−F4/80− APCs.
Intestinal F4/80+ macrophage subsets, regardless of CD11c expression, were the major IL-10 producers in the LP and promoted the differentiation of inducible FoxP3+ Treg cells, but not TH-17 cells. Furthermore, all DC subsets could promote Foxp3+ Treg cells, albeit less efficiently than the macrophages, and particularly at the higher T:APC ratios. The CD11b+CD103+ LP DCs from C57BL/6 mice from Charles River efficiently induced the differentiation of TH-17 cells. However, such DCs from C57BL/6 mice from Jackson Laboratories were not efficient inducers of TH-17 cells, and in fact all DC subsets from this strain were much more potent inducers of Foxp3+ Treg cells.
The CD103+ LP DC subsets displayed striking region-specific localization along the intestine with the presence of CD11b+CD103+ LP DCs strongly correlating with the abundance of TH-17 cells in the healthy, upper small intestine and the inflamed large intestine. Our results emphasize the complexities of the intestinal LP DC and macrophage network in terms of phenotypic characterization, regional localization and unique functional properties and suggest that these cells may be targets for modulating Treg and TH-17 responses in the intestine during health and disease.
Recently, several studies have analyzed LP DC and macrophages based on the expression of cell surface antigens that are well appreciated to demarcate each cell population in other tissues (19, 20, 23, 24, 45-47). We and others have relied upon CD11c as the bona fide marker of DCs and CD11b expression (in the absence of CD11c) as a useful, although not specific, marker for LP macrophages. Here it is demonstrated that CD11b and CD11c each have limitations for identifying CD45+IAb+ LP DCs and macrophages when used in the absence of CD103 and F4/80 co-staining. It has been argued that DCs and macrophages (of the LP and elsewhere) are not different cell types or lineages but rather a continuum of cells of the mononuclear phagocyte system (48). While our data does not specifically address this issue, we operationally define LP DCs as cells that are CD45+IAb+CD11c+CD103+CD11b+F4/80− or CD45+IAb+CD11c+CD11b-F4/80−CD103+, protrude many dendrites, few phagocytic vacuoles, and induce robust CD4+ T cell proliferation in vitro. In contrast, we operationally define LP macrophages as CD45+IAb+CD11c−CD103−F4/80+ cells or CD45+IAb+CD11c+CD103−F4/80+ cells that protrude few dendrites, many phagocytic vacuoles, and induce CFSE dilution of responding CD4+ T cell proliferation in vitro that is equivalent to that induced by LP DCs, albeit with less T cell clustering and overall expansion. Importantly, several recent reports demonstrated that CX3CR1+ and CD103+ LP DCs are of distinct origin and perform separate roles, with the CD103+ subset performing classical DC functions (49-51).
The functional capacity of LP DCs and macrophages in modulating CD4+ T cell differentiation has gained significant attention recently. Several groups reported that LP (24, 27, 52) or mesenteric lymph node (MLN; (23, 25, 53, 54)) DCs express retinoic acid generating enzymes and can induce gut-homing molecules and the in vitro differentiation of Foxp3+ Treg via retinoic acid. Furthermore, we demonstrated a potent role for LP macrophages at inducing Foxp3+ Treg via the retinoic acid pathway (18). Here we reveal several previously unappreciated complexities in determining the functions of intestinal APC subsets. Our data clearly shows that the ability of intestinal DC and macrophage subsets to induce Foxp3+ Tregs versus TH-17 cells is dependent on several factors, including: (i) the ratio of T cells to APCs, (ii) the regional localization of in the intestine, and (iii) the source of the mouse strain. With regards to the latter, our results indicate that DC subsets from C57BL/6 mice from Jackson Laboratories are much more efficient at inducing Foxp3+ Tregs and less efficient at inducing TH-17 than those from the same strain from Charles River. This is consistent with recent work from our lab in which we had observed robust induction of Foxp3+ Tregs and weak induction of TH-17 by intestinal DCs obtained from Jackson Laboratories (26). Recently it has been established that specific bacterial species, especially segmented filamentous bacteria (SFB), may preferentially drive TH-17 cell differentiation in the intestinal LP (1, 2, 55) and that SFB are absent in Jackson mice. Therefore the relative absence of TH-17 cells in the LP of Jackson mice may be due to the lack of CD11b+CD103+ LP DCs stimulation by SFB or other TH-17 inducing factors. Thus the differences observed between LP DCs subsets in Charles River mice might be less striking in Jackson mice and potentially other strains of mice lacking TH-17 inducing bacteria/factors.
With regards to the T: APC ratio, when LP DCs were reduced in number relative to CD4+ T cells, such conditions permitted the differentiation of Foxp3+ Treg although not as efficiently as CD11c+ and CD11c− F4/80+ LP macrophages that induced Foxp3+ Treg at all ratios tested. The exact factor(s) that account for this ratio dependent induction of Foxp3+ Treg by LP DCs are not clear, but these data suggest that at ratios near 1:1 there are either DC-expressed cell surface molecules and/or cytokines that are capable to preventing Foxp3 induction in responding CD4+ T cells. IL-6 is one candidate cytokine secreted by CD11b+ CD103+ LP DCs that may inhibit Foxp3 induction (56) at higher concentrations and lose such an effect as it is diluted away with increasing T:DC ratios. However, we did not observe robust IL-6 mRNA expression by CD11b+ CD103+ LP DCs suggesting that other factors may play a role in this process as well. Another important difference between our culture conditions and those reported previously (27) is that those studies supplemented cultures with exogenous IL-2, which is a known inducer of Foxp3+ Treg (57) cells and an inhibitor of TH-17 differentiation (58). It remains to be determined how well each of these in vitro culture conditions mimic actual T:DC ratios and local concentrations of TGF-β and IL-2 in the intestinal milieu. Additionally, while all LP DCs and macrophage subsets can promote Foxp3+ Treg differentiation under certain conditions in vitro, which of these populations/subsets are capable of inducing Foxp3+ Treg cells in the LP in vivo remains unresolved. Beyond de novo induction of Foxp3 in responding T cells by LP APC subsets, it has recently been demonstrated that IL-10 producing myeloid cells in the intestinal LP are crucial for maintaining Foxp3 expression and suppressor function in Treg cells in the CD45RBhi model of colitis (59). Therefore, LP DCs and macrophages may play a dual role in induction and maintenance of Foxp3 in CD4+ T cells.
We have demonstrated here that CD11b+CD103+ LP DCs from C57BL/6 mice from Charles River expressed il6 and tgfb1 as well as retinoic acid-generating enzymes and efficiently induced the differentiation of TH-17 cells in vitro. These data are consistent with our original report (18) that CD11b+CD11c+ LP DCs preferentially induced TH-17 cells and further defines the specific role for the CD103+ LP DC subsets in this process. Without methods to specifically deplete CD11b+CD103+ LP DCs, however, the ability of this DC subset to induce TH-17 responses in vivo remains to be formally demonstrated. Additional signals that condition CD11b+CD103+ LP DCs to efficiently generate TH-17 cells are beginning to be defined and may involve sensing of bacterial components such as flagellin via TLR5. Uematsu and colleagues recently reported that CD11b+CD11c+ LP DCs expressed TLR5 and induced TH-17 and TH1 cells when stimulated with flagellin (19). Interestingly, we did not detect robust levels of tlr5 mRNA in CD11b+CD103+ LP DCs and our analysis of TLR5 deficient mice revealed no detectable differences in TH-17 cells in the intestinal lamina propria (data not shown). Additionally, it is now appreciated that constitutive TH-17 cell development in the LP (60) is independent of MyD88 and Trif pathways, yet highly dependent on the bacterial microbiota (46). Thus, it is possible that TLR5 detection of bacterial flagellin by CD11b+CD103+ LP DCs may play a role in exacerbating TH-17 responses during intestinal inflammation but not in the homeostasis of natural TH-17 cells resident in the normal intestine.
Although in vitro modeling of intestinal LP DCs and macrophage functions is important, we also complimented our in vitro observations with characterization of the regional distribution of these cells, and TH-17 and Foxp3+ Treg cells ex vivo. While the entire small and large intestines drain to the mesenteric lymph nodes, region-specific differences were noted in DCs, macrophages and CD4+ T cells in the LP. In particular, CD11b+CD103+ LP DCs followed a gradient where they were enriched in the duodenum and rare in the large intestine, a pattern that mimicked TH-17 cells. Alternatively, CD11b−CD103+ LP DCs as well as CD11c+ and CD11c− F4/80+ LP macrophages followed an opposite gradient and were enriched in the large intestine where Foxp3+ Treg cells were abundant. These data suggest that unique factors in the upper small intestine favor the differentiation and/or retention of TH-17 inducing LP DCs, while the large intestinal milieu preferentially supports the differentiation and/or retention of Foxp3+ Treg inducing LP DCs and macrophages. Since TH-17 cells may mediate anti-bacterial activities it is possible that these cells are enriched in the relatively low bacterial load of the upper small intestine in order to quickly screen for potentially pathogenic bacteria that emigrate from the stomach. The vast number of commensal bacteria in the large intestine may induce a immunoregulatory developmental program that conditions DCs and/or macrophages to generate Foxp3+ Treg cells that restrain immunoreactivity towards the microbiota. Another potential explanation for the region-specific differences observed is that some “LP” DCs and/or macrophages may reside in gut associated lymphoid tissue such as ILFs or lymphoid follicles. These structures are more abundant in the descending gastrointestinal tract (61) and could explain the abundance of CD11b− LP DCs found in these regions. Alternatively, some of these “LP” DCs and/or macrophages may reside within sub-mucosal regions (49).
The important issue of where steady-state LP T cells differentiate in vivo remains incompletely understood. If differentiation were to take place primarily in the MLN as a result of LP DCs draining, one would expect to find increased frequencies of TH-17 cells and Foxp3+ Treg in this locale, which is not observed (data not shown). Thus, an alternative possibility is that, in the steady state, TH-17 and Foxp3+ Treg cell differentiation does not efficiently take place in the MLN, but rather in the LP of the intestine following migration of naïve T cells to this site (62). Another possibility is that T cells may initiate differentiation in the MLN and up regulate homing receptors that will enhance migration to specific regions of the intestine. In support of this, CCR6 has been demonstrated to be an important homing receptor involved in recruitment and/or retention of TH-17 cells during inflammation (63) and in Peyer’s patches (64) in normal mice. CCR6 does not appear to significantly affect the presence of TH-17 cells in the normal small intestinal LP however, so the enrichment of TH-17 cells we have observed in the upper small intestine is not easily explainable by differential homing due to CCR6. It is possible that in the steady state TH-17 and Foxp3+ Treg cells differentiate in the LP under the influence of region specific DC and macrophage subsets presenting antigen and that during intestinal inflammation TH-17 and Foxp3+ Treg cells and other effector T cells undergo more complete differentiation in the MLN due to increased LP DC trafficking to the MLN (65-68).
When we induced chronic colitis in mice an interesting change in the composition of large intestinal LP DCs and TH-17 cells was noted. Not only did TH-17 cells significantly increase in the diseased tissue but also so did CD11b+CD103+ LP DCs. Thus, altered frequencies and functions of large intestine LP DCs during colitis may contribute to inflammation (69). Additionally, DSS-induced colitis altered the normal decreasing gradient of TH-17 cells along the length of the small intestine and promoted high levels of TH-17 cells in the duodenum, jejunum, and ileum. Thus, increases in TH-17 cells and/or CD11b+CD103+ LP DCs and loss of the typical gradient of these cells may be involved in intestinal inflammation. These data are intriguing in light of the recent description of increased IL-17 levels in the colonic mucosa of patients with Crohn’s disease and ulcerative colitis and in mouse models of colitis (70). Whether IL-17 plays a protective (71, 72) or pathogenic (73) role in intestinal inflammation remains a complex issue since it is secreted by several cell types and exhibits anti-microbial activity (74). IL-23 is a key cytokine of the IL-12 family that can be secreted by DCs (75) and induce IL-17 secretion by responding CD4+ T lymphocytes. IL-23 is now documented as playing a critical role in the pathogenesis of several autoimmune diseases (76) and il23r has recently been shown to be an inflammatory bowel disease gene in adult (77) and pediatric (78) Crohn’s disease patients. Thus, LP DCs may also be a source of IL-23 that drive TH-17 cells during intestinal inflammation.
In addition to the increased frequency and cell number of CD11b+CD103+ LP DCs in the inflamed colon, the other antigen presenting cell subsets were also increased numerically due to increased cellularity in inflamed tissue. Thus, it cannot be excluded that cells other than CD11b+CD103+ LP DCs contribute to TH-17 differentiation during intestinal inflammation. Additionally, intestinal macrophages are well appreciated to contribute to the pathogenesis of intestinal inflammation (39, 79, 80). Therefore, while intestinal macrophages appear to play an anti-inflammatory role in the steady state they can mediate pro-inflammatory functions during colitis. Whether such diverse functions are performed by different subsets of macrophages or rather the plasticity of individual cells is unclear, however recent data suggests that a unique subset of TLR2+CCR2+CX3CR1int Ly6ChiGr-1+ macrophages secrete TNF-α and promote intestinal inflammation (81). Another recent report suggested that E-cadherin+ DCs are inflammatory APCs that may contribute to the pathogenesis of T cell-mediated colitis (82). Future studies should help clarify whether these inflammatory E-cadherin+ cells are DCs or macrophages.
Overall, our data demonstrate that LP DCs and macrophages play a fundamentally important, yet distinct, roles in directing CD4+ T cell differentiation. Importantly, we show that the functional specializations of DC and macrophage subsets are dependent on the ratio of T:APCs, their regional localization, and the source of the mouse strain from which they were isolated. Further clarifying the developmental and functional aspects of these and other LP DC and macrophage subsets along the length of the gastrointestinal tract during health and disease will likely contribute to a better understanding of mucosal immune regulation, which may direct efforts aimed at improving mucosal vaccination regimens and therapeutic intervention for Crohn’s disease and ulcerative colitis.
We thank Youliang Wang, Sommer Dunham, and Aaron Rae for cell sorting, Brent Greene for H&E staining, and Ifor Williams for helpful discussions.
1This work was supported by NIH grants (AI0564499; AI048638; AI05726601; DK057665; AI057157, AI-50019) to B.P., NIH grant (AA01787001; AI083554), a Career Development Award from the Crohn’s and Colitis Foundation of America, and an Emory Egleston Children’s Research Center seed grant to T.L.D.