Aryl hydrocarbon receptor (AhR) has been shown to have profound influence on T cell differentiation, and use of distinct AhR ligands has shown that whereas some ligands induce regulatory T cells (Tregs), others induce Th17 cells. In the present study, we tested the ability of dietary AhR ligands (indole-3-carbinol [I3C] and 3,3′-diindolylmethane [DIM]) and an endogenous AhR ligand, 6-formylindolo(3,2-b)carbazole (FICZ), on the differentiation and functions of Tregs and Th17 cells. Treatment of C57BL/6 mice with indoles (I3C or DIM) attenuated delayed-type hypersensitivity (DTH) response to methylated BSA and generation of Th17 cells while promoting Tregs. In contrast, FICZ exacerbated the DTH response and promoted Th17 cells. Indoles decreased the induction of IL-17 but promoted IL-10 and Foxp3 expression. Also, indoles caused reciprocal induction of Tregs and Th17 cells only in wild-type (AhR+/+) but not in AhR knockout (AhR−/−) mice. Upon analysis of microRNA (miR) profile in draining lymph nodes of mice with DTH, treatment with I3C and DIM decreased the expression of several miRs (miR-31, miR-219, and miR-490) that targeted Foxp3, whereas it increased the expression of miR-495 and miR-1192 that were specific to IL-17. Interestingly, treatment with FICZ had precisely the opposite effects on these miRs. Transfection studies using mature miR mimics of miR-490 and miR-1192 that target Foxp3 and IL-17, respectively, or scrambled miR (mock) or inhibitors confirmed that these miRs specifically targeted Foxp3 and IL-17 genes. Our studies demonstrate, to our knowledge for the first time, that the ability of AhR ligands to regulate the differentiation of Tregs versus Th17 cells may depend on miR signature profile.
This article is featured in In This Issue, p.937
Indoles (indole-3-carbinol [I3C] and 3,3′-diindolylmethane [DIM]) are phytochemicals found abundantly in the Brassica (cruciferous) group of vegetables, including broccoli, Brussels sprouts, cabbage, and cauliflower. Cruciferous vegetables are rich in glucosinolates that can endogenously be converted into indole compounds, including I3C and DIM, upon ingestion (1). Also, I3C under acidic conditions gets converted into DIM (2). Both I3C and DIM are two important indoles and have been shown to possess anti-inflammatory, anticancer, and antioxidant properties (3–7). There are studies demonstrating that dietary consumption of indoles in the form of cruciferous vegetables would reduce cancer risk (8–10) and possess potent cancer chemopreventive properties (11, 12). The studies suggest that indoles mediate cancer chemopreventive effects potentially via multiple targets. For instance, indoles are capable of inducing antioxidant activity, regulating cellular genes, inducing cell cycle arrest/apoptosis, and regulating hormone metabolism (13–20). Although the chemopreventive effects of indoles have been well characterized, their effect on the immune system is unclear.
Both I3C and DIM are ligands for aryl hydrocarbon receptor (AhR) and can cause AhR activation (21). AhR is a transcription factor and belongs to the bHLH-PAS family (22, 23). Activated AhR translocates to the nucleus where binding to AhR nuclear translocator generates the AhR–AhR nuclear translocator complex that regulates genes through promoters containing a dioxin-responsive element consensus sequence (11, 12, 24–27). AhR was characterized as a transcription factor responsible for the activation of genes encoding a number of xenobiotic metabolizing enzymes, and it mediates the toxicity induced by environmental chemicals such as 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) (28). Interestingly, recent studies indicated that AhR activation plays diverse roles in cellular functions, including the regulation of the immune system (29, 30). In recent years, there has been significant interest in understanding the role of AhR in T cell differentiation (29, 31).
Regulatory T cells (Tregs) and Th17 cells express AhR, and activation of AhR promoted their differentiation (32). Also, there have been reports that Treg and Th17 cell differentiation is reciprocally regulated (33). Also, whereas TGF (TGF-β1) induces the differentiation of Tregs (34), TGF-β1 along with the combination of IL-6 or IL-21 results in the differentiation of Th17 cells (32, 35). Although many studies have indicated that AhR activation preferentially induces Tregs while suppressing Th17 cells (33, 35), others have shown increased induction of Th17 cells as well (36). It is not clear how activation of AhR causes induction of both Tregs and Th17 cells. One possibility is that the type of ligand and its affinity to AhR may determine whether Tregs or Th17 will be induced (35). Recently, we have shown that activation of AhR by TCDD triggered differential induction of Tregs and Th17 cells through epigenetic regulation (33). Although indoles can activate AhR, the impact of indoles (I3C and DIM) on T cell differentiation in vitro or in vivo has not been fully explored. To this end, we investigated the potential of I3C and DIM in differentiation of Tregs and Th17 cells and evaluated their use and efficacy in the treatment of delayed-type hypersensitivity (DTH) using a mouse model. DTH reactions provide protection against intracellular pathogens. Additionally, DTH response against some innocuous Ags can sometimes trigger severe tissue damage as seen in contact hypersensitivity. In the present study, we used methylated BSA (mBSA) as an Ag to trigger DTH response that has been shown to predominantly trigger Th17 and Th1 cells (37).
MicroRNAs (miRs) have been shown to play a critical role in the regulation of gene expression (7, 38–40). miRs are highly conserved noncoding single-stranded small RNA molecules (17–27 nt) that control gene expression posttranscriptionally. They bind at the 3′ untranslated region (UTR) of the target gene mRNA, resulting in the degradation of the target mRNA or inhibition of the translation of the mRNA (38, 41). There are several studies showing miRs to play a critical role in the regulation of T cell maturation and differentiation (42, 43). Recent studies from our laboratory have also shown miRs playing an important role in T cell differentiation (3, 33, 44–47).
In the present study, we investigated the effect of dietary AhR ligands such as I3C and DIM as well as an endogenous AhR ligand, 6-formylindolo(3,2-b)carbazole (FICZ), on Th17/Treg differentiation using mBSA as an Ag to trigger DTH reaction. Interestingly, we found that whereas I3C and DIM suppressed the DTH response through induction of Tregs and suppression of Th-17 cells, FICZ had opposite effects. We found that the mechanism of action of these AhR ligands was related to the nature of miRs induced by these ligands.
Materials and Methods
Female mice (C57BL/6), aged 10–12 wk, were obtained from the National Institutes of Health (Bethesda, MD). We received AhR knockout (AhR−/−) mice as a breeding pair from Dr. Gary H. Perdew (Pennsylvania State University, College Park, PA) as a gift. AhR−/− mice were bred in our animal facility. All mice were housed and maintained in a temperature- and light-controlled facility and had free access to water and diet at the University of South Carolina School of Medicine animal facility, which is an American Association for the Accreditation of Laboratory Animal Care–accredited and –certified animal facility.
Chemical reagents and Abs
mBSA, I-3C, and DIM were purchased from Sigma-Aldrich (St. Louis, MO). Both I3C and DIM suspended in DMSO were used in in vitro studies, whereas when diluted in corn oil they were used for in vivo studies. FICZ was purchased from Enzo Life Sciences (Farmingdale, NY). We purchased RPMI 1640, penicillin-streptomycin, HEPES, l-Glutamine, FBS, and PBS from Invitrogen Life Technologies (Carlsbad, CA). We purchased the following reagents: anti-mouse IgG-FITC, IgG-PE, CD4-FITC, Foxp3-PE, and IL-17-PE mAbs and Fc Block from BD Pharmingen (Carlsbad, CA); Cytofix/Cytoperm fixation/permeabilization kit from BD Biosciences (San Diego, CA); RNeasy and miRNeasy Mini kits, miScript primer assays kit, and miScript SYBR Green PCR kit from Qiagen (Valencia, CA); iScript and miScript cDNA synthesis kits from Bio-Rad (Madison, WI); Epicentre’s PCR premix F and Platinum Taq DNA polymerase kits from Invitrogen Life Technologies; ELISA kits for IL-17 (ELISA MAX standard set mouse IL-17) and IL-10 (ELISA MAX standard set mouse IL-10) from BioLegend (San Diego, CA).
Generation of DTH in mice
DTH was induced in C57BL/6 mice using mBSA as described previously (48–51). In brief, C57BL/6 mice were sensitized with 200 μl (1.25 mg/ml) mBSA (Sigma-Aldrich, MO) emulsified in CFA (Difco Laboratories, Detroit, MI) by s.c. injection at the base of the tail. Six days later, the mice were rechallenged by intradermal injection of 20 μl 10 mg/ml mBSA suspended in PBS into both footpads. The footpad thickness was measured before and after mBSA or PBS challenge using engineers calipers. The data were calculated as follows: increase in footpad thickness (mm) = footpad thickness after challenge − footpad thickness before challenge.
Treatment with AhR ligands
The mice were treated with vehicle (VEH; corn oil), I3C (50 mg/kg body weight), DIM (50 mg/kg body weight), or FICZ (50 μg/kg body weight) by oral gavage, as described (52, 53). We used these doses based on a dose-response pilot study. We used the following groups of randomized mice: VEH, mice that received corn oil and were used as controls for mice that received either I3C or DIM alone; I3C, mice that received I3C only; DIM, mice that received DIM only; mBSA-VEH, mice immunized and rechallenged with mBSA in footpad and received VEH treatment; mBSA-I3C, mice immunized and rechallenged with mBSA in footpad and received I3C treatment; mBSA-DIM, mice immunized and rechallenged with mBSA in footpad and received DIM treatment. Each group contained at least five mice and each experiment was repeated at least three times.
In vitro culture of draining inguinal lymph node cells and proliferation assays
To determine the effects of I3C or DIM on T cells, proliferation assays were performed. To this end, as described earlier (37), mice (C57BL/6) were s.c. sensitized with 200 μl 1.25 mg/ml mBSA emulsified in CFA and 6 d later inguinal lymph nodes (LNs) were harvested. Single-cell suspensions were prepared and LN cells (4 × 105 cells/well of 96-well round-bottom plates) from various treatment groups were cultured in the absence or presence of mBSA (40 μg/ml) at 37°C for 72 h. The mBSA-specific proliferative responses of LN cells were determined by pulsing with 0.2 μCi–labeled thymidine (54) for 12 h.
mBSA-mediated generation of Tregs and Th17 cells in vitro
To determine the generation of Tregs and Th17 cells in the presence or absence of mBSA, intracellular staining of T cells was performed. In brief, LN cells (3 × 106 cells/well in 24-well plates) were cultured in the absence or presence of mBSA (40 μg/ml) at 37°C for 72 h. In last 6 h of the culture, the cells were stimulated with 1 μg/ml ionomycin and 100 ng/ml PMA (Sigma-Aldrich). GolgiPlug (1 μM) was added to the culture during the last 4 h. After washing twice with FACS buffer (PBS and 1% BSA), LN cells were preblocked with Fc receptors for 15 min at 4°C. The cells were washed with FACS buffer and then stained with FITC-conjugated anti-CD4, PE-conjugated anti-Foxp3, and PE-conjugated anti–IL-17 mAbs (BD Pharmingen) for 30 min with occasional shaking at 4°C. The cells were washed twice with FACS staining buffer and resuspended in BD Cytofix/Cytoperm solution (BD Biosciences) for 20 min. The cells were washed again with BD Perm/Wash solution (BD Biosciences) after incubating them for 10 min at 4°C. LN cells were then washed twice with FACS buffer and analyzed by flow cytometry (FC 500) and using CXP software (Beckman Coulter, Ft. Collins, CO).
Leptin concentration in sera
The concentration of leptin in sera was measured by performing ELISA using a leptin ELISA kit from Crystal Chem (Chicago, IL) and following the protocol of the company. In brief, wells were washed twice using 300 μl washing buffer. Fifty microliters guinea pig anti-mouse leptin serum and 45 μl sample diluents were added to each well. Standards and serum samples (5 μl) from various groups were added to the wells in duplicate and incubated overnight (16–18 h) at 4°C. The wells were washed five times using 300 μl washing buffer. After washing, 100 μl anti–guinea pig IgG enzyme conjugate was added and incubated for 3 h at 4°C. After washing seven times with washing buffer, 100 μl enzyme substrate was added and incubated for 30 min at room temperature. The reaction was stopped after 30 min incubation at room temperature by adding 100 μl enzyme reaction stopping solution. The ELISA plate was read at 450 nm within 30 min of addition of stop solution.
Footpads of various treated groups were horizontally cut and preserved in Cal-Rite purchased from Richard-Allan Scientific (Kalamazoo, MI) for at least 3 wk to decalcify the footpad. Decalcified footpads were sectioned at 5 μm and stained with H&E for microscopic examination in the core facility of University of South Carolina School of Medicine (Columbia, SC).
Role of AhR in ligand-mediated differential regulation of Tregs and Th17 cells and suppression of DTH in mice
To understand the role of AhR in I3C-, DIM-, or FICZ-induced differentiation of Tregs and Th17 cells, in vitro and in vivo assays were performed. In vitro assays were performed using purified T cells from C57BL/6 wild-type (AhR+/+) and AhR knockout (AhR−/−) mice that were cultured in the presence of purified anti-mouse anti-CD3 (2 μg/ml) and anti-CD28 (1 μg/ml) mAb for 3 d. During culture, the T cells were also treated with VEH (DMSO), I3C (50 μM/ml), DIM (50 μM/ml), or FICZ (500 nM/ml). On day 3, the last 5–6 h of culture, PMA (100 ng/ml) and ionomycin (1 μg/ml) from Sigma-Aldrich were added, and in last 4 h of culture, GolgiStop (1 μM/ml) was added into the culture. Intracellular staining of cells was performed as described above. The cells were then analyzed for the presence of Tregs (CD4+/Foxp3+) and Th17 (CD4+/IL-17+) cells in the culture. In some experiments, we also tested FICZ, an endogenous AhR ligand. Furthermore, we also performed in vivo assays to investigate the role of AhR in I3C- or DIM- or FICZ-induced differentiation of Tregs and Th17 cells. To this end, we injected (i.p.) wild-type C57BL/6 (AhR+/+) and AhR knockout (AhR−/−) mice with VEH (corn oil), I3C (50 mg/kg body weight), DIM (50 mg/kg body weight), or FICZ (10 μg/kg body weight). On day 6, spleens from mice treated with VEH, I3C, DIM, or FICZ were harvested and a single-cell suspension was prepared. The cells were then cultured in the presence of PMA (100 ng/ml) and ionomycin (1 μg/ml) for 5–6 h. In the last 4 h of culture, GolgiStop was added into the culture. The cells were stained using anti-mouse CD4-FITC and Foxp3-PE for Tregs and anti-mouse CD4-FITC and IL-17–PE for IL-17+ Th17 cells and were analyzed using flow cytometry.
Furthermore, we used AhR-specific antagonist to determine whether blocking of AhR reverses the effect of I3C, DIM, or FICZ treatment in vivo. To this end, we used AhR-specific antagonist CH223191 (CH). DTH was developed as described above, and AhR-specific antagonist (CH; 5 mg/kg body weight) was injected i.p at least 2 h prior to I3C, DIM, or FICZ treatment. The mice were monitored for DTH responses as described above. The mice were sacrificed 72 h after treatment and spleens were harvested, single-cell suspensions were prepared, and the cells were analyzed for the presence of Tregs and Th17 as described above.
Evaluation of miRs profile in draining (inguinal) LNs
To evaluate miR profiles in mice with DTH and treated with VEH, I3C, DIM, or FICZ, we used draining (inguinal) LNs of mice. In brief, total RNA including miRs from inguinal LNs were isolated using a miRNeasy kit and following the protocol of the company (Qiagen). Next, miR arrays were performed and data were analyzed using hierarchical clustering, as described previously (33, 44, 55). miR arrays data were submitted to Array Express (accession no. E-MTAB-4076; http://www.ebi.ac.uk/arrayexpress/submit/overview.html; https://www.ebi.ac.uk/fg/annotare/). The expression of miRs and pathway network analysis were performed using a two-sample t test, and a p value of <0.01 in the t test was considered significant. In this study, we considered a >1.5-fold change in the expression of miRs as positive.
Real-time PCR to validate the expression of miRs in LNs
After analyzing miR profiles and identifying miRs that were dysregulated >1.5-fold, we validated the expression of some of the miRs. To this end, we selected five downregulated miRs (miR-203, -340, -31, -219a, and -490) and two upregulated miRs (miR-495 and -1192). Snord96a was used as an internal control for miR. Real-time PCR assays were performed on cDNAs generated from total miRs isolated from LNs from mice with DTH and treated with VEH, I3C, DIM, or FICZ, as described earlier. We used a miScript primer assays kit (details in Table I) and a miScript SYBR Green PCR kit from Qiagen and followed the protocol of the company.
|miRBase ID .||Target Sequences (5′→3′) .||Qiagen Catalog No. .|
|miRBase ID .||Target Sequences (5′→3′) .||Qiagen Catalog No. .|
We used a StepOnePlus real-time PCR system v2.1 (Applied Biosystems, Carlsbad, CA) and ViiA 7 RUO software. Real-time PCR was performed with 40 cycles using the following conditions: 15 min at 95°C (initial activation step), 15 s at 94°C (denaturing temperature), 30 s at 55°C (annealing temperature), and 30 s at 70°C (extension temperature and fluorescence data collection). Normalized expression of miRs was calculated using normalized expression 1/4 2_DDCt, where Ct is the threshold cycle to detect fluorescence. The data were normalized to various miRs against internal control miR (Snord96a), fold change of miRs was calculated against control miR, and treatment groups (mBSA plus FICZ, mBSA plus I3C, and mBSA plus DIM) were compared with the mBSA plus VEH group. To define significant differences in miR levels in the LN of mBSA plus FICZ–, mBSA plus I3C–, mBSA plus DIM–, or mBSA plus VEH–treated groups, ANOVA was performed using GraphPad version 4.0 (GraphPad Software, San Diego, CA). Differences between treatment groups were considered significant when a p value was <0.05.
miR–mRNA target interactions
We used miRWalk, microRNA.org, TargetScan mouse 5.2, and miRGen (version 3) software and their database to identify miR-specific mRNA targets. Computational algorithms aid this task by examining base-pairing rules between miR and mRNA target sites, location of binding sites within the target’s 3′-UTR, and conservation of target binding sequences within related genomes. The details of some of miRs and UTR region of their target gene (RNA targets) are described in Fig. 10A.
Determination of miR-specific expression of Foxp3 and IL-17 in LN
To study the role of miRs in the regulation of Foxp3 and IL-17 expression, we selected two miRs: miR-490 that was highly downregulated and miR-1192 that was upregulated after indole treatment. The selection of these two miRs was also based on their binding affinity with complementary sequences of their respective genes (miR-490, Foxp3; miR-1192, IL-17). The transfections were pursued as described earlier (33, 44, 55, 56). Briefly, T cells purified from LNs of mBSA-immunized mice were either transfected with scrambled miR (Mock) or with mature miR-490 mimic or miR-490 inhibitor using transfection kit from Qiagen and following the protocol of the company. Similarly, T cells were either transfected with scrambled miR (Mock) or with mature miR-1192 mimic or miR-1192 inhibitor. The details of miR-490 mimic, miR-490 inhibitor, miR-1192 mimic, and miR-1192 inhibitor are described in Table II. We also used a vector containing GFP gene as a positive control for transfection. After 48 h, GFP expression in LN cells was determined by flow cytometry. The transfected cells cultured for 48 h posttransfection were collected and after washing twice with cold PBS, the cells were either immediately used for total RNA, including miR extraction, or were stored at −80°C for use in future.
|miRBase ID .||Qiagen Catalog No. .|
|miRBase ID .||Qiagen Catalog No. .|
Real-time PCR using miR490- and miR-1192-specific primer assays purchased from Qiagen (Table II) was performed on cDNAs generated from total RNAs, including miRs from transfected LN cells. Snord96a was used as an internal control for miRs. Also, RT-PCR was performed to determine Foxp3 and IL-17 expression using mouse Foxp3- and mouse IL-17–specific sets of forward and reverse primes as described earlier (33). Mouse-specific 18S primer pairs were used as an internal control.
Statistical analyses were performed using GraphPad Prism software (GraphPad Software). The data were expressed as the means ± SEM and compared using a two-tailed paired Student t test or an unpaired Mann–Whitney U test. Single-factor variance ANOVA analyses were used to evaluate groups. Differential (upregulated or downregulated) expression of miRs was analyzed using the two-sample t test method. The significance of analysis of microarrays was performed using the Kaplan–Meier method. Results were considered statistically significant when p values were <0.05 between the control and the experimental groups.
Indoles (I3C and DIM) suppress mBSA-induced DTH response in mice
To study DTH response, C57BL/6 mice were sensitized with mBSA and 6 d later they were rechallenged with mBSA in both footpads, as described in 2Materials and Methods. The following groups of mice were included: VEH, I3C, DIM, mBSA plus VEH, mBSA plus I3C, and mBSA plus DIM. Each group consisted of at least five mice and the experiments were repeated at least three times. The secondary challenge with mBSA into the footpads triggered severe DTH responses in C57BL/6 mice (mBSA plus VEH), as shown by a significant increase in footpad thickness (Fig. 1A, 1B). Histological analysis of footpads also showed significant numbers of infiltrating cells in mBSA plus VEH mice when compared with footpads of mice that received only VEH, I3C, or DIM treatment (Fig. 1C). Interestingly, both mBSA plus I3C and mBSA plus DIM groups showed significant decrease in the DTH response and cellular infiltration (Fig. 1A–C). Taken together, these data suggested that both I3C and DIM can suppress mBSA-induced DTH responses.
Indoles trigger reciprocal differentiation of Tregs and Th17 cells during mBSA-induced DTH response
Next, we examined the impact of treatment with I3C or DIM on Tregs and Th17 cells in mBSA-induced DTH in mice. Treatment of mice with I3C or DIM alone caused small but significant increase in Tregs (CD4+/Foxp3+) when compared with VEH alone group (Fig. 2A). Moreover, mBSA plus I3C and mBSA plus DIM groups showed further increases in the percentage of Tregs when compared with the mBSA plus VEH group. In contrast, mBSA-immunized mice treated with I3C or DIM showed significant decrease in the percentage of Th17 cells when compared with the mBSA plus VEH group (Fig. 2B).
When we measured the levels of IL-17 in the sera of these mice, we noted that mBSA plus I3C and mBSA plus DIM groups showed a significant decrease in IL-17 when compared with the mBSA plus VEH group (Fig. 2C). Also, there was a significant increase in IL-10 (Fig. 2D) in the sera of mice that received mBSA plus I3C or DIM groups when compared with mBSA alone. Interestingly, note that treatment with I3C or DIM alone caused a significant induction of IL-10 when compared with VEH controls.
Indoles (I3C and DIM) suppress proliferation of mBSA-sensitized T cells in vitro
To test whether the indoles (I3C or DIM) had a direct effect on mBSA-sensitized T cells, we performed proliferation assays using T cells from the draining LN of mice sensitized with mBSA and treated with VEH, I3C, or DIM. These in vitro cultures were incubated with or without mBSA. The data shown in Fig. 3A indicated that in the absence of additional stimulation with mBSA, cells from mBSA-immunized mice showed higher levels of proliferation than did those that were obtained from mice challenged with VEH, I3C, or DIM alone. Also, cells from mice treated with mBSA plus I3C or mBSA plus DIM showed lower levels of proliferation than did the mBSA plus VEH group. These data correlated with the in vivo observations that I3C and DIM suppress mBSA-induced activation of T cells. When these cultures were activated in vitro with mBSA in the presence of VEH, I3C or DIM, we noted that there was a much stronger proliferative response in the mBSA plus VEH group when compared with the mBSA plus I3C or mBSA plus DIM groups. These data suggested that addition of I3C or DIM in vitro can further decrease the mBSA-specific T cell proliferation (Fig. 3A). We used Con A in the culture as a positive control.
Indoles regulate differential expression of Tregs and Th17 cells after mBSA sensitization in vitro
Next, we directly tested whether the addition of indoles to in vitro cultures of sensitized T cells would alter their differentiation in vitro. To this end, T cells obtained from draining LNs of mBSA-sensitized mice were cultured in the absence or presence of mBSA and treated with VEH, I3C, or DIM. The analysis of T cell differentiation revealed that both I3C and DIM caused significant induction of Tregs (∼9–11%) when compared with VEH (∼3.5%) (Fig. 3B). However, both I3C and DIM caused suppression of Th17 cell differentiation (Fig. 3C) when compared with VEH. These data demonstrated that both I3C and DIM can promote the differentiation of Tregs while inhibiting that of Th17 cells.
Upon analysis of IL-17 and IL-10 cytokines in the culture supernatant by ELISA, we observed a significant increase in IL-17 (Fig. 3D) but a significant decrease in IL-10 in the mBSA plus VEH group (Fig. 3E). However, for treatment in the mBSA plus I3C or mBSA plus DIM groups, we noted a decrease in IL-17 (Fig. 3D) but an increase in IL-10 (Fig. 3E). Taken together, these data indicated that both I3C and DIM may induce a switch in T cell differentiation from Th17 to Tregs.
Indoles (I3C and DIM) do not affect leptin in mice with DTH
Because dietary indoles can affect metabolism, including leptin concentrations, which can also regulate the immune response, we examined the presence of leptin in sera of mice treated with I3C, DIM, or FICZ with or without DTH. Data obtained from this study showed no significant change in leptin concentration in sera of control or DTH mice and between DTH mice treated with VEH or I3C, DIM, or FICZ (Supplemental Fig. 1). Also, there was no significant difference in the weight of mice from various groups (data not shown). These data demonstrated that the effects of indoles in the present study may be independent of leptin because the DTH assay that we used has short-term peaking at 24–48 h and declines by 72 h.
Indoles (I3C and DIM) induce differentiation of Tregs and Th17 cells via AhR
Because indoles are ligands for AhR, we investigated whether the ability of I3C and DIM to induce Tregs and inhibit Th17 cells, in response to mBSA, resulted from the activation of AhR. To this end, we performed a series of in vitro and in vivo assays using wild-type AhR+/+ and AhR knockout (AhR−/−) mice. We used purified T cells from AhR+/+ and AhR−/− mice to perform in vitro assays. The cells were activated with anti-CD3 plus CD28 Abs and treated with VEH or I3C (50 μM/ml) or DIM (50 μM/ml). We also used FICZ, a potent AhR ligand, to examine its role in the generation of Tregs and Th17 cells. As seen before, I3C and DIM triggered strong induction of Tregs while suppressing Th17 cells in an AhR-dependent manner, whereas, interestingly, FICZ had opposite effects (Fig. 4A, 4B).
We next investigated whether administration of indoles in naive mice would also trigger AhR-mediated induction of Tregs. To this end, we studied the effect of indoles in wild-type (AhR+/+) and AhR knockout (AhR−/−) mice. Spleens from mice treated with indoles or FICZ were examined for Tregs and IL-17+ Th17 cells using flow cytometry. There were significantly higher percentages and numbers of Tregs in I3C- or DIM-treated wild-type mice when compared with VEH-treated mice (Fig. 4C). Upon examination of Th17 cells in spleen, we observed very low numbers of Th17 cells, as expected because these were naive mice (Fig. 4D). Interestingly, however, following FICZ administration, we observed no Treg induction in either AhR+/+ or AhR−/− mice (Fig. 4C, 4D), but we observed marked induction of Th17 cells in AhR+/+ but not in AhR−/− mice (Fig. 4D). Taken together, these data demonstrated that both I3C and DIM trigger Tregs in naive mice via AhR without affecting Th17 cells whereas FICZ was able to induce Th17 cells even in naive mice.
To further confirm that I3C, DIM, or FICZ act through AhR, we used an AhR-specific antagonist (CH) to block AhR. When CH was used prior to I3C, DIM, or FICZ treatments, we observed significant reversal of the action of I3C (Fig. 5A), DIM (Fig. 5B), and FICZ (Fig. 5C) on DTH responses. Also, CH did not have much effect on DTH symptoms when used alone (data not shown). Upon analysis of Tregs and Th17 cells in spleens of various groups of mice with DTH, we found that the ability of I3C and DIM to induce Tregs and suppress Th17 cells and that of FICZ to induce Th17 was reversed following treatment with AhR antagonist (Fig. 5D, 5E). The data obtained from these studies confirmed that I3C, DIM, and FICZ act through AhR.
FICZ augments mBSA-mediated DTH in mice
Because FICZ increased Th17 cells while decreasing Tregs, as was also reported in other studies (35), we tested the effect of FICZ on mBSA-induced DTH in mice. To this end, DTH was induced in C57BL/6 mice as described above and were treated with VEH or FICZ. There was a significant increase in DTH response in mice that received mBSA plus FICZ when compared with mice that received mBSA plus VEH (Fig. 6A). Upon histopathological analysis of the footpad, mBSA plus FICZ groups showed a significant increase in the number of infiltrating cells when compared with mBSA plus VEH groups (Fig. 6B).
FICZ increases generation of Th17 cells in mice with mBSA-induced DTH
We next tested whether the ability of FICZ to enhance DTH response correlated with its ability to induce Th17 cells in vivo. To this end, we used the following groups: VEH only, mBSA plus VEH, FICZ, and mBSA plus FICZ. Additionally, we also included mBSA plus I3C or mBSA plus DIM groups to enable simultaneous comparisons. The data showed that there was significant increase in the absolute number of cells in the draining LNs of mice that received mBSA plus FICZ when compared with mice that received mBSA plus VEH (Fig. 7A). However, in both mBSA plus I3C and mBSA plus DIM groups, there was a significant decrease in the absolute number LN cells (Fig. 7A). Upon analysis of Tregs and Th17 cells in LNs, there was no significant change in the Tregs (Fig. 7B, 7C) in mice that received mBSA plus FICZ when compared with the mBSA plus VEH group (Fig. 7B, 7C). However, there was a significant increase both in the percentage and absolute number of Tregs in LNs of the mBSA plus I3C and mBSA plus DIM groups when compared with the mBSA plus VEH group. In contrast, there was a significant increase in both the percentage and absolute number of Th17 cells in LNs of mice that received mBSA plus FICZ when compared with mBSA plus VEH (Fig. 7D, 7E). Moreover, the mBSA plus I3C and mBSA plus DIM groups showed a significant decrease in the percentage and number of Th17 cells (Fig. 7D, 7E). These data demonstrate that FICZ exacerbates mBSA-induced DTH symptoms by increasing the induction of Th17 cells without affecting Tregs, whereas I3C and DIM suppress DTH by promoting the differentiation of Tregs while decreasing that of Th17 cells.
Analysis of miR profile in LN of mice treated with various compounds
To test whether the differences in the mode of action of the compounds that we tested resulted from differential miR induction, we examined the miR profile in the draining (inguinal) LN cells of mice with DTH and treated with VEH, FICZ, I3C, or DIM. Cluster analysis of 3109 miRs (Fig. 8A), as defined by median absolute deviation in mBSA plus VEH, mBSA plus I3C, mBSA plus DIM, and mBSA plus FICZ, was performed using Ward’s method. Similarity measure of miRs in the four groups was done using half square Euclidean distance method, and ordering function of miRs was done on the basis of input rank. In Fig. 8A, the visualization of cluster analysis of miRs is shown as a tree graph (a dendrogram) and their expression pattern is reflected. Upon comparison of miRs using a Venn diagram (Fig. 8B), we observed that the miRs induced in the mBSA plus VEH group differed significantly from that seen in the mBSA plus I3C, mBSA plus DIM, mBSA plus FICZ, and mBSA plus I3C groups, thereby suggesting that each treatment group (I3C, DIM, FICZ) had a profound impact on miR expression (Fig. 8B). Among the three compounds tested, I3C and DIM seemed to share significantly larger number (n = 113) of miRs. FICZ shared 32 miRs with I3C and 19 miRs with DIM group. There were four miRs that were shared by the three treatment groups that were not expressed in the VEH group. Taken together, these data demonstrated that I3C, DIM, and FICZ induce unique profiles of miR expression when compared with VEH controls and that there was more commonality between I3C and DIM when compared with FICZ.
Differential expression (fold change) of miRs in LN of mice with DTH treated with I3C, DIM, or FICZ
Next, differential (upregulated or downregulated) expression of miRs was analyzed using the two-sample t test method. The significance of analysis of microarrays (Fig. 8C) was performed using the Kaplan–Meier method. A p value of <0.01 in the t test was considered significant. Of the total 3109 miRs screened, there were 102 miRs that were differentially regulated in various groups (Fig. 8D). Upon analysis of >1.5-fold upregulated miRs in the above groups, we found that there were 41 miRs that were unique to I3C, 44 miRs to DIM, and 15 miRs to the FICZ group (Fig. 8E). Upon analysis of 1.5-fold downregulated miRs, there were 20 that were unique to I3C, 29 miRs to DIM, and 12 miRs to the FICZ group (Fig. 8F). Interestingly, note that of the three compounds tested, there was not a single miR that was upregulated that was common to all, and of the downregulated miRs, there was only one that was common to all three compounds.
Validation of miR expression by real-time PCR
Based on the analysis of miR array data, we chose five downregulated miRs (miR-203, -340, -31, -219, and -490) and two upregulated miRs (miR-495 and -1192) to verify and validate their expression in LNs of four (mBSA plus VEH, mBSA plus I3C, mBSA plus DIM, and mBSA plus FICZ) groups of mice (Fig. 9A). To this end, real-time PCR was performed on cDNA samples converted from total RNA, including miRs from LN harvested from various groups treated with VEH, FICZ, I3C, or DIM as described in 2Materials and Methods (Table I). Data obtained from real-time PCR demonstrated downregulated expression of miR-203 and miR-340 in the LN cells of all three treatment groups (Fig. 9B). However, expression of miR-31, miR-219, and miR-490 was upregulated in the FICZ-treated groups (Fig. 9B) but was significantly downregulated in the I3C- and DIM-treated groups (Fig. 9B). In contrast, miR-495 and miR-1192 were downregulated in FICZ-treated groups (Fig. 9B) but significantly upregulated in the I3C- and DIM-treated groups (Fig. 9A). These real-time PCR data validated the expression profile of selected miRs (miR-203, -340, -31, -219, -490, 594, and -1192) obtained from miR arrays of LN cells of the four groups tested.
Analysis of miR-specific Foxp3 (miR-490) and IL-17 (miR-1192) expression in T cells
Upon analysis of binding affinity of the miRs using miRWalk, microRNA.org, and TargetScanMouse 5.2 software, we found a strong binding affinity of miR-31, miR-219, and miR-490 with the 3′-UTR region of the Foxp3 gene and binding affinity of miR-495 and miR-1192 with the 3′-UTR region of the IL-17 gene, as shown in Fig. 10A. These data suggested that the decrease in miR-31, miR-219, and miR-490, caused by I3C and DIM, may account for the increased expression of Foxp3, and the increase in miR-495 and miR-1192 may lead to suppression of IL-17 production. Interestingly, the expression of these clusters of miRs were reversed following FICZ treatment, which may account for its ability to suppress Foxp3 and enhance IL-17.
To further confirm the role of these miRs in the regulation of Foxp3 and Il-17, we performed transfection studies using selected miRs such as miR-490 and miR-1192 that target Foxp3 and IL-17, respectively. To this end, LN cells harvested from mBSA-immunized mice were either transfected with scrambled miR (mock) or with mature miR mimic and inhibitors. GFP expression in LN cells was determined by flow cytometry 48 h posttransfection and >80% LN cells were transfected (Fig. 10B). The expression of miR-490 and miR-1192 in the transfected cells was determined by real-time PCR using miR-specific primer assays (Table II). Snord96a was used as an internal control. There was moderate expression of miR-490 in mock-transfected LN cells but it was significantly upregulated in cells transfected with miR490 mimic (Fig. 10C). However, the expression of miR490 remained low in cells transfected with the inhibitor (Fig. 10C). Upon examination of miR-1192 expression in the transfected LN cells, there was moderate expression of miR-1192 in mock-transfected LN cells but significantly higher expression of miR-1192 in LN cells transfected with miR-1192 mimic (Fig. 10C). However, the expression of miR-1192 was significantly downregulated in miR-1192 inhibitor–transfected LN cells (Fig. 10C). These data demonstrated successful transfection of both miRs (miR-490 and miR-1192) and inhibitors.
Next, expression of Foxp3 and IL-17 in transfected LN cells was determined by performing RT-PCR. There was significantly decreased expression of Foxp3 (Fig. 10D, 10E) and IL-17 (Fig. 10F, 10G) in LN cells transfected with miR-490 mimic and miR-1192 mimic, respectively, when compared with mock-transfected LN cells. In contrast, expression of Foxp3 (Fig. 10D, 10E) and IL-17 (Fig. 10F, 10G) was significantly higher in LN cells transfected with miR-490 inhibitor and miR-1192 inhibitor, respectively. These data demonstrated that miR-490 and miR-1192 have a direct effect on Foxp3 and IL-17 expression.
Brassica (cruciferous) group of vegetables, such as broccoli, Brussels sprouts, cabbage, cauliflower, and the like contain abundant indoles such as I3C (1). DIM is derived from the digestion of I3C. Several studies have shown that dietary consumption of indoles could reduce cancer risk (8–10), as they possess an anticancer property (11, 12), which has led to the pursuit of clinical trials. There are several studies showing indoles achieving chemopreventive effects against cancer by inducing antioxidant activity, regulating cellular proliferative genes, inducing cell cycle arrest, causing apoptosis, and regulating hormone metabolism (14, 20, 57–60). However, the effect of indoles on the immune response has not been well characterized. Indirubin and indigo that are present in human urine have been shown to be potent AhR ligands (61). However, little is known about the impact of dietary indoles on AhR signaling in cells, particularly those of non–tumorigenic origin, as the vast majority of data concerning the intracellular action of indoles have been obtained from cancer cells (61, 62). Recently, studies from our laboratory and elsewhere have shown that AhR plays an important role in the regulation of immune responses specifically promoting the generation of Tregs while suppressing Th17 cells (33, 35). Also, we recently demonstrated that dietary indoles can suppress experimental encephalomyelitis as well as inflammation induced by staphylococcal enterotoxin (46, 63).
DTH plays a critical role in protection of the host against intracellular infections. Moreover, it is involved in mediating tissue injury following exposure to certain chemicals, referred to as contact hypersensitivity. While the DTH response is primarily mediated by Th1 cells, recent studies have suggested that some forms of DTH may also be mediated by Th17 cells. The mBSA-mediated DTH reaction has been shown to be primarily mediated by Th17 cells (37). The effect of indoles on DTH mediated by Th17 cells has not been investigated before. We undertook these studies not only to understand the mechanism of immunomodulation caused by indoles, but also to delineate whether indoles can be used to suppress contact hypersensitivity. Our studies demonstrated that indoles were very effective in suppressing mBSA-mediated DTH in mice through suppression of Th17 cells. These data correlated with significantly increased induction of IL-10 and decreased production of IL-17 in these mice. Moreover, treatment with indoles caused increased induction of Foxp3+ Tregs.
Importantly, we noted that the effects of indoles were mediated through AhR inasmuch as these effects were not seen in AhR−/− mice. In another study, Quintana et al. (35) showed that an endogenous AhR ligand [2-(1′H-indole-3′-carbonyl)-thiazole-4-carboxylic acid methyl ester, or ITE] acting on dendritic cells and T cells suppressed experimental autoimmune encephalomyelitis (EAE). They showed that ITE induced Foxp3+ Tregs that suppressed EAE. Also, ITE acted not only on T cells, but also directly on dendritic cells to induce tolerogenic dendritic cells that supported Treg differentiation in a retinoic acid–dependent manner (35). It was also reported that the presence of natural agonists for AhR in culture medium is essential for optimal differentiation of Th17 T cells (32).
Note that AhR ligands have been shown to exhibit contrasting effects in their ability to trigger either Tregs or Th17 cells. For example, TCDD, a potent AhR ligand, suppresses diabetes in NOD mice by increasing Foxp3+ T cells in pancreatic LNs (31). In another study, TCDD was shown to inhibit EAE symptoms in mice by induction of Tregs (14). In the same study, interestingly, FICZ, which also binds to AhR with high affinity, interfered with Treg cell development and boosted Th17 cell differentiation, thereby increasing the severity of EAE. The precise reasons for such differential effects of AhR ligands are not clear. The above findings with FICZ are consistent with our present study wherein we noted that FICZ, unlike the indoles, promoted Th17 while suppressing Tregs. Also, FICZ enhanced the mBSA-induced DTH in mice (Fig. 5A–C). Okino et al. (64) determined that the two AhR ligands, TCDD and DIM, were able to exert different effects through activation of distinct AhR-controlled pathways. Genetic knockdown of AhR confirmed that the effects of TCDD and DIM were indeed AhR-dependent. They showed that TCDD strongly induced AhR-dependent CYP1 gene expression, whereas DIM was a relatively weak cytochrome P450 inducer. DIM also strongly inhibited estrogen receptor-α expression and estrogen signaling, whereas TCDD had a notably weaker effect on these processes. Additionally, DIM promoted apoptosis and through perturbation of cyclin D caused cell cycle inhibition, similar to our previous findings that AhR activation by TCDD can induce apoptosis in activated but not naive T cells (65–67). In the present study, although we did not study whether I3C and DIM could induce apoptosis in T cells, we noted that indoles decreased Ag-specific T cell proliferation, which could result from induction of apoptosis.
The precise mechanisms that lead to differential induction of Tregs versus Th17 cells by AhR ligands remain unclear. It has been hypothesized that this effect may depend on a variety of factors, including the binding affinity to AhR, route of exposure, and nature of the immune response being triggered at specific sites (68). For example, FICZ was shown to enhance Th17 differentiation and when given s.c. it enhanced EAE (14). However, when FICZ was given i.p., it promoted Tregs and suppressed EAE (69). In the present study, we administered FICZ orally and it appeared to have a similar effect as did the s.c. administration. The present study also demonstrates, to our knowledge for the first time, that the differential effects of AhR ligands such as TCDD, FICZ, and indoles may result from their ability to modulate miR pathways. In this study, we have noted that indoles (I3C and DIM) regulated a large number of miRs in the draining (inguinal) LNs of mice with DTH (Fig. 7). Moreover, some of these dysregulated (upregulated/downregulated) miRs had strong binding affinity with complementary sequences of UTR regions of Foxp3 (miR-31, miR-219, and miR-490) and IL-17 (miR-495 and miR-1192), as shown in Fig. 10A. Importantly, note that whereas I3C and DIM downregulated miR-31, miR-219, and miR-490 and upregulated miR-495 and miR-1192, thereby suggesting how these compounds may increase Foxp3 and decrease IL-17 expression, treatment with FICZ had the opposite effect of on these miRs, which may explain why it decreases Foxp3 and increases IL-17. The role of some of these miRs to regulate Foxp3 and IL-17 was confirmed by transfection experiments. For example, there was significant suppression of Foxp3 expression in T cells that were transfected with miR-490 mimic (Fig. 10D, 10E) and suppression of IL-17 in T cells that were transfected with miR-1192 mimic (Fig. 10F, 10G). However, the expression of Foxp3 and IL-17 was reversed (significantly upregulated) in T cells that were transfected with miR-490 inhibitor and miR-1192 inhibitor, respectively.
We have previously shown that AhR activation by TCDD can trigger epigenetic regulation such as increased methylation of CpG islands of Foxp3 and demethylation of IL-17 promoters, thereby facilitating a switch from Th17 to Tregs (33). TCDD was also found to alter the expression of the miR profile in T cells (33). Additionally, we have also noted that I3C and DIM are potent suppressors of staphylococcal enterotoxin B–induced T cell activation and cytokine storm, and that they mediate these effects by acting as HDAC inhibitors (63). Thus, further studies are necessary to delineate whether the differential effects of AhR agonists on T cell differentiation are caused by epigenetic alterations.
An ideal immunosuppressive drug for the treatment of inflammatory or autoimmune diseases as well as contact hypersensitivity should possess properties of suppressing proinflammatory T cells such as Th1 and Th17 cells while promoting the differentiation and functions of Tregs or Th2 cells. In this context, AhR activation by certain ligands such as plant-derived indoles seems to meet these needs. Moreover, our studies also have implications in the prevention of certain cancers. For example, cruciferous vegetables have been associated with lowering the risk of cancer. Because certain cancers such as colon cancer are linked to chronic inflammation, our studies also provide a useful link between use of cruciferous vegetables, suppression of inflammation, and decreased incidence of colon cancer.
This work was supported in part by National Institutes of Health Grants P01AT003961, R01ES019313, R01MH094755, P20GM103641, and Veterans Affairs Merit Award BX001357, as well as by a University of South Carolina ASPIRE 1 (A011) grant.
The sequences presented in this article have been submitted to The European Bioinformatics Institute’s Array Express database (http://www.ebi.ac.uk/arrayexpress/submit/overview.html) under accession number E-MTAB-4076.
The online version of this article contains supplemental material.
Abbreviations used in this article:
aryl hydrocarbon receptor
experimental autoimmune encephalomyelitis
2-(1′H-indole-3′-carbonyl)-thiazole-4-carboxylic acid methyl ester
regulatory T cell
The authors have no financial conflicts of interest.