Abstract
Activation of the ligand inducible aryl hydrocarbon receptor (AhR) during primary influenza A virus infection diminishes host responses by negatively regulating the ability of dendritic cells (DC) to prime naive CD8+ T cells, which reduces the generation of CTL. However, AhR-regulated genes and signaling pathways in DCs are not fully known. In this study, we used unbiased gene expression profiling to identify differentially expressed genes and signaling pathways in DCs that are modulated by AhR activation in vivo. Using the prototype AhR agonist TCDD, we identified the lectin receptor Cd209a (DC-SIGN) and chemokine Ccl17 as novel AhR target genes. We further show the percentage of DCs expressing CD209a on their surface was significantly decreased by AhR activation during infection. Whereas influenza A virus infection increased CCL17 protein levels in the lung and lung-draining lymph nodes, this was significantly reduced following AhR activation. Targeted excision of AhR in the hematopoietic compartment confirmed AhR is required for downregulation of CCL17 and CD209a. Loss of AhR’s functional DNA-binding domain demonstrates that AhR activation alone is necessary but not sufficient to drive downregulation. AhR activation induced similar changes in gene expression in human monocyte-derived DCs. Analysis of the murine and human upstream regulatory regions of Cd209a and Ccl17 revealed a suite of potential transcription factor partners for AhR, which may coregulate these genes in vivo. This study highlights the breadth of AhR-regulated pathways within DCs, and that AhR likely interacts with other transcription factors to modulate DC functions during infection.
Introduction
Viral infections continue to pose a major disease burden, accounting for millions of lost disability-adjusted life years (1). Emerging evidence points to environmental factors as important modulators of host responses to infection; yet, precisely how signals from the environment influence immune responses to infection are not fully understood (2). However, there is clear evidence leukocyte function is influenced by environmental factors that bind the aryl hydrocarbon receptor (AhR). The AhR is an environment sensing, ligand inducible transcriptional regulator that is broadly expressed in immune cells, including dendritic cells (DCs) (3, 4). DCs are central mediators of antiviral immune defenses. At mucosal surfaces, DCs interact with the environment and provide a key surveillance system that senses and orchestrates responses to infection (5, 6). As part of their response, DCs take up viruses and traffic to lymphoid organs, where they serve as highly specialized professional APCs, attract naive T cells, and direct the adaptive immune responses that ultimately clear the infection.
Over 500 synthetic and naturally derived AhR ligands have been identified via a combination of in vitro reporter-based screening assays and targeted studies using cultured cells and animal models (7). AhR-binding chemicals are structurally diverse, and come from a range of sources, including environmental pollution and pharmaceuticals as well as metabolites from foods and microorganisms (3, 7, 8). Although associations between human exposure to many AhR ligands and alterations to health status remain to be carefully delineated, several epidemiological reports show associations between exposure to AhR-binding pollutants (dioxins and dioxin-like chemicals [DLCs]) and increased respiratory infection infections, increased prevalence of wheezing, and also poorer vaccine responses (9–14). Humans are regularly exposed to DLCs through the diet (15–17), and they remain high priority chemicals of concern because of their bioaccumulation, persistence in the environment, and ability to adversely affect human health around the globe (18). Among DLCs, 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) is commonly used to study the AhR because it not only represents this broad category of chemicals with documented human exposure but also has high affinity for AhR, does not bind other cellular receptors, and, unlike other AhR-binding compounds, is poorly metabolized into other moieties. Thus, TCDD provides a powerful tool to interrogate AhR-mediated changes in immune cells with well-defined in vivo parameters and the absence of potential immune modulation by secondary metabolites.
In the context of infection, AhR activation by TCDD modulates host responses in rodent models of infection, such as influenza A virus (IAV) and HSV (2, 19, 20). For instance, during infection with IAV, AhR activation significantly reduced the response of CD8+ T cells. This is meaningful because CD8+ CTL are essential for fighting primary viral infection. Thus, even when a mild, sublethal inoculum is administered, AhR activation influences morbidity, mortality, and bronchopulmonary inflammation (21–24). Yet, other prior work established that AhR regulates CD8+ T cell responses indirectly, via a mechanism that requires AhR in immune cells, but not in CD8+ T cells (4, 25). TCDD, and other AhR ligands, modulate DC function and influence the CTL response (4, 25–32). During IAV infection, AhR activation reduces DC trafficking from the lung to lymph nodes and dampens the ability of DCs to act as APCs (25, 33) on day three postinfection. In other model systems, AhR ligands, including TCDD, influence steady-state splenic DCs and DC interactions with T cells (4, 25, 26, 30, 31). Moreover, selective deletion of the AhR from the CD11c lineage prevented the dampened CD8 T cell response to IAV in the lung, providing strong evidence that AhR causes intrinsic changes in DCs, which in turn influence antiviral T cell responses (4). Although these studies support that DCs are central mediators of the immunomodulatory action of AhR signaling, how AhR activation influences DC function remains uncertain.
Given that AhR is a transcription factor (TF), it is logical to postulate that changes in DC function driven by the AhR are derived, at least in part, via changes in gene expression. Indeed, AhR activation consistently induces the expression of several known genes in monocytes and DCs, including Ahrr, Cyp1a1, and Ido1 (4, 26, 28, 29, 31, 34, 35). However, using targeted gene approaches, several genes encoding proteins that are involved in DC function, such as costimulatory and adhesion molecules, are unchanged by AhR activation, or changes in expression are inconsistent across studies (28, 32, 36). This suggests that AhR-driven changes in DC function may result from alterations in the expression of other genes and that the context in which the DCs are activated may converge with AhR signaling, leading to distinct profiles of altered genes in different situations. In the current study, we used unbiased transcriptomics to reveal genes and pathways affected by AhR in DCs during acute primary IAV infection. Thus, our data provide evidence as to how AhR activation impacts expression of novel AhR gene targets critical to DC function.
Materials and Methods
Mice and treatments
Five- to eight-week-old C57BL/6 mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Dr. C. Bradfield (University of Wisconsin, Madison, WI) provided breeding stock of Ahrdbd, AhrKO, and Ahrfx/fx mice, and colonies are continually maintained at the University of Rochester Medical Center. Vav1creAhrfx/fx (AhrΔVav1) were created by crossing female Ahrfx/fx mice with male Vav1cre transgenic mice (37). All mice are housed in microisolator cages in a specified pathogen?–free facility at the University of Rochester Medical Center and are provided food and water ad libitum. For all experiments reported in this study, female mice were used. Homozygous gene expression in Ahr knockout (global and conditional) and Ahr mutant mice is continually confirmed using PCR, as previously described (38).
TCDD (>99% purity, Cambridge Isotope Laboratories, Woburn, MA) was dissolved in anisole and diluted in peanut oil. C57BL/6 mice, which express the Ahrb allele, were given a single oral dose of 10 μg TCDD/kg body weight by gavage 1 d before infection. Control mice received the peanut oil–anisole vehicle in the same manner. Ahrfx/fx, AhrΔVav1, and Ahrdbd mice express the Ahrd allele, which has a 10-fold lower affinity for TCDD than protein encoded by the Ahrb allele. Thus, these strains receive a 10-fold higher oral dose of 100 μg TCDD/kg body weight. These doses do not cause overt toxicity, nor do they elicit frank immunotoxicity, such as thymic atrophy (39).
Influenza virus strain A/HKx31 (IAV, x31; H3N2) was prepared, tittered, and stored at −80°C as previously described (24). Mice were anesthetized by i.p. administration of avertin (2,2,2-tribromoethanol; Sigma-Aldrich, Milwaukee, WI) and inoculated intranasally with 20–25 μl containing 120 hemagglutinating units of virus diluted in sterile, endotoxin-tested PBS. This viral inoculation does not cause mortality in immunocompetent mice.
The University of Rochester Institutional Animal Care and Use and Institutional Biosafety Committees reviewed and approved all procedures involving laboratory animals and infectious agents. The University of Rochester is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care, and handling of vertebrate animals is conducted following guidelines set for by the U.S. Public Health Service Policy on Humane Care and Use of Laboratory Animals. All work with infectious agents was conducted with prior approval of the Institutional Biosafety Committee of the University of Rochester, following guidelines of the U.S. National Institutes of Health and the Center for Disease Control and Prevention.
Collection and preparation of immune cells from mice
To obtain lung-derived immune cells, pulmonary vessels were perfused with 5 ml of PBS containing 0.6 mM EDTA and then lungs were digested with collagenase (40). Single-cell suspensions of mediastinal lymph nodes (MLN) cells were prepared as previously described (41, 42). Erythrocytes were removed using an ammonium chloride lysing solution and then cells were washed, passed through a cell strainer, and kept at 4°C until labeling, further cell isolation, or use in an assay. The number of viable cells in each sample was determined using a TC10 automated cell counter (Bio-Rad Laboratories, Hercules, CA). Collection of total RNA from isolated cells was performed using TRIzol extraction and RNeasy Mini Kits (QIAGEN, Germantown, MD).
Flow cytometry and cell sorting
Single-cell suspensions of leukocytes isolated from whole lung tissue were incubated with anti-mouse CD16/CD32 and stained with previously determined optimal concentrations of the following fluorochrome-conjugated mAbs: CD11b (M1/70), CD11c (N418), CD103 (2E7), MHC class II (MHC II; M5/114.15.2), and CD209a (5H10); purchased from eBioscience, BD Biosciences, or BioLegend. Fluorescence minus one controls were used to define gating parameters. Samples were analyzed on an LSR II (BD Bioscience). The total DC population was identified as MHC II+CD11c+ cells, and conventional DC subsets 1 and 2 (cDC1 and cDC2) were identified by expression of CD103 or CD11b, respectively. CD209a expression was then analyzed on total DCs and cDC1 and cDC2 subsets. Data analyses were performed using FlowJo software (Tree Star, Ashland, OR). To purify mouse DCs for gene expression analyses, immune cells from lungs of infected mice (±AhR activation) were pooled, and CD11c+ cells were enriched using a Mouse CD11c Microbeads Kit (Miltenyi Biotech, Auburn, CA). Enriched cells were stained with fluorochrome-conjugated mAbs against CD11c and MHC II for sorting (FACS Aria). The purity of isolated DCs (CD11c+MHC IIhi cells) was >94% (Fig. 1B). This population excluded CD11c+MHCnull/lo cells, which lack the ability to activate naive T cells and are often considered to be alveolar macrophages (4). Collection of total RNA from isolated DCs was performed using TRIzol extraction and RNeasy Mini Kits (QIAGEN).
RNA-sequencing library construction
The total RNA concentration from sorted mouse DCs was determined with the NanoDrop 1000 Spectrophotometer (NanoDrop, Wilmington, DE) and RNA quality assessed with an Agilent Bioanalyzer 2100 (Agilent, Santa Clara, CA). There were three independent DC samples for each treatment group. The TruSeq RNA Sample Preparation Kit v2 (Illumina, San Diego, CA) was used for next-generation sequencing library construction per the manufacturer’s protocols. Briefly, mRNA was purified from 100 ng of total RNA with oligo-dT magnetic beads and fragmented. First-strand cDNA synthesis was performed with random hexamer priming followed by second-strand cDNA synthesis. End repair and 3′ adenylation was then performed on the dscDNA. Illumina adaptors were ligated to both ends of the cDNA, purified by gel electrophoresis and amplified with PCR primers specific to the adaptor sequences to generate amplicons of∼200–500 bp in size. The amplified libraries were hybridized to the Illumina single-end flow cell and amplified using the cBot System (Illumina) and sequenced using the Illumina Genome Analyzer IIx. All sequence data have been submitted to the Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE123996) under accession number GSE123996.
Computational analysis of RNA-Sequencing data
Raw reads (65 bp) were generated from Illumina GAIIx sequencing. Quality filtering and adapter removal were performed using Trimmomatic version 0.32 (43) with the following parameters: “SLIDINGWINDOW:4:20 TRAILING:13 LEADING:13 ILLUMINACLIP:adapters.fasta:2:30:10 MINLEN:15.” Processed/cleaned reads were then mapped to the Mus musculus reference sequence (mm9) using the STAR version 2.4.2a (44) and the following parameters: “–twopassMode Basic–runMode alignReads–genomeDir ${GENOME}ds–readFilesIn ${SAMPLE}–outSAMtype BAM SortedByCoordinate–outSAMstrandField intronMotif–outFilterIntronMotifs RemoveNoncanonical.” Uniquely aligned reads were counted within the mm9 gene annotations in a nonstrand-specific manner using htseq-count version 0.6.1 (45) and the following parameters: -q -f bam -s no -r pos -i gene_name.” In total, the processed reads were mapped to a total of 23,024 genes with 14,156 genes detected in all three samples of the vehicle and/or TCDD-exposed group. Differential expression analyses and data normalization were performed using DESeq2 version 1.12.4 (46, 47) R/Bioconductor package (48) with an adjusted p value (Benjamini–Hochberg) threshold of 0.05 within the R version 3.2.3 environment (https://Ahr.R-project.org). Heat maps were produced using the pHeatmap version 1.0.8 (https://CRAN.Rproject.org/package=pheatmap) package with row scaling and hierarchical clustering of the rlog transformed expression values. Principal component analysis (PCA) plots were created within R using the pcaExplorer package, also giving the rlog transformed expression values (49). Circos plots were generated using the OmicCircos R package (50) using RNA-sequencing (RNA-Seq) fold changes and binding site information. A full listing of identified differentially expressed genes (DEGs) is provided in Supplemental Table I.
Real-time RT-PCR
Real-time RT-PCR was performed using a Bio-Rad iCycler MyiQ2 with IQ SYBR Green Supermix (Bio-Rad Laboratories). The primers used for gene-specific amplification are listed in Supplemental Table II. Changes in the expression of a particular gene, compared with the same DC from vehicle-treated mice and normalized to the housekeeping gene L13, were calculated using the 2−ΔΔCT method (51).
Lung and MLN isolation for CCL17 ELISA
At the time of harvest, whole lung tissue was snap frozen in liquid nitrogen and stored at −80°C for future analysis. Frozen lung tissue was then homogenized in 5% BSA in Dulbecco’s PBS and proteinase inhibitors (Pierce Protease Inhibitor Mini tablets; Thermo Fisher Scientific), centrifuged to remove tissue debris, and supernatants were used in the analysis. MLN explants were prepared as described by Legge et al. (52, 53). Briefly, MLN explants were prepared using three lymph nodes per well, in IMDM supplemented with 10% FBS, penicillin, streptomycin, and amphotericin B. MLNs were cultured at 37°C in 5% CO2 for 48 h.
MLNs were washed and then placed in 1 ml of fresh media and allowed to incubate at 37°C for 48 h. At the time of collection, supernatants were collected, centrifuged, and stored at −80°C until used. Levels of CCL17/TARC in lung homogenates and MLN explant supernatants were measured by DuoSet ELISA (DY529; R&D Systems). Each sample was run in duplicate. Sample OD was measured at 450 nm, and the concentration determined by comparison with a standard curve.
Aryl hydrocarbon responsive element binding site motif discovery
Genome-scale pattern matching was performed using Regulatory Sequence Analysis Tools program (RSAT; http://rsat.sb-roscoff.fr/index.php) (54–57). Briefly, known aryl hydrocarbon responsive element (AHRE) sequences (AHREI: 5′-GCGTG-3′, AHREII: 5′-CATGnnnnnnC(T/A)TG-3′, and RelBAHRE: 5′-GGGTGCAT-3′) were input and referenced against the full length C57BL/6 mouse genome (GRCm38, mm10). Gene lists with presumptive AHRE were cross-referenced to DEGs identified by RNA-Seq. The number of AHREs on the direct and reverse strand of DNA were enumerated for each DEG and included in the analysis. A full listing of identified AHREs in upstream regulatory region of each DEG is located in Supplemental Table III.
TF binding motif discovery and analysis
The 5-kb upstream regulatory region of Ccl17 and Cd209a was identified and exported via RSAT using reads of the M. musculus GRCm38 (mm10) or Homo sapiens genome. TF binding sites were identified using PROMO (http://alggen.lsi.upc.edu/), a digital tool that uses positional weight matrices constructed from known TF binding sites to search for matches in DNA sequences (58, 59). Only TFs that were present in both regulatory regions of Ccl17 and Cd209a and were expressed in all samples by RNA-Seq were interrogated further. For human DCs, TF gene expression data were unavailable; thus, all TFs identified in the upstream regulatory region of Ccl17 and Cd209a were included in this analysis. A TF binding site identified within 20 bp of at least two independent AHREs in the upstream promoter region of each gene was included in the analysis. Identified TFs and their location in the upstream promoter region relative to the transcriptional start site of each DEG is located in Supplemental Table III.
Human primary DC cultures
Primary human monocytes were obtained by leukophoresis from ungendered healthy anonymous donors who were known nonsmokers from the University of Nebraska Medical Center. Collection of PBMC was performed in full compliance and approval of the University of Nebraska Medical Center Institution Review Board. Monocytes were purified using counter-current centrifugal elutriation and shipped overnight to the University of Rochester on ice in RPMI 1640, containing 10% FBS. Upon receipt, 4 × 106 PBMCs per well were cultured with 50 ng/ml recombinant human GM-CSF (PeproTech) and 25 ng/ml recombinant human IL-4 (PeproTech) for 6 d in vitro. Nonadherent immature DCs were then collected, counted, and used for experiments. Immature DCs were treated with TCDD (10 nM) or 0.01% DMSO control for 24 h prior to the addition of IAV (200 hemagglutinating units/ml) and incubation for another 24 h. Other wells of DCs were not treated with IAV, providing controls for infection. DCs were collected, RNA was isolated using TRIzol and RNeasy Mini Kits (QIAGEN), and quantitative real-time RT-PCR (qRT-PCR) was performed.
Statistical analyses
With the exception of sequencing data, statistical analyses were performed using JMP 9.0.0 (SAS Institute, Cary, NC). Differences between means of multiple independent variables were compared using one-way ANOVA followed by post hoc tests (Tukey honestly significant difference). Differences between vehicle and TCDD treatment groups within the same genotype and at a single point in time were analyzed using a Student t test. Differences in mean values were considered statistically significant when p < 0.05. Error bars on all graphs represent the SEM.
Results
Transcriptome analysis of isolated DCs reveals novel AhR target genes and pathways
To identify genes affected by AhR activation in DCs in the context of respiratory viral infection, we used unbiased transcriptome analysis. Specifically, on day three postinfection, DCs were isolated from the lungs of IAV-infected mice treated with either the prototype AhR ligand TCDD or vehicle control and purified using immunomagnetic enrichment followed by FACS to yield highly enriched DCs (Fig. 1A, 1B). Transcriptome analysis was performed using high-throughput RNA-Seq, and downstream PCA was used to determine the extent of variability between each sample and group. Principle component 1, which explains 94.3% of the variation, clearly separates the vehicle and TCDD-exposed groups from each other (Fig. 1C). The 50 most upregulated and 50 most downregulated genes were identified and arranged into a heatmap with unsupervised clustering (Fig. 1D). Overall, AhR activation resulted in 631 DEGs, with 310 upregulated genes and downregulation of 321 genes in DCs from infected mice (Fig. 1E). The full list of DEGs is provided in Supplemental Table I. Ingenuity Pathway Analysis revealed that the top five cellular pathways influenced by AhR activation include immune cell trafficking, cellular movement, and hematological system development and function (Fig. 1F). The scope of genes and pathways altered highlights the diverse role that the AhR plays within DCs, ranging from influencing pathogen receptors, chemokines, and cell-signaling proteins linked to DC development and function.
AhR activation alters gene transcription in purified Lung DCs.
(A) Mice (180 female, C57BL/6, 8 wk of age) were exposed to either 10 μg TCDD per kilogram of body weight (T) or peanut oil vehicle control (V) orally 1 d prior to intranasal IAV infection. Three days postinfection, lungs were collected, and DCs (CD11c+MHC II+ cells) were purified by FACS. RNA-Seq was performed using three independent biological replicates for each treatment group (six total samples). (B) Representative FACS plots of sorted DC samples from each treatment group are shown. Numbers represent mean percentage (±SEM) for each group. (C) PCA of gene expression in DCs from vehicle- and TCDD-treated mice. (D) Volcano plot of the 631 DEGs reporting p value (−log10) as a function of log2-fold change (horizontal gray line denotes p = 0.05). (E) Heatmap of the top 50 upregulated and the top 50 downregulated genes influenced by AhR activation. Genes were grouped by unsupervised hierarchical clustering. Fold change (red, upregulated; blue, downregulated; white, no change) for each independent sample within a treatment group (orange, vehicle; green, TCDD) are shown, and gene identification is listed along the right side of the heatmap. The complete list of all DEGs is in Supplemental Table I. (F) All 631 DEGs were analyzed using Ingenuity Pathway Analysis. The top five cellular functions and physiological systems altered by AhR activation in DCs are shown.
AhR activation alters gene transcription in purified Lung DCs.
(A) Mice (180 female, C57BL/6, 8 wk of age) were exposed to either 10 μg TCDD per kilogram of body weight (T) or peanut oil vehicle control (V) orally 1 d prior to intranasal IAV infection. Three days postinfection, lungs were collected, and DCs (CD11c+MHC II+ cells) were purified by FACS. RNA-Seq was performed using three independent biological replicates for each treatment group (six total samples). (B) Representative FACS plots of sorted DC samples from each treatment group are shown. Numbers represent mean percentage (±SEM) for each group. (C) PCA of gene expression in DCs from vehicle- and TCDD-treated mice. (D) Volcano plot of the 631 DEGs reporting p value (−log10) as a function of log2-fold change (horizontal gray line denotes p = 0.05). (E) Heatmap of the top 50 upregulated and the top 50 downregulated genes influenced by AhR activation. Genes were grouped by unsupervised hierarchical clustering. Fold change (red, upregulated; blue, downregulated; white, no change) for each independent sample within a treatment group (orange, vehicle; green, TCDD) are shown, and gene identification is listed along the right side of the heatmap. The complete list of all DEGs is in Supplemental Table I. (F) All 631 DEGs were analyzed using Ingenuity Pathway Analysis. The top five cellular functions and physiological systems altered by AhR activation in DCs are shown.
Among the most upregulated DEGs (Table I) were bona fide AhR target genes, such as Ahrr and Cyp1a1, as well as Ido1, a gene that is consistently elevated in DCs following AhR stimulation (4, 26, 34, 35). Within the most downregulated genes were members of the Cd209 family, including Cd209a, the murine homolog of DC-specific ICAM 3–grabbing nonintegrin (DC-SIGN), and several chemokine genes, including chemokine (C-C motif) 17 (Ccl17, also known as TARC) (Table I). Both CD209a and CCL17 play important roles in the attraction and maintenance of DC contacts with lymphocytes (60, 61). Additionally, CD209a is critical during viral Ag uptake and processing by DCs. Therefore, reduced expression of these molecules may dampen the overall ability of DCs to attract and properly activate naive T cells, thereby blunting the overall primary adaptive immune response to IAV infection. Thus, we sought to determine the dependence of these changes on AhR expression and function.
Gene Name . | Log2-Fold Change (TCDD/Vehicle) . | p Value . |
---|---|---|
Ahrr | 3.874 | 1.25 × 10−123 |
Cyp1a1 | 2.415 | 8.29 × 10−20 |
Nos2 | 1.570 | 8.86 × 10−43 |
Clec4d | 1.254 | 2.07 × 10−22 |
Ido1 | 1.238 | 3.07 × 10−7 |
Clec4a4 | 0.990 | 5.20 × 10−6 |
Il10 | 0.809 | 2.35 × 10−4 |
Cd274 | 0.737 | 6.75 × 10−9 |
Cxcl3 | 0.658 | 4.34 × 10−10 |
Cxcl16 | 0.616 | 4.66 × 10−10 |
Clec7a | −0.666 | 1.29 × 10−7 |
Ccl6 | −0.717 | 6.82 × 10−4 |
Ccl9 | −0.790 | 2.02 × 10−6 |
Pdcd1 | −0.846 | 3.27 × 10−3 |
Arg1 | −0.924 | 1.57 × 10−6 |
Ccl22 | −1.029 | 2.68 × 10−7 |
Clec4m | −1.218 | 1.08 × 10−9 |
Cd209d | −1.281 | 6.02 × 10−7 |
Cd209a | −1.337 | 2.88 × 10−6 |
Ccl17 | −1.742 | 2.88 × 10−6 |
Gene Name . | Log2-Fold Change (TCDD/Vehicle) . | p Value . |
---|---|---|
Ahrr | 3.874 | 1.25 × 10−123 |
Cyp1a1 | 2.415 | 8.29 × 10−20 |
Nos2 | 1.570 | 8.86 × 10−43 |
Clec4d | 1.254 | 2.07 × 10−22 |
Ido1 | 1.238 | 3.07 × 10−7 |
Clec4a4 | 0.990 | 5.20 × 10−6 |
Il10 | 0.809 | 2.35 × 10−4 |
Cd274 | 0.737 | 6.75 × 10−9 |
Cxcl3 | 0.658 | 4.34 × 10−10 |
Cxcl16 | 0.616 | 4.66 × 10−10 |
Clec7a | −0.666 | 1.29 × 10−7 |
Ccl6 | −0.717 | 6.82 × 10−4 |
Ccl9 | −0.790 | 2.02 × 10−6 |
Pdcd1 | −0.846 | 3.27 × 10−3 |
Arg1 | −0.924 | 1.57 × 10−6 |
Ccl22 | −1.029 | 2.68 × 10−7 |
Clec4m | −1.218 | 1.08 × 10−9 |
Cd209d | −1.281 | 6.02 × 10−7 |
Cd209a | −1.337 | 2.88 × 10−6 |
Ccl17 | −1.742 | 2.88 × 10−6 |
Log2 (fold change) and p value of 10 upregulated and 10 downregulated genes of interest expressed in DCs from infected mice following AhR activation.
AhR activation reduces CD209a expression
CD209 family members, including DC-SIGN, promote the adhesion of DCs with naive T cells and support DC migration to T cell–rich areas (62, 63). CD209 family members also act as a broad-spectrum pathogen uptake receptor (64–66). Given that CD209 molecules play roles in multiple aspects of DC functions and that aspects of DC functions are altered upon AhR activation, it is possible that Cd209a represents a novel AhR target gene. Consistent with RNA-Seq data, in vivo AhR activation reduced Cd209a levels in lung DCs∼3-fold relative to DCs isolated from infected mice treated with the vehicle control (Fig. 2A). To determine whether changes in gene expression correspond with reduced protein levels, flow cytometry was used. Consistent with the reduced gene expression, activation of the AhR caused a 33% reduction in the percentage of DCs that express CD209a (Fig. 2B). Global loss of AhR ameliorated the reduction in the proportion of DCs that were CD209a+ following TCDD exposure (Fig. 2B). Although the mean fluorescence intensity (MFI) of CD209 was also∼30% lower, the differences in the CD209a MFI on lung DCs from the two groups was not statistically significant (Fig. 2C). Global loss of AhR also did not alter CD209a MFI upon infection (Fig. 2D). When the two main conventional DC subsets in the lung were examined separately, the decline in CD209a was not observed on cDC1s (Fig. 2E) but was observed on cDC2s (Fig. 2G). Consistent with total DCs, there was not a statistically significant difference in the level of CD209a expression on cDC1s (Fig. 2F) or the cDC2s (Fig. 2H), although there was a 20% reduction in the CD209a MFI on cDC2s. To determine if the reduction in CD209a was directly because of AhR activation within hematopoietic cells, AhrΔVav1 mice were used. Similar to wild-type (WT) B6 mice, AhR activation in Ahrfx/fx mice significantly reduced the percentage of CD209a+ lung DCs (Fig. 2I). Loss of AhR in the hematopoietic compartment (AhrΔVav1) abrogated this down-modulatory effect of AhR activation on the percentage of DCs expressing CD209a (Fig. 2J). These data indicate that AhR regulates CD209a expression on DCs via a mechanism that requires AhR in hematopoietic cells, and AhR expression in nonhematopoietic cells does not play a role in regulating CD209a surface expression.
AhR activation reduces lung DC gene and surface expression of DC-SIGN.
Mice were treated with vehicle control (V) or TCDD (T) 1 d prior to IAV infection, and all endpoints measured 3 d postinfection. (A) Cd209a gene expression in isolated lung DCs (CD11c+MHC II+ cells). (B) CD209a surface expression by percentage of total lung DCs from AhrWT and AhrKO. (C and D) Representative histograms depict the expression of CD209a on lung DCs from (C) AhrWT mice and (D) AhrKO mice. Numbers on the histograms indicate the MFI; fluorescence minus one (FMO) control (gray fill), V (blue line), and T (red line). (E) The percentage of cDC1s in the lung that were CD209a+. (F) The MFI of CD209 on cDC1s from the infected lung. (G) The percentage of lung cDC2s that were CD209a+. (H) The MFI of CD209a on cDC2s from the lung. (I) The percentage of lung DCs from infected Ahrfx/fx and AhrΔVav1 mice that express CD209a. (J) Histograms depict the expression of CD209a on lung DCs from Ahrfx/fx and AhrΔVav1 mice; numbers denote MFI in each group: FMO control (gray fill), V (Ahrfx/fx solid blue line, AhrΔVav1 light blue), and T (Ahrfx/fx solid red, AhrΔVav1 light red). Data are representative of three independent experiments. Group sizes within each experiment ranged from three to six mice per treatment group. The data represent the mean ± SEM for each group. *p < 0.05 by Student t test or one-way ANOVA.
AhR activation reduces lung DC gene and surface expression of DC-SIGN.
Mice were treated with vehicle control (V) or TCDD (T) 1 d prior to IAV infection, and all endpoints measured 3 d postinfection. (A) Cd209a gene expression in isolated lung DCs (CD11c+MHC II+ cells). (B) CD209a surface expression by percentage of total lung DCs from AhrWT and AhrKO. (C and D) Representative histograms depict the expression of CD209a on lung DCs from (C) AhrWT mice and (D) AhrKO mice. Numbers on the histograms indicate the MFI; fluorescence minus one (FMO) control (gray fill), V (blue line), and T (red line). (E) The percentage of cDC1s in the lung that were CD209a+. (F) The MFI of CD209 on cDC1s from the infected lung. (G) The percentage of lung cDC2s that were CD209a+. (H) The MFI of CD209a on cDC2s from the lung. (I) The percentage of lung DCs from infected Ahrfx/fx and AhrΔVav1 mice that express CD209a. (J) Histograms depict the expression of CD209a on lung DCs from Ahrfx/fx and AhrΔVav1 mice; numbers denote MFI in each group: FMO control (gray fill), V (Ahrfx/fx solid blue line, AhrΔVav1 light blue), and T (Ahrfx/fx solid red, AhrΔVav1 light red). Data are representative of three independent experiments. Group sizes within each experiment ranged from three to six mice per treatment group. The data represent the mean ± SEM for each group. *p < 0.05 by Student t test or one-way ANOVA.
AhR activation reduces CCL17 production in the lung and MLN during IAV infection
To effectively present Ag and initiate an adaptive immune response, DCs need to attract and retain naive T cells. One means by which DCs create stable interactions with T cells is by producing the chemokine CCL17 (61, 67, 68). CCL17 is one of the chemokines made by DCs that has been associated with promoting T cell recruitment toward DC-rich areas and supporting stable cognate T cell–DC interactions (61). Ccl17 was one of the most downregulated genes following AhR activation (Fig. 1, Table I). To extend this observation, DCs from lungs of IAV-infected mice were isolated, and Ccl17 gene expression and protein levels were measured using real-time RT-PCR and ELISA. Compared with Ccl17 expression in DCs from infected mice given the vehicle control, AhR activation reduced the level of Ccl17 expression in DCs by one third (Fig. 3A). Similarly, CCL17 protein levels in the lung were reduced upon AhR activation (Fig. 3B). Specifically, in IAV-infected mice treated with the vehicle control, the average concentration of CCL17 in lung homogenates was 302.3 (±38.7) pg/ml, whereas the average CCL17 level dropped to 157.1 (±22.6) pg/ml in lungs from infected, TCDD-exposed mice, a roughly 50% reduction (Fig. 3B). Similar to CD209a, the absence of AhR prevented the ability of TCDD to cause a decline in CCL17 levels (Fig. 3B). DCs take up virus in the lung and then emigrate to lymphoid organs, which is the primary site where they interact with naive T cells. Thus, we determined whether the reduction in CCL17 in lung was also observed in the lung-draining MLN. To capture the level of in vivo CCL17 production, whole MLN explant cultures were employed (52, 53). CCL17 was detected in all explant cultures derived from IAV-infected WT mice that were treated with vehicle control (Fig. 3C). In contrast, AhR activation significantly reduced CCL17 levels in explant cultures of MLNs from TCDD-treated WT mice. (Fig. 3C). Given that hematopoietic and nonhematopoietic cells have been reported to express Ccl17 (61, 69–71), to determine whether AhR is required in the hematopoietic compartment (AhrΔVav1) abrogated, we used AhrΔVav1 mice. Lung CCL17 levels were comparable in infected Ahrfx/fx and AhrΔVav1 vehicle control groups, indicating that conditional deletion of AhR in hematopoietic cells did not affect production of CCL17. Yet, in contrast to significant reductions of lung CCL17 in TCDD-treated Ahrfx/fx mice, treatment of infected AhrΔVav1 mice with TCDD did not reduce lung CCL17 levels (Fig. 3D). This indicates that AhR modulates CCL17 through a mechanism that requires AhR in hematopoietic cells. In contrast to the lung, CCL17 levels in MLN explants from Ahrfx/fx and AhrΔVav1 mice were consistently below the limit of detection (data not shown). Consequently, hematopoietic compartment-intrinsic effects of AhR activation on CCL17 levels in MLN are indeterminant. Despite this limitation, these data suggest that AhR activation regulates CCL17 gene expression and protein production via a mechanism that requires the hematopoietic expression AhR in the in the lung compartment.
AhR activation reduces CCL17 production in the Lung and MLN.
Mice were treated with vehicle control (V) or TCDD (T) 1 d before IAV infection, and Ccl17/CCL17 was measured 3 d postinfection. (A) Ccl17 gene expression in isolated lung DCs was measured by qRT-PCR. The data represent the average ± SEM for each group. CCL17 protein levels in lung homogenates from IAV-infected (B) WT and AhrKO mice were measured by ELISA. Data presented are representative of at least two independent experiments that yield similar results. Each experiment had three to nine mice per treatment group. (C) Intact MLNs from three mice in the same group were combined and cultured ex vivo for 3 d, and CCL17 levels in explant culture supernatants were measured by ELISA. (D) CCL17 levels in the lung homogenates of Ahrfx/fx and AhrΔVav1 (four to eight per group) were measured by ELISA. For lung homogenates (A, B, and D), each symbol indicates an individual mouse. For MLN explant cultures (C), each symbol represents a sample consisting of three pooled MLNs. The cross marks (B–D) indicate the mean and SEM for each group. *p < 0.05 by Student t test or one-way ANOVA.
AhR activation reduces CCL17 production in the Lung and MLN.
Mice were treated with vehicle control (V) or TCDD (T) 1 d before IAV infection, and Ccl17/CCL17 was measured 3 d postinfection. (A) Ccl17 gene expression in isolated lung DCs was measured by qRT-PCR. The data represent the average ± SEM for each group. CCL17 protein levels in lung homogenates from IAV-infected (B) WT and AhrKO mice were measured by ELISA. Data presented are representative of at least two independent experiments that yield similar results. Each experiment had three to nine mice per treatment group. (C) Intact MLNs from three mice in the same group were combined and cultured ex vivo for 3 d, and CCL17 levels in explant culture supernatants were measured by ELISA. (D) CCL17 levels in the lung homogenates of Ahrfx/fx and AhrΔVav1 (four to eight per group) were measured by ELISA. For lung homogenates (A, B, and D), each symbol indicates an individual mouse. For MLN explant cultures (C), each symbol represents a sample consisting of three pooled MLNs. The cross marks (B–D) indicate the mean and SEM for each group. *p < 0.05 by Student t test or one-way ANOVA.
Identification of presumptive AHRE in the upstream regulatory region
Given that the AhR is a ligand activated TF, we reasoned that at least a portion of the DEGs identified in DCs using transcriptomics represent AhR target genes that have not been described before. However, the extent of AhR bindings sites among the 631 DEGs identified has not been thoroughly cataloged, including possible AhR-binding sites in the upstream regulatory regions of Ccl17 and Cd209a. Several AHRE have been described, including the canonical AHRE (5′-GCGTG-3′, also referred to as AHREI), a noncanonical AHREII (5′-CATGnnnnnnC(T/A)TG-3′), and a RelBAHRE motif (5′-GGGTGCAT-3′) (72–74). We employed publicly available genomic tools to map each DEG to its chromosomal location throughout the murine genome (Fig. 4A, outer rings). We then identified and mapped putative AHREs in the upstream regulatory region of each DEG based on sequence analysis (Fig. 4A, inner three rings). Of the 631 DEGs, 87% (549/631) contained at least one AHREI the upstream regulatory, 10.1% (64/631) contained a presumptive RelBAHRE in their upstream promoter regions, and only 1/631 DEG (0.15%) contained an AHREII (Fig. 4B). Of the presumptive AHREI containing genes, 58 of them also contained a RelBAHRE (58/549, 10.6%). The single gene containing an upstream AHREII motif, gab2, also contained five AHREIs but no RelBAHREs. This suggests that the dominant binding motif used by AhR to regulate gene expression in DCs, at least in the context of TCDD exposure and IAV infection, is the canonical AHREI. This sequence-based analysis also demonstrates dispersion of AhR-regulated genes and AHREs throughout the genome and suggests no bias toward a specific genomic region.
Identification and distribution of AHREs in the upstream regulatory region of AhR responsive genes in DCs.
All 631 DEGs identified by RNA-Seq were mapped to the genome, and sequences for the canonical AHRE, RelBAHRE, or AHREII binding sites in upstream regulatory regions were identified and mapped. (A) Genomic distribution of DEGs throughout the mouse genome denotes theoretical AHRE. Rings depict the following: DEG name (outer most ring), log2-fold change (red, upregulation; blue, downregulation; second inner ring), genomic ideogram (first inner ring), number of canonical AHRE (5′-GCGTG-3′, black inner ring), number of RelBAHRE sites (5′-GGGTGCAT-3′, green ring), and number of AHREII (5′-CATG{N6}C[T/A]TG-3′, purple ring). (B) Venn diagram shows the distribution of different AHREs identified by analysis of the upstream regulatory region. (C–E) Each individual DEG was plotted versus the number of presumptive (C) AHRE, (D) RelBAHRE, and (E) AHREII binding sites identified by RSAT. Circles indicate a single DEG, and red filled circles designate ccl17 and Cd209a on each graph.
Identification and distribution of AHREs in the upstream regulatory region of AhR responsive genes in DCs.
All 631 DEGs identified by RNA-Seq were mapped to the genome, and sequences for the canonical AHRE, RelBAHRE, or AHREII binding sites in upstream regulatory regions were identified and mapped. (A) Genomic distribution of DEGs throughout the mouse genome denotes theoretical AHRE. Rings depict the following: DEG name (outer most ring), log2-fold change (red, upregulation; blue, downregulation; second inner ring), genomic ideogram (first inner ring), number of canonical AHRE (5′-GCGTG-3′, black inner ring), number of RelBAHRE sites (5′-GGGTGCAT-3′, green ring), and number of AHREII (5′-CATG{N6}C[T/A]TG-3′, purple ring). (B) Venn diagram shows the distribution of different AHREs identified by analysis of the upstream regulatory region. (C–E) Each individual DEG was plotted versus the number of presumptive (C) AHRE, (D) RelBAHRE, and (E) AHREII binding sites identified by RSAT. Circles indicate a single DEG, and red filled circles designate ccl17 and Cd209a on each graph.
We also determined the number of each type of AHRE identified within the upstream regulatory region of all DEGs (Fig. 4C–E). Sequence-based mapping identified 14 AHREI on the direct and reverse strand upstream of the Cyp1a1 transcriptional start site and zero AHREII or RelBAHREs. This aligns with prior mapping of the mouse Cyp1a1 gene (75). The upstream regulatory region of mouse Ccl17 contains six potential AHREIs, and no AHREII or RelBAHREs (Fig. 4C–E). In contrast, the upstream region of Cd209a did not contain any of these three known AHRE sequences (Fig. 4C–E). The mapping of each AhR-binding motif for all DEGs can be found in Supplemental Table III.
Given that conditional AhrΔVav1 mice indicated that the AhR was necessary for the downregulation of CD209a and CCL17, we next sought to determine whether these changes were mediated through AhR’s DNA-binding domain (DBD). To accomplish this, we used mice that express an AhR protein that binds ligands and translocates to the nucleus but lacks a functional DBD (Ahrdbd mice) (76). Consistent with prior reports, Ahrdbd mice do not respond to TCDD treatment by upregulating Cyp1a1 (Fig. 5A) (76). When exposed to TCDD, however, Ahrdbd mutant mice demonstrated significant reductions in the percentage of CD209a+DCs (Fig. 5B) and CCL17 levels (Fig. 5C), changes that are similar to observations in WT mice. These findings suggest that AhR does not regulate Ccl17 and CD209a via direct binding to DNA via its known DBD. Therefore, our results suggest AhR likely controls Ccl17 and Cd209a via cross-talk with other mechanisms, such as via interactions with other transcriptional regulators and signaling molecules.
The DBD of AhR is not required for regulation of Cd209a and Ccl17.
(A) Spleens from WT B6 or Ahrdbd mice treated with either vehicle control (V) or TCDD (T) 1 d prior to IAV infection were isolated, and cyp1a1 gene expression was measured by qRT-PCR. (B) Expression of CD209a on lung DCs from IAV-infected Ahrdbd mice was determined using flow cytometry. (C) The level of CCL17 in lung homogenates was measured using ELISA. (D) Schematic diagram of the mouse Ccl17 upstream regulatory region. Presumptive AHREs were mapped to the Ccl17 upstream promoter sequence along with proximal TF binding sites (within 20 bp of >1 different AHRE) were identified using PROMO software. NF-κB family member binding sites, regardless of AHRE proximity are also shown. (E) The upstream regulatory region of murine Cd209a does not contain any AHREs. TF binding sites for C/EBPβ, c-Fos, HES-1, and HOXA5 were ubiquitous throughout the regulatory region and excluded from the diagram. For (A)–(C), these data are representative of three independent experiments. For data in (A)–(C), there were five to six mice per treatment per experiment. The data represent the average ± SEM for each group. *p < 0.05, **p <0.01 by Student t test.
The DBD of AhR is not required for regulation of Cd209a and Ccl17.
(A) Spleens from WT B6 or Ahrdbd mice treated with either vehicle control (V) or TCDD (T) 1 d prior to IAV infection were isolated, and cyp1a1 gene expression was measured by qRT-PCR. (B) Expression of CD209a on lung DCs from IAV-infected Ahrdbd mice was determined using flow cytometry. (C) The level of CCL17 in lung homogenates was measured using ELISA. (D) Schematic diagram of the mouse Ccl17 upstream regulatory region. Presumptive AHREs were mapped to the Ccl17 upstream promoter sequence along with proximal TF binding sites (within 20 bp of >1 different AHRE) were identified using PROMO software. NF-κB family member binding sites, regardless of AHRE proximity are also shown. (E) The upstream regulatory region of murine Cd209a does not contain any AHREs. TF binding sites for C/EBPβ, c-Fos, HES-1, and HOXA5 were ubiquitous throughout the regulatory region and excluded from the diagram. For (A)–(C), these data are representative of three independent experiments. For data in (A)–(C), there were five to six mice per treatment per experiment. The data represent the average ± SEM for each group. *p < 0.05, **p <0.01 by Student t test.
Others have shown that AhR can form complexes with other TFs to influence gene expression (77). However, the full extent by which AhR interacts with other TFs remains unclear. To identify other TFs that may interact with AhR to impact expression of Ccl117 and Cd209a, we used the RSAT tool to identify TF binding sites throughout a 5-kb upstream region these genes. RSAT identified six canonical AHREs in the upstream regulatory region of the mouse Ccl17 gene (Fig. 5D). Sequences exported from RSAT were input into PROMO, which uses positional weight matrices to identify TF binding sites. In total, binding sites for 62 TFs were identified by PROMO analysis (Supplemental Table III). To predict whether AhR may potentially partner with these TFs, we filtered the analysis to identify TFs that were found in upstream regulatory regions of both Ccl17 and Cd209a, have binding motifs within 20 bp of at least two different presumptive AHREs in the upstream regulatory region, and were expressed in DCs with and without AhR activation. This resulted in seven candidate TFs with which AhR could potentially interact to regulate Ccl17 (Fig. 5D): C/EBPβ, c-Fos, HES-1, HOXA5, MyoD, NF-1, and TCF-1. Given that there is evidence that AhR can partner with NF-κB family members to regulate gene expression (28, 36, 78, 79), we identified binding sites for c-Rel, RelA, NF-κB, NF-κB1. However, no binding sites for any of these NF-κB family members were identified within close proximity (20 bp) of a presumptive AHRE in the mouse Ccl17 gene (Supplemental Table III). There is also evidence that AhR signaling induces transactivation of estrogen receptor α (ERα) (80); however, neither Ccl17 nor Cd209a have ERα binding sites in their upstream regulatory regions (Supplemental Table III). Consistent with sequence-based mapping, analysis of the upstream regulatory region of Cd209a by RSAT did not reveal the presence of any AHREs. However, PROMO analysis yielded binding sites for 55 other TFs. These include the same eight TFs identified upstream of the Ccl17 gene that have binding sites proximal to the AHREIs, as well as binding sites for NF-κB family members (Fig. 5E). Thus, one avenue by which AhR may regulate Ccl17 and Cd209a gene expression in DCs is through interaction with one or more TFs, overcoming the need for AhR to directly recognize its DNA-binding motif within the upstream regulatory region.
AhR activation modulates gene expression in human monocyte-derived dendritic cells
We expanded our investigation to determine whether AhR activation caused similar changes in Ccl17 and Cd209a expression in human primary DCs. AhR activation upregulated the AhR target gene Cyp1a1 (Fig. 6A) and induced Ido1 (Fig. 6B) in human DCs derived from peripheral blood monocytes from ungendered healthy donors (moDCs). The majority of DCs in the mouse lung 3 d after IAV infection are cDC2 and monocyte derived (4, 81–83). Although it is challenging to study precisely the same DC populations from mouse and human, cDC2s have similarities to human moDCs (84). AhR activation dampened the expression of Ccl17 in human DCs from several donor samples but not in others (Fig. 6C). Thus, Ccl17 expression did not demonstrate a consistent pattern of change in human moDCs. In contrast, in these same human moDCs, AhR activation consistently diminished Cd209a expression in all donor samples tested (Fig. 6D). We next analyzed the upstream regulatory region of the human Ccl17 and Cd209a genes to determine whether they contained AHREs and to define nearby TF binding motifs. Similar to mouse, human Ccl17 contained six presumptive AHREI sequences in its upstream regulatory region (Fig. 6E). In contrast to mouse, human Cd209a contained five possible AHREI binding sites in its upstream regulatory region (Fig. 6F). Analogous to the analysis of mouse genes, proximal TF binding sites in the upstream regulatory regions of human Ccl17 and Cd209a were identified using PROMO. In total, 21 different TFs were identified by PROMO with close proximity to the AHREs in upstream regulatory regions of human Ccl17 and Cd209a. By our search criteria, 12 different TFs of interest were identified within the Ccl17 upstream regulatory region (Fig. 6E). Within the Cd209a upstream region, 17 different TFs had close proximity to more than one AHRE (Fig. 6F). Significant overlap between the two sets of TFs was observed, and 10 TFs were identified in the regulatory regions of both genes (AP-2α, C/EBPβ, ENKTF-1, Foxp3, GR-α, GR-β, NF-κB, NF-κB1, TFII-I, RelA). These findings demonstrate that AhR activation affects human Cd209a expression but may not affect Ccl17 expression in DCs, although the mechanism of regulation may be significantly different in humans compared with the mouse.
AhR activation alters gene expression in human monocyte-derived DCs.
Immature human monocyte-derived DCs were cultured in the presence of DMSO (V) or 10 nM TCDD (T) and exposed to IAV for 24 h. RNA was isolated 24 h postinfection, and qRT-PCR was performed to quantify gene expression of (A) Cyp1a1, (B) Ido1, (C) Ccl17, and (D) Cd209a. Lines connect vehicle- and TCDD-treated samples from same donor pool. Presumptive AHREs were mapped to the (E) Ccl17 and (F) Cd209a upstream regulatory region, and TF binding sites proximal (within 20 bp) to at least two different AHRE were identified using PROMO software. NF-κB family member binding sites, regardless of AHRE proximity are shown. For (A)–(D), a total of four to five donor pools were used per experiment.
AhR activation alters gene expression in human monocyte-derived DCs.
Immature human monocyte-derived DCs were cultured in the presence of DMSO (V) or 10 nM TCDD (T) and exposed to IAV for 24 h. RNA was isolated 24 h postinfection, and qRT-PCR was performed to quantify gene expression of (A) Cyp1a1, (B) Ido1, (C) Ccl17, and (D) Cd209a. Lines connect vehicle- and TCDD-treated samples from same donor pool. Presumptive AHREs were mapped to the (E) Ccl17 and (F) Cd209a upstream regulatory region, and TF binding sites proximal (within 20 bp) to at least two different AHRE were identified using PROMO software. NF-κB family member binding sites, regardless of AHRE proximity are shown. For (A)–(D), a total of four to five donor pools were used per experiment.
Discussion
The respiratory tract has complex mechanisms to protect the host from infection. Yet, these defenses are susceptible to modification by environmental factors, which could increase or decrease pathophysiological and clinical sequelae of infection. Better defining how environmental signals modulate critical defense mechanisms against common respiratory pathogens is necessary to design new and improved strategies to prevent and treat respiratory tract infections. The environment sensing TF AhR is a potent regulator of immune cell function. Yet, the scope of genes and signaling networks that AhR modulates within specific immune cell types are not fully defined. In the current study, we used a signature AhR agonist to interrogate changes in the DC transcriptome during acute primary IAV infection. We also used publicly available databases to map potential AhR-binding sites in all differentially regulated genes. In addition to genes known to be changed by AhR activation, we identified new genes and pathways that are modulated following in vivo AhR activation. We further explored how AhR affects the expression of two of the most downregulated genes among this dataset and compared expression changes in mouse and human DCs. Considered together, these new findings indicate that the AhR regulates an integrated network of pathways in DCs and influences antiviral immune responses by altering the ability of DCs to downregulate receptors and lymphocyte-attracting chemokines used to attract and retain naive T cells.
In the work reported in this study, AhR activation significantly blunted the infection-associated rise in CCL17 levels. Although CCL17 is less studied than other chemokines, several reports indicate that it plays an important role in T cell migration and retention (85). This suggests that AhR-mediated changes in CCL17 could contribute to the ways via which AhR signaling in DCs regulates antiviral T cell responses to infection. It is primarily produced by DCs in the lung, although there are reports of CCL17 expression by epithelial cells following exposure to the respiratory syncytial virus (69). Thus, lower levels of CCL17 would dampen the ability of DCs to recruit and retain naive T cells, which would in turn lead to reduced activation of naive, virus-specific T cells. Consistent with this idea, in the context of IAV infection AhR activation significantly impairs the ability of DCs to activate virus-specific CD8+ T cells via a mechanism that is in part because of to AhR-mediated events in DCs (25). This is likely not an effect that is unique to the prototype AhR agonist TCDD, as other AhR ligands cause similar changes in T cell responses to IAV infection (37). Also, soybean tar extract, which contains AhR agonists, dampened CCL17 production by mouse bone marrow–derived DCs, although it is not certain that this occurred in an AhR-dependent mechanism (86). Furthermore, exposure to other AhR ligands, such as diesel exhaust particles or tryptophan metabolites, impacts the expression of other chemokines in the CNS or lung-derived cells (71, 87, 88).
Very few studies have defined what controls expression of CCL17 during infection. Takemura et al. (86) suggest a role for IL-4/STAT6 in driving Ccl17 expression in DCs. Although the mechanism is not yet known, paracrine activation of DCs by other immune cells, such as NKT cells, has also been shown to enhance CCL17 production (70). This suggests that other immune cells may license DCs to produce chemokines. Another way in which AhR may influence CCL17 levels is via IL-12. DCs produce IL-12 in response to recognition of viral Ags, and IL-12 from DCs augments CCL17 production, which bolsters DC–CD8+ T cell interactions (89–91). We have previously shown that IL-12 levels in the IAV-infected lung are significantly reduced by AhR activation (24). Human moDCs treated with TCDD or other AhR agonists, such as indirubin or indolo[3,2-b]barbazole, also demonstrated reduced IL-12 (92). Moreover, the loss of AhR enhanced the levels of IL-12 production by splenocytes in vitro, providing further evidence that AhR regulates IL-12 (93). It is therefore plausible that AhR activation attenuates CCL17 via a mechanism that involves altered signaling in DCs, including changes in IL-12 and IL-12–regulated pathways, which in turn influence the ability of DCs recruit and retain naive T cells. Thus, there is an accumulation of evidence that AhR contributes to regulation of chemokine expression, including CCL17, via multiple mechanisms.
In addition to regulating chemokine levels, AhR signaling downmodulated the expression of the CD209 family of genes in DCs. CD209a/DC-SIGN binds high mannose glycoproteins present on the surface of an array of viruses, including IAV (94–97). CD209a is also involved in trafficking bound viral Ags to the proteasome for degradation and cross-presentation (98). Along with pathogen uptake and Ag presentation, it also plays a role in promoting the adhesion of DCs to naive T cells (62). The majority of CD209a studies have focused on delineating downstream functions after CD209a is engaged (99). Consequently, less is known about how CD209a expression and that of related family members is regulated. Our studies suggest that engagement of AhR, perhaps in combination with other signaling events triggered by infection, influences CD209a expression by DCs. The most pronounced changes were observed in the cDC2 subset. This could reflect differences in the maturation or egress of the different DC subsets from the infected lung to the draining lymph nodes (4, 82, 83, 100). In support of this idea, both the “cellular movement” and “immune cell–trafficking” pathways were two of the most influenced pathways within the DC transcriptome following AhR activation (Fig. 1F). Thus, changes in CD209a may reflect a shift in the DC populations in the infected lung. Related to this population shift in the lung, AhR activation may also influence CD209a expression in distinct DC subsets at different times during the response to infection. Given that there are likely differences in the transcriptomes of cDC1s and cDC2s (101, 102), AhR may regulate CD209a expression in a subset-specific manner. Although there are not differences in the expression of Ahr in cDC1 and cDC2 (4), this could reflect differences in transcriptional regulators with which AhR interacts. In addition to cross-talk with other transcriptional regulators, CD209a levels could be regulated by AhR via an indirect mechanism, such as via AHR-mediated changes in IDO1. AhR activation induces Ido1 expression in DCs during IAV infection (4), and the AhR-IDO1 pathway has been implicated in diminished DC-SIGN expression levels in LPS-stimulated human DCs (103). While examining how AhR signaling influences CD209a expression, it is important to bear in mind that CD209a is one member of a larger gene family; thus, AhR likely also regulates the expression of other C-type lectin receptor genes.
The spectrum of AhR target genes varies by cell type, tissue, disease state, and ligand (37, 104–106). Regarding the latter, humans are exposed daily to a wide variety of AhR ligands beyond TCDD through dietary intake and exposure to different pollutants (18). We chose to expose mice to TCDD has a representative ligand that binds solely to AhR and does not induce off-target effects that other ligands do. Further investigation of other known AhR ligands or chemical admixtures on the impact of immune function are underway and are of continued interest to our group. In this study, we identified 631 genes which were upregulated or downregulated by AhR activation (Supplemental Table I). Many of these genes are involved in immune function and contain AHREs in their upstream regulatory region. By pathways analysis, the most impactful changes were found to be immune cell trafficking and movement of DCs and cell-to-cell signaling (Fig. 1F). Although genes such as Cyp1a1 and Ido1 are consistently induced in a broad spectrum of mouse and human cells, AhR may regulate expression of other genes through distinct mechanisms. The identification of presumptive AHREs across all DEGs provides new insight as to how AhR could influence gene expression and suggests that some AhR-dependent changes reflect cross-talk between AhR and other context-associated signals. For example, mouse Cd209a has no AHRE, whereas human Cd209a gene has seven AHREI sites; yet, gene expression is dampened in mouse and human DCs upon AhR activation with the same agonist. A simplistic interpretation is that there are differences in how the AhR regulates gene expression in mouse versus human cells. Yet another explanation is that the AhR induces and dampens gene expression via multiple mechanisms. For instance, AhR induces Cyp1a1 via binding AHREI sites (107); yet how AhR regulates expression of many other genes remains to be determined. The majority of DEGs in mouse DCs contained binding motifs for the canonical AHRE. Other studies have analyzed the mouse genome for the presence of canonical AHREs (108) and used a genome-wide approach to identify novel AhR-binding motifs (i.e., AHREII) (73). Other recent studies have identified AhR-specific gene targets and AhR-binding sites in B cells and hematopoietic stem cells (109, 110). These other studies did not include DCs and used a narrower definition of the promoter region (−1500 to +500 bp of the transcription start site) (73, 107, 109, 110). Thus, our understanding of which genes are shared across cell types as direct AhR target genes, modulated by AhR–AHRE interactions, and which DEGs reflect indirect effects, because of cross-talk with other TFs and pathways is somewhat limited. Integration of information from genome-wide analyses with emerging findings using conditional AhR knockout mice and AhR–DBD mutant mice will further distinguish between direct changes that are a result of AhR–DNA binding versus those caused by AhR-mediated changes in cellular signaling.
An unanticipated finding from these studies was that AhR is necessary but not sufficient to diminish CCL17 and CD209a levels during infection. This is in contrast to established AhR-regulated genes, such as Cyp1a1, for which AhR is necessary and sufficient to induce (76). This suggests that AhR regulates expression of these genes via interaction with other transcriptional regulators or via other gene regulatory mechanisms. AhR has been shown to interact with other TFs, such as NF-κB, ERα, and AP-1 (36, 111–114). Models of TF binding in the Ccl17 upstream regulatory region did not predict sites for these TFs in proximity to AHRE. Likewise, STAT6 binding sites were present, but were not in close proximity to any of the presumptive AHREs identified. Thus, if AhR is interacting with STAT6 or other STAT family members downstream of IL-12, its likely to occur upstream of DNA binding. Whereas a limitation of using computational approaches is that they do not prove occupancy, we identified seven additional TFs with which AhR may partner to regulate Ccl17 expression in DCs. In contrast, analysis of the upstream regulatory region of Cd209a suggests that AhR may regulate its expression, at least in mouse cells, through an indirect mechanism. For example, although not studied in DCs per se, AhR activation influences gene expression via epigenetic mechanisms, including histone acetylation/deacetylation and DNA methylation (42, 115, 116). In contrast, whether AhR directly or indirectly regulates Cd209a in human DCs is less clear because our analysis of the human genome identified multiple AHREs upstream of the human Cd209a transcription start site. Thus, AhR may influence the DC transcriptome via multiple mechanisms that include direct AhR binding to DNA, interactions with other TFs, and epigenetic regulatory machinery, and these may be the same or different in mouse and human.
Although the focus of the research presented in this study is to better understand how AhR activation affects DCs during infection, the changes in gene expression are likely relevant to AhR regulation of DCs more broadly. For instance, elevation of Ido1 and Ido2 are consistently observed in mouse and human DCs treated with AhR agonists in vitro (26, 28, 35, 117) and in vivo (31, 34). Cd274 expression was significantly elevated in this current study, and PDL1 (CD274) levels were also increased in human DCs via a pathway that involves AhR signaling (103). In contrast, genes for several common costimulatory molecules and cytokines, such as CD40L, CD54, CD80, CD86, and IL-6, which have been reported to be modulated on DCs treated with AhR ligands in vitro (4, 28, 31), were not among the DEGs in DCs isolated from IAV-infected mice. This is consistent with the idea that some of the downstream effects of AhR agonism or antagonism are species, organ, cell type, and even immune state specific. Therefore, that we found differences in AHRE number, distribution, and proximity to other TF binding sites in human and mouse genes for Ccl17 and Cd209a is not surprising. Likewise, there are differences TCDD binding affinity between human and mouse AhR that may contribute to differences in regulatory targets, even within the same cell type (118). This highlights the need for more research that directly compares AhR regulation of genes and functions in human and mouse leukocytes within the same immune challenge.
In the current study, we showed the extent by which AhR activation during acute primary IAV infection modifies gene expression in DCs, revealing several new gene targets which likely contribute to overall ways in which AhR influences antiviral defense mechanisms. Considering that several reports show that populations exposed to AhR-binding environmental chemicals, including polychlorinated biphenyls and other DLCs, have increased prevalence or severity of infection (9, 11, 12, 119), the present findings predict that some AhR ligands will influence human DC functions and modulate their response to viruses. Whether and how the myriad AhR ligands that have been discovered have distinct or similar effects on gene expression in DCs will be important to analyze. Likewise, strategic evaluation of the gene expression changes triggered by AhR signaling in response to IAV occur during other immune challenges will better inform predictive analytics, including the development of biomarkers for AhR modulation. A better understanding of how AhR contributes to antiviral defenses will be helpful to developing therapeutic tools that leverage the immunomodulatory power of AhR and to public health strategies aimed at better protections from viral infections.
Acknowledgements
We thank the Integrated Health Sciences Facility Core of the Rochester Environmental Health Science Center for support of this project and Dr. Allison Ehrlich and Dr. Americo Lopez-Yglesias for thoughtful comments during the preparation of this manuscript. We also thank Dr. Timothy P. Bushnell and the Flow Cytometry Core for assistance with cell sorting and Dr. John Ashton and the team at the University of Rochester Genomics Research Core for processing samples and RNA sequencing.
Footnotes
This work was supported by Grants R01-ES0004862, R01-ES023260, T32-ES07026, and P30-ES01247 (to B.P.L.) and R01-ES013784 (to D.M.S.) from the National Institutes of Health.
The sequences presented in this article have been submitted to the Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE123996) under accession number GSE123996.
The online version of this article contains supplemental material.
Abbreviations used in this article:
- AhR
aryl hydrocarbon receptor
- AHRE
aryl hydrocarbon responsive element
- cDC1
conventional DC subset 1
- cDC2
conventional DC subset 2
- DBD
DNA-binding domain
- DC
dendritic cell
- DC-SIGN
DC-specific ICAM 3–grabbing nonintegrin
- DEG
differentially expressed gene
- DLC
dioxin-like chemical
- ERα
estrogen receptor α
- IAV
influenza A virus
- MFI
mean fluorescence intensity
- MHC II
MHC class II
- MLN
mediastinal lymph node
- moDC
DC derived from peripheral blood monocytes from ungendered healthy donor
- PCA
principal component analysis
- qRT-PCR
quantitative real-time RT-PCR
- RNA-Seq
RNA sequencing
- RSAT
Regulatory Sequence Analysis Tools program
- TCDD
2,3,7,8-tetrachlorodibenzo-p-dioxin
- TF
transcription factor
- WT
wild-type.
References
Disclosures
The authors have no financial conflicts of interest.