Timing influences the impact of aryl hydrocarbon receptor activation on the humoral immune response to respiratory viral infection

Abstract

Humoral responses to respiratory viruses, such as influenza viruses, develop over time and are central to protection from repeated infection with the same or similar viruses. Epidemiological and experimental studies have linked exposures to environmental contaminants that bind the aryl hydrocarbon receptor (AHR) with modulated antibody responses to pathogenic microorganisms and common vaccinations. Other studies have prompted investigation into the potential therapeutic applications of compounds that activate AHR. Herein, using two different AHR ligands [2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) and 2-(1H-Indol-3-ylcarbonyl)-4-thiazolecarboxylic acid methyl ester (ITE), to modulate the duration of AHR activity, we show that the humoral response to viral infection is dependent upon the duration and timing of AHR signaling, and that different cellular elements of the response have different sensitivities. When AHR activation was initiated prior to infection with influenza A virus, there was suppression of all measured elements of the humoral response (i.e., the frequency of T follicular helper cells, germinal center B cells, plasma cells, and circulating virus-specific antibody). However, when the timing of AHR activation was adjusted to either early (days −1 to +5 relative to infection) or later (days +5 onwards), then AHR activation affected different aspects of the overall humoral response. These findings highlight the importance of considering the timing of AHR activation in relation to triggering an immune response, particularly when targeting the AHR to manipulate disease processes.

1. Introduction

The production of antigen-specific antibodies is among our greatest defenses against infectious diseases. Underpinning antibody production is the humoral immune response, which is comprised of the interplay between and responses of CD4⁺ T cells and B cells. Furthermore, a specific sub-type of CD4⁺ T cells, called T follicular helper (Tfh) cells, is necessary for humoral immunity (1, 2). Tfh cells establish and maintain germinal centers, which are the main sites where B cells undergo isotype switching, mature into antibody secreting plasma cells, and form memory B cells (1–3). Thus, extrinsic factors that dampen or impede humoral immune responses have the potential for tremendous impact on public health.

This is not a theoretical concept: there is a high degree of variability in humoral responses to infection and vaccination that cannot be accounted for by intrinsic factors, such as host or pathogen genetics (4–7). Sources of extrinsic factors are highly diverse, and include substances from consumer goods, medicines, and pollution (8). Among these many sources, small molecules that bind to the aryl hydrocarbon receptor (AHR) have been consistently associated with altered, generally suppressed, humoral responses (9–18). The AHR is a ligand-regulated transcription factor that is expressed throughout the immune system, including in T and B lymphocytes (19). Among the more heavily studied AHR agonists are pollutants, such as polycyclic aromatic hydrocarbons derived from industrial and combustion processes (20, 21). Human cohort studies have associated exposure to pollutants containing AHR ligands with dampened antibody responses to vaccination (9, 11). Studies in mice align with this pattern, and have shown that AHR activation with dioxin 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), or with dioxin-like polychlorinated biphenyls (PCB, e.g., PCB126) generally diminish antibody levels (12, 13, 15, 18). TCDD is the prototype AHR ligand to study AHR biology for many reasons, including that it acts exclusively via the AHR, binds AHR with high affinity (K^D ~0.48 nM), and does not break down into downstream metabolites (22). In addition to pollutants, many other synthetic and naturally derived small molecules bind AHR, including metabolites and derivatives of naturally derived substances, such as tryptophan and certain indoles (23–25). Moreover, some pharmaceuticals bind to AHR (23–25), and there is burgeoning interest in targeting AHR to modulate disease course or attenuate disease severity. However, a key concern using AHR agonists therapeutically is unintended repression of humoral immune responses.

Prior studies comparing immune modulation by AHR agonists suggest that a range of factors likely contributes to the ability of AHR ligands to modulate adaptive immune responses; however, the majority of these studies focused on T cells and cell-mediated immune responses (14, 26–28). For instance, modulation of the response of CD8⁺ T cells to virus infection or allogeneic tumor cells required that AHR activation is initiated either prior to or no later than the first 4–5 days after antigen challenge (17, 29). Other prior studies indicated that AHR-driven changes to anti-viral humoral responses could begin to be detected as early as 5 days post-infection (18). However, in most of these studies, AHR activation was initiated prior to immune challenge. This limits the ability to understand whether humoral responses are similarly sensitive to the timing of AHR activation relative to antigenic challenge. Yet, this issue has tremendous relevance to the potential use of AHR agonists as therapeutic agents. Deciphering this also sharpens understanding of whether there are particularly sensitive windows of time during infection in which exposure to environmentally derived AHR agonists are more likely to derail antibody production.

The research reported herein evaluated the impact of the timing of AHR activation on humoral responses during respiratory viral infection using two well-known AHR ligands: the canonical agonist 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) and the tryptophan derivative 2-(1H-Indol-3-ylcarbonyl)-4-thiazolecarboxylic acid methyl ester (ITE) and a common respiratory pathogen: influenza A virus (IAV). Altogether, the findings reported provide evidence that AHR activation modulates humoral immunity via events that occur early as well as later after infection, which collectively influence the antiviral antibody response generated.

2. Materials and Methods

Animals and Treatments

C57Bl/6 (B6) mice (age 5–6 weeks) were purchased from The Jackson Laboratory (Bar Harbor, ME). All mice were housed in microisolator cages in a specific-pathogen free facility, and were provided food (LabDiet 5010, St. Louis MO) and water ad libitum. The animals were housed on a 12 h light/dark cycle, and ambient temperatures were between 20–22°C. Upon receipt from supplier, mice were randomly housed with 3–5 mice per cage. All data presented are from female mice that were 6–12 weeks of age at the time of experiments, and each point in time relative to infection there were 4–8 mice in each treatment group. Since this paper reports the results of multiple independent experiments, the number of mice used within an experiment, and the number of mice within each group is provided within the figure legends.

2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD; ≥99% purity, Cambridge Isotope Laboratories, Tewksbury, MA) was dissolved in anisole and diluted in food grade peanut oil (Wegmans, Rochester, NY). Mice were randomly assigned to treatment groups. Mice that were administered the vehicle control and intranasally infected with IAV are described as “control infected mice” and this group is designated as Vehicle or Veh in figures. For studies using TCDD, mice were administered a single oral dose of either TCDD (10 μg/kg body weight) or the vehicle control (peanut oil containing 0.1% anisole). TCDD was either administered one day prior to infection (D-1) or five days after infection (D+5). TCDD was selected because it is the prototype synthetic AHR ligand. Key features that make it an excellent tool compound for studying AHR include that it is an AHR-specific agonist that does not generate secondary metabolites (30–33), and it sustains AHR activation in vivo for at least 2 weeks in mice (34–36). Moreover, TCDD represents a large class of environmentally derived AHR ligands to which humans are regularly exposed (37, 38). The administered dose used is relevant to human exposure to this broad class of chemicals (38, 39), is not overtly toxic, is immunomodulatory (14, 18, 40) and the effects of TCDD treatment are not observed in AHR deficient mice (14, 31, 33).

For some experiments, mice were treated with 2-(1H-Indol-3-ylcarbonyl)-4-thiazolecarboxylic acid methyl ester (ITE; Tocris Bioscience, Minneapolis, MN). ITE is an indole-based AHR ligand (14, 25, 41). Current evidence indicates that ITE can be derived from condensation reactions between tryptophan and cysteine, or from the conversion of glucobrassicins, which are derived from foods such as cruciferous vegetables (42–44). ITE was dissolved in acetone and diluted in peanut oil, and acetone was evaporated using a nitrogen evaporator. Mice were injected intraperitoneally (i.p.) with 10 mg ITE/kg body weight or the peanut oil acetone vehicle control (14). Mice were dosed with ITE once, or daily starting one day prior to infection and continuing until either 5 or 9 days after infection. ITE is also immunomodulatory (14, 18) and the effects of ITE treatment are observed in global AHR deficient mice (14).

For TCDD and ITE treatments, the dosing period used within a particular experiment is indicated in the figure legend. Experimental work was planned and performed following ARRIVE guidelines for reporting in vivo experimental research. All animal treatments had prior approval of the Institutional Animal Care and Use Committee (IACUC) of the University of Rochester (approval number 2006–078/100101). The University is accredited through the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC). Animals were treated humanely and with due consideration to alleviate any distress and discomfort. All U.S. Public Health Service Policy on Human Care and Use of Laboratory Animals guidelines for the handling of vertebrate animals were followed.

Influenza A virus

Influenza A virus (IAV) strain A/HKx31 (HKx31; H3N2) was propagated in embryonated chicken eggs, according to previously described method (12). Briefly, fertilized eggs (Charles River, Wilmington, MA) were inoculated with 0.05 hemagglutinating units (HAU) IAV in 100 μL of Hank’s balanced salt solution (HBSS), containing 10 mM HEPES. Inoculated eggs were incubated for 48 h at 37°C, followed by refrigeration overnight at 4°C. Allantoic fluid was harvested under aseptic conditions, centrifuged (300 xg, 10 min, 4°C), aliquoted, snap frozen using liquid nitrogen, and then stored at −80°C. Titer was determined by hemagglutination of avian erythrocytes (12). Mice were anesthetized by i.p. injection of avertin (2,2,2-tribromoethanol; Sigma Aldrich, St. Louis, MO), and infected intranasally (i.n.) with a sublethal dose (120 HAU virus in 25 μL) diluted in endotoxin-tested PBS (8, 10). Infections were performed in the morning (between 8:00 and 10:00 AM). All work with infectious agents was conducted with prior approval of the Institutional Biosafety Committee of the University of Rochester, following guidelines of the NIH/CDC.

Tissue collection and preparation of immune cells

Tissue collection was initiated in the morning. Mice were sacrificed using lethal dose of anesthetic (e.g., Euthasol, Virbac, Greeley, CO) or cervical dislocation and exsanguination. Blood was collected by cardiac puncture, with syringes containing 100 μL of heparin sodium salt (250 U/mL, Grade I-A, ≥180 USP units/mg; Sigma Aldrich), and plasma was stored at −80 °C. Immune cells were isolated from the mediastinal lymph nodes (MLN). Briefly, single-cell suspensions were prepared by pressing lymph nodes from a single mouse between the frosted ends of two microscope slides. MLN cells were suspended in cold HBSS containing 2.5% FBS (Hyclone, Logan, UT), and cells were collected using centrifugation (200 x g; 6 min). Erythrocytes were removed by incubating pelleted cells in a solution containing 0.15 M NH₄Cl, 10 mM NaHCO₃, and 1 mM EDTA for 5 min at room temperature. Erythrocyte lysis was terminated by adding equal volume of cold HBSS containing 2.5% FBS. Cell suspensions were then passed through a 70 μm cell strainer. After centrifugation (200 x g; 6 min), cell pellets were re-suspended in cold HBSS, 2.5% FBS for enumeration using a TC10 automated cell counter (BioRad, Hercules, CA), or a hemocytometer and Trypan blue dye. Cells were maintained at 4°C until labeling with fluorescent-conjugated antibodies, which was performed the same day.

Flow cytometry

Single-cell suspensions containing 2 x 10⁶ lymph node cells were incubated with anti-mouse CD16/32 monoclonal antibody (mAb; clone93) for 10 min at 4º C to block non-specific binding. Cells were then incubated for 20 min at 4º C with fluorochrome-conjugated mAb against extracellular antigens. Table 1 indicates the specific mAb (clone), fluorochrome, vendor and amount of each reagent used. After cell surface labeling, cells were fixed using 2% formaldehyde in PBS for 20 min and analyzed directly by flow cytometry. To identify live cells, prior to extracellular staining, cells were incubated for 30 min in PBS at 4ºC with fixable viability dye eFluor 506 (eBioscience, Cat# 65–0866-14). Fluorescence minus one (FMO) controls were used to determine non-specific fluorescence and define gating parameters. Data (500,000 events per sample) were collected using an LSRII flow cytometer (BD Biosciences). Data were analyzed using the FlowJo software program (Version 10; Tree Star, Ashland, OR). Tfh cells were defined as CD4⁺ T cells that were CD44^hiCXCR5^hiPD1^hi (2, 18). T follicular regulatory (Tfr) cells were defined as CD4⁺CD44^hiCXCR5^hiPD1^hiFOXP3⁺ cells (18, 45). Germinal center (GC) B cells were defined as CD3^-B220⁺CD95⁺GL7⁺ cells (46). Plasma cells were identified as CD3^-B220⁺CD138^hi cells (2). Gating strategies and FMO controls can be found in Supplementary Figure 1.

Table 1:

Antibodies used for flow cytometry

Reagent	Clone	Vendor	Catalog No.	Amount (μg) per 2 x 106 cells
Anti-CD4 PerCP-Cy5.5	RM4–5	Invitrogen	45–0042-82	0.128
Anti-CD44 BV711	IM7	BD Biosciences	563971	0.064
Anti-CXCR5-biotin	2G8	BD Biosciences	551960	2.5
Streptavidin PE	n.a.	BD Biosciences	554061	0.16
Anti-PD1 FITC	J43	Invitrogen	11–9985-82	0.625
Anti-CD3 PE	145-2C11	Biolegend	100206	0.50
Anti-B220 FITC	RA3-6B2	Invitrogen	11–0452-82	0.32
Anti-CD95 PE-Cy7	J02	BD Biosciences	557653	0.50
Anti-GL7 ef450	GL7	Invitrogen	45–5902-82	0.25
Anti-CD138 BV711	S81–2	Biolegend	142519	0.064
Anti-FOXP3 af700	FJK-165	Invitrogen	56–5773-82	1.0

Anti-influenza virus antibody ELISA

Relative levels of IAV-specific antibodies in plasma were measured using enzyme linked immunosorbent assays (ELISA), as previously described (13, 14). In summary, 96-well ELISA microplates (MICROLON^®, flat-bottom, Greiner Bio-one, Monroe, NC) were coated with purified HKx31 IAV (5 μg/mL, Charles River, Wilmington, MA) by incubating plates at 37ºC overnight. The following day, plates were washed with PBS containing 0.05% Tween 20, and this same wash buffer was used in between each step. To reduce non-specific binding, PBS containing 5% BSA was added to all wells, and plates were incubated at 4ºC for 2 h. Plasma samples from infected or immunologically naïve mice were serially diluted in PBS 5% BSA from 1:100 to 1:25600. Diluted plasma (100 μL) was added to the plates and incubated overnight (16–20 h) at 4ºC. Biotinylated goat anti-mouse IgG2a antibodies (Southern Biotechnology) were diluted 1:5000 from commercial stock in PBS 5% BSA, and added to the plates (100 μL per well), which were incubated at room temperature for 45 min. Avidin-peroxidase (1:400 dilution in PBS 5% BSA; Sigma Aldrich) was added directly to the wells, and plates were incubated at room temperature for 30 min. 2,2′-Azino-bis(3-ethylbenzothiazoline-6-sulphonic acid) with 0.03% H₂O₂ was used to induce colorimetric change. Absorbance values were read at 405 nm using a SpectraMax Plate reader (Molecular Devices). To measure antibody affinity, 100 μL of PBS containing 1.0 M guanidine hydrochloride (GuHCl; Sigma Aldrich), 0.2% Tween-20 and 10 mg/mL BSA was added to samples, which were incubated for 15 min at room temperature prior to adding biotinylated isotype-specific antibodies (47).

RT-qPCR

For some experiments, RNA was isolated from the liver of mice using the QIAGEN RNeasy Mini Kit (Qiagen, Valencia, CA). Total RNA concentration was determined with a NanopDrop 1000 spectrophotometer (NanoDrop, Wilmington, DE), and RNA quality assessed with the Agilent Bioanalyzer 2100 (Agilent, Santa Clara, CA). RNA was reverse transcribed using the iScript cDNA Synthesis Kit (BioRad; Hercules, CA) and RT-qPCR performed with the iQ SYBER Green Supermix (BioRad). The following primers (IDT; San Diego, CA) were used: Cyp1a1: Forward primer, 5’-TTTGGAGCTGGGTTTGACAC-3’, Reverse primer, 5’-CTGCCAATCACTGTGTCTA-3’; L13: Forward primer, 5’-CTACAGTGAGATACCACACCAAG-3’, Reverse primer, 5’-TGGACTTGTTTCGCCTCCTC-3’. RT-qPCR was performed on a C1000 Touch Thermal Cycler (BioRad). The fold change in Cyp1a1 was calculated relative to vehicle control group from the same day and/or tissue and normalized to the reference gene L13 using the 2^−ΔΔCt method (48).

Statistical analysis

Data were analyzed using JMP Pro 15 software (SAS Institute Inc.). Differences between means of two treatment groups at a single point in time were evaluated using a Student’s t-test. A two-way ANOVA followed by a Tukey honest significant difference (HSD) test was used when comparing means between more than two groups. A one-way ANOVA followed by a Dunnett’s post hoc test was used to compare means to control (vehicle). Differences were considered statistically significant when p-values were ≤ 0.05. Error bars on all graphs represent the standard error of mean (SEM).

3. Results

Delaying AHR activation reduced the attenuation of the humoral response to infection.

Consistent with previous findings (14, 18), commencing AHR activation by administering TCDD one day prior to acute primary infection significantly dampened the antibody response (Figure 1A). In fact, in TCDD-treated mice, the level of virus-specific antibodies was not significantly different from immunologically naïve animals. When AHR activation was not initiated until the 5^th day after infection, the level of circulating virus-specific antibodies was higher than immunologically naïve mice and in mice in which AHR activation was started just prior to infection; however, levels were still significantly less than control infected mice (Figure 1B). Thus, delaying AHR activation until 5 days after infection still suppressed antibody levels, but to a lesser extent than when AHR was activated just prior to viral infection (Figure 1C).

Figure 1. AHR activation prior to or 5 days after infection diminished levels of circulating anti-viral antibodies, but not antibody avidity.

C57BL/6 mice (6 weeks old, female) were administered vehicle control or TCDD (10 μg/kg body weight) once by gavage one day prior to (TCDD D-1) or five days after (TCDD D+5) infection. Mice (7 per treatment group) were infected intranasally (i.n.) with 120 hemagglutinating units (HAU) influenza A virus (IAV, strain HKx31; H3N2). Blood was harvested on day 14 post-infection. As a negative control for infection, blood was collected from 3 immunologically naïve mice. (A-C) Anti-IAV antibody levels in serially diluted plasma (1:100 to 1:25600) were determined using ELISA. Line graphs show levels of IAV-specific IgG2a in serially diluted plasma from mice treated with (A) vehicle and TCDD prior to infection (D-1), or (B) vehicle and TCDD 5 days after infection (D+5). Closed circles indicate vehicle, open squares indicate TCDD D-1, closed squares indicate TCDD D+5, and triangles indicate uninfected (immunologically naïve) mice. (C) Bar graph shows the mean area under curve (AUC) for each group. Symbols within each bar represent data from individual mice 14 days after infection. The dashed line denotes AUC for plasma samples from naïve mice. (D-E) Line graphs depict levels IAV-specific IgG2a in serially diluted plasma (1:100 to 1:25600) that was either untreated (closed symbols) or treated with 1M GuHCl (open symbols). Samples from (D) vehicle treated mice are denoted with circles, and samples from (E) TCDD D+5 treated mice with squares. The p-values on each graph indicate a 2-way ANOVA (dilution series with and without chaotropic agent). (F) Bar graph shows the mean AUC for all plasma dilutions. The percentage on the graphs denotes the mean reduction due to 1M GuHCl treatment in each group. The horizontal dashed line indicates the AUC for all plasma dilutions of the TCDD D-1 treatment group. Different letters on the bar graph indicate means that were significantly different from one another (p≤ 0.05), whereas groups that have the same letter were not significantly different from each other (ANOVA, Tukey HSD). All error bars indicate the SEM. Data are representative of at least one experiment with 6–8 infected mice in each treatment group.

Antibody affinity maturation is another critical aspect of the antibody response (49). To determine whether delaying AHR activation affected the relative affinity of antibodies produced, we measured antibody avidity for IAV using a modified ELISA with a chaotropic agent (guanidine hydrochloride; GuHCl) (47). In plasma from infected mice given the vehicle control, treatment with GuHCl significantly reduced the level of IAV-specific antibodies across plasma dilutions (Figure 1D). GuHCl treatment also reduced the level of antiviral antibodies in plasma from mice given TCDD on day 5 post-infection (TCDD D+5) (Figure 1E). Overall, the effect of adding a chaotropic agent was similar in the vehicle and TCDD D+5 treatment groups (Figure 1F). In contrast, when AHR was activated one day before infection (TCDD D-1), antibody levels were very low (Figure 1C), and GuHCl treatment was unable to reduce further (Figure 1F, horizontal line). These findings indicate that although there is less antibody produced, starting AHR activation five days after infection did not appear to affect processes that drive antibodies to develop higher affinity.

Plasma cells are a key producer of IAV-specific antibodies (49). AHR activation prior to infection significantly reduced the percentage of plasma cells in mice by 6.7-fold compared to infected controls (Figure 2A–B). When AHR activation began on day 5 of infection, the percent of plasma cells was also significantly less than in IAV-infected controls (Figure 2C). Yet, the impact on the frequency of plasma cells was tempered (3.6-fold lower rather than 6.7-fold lower) when initiating AHR activation was delayed. The number of plasma cells was reduced when AHR activation was initiated before or starting on day 5 of infection (Figure 2D). Thus, similar to the relative level of circulating virus-specific antibodies, delaying ligand administration attenuated, but did not abolish, the impact of AHR activation on the frequency of plasma cells.

Figure 2. AHR activation prior to and five days after infection impacts the generation of plasma cells.

C57BL/6 mice (6 weeks old, female) were given vehicle control or TCDD (10 μg/kg body weight) once by gavage one day prior to (TCDD D-1) or five days after (TCDD D+5) infection. Mice were infected (i.n.) with 120 HAU of IAV. Blood and MLN were harvested on day 14 post-infection for analysis by ELISA and flow cytometry. (A-C) FACS plots depict plasma cells (CD3⁻B220⁺CD138^hi) from mice treated with (A) vehicle, (B) TCDD D-1, or (C) TCDD D+5. The number on each plot denotes the mean (±SEM) percentage of plasma cells out of total MLN cellularity. Prior gating excluded dead cells, debris and doublets. Figure S1 illustrates the gating strategy and FMO controls. Asterisk (*) denotes p≤ 0.05 compared to vehicle treatment group (ANOVA, Dunnett’s test). (D) Bar graph shows the mean number (± SEM) of plasma cells in the MLN. The p-values are a comparison between the TCDD treatment groups compared to vehicle control (ANOVA, Dunnett’s test). Data are representative of at least one independent experiment with 6–8 mice per treatment group.

Germinal centers (GC) are an important source of plasma cells, and a key site of affinity maturation (50). When AHR activation was initiated one day prior to infection, the percent of GC B cells was reduced 5.9-fold compared to infected controls (Figure 3A–B). In contrast, delaying the initiation of AHR activation until day 5 after infection no longer reduced the percent of GC B cells (Figure 3C). Yet, whether AHR was activated one day before or 5 days after infection, the number of GC B cells was significantly less than in infected controls (Figure 3D). Similar to anti-viral antibody levels, the overall magnitude of the effect was blunted when AHR activation was delayed, such that rather than 22.7-fold fewer number of GC B cells, there were only 2.9-fold less than mice in the control group. Collectively, these results indicate that AHR activation affects the production of GC B cells and plasma cells by influencing key events before and after the fifth day of viral infection.

Figure 3. Delayed initiation of AHR activation relative to onset of infection blunts impact on generation of germinal center B cells.

C57BL/6 mice were administered vehicle or TCDD once one day prior to or five days after IAV infection. MLN were harvested for analysis by flow cytometry on day 14 post-infection. (A-C) FACS plots show germinal center (GC) B cells (CD3⁻ B220⁺CD95⁺GL7⁺) from mice given (A) vehicle, (B) TCDD one day before (TCDD D-1), and (C) five days after (TCDD D+5) infection. Prior gating excluded dead cells, debris and doublets, and the gating strategy and FMO controls are in Figure S1. The number on each plot denotes the mean percentage of all MLN cells that are GC B cells. *Denotes p≤ 0.05 compared to vehicle treatment group (ANOVA, Dunnett’s test). (D) Bar graph shows the mean number of GC B cells in the MLN. Different letters indicate that the means were statistically different from one another (ANOVA, Tukey HSD). All error bars denote the SEM, and data are representative of one independent experiment with 6–8 mice per treatment group.

Tfh cells are another critical cellular component of humoral immune responses (2). When AHR activation was initiated prior to IAV infection there was a 10-fold reduction in the percent of Tfh cells compared to infected vehicle controls (Figure 4A–B). There was also a diminution in the percentage of Tfh cells when AHR activation was initiated on the 5^th day of infection (Figure 4C). The number of Tfh cells was also reduced in mice treated with TCDD one day before and or five days after (Figure 4D). Even though the two different TCDD treatment groups were not significantly different from one another (ANOVA, p=0.47), instead of being 10-fold fewer, the frequency of Tfh cells was about 3-fold lower than the vehicle control (Figure 4A–C). Although significantly less than vehicle control, the number of Tfh cells in the two different TCDD treatment groups were not significantly different from one another (Figure 2D; ANOVA, p=0.92). However, instead of being 21-fold lower (TCDD D-1), the number of Tfh cells was 6-fold lower (TCDD D+5) than in infected controls. These findings are consistent with observations in B cells and suggest that AHR activation affects events that are critical early on (before day 5) of immune challenge and delaying AHR activation attenuates, but does not fully abrogate, an impact on the humoral response to infection.

Figure 4. AHR activation prior to or five days after viral infection affects Tfh cells.

Mice were administered vehicle or TCDD once, either one day prior (D-1) or five days after (D+5) infection with IAV. MLN were harvested for analysis by flow cytometry 9 days after infection. (A-C) FACS plots depict the percentage of CD4⁺ T cells that were defined as Tfh cells (CD4⁺CD44^hiCXCR5^hiPD1^hi cells) from (A) vehicle, (B) TCDD D-1, and (C) TCDD D+5 mice. The number on each plot denotes the mean percentage of CD4⁺ T cells that were Tfh cells. (D) Bar graph shows the number of Tfh cells in the MLN. Tfr cells were defined as CD4⁺CD44^hiCXCR5^hiPD1^hiFOXP3⁺ cells. (E, G) The graph denotes the (E) percent of follicular CD4+ T cells (CD4⁺CD44^hiCXCR5^hiPD1^hi) that were FOXP3⁺ (Tfr cells). (F) Stacked bar graph depicts the proportion of follicular CD4⁺ T cells that were Tfr and Tfh cells. (G) The graph shows the number of Tfr cells in the MLN. The gating strategy and FMO controls are found in Figure S1. *Denotes p≤ 0.05 compared to vehicle treatment group (ANOVA, Dunnett’s test). Different letters indicate that the means were statistically different from one another, while the same letters indicate that the means were not different from one another (ANOVA, Tukey HSD). Data are shown as mean ± SEM, and are representative of at least one independent experiment, which was initiated with 7 mice per group in each experiment.

Another category of CD4⁺CXCR5^hiPD1^hi T cells are T follicular regulatory(Tfr) cells, which can be distinguished from Tfh cells based on the expression of the transcription factor, FOXP3 (18, 45, 51) Consistent with previous findings (18), AHR activation prior to infection (TCDD D-1) significantly increased the percentage of Tfr cells compared to infected vehicle controls (Figure 4E). However, the percent of Tfr cells was not significantly different from infected controls when AHR activation was not initiated till the 5^th day of infection (Figure 4E). In contrast to the percent of Tfh cells (Figure 4B–C) the percent of Tfr cells was significantly different between the two TCDD treatment groups (ANOVA, p = 0.012) (Figure 4E). The changes in the percentages of Tfh cells and Tfr cells were reflected in the lymph node, as the proportion of Tfr to Tfh cells increased when AHR was activated prior to infection, but not when AHR activation was not initiated until the 5^th day after infection (Figure 4F). Furthermore, regardless of the timing of initiating AHR activation, the number of Tfr cells was significantly reduced compared to vehicle (Figure 4G). These findings indicate that the timing of AHR activation influences the balance of immunostimulatory and immunosuppressive follicular T cells in response to IAV infection.

Activating the AHR during the initial days of infection impacts the antibody response later on during IAV infection.

TCDD is poorly metabolized, with a half-life in mice of 7–10 days (14, 27, 34). The in vivo persistence of TCDD means that a single dose causes sustained AHR activation (27, 52, 53). For example, using expression of Cyp1a1 as a marker of AHR activation, elevated gene expression continues for at least three weeks after administration of a single oral dose of TCDD (27, 52, 53). This enduring in vivo AHR activation poses a challenge to using TCDD to determine the contribution of AHR activation to modulation of events that occur during the first few days of infection because there is no facile way to end this signal. In contrast, a single dose of ITE to mice leads to AHR activation for ~24 h in vivo (Figure 5A), whereas daily ITE administration sustained AHR activation (Figure 5A). Therefore, to determine whether AHR activation only during the initial days after infection contributes to dampening of the antiviral humoral response, mice received a daily dose of ITE or vehicle control for different durations starting one day prior to IAV infection. Compared to infected mice that received the vehicle control, levels of IAV-specific antibodies were reduced 1.5-fold in mice that received daily ITE on days −1 to 9 relative to infection (Figure 5B). Similarly, daily administration of ITE from days −1 to 5 relative to infection significantly reduced the level of IAV-specific antibodies compared to infected controls (1.4-fold reduction) (Figure 5C). Moreover, IAV-specific antibodies levels in infected mice dosed with ITE daily on days −1 to 5 were similar to that of mice receiving ITE daily between days −1 to 9 relative to infection (p=0.65). While there was a reduction in IAV-specific antibodies levels, antibody avidity was not different in mice that received ITE days −1 to 5 compared to infected controls (Figure 5D).

Figure 5. Activation of the AHR by ITE during the first 5 days of infection reduces circulating IAV-specific antibodies.

C57BL/6 mice were given vehicle or ITE (10 mg/kg; i.p.). (A) Bar graph shows the relative level of Cyp1a1 in the liver after ITE was given either every day for 10 days (Daily) or once. For the mice (5 per group) given ITE one-time, separate groups of mice were sacrificed 12, 24 and 48 hr later. Livers were collected from mice given control daily or once. There was no difference in Cyp1a1 expression among these two groups of vehicle controls; therefore, the bar graph depicts a single Vehicle group. Asterisks (*) indicates p-value ≤ 0.05 from a one-way ANOVA comparing ITE treatment to vehicle. (B) Mice (7–8 per group) were dosed daily with vehicle or ITE starting on day −1 relative to infection, infected with IAV, and blood was harvested on day 9 post infection. IAV-specific IgG2a was determined by ELISA and bar graph shows the mean area under the curve (AUC). (C) Mice (5–8 per group) were given vehicle or ITE daily starting on day −1 relative to infection until 5 days after infection, infected with IAV and blood was harvested on day 14 after infection. IAV-specific IgG2a was determined by ELISA and bar graph shows the mean area under the curve (AUC). For panels B and C: * denotes p-value≤ 0.05 from a two-sided Student’s t-test. (D) Mice were given vehicle or ITE daily from day −1 to day +5 relative to infection. Bar graph shows the mean (± SEM) AUC of IAV-specific IgG2a that was untreated (closed symbols) or treated with 1M GuHCl (open symbols). Samples from vehicle treated mice are denoted with circles and ITE treated mice with squares. * denotes p-value≤ 0.05 and ns denotes non-significance from a two-way ANOVA followed by a Tukey HSD.

Unlike the level of virus-specific antibodies, the percent (Figure 6A) and number (Figure 6B) of GC B cells were not affected when the administration of ITE did not continue past the 5^th day of infection. Similarly, the percent (Figure 6C) and number (Figure 6D) of plasma cells were no longer different from infected controls. In contrast, compared to infected controls, the number of Tfh cells was significantly lower (3-fold) in mice that received daily ITE from day −1 until the 5^th day of infection (Figure 6F). These findings indicate that AHR activation within the first 5 days of viral infection impacts the production of IAV-specific antibodies in a manner that is not reflected in a change to frequency of GC B cells or plasma cells, but rather, is likely influenced by the generation of the Tfh cell population.

Figure 6. AHR activation during the first 5 days of infection influences the outcome of IAV-specific humoral responses.

C57BL/6 mice were given vehicle or ITE (10 mg/kg; i.p.) daily from one day prior to infection to five days post-infection. MLN were harvested on days 9 and 14 post infection, and single cell suspensions analyzed using flow cytometry. (A-B) Bar graphs show the (A) percent of MLN cells that were GC B cells and (B) number of GC B cells in the MLN on day 14 post-infection. (C-D) Graphs depict the (C) percent and (D) number of MLN cells that were plasma cells on day 14 post-infection. (E-F) Graphs depict the (E) percent of Tfh cells that are CD4+ T cells and the (F) number of Tfh cells out of total MLN on day 9 post-infection. *Denotes p≤ 0.05 compared to vehicle treatment group (two-sided Student’s t-test). Data are shown as mean ± SEM, and are representative of at least one independent experiments with 6–8 mice per group in each experiment.

4. Discussion

The AHR sits at the nexus of toxicology and pharmacology. Environmental contaminants that bind to and activate AHR have been associated with deleterious impacts on human health, including lower antibody responses and greater incidence or severity of infectious disease (9, 11, 54). However, a broad range of synthetic and naturally derived chemicals that are not pollutants also activate AHR (8) and hold tremendous promise as potential therapeutic agents. Thus, unintended effects on humoral immune defenses are a concern when using small molecules that target the AHR. One of the central findings was that the timing of AHR activation relative to acute primary viral infection matters. That is, different elements of the humoral response were more, or less, sensitive to AHR activation during earlier or later periods after IAV infection (Figure 7). For instance, when AHR activation was delayed, circulating levels of antiviral antibodies were lower than infected controls, but not as low as when AHR was activated prior to infection. In contrast, the number of Tfh cells was blunted by AHR activation regardless of whether it was initiated one day before or 5 days after infection.

Figure 7. Timing of AHR activation impacts multiple aspects of humoral immunity.

(A) Graphical summary of timing, relative to infection in which AHR was activated by the administration of an exogenous agonist. (B) Summary of observed modulation of the humoral response in response to AHR activation. The direction of the arrow denotes the change relative to infected controls, and the number of arrows indicate the magnitude of the difference. ND denotes “not detected” and indicates the impact of AHR activation was unchanged compared to infected mice in that received vehicle control.

When AHR activation was initiated just prior to infection and is sustained, antibody responses were dampened. This is consistent with many prior reports that AHR activation by persistent pollutants has been associated with suppression of antibody responses to a range of pathogenic and non-pathogenic antigens (9, 11, 12, 14, 17–19, 54, 55). Yet, in this current study, when AHR activation was not started until the 5^th day of infection, humoral responses were still dampened; however, the overall magnitude of the effect was not as substantial as it was when AHR was triggered earlier. That is, delaying AHR activation attenuated, but did not abrogate, suppression of the antibody response compared to when AHR activation began one day prior to infection. Conversely, AHR activation at the start of infection through day 5 still reduced IAV-specific antibody levels and diminished the number of Tfh cells, but no longer affected the frequency of germinal center B cells, plasma cells, Tfr cells, nor did not reduce the proportion of anti-viral antibodies with higher avidity. This suggests that in addition to directly affecting B cells via cell autonomous mechanisms (10, 15, 16, 56), AHR activation affects the function of other immune cell types that feed signals into the germinal center, such as Tfh cells (18). When examined in a time course study, Tfh cells in the MLN can be detected in mice starting on the 5^th day after infection, and AHR activation with TCDD dampened this early Tfh cell production (18). That there are scant Tfh cells (or GC B cells or plasma cells) in the absence of infection (18) is likely due to the requirement for an immune stimulus, such as infection, to generate these cells, rather than an impact of AHR on these cell types in immunological naïve mice. These new findings suggest that AHR activation affects cellular responses within the first few days after IAV infection.

Similar temporal effects of TCDD on the antibody response have been reported in vivo and in vitro (17, 57, 58). For example, a single dose of TCDD given any time from one day prior to and until 5 days after injection of allogeneic tumor cells suppressed the alloantibody response in mice (17). Yet, when TCDD was administered 5 days after tumor cell injection, it had no impact on the antibody response (17). In other studies, treatment with TCDD suppressed the formation of plasma cells when administered within 24 h after the addition of sheep red blood cells (SRBCs) in vitro, but not when it was added 48–120 h later (57). Similarly, TCDD treatment inhibited IgM secretion if it was added within 24h of B cell activation using LPS (58). Temporal consequences suggest that AHR likely regulates multiple events that are critical to humoral immune responses, including some that are particularly important during the early stages after infection, and other events that occur at later stages.

Among the earlier stage events during infection that are likely modulated by AHR activation are DC functions and interactions with CD4+ T cells. AHR ligands modulate DC responses during IAV infection (6, 29, 59, 60), and influence DC:CD4+ T cell interactions in the context of other antigens (60–65). While DC:T cell interactions are crucial during the very early stages after antigen encounter, in the case of a replicating antigen DC:T communications appear to play a central role for a longer period of time (59, 66). For example, during IAV infection, DCs from the respiratory tract continue to accumulate in the lung-draining lymph nodes beyond the 5^th day of infection (66). Thus, AHR activation may be affecting interactions between DCs and CD4⁺ T cells throughout infection. These interactions could manifest in alterations to aspects of CD4:B cell interactions that underpin the generation of germinal center B cells, plasma cells and antibody isotype switching. This would explain the observation that the frequencies of GC B cells, as well as plasma cells and Tfh cells were still largely suppressed even when AHR activation was not initiated until day 5 of infection.

Extrafollicular responses offer another potential explanation for differential sensitivities to AHR activation based on time relative to infection. Extrafollicular responses are initiated early after viral infection (within 2–3 days), and are composed of rapidly dividing, short-lived plasma cells in foci outside of the B cell follicle (49, 67, 68). Extrafollicular responses are an important source of virus-specific antibody during infection, and they can occur independently of T cell help (49). Thus, it is possible that extrafollicular responses are able to compensate for the loss of GC B cells, plasma cells, and Tfh cells in mice in which AHR was activated on the fifth day of IAV infection. It is also interesting that in mice exposed to TCDD starting on the fifth day of infection produced antibodies with relatively similar avidity compared to control mice. This is consistent with data showing that high-affinity BCR interactions with antigen favor extrafollicular responses over germinal center responses (49). Furthermore, although germinal centers are the main sites of somatic hypermutation, infection with other microorganisms, such as Ehrlichia muris and Salmonella, which elicit predominant extrafollicular responses, can also initiate low levels of somatic hypermutation (69, 70). Whether this occurs during IAV infection, which induces both extrafollicular and GC responses, is not entirely clear. Nonetheless, the new findings reported in the current study indicate that the timing of AHR activation can shape the magnitude and type of B cell response to IAV infection.

For rapidly metabolized AHR ligands, such as ITE, it is likely that duration of AHR activation is another factor. The consequences of the length of AHR activation on humoral responses have not been thoroughly investigated. That is, whether short term AHR activation, such as by a single dose of a rapidly metabolized AHR agonist is sufficient to disrupt a multi-day immune response is uncertain. Administration of ITE only once on the fifth day after IAV infection had no discernable impact on the frequency of GC B cells and plasma cells, Tfh cells or virus-specific antibody levels (Figure S2). This suggests that transient AHR activation may be less able to disrupt the complex and on-going processes that underpin humoral immune defenses. However, factors such as dose, metabolism and timing are important. The current studies used a single dose of TCDD or ITE. Whether and how different administered doses or a short burst of AHR activation at other points in time relative to infection may elicit a different consequence, or may dampen some aspects of the response while sparing others is uncertain. Prior studies have shown that the antibody response to IAV is affected by even lower administered doses of TCDD (71). Moreover, daily administration of readily metabolized AHR ligands, such as ITE and FICZ, had a clear impact (14, 18). In addition to reducing levels of virus-specific antibodies during IAV infection, daily treatment of mice with ITE dampened the number of Tfh cells (14). Moreover, treatment with FICZ dampened Tfh cell number when its metabolism was reduced by using Cyp1a1 knock-out mice (14). A comparison of AHR ligands, including TCDD and ITE, in the context of an allogeneic tumor model also supports that the duration of AHR activation is a key factor in influencing how AHR ligands impact T cell immune responses to alloantigen (28). Thus, it is likely that the ability of AHR ligands to modulate humoral immune responses reflects a combination of the timing and duration of AHR activation relative to antigen exposure.

When considered in a broader context, the new findings reported have several implications related to human health. The cellular responses to IAV infection measured in mice also occur in humans infected with IAV and other respiratory viruses (72, 73). In particular, the kinetics of the antibody response is similar between mice and humans (73–75). While not fully understood, interspecies differences in AHR ligand binding affinity and metabolism exist (76–78). For example, the half-life of TCDD is ~7–10 years in humans and ~7–10 days in mice (34, 79). While the human AHR may have lower binding affinity for TCDD compared to the AHR protein encoded by mouse Ahr^b alleles, the human AHR may have higher affinity for other ligands, compared to the mouse AHR (78, 80). While it is likely that AHR ligands influence antibody responses to infection in human populations, as immunomodulation is a common property of AHR agonists, prior comparative studies of AHR ligands have, for the most part, not examined humoral responses (14, 24–26, 55). Exposure to environmental contaminants is often continuous or difficult to control, and the temporal aspects indicate that therapeutic agents that act via AHR could be administered intermittently or during certain time periods, to intentionally modulate or avoid affecting humoral immune responses. These new findings indicate that timing and duration of AHR activation are important considerations relevant to how AHR ligands influence humoral immunity to a common respiratory pathogen in humans.

In the context of IAV infection, the principal role of virus-specific antibodies is not clearing acute primary infection, but establishing circulating antibodies and long-lived antibody-producing cells that are poised to protect against repeated infections by the same, or very similar, virus in the future. Nonetheless, another broad aspect of the current work relates to viral clearance. Several prior studies have demonstrated that AHR activation with different agonists, including TCDD and ITE, do not ultimately affect pulmonary viral clearance (12, 14, 81). While the precise mechanisms via which IAV is cleared when adaptive immune defenses are dampened by AHR ligands is uncertain, this probably reflects a combination of factors, including that some other aspects of the immune response are elevated, which may ultimately help eliminate virus (82–84). During infection with other agents, AHR activation consistently affects the immune response, but only sometimes affects pathogen burden (84, 85). When considering host defenses more broadly, these collective observations suggest that pathogen clearance is not necessarily a reliable indicator of whether AHR binding chemicals affect adaptive immune responses.

Given that antibody responses are crucial for immune defenses to a myriad of agents, understanding the cellular aspects that are affected by exogenous factors strengthens targeted strategies for harnessing AHR ligands for therapies, while minimizing off-target effects, such as unwanted immune suppression. Moreover, understanding how AHR agonists dampen humoral responses will help to improve strategies to attenuate pathogenic antibody production in autoimmune diseases. On the other hand, better understanding of how pollutants that activate AHR drive immunosuppression strengthens overall approaches to reduce their deleterious impact on vulnerable populations.

Highlights.

• The duration of AHR signaling influences antibody responses to viral infection

• AHR activation by TCDD or ITE suppresses humoral immune response

• The initial days after infection are a key window for AHR activation to influence anti-viral humoral immunity

Abbreviations:

TCDD

2,3,7,8-tetrachlorodibenzo-p-dioxin

ITE

2-(1H-Indol-3-ylcarbonyl)-4-thiazolecarboxylic acid methyl ester

AHR

aryl hydrocarbon receptor

Tfh

T follicular helper

GC

germinal center

PC

plasma cell

Ab

antibody