Female BALB/c mice (6–8 weeks of age) purchased from Taconic (Germantown, New York) were acclimated for 5 days and then randomly assigned to treatment group. Homogenous weight distribution was insured across treatment groups. BALB/c mice were selected on the basis of their Th2 bias, robust IgE production, and the historical use of these mice in the laboratory to investigate chemical sensitization (Anderson et al., 2010). Mice were housed in ventilated plastic shoebox cages with hardwood chip bedding at a maximum of 5 animals per cage. A NIH-31 modified 6% irradiated rodent diet (Harlan Teklad) and tap water were administered ad libitum. Housing facilities were maintained at 68–72°F and 36–57% relative humidity, and a 12-h light–dark cycle was maintained. All animal experiments were performed in the Association for Assessment and Accreditation of Laboratory Animal Care accredited National Institute for Occupational Safety and Health animal facility in accordance with an Institutional Animal Care and Use Committee-approved protocol.

Toluene 2,4-diisocyanate (TDI, CAS no. 584-84-9) was purchased from Sigma-Aldrich Chemical Company (Milwaukee, Wisconsin). Animals (n = 4–5) were exposed to a single dose of 0.5–4% TDI (v/v) on the dorsal surface of each ear (25 μl per ear). The highest concentration (4% v/v; the maximum sensitizing concentration with minimal toxicity) and dosing regimen was previously shown to induce sensitization (Anderson et al., 2013) and sensitization was confirmed in the lower dose based on total serum IgE levels and dLN allergic cytokine mRNA expression (0.5% v/v; data not shown). Acetone was selected as the vehicle control to minimize chemical reactivity as diisocyanates react with OH groups which are present in other potential vehicles such as olive oil. Ear thickness was measured 4 days following TDI exposure using a modified engineer's micrometer (Mitutoyo Corporation, Japan) and measurements were collected in millimeters (mm). Average ear swelling was calculated as previously described in Anderson et al. (2012). Mice were euthanized via CO₂ asphyxiation at time points ranging from 1 to 11 days postchemical exposure.

Animals were weighed, euthanized via CO₂ asphyxiation, and examined for gross pathology at the designated time point. Left and right auricular draining lymph nodes (dLNs; drain site of chemical application) were collected in sterile phosphate-buffered saline (pH 7.4) and manually dissociated using the frosted ends of 2 microscope slides. dLN cellularity was determined using a Cellometer (Nexcelom Bioscience, Lawrence, Massa chusetts) with size exclusion parameters (3.5–36 μm) and a combined Acridine Orange/Propidium Iodide solution to identify viable cells. Blood was collected via cardiac puncture, placed into serum collection tubes, centrifuged, and serum was removed and stored at −80°C for subsequent total IgE analysis via ELISA. Ears were collected in 1 mL RNALater (Ambion, Pittsburgh, Pennsylvania), stored at −80°C for subsequent gene expression analysis.

Ears were processed for RNA isolation using a Tissue Lyser II in Qiagen lysis buffer. Total RNA was isolated from the ears and dLN using the Qiagen RNeasy and miRNeasy kits, respectively. A QiaCube (Qiagen, Hilden, Germany) automated RNA isolation machine was utilized in conjunction with the specified RNA isolation kits. A DNase treatment was performed for removal of residual DNA. The concentration and purity of the RNA was determined using a ND-1000 spectrophotometer (Thermo Scientific Nanodrop, Wilmington, Delaware). First strand cDNA synthesis was performed using a High-Capacity cDNA Synthesis Kit (Applied Biosystems, Carlsbad, California) according to manufacturer recommendations. Ultimately, the cDNA was analyzed for mRNA expression as described in the real-time PCR (RT-PCR) methods section below.

For analysis of mRNA expression, TaqMan Universal Fast master mix (Applied Biosystems, Calsbad, California), cDNA, and mouse-specific mRNA primers (TaqMan Custom PCR Arrays, Carlsbad, California) were combined and PCR was performed according to manufacturer protocol. Primers used include: il-1β, il-6, and tnf-α. Master mix, primers, and cDNA were added to a MicroAmp Fast Optical 96-well reaction plate and analyzed on an Applied Biosystems 7500 Fast RT PCR system using cycling conditions as specified by the manufacturer. β-actin was used as the endogenous reference control gene as expression was determined to be stable following chemical exposure (data not shown). RT-PCR data were collected and represented as relative fold change over vehicle control, calculated by the following formula: 2^−ΔΔCt = ΔCt_Sample – ΔCt_Control. ΔCt = Ct_Target – Ct_β-ACTIN, where Ct = cycle threshold.

Single cell suspensions were prepared from tissues and a minimum of 150 000 dLN cells were aliquoted into 96-well U-bottom plates and washed in staining buffer (PBS + 1% bovine serum albumin + 0.1% sodium azide). Cells were resuspended in staining buffer containing anti-mouse CD16/32 antibody (clone 2.4G2; BD Biosciences, San Jose, California) for blocking of F_c receptors to minimize nonspecific binding. Cells were resuspended in staining buffer containing a cocktail of fluorochrome-conjugated antibodies specific for cell surface antigens including: CCR6 (clone: 29-2L17, fluorophore: BV605, BioLegend), CD3 (500A2, V500, BD), CD4 (RM4-5, AF700, BD), CD8a (53-6.7, AF488, BioLegend), CD25 (PC61, APC Cy7, BioLegend), CD45 (30-F11, PE, BD), CD103 (2E7, PerCP Cy5.5, BioLegend), Inducible T-cell costimulator (ICOS) (C938.4A, PE Cy7, eBioscience), Neu-1 (3E12, PE, BioLegend), and Ter-119 (TER-119, FITC, eBioscience).

Following surface staining, cells were washed in staining buffer and fixed using the Foxp3 fixation buffer set (eBioscience, San Diego, California). After overnight incubation in staining buffer, cells were permeablilized using the Foxp3 fixation buffer set (eBioscience, San Diego, Calfornia) and re-suspended in permeabilization buffer containing a cocktail of fluorochrome-conjugated antibodies specific for intracellular antigens including: CTLA4 (UC10-4B9, BV421, BioLegend), Foxp3 (FLK-16s, eF450, and APC, eBioscience), Gata3 (TWAJ, eF660, eBioscience), Rorγt (Q31-378, PerCP Cy5.5, eBioscience), and Tbet (4B10, PE Cy7, eBioscience). Following staining, cells were re-suspended in staining buffer and analyzed on an LSR II flow cytometer using FacsDiva software (BD Biosciences). Data analysis was performed with FlowJo 10.0 software (TreeStar Inc., Ashland, Oregon). A minimum of 10 000 events were captured for each sample. Leukocytes were first identified by their expression of CD45. The T_reg subset was further identified as CD3⁺ CD4⁺ CD8⁻ CD25⁺ Foxp3⁺. Numerical population values were calculated by applying subset frequencies to the initial cell count obtained following lymph node homogenization. Compensation controls were performed using single stained cellular suspensions and OneComp beads (eBioscience, San Diego, California) and fluorescence minus one staining controls were included to help set gating boundaries.

The suppressive ability of T_regs was analyzed using an ex vivo T_reg suppression assay as described by Kruisbeek et al. (2001) and Marshall et al. (2008) with some modifications. This assay evaluates the ability of naïve, conventional dLN-derived T cells (T_cons) to proliferate in the presence of varying numbers of T_regs isolated from acetone- or TDI-exposed mice. Mice were exposed to acetone (n = 7–11) or TDI (4%) (n = 4–5) as previously described and following sacrifice at 4 and 7 days postTDI (peak of the expansion) exposure the dLN and spleens were removed. T_regs (CD4⁺ CD25⁺) and T_cons (CD4⁺ CD25⁻) were isolated from the lymph nodes and CD4⁻ accessory cells were isolated from naïve spleens using CD4 negative and CD25 positive selection-based magnetic separation kits (Stemcell, Vancouver, British of Columbia). Average T_reg purity is as follows for 4 days: Acetone-70.65% of CD3⁺ CD4⁺ cells and 4% TDI-63.05% of CD3⁺ CD4⁺ cells and 7 days: Acetone-91.35% of CD3⁺ CD4⁺ cells and 4% TDI-82.8% of CD3⁺ CD4⁺ cells post 4% TDI exposure. Following isolation from naïve mouse dLNs, T_cons were labeled with 2 μM carboxyfluorescein succinimidyl ester (CFSE). T_cons and T_regs were cultured in a 96-well U-bottom plate with anti-CD3 (0.2 μg/ml; BD Biosciences) and accessory cells at a variety of T_con:T_reg ratios (1:1, 2:1, 4:1, and 8:1). Naïve CD4⁻ splenocytes were treated with Mitomycin C (Sigma Aldrich, St Louis, Missouri) and utilized as accessory cells. Additional controls included stimulated T_cons only to assess baseline proliferation, T_regs only, accessory cells only, and T_cons only with no stimulation nor accessory cells. Cells from each treatment group were pooled and added to triplicate wells of the culture plate. Seventy-two hours following plating, cells were stained with anti-CD4 and Live/Dead Violet (Life Technologies, Carlsbad, California). T_cons were defined as CD4⁺ CFSE⁺ cells and suppression was measured based on changes in the frequency of dividing CFSE⁺ cells based on the dilution of CFSE. T_regs were analyzed for purity based on their expression of CD3, CD4, and Foxp3 as determined by flow cytometric analysis as previously described.

In order to deplete T_regs before and during TDI sensitization, InvivoMAb anti m CD25 (BioXCell, PC-61.5.3) was administered in vivo. The use of this antibody as a T_reg depletion strategy is validated in vivo (Felonato et al., 2012; Setiady et al., 2010). InVivoMAb Rat IgG₁ (Bioxcell, HRPN) was utilized as a negative, isotype control. Treatment groups for this study (n = 5 mice) are as follows: isotype/acetone, isotype/0.5% TDI, isotype/2% TDI, antiCD25/acetone, antiCD25/0.5% TDI, and antiCD25/2% TDI. 200 μg of the respective antibody in USP grade saline was administered intraperitoneally at days −11 and −8 during the depletion study (see Figure 6A for study timeline). Animal weights were recorded throughout the duration of the experiment to monitor potential toxicity. In order to confirm the effectiveness of the antibody, blood was collected from the lateral tail vein at days 2 and 7, and T_regs were measured by flow cytometry as previously described. Baseline blood T_reg levels were assayed at day 12 to ensure equal pretreatment frequencies across all groups. Mice were exposed to a single dermal application of 0.5 or 2% TDI on day 0. The high dose of 2% TDI was selected in an effort to allow for a measurable increase in the sensitization response, as 4% TDI elicits a maximum sensitization response. Mice were euthanized 7 days following dermal chemical exposure. Specific measures of sensitization were evaluated, including examination of the Th2 population by flow cytometry (CD3⁺ CD4⁺ Gata3⁺), IL-4 dLN mRNA levels by RT-PCR, and total IgE levels in the serum. Total serum IgE was quantified following serum separation from whole blood (centrifugation) and analysis using a Mouse IgE Ready-SET-Go! Kit (Affymetrix eBioscience, San Diego, Califorrnia) according to the manufacturer's protocol. Absorbance was determined using a Spectramax V_max plate reader (Molecular Devices, Sunnyvale, California) at 450 and 650 nm. Data analysis was performed using the IBM Softmax Pro 3.1 program (Molecular Devices, Sunnyvale, California) and the IgE concentrations for each sample were interpolated from a standard curve derived from multipoint analysis.

Statistical analyses were generated using SAS/STAT software, version 9.3 (SAS Institute, Cary, North Carolina) and GraphPad Prism version 5.0 (San Diego, California). For irritancy and inflammatory gene expression analysis (Figure 1), a 1-way analysis of variance (ANOVA) was conducted. If the ANOVA showed significance at P < .05 or less, the Dunnett's Multiple Comparison Test was used to compare values from groups of mice treated with varying concentrations of TDI to the acetone control group. Figures 2–6 and Table 2 were analyzed by analysis of variance using PROC MIXED. In some cases, data were transformed using the natural log to meet the assumptions of the analysis. Significant interactions were explored utilizing the ‘slice’ option in PROC MIXED and pairwise differences were assessed using a Fishers Least Significant Difference Test. Supplemental data was analyzed by a Student t-test comparing groups as indicated in the figure legends. All differences were considered significant at P < .05; representative significance symbols varied by figure, as indicated in the figure legend.

To confirm that a single dose exposure to TDI (0.5 and 4%) would sensitize animals, total serum IgE was evaluated following TDI exposure. Although not initially statistically significant, IgE levels appeared to increase in a dose-dependent manner, reaching significance following 4% exposure (Figure 1A). Because TDI is a known irritant (Daftarian et al., 2002), the selected doses of TDI (0.5–4% v/v) were assayed for dermal irritancy potential via ear swelling measurements and ear inflammatory cytokine mRNA production quantified via RT-PCR. Average ear swelling was significantly increased four days following 2 and 4% TDI exposure (Figure 1B); however, neither 0.5 nor 1% TDI exposure induced significant increases in ear swelling (Figure 1B). Inflammatory cytokine (IL-1β, IL-6, and TNF-α) mRNA levels were significantly increased in the ear four days following 4% TDI exposure (Figs. 1C–E). These data suggest that 2 and 4% TDI exposure induce a significant irritation response in the ear compared with 0.5 and 1% TDI exposure. Based on this data, 0.5 and 4% TDI were selected as the exposure concentrations in subsequent studies to represent sensitizing concentrations that encompassed both a low dose exhibiting a lack of irritation (0.5%) along with a high dose exhibiting significant irritation (4%).

In order to profile the expression kinetics of the T_reg subset during TDI sensitization we examined dLN cell populations at 1, 2, 4, 7, and 9 days postTDI exposure. T_regs were identified as CD3⁺ CD4⁺ CD25⁺ Foxp3⁺ cells by flow cytometry (Figure 2A). The frequency of T_regs was unchanged compared with control cells from acetone-treated mice during 0.5% TDI sensitization; however, T_reg frequency increased significantly at 4, 7, and 9 days post 4% TDI exposure (Figure 2B). T_regs also significantly increased in number at all time points analyzed during 0.5 and 4% TDI sensitization (Figure 2C). The peak numbers of T_regs in the dLN appeared to occur at day 4 post TDI exposure for both concentrations (mean ± SEM; 1.17 × 10⁶ cells ± 0.08 (0.5%) and 1.7 × 10⁶ cells ± 0.08 (4%)) compared with the acetone control (0.18 × 10⁶ cells ± 0.02). The T_reg population remained elevated compared with the acetone control, but began to retract at days 7 (0.5% TDI- 0.88× 10⁶±0.09, and 4% TDI- 1.39 × 10⁶±0.1) and 9 (0.5% TDI- 0.37 × 10⁶±0.06, and 4% TDI- 0.79 × 10⁶±0.05) following TDI exposure relative to their peak at day 4.

In order to better elucidate the origin of the T_regs involved in the TDI sensitization response, natural T_regs (nT_regs) were identified based on their expression of neuropilin-1. Interestingly, the frequency of the nT_reg subset as a percentage of total T_regs significantly decreased at 7 and 9 days during 0.5% and 4% TDI sensitization (Figure 3C). Although the frequency of this subset decreased in relation to the acetone control group, the numbers of nT_regs in the dLN increased during 0.5 and 4% TDI sensitization (Figure 3D). This increase was significant for 4–9 days during 0.5% TDI sensitization and at all time points measured during 4% TDI sensitization. It is important to note that both the nT_reg and the neuropilin-1^neg T_reg population (presumably induced T_regs [iT_regs]) were represented as co-expressing populations examined using additional markers.dLN T_reg subsets were further phenotyped by flow cytometry (Table 1). The phenotyping analysis revealed a number of T_reg subpopulations that exhibited the potential to be functionally diverse in relation to their mechanism(s) of suppression (Table 1 and Supplementary Figure 1). An expansion in the frequency and number of the CTLA-4⁺T_reg population was observed at all measured time points following both 0.5 and 4% TDI exposure (Figs. 3A and B). This population peaked in expression of both frequency (Figure 3A) and number (Figure 3B) at day 4 post TDI exposure. The emergence of the CTLA4⁺T_reg population during TDI sensitization suggests the engagement of CTLA-4, a negative costimulatory molecule, as a potential suppressive mechanism during this response. As expected, presumably due to compensatory mechanisms during T cell activation, there was also an increase in CTLA4 expression in CD4⁺non-T_regs; however, at the peak of expression (4 days) the mean frequency of expression among all CD4⁺ cells was 0.54% ± 0.024 for acetone, 4.6% ± 0.24 for 0.5% TDI, and 7.85% ± 0.66 for 4% TDI groups. When compared with the T_regs expressing CTLA4 at the same time point, these numbers reveal a sizably smaller frequency of nonT_regs expressing this marker, emphasizing the specificity of this marker to the T_reg population during this response.

In addition to the CTLA4⁺T_regs and nT_regs, cells expressing the homing molecules CD103, CCR6, and ICOS were analyzed during TDI sensitization. T_reg expression of CD103, CCR6, and ICOS increased in both frequency and number throughout 0.5 and 4% TDI sensitization (Table 2). The CCR6⁺ population's expression frequency was highest at 7 days postTDI exposure, while the CD103⁺ and ICOS⁺ populations’ frequency peaked at 4 days postexposure. It is important to note that while there are significant changes in these subsets, they represent a small portion of the total CD4⁺ population, as their expression is presented on T_regs, which themselves constitute a minority of the total CD4⁺ population. Although there was robust T_reg expression of CD103, CCR6, and ICOS as single markers, there was a significant population of dLN T_regs that co-expressed these molecules (Figs. 4A and B) with kinetics were similar to those exhibited by each single marker. This population likely represents a migratory subset of effector T_regs that have been activated during TDI sensitization. This co-expressing population expanded in the dLN in both number and percent during TDI sensitization and represented 6.34% ± 1.2 of T_regs following acetone exposure, 12.82% ± 1.2 of T_regs during 0.5% TDI sensitization and 20.7% ± 1.8 of T_regs during 4% TDI sensitization at 7 days postexposure (peak expression).

Because the T_reg population expanded during TDI sensitization, the functional role of T_regs was further examined by performing a CFSE-based T_reg suppression assay with T_regs isolated from acetone or TDI-treated mice (Figure 5A). T_regs from acetone-treated mice were significantly suppressive at all T_con:T_reg ratios tested when isolated at 4 (Figure 5B) and 7 days (Figure 5C) postexposure. Acetone T_reg-induced suppression was found to be equivalent to naïve-derived T_reg-induced suppression (data not shown). T_regs from TDI-exposed mice exhibited increased suppressive ability compared with the acetone controls when isolated at both 4 (Figure 5B) and 7 days (Figure 5C) at all T_con:T_reg ratios tested. The heightened suppressive ability of T_regs was surprising, given the progression of sensitization at the selected concentrations of TDI. This data suggested a functional, suppressive role for T_regs in TDI sensitization.

In order to analyze the functional potential of T_regs in vivo during TDI sensitization, this subset was depleted by injecting mice with anti-CD25 antibody days 11 and 8 prior to TDI exposure (Figure 6A). The high dose of 2% TDI was selected in an effort to allow for a measurable increase in the sensitization response, as 4% TDI elicits a maximum sensitization response. The basal levels of T_regs in the blood of mice were determined to be comparable among all groups prior to dosing (Supplementary Figure 2A) and depletion was confirmed in the blood (Sup 2A) at days −2 and 7 and in the dLN (Supplementary Figure 2B) at day 7. Mice dosed with 2% TDI that received anti-CD25 lost significantly more body weight (grams; mean decrease 5.33% 6 3.9) than mice exposed to 2% TDI and isotype control antibody (mean increase 1.01% ± 4.3), suggesting enhanced toxicity following T_reg depletion and high-dose TDI administration. Because local irritation is thought to influence allergic sensitization (Pauluhn, 2014) and TDI is a known irritant (Duprat et al., 1976), we analyzed the dermal irritation response at the site of TDI exposure by ear swelling measurements following T_reg depletion. As expected, there was a dose-responsive increase (Linear trend test P < .01) in ear swelling following exposure to TDI for both the isotype control and anti-CD25 groups (Figure 6B). Although not statistically significant, ears from animals treated with anti-CD25 and either concentration of TDI exhibited increased swelling (12.4% ± 3.7 for 0.5%; 54.4% ± 10.7 for 2% TDI) compared with their isotype-treated counterparts (1.4% ± 6.6 for 0.5%; 35.6% ± 8.9 for 2% TDI). At day 7, dLN cellularity increased dose-responsively (Linear trend test P < .05) for both the isotype control and antiCD25-treated groups during TDI sensitization (Figure 6C). For both the 0.5 and 2% TDI-treated groups statistically significant increases in the dLN cellularity with anti-CD25 treatment compared with isotype were observed. Another effector CD4⁺T cell subset, Th2, which play an important role in allergic responses, was identified by GATA-3 expression and was found to dose-responsively expand in number during 0.5 and 2% TDI sensitization for both the isotype and anti-CD25-treated groups (Figure 6D; Linear trend test P < .01). When comparisons were made between the isotype control and anti-CD25 treated groups this population appeared to further expand during TDI sensitization, although these perceived trends were not statistically significant. Similar observations were made for the Th2 population's frequency (data not shown). IL-4 mRNA expression levels increased dose-responsively in the dLN following 0.5 and 2% TDI exposure (Linear trend test P < .05; isotype and anti-CD25 groups) and were significantly augmented in groups treated with anti-CD25 compared with isotype-treated controls (Figure 6E). Similarly, dose-responsive increases in total IgE levels in the blood appeared to occur at 7 days following exposure to TDI in mice treated with both isotype control and anti-CD25, although significance was only observed for the 2% TDI anti-CD25-treated group (Linear trend test P < .01; Figure 6F). Statistically significant increases in serum IgE levels were observed after anti-CD25 treatment following exposure to 0.5 and 2% TDI.

Other T cell-related dLN phenotyping was performed at day 7 including CD3, CD4, and CD8 expression, along with other effector CD4⁺Th1 and Th17 populations based on their expression of Tbet and Rorγt, respectively (Supplementary Figure 3). CD3⁺, CD3⁺ CD4⁺, CD3⁺ CD8⁺, Th1, and Th17 cells dose-responsively expanded in number during 0.5 and 2% TDI sensitization and antiCD25 depleted animals exhibited higher numbers of cells during TDI sensitization (Supplementary Figs. 3B, D, F, H, J). dLN CD3⁺ cellular frequency decreased in isotype and antiCD25-treated groups during 0.5 and 2% TDI sensitization compared with their respective acetone control (Supplementary Figure 3A). Interestingly, CD3⁺ cellular frequency was significantly increased in groups treated with anti-CD25 compared with isotype controls for acetone and 0.5% TDI, indicating that antiCD25 treatment did not affect general T cell frequencies in the dLN. CD4⁺ and CD8⁺T cell frequency (as a percentage of CD3⁺ cells) remained stable during TDI sensitization; however, the CD4+ frequency decreased following anti-CD25 treatment while the CD8+ frequency increased (Supplementary Figs.3C and E). The Th1 dLN population significantly expanded in frequency during TDI sensitization (Supplementary Figure 3C). This subset further expanded in mice treated with antiCD25 in frequency (2%) compared with the isotype control groups. The Th17 subset was also profiled and was not significantly altered in frequency during TDI sensitization in the isotype control treated groups; however, the frequency of Th17 cells was significantly increased above both the corresponding acetone and isotype-treated control groups during 2% TDI sensitization following anti-CD25 treatment (Supplementary Figure 3D).