Our study was conducted in accordance with the guidelines of the Declaration of Helsinki and the Principles of Good Clinical Practice. All patients and healthy controls who participated in this study provided written informed consent, and our research was approved by the ethics committee of the First Affiliated Hospital of Jinzhou Medical University, China. The clinical trial registration number is ChiCTR-BOC-16010279, and the URL is http://www.chictr.org.cn/. The registration date was December 28th 2016.

Seventeen healthy controls (HC) with no history of allergy or asthma, 42 allergic asthmatics (aAS), and five non-allergic asthmatics (naAS) were recruited for this study. Allergic asthma was diagnosed according to the criteria of the Global Initiative for Asthma (GINA reports). In brief, aAS subjects were diagnosed based on physical examination, history, serum IgE levels, and skin test responses to allergens. And naAS subjects were diagnosed with asthma who had the negativity of 14 skin prick tests to common environmental allergens(such as Artemisia, house dust mite, Platanus pollen, cat fur, dog fur and so on). HC were defined as nonsmoking, negative skin test results and no evidence of lung disease. HC were recruited from the first affiliated hospital of Jinzhou medical university. Patients with asthma were recruited from the general hospital of Shenyang military area command and the second affiliated hospital of Shenyang medical college. Peripheral blood samples were collected from all individuals. The clinical characteristics of enrolled subjects are listed in Table 1 .

Fresh peripheral blood mononuclear cells (PBMCs) were isolated by density gradient centrifugation using Ficoll-Paque Plus (GE Healthcare). PBMCs were then washed twice with 1_x phosphate buffered saline (PBS) and resuspended in RPMI 1640 medium (Gibco, Grand Island, NY, USA) supplemented with 10% heat-inactivated foetal bovine serum (HyClone, Logan, UT, USA) and 100 units/mL penicillin/streptomycin.

PBMCs (1 × 10⁶ cells) were cultured for five hours with PMA (50 ng/mL; Sigma Chemical, Saint-Louis, MO, USA); Ionomycin (1 μg/mL; Euromedex) and Brefeldin A (10 μM; BD Bioscience) were added during the last 1 h of culture. Cells were harvested and stained with Panel A antibodies: CD3-Percp-Cy5.5, CD8-FITC, CD25-PE, and CD127-Brilliant Violet 510; Panel B antibodies: CD3-Percp-Cy5.5, CD8-FITC, CCR6-PE, CD278-PE/Cy7, and CD25-Brilliant Violet 510. After staining of cell-surface markers, cells were fixed and permeabilized with the Foxp3 Fixation/Permeabilization Buffer (BioLegend) according to the manufacturer’s protocols. Subsequently, the cells were stained with Panel A: IL-10-PE/Cy7, TGF-β-Brilliant Violet 421, and Foxp3-Alexa-Fluor 647; Panel B: Foxp3-Alexa-Fluor 647, and IL17A-Brilliant Violet 421. All antibodies were purchased from BioLegend (San Diego, CA, USA). The detailed antibodies information were listed in Table 2 . In humans, Treg cells are classically defined as CD4⁺CD25⁺Foxp3⁺ T cells. In this study, Treg cells were first defined as single cells within the lymphocyte gate. Second, these cells were CD3⁺CD25⁺Foxp3⁺ yet Zombie⁻ and CD8⁻ to exclude dead cells and CD8⁺ Treg cells, respectively ( Supplementary Figure 1 ).

Fresh human PBMCs were stained with monoclonal antibodies, such as PercP-Cy5.5-anti-human CD3, APC/Cy7-anti-human CD4, PE-anti-human CD25, and Brilliant Violet 510-anti-human CD127 (all purchased from BioLegend, San Diego, CA, USA), in 0.5% BSA + 2 mM EDTA in PBS. Splenocytes were obtained from the spleens of mice by mechanical disruption and through a 200-gauge steel mesh. Splenocytes were lysed with 1_X red blood cells (RBCs) lysis buffer to remove RBCs. First, magnetic microbeads (Miltenyi Biotec) bead-based enrichment of CD4⁺T cells was performed. Then, cells were stained with monoclonal antibodies, such as PE-anti-mice CD4, APC/Cyanine7-anti-mice CD25, and Alexa-Fluor647-anti-mice CCR6 (all purchased from BioLegend), in 0.5% BSA + 2 mM EDTA in PBS. Human Treg cells and murine Treg cells were sorted using a SH800S cell sorter (SONY, Japan) and analyzed with the SH800 software (SONY, Japan). Human Treg cells were gated as CD3⁺CD4⁺CD25⁺CD127^- within the lymphocyte gate, which excluded cell debris, doublets, and dead cells ( Supplementary Figure 2 ). Mice Treg cells were gated as CD4⁺CD25⁺ within the lymphocyte gate which excluded cell debris, doublets, and dead cells ( Supplementary Figure 2 ). The purity of the sorted cell population was above 95% for human and 91% for mice, which was verified using post-sorting flow cytometry.

Sputum was induced with hypertonic saline (4.5%) and processed as previously described (28). The amount of CCL20 of induced sputum was determined using the human CCL20/MIP-3 alpha Quantikine ELISA Kit (R&D Systems), and the mouse MIP-3 alpha/CCL20 AccuSignal ELISA Kit (Rockland) was used for the analysis of mice CCL20. The levels of IL-17A, IL-4, and IgE were determined using the ELISA kits from Dakewe Biotech (Shenzhen, China).

To assess the migratory properties, Treg cells (1 × 10⁵) were seeded in the upper chamber of a 5-μm 24-well transwell plate (Corning Costar, Corning, NY) in 100 μL of RPMI 1640 medium containing 10% FBS. A gradient of 100 ng/mL recombinant chemokines CCL20 (PeproTech, Germany) was added to the lower compartment. Migration was assayed for 2.5 h at 37 °C, 5% CO₂, and subsequently the inserts were removed. Migrated Treg cells were enumerated with an automatic cell counter (Countstar BioMed, ALIT Life Science, Shanghai, China).

Expression of CD4, Foxp3, CD25, CD8A, and CD8B was assessed in our 10X Genomics datasets combining 3’-mRNA and surface protein expression. Foxp3^hi cells were defined as cells expressing one or more copies of Foxp3. More than one thousand Foxp3^hi cells obtained from two HC and two patients with allergic asthma were analyzed in this study. The sequencing depth was 20000 to 50000 reads per cell or nuclei. We used the Seurat package to perform ScRNA-seq analysis. The function of UMAPplot, RidgePlot, and FeatureScatter were used to plot the figures. The detailed procedure on single-cell RNA-seq and ScRNA-seq data processing is provided in the Supplementary Material .

C57BL/6 mice (6–8 weeks old) were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). eGFP transgenic mice and CCR6^-/- mice were obtained from the Model Animal Research Center (Nanjing, China), provided by the Jackson Laboratory. All mice were raised under specific pathogen-free conditions. The Committee on Animal Experimentation of Jinzhou Medical University approved the animal procedures used in this study.

For ovalbumin (OVA)-induced airway inflammation, mice were sensitized by intraperitoneal injection with 50 mg of OVA in 1 mg of Alhydrogel (CSL), and after one week, they were challenged with 50 mg of OVA intranasally (i.n.) for seven days. Non-sensitized mice received 1 mg of Alhydrogel in 0.9% saline and were challenged with sterile saline. Mice were harvested and examined on Day 19. For house dust mite (HDM)-induced airway inflammation, mice were sensitized and challenged with intratracheal administration of HDM extracts (Huipuyuan, Liaoning, China) with 50 μg HDM in 25 μL of 0.9% saline twice at seven day intervals. Seven days after the last sensitization, mice were challenged with 5 μg of HDM for four consecutive days, and 24 h after the last HDM challenge, mice were harvested and analyzed. Their blood was collected, after which the mice were tracheotomized. Bronchoalveolar lavage fluid (BALF) was collected for analysis using FACS, ELISA, and inflammation cell counting. To obtain cells in the bronchoalveolar lavage fluid (BALF), we performed bronchoalveolar lavage as described previously (29) with slight modifications. The whole lung was perfused through the trachea three times with 0.5 ml sterile phosphate buffered saline (PBS). After centrifugation of the BALF, cell pellets were harvested for FACS. To obtain cells in whole blood, the red blood cells (RBCs) were removed by RBC lysis buffer (BioLegend). The harvested cells were washed with 1_x PBS and collected for further experiments. Serum was collected, and total IgE were measured by ELISA.

For surface staining, single cells were incubated with antibodies, such as anti-mice CD4 and anti-mice CD25 for 30 min at 4°C after blocking the Fc receptor, then washed twice with 1_x phosphate buffered saline (PBS). After staining of cell-surface markers, cells were fixed and permeabilized with a Foxp3 Staining Buffer Set (BioLegend) according to the manufacturer’s protocol and then stained with anti-mouse IL-17 and anti-mouse Foxp3 or isotype control for 1 h at 4°C. Samples were analyzed by the BD verse system (BD Biosciences). In mice, Treg cells were defined as CD4⁺CD25⁺Foxp3⁺ within the lymphocytes gate.

Mouse lung tissues without lavage were fixed in 10% formalin and subsequently embedded in paraffin. The lung sections were stained with hematoxylin and eosin (H&E). Histopathology images were acquired on a Leica microscope (DM4000 B LED) using magnification (200×) by a DFC450C camera and Leica application suite V4.12 software. More than six fields from left and right of the lung samples were selected randomly to determine inflammatory injury, and the researcher was blinded to the treatments. Histological scores were assessed as described previously (30). Each set of sections was given a score of 0–5 for inflammation (0, no inflammation; 1, few inflammatory cells; 2, a ring of inflammatory cells one cell layer deep; 3, a ring of inflammatory cells two to four cells deep; and 4, a ring of inflammatory cells of more than four cells deep).

All flow cytometry data were analyzed using FlowJo X 10.0.7r2 (Tree Star). Statistical analysis was performed using Prism 6.0 (GraphPad Software). Pearson’s correlations were calculated using logarithmically transformed data. Comparisons of continuous variables between groups were conducted using unpaired t-tests (two-tailed), one-way and two-way analysis of variance, and linear regression according to the type of experiment. The P values of <0.05 were considered significant (*P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001). Data were summarized as the mean ± SEM, unless otherwise indicated.

Compared with the HC (n = 17) and non-allergic asthma subjects (naAS, n = 5), the percentage and number of CD25⁺Foxp3⁺ Treg cells was markedly decreased in patients with allergic asthma (aAS, n = 42, Figures 1A–C ). Treg cells were defined as described in the Materials and Methods section. Next, we compared the percentages of circulating CCR6⁺Foxp3⁺ Treg cells and CCR6^-Foxp3⁺ Treg cells among peripheral CD3⁺CD8^- T cells between HC and aAS. In comparison with the HC (n = 12), the percentage of CCR6⁺Foxp3⁺ Treg cells was markedly decreased in aAS (n = 15). Meanwhile, there were no changes in the percentage of CCR6^-Foxp3⁺ Treg cells ( Figures 1D, E ). Figure 1F shows that the mean fluorescence intensity (MFI) of CCR6 in CD25⁺Foxp3⁺ Treg cells of aAS is lower than that of HC. Furthermore, there was a marked decrease in CCR6 expression on CD25⁺Foxp3⁺ Treg cells in asthma patients as compared to the HC ( Figure 1G ). These observations suggest that the circulating CCR6⁺ Treg cells might have migrated to the airways of patients with allergic asthma. Subsequently, we collected induced sputum from five patients with aAS and five patients with naAS, and also detected CCL20, a unique ligand for CCR6. Notably, we found an increased level of CCL20 in the sputum of patients with aAS ( Figure 1H ). To identify the role of CCR6-CCL20 in Treg cells migration, a high proportion of Treg cells (> 95%) was sorted from HC ( Figure 1I and Supplementary Figure 2A ), and a transwell assay was used. Figure 1J shows that CCL20 promotes Treg cells migration in a dose-dependent manner. By comparing the differences in Treg cells migration, we found that Treg cells from HC had a higher migration ability than those from patients with aAS ( Figure 1K ). These results suggest that CCL20, which was markedly increased in the lungs of patients with aAS, might promote CCR6⁺ Treg cells migration. The reduced peripheral CCR6⁺ Treg cells might influence the development of allergic asthma.

Treg cells expressing CD8 have been previously reported (30). These cells were observed in tonsils, but rarely detected in peripheral blood (31). These CD8⁺ Treg cells were also rarely reported in allergic asthma. To explore whether altered peripheral Treg cells originate mainly from CD4⁺ or CD8⁺ Treg cell subsets, we gated on Foxp3⁺CD25⁺ Treg cells and separated them into CD8⁺ and CD8^- Treg (mainly CD4⁺ Treg) cells ( Supplementary Figure 3A ). We found CD8^- Treg (mainly CD4⁺ Treg) cells were more dominant than CD8⁺ Treg cells. And no significant difference was observed between HC and aAS ( Supplementary Figure 3B ). Next, we used ScRNA-seq to capture the transcriptional landscape at the single-cell resolution. The samples of ScRNA-seq were obtained from two HC and two patients with aAS. The PBMCs of these subjects were collected, and the CD3⁺ T cells were then sorted and subjected to scRNA-seq using the 10 × platform. Then, we used Seurat to normalize and cluster the gene expression matrix and identified the cell subsets. The results were visualized using uniform manifold approximation and projection (UMAP) for dimension reduction. The three genes (CD4, CD25, and Foxp3) enriched in the cluster were used to assign cluster-specific cell identities ( Figure 2A ). Significantly more differentially expressed genes (DEGs) were shown in Supplementary Figure 4 . Among them, CD25 was highly co-expressed with Foxp3 in nearly 50% of Foxp3⁺ Treg cells ( Figure 2B ). Moreover, the most prominent features were the high levels of CD25, Foxp3, and CD4 expression, but not CD8A or CD8B ( Figure 2C ). Collectively, RNA-seq provided evidence of circulating Treg cell-specific expressed of CD4, CD25, and Foxp3, but not CD8A/CD8B. Subsequently, we explored whether a small amount of CD8⁺ Treg cells altered in aAS in comparison with HC. Our data showed that there was no significant difference in the percentage of CD8⁺CD25⁺Foxp3⁺ Treg cells between these two groups ( Figures 2D, E ).

It has been reported that CCR6 induces the phosphorylation of Akt, mTOR, and STAT3 molecules, which are involved in suppressing Foxp3 expression and promoting Th17 differentiation (32, 33). We measured the function of Treg cells in the enrolled subjects. First, we compared the co-expression of IL-17 vs. CCR6 ( Figure 3A ) and ICOS vs CCR6 ( Figure 3C ) on the CD25⁺Foxp3⁺ Treg cells (gated on CD3⁺CD8^- T cells) between HC (n =12) and patients with aAS (n = 15). Interestingly, we found that different Treg cells subtypes separate populations. The CCR6⁺IL-17⁺ Treg cells subsets ( Figure 3B ) and the CCR6⁺ICOS⁺ Treg cells subsets ( Figure 3D ) decreased significantly in patients with allergic asthma compared with HC. However, the CCR6^-IL-17^- Treg cells subsets increased in patients with allergic asthma ( Figure 3D ).The percentage of Foxp3⁺IL-17A⁺ Treg cells was significantly lower in patients with aAS than in HC subjects (1.79% vs. 9.06%, median, respectively; P < 0.05, Figure 3B ). Furthermore, patients with aAS displayed a lower percentage of Foxp3⁺ICOS⁺ Treg cells than HC subjects (15.9% vs. 30.5%, median, respectively; P < 0.05; Figure 3D ). Consistently with previous studies (34), when comparing patients with aAS, Treg cells from HC subjects produced significant amounts of IL-10 (2.7% vs. 1.1%, median, respectively; P < 0.05; Figures 3E, F ). No significant difference in TGF-β production was found between the two groups (2.5% vs. 1.8%, respectively; P > 0.05; Figures 3G, H ). These results suggest that the Treg cells subsets are altered and inhibitory function of circulating Treg cells was weakened in patients with aAS.

To further elucidate changes in the phenotype and function of Treg cells in inflammatory lungs, we used the classic allergen-induced airway high reactivity (AHR) mouse models. First, an experimental model of OVA-induced AHR, airway eosinophilia, and type 2 cytokine production was used to monitor the activities of CCL20 and CCR6⁺ Treg cells ( Figure 4A, B ). Regardless of the treatment, the amount of CCL20 detected in the BALF was similar between the control and OVA-induced mice ( Figure 4C ). Moreover, the concentration of CCL20 protein in the lungs was not correlated with the frequency of CCR6⁺ Treg cells in the BALF ( Figures 4D, E ).

We then used a mice model with HDM-induced asthma. Mice were sensitized with HDM in alum and intranasally challenged with HDM ( Figure 5A ). On Day 25, histological studies of the lungs showed that HDM exposure induced significantly higher peri-bronchial and peri-vascular leukocyte infiltrates in the lungs ( Figures 5B, C ). By detecting the CD11b⁺Ly6G/Ly6C⁺ neutrophils from BALF and blood ( Supplementary Figure 5 ), we confirmed that HDM induced the neutrophilic inflammation in these mice ( Figures 5D, E ). Total IgE and cytokines from the BALF and serum were detected by ELISA. IgE in serum, but not in the BALF, increased significantly from the mice induced by HDM compared with the controls ( Figure 5F ). Moreover, HDM-sensitized mice primarily produced IL-4, IL-17, and CCL20 ( Figure 5G ). In agreement with these findings, after HDM sensitization, the mouse lungs had markedly more CCR6⁺ Treg cells than in PBS-injected control mice ( Figure 5H ). IL-17 producing Foxp3⁺ Treg (Th17 like Treg) cells were significantly elevated in the lungs of the HDM-induced asthma mice, whereas Treg cells did not express Th17-type cytokines not altered ( Figures 5I, J ). These data indicated that these Th17-like Treg cells might increase in neutrophilic inflammation. In contrast to the suppression of lung allergic responses, CCR6⁺ Treg cells enhanced this response. This was associated with increased levels of IgE in serum, IL-4, IL-17, and CCL20 in the BALF.

To determine whether CCL20 is a chemoattractant for CCR6⁺ Treg cells in vivo, we crossed eGFP transgenic mice expressing the GFP protein with CCR6^-/- mice; GFP marked cells that did not express CCR6 ( Figure 6A ). We separately sorted high proportion CD4⁺CD25⁺ Treg cells from these donors ( Figure 6B and Supplementary Figure 2B ) and adoptively transferred GFP⁺CCR6^-/- or GFP⁺CCR6^+/+ donor derived CD4⁺CD25⁺ T cells into immunized WT mice, which were analyzed three days later ( Figure 6A ). In this migration assay, CCR6^+/+ Treg cells showed a higher ability to migrate to the inflamed lungs than CCR6^-/- Treg cells ( Figures 6C, D ). Finally, we detected the CCL20 in BALF after the eGFP⁺ Treg cells were transferred on Day 25. These data confirmed that CCL20 in the lungs of recipient mice was higher after the Treg cells transfer ( Figure 6E ). Collectively, these findings suggest that CCR6⁺ Treg cells trafficking via CCL20, may be associated with the pathology of asthma ( Figure 7 ).