Increased Frequency of Dual Positive Th2/Th17 cells in Bronchoalveolar Lavage Characterizes a Population of Severe Asthmatic Patients

Abstract

Background

Th2 cells can further differentiate into dual positive Th2/Th17 cells. The presence of dual positive Th2/Th17 cells in the airways and its impact on asthma severity are unknown.

Objective

To study dual positive Th2/Th17 cells in bronchoalveolar lavage (BAL) from asthmatic patients, examine their response to glucocorticoids, and define their relevance for disease severity.

Methods

Bronchoscopy and lavage were performed on 52 asthmatic patients and 25 disease controls. Th2 and Th2/Th17 cells were analyzed by multi-color flow cytometry and confocal immunofluorescence microscopy. Cytokines were assayed by ELISA.

Results

Dual positive Th2/Th17 cells were present at a higher frequency in BAL from asthmatic patients as compared to disease controls. High-level IL4 production was typically accompanied by high-level IL17 production and co-expression of GATA3 and RORγT. Increased presence of Th2/Th17 cells was associated with elevated IL17 in the lavage fluid. Th2/Th17 cells and IL17 correlated with PC20 for methacholine, eosinophils and FEV1. Th2/Th17 cells, unlike Th2 cells, were resistant to dexamethasone-induced cell death. They expressed higher levels of MEK1, a molecule that induces glucocorticoid resistance. Based upon the dominance of BAL Th2 or Th2/Th17 cells, we identified three subgroups of asthma—Th2^predominant, Th2/Th17^predominant and Th2/Th17^low. The Th2/Th17^predominant subgroup manifested the most severe form of asthma whereas the Th2/Th17^low subgroup had the mildest asthma.

Conclusion

Asthma is associated with a higher frequency of dual positive Th2/Th17 cells in BAL. The Th2/Th17^predominant subgroup of asthmatic patients manifests glucocorticoid resistance in vitro. They also have the greatest airway obstruction and hyperreactivity as compared to Th2^predominant and Th2/Th17^low subgroups.

Introduction

Allergic asthma is characterized by a Th2 immune response and the presence of allergen-specific IgE antibodies.^{1, 2} IgE-mediated mast cell activation plays an important role in asthma. This conclusion is supported by the efficacy of omalizumab in many patients with refractory asthma.³ The differentiation of T cells into various T helper cell populations is a functional but not a terminal differentiation state. This was clearly demonstrated in early studies by Coffman et al, who showed that fully differentiated Th2 cells could produce Th1 cytokines in the presence of appropriate stimulants.^{4, 5} In recent years this plasticity of T helper cell function has drawn attention.^6–11 Studies have shown interconversion of T helper cells, especially interconversion of Tregs and Th17. Further, studies have demonstrated the presence of dual positive Th2/Th17 cells in the blood and tissue from subjects with asthma^{12, 13} and healthy subjects.¹⁴

Although allergens play an important role, there are other environmental factors such as infection and a broad range of chemical and physical factors that contribute to exacerbation of asthma. Many of the latter factors are likely to elicit a Th17-type immune response. A particular matter of interest is the qualitative difference between Th2 and Th17 cells in their response to glucocorticoids. IL17 production by Th17 has been shown to be less susceptible to inhibition by glucocorticoids as compared to IL4 and IL5 production by Th2 cells.¹⁵ Thus, the emergence of Th2/Th17 cells in the airways could make asthma less responsive to glucocorticoid treatment. A number of papers have reported increased presence of IL17 in the lung biopsy specimen and sputum form asthmatic patients.^16–22 Increased expression of IL17 was associated with severe asthma. However, there are no reports on the presence of dual positive Th2/Th17 cells in BAL from asthmatic patients. Thus, there are a number of reasons to study Th2/Th17 cells in asthma. The objectives of this study were to examine the expression of Th2, Th17 and dual positive Th2/Th17 cells in BAL from treatment-refractory asthmatic patients and their response to glucocorticoids. We also determined the clinical relevance of Th2/Th17 cells in asthma.

Methods

Human subjects

The study subjects were recruited from the outpatient clinics of National Jewish Health. The study protocol for bronchoscopy and BAL was approved by the Institutional Review Board (IRB). An informed consent was obtained from the study subjects. Patients were allowed to continue their routine medication.

Processing of BAL cells and flow cytometry

Bronchoscopy and BAL were performed as described previously.²³ BAL was processed immediately. Cells were isolated by centrifugation. Supernatant fluid was aliquoted into small samples and frozen. Cells were either cultured or fixed immediately in 4% paraformaldehyde and processed for flow cytometry. For cultures, cells were washed and divided into two treatment groups—medium or dexamethasone (10⁻⁷M). The cells were cultured in RPMI 1640 with 10% FBS overnight. Next day, cells were washed and fixed in paraformaldehyde. Monensin (2 µM) was added 6 hr before fixing.

Flow cytometry

Cells were stained with the following antibodies: APC-labeled mouse anti-human CD4 (clone RPA-T4), PE-Cy7 labeled mouse anti-human IL4 (clone MP4 25D2), APC/Cy7-labeled mouse anti-human IL17 (clone BL168), mouse anti-CD3ε (clone OKT3), mouse anti-CD68 (clone Y1/82A), mouse anti-CD163 (clone GH1/61), anti-CRTH2 (clone BM16), mouse anti-CCR6 (clone G034E3), rat anti-IL5 (clone TRFK5) were from Biolegend, Inc. (San Diego, CA). PE-labeled anti-MEK1 (clone 25/MEK1), PerCP-Cy5.5-labeled mouse anti-STAT3 (pY705) (clone 4/P-STAT3) and Alexa Fluor 488-labeled mouse anti-STAT6 (pY641) (clone 18/P-STAT6) were from BD Biosciences (San Jose, CA). A rabbit anti-MKP1 antibody was from Santa Cruz, Inc. This was detected using an Alexa Fluor 488-labeled anti-rabbit secondary antibody. The isotype controls were rat IgG1 for the anti-IL4 antibody, mouse IgG1 for anti-CD4 and anti-IL17 antibodies, and mouse IgG2a for anti-pSTAT3 and anti-pSTAT6 antibodies. We blocked the Fc receptors by incubating cells first with 10% goat serum and the conducting immunostaining with specific antibodies in 5% goat serum. Flow cytometry was performed using a CyAn ADP Analyzer 9 color flow cytometer (Beckman Coulter, Brea, CA) in the NJH Flow Core Facility as described previously.²⁴ Flow cytometry data were analyzed with the software FlowJo (Tree Star, Inc. Ashland, OR). We carefully gated only small and non-granular cells by forward and side scatters, respectively, and excluded the large and highly granular BAL cells from analysis as they tend to bind many antibodies non-specifically as reported previously.²⁵ The threshold line for identification of positively stained cells was set conservatively based upon the control isotype antibody staining pattern. The emphasis was on exclusion of non-specifically stained cells. Less than 1% (usually less than 0.5%) of the cells stained positively using this control antibody-based thresholding strategy. We first identified specific cell populations (CD4, CD3, CD163, and CD68) and then analyzed the cells for presence of intracellular cytokines. Isotype control antibodies were run in all experiments and the aforementioned gating and thresholding strategy was applied to all BAL samples.

Double immunofluorescence staining

BAL cells were fixed and cytospins were immunostained with a combination of mouse monoclonal anti-GATA3 (clone TWAJ, Ebiosciences, Inc., San Diego, CA) and a rabbit polyclonal anti-RORγT (clone H-190, Santa Cruz Biotech, Inc., Santa Cruz, CA) antibodies or mouse anti-IL4 (clone 8D4-8, BD Pharmingen, San Jose, CA) and rabbit anti-IL17 antibodies (Santa Cruz Biotech, Inc.) as described previously.²⁶ FITC and PE labeled secondary antibodies were directed against GATA3 and RORγT, respectively. The cytospins were counterstained with DAPI. Z-series images were captured using a Zeiss confocal microscope in the NJH Microscope Core Laboratory.

ELISA

IL17 (IL17A) was assayed in undiluted BAL fluid by an ELISA kit from R&D systems, Inc. (Minneapolis, MN) as per the supplier’s instruction.

Statistical Analyses

Comparison between study groups were done by Mann-Whitney U test. Comparison among multiple study groups was performed by Kruskal Wallis test. Pearson correlation coefficient was used to calculate for correlation coefficient.

Results

Detection of single Th2 and Th17, and dual Th2/Th17 cells in BAL from asthmatic patients

We studied bronchoalveolar lavage cells from 52 asthmatic patients and 25 disease controls. Most of the patients were referred to National Jewish Health for diagnosis and management of refractory asthma. Others were referred for routine asthma care. We will use the term “refractory” in this manuscript to indicate uncontrolled asthma, which could be moderate or severe by the Expert Panel Report 3 (EPR3) criteria. The clinical characteristics of the study subjects are shown in Table 1. Bronchoscopy and BAL were performed as described previously.²³ BAL was processed immediately for flow cytometry or for culture overnight as described below. Following Immunostaining cells were first gated for small and non-granular cells by forward and side scatters (Repository figure E1A). Next, we identified cells positive for CD4 and negative for the macrophage marker CD163 (Fig. 1A). These CD4+CD163− cells were analyzed for intracellular IL17 and IL4 or IL5. Figure 1B and 1C show detection of largely IL4 and IL5 positive CD4 T cells in BAL from an asthmatic subject. Very few CD4 T cells stained for IL17. We analyzed co-expression of IL4 and IL5 in these BAL cells. We observed partial overlap between IL4+ and IL5+ (Repository figure E1B) suggesting both synchronous and non-synchronous production of these two Th2 cytokines. Figure 1D represents a BAL flow cytogram with a Th17^predominant pattern, which comes from a patient with chronic pulmonary aspiration. Figure 1E represents a Th2/Th17^predominant staining pattern obtained from an asthmatic patient. Flow cytograms of BAL cells with isotype control antibodies are shown in online repository Figure E2A–D.

Table 1.

Characteristics of the study subjects

Parameters	Asthmatic patients	Disease controls
N	52	25
Diagnoses	52 patients with asthma; co-morbidities: 41 patients with allergic rhinitis, 29 with chronic sinusitis, 38 with GERD, 3 with bronchiectasis and 5 with aspiration	16 patients with chronic cough and concurrent allergic rhinitis & GERD, 3 patients with bronchiectasis, 5 with chronic aspiration and 1 with COPD
Male/female	23/29	12/13
Age	52.88 ± 3 (52)	52.2 ± 3 (53)
FEV1 (%)	71 ± 2 (71.2α)*	93.5 ± 4 (88.5)
Reversibility (%)	18.0 ± 2.2 (13.5)*	2.1 ± 0.6 (1)
PC20 (mg/ml) for methacholine	2.95 ± 0.3 (2.2)*	24.5 ± 0.5 (25)
Eosinophils /µL blood	336 ± 50 (300)*	100 ± 17 (100)
Total IgE (KIU/L)	215 ± 46 (99)	165 ± 78 (36)
BMI	28.7 ± 1 (30)	25.3 ± 1 (25)

^α: number in the parenthesis indicates median;

^*: P<0.05, Mann-Whitney U test

Figure 1.

Expression of Th2, Th17 and dual Th2/Th17 cells in BAL. A–C: Representative flow cytograms from a Th2 dominant asthma patient. BAL cells from an asthmatic patient was gated as shown in repository figure E1A and then analyzed for CD4 T cells and CD163+ macrophages (A). The CD4+CD163− cell population was then analyzed for expression of IL17 and IL4 (B) or IL5 (C). D & E: Representative flow cytograms from a Th17^predominant (patient with recurrent pulmonary aspiration) and Th2/Th17^predominant (asthma) patient. BAL cells were processed as per A-C and then analyzed for IL4 and IL17. F: A heat map of the frequency of IL4+, IL17+ and dual IL4/IL17+ CD4 T cells in BAL from 52 asthmatic patients. Each row is a single BAL sample. The embedded number indicates the actual frequency of cells. Based upon the dominance of the cell type asthma patients can be divided into Th2^predominant, Th2/Th17^predominant and Th2/Th17^low. G-H: Expression of IL4 and IL17 (G), and co-expression of GATA3 and RORγT (H) in BAL cells by immunofluorescence staining. Cytospins of BAL cells were double-stained for IL17 or GATA3 (green) and IL4 or RORγT (red) and then counterstained with DAP (blue) for nuclear staining. For GATA3/RORγT, Z-series images of a single lymphocyte were captured using a confocal microscope. The images from a mid-section show co-expression of GATA3 and RORγT in the nucleus of a single cell. Representative images from 3 separate experiments done with BAL cells from 3 different donors are shown.

In another approach we analyzed BAL cells for CD4 and CD68 (a different macrophage marker) or CD4 and CD3ε (Figure E3A–C) The vast majority of CD4 cells were CD68-CD3ε+, supporting their T cell origin. There was a smaller population of cells that was CD4-CD3ε+, which are likely CD8 and gamma/delta T cells. Both CD4+CD3+ and CD4-CD3+ T cell populations contained variable numbers of single IL4+ and IL17+ cells and dual IL4/IL17+ cells (Figure E3D&E). In general we observed two staining patterns—in one pattern the cytokine+ cells were clearly separated from the non-stained cytokine- cells (Figure E4A&B for IL4, Fig. 1D&E and Figure E4C for IL4/IL17, and Figure E4-E for IL17). In another pattern the cytokine+ cells showed a small but measurable shift in staining intensity, which made them positive for cytokine expression but they did not clearly separated from the cytokine negative cells (Fig. 1B, C&E for IL4, E4D for IL4/IL17, and Fig 1D and E4F for IL17). The dual IL4/IL17+ were usually better separated from the cytokine negative cells and showed a higher level of MFI. Figure 1F presents a heat map profile of single IL4+ and IL17+ cells and dual IL4/IL17+ CD4 T cells in BAL from all 52 asthmatic patients as studied flow cytometry. The heat map demonstrates 3 different BAL CD4 T cell profiles in asthma: 1). Th2 (IL4)^predominant, 2). Th2/Th17 (IL4/IL17)^predominant, and 3). Th2/Th17^low.

In order to confirm the flow cytometric result of co-expression of IL4 and IL17 we performed immunocytochemical staining of BAL cells (Fig. 1G). The frequency of cytokine+ cells in 3 different BAL samples was 22 ± 7%. We detected cells positive for IL4 or IL17 as well as cells co-expressing IL4 and IL17. These cells were relatively small in size and round in shape with a large nucleus and small perinuclear cytoplasm. IL4 showed a large vesicular staining pattern. In contrast, IL17 presented a diffuse and small vesicular staining pattern. The staining of single as well as double positive cells in the same BAL specimen further validated the staining specificity of the employed antibodies.

Co-expression of GATA3 and RORγT in BAL lymphocytes

The differentiation of Th2 and Th17 cells are controlled by the transcriptional regulators—GATA3 and RORγT, respectively. We examined co-expression of GATA3 and RORγT by immunofluorescence staining and confocal microscopy in BAL lymphocytes from 4 asthmatic patients. Figure 1H shows GATA3 and RORγT staining pattern of a representative BAL lymphocyte from an asthmatic patient. This image represents the mid-section from a Z-series. The entire Z-series images from this patient and the representative image from all 4 donors are shown in repository figure E5A&B. The staining with the DNA-binding dye DAPI shows the nuclear staining pattern. The less dense and transcriptionally active euchromatin stained light blue whereas the tightly packed and transcriptionally inactive heterochromatin stained dark blue. Both GATA3 and RORγT stained primarily the nucleus. The staining was most prominent with euchromatin and negligible with heterochromatin. Immunostaining with isotype control antibodies is shown in repository figure E2E.

Expression of dual positive IL4/IL17 cells is associated dual positive pSTAT6/pSTAT3 and CCR6/CRTH2 expressing cells

The phosphorylation and activation of STAT6 by IL4 and STAT3 by IL6 and IL21 are an early event during differentiation of Th2 and Th17 cells.^{27, 28} We examined the expression of activating phosphorylation of STAT3 and STAT6 in BAL CD4 T cells. We observed significant co-expression of pSTAT3 and pSTAT6, which followed the pattern of IL4 and IL17 co-expression (Repository figure E6A&B). There was a strong correlation between pSTAT3 and pSTAT6 (Repository figure E6C). These results suggest that the phosphorylation of STAT3 and STAT6 is an active and ongoing process in the BAL milieu from some patients with asthma. Dual positive IL4/IL17cells in the peripheral blood express the Th2 marker CRTH2+ and the Th17 marker CCR6.¹² In agreement, we detected CCR6+CRTH2+ cells in BAL (Repository figure E7A). The frequency of CCR6+CRTH2+ cells correlated with IL4/IL17+ cells (Repository figure E7B&C).

Increased expression of dual positive Th2/Th17 cells correlates with airway hyperreactivity, blood and BAL eosinophil counts and BAL lymphocytes

We compared the expression of BAL dual positive Th2/Th17 cells between 52 asthmatic patients and 25 disease controls. The disease controls included 16 patients with chronic cough and with a concurrent diagnosis of allergic rhinitis and GERD and were refractory to treatment. Additional controls included 3 patients with bronchiectasis, 5 patients with chronic aspiration and 1 patient with COPD. The number of dual positive cells was significantly (P=0.0006) increased in asthma as compared disease controls (Figure 2A). The median frequency of dual (IL4 and IL17) positive cells was 5% and 1% in asthma and disease controls, respectively. The median frequency of single IL4+ CD4 T cells was 11% and 1% in asthma and disease controls, respectively. Th2/Th17 cells could differentiate from either single positive Th2 cells or single positive Th17 cells. A previous study showed that Th2 cells differentiated into dual positive Th2/Th17 in vitro under the influence of Th17-inducing cytokines.¹² We observed a strong correlation between Th2 cells and Th2/Th17 cells (Figure 2B) supporting this notion. Both Th2 and Th2/Th17 cells correlated negatively with bronchial hyperreactivity (PC20 for methacholine), BAL and blood eosinophils and BAL lymphocytes (Figure 2C). In each case the correlation was stronger for Th2/Th17. There was no correlation with FEV1 or BAL neutrophils.

Figure 2.

BAL Th2/Th17 cells and clinical correlations. A: Comparison of BAL dual positive Th2/Th17 cells between asthmatic patients and disease controls (see results for description of disease controls). B: Correlation between Th2/Th17 and Th2 cells in BAL. C: Correlation of BAL Th2, Th17 and Th2/Th17 cells and clinical parameters. D: Comparison of BAL IL17A between asthmatic patients and disease controls. E. Correlation between BAL IL17A and FEV1.

The IL17 level is increased in BAL fluid from asthmatic patients and negatively correlates with FEV1

Since we detected expression of IL17 by Th2/Th17 cells, we measured secreted IL17 in the BAL fluid by ELISA. The IL17 level was significantly elevated in asthma as compared to disease controls (Figure 2D). The IL17 level negatively correlated with FEV1 (%) in asthmatic patients (Figure 2E). This correlation is not unexpected as IL17 enhances airway smooth muscle contraction.²⁹ Note that the Th2/Th17 number showed a trend to correlate with FEV1 but did not reach the statistical significance (P=0.07). It is likely that the amount of secreted IL17 is variable among Th2/Th17 cells, which could help explain this minor discrepancy.

Dexamethasone fails to inhibit BAL-derived dual positive Th2/Th17 cells

Most of the patients were referred to us for refractory asthma and relative steroid resistance. For this reason we examined sensitivity of Th2/Th17 to dexamethasone. One of the dexamethasone-responsive genes is MAP kinase phosphatase 1(MKP1).³⁰ We observed differential expression of MKP1 in T cells after treatment with dexamethasone. Dexamethasone increased the number of cells that expressed higher levels of MKP1 (Figure 3A&B). This population distinguished itself by expressing an intermediate level of CD4. On the other hand the number of T cells that expressed high levels of CD4 and low levels of MKP1 decreased after dexamethasone. Figure 3C&D show a similar differential pattern of dexamethasone response in 6 study subjects. These studies suggest that different populations of CD4 T cells respond differently to dexamethasone. Next we studied the effect of dexamethasone on Th2, Th17 and dual positive Th2/Th17 cells. Dexamethasone downregulated IL4+ Th2 cells (Figure 4A–E) but failed to show any inhibitory effect on dual positive Th2/Th17 cells (Figure 4C, D&F). Thus, Th2/Th17 cells are relatively resistant to the apoptotic effect of dexamethasone. Dexamethasone actually increased the number of Th2/Th17 cells in some patients.

Figure 3.

Effect of Dexamethasone on BAL cell expression of MKP1. BAL cells were cultured overnight (16 hr) with medium alone or dexamethasone (10⁻⁷ M). The expression of MKP1 in CD4^high and CD4^low T cells were assessed by flow cytometry (A&B) and quantified (C&D).

Figure 4.

Effect of dexamethasone on BAL single IL4+ and dual IL4/IL17+ CD4 T cells. BAL cells were cultured with medium or dexamethasone as above in Figure 3 and then analyzed for expression of IL4 and IL17 by flow cytometry as per Figure 1. Representative flow cytograms from two patients, one with dominant IL4+ cells (A&B) and the other with both IL4 and dual IL4/IL17+ cells (C&D) are shown. E–F: Effect of dexamethasone (dex) on Th2 and Th2/Th17 cells from 14 asthmatic patients. The numbers on the top of the graphs represent statistical significance.

BAL dual positive Th2/Th17 cells express higher levels of MEK1

One of the signaling pathways that antagonize the inhibitory action of glucocorticoids is the MEK-ERK1/2 pathway.^{24, 31, 32} This pathway induces the activating protein 1 (AP1) transcription factors, which antagonize glucocorticoids. Conversely, glucocorticoids antagonize AP1 by inducing GILZ.³³ We have previously reported that CD4 T cells from moderate to severe asthmatic patients have increased expression of MEK1.²⁴ Inhibition of MEK1 reverses T cell resistance against dexamethasone. We analyzed BAL CD4 T cells for MEK1 (referred to as MEK) expression and the sensitivity of MEK positive cells to dexamethasone. We observed a small CD4 population in BAL that expressed a high level (MFI 133) of MEK (MEK^hi) (Figure 5A). There was a larger population of CD4 T cells that expressed a low level (MFI 10) of MEK (MEK^low). The MEK^high CD4 T cell population was completely resistant to dexamethasone-induced cell death (Figure 5A&B). This contrasted with the MEK^low population, whose cell count decreased by 2.7 fold after dexamethasone treatment. Next we asked if MEK expression was differentially regulated in dual positive Th2/Th17 cells. To this objective we analyzed the MEK^high and MEK^low cell populations (Figure 5C&D) for enrichment of Th2/Th17 cells. Our analysis showed that MEK^high CD4 T cells were disproportionately enriched for dual positive Th2/Th17 cells (Figure 5E–G). In contrast, the MEK^low T cell population had significantly reduced number of Th2/Th17 cells. The mean frequency of Th2/Th17 cells was 66 ± 12% (median 80) in the MEK^high population as compared to 36 ± 9% (median 22) in the MEK^low population. As anticipated, the MEK^high population was also enriched for dual pSTAT3/pSTAT6 positive cells (Figure 5H).

Figure 5.

Preferential expression of dual positive Th2/Th17 cells in the BAL MEK^high CD4 T cell population. A&B: Sensitivity of MEK^high and MEK^low CD4 populations to dexamethasone-induced cell death. MFI: mean fluorescence intensity; data is presented as percentage and absolute cell count (in the parenthesis) per boxed area. C: Gating strategy for BAL CD4 T cells. D: Gating strategy for separation of MEK^high and MEK^low cells in CD4 gated cells. E&F: Expression of dual positive Th2/Th17 cells in MEK^high and MEK^low BAL CD4 T cell populations. G&H: Comparison of expression of dual IL4/IL17 positive (G) and pSTAT3/pSTAT6 positive (H) cells in MEK^high and MEK^low cell populations.

The Th2/Th17^predominant subgroup has the most severe airway hyperreactivity and obstruction

Based upon the frequency of Th2 and Th2/17 cells in BAL our patient population can be divided into three separate subgroups (Figure 6A–C)—1). Th2^predominant; 2). Th2/Th17^predominant, and 3). Th2/Th17^low. The Th2^predominant subgroup (22 patients) had the highest frequency of single IL4+ cells as compared to Th2/Th17^predominant and Th2/Th17^low groups (Table-2). The Th2/Th17^predominant subgroup (15 patients) had the highest number of dual positive IL4+IL17+ cells as compared to Th2^predominant and Th2/Th17^low groups. The Th2/Th17^low subgroup (15 patients) had 5% or less of each of these cell populations in BAL. Interestingly, although the Th2/Th17^predominant group had less single IL4+ cell number than the Th2^predominant group (12.6±2% vs 22.7±2%), the mean fluorescence intensity (MFI) of IL4 in IL4/IL17+ cells in this group was higher (304±65 vs 180±85) suggesting that these cells produce higher quantities of IL4 on a per cell basis. Next, we examined clinical and biochemical features of the identified subgroups. Airway hyperreactivity was most severe in the Th2/Th17^predominant subgroup (Figure 6D). The PC20 for methacholine were 1.24± 0.2, 2.9 ± 0.4 and 4.6 ± 0.5 mg/ml in the Th2/Th17^predominant, Th2^predominant and Th2/Th17^low subgroups, respectively. The differences were statistically significant. Airway obstruction was most severe in the Th2/Th17^predominant subgroup as compared to other subgroups (Figure 6E). FEV1 were 59.6 ± 2.7%, 73.9 ± 3% and 79.6 ± 2% in the Th2/Th17^predominant, Th2^predominant and Th2/Th17^low subgroups, respectively. The differences were statistically significant except for Th2^predominant vs. Th2/Th17^low. Blood eosinophilia was present in both Th2^predominant and Th2/Th17^predominant subgroups but was absent in the Th2/Th17^low subgroup (Figure 6F). There were no differences in the total IgE level among the subgroups. All study subjects were on a high dose inhaled corticosteroid (ICS)/long-acting bronchodilator (LABA) therapy. The number of patients on chronic systemic steroid therapy and omalizumab was highest in the Th2/Th17^predominant group.

Figure 6.

Th2^predominant, Th2/Th17^predominant and Th2/Th17^low subgroups of asthmatic patients and their clinical features. A–C: All asthmatic patients (52) were subgrouped based upon the dominant expression of IL4+ cells (Th2^predominant), dual IL4/IL17+ cells (Th2/Th17^predominant) or 5% or less of either cell types (Th2/Th17^low). D–F: Comparison of PC20 for methacholine, FEV1 and blood eosinophilia among the 3 subgroups.

Table 2.

Comparison of clinical features of Th2/Th17^low, Th2^predominant and Th2/Th17^predominant subgroups of asthmatic patients

Parameters	Th2/Th17low	Th2predominant	Th2/Th17predominant
N	15	22	15
Co-morbidities:
-Allergic rhinitis	11	19	8
-Chronic sinusitis	4	7	7
-GERD	9	9	6
-Bronchiectasis	0	1	0
-Aspiration	0	1	1
-Smoking	0	1	0
%Cytokine+ cells in the gated BAL cell population Mean ± SEM (Median)	IL4: 3.0±0.5 (3.5) IL17: 1.2±0.3 (1) IL4/IL17: 1.1±0.3 (1)	IL4: 22.7±2 (23)Ω IL17: 5.9±1.6 (4) IL4/IL17: 7.3±1.5 (5)	IL4: 12.6±2 (13) IL17: 6.5±1.5 (6) IL4/IL17: 20.4±4(16)*
MFI of Cytokine+ cells in the gated BAL cell population	ND	IL4: 125±43 (68) IL17: 34±7 (27) IL4 in IL4/IL17+ cells: 180±85 (83)	IL4: 103±11 (92) IL17: 34±2 (34) IL4 in IL4/IL17+ cells: 304±65(304)#
Positive skin test	8	19	12
Total IgE (KIU/L)	111 ± 26 (100)	123 ± 26(98)	230 ± 53 (167)
BMI	26.5 ± 2 (27)	29.0 ± 1.8 (30)	25.9 ± 1.2 (26)
FEV1 (%)	79.6± 2 (80)	73.9 ± 3 (74)	59.6 ± 2.7 (62)*#α
PC20 (mg/ml) for methacholine	4.6 ± 0.5 (4.5)	2.9 ± 0.4 (2.85)∞	1.24 ± 0.2(1.3)*#β
Eosinophils/ µL blood	113 ± 9 (100)	422 ± 85 (200)∞	433 ± 57 (400)*μ
Asthma medications
-systemic steroids	1	2	4
-omalizumab	0	1	3
-ICS/LABA	15	22	15
-LTI	2	5	1
-SABA	15	22	15
-Tiotropium	1	1	0

^*: P<0.05 compared to Th2^predominant and Th2/Th17^low;

^#: P<0.05 compared to Th2^predominant;

^Ω: P<0.05 compared to Th2/Th17^predominant and Th2/Th17^low;

^∞: P<0.05 compared to Th2/Th17^low, Mann-Whitney U test;

^α: P=0.0004;

^β: P=0.00007;

^μ: P=0.002 all Kruskal Wallis test for multiple comparisons.

Abbreviations: ND: not done; PC20: provocation concentration causing a 20% drop in FEV1; ICS: inhaled corticosteroid; LABA: long-acting beta-adrenergic agonist; LTI: leukotriene receptor inhibitor; and SABA: short-acting beta-adrenergic agonist

Discussion

The presence of allergic sensitivity (skin test positivity or IgE-specific antibody in the serum) can be detected in a vast majority but not all asthmatic patients.³⁴ In addition to IgE, mild to moderate eosinophilia is a characteristic feature of asthma. Interestingly, some non-allergic asthmatic patients also manifest blood eosinophilia. Both IgE antibody and eosinophilia are driven by a Th2-type immune response. Thus, the presence of Th2 cells in the lung and BAL is anticipated and has previously been reported.^35–41 The median frequency of BAL IL4+ CD4 T cells in asthmatic patients in the published reports has ranged from 5%⁴⁰ to 9%³⁶. We observed a median frequency of 11% IL4+ CD4 T cells in BAL from asthmatic patients. The relatively higher frequency could be due to the difference in severity of asthma and the number of patients studied. The previous studies were performed on a small number (N=11–12) of patients, whose asthma was relatively mild (median FEV1 90–100%). We studied 52 severe asthmatic patients whose median FEV1 was 73.5%. Th2-type cytokines, especially IL5 and IL13 have been recovered from BAL fluid obtained from asthmatic patients,^{36, 37} although some studies did not observe any increase in these cytokines.^{42, 43} IL4 and especially, IL13 acts upon epithelial cells and induces transcription of specific genes.^{44, 45} Microarray analysis of airway epithelial cells from asthmatic patients demonstrated an increase in IL4/IL13-responsive genes.⁴⁶ However, a significant number of asthma patients failed to show an increase in IL4/IL13-responsive genes. Based upon these findings Woodruff and colleagues have identified two subgroups of asthma—Th2^high and Th2^low.⁴⁶

As mentioned, most differentiated T helper cells manifest plasticity and acquire additional functional features.^5–10 Th2 cells can acquire the ability to produce Th17 cytokines without losing their ability to produce Th2 cytokines.¹¹ The dual positive cells emerge from Th2 cells in the presence of Th17-inducing cytokines—IL1β, IL6 and IL21.¹² These blood dual positive cells also express CCR6. The frequency of the Th2/Th17 population is increased in peripheral blood from asthmatic patients.¹²

Using a multi-color flow cytometry approach we detected Th2, Th17 and dual positive Th2/Th17 cells in BAL from asthmatic patients. The differentiation of Th2 and Th17 is regulated by GATA3 and RORγT, respectively. Immunofluorescence studies of BAL cells showed co-expression of IL4 and IL17 as well as nuclear co-localization of GATA3 and RORγT supporting the Th2/Th17 phenotype. IL4 and IL6 play a crucial in induction of Th2 and Th17, respectively. They do so by inducing phosphorylation of STAT6 and STAT3. We demonstrated the presence of dual positive pSTAT3/pSTAT6 cells further confirming the Th2/Th17 phenotype.

One question of interest is the mechanism of differentiation of dual positive Th2/Th17 cells. Both Th2 and Th17 cells could give rise to Th2/Th17 cells. However, the frequency of single positive Th17 was either lower than that of Th2 cells or undetectable in our study subjects. This contrasted with much higher frequency of Th2 cells in BAL from most asthmatic patients. Further, IL17 production in dual positive cells was usually associated with IL4^high CD4 T cells. These findings favor Th2 cells as the precursors for Th2/Th17 cells. The generation and sustenance of Th2 cells require the AP1 transcription factor JunB.⁴⁷ Some recent publications have demonstrated that JunB forms a trimolecular complex with Batf and IRF4.^48,49 This complex binds to the so-called AP1-IRF composite element (AICE) in the IL17 gene promoter and plays an important role in IL17 gene induction.⁵⁰ We have previously shown that JunB is a MEK inducible protein.²⁴ Thus, IL4^high Th2 cells with a high level of JunB could, under Batf-inducing condition, (e.g. stimulation with IL1) begin to form complexes with Batf and IRF4, and lead to IL17 production. MEK also induces the PEA family transcription factor Etv4.^51,52 The latter is required for Th17 induction.⁵³ Thus, MEK-driven JunB and Etv4 could promote IL17 production in IL4^high Th2 cells. The validity of this thesis needs to be proven in future studies.

A previous study has demonstrated that Th2/Th17 cells induce a more severe form of experimental asthma in an adoptive transfer model in mice when compared to Th2 and Th17 cells.¹² We observed more severe airway obstruction and airway hyperreactivity in the Th2/Th17^predominant subgroup. A simple explanation for increased severity of asthma could be the presence of Th2/Th17 cells. However, the MFI of IL4 in Th2/Th17 cells from the Th2/Th17^predominant subgroup is higher than that in Th2 cells from the Th2^predominant subgroup. This data is suggestive of a higher level of IL4 production by the Th2/Th17 cells in the Th2/Th17^predominant subgroup. Thus, the severity of asthma in this subgroup could be due to increased IL4 production. On the other hand, IL17 production by Th2/Th17 cells is likely to change the quality of airway inflammation and function. IL17 is known to directly affect airway epithelial cells, fibroblasts and smooth muscle cells.^{16–20, 29} In agreement, we observed a significant negative correlation between BAL IL17 levels and FEV1.

All asthma patients were on high dose inhaled steroids at the time of the study. Despites this treatment they had sustained airway obstruction suggesting relative steroid resistance. Th17 cells have previously been shown to be resistant to glucocorticoid.¹⁴ Our results suggest that the dual positive Th2/Th17 cells are also resistant to dexamethasone. This resistance is likely to contribute to the refractory nature of their asthma. A number of molecular mechanisms have been implicated in glucocorticoid resistance. Reduced glucocorticoid receptor (GR) phosphorylation by p38 MAPK (α and γ) leads to reduced nuclear translocation and confers resistance.⁵⁴ However, Th17 cells manifest normal nuclear translocation of GR receptor.¹⁴ The anti-inflammatory action of GR is facilitated by histone deacetylase 2 (HDAC2), a chromatic remodeling enzyme. Patients with severe asthma have reduced HDAC2, which could contribute to glucocorticoid resistance.⁵⁵ The delta isoform of phosphatidyl inositol-3 kinase phosphorylates HDAC2 and contributes to its degradation, which results in steroid resistance.⁵⁶ The transcription factor c-Fos (a component of AP1) directly antagonizes GR and confers resistance.³² c-Fos is downstream of the MEK-ERK1/2 signaling pathway. We have reported that severe asthma associated with increased expression of MEK1 in CD4 T cells.²⁴ Increased expression of MEK1 confers resistance against a broad spectrum of endogenous homeostatic regulators—TGFβ, IL10 and glucocorticoids through induction of c-Fos and JunB. Inhibition of MEK1 reverses this resistance. In this study we show that Th2/Th17 cells express higher levels of MEK1 and these MEK1^high CD4 T cells are resistant to dexamethasone.

Our BAL CD4 T cell analysis allows us to identify three distinct subgroups of asthma—Th2^predominant, Th2/Th17^predominant and Th2/Th17^low. Conceptually this is similar to the findings of Woodruff et al.⁴⁶ These investigators have identified Th2^high and Th2^low subgroups of asthma based upon epithelial microarray data. We have identified a subgroup of Th2^predominant patients whose BAL Th2 cells co-express IL17. This finding has mechanistic and clinical implications. It is well known that asthma exacerbation is frequently triggered by respiratory infections.⁵⁷ A subgroup of asthmatic patients has chronic lung infections with Mycoplasma and Chlamydia sp.⁵⁸ Many environmental factors— air pollution, smoke, and chemicals nonspecifically aggravate asthma.^{59, 60} Infections and these environmental factors are known to elicit inflammatory cytokines such as IL1β and IL6.^61,62 Both IL1β and IL6 induce differentiation of Th2/Th17 cells from Th2 cells. Thus, a typical allergic asthmatic patient with a Th2^predominant endotype could trend toward a Th2/Th17^predominant endotype over time if he or she suffers from recurrent respiratory infections and/or exposed to the aforementioned environmental toxicants.

The major weakness of this paper is the low sample size. The invasive nature of bronchoscopy and the potential for adverse events are a major deterrence for patient participation. Induced sputum, which is less invasive, is an alternative approach to obtaining airway cells. However, the number of cells obtained through induced sputum is much less than that through BAL. Isolation of cells from the sputum requires harsh mucolytic treatment, which may affect cellular function and viability. Flow cytometric analysis of sputum cells has showed greater variability in our hand. Another limitation is the low number of BAL CD4 T cells, which restricted our ability to further characterize Th2/Th17 cells. Unfortunately, there is no reasonable alternative to increasing BAL cell yield. Finally, all patients took routine medications at the time of the BAL. Medications can affect the T cell number and their cytokine expression. Unfortunately, there is no alternative. The IRB does not allow discontinuation of medications, especially in severe patients, whose asthma may deteriorate without medication.

In summary, we have demonstrated increased frequency of dual positive Th2/Th17 cells in BAL from asthmatic patients. The increased expression of Th2/Th17 cells and one of its cytokines—IL17 are associated with heightened airway hyperreactivity and airway obstruction, two objective features of asthma severity. Severe asthma manifests relative steroid resistance. We provide a mechanistic explanation for this resistance. Th2/Th17 cells express high levels of MEK1, which is associated with steroid resistance. The identification of a Th2/Th17^predominant endotype in addition to the previously recognized Th2^predominant and Th2^low endotypes of asthma has pathogenetic and therapeutic implications.

Abbreviation used

AP1

activating protein-1

BAL

bronchoalveolar lavage

FEV1

forced expiratory volume in 1 second

MEK1

MAP-ERK kinase 1

MKP1

MAP kinase phosphatase 1

RORγT

retinoic acid receptor-related orphan receptor gamma T