Introduction

Allergic bronchopulmonary aspergillosis (ABPA), an allergic pulmonary disease with Aspergillus fumigatus sensitisation, is usually diagnosed in patients with asthma and/or cystic fibrosis. The estimated global burden of ABPA is 5 million [1, 2]. ABPA patients are characterised by elevated total IgE and A. fumigatus-specific IgE in the serum. Central bronchiectasis and high attenuation mucus on high-resolution computed tomography (HRCT) are also indicative of the disease [3]. ABPA patients usually have high levels of blood eosinophil counts and viscous sputum impaction rich in eosinophils. Thus, ABPA is deemed a type 2 inflammation dominant disease. We demonstrated that the single nucleotide polymorphism of rs4077515 at CARD9 (S12N mutation) in ABPA patients facilitates the production of interleukin (IL)-5 and type 2 T-helper (Th2) cell-mediated pathological responses [4].

Th2-cells produce type 2 cytokines such as IL-4, IL-5 and IL-13 [5]. IL-5 plays a critical role in the differentiation, proliferation and chemotaxis of eosinophils [6]. IL-13 mediates mucus hypersecretion, airway hyperresponsiveness and tissue homing of eosinophils [7]. IL-5+ Th2 cells comprise only 20% of Th2 cells and appear after multiple rounds of induction in vitro, while IL-13+ Th2 cells are generated earlier [8]. Animal experiments showed stronger pro-inflammatory function of IL-5+ Th2 cells [9, 10]. IL-5+ Th2 cells were increased in the blood of patients with allergic diseases such as eosinophilic gastrointestinal disease [11, 12]. Thus, IL-5+ Th2 cells are recognised as highly differentiated [8] and pathogenic [12] subpopulations of Th2 cells. It was assumed for a long time that Th2 cells are essential to the production of IgE. However, recent studies showed that IgE response was entirely absent without follicular T-helper (Tfh) cells [13–15], while the Th2 cell responses such as the infiltration of eosinophils in the target tissue were unaffected. Tfh cells are essential for the generation of high-affinity, long-lived plasma cells in the germinal centre of lymph nodes [16]. One of the key molecules expressed by Tfh cells is IL-21, the most potent cytokine for the differentiation of plasma cells [17]. Animal asthma experiments showed that IL-21 could promote type 2 immunity to house dust mite [18].

Distinct subsets of Tfh cells have been identified and their functions illustrated [19, 20]. Tfh1 cells promote IgG production through the production of IL-21 and interferon (IFN)-γ. Tfh2 cells that co-express IL-4 and IL-21 are induced in helminth infection. Tfh13 cells, which co-express IL-4, IL-5 (not always), IL-13 and IL-21, were induced by a variety of allergens, but not helminth infection [20]. Tfh13 cells contribute to anaphylactic reaction through high-affinity IgE production. In the absence of Tfh13 cells, although a large quantity of total IgE was detected in the serum after allergen stimulation, anaphylactic response was abrogated due to the seriously impaired production of high-affinity IgE.

However, the Tfh cell is not the only T-cell subset that contributes to antibody production. Peripheral T-helper (Tph) cells (PD-1+CXCR5−CD4+ T-cells) boost IgG secretion by B-cells and contribute to the pathogenesis of rheumatoid arthritis [21]. Unlike Tfh cells, Tph cells were more abundant in synovium, synovial fluid and blood instead of the lymphoid tissue. The different expressions of CXCR5 (lymphoid follicle homing chemokine) and other chemokines/receptors might contribute to the distinct location preference of Tfh cells and Tph cells.

Classic therapies for ABPA include long-term usage of oral glucocorticoids and antifungal agents. After the treatment, the symptoms are relieved, the mucus impaction is alleviated and the serum total IgE decline. In most cases, patients at remission still have excessive serum total IgE. However, most patients experience repeated exacerbations manifested as significantly elevated serum total IgE, new pulmonary infiltrations and worsened symptoms. Some patients could be refractory to or dependent on classic drugs. The mechanism of ABPA pathogenesis and relapse still remain unclear. New therapies have been applied in ABPA patients and summarised in a review article [22]. For patients who were unresponsive to or dependent on corticosteroids and patients who suffered from the side-effects of corticosteroids, benefits were seen with the application of biologic drugs such as omalizumab, an anti-IgE monoclonal antibody (mAb); mepolizumab, an anti-IL-5 mAb; and dupilumab, an anti-IL-4 receptor-α mAb that inhibits both IL-4 and IL-13 cytokine signal. Thus, we wonder if there is an upstream target that might block these pathways.

To explore the crucial T-cell subsets in the pathogenesis of ABPA, we analysed circulating regulatory T-cells (Treg) cells and CD4+T-cells that express IFN-γ, IL-17, IL-5, IL-13 and IL-21 in ABPA patients with different clinical statuses. A subpopulation of Tph cells that co-expressed IL-5, IL-13 and IL-21 was identified. In addition, Tph cells were found in the airways of ABPA patients. Transcriptome data revealed the unique expression profile of circulating Tph cells in ABPA patients. In vitro co-culture showed that Tph cells of ABPA patients could induce the differentiation of B-cells and promote the secretion of IgE. These findings showed that circulating IL-5+IL-13+IL-21+ Tph cells may play a critical role in ABPA pathogenesis and might be a potential biomarker and therapeutic target.

ResultsIL-5+IL-13+IL-21+CD4+T-cells were identified in ABPA

The gating strategies of circulating CD4+T-cells are shown in supplementary figure S1a. We tested IFN-γ+Th1, IL-17+Th17, IL-5+/IL-13+Th2 and Treg cells. ABPA patients had decreased Th1 and Th17 cells (median 5.07% and 0.29%, respectively) when compared to healthy controls (median 8.82% and 0.53%, respectively) (supplementary figure S1b). However, the difference in Th1 and Th17 cells between exacerbated (median 5.70% and 0.32%, respectively) and nonexacerbated (median 5.03% and 0.29%, respectively) patients was not significant. Treg cells were similar between ABPA patients (median 2.03%) and healthy controls (median 2.36%) (supplementary figure S1c). ABPA patients had more IL-5+ (supplementary figure S1d), IL-13+ (supplementary figure S1e) and IL-5+IL-13+CD4+T-cells (figure 1a) (median 0.79%, 2.31% and 0.72%, respectively) than healthy controls (median 0.0935%, 0.8% and 0.08%, respectively) and asthma patients (median 0.26%, 1.6% and 0.23%, respectively). In subgroup analysis, these cells were more abundant in exacerbated ABPA patients (median 1.13%, 2.94%, 1.07%, respectively) than nonexacerbated patients (median 0.43%, 1.64% and 0.37%, respectively). Most IL-5+CD4+T-cells co-expressed IL-13 (supplementary figure S1f), while fewer IL-13+CD4+T-cells co-expressed IL-5 (figure 1b). The ratio of IL-5+IL-13+ cells to IL-13+CD4+T-cells of ABPA patients (median 33.90%) was double that of asthma patients (median 16.68%; p<0.0001) and triple that of healthy controls (median 11.75%; p<0.0001) (figure 1b). This discrepancy was greater in exacerbated ABPA patients (median 37.64%) than nonexacerbated ones (median 22.47%, p<0.0001) (figure 1b). Thus, we found abundant circulating IL-5+/IL-13+ Th2 cells in ABPA patients, especially exacerbated ones.

FIGURE 1

Interleukin (IL)-5+IL-13+IL-21+CD4+T-cells were identified in allergic bronchopulmonary aspergillosis (ABPA). a) Gating (left) and the ratio of IL-5+IL-13+CD4+ T-cells in the blood. b) The percentage of IL-5+IL-13+ cells in IL-13+CD4+T-cells in the blood. c) Gating (top) and the ratio of IL-21+CD4+ T-cells in the blood. d) Gating (left) and the ratio of IL-13+IL-21+CD4+ T-cells in the blood. e) Gating (left) and the ratio of IL-5+IL-13+IL-21+CD4+ T-cells in the blood. f) Receiver operating characteristics (ROC) curve analyses of IL-5+IL-13+IL-21+CD4+T-cells in the blood for the discrimination of ABPA patients from healthy controls (HC)/asthma patients (left) and ABPA exacerbation (right). g) The percentage of IL-5+IL-13+IL-21+ cells in IL-21+CD4+T-cells in the blood. HC (n=22), asthma patients (n=23) and ABPA patients (n=67, exacerbated (E, n=36), nonexacerbated (non-E, n=31)). Data are presented as median (interquartile range). ns: nonsignificant (p>0.05); AUC: area under the curve. *: p<0.05, **: p<0.01, ***: p<0.001, ****: p<0.0001. ANOVA with Dunnett's multiple comparisons or Mann–Whitney test.

Unlike Th2 cells, the proportions of IL-21+CD4+T-cells in the blood were similar among ABPA (median 2.44%), asthma patients (median 2.05%) and healthy controls (median 1.75%) (figure 1c). Exacerbated ABPA patients (median 3.05%) had more IL-21+CD4+T-cells than nonexacerbated patients (median 2.05%) (p<0.001). Unexpectedly, ABPA patients, especially the exacerbated ones, contained a distinct subgroup of CD4+T-cells that co-expressed IL-21 and Th2 cell cytokines IL-5/IL-13, while this was rarely seen in the blood of healthy controls (figure 1d and e). The frequency of IL-5+IL-13+IL-21+CD4+T-cells in ABPA patients (median 15 000 per 107 lymphocytes) was 19.38 times that of healthy controls (median 774 per 107 lymphocytes, p<0.0001) and 3.13 times that of asthma patients (median 4800 per 107 lymphocytes, p<0.01) (figure 1e). Moreover, the frequency of IL-5+IL-13+IL-21+CD4+T-cells in exacerbated ABPA patients (median 31 000 per 107 lymphocytes) was 5.17 times that of nonexacerbated patients (median 6000 per 107 lymphocytes, p<0.0001) (figure 1e). The ROC analysis showed that the frequency of IL-5+IL-13+IL-21+CD4+T-cells could help distinguish ABPA patients from healthy controls (area under the curve (AUC) 0.9739, 95% CI 0.9402–1.008; p<0.0001) and asthma patients (AUC 0.7992, 95% CI 0.7057–0.8926; p<0.0001) (figure 1f). The cut-off value of the frequency of IL-5+IL-13+IL-21+CD4+T-cells per 107 lymphocytes to discriminate ABPA from healthy control and asthma was 3200 (sensitivity 95.45%, specificity 95.45%) and 5850 (sensitivity 76.12%, specificity 78.26%). The threshold of 17 500 (the frequency of IL-5+IL-13+IL-21+CD4+T-cells per 107 lymphocytes) resulted in a retrospective sensitivity and specificity of 83.33% and 93.55%, respectively, for the exacerbation of ABPA (AUC 0.9117, 95% CI 0.8407–0.9828; p<0.0001) (figure 1f). The percentage of IL-5+IL-13+IL-21+ cells in the IL-21+CD4+T-cell was higher in ABPA patients, especially the exacerbated ones (figure 1g). Thus, the levels of IL-5+IL-13+IL-21+CD4+T-cells might help identify ABPA patients and their exacerbation status.

Most IL-5+IL-13+IL-21+CD4+T-cells in the blood were Tph cells

Gowthaman et al. [20] reported a new subgroup of Tfh cells essential to the production of anaphylaxis-inducing high-affinity IgE and named it Tfh13 for its co-expression of IL-13 and IL-21. Some Tfh13 cells also express IL-5. Thus, we wondered if the IL-13+IL-21+ and IL-5+IL-13+IL-21+CD4+T-cells in the blood of ABPA patients were PD-1+CXCR5+ Tfh13 cells. Since phorbol 12-myristate 13-acetate and ionomycin may induce the loss of CXCR5 (data not shown), CD4+ T-cells were sorted according to the expression of PD-1 and CXCR5 before cytokine induction (figure 2a). Unexpectedly, most IL-13+IL-21+ and IL-5+IL-13+IL-21+CD4+T-cells in the blood were PD-1+CXCR5− Tph cells (figure 2b).

FIGURE 2

Most interleukin (IL)-5+IL-13+IL-21+CD4+T-cells in the blood were peripheral T-helper (Tph) cells. a) Multigating strategy for flow cytometry sorting of PD-1+CXCR5+, PD-1+CXCR5−, PD-1−CXCR5+ and PD-1−CXCR5−CD4+T-cells in the blood. b) Comparing the percentages of IL-13+IL-21+ (top) and IL-5+IL-13+IL-21+ (bottom) PD-1+CXCR5+, PD-1+CXCR5−, PD-1−CXCR5+ and PD-1−CXCR5−CD4+T-cells in the blood of exacerbated allergic bronchopulmonary aspergillosis (ABPA) patients, nonexacerbated ABPA patients and asthma patients. Data were paired in each plot. c) Gating of PD-1+CXCR5−Tph cells in the bronchoalveolar lavage fluid (BALF) of ABPA patients. d) Gating of PD-1+CXCR5+, PD-1+CXCR5−, PD-1−CXCR5+ and PD-1−CXCR5−CD4+T-cells in the blood (left) and the percentages of PD-1+CXCR5−Tph cells in the blood (right). e) Summary plots of the percentages of IL-13+IL-21+ (left) and IL-5+IL-13+IL-21+ (right) cells in Tph cells (top). Summary plots of the percentages of IL-13+IL-21+ (left) and IL-5+IL-13+IL-21+ (right) Tph cells in the blood (bottom). Healthy controls (HC, n=7), asthma patients (n=6), non-exacerbated ABPA patients (non-E, n=10) and exacerbated ABPA patients (E, n=11). Data are presented as mean±sd. SSC: side scatter; FSC: forward scatter; ns: nonsignificant (p>0.05). *: p<0.05, **: p<0.01, ***: p<0.001, ****: p<0.0001. Dunnett's multiple-comparison test.

Next, we analysed the bronchoalveolar lavage fluid (BALF) of three ABPA patients with pulmonary infiltrates and mucoid impaction on HRCT. As expected, Tph cells were also found in the bronchial lumen of ABPA patients, while CXCR5+CD4+T-cells were rarely seen (figure 2c).

Neither the percentages of Tph cells (figure 2d) nor the three subsets gated by the expression of PD-1 and CXCR5 (supplementary figure S2a) in the blood significantly differed among ABPA patients, asthma patients and healthy controls. Besides, the median fluorescence index (MFI) of PD-1 on the Tph cells was dispersive and similar between the three groups (supplementary figure S2b). However, ABPA patients had more Tph cells that co-expressed IL-5, IL-13 and IL-21 (figure 2e), suggesting that Tph cells were heterogeneous and ABPA patients had more Th2-skewed Tph cells.

The heterogeneity of Tph cells was affirmed by transcriptome analysis

We conducted bulk RNA sequencing of PD-1+CXCR5− Tph cells and PD-1−CXCR5−CD4+T-cells from exacerbated ABPA patients (n=2) and healthy controls (n=2), respectively. As we expected, the transcriptome profiles of Tph cells in both groups shared some features of Tph cells reported earlier, yet the Tph cells of ABPA patients were distinctly different to those of healthy controls. They varied in gene ontology terms related to inflammatory response, cytokine activity, cytokine-mediated signalling pathway, etc. (figure 3a and supplementary table S2). Pathway interactions were observed (figure 3b). >700 genes were dysregulated (false discovery rate adjusted p<0.05, fold change >2; supplementary table S3).

FIGURE 3

The heterogeneity of peripheral T-helper (Tph) cells was affirmed by transcriptome analysis. a) The scatterplot and b) chord diagram of gene ontology (GO) enrichment analysis between the Tph cells of allergic bronchopulmonary aspergillosis (ABPA) patients and healthy controls (HC). c) Heatmap of differentially expressed genes between Tph cells and PD-1−CXCR5−CD4+T-cells in ABPA patients (n=2) and healthy controls (n=2). d) Volcano plot of differentially expressed genes in the Tph cells of exacerbated ABPA patients compared to healthy controls. Red datapoints represent genes significantly upregulated in ABPA patients. Blue datapoints represent genes significantly downregulated in ABPA patients. FC: fold change

Tph cells of ABPA patients and healthy controls shared some common expression profiles when compared to PD-1−CXCR5−CD4+T-cells (figure 3c). Tph cells of both groups expressed elevated Tfh cell related genes such as MAF, TIGIT, SLAMF6 and TOX (figure 3c). Tph cells of both groups downregulated CCR7 and upregulated genes such as HLA-DRB5, KLRB1(also known as CD161), IL2RB, etc.

Meanwhile, the Tph cells of ABPA patients were remarkably different from the Tph cells of healthy controls. ABPA patients upregulated genes related to chemotaxis, type 2 inflammation and T-cell activation, including CXCL2, CXCL3, CXCL8, IL5, IL13, IL6, SOCS1, HPGDS, GATA3-AS1, STAT6, JUNB, CCR3, IL9R, CD80 and S100A8 (figure 3d). Although PD-1−CXCR5−CD4+T-cells of ABPA also upregulated genes such as PIM3, CXCL16, CD69, CXCR4 and JUNB as autologous Tph cells (figure 3c), a number of type 2 inflammation relative genes such as FCER1G, HPGDS, IL17RB (IL-25 receptor), FFAR2 and TNFRSF4 (also known as OX40) were only upregulated in Tph cells of ABPA patients. We further sorted OX40+ and OX40− Tph cells (supplementary figure S3a). IL-5+IL-13+IL-21+CD4+T-cells distributed into both subsets, but the ratio was significantly higher in OX40+ Tph cells (supplementary figure S3b).

Circulating IL-5+IL-13+IL-21+CD4+Th cells were relative to ABPA disease status

ABPA patients were followed-up to record the dynamic change of cytokine-expressing CD4+T-cells in the circulation. For exacerbated ABPA patients, the percentages of IL-5+CD4+T-cells, IL-5+IL-13+IL-21−CD4+T-cells and IL-5+IL-13+IL-21+CD4+T-cells declined after the treatment (figure 4a and supplementary figure S4a). Two ABPA patients relapsed during follow-up. The level of their IL-5+CD4+T-cells, IL-5+IL-13+IL-21−CD4+T-cells and IL-5+IL-13+IL-21+CD4+T-cells in the second exacerbation exceeded the previous one. For clinically stable patients, the levels of these cells fluctuated within certain ranges, but were not high enough for another exacerbation. After 1 month of treatment, the MFI of IL-5, IL-13 and IL-21 also decreased in IL-5+, IL-13+ and IL-21+CD4+T-cells, respectively (figure 4b and c), indicating that the expression of these cytokines was also weakened after the therapy.

FIGURE 4

Circulating interleukin (IL)-5+IL-13+IL-21+ T-helper cells were relative to allergic bronchopulmonary aspergillosis (ABPA) disease status. a) The dynamic change of circulating IL-5+CD4+T-cells (left) and IL-5+IL-13+IL-21+CD4+T-cells (right) of ABPA patients. Patients in the repeated exacerbation (repeated E) group (n=2) experienced exacerbation twice; single exacerbation (single E) (n=13) and stable (n=3). Time point interval is 1 month. For patients in “repeated E” and “single E” groups, time point 1 is before treatment. b) The representative plot of IL-5 fluorescence distribution of IL-5+CD4+ T-cells before and 1 month after the treatment of an exacerbated ABPA patient. c) The median fluorescence index (MFI) of IL-5 (top), IL-13 (centre) and IL-21 (bottom) in IL-5+, IL-13+ and IL-21+CD4+T-cells, respectively, (n=11) at exacerbation and 1 month after the treatment. *: p<0.05, **: p<0.01 (t-test). d) Correlation of blood eosinophil counts with IL-5+IL-13+IL-21+CD4+T-cells (left) in exacerbated ABPA (n=36), nonexacerbated ABPA (ABPA non-E, n=29) and asthma patients (n=23). Correlation of serum total IgE with IL-5+IL-13+IL-21+CD4+T-cells (centre) in exacerbated ABPA (n=35), nonexacerbated ABPA (n=30) and asthma patients (n=23). Correlation of serum Asf-IgE with IL-5+IL-13+IL-21+CD4+T-cells (right) in exacerbated ABPA patients (n=34). Spearman test.

We analysed the correlation of lab parameters with cell subset abundance. For exacerbated ABPA patients, the serum total IgE level correlated well with the A. fumigatus specific IgE level (supplementary figure S5a). Thus, patients at severe A. fumigatus sensitisation status also produced more unspecific IgE. Blood eosinophil counts were significantly relative to IL-5+CD4+T-cells in nonexacerbated ABPA patients (supplementary figure S5b) and IL-5+IL-13+IL-21+CD4+T-cells in ABPA patients (figure 4d). Notably, IL-5+CD4+T-cells and IL-5+IL-13+IL-21+CD4+T-cells were much less abundant in the blood of asthma patients who have comparable blood eosinophil counts with ABPA patients. The IL-5+IL-13+IL-21−CD4+T-cells correlated well with IL-5+IL-13+IL-21+CD4+T-cells (supplementary figure S5c). However, there was no significant correlation between serum total/A. fumigatus specific IgE and CD4+ T-cells co-expressing IL-5/IL-13 and IL-21 (figure 4d and supplementary figure 5d and e), indicating that the amounts of these cells were more relevant to ABPA exacerbation events than the sensitisation severity.

FIGURE 5

Peripheral T-helper (Tph) cells of allergic bronchopulmonary aspergillosis (ABPA) patients induced B-cell differentiation and IgE secretion in vitro. a) Schematic representation of Tph, follicular T-helper (Tfh) and B-cell sorting from peripheral blood of ABPA patients/healthy donors and T-/B-cell co-culture in vitro. b) Multigating strategy for sorting of B-cells. c) Multigating strategy of CD38+CD27+ plasmablasts after co-culture in vitro for 7 days. d) The percentages of CD38+CD27+ plasmablasts after co-culture in vitro for 7 days using cells from five ABPA donors and seven healthy donors. ANOVA with Tukey's multiple comparisons. e) IgG and IgE concentration in the supernatants of co-cultures using cells from four ABPA donors. Dunnett's multiple comparisons test. f) Representative ELISpot dot plots of IgG plasmablasts after co-culture in vitro. g) The frequency of IgG plasmablasts after co-culture in vitro from four ABPA and three healthy donors. SSC: side scatter; FSC: forward scatter; SEB: staphylococcal enterotoxin B; ns: nonsignificant (p>0.05). *: p<0.05, **: p<0.01. Dunnett's multiple comparisons test.

Tph cells of ABPA patients induced B-cell differentiation and IgE secretion in vitro

We tested whether Tph cells of ABPA patients could provide B-cells help in regard to B-cell differentiation and antibody production in vitro. Tph cells and Tfh cells of ABPA patients were co-cultured with B-cells from ABPA patients and healthy controls (figure 5a). The sorting strategy of Tph/Tfh cells (figure 2a) and B-cells is shown (figure 5b). PD-1+CXCR5− Tph cells isolated from the peripheral blood mononuclear cells of ABPA patients induced the differentiation of co-cultured autologous and heterologous B-cells into plasmablasts as did PD-1+CXCR5+ Tfh cells (figure 5c and d). We tested the cell culture fluid using ELISA and found that the B-cells of ABPA patients produced quite an amount of IgG antibodies without constant T-cell help in vitro (figure 5e). Autologous Tph cells slightly enhanced the production of IgG (figure 5e). It was similar when we further measured the frequency of IgG-secreting cells using ELISpot and observed that Tph cells induced B-cells to secrete IgG antibody (figure 5f and g). However, without the help of Tph cells, IgE produced by B-cells of ABPA patients was scarce (figure 5e). Tph cells from the blood of ABPA patients significantly enhanced the IgE production by co-cultured autologous B-cells (figure 5e). The amount of IgE produced was even higher when compared to that of Tfh cell co-culture (figure 5e).

Discussion

In this study, we reported the expansion of Th2 cells and Th2-skewed Tph cells in the circulation of ABPA patients, especially the exacerbated ones. Tph cells were also found in the airway of ABPA patients. Flow cytometry and transcriptome data co-ordinately suggested that Th2-skewed Tph cells prevailed in exacerbated ABPA patients. When co-cultured, Tph cells of ABPA patients induced B-cell differentiation and IgE secretion. These findings helped to reveal the pathogenesis and exacerbation mechanism of ABPA. Tph cells might be a useful biomarker and potential target in the treatment of ABPA.

In a recent study by Chen et al. [24] in which the exacerbation status of ABPA was not specified and the markers of cell subsets were different, ABPA patients had increased IFN-γ+Th1 cells and similar percentages of IL-4+Th2 cells and Treg cells when compared with asthma patients. However, in our study, ABPA patients had decreased Th1/Th17 cells, increased IL-5+Th2 cells and similar Treg cells when compared with healthy controls. The difference in the study design and markers selected might have caused the discrepancy. Previous research showed that only a small part of Th2 cells were IL-5+ [8]. IL-5+Th2 cells have enhanced pro-inflammatory function in comparison to IL-5−Th2 cells [11]. In our research, ABPA patients, especially the exacerbated ones, had more IL-5+ cells and higher percentages of IL-5+ cells within Th2 cells, suggesting robust type 2 inflammation status in ABPA patients.

It was previously assumed that Th2 cells contribute to the production of IgE. However, research [13, 14] proved the importance of Tfh cells in IgE production. A new type of Tfh cells named Tfh13 cells has been reported [20]. This was the first research reporting IL-21+ Tfh cells co-expressing IL-13 and even IL-5. When Tfh13 cells were depleted, the amount of total IgE detected after allergen exposure declined slightly, but the production of high-affinity IgE was severely impaired and anaphylactic response was abrogated. Serum total IgE levels of ABPA patients usually decrease after the onset of glucocorticoid therapy and increase again when relapse happens. During the follow-up of the ABPA cohort, we observed that the fluctuation of circulating IL-13+IL-21+ and IL-5+IL-13+IL-21+CD4+T-cells synchronised with disease status and treatment events. Therefore, we wonder whether the expanded IL-13+IL-21+ and IL-5+IL-13+IL-21+CD4+T-cells in the blood of exacerbated ABPA patients were Tfh13 cells. Unexpectedly, most IL-13+IL-21+ and IL-5+IL-13+IL-21+CD4+T-cells in the blood were Tph cells rather than Tfh cells. Tph cells were initially reported in rheumatoid arthritis and induced the production of IgG antibodies, the biomarker and pathogenic molecules of rheumatoid arthritis [21]. The name “peripheral” was used as Tph cells lacked lymphoid follicle orienting chemokine receptor CXCR5 and were abundant in the blood, synovial fluid and synovial tissue of rheumatoid arthritis patients. The co-culture of Tph cells and B-cells from rheumatoid arthritis patients promoted plasma cell differentiation and IgG antibody secretion. Then, Tph cells were also reported in other autoimmune diseases or infection disease [25–27]. We confirmed that most IL-5+IL-13+IL-21+CD4+T-cells in the blood of ABPA patients were PD-1+CXCR5− Tph cells. It was similar in asthma, another allergic pulmonary disease, although IL-5+IL-13+IL-21+ Tph cells were less abundant in asthma patients. These “Tph13” cells that co-express IL-21 and Th2 cytokines IL-13 and/or IL-5 might be the peripheral counterparts of Tfh13 cells and promote the local production of IgE. Although PD-1−CXCR5−CD4+T-cells in the blood also contained cells expressing IL-5, IL-13 and/or IL-21, but the ratio was much lower than that of Tph cells. Besides, PD-1−CXCR5−CD4+T-cells did not upregulate Tfh cell related genes, thus they were less likely to help B-cells with IgE production.

In this study, general Tph cells were not significantly elevated in the circulation of ABPA patients. However, compared to healthy controls, the Tph cells of ABPA patients had a Th2-skewed gene expression profile and comprised more cells expressing Th2 cytokines IL-5 and IL-13, indicating that Th2-skewed Tph subgroup expanded. RNA-sequencing identified type 2 inflammation relative genes in the Tph cells of ABPA patients, such as HPGDS [12, 28], GATA3-AS1 [29], IL9R [30, 31], FCER1G [32], IL17RB (IL-25 receptor) [28], FFAR2 (GPR43) [33] and TNFRSF4 (OX40) [34, 35]. A number of chemokines and inflammation genes that may assist type 2 inflammation, such as neutrophil chemokines CXCL2, CXCL3 and CXCL8 [36, 37], and factors promoting the differentiation and proliferation of Th2 cells, such as IL-6 and SOCS1 [38], were upregulated in the Tph cells of ABPA patients. IL-6 could inhibit Th1 differentiation through the upregulation of suppressor of cytokine signalling (SOCS)-1 expression to interfere with IFN-γ signalling [38]. The gene profiles of the Tph cells of ABPA suggested that it contributed to the strong type 2 inflammation.

In vitro co-culture of Tph cells and B-cells verified our surmise on the function of Th2-skewed Tph cells of ABPA patients. The Tph cells of ABPA patients induced both the differentiation of plasmablasts and the production of IgE, despite the frequency of IgE-producing cells was too low to be detected using the ELISpot method (data not shown). Although the Tfh cells of ABPA patients also induced the differentiation of plasmablasts, their ability of promoting IgE production could not compare to that of Tph cells. This is in accord with the flow cytometry results, which showed that the Tfh cells of ABPA patients contained fewer IL-5/IL-13-expressing cells compared to the Tph cells. In this study, the production of IgG was not as dependent on the constant help of Tph cells as the production of IgE, which indicated the crucial role of Tph cells in the pathogenesis of ABPA. After all, serum IgE level was relative to the exacerbation status of ABPA. The abundance of pathologic Tph cell subsets in the circulation of ABPA patients might explain why inhaled corticosteroids was unable to control the exacerbation of ABPA.

Additionally, we identified Tph cells in the BALF of ABPA patients with lung infiltrations and mucus impactions, which indicated that Tph cells might circulate and migrate to the lung and contribute to local inflammation. The Tph cells of ABPA patients did not express CCR2 as in rheumatoid arthritis [21], which indicate that for ABPA, other chemokines were important for the homing of Tph cells to the lung. A previous study showed that CD69 is one of the lung-resident markers [39]. CXCL12, the ligand of CXCR4, is abundant in the lung [40] and the CXCR4/CXCL12 axis was important in lung injury/fibrosis [41]. Thus, CD69, CXCR4 and other markers expressed by Tph cells could promote the chemotaxis of Tph cells to the lung in ABPA coordinately. The Tph cells of ABPA patients expressed chemokines for neutrophils such as CXCL2, CXCL3 and CXCL8. Studies show that they are involved in the allergic inflammation and could exaggerate type 2 response [41–43]. Besides, CXCL8 might increase the bronchoconstriction by stimulating airway smooth muscle cells in cystic fibrosis and asthma [42], both of which are major concomitant diseases of ABPA. Although whether the Tph cells in the BALF were Th2-skewed was unverified, we speculate that the gathering of activated Tph cells in the lung of ABPA patients might lead to uncontrolled local inflammation responses, IgE production and excessive secretion of sputum, manifesting as infiltrations and mucus impaction on HRCT. Cross-reactivity against Candida albicans was observed in the A. fumigatus-reactive T-cells of untreated acute ABPA patients [44]. Since the stimulation factor of ABPA exacerbation is still unclear, whether the Tph cells of ABPA patients also have the cross-reactivity with antigens other than A. fumigatus are also worthy of investigating in the future.

With the development of biologic drugs, ABPA patients have more choices other than corticosteroids and antifungal agents. The benefits of mAbs to IL-5, IL-4/IL-13 and IgE were reported in several ABPA cases [22]. Anti-IL-21 mAb was recently applied in a phase 2 trial of type 1 diabetes and showed benefits when combined with liraglutide [45]. Since Tph cells of ABPA patients produce cytokines such as IL-5/IL-13/IL-21 and promote the production of IgE, the blockade of Tph cells could cut the upstream pathological immune response. Considering the importance of IL-21 in B-cell activation and the production of antibodies [46], the overall blocking of IL-21 might lead to impaired humoral immune response and infection. The precise target at Th2-skewed subsets of Tph cells might decrease the risk. Thus, the Th2-skewed Tph cells of ABPA might be a promising target for treatment.

To better supervise or target at Tph Th2-skewed subgroup cells, it would be useful to identify its surface marker. OX40, one of the T-cell co-stimulatory molecules, facilitates Tfh cell development [47]. The deficiency of OX40 and its ligand OX40L impairs Th2 polarisation and relevant immune response in an asthma model [48]. In our study, both OX40+ and OX40− Tph cells expressed IL-5, IL-13 and/or IL-21, except OX40+ cells had more cells co-expressing IL-21 and Th2 cytokines. Thus, OX40 might not be a suitable marker here.

Our study has limitations. We only showed that Tph cells could be found in the BALF of three ABPA patients. Due to the microbiota in lungs, BALF is not sterile. Thus, it is hard to detect the intracellular cytokines and co-culture Tph cells from the BALF with B-cells without extra interference. Whether the Tph cells in the BALF are Th2-skewed needs to be studied in the future. Besides, more studies are needed to investigate the surface markers of Tph cell subsets expressing IL-5, IL-13 and/or IL-21 so as to detect and sort these cells more easily for studies. The original article on Tfh13 cells found a lot of Tfh13 cells in the lymph nodes of sensitised mice [49]. Although Tfh13 cells in the blood of patients were tested in that study, the gating marker PD-1 was so