Introduction

Bronchial asthma, affecting more than 260 million people worldwide, is a critical chronic airway disease with its pathogenesis still incompletely understood [1]. Accurate assessment of asthma severity, monitoring of disease progression and development of precise individualised treatment strategies are paramount in clinical practice. Biomarkers, including those for asthma, are reliable tools widely used for disease diagnosis, classification and prognostic assessment. CD4+ T lymphocyte-driven inflammation is the principal pathophysiological mechanism of asthma. T helper (Th)2 cells play a central role in allergic asthma, characterised by high levels of IgE antibodies, type 2 (T2) cytokines (including interleukin (IL)-4, -5 and -13), and eosinophilia. The quantity of eosinophils in peripheral blood, bronchoalveolar lavage fluid (BALF) and bronchial biopsy samples directly correlates with the severity of asthma [2].

Beside Th2 cells, other CD4+ cell types also play a significant role in asthma. Th1 cells secrete interferon-γ, serving as a protective factor and antagonist in the Th2 response by restraining T2 cytokine IL-4 [3–5]. Th17 cells secrete IL-17 to promote neutrophil recruitment and expansion, thereby initiating neutrophilic airway inflammation [3]. Moreover, IL-4+IL-17+ T-cells have been identified in the BALF of patients with severe asthma [6]. Regulatory T-cells (Tregs) maintain immune tolerance to allergens in the airways’ environmental interfaces [7]. Tregs maintain high levels of Foxp3 and secrete cytokines, including IL-10 and transforming growth factor (TGF)-β, to inhibit abnormal Th1, Th2 and Th17 airway responses [8, 9]. However, regarding the Treg ratio in asthmatics, studies have found inconsistent results due to varied identification methods [5, 10]. Even within the same sample batch, ratios of CD4+FOXP3+ Tregs in the BALF were higher in patients with moderate-to-severe asthma compared to healthy subjects. Similar findings applied to CD4+CD25+CD127− Tregs, but not CD4+CD25bright Tregs [11]. Furthermore, Harb et al. [12] showed that Notch4 signalling subverted lung tissue Treg cells into Th2 and Th17 cells in allergen- and pollutant-induced airway inflammation. Specific mutations in Tregs also lead to their conversion into Th17 cells and exacerbate airway inflammation [13]. Consequently, the ratios of helper T-cell subsets and concentrations of related cytokines in asthmatic inflammation are dynamic and intricate, warranting specific biomarkers for assessing asthma severity.

Clinical data have been accumulating for drugs targeting key biological molecules in asthma pathogenesis, such as dupilumab (IL-4Rα antibody), mepolizumab (IL-5 antibody), benralizumab (IL-5Rα antibody), tezepelumab (anti-thymic stromal lymphopoietin antibody) and omalizumab (IgE antibody) [14]. However, asthma exhibits biological heterogeneity [15]. Many biomarkers and therapeutic targets are primarily focused on T2-asthma, with other asthma types lacking well-defined markers. Consequently, a comprehensive spectrum of asthma biomarkers is lacking. The advancement of powerful omics technologies and the development of multi-omics data analysis methods are expected to improve asthma classification and biomarker discovery [16].

Glycerophospholipids, the main components of biological membranes, can be hydrolysed by phospholipases into lysophospholipids: lysophosphatidic acid (LPA), lysophosphatidylcholine (LPC), lysophosphatidylserine (LPS), lysophosphatidylglycerol (LPG), lysophosphatidylinositol (LPI) and lysophosphoethanolamine (LPE). Depending on carbon chain length and saturation, each class of lysophospholipids contains many species that have been shown to be important cell signalling mediators [17]. LPA promotes the differentiation of Th2 cells and acute allergic inflammation in a house dust mite-induced mouse model, which can be mitigated with an LPA receptor 2 (LPAR2) antagonist [18]. In addition, LPA induces IL-17 secretion by Th17 cells and inhibits IL-10 production by Tregs in a pulmonary arterial hypertension rat model [19]. LPS 18:0 promotes the development and function of mouse Tregs via GPR174, whereas LPS 18:1 suppresses IL-2 production and activation of murine CD4+ T-cells through the GPR174/Gαs pathway [20, 21]. LPI 18:0 inhibits accumulation of TCRγδ intraepithelial lymphocytes in the murine small intestine via its receptor GPR55 [22]. Although LPC, LPE and LPI levels are significantly elevated in the serum of corticosteroid-uncontrolled asthma patients [23], the specific effects of these lysophospholipids, particularly those with differing fatty acid chains, on CD4+ T-cell homeostasis in the asthmatic microenvironment remain unexplored.

In this study, we analysed the glycerophospholipid profile in serum from asthma patients and identified a universal elevation of the lysophospholipid LPG 18:0 as a biomarker for asthma. We further investigated the effects of LPG 18:0 on the differentiation and function of human CD4+ Tregs through in vitro assays. Our results establish a connection between lysophospholipids in the asthmatic microenvironment and CD4+ T-cell homeostasis, offering a novel strategy for asthma treatment by targeting the small-molecule phospholipid signalling pathway.

MethodsPatients

270 patients with asthma who attended Peking University Third Hospital (Beijing, China), along with 98 healthy controls, were included in this study for untargeted and targeted metabolomics analysis. All patients were diagnosed with asthma by physicians according to the Global Initiative for Asthma (GINA) 2023 guideline [24]. The diagnosis of asthma was based on symptoms of cough, shortness of breath, wheezing or chest tightness, in conjunction with variable airflow limitation. Exclusion criteria included patients with acute asthma exacerbation within 4 weeks, patients with coexisting respiratory diseases, heart-, renal- or liver-related diseases, rheumatoid immune disease, metabolic diseases, malignant diseases, pregnant women and patients with immunodeficiency or any other medical illnesses. Healthy subjects were those with no history or symptoms of chronic respiratory diseases, metabolic-related diseases, allergic diseases or other conditions that could affect the outcome. This study was approved by the ethics committee of Peking University Third Hospital (approval numbers 2014071 and M2020023). All subjects included in the study signed an informed consent form. Details about clinical data, classification of asthma subgroups and sputum induction cytology can be found in the supplementary material.

Lipidomics

Serum phospholipid profiles were determined using liquid chromatography-mass spectrometry-based lipidomics, as described previously [25]. In brief, for lipid extraction, 100 μL of serum (mixed with an internal standard mixture) was combined with 400 μL of 75% frozen methanol and vortexed for 2 min. Subsequently, 1 mL of methyl tert-butyl ether was added and the mixture was vortexed for 1 h at room temperature before centrifugation at 12 000×g for 10 min. The upper organic phase was collected and evaporated to dryness. The lipid analysis was conducted using an ACQUITY UPLC liquid chromatography system (Waters, Milford, MA, USA) and a UPLC BEH C18 column (1.7 μm, 100×2.1 mm ID; Waters). A 5500 QTRAP mass spectrometer (AB Sciex, Framingham, MA, USA) was used with Turbo Ion Spray electrospray ionisation as the ion source. The scanning mode was multiple reaction monitoring with ion source parameters set as follows: CUR=40 psi, GS1=30 psi, GS2=30 psi, IS=−4500 V, CAD=medium and temp=350°C.

Cell sorting and in vitro differentiation

EDTA anticoagulated blood from healthy volunteers was diluted 1:1 with PBS and centrifuged using a density gradient at 400×g for 20 min at room temperature using Ficoll-Paque PREMIUM 1.073 (GE Healthcare, Chicago, IL, USA; Cat#17-5446-52), with a slow rise and fall. Subsequently, the peripheral blood mononuclear cells (PBMCs) were collected and washed. Naïve CD4+ T-cells were enriched using a magnetic bead sorting reagent kit (Miltenyi Biotec, Berlin, Germany; Cat#130-094-131). The enriched cells were induced for differentiation based on anti-CD3 (1 μg·mL−1, pre-coated) and anti-CD28 (2 μg·mL−1) stimulation for 3–7 days. Cell differentiation conditions are detailed in the supplementary material.

Mouse asthma model

Female BALB/c mice aged 6–8 weeks were obtained from Peking University Health Science Center (Beijing, China). All mice were maintained in specific pathogen-free conditions at the Peking University Health Science Center animal feeding facility. The animal experiment procedure received approval from the ethics committee of Peking University Health Science Center (approval number LA2020391). The mouse model of asthma was established based on ovalbumin (OVA) or house dust mite (HDM) sensitisation and challenge. In the OVA model, mice were injected intraperitoneally with 200 μg OVA (Sigma-Aldrich, St. Louis, MO, USA; Cat#A5503) and 2.25 mg aluminium hydroxide (Sigma-Aldrich; Cat#239186) on day 1 and day 14, respectively. On day 28, mice were placed in the stimulation chamber and given 1% OVA saline solution for aerosol inhalation for 30 min for each of three consecutive days. The control group received only injections and atomisation of normal saline. BALF, serum and lungs were collected 24 h after the final nebulisation. For HDM-induced allergic inflammation, mice were anaesthetised with isoflurane and sensitised intranasally with 1 μg HDM (Greer Laboratories, Lenoir, NC, USA) in 50 µL saline or PBS (Gibco Life Technologies, NY, USA). 7 days later, mice were subjected to daily challenges for five consecutive days with 10 μg HDM in 50 µL saline intranasally. Mice were euthanised for analysis 4 days after the final challenge.

Detection of Treg suppressive function

Naïve CD4+ T-cells from human PBMCs or mouse spleen were isolated and induced under Treg-inducing conditions, treated with different concentrations of LPG 18:0 (0 nM, 10 nM and 50 nM) for 7 days. Foxp3 positivity was tested on day 7. CD4+CD25- T-cells were isolated from the peripheral blood of the same volunteer and stained with carboxyfluorescein succinimidyl ester. Depending on the Foxp3 positive rate, the induced Tregs were co-cultured with CD4+CD25- T-cells at ratios of 1:1, 1:2, 1:4 and 1:8. CD4+CD25− T-cell proliferation was assessed after 3 days.

Detection of mitochondrial damage and function

Mitochondrial reactive oxygen species (mROS) were assessed using MitoSOX Red Mitochondrial SUPE (Invitrogen, Waltham, MA, USA; Cat#M36008), while mitochondrial membrane potential was assessed using a tetramethylrhodamine ethyl ester probe (Thermo Fisher, Waltham, MA, USA; Cat#T669) according to the manufacturer's instructions. Live cells were incubated at 37°C for 30–60 min at the appropriate concentrations of the above probes for subsequent assays. ATP levels in cells were measured using an ATP assay kit (Biyuntian, Shanghai, China; Cat#S0026). NAD+ and NADH levels were also measured using a NAD+/NADH kit (Biyuntian; Cat#S0175).

Statistical analysis

Glycerophospholipid metabolic profiles were analysed using an online analysis tool (www.metaboanalyst.ca). Flow cytometry data were analysed using FlowJo 10. Statistical analysis and graphing were conducted using SPSS 26.0 and GraphPad Prism 8. Continuous variables are presented as mean±sd when normally distributed and presented as median (interquartile range) when abnormally distributed. The abnormally distributed data were analysed using the Mann–Whitney U-test or the Kruskal–Wallis test, while correlations were assessed using Spearman correlation analysis. Data conforming to normal distribution were analysed using the t-test or one-way ANOVA, while correlations were assessed using Pearson correlation analysis. The Pearson Chi-squared test was used for qualitative data comparison. p-values <0.05 was considered statistically significant.

ResultsLPG 18:0 level is elevated in the serum of asthmatic patients

To elucidate variations in serum phospholipid metabolites among asthmatic patients, we conducted a comparative analysis involving 270 asthmatic patients and 98 healthy controls. Demographic characteristics, including age and gender, were evenly distributed across the two groups. In the latest cohort, 45 asthma patients (male/female 21/24; mean±sd age 47.82±15.53 years) and 45 healthy volunteers (male/female 20/25; mean±sd age 47.40±16.22 years) were included for targeted lipidomic analysis. The results of cohort 1 are shown in supplementary figure S1a–c. The lipidomic profile showed clear separation between the asthmatic group and healthy controls, based on the sparse partial least squares discriminant analysis score plot (figure 1a). The heatmap and volcano plot indicated distinct phospholipid species between the asthma and control groups (figure 1b–d). The upregulated phospholipid species were primarily LPG 18:0, LPG 18:2 and LPG 20:0, while the downregulated phospholipids consisted mostly of LPE 22:0, LPE 20:0 and phosphatidylinositol (PI) 36:1 with long fatty acid chains. Additionally, barplots presented the differences in LPG species, showing significantly higher LPG 18:0 levels in asthma patients (figure 1e). When comparing the results with those from cohort 1, LPG 18:0 emerged as a stable and significantly changing factor (supplementary figure S1a–c). To confirm these findings, more subjects were enrolled for absolute quantification of serum LPG 18:0, involving 174 asthma patients and 43 healthy controls (table 1), substantiating higher LPG18:0 concentrations in asthma patients compared to healthy individuals (figure 1f). The area under the curve, based on the absolute concentration of LPG 18:0, was 0.812, indicating that an increased level of LPG 18:0 in asthma patients’ serum could serve as a potential biomarker for asthma (figure 1g).

FIGURE 1

a) An overall metabolic composition difference in serum between the asthma group and the healthy control group is illustrated in the sparse partial least squares-discriminant analysis (sPLS-DA) score plot. b) Heatmap of lipid species in asthmatics and healthy controls. c) The sPLS-DA loading plot displays the metabolites that mainly contributed to the differences. The vertical coordinates indicate the metabolites, while the horizontal coordinates represent the impact on the distinction between the two metabolic groups (asthma group (A), healthy control (C)). High concentrations are denoted by red squares while low concentrations are denoted by blue squares. d) A volcano plot demonstrating significant alterations in serum metabolites between the asthma group and healthy controls. e) The relative intensity of lysophosphatidylglycerol (LPG) phospholipids in the serum of asthmatics and healthy controls is depicted in a histogram. f) Liquid chromatography-mass spectrometry was employed to quantify the levels of LPG 18:0 in the serum of 174 asthma patients and 43 healthy controls. g) Receiver operating curve of LPG 18:0 based on f. PI: phosphatidylinositol; LPE: lysophosphoethanolamine; PC: phosphatidylcholine; FC: fold change; AUC: area under the curve. *: p<0.05, ***: p<0.001 (unpaired t-test or Mann–Whitney U-test in e; Mann–Whitney U-test in f).

TABLE 1

Clinical characteristics of asthma and healthy subjects

Sample sizeAsthmaHealthyp-valueSubjects17443Age years21742.50 (27.00)45.00 (25.00)0.744Female n (%)217100 (57.5)25 (58.1)0.937Age at first diagnosis years17436.54±16.45NABMI kg·m−221123.90±3.6124.30±3.290.532Pre-BD FEV1 % predicted21087.20 (21.20)96.00 (25.75)0.017*Pre-BD FEV1/FVC21076.06 (14.68)83.00 (7.87)<0.001*Blood eosinophils μL193250.00 (355.00)100.00 (97.50)<0.001*Sputum cell differential counts69 Neutrophils %63.10 (62.85)NA Eosinophils %5.40 (33.20)NA Macrophages %7.20 (23.20)NA Lymphocytes %0.50 (3.80)NATotal IgE kU·L−1165148.77 (322.09)NAFENO ppb12141.00 (64.50)NALPG 18:0 is elevated in different phenotypes of asthma and related to severity

Asthma is a heterogeneous disease that has been recognised as exhibiting different phenotypes by demographic, clinical and pathophysiological characteristics, including allergic asthma, nonallergic asthma, late-onset asthma, asthma with fixed airflow limitation and asthma with obesity, as well as different endotypes identified by their underlying biological immunological mechanisms [26]. To explore the relationship between serum LPG 18:0 and patients’ clinical phenotypes, an analysis was conducted on the absolute LPG 18:0 content in various subgroups of asthmatic patients.

LPG 18:0 levels are elevated in different phenotypes and endotypes of asthma compared to controls, yet no significant difference in serum LPG 18:0 levels was observed among underweight and normal weight, overweight and obese individuals with asthma (figure 2a), the smoker and nonsmoker asthma groups (figure 2b), the early-onset and late-onset asthma groups (figure 2c), patients with allergic or nonallergic states (figure 2d) and different lung function states (figure 2e and f). T2-high and non-T2-high asthma endotypes were distinguished by fractional exhaled nitric oxide (FENO) (figure 2g), blood eosinophils (figure 2h) and sputum eosinophil counts (figure 2i), with no significant difference observed in T2-high or non-T2-high asthma groups. However, all the mentioned asthma subgroups demonstrated elevated LPG 18:0 levels compared to healthy controls. Clinical treatment information was collected for asthma patients, and they were classified into different treatment steps according to GINA 2023. Serum LPG 18:0 levels were found to be significantly elevated in patients undergoing GINA treatment step 5 compared to those with GINA steps 1–4. This suggests that higher levels of LPG 18:0 in asthma patients may indicate more severe conditions (figure 2j). The proportion of patients in our cohort who treated with the GINA step 5 was also consistent with the reported proportion of severe asthma patients in the GINA report [24]. Furthermore, treated patients were categorised into controlled and uncontrolled groups based on ACT levels, and it was observed that asthmatics in the uncontrolled group exhibited higher LPG 18:0 levels (figure 2k), predicting poorer responses to treatment and worse control.

FIGURE 2

Analysis of lysophosphatidylglycerol (LPG) 18:0 concentration in the serum of different clinical characteristics. a) LPG 18:0 concentration in various body mass index (BMI) categories of asthma patients: normal or underweight (BMI <24.0 kg·m−2), overweight (BMI 24.0–27.9 kg·m−2) and obese (BMI ≥28.0 kg·m−2). b) LPG 18:0 concentration comparison between smokers and nonsmokers among asthma patients. c) LPG 18:0 concentration differentiation between late-onset and early-onset asthma. d) LPG 18:0 concentration comparison between allergic asthma and nonallergic asthma. e) LPG 18:0 concentration comparison in asthma patients with pre-bronchodilator (BD) forced expiratory volume in 1 s (FEV1) <80% predicted and pre-BD FEV1 ≥80% pred. f) LPG 18:0 concentration in asthma patients with pre-BD FEV1/FVC <70% and pre-BD FEV1/FVC ≥70%. g) LPG 18:0 concentration comparison in asthma patients with fractional exhaled nitric oxide (FENO) ≥25 ppb and FENO <25 ppb. h) LPG 18:0 concentration comparison in asthma patients with blood eosinophils (bEOS) ≥300 cells·μL−1 and <300 cells·μL−1. i) LPG 18:0 concentration in asthma patients with sputum eosinophils (sEOS) ≥3% and <3%. j) LPG 18:0 concentration comparison in asthma patients at Global Initiative for Asthma (GINA) treatment stages 1–4 and GINA treatment stage 5. k) LPG 18:0 concentration in asthma patients with controlled or uncontrolled symptoms post-treatment. Each data point in the graph represents an individual sample. *: p<0.05, **: p<0.01, ***: p<0.001, determined through the Kruskal–Wallis test or Mann–Whitney test.

In addition, correlation analysis was performed between LPG 18:0 levels and various laboratory indices, and serum Th cytokines were assessed in asthma patients. Results revealed that LPG 18:0 levels had no significant correlation with the number of blood eosinophils, percentages of sputum eosinophils or neutrophils, total serum IgE, FENO or lung function parameters (pre-bronchodilator (BD) forced expiratory volume in 1 s (FEV1) % predicted and pre-BD FEV1/forced vital capacity (FVC) ratio) (supplementary figure S1d). This finding supports the notion that LPG 18:0 elevation in asthma is consistent across different types of inflammation and lung function states. However, serum levels of IL-5, IL-10 and IL-17A demonstrated a very weak negative correlation with LPG 18:0 content (supplementary figure S1d). Although this correlation might be influenced by individual clinical variability and prior medical treatments, the results suggest that LPG 18:0 could play a universal role in CD4+ T-cell-related airway inflammation.

LPG 18:0 is elevated in asthmatic mice

To validate the significance of LPG 18:0 in asthma and elucidate its mechanism of action, mouse models were established using OVA and HDM. In these models, increased airway resistance (figure 3a) increased mucous secretion in the airway surrounding the asthmatic mice, along with increased infiltration of inflammatory immune cells (figure 3b and c), and elevated serum IgE levels (figure 3d) were detected. Furthermore, elevated concentrations of IL-4 and IL-5 were detected in mouse bronchoalveolar lavage fluid (figure 3e). These findings suggest that a typical mouse model of asthma has been successfully established. In addition, the serum concentration of LPG 18:0 in mouse was examined using liquid chromatography mass spectrometry. LPG 18:0 was found to be notably elevated compared to the control group in both OVA and HDM mouse models, mirroring findings in clinical human samples (figure 3f).

FIGURE 3

Lysophosphatidylglycerol (LPG) 18:0 is elevated in asthmatic mice. a) Airway hyperresponsiveness in the respective mouse groups in response to acetylcholine chloride. Rrs represents the resistance of the respiratory system, which is a measure of airway resistance. b) Haematoxylin and eosin (HE) and periodic acid–Schiff (PAS) staining of mouse lung samples from the house dust mite (HDM)-induced asthma model. Scale bar=100 μm. c) Inflammation scores in the corresponding lung tissues (n=10). d) The concentration of IgE in mouse serum was determined using ELISA. e) Mouse bronchoalveolar lavage fluid (BALF) was collected after asthma induction, and levels of interleukin (IL)-5 and IL-4 were quantified using ELISA. f) Concentrations of LPG 18:0 in the serum of the ovalbumin (OVA) and HDM mouse models were quantified using liquid chromatography mass spectrometry. Data represent results from two or three independent experiments. *: p<0.05, **: p<0.01, ***: p<0.001.

High concentrations of LPG 18:0 inhibit the differentiation and suppressive function of Tregs in vitro

Subsets of CD4+ helper T-cells, including Th1, Th2, Th17 and Treg cells, are involved in the pathogenesis of asthma [3]. To explore the effects of LPG 18:0 on CD4+ T-cell subsets, naïve CD4+ T-cells were sorted from healthy human peripheral blood using magnetic beads, and their proliferation, apoptosis and differentiation abilities were examined in vitro. Results revealed that LPG 18:0 specifically inhibited the differentiation of naïve CD4+ T-cells into Tregs by reducing the expression of the master transcription factor FOXP3. Interestingly, LPG 18:0 does not impact the differentiation of Th1, Th2 and Th17 cells (figure 4a and b). At concentrations of 10–50 nM, LPG 18:0 had no significant effect on the proliferation and apoptosis (figure 4c and d) levels of Treg and Th17 cells.

FIGURE 4

The inhibitory effects of high concentrations of lysophosphatidylglycerol (LPG) 18:0 on the differentiation and suppressive function of regulatory T-cells (Tregs). a, b) Human naïve CD4+ T-cells were cultured for 5–7 days in a differentiation medium containing various concentrations of LPG (0, 10 and 50 nM). Differentiation into type 1, 2 and 17 T-helper cells (Th 1, Th2 and Th17, respectively) and Treg cells was assessed via flow cytometry, employing T-bet, GATA3, RORγT and Foxp3 markers. The experiment was replicated independently at least three times. c, d) The impact of LPG on the proliferative capacity of various Th cell subsets was quantified using carboxyfluorescein succinimidyl ester (CFSE) labelling and flow cytometry analysis. Apoptosis among Th cells was quantified utilising annexin-V and 7-AAD staining. Three separate runs of this experiment were conducted. The experiment was independently replicated three times. e–g) In vitro induction of human Tregs was performed with varying concentrations of LPG 18:0. Levels of transforming growth factor (TGF)-β and interleukin (IL)-10 in Tregs were measured on day 5 using flow cytometry (e, f) and ELISA for supernatants (g). h, i) The induced Tregs from healthy donors and CD4+ CD25− effector T-cells (Teff cells) were co-cultured in vitro at ratios of 1:8, 1:4, 1:2 and 1:1 with LPG, and the proliferation levels of Teff cells was measured after 3 days. These experiments were repeated independently at least three times. j–l) Lung Treg cell frequencies and flow cytometric analysis in the corresponding groups (n=10). Treg cells gated in CD45+ CD3+ CD4+ FOXP3+. An additional intraperitoneal injection of 50 nM LPG 18:0 was administered every other day to mice during the induction of asthma (n=5 per group). ns: nonsignificant; IFN: interferon. The statistical results employed the unpaired t-test and the results are displayed as mean±sd. *: p<0.05, **: p<0.01, ***: p<0.001.

We further examined the effects of LPG 18:0 on Treg functions. The results showed that LPG 18:0 significantly inhibited TGF-β and IL-10 secretion in Tregs (figure 4e and f). Meanwhile, ELISA detected a reduction in IL-10 levels in cell culture supernatants after LPG 18:0 administration (figure 4g). Consistent with these results, the inhibitory capacity of Tregs on autologous CD4+CD25− T-cells was significantly decreased with LPG 18:0 treatment (figure 4h and i). Therefore, our results showed that LPG 18:0 specifically inhibited the differentiation and suppressive function of human Tregs, without directly affecting the differentiation of other Th cells. Based on mouse models, additional LPG 18:0 was administered through intraperitoneal injections to asthmatic mice. The results indicated that this supplementation resulted in a decreased number of Tregs in the lung tissue compared to the standard asthma model group, demonstrating that LPG 18:0 affects lung Tregs (figure 4j–l). These discoveries suggest a potential connection between LPG18:0 and the function of Treg cells in asthma.

LPG 18:0 induces mitochondrial dysfunction of human Tregs

Unlike Th1, Th2 and Th17 cells, which mainly rely on glycolysis and oxidative phosphorylation for energy supply, human Tregs rely on additional fatty acid oxidation for their differentiation and function [27, 28]. The mitochondrion is a critical organelle for fatty acid oxidation that shapes the cell fate and function of Tregs [29]. Inhibition or deletion of electron transport chain components remarkably reduces Foxp3 expression and impairs Treg suppressive function [30–32]. Therefore, we sought to explore whether LPG 18:0 altered the mitochondrial function of Tregs.

LPG 18:0 possesses a glycerol backbone, with a single fatty acid chain C18:0 at the sn-1 position, a hydroxyl group at the sn-2, and a phosphate head group at the sn-3 position. LPG can incorporate appropriate acyl content through LPG acyltransferase 1 or cardiolipin synthase 1 to be remodelled into phosphatidylglycerol, a precursor for the synthesis of cardiolipin in the mitochondrial membrane [33]. Therefore, we hypothesise that excessive production of LPG may impact the mitochondrial function of Treg cells. Indeed, LPG 18:0 reduced the total ATP production in human Tregs (figure 5a), accompanied by a reduction of oxidative phosphorylation as measured by mitochondrial oxygen consumption rate at both basal and ATP-linked levels using Seahorse technology (figure 5b and c). Moreover, LPG 18:0 treatment increased mROS levels and decreased mitochondrial membrane potential (figure 5d and e). These results indicate that LPG 18:0 induces mitochondrial dysfunction in human Tregs.

FIGURE 5

Lysophosphatidylglycerol (LPG) 18:0 induces mitochondrial dysfunction in human regulatory T-cells (Tregs). a) The intracellular ATP level was measured on day 7 following in vitro Treg induction with a 50 nM LPG 18:0 treatment. This experiment was replicated independently three times, employing a t-test for statistical analysis. b, c) ATP-linked and basal mitochondrial respiration in in vitro-induced Tregs, both with and without LPG 18:0 treatment, were assessed using Seahorse oxygen consumption rate (OCR) analysis. d, e) Mitochondrial reactive oxygen species (mROS) within Tregs were quantified using mitoSOX (d), and their mitochondrial membrane potential was measured using tetramethylrhodamine, ethyl ester (TMRE) (e) through flow cytometry. This experiment was independently replicated three times (n=9). Statistical analyses were performed using a two-tailed, unpaired t-test. *: p<0.05, **: p<0.01, ***: p<0.001.

LPG 18:0 reduces the protein level of FOXP3 by regulating intracellular NAD+/NADH and SIRT1 level

FOXP3, the master transcription factor for Tregs differentiation and suppressive function, is regulated at both transcriptional and post-translational levels [34]. Our study revealed that LPG 18:0 treatment led to reduced FOXP3 protein levels in Tregs, as determined by flow cytometric analysis (figure 4a and b), while its mRNA levels remained unchanged, as evidenced by real-time PCR (figure 6a). Acetylation of FOXP3 is critical for both protein stability and transcriptional activity [34]. Deacetylation of FOXP3 not only reduces its protein expression, but also impairs the suppressive function of Tregs [35]. It has been reported that in human Tregs, FOXP3 can be reciprocally regulated by the histone acetyltransferase p300 and the histone deacetylase sirtuin-1 (SIRT1) [36]. SIRT1, an NAD+-dependent histone deacetylase, has its activity induced by an elevation of NAD+ levels [37].

FIGURE 6

Lysophosphatidylglycerol (LPG) reduces FOXP3 protein acetylation via the NAD/sirtuin-1 (SIRT1) pathway in induced regulatory T-cells (iTregs) a) Foxp3 mRNA levels in human iTregs were measured using quantitative (q)PCR after 7 days of treatment with or without 50 nM LPG. The experiment was independently repeated three times. b) NAD+/NADH levels were determined using the NAD+/NADH Assay Kit (n=3). This experiment was independently repeated three times. c) SIRT1 mRNA levels were assessed via qPCR (n=3). The experiment was independently repeated three times. d) Intracellular SIRT1 protein levels were assessed using Western blot analysis. e) After 7 days of stimulation with 50 nM LPG, the acetylation level of FOXP3 was determined. Cells were pre-treated with EX-527 (1 μM) for 72 h or TSA (0.3 μM) for 24 h prior to harvesting. This experiment was independently replicated three times. f, g) Foxp3 levels in iTregs treated with 50 nM LPG and SIRT1 or histone deacetylase (HDAC) inhibitors. This experiment was independently repeated three times. h, i) Mouse CD4+-naïve T-cells were differentiated into Tregs with or without 50 nM LPG and SIRT1 or HDAC inhibitors treatment for 4 days. The resulting CD4+ CD25+ Tregs were then co-cultured with carboxyfluorescein succinimidyl ester (CFSE)-labelled CD4+ CD25− effector T-cells (Teff cells) for an additional 3 days. Teff cell proliferation was evaluated by flow cytometry. Three independent repetitions of this experiment were performed. ns: nonsignificant. Statistical analysis was conducted using one-way ANOVA, and results are presented as mean±sem. *: p<0.05, **: p<0.01, ***: p<0.001.

In mitochondria, NAD+/NADH couples the tricarboxylic acid cycle and electron transfer chain for ATP production. We assessed the NAD+ and NADH level in human Tregs and discovered that LPG 18:0 treatment elevated the NAD+/NADH ratio (figure 6b), suggesting possible alterations in SIRT1 expression and activity. Indeed, upon culturing human Tregs in vitro with LPG 18:0 administration, we observed an increase in SIRT1 expression at both the mRNA (figure 6c) and protein levels (figure 6d). Upon inhibiting SIRT1 activity with EX-527 and simultaneously treating with LPG 18:0, both the acetylation and protein levels of FOXP3 were restored (figure 6e–g). Additionally, alterations in SIRT1 may affect the function of DNA methyltransferases and the acetylation of histones [38]. To ascertain whether LPG 18:0 affects the methylation of the Foxp3 gene and the acetylation of histones via the SIRT1 pathway, we assessed the influence of LPG 18:0 on the acetylation levels of histones at the Foxp3 locus and on DNA methylation at the CNS-2 locus of the Foxp3 gene. The results indicated that administration of LPG 18:0 did not affect these processes (supplementary figure S2d and e).

In an in vitro experiment, naïve CD4+ T-cells from mice were induced into Tregs and cocultured with CD4+ CD25− effector T-cells at various ratios. Following LPG 18:0 treatment, the inhibitory capacity of Tregs diminished, but the concomitant addition of EX-527 and LPG 18:0 restored the inhibitory capacity of Tregs to normal levels (figure 6h and i). Furthermore, the simultaneous addition of TSA, a histone deacetylase inhibitor, with LPG 18:0, preserved both FOXP3 protein levels and the suppressive function of human Tregs (figure 6e–i). These results demonstrated that LPG 18:0 reduces the protein level of Foxp3 by regulating intracellular NAD+/NADH and SIRT1 level.

Discussion

Diagnosing asthma and predicting its trajectory via biomarkers holds substantial clinical significance. In recent years, comprehensive research has focused on identifying efficacious biomarkers for asthma, ranging from molecular and cellular levels to tissue involvement and imaging characteristics [39]. Among these biomarkers, cytokines such as IL-4, IL-5 and IL-13 play pivotal roles in the inflammatory cascade characteristic of asthma [14]. However, the heterogeneity of T-cells and the dynamic interactions among Th cells and lung tissue cells induce variations in these cytokines, thereby compromising the biomarkers’ accuracy [2]. There exists an urgent need for biomarkers that can precisely predict asthma's severity and control levels (treatment response). Recent lipidomics analyses have revealed elevated levels of LPC (16:0, 16:1, 18:0, 18:1, 20:3, 20:4, 22:4) and LPE (18:0, 20:3, 20:4, 22:5) in the serum of patients with uncontrolled asthma as compared to those with corticosteroid-managed asthma [23]. In our prior study, we demonstrated that LPG 22:6 and LPG 20:5 levels were significantly elevated in patients with eosinophilic asthma [25]. Elevated levels of LPA 22:5 and LPA 22:6 were identified in the BALF of patients with allergic asthma [40]. However, although these results indicated lysophospholipid elevation in specific asthma subgroups or statuses, few lysophospholipids were consistently altered across asthma presentations.

In this study, we analysed the absolute concentration of LPG 18:0 across various asthma subgroups and observed no significant differences among subgroups delineated by clinical features. This finding suggests that LPG 18:0 may serve as a universal biomarker for asthma pathogenesis. Based on the relationship between LPG 18:0 and GINA steps or asthma control status, we highlight the potential role of LPG 18:0 as a marker covering a spectrum of asthma manifestations and as a predictor of increased treatment needs and poorer treatment responses.

An elevation in total LPG levels has been reported in tumour tissues from patients with nonsmall cell lung cancer [41]. Our prior research revealed a significant reduction in LPC and LPE levels during the acute exacerbation stage of COPD [42]. An excessive accumulation of LPC, LPE, LPA was observed in human lung fibroblasts upon exposure to fine particulate matter [43]. However, no alteration in LPG 18:0 levels was observed in this setting. The role of LPG 18:0 as a specific mediator in asthma, as opposed to other lung diseases, warrants further investigation. In this study, we utilised human blood samples as the primary research subject to more accurately simulate the impact of lipids on human immune cells. Our in vitro analyses demonstrated that exogenous LPG 18:0 inhibited Treg differentiation and suppressive function. This effect appeared to be specific to Tregs, as no significant changes were observed in the expression of specific transcription factors or the secretion of cytokines in Th1, Th2 and Th17 cells. The involvement of LPG 18:0 in other Treg-related diseases necessitates further investigation.

LPG 18:0 specifically induced mitochondrial dysfunction in Tregs by elevating mROS levels and reducing mitochondrial membrane potential. Previous reports indicate that impaired lysosomal degradation function exacerbates mROS, affecting mitophagy in Tregs during autoimmune responses [29]. Enhanced mitochondrial-lysosomal colocalisation was observed in induced Tregs from asthma patients compared to those from healthy subjects (supplementary figure S3a). Nevertheless, neither the mROS scavenger (MitoTEMPO), the mammalian target of rapamycin (mTOR) agonist (MHY1485) nor the autophagy inhibitor (BafA1) ameliorated the reduction in FOXP3 and TGF-β levels in human Tregs treated with LPG 18:0 (supplementary figure S3b and c). These findings imply that mROS, mTOR and autophagy pathways might not be the pivotal factors in the LPG 18:0-mediated inhibition of Treg differentiation and function.

Reduced ATP production and an elevated NAD+/NADH ratio in human Tregs treated with LPG 18:0 suggest a disruption of mitochondrial energy metabolism by LPG 18:0. Despite the reduction in FOXP3 protein expression with LPG 18:0 treatment, FOXP3 gene transcription remained unchanged. These observations may be attributed to post-translational modifications, including acetylation, phosphorylation, ubiquitination and methylation [44, 45]. SIRT1, a pivotal NAD+-dependent histone/protein deacetylase, mediates the deacetylation and subsequent proteasomal degradation of FOXP3 [36, 46]. Within our in vitro system, an increase in SIRT1 expression and activity accounted for the decreased FOXP3 protein levels in human Tregs. This aligns with prior findings that phospholipids enriched in docosahexaenoic acid, eicosapentaenoic acid, or LPA can modulate deacetylase activity and SIRT1 expression [47, 48].

However, both the specific SIRT1 inhibitor EX-527 and the synthetic SIRT1 activator SRT1720 mitigate T2 asthmatic inflammation [49]. This phenomenon may be attributed to SIRT1's differential effects on various cell types implicated in the pathogenesis of asthma. SIRT1 exerts a dual role by reducing GATA3 acetylation in Th2 cells and inhibiting Akt/NF-κB signalling in human lung epithelial cells, conferring protective effects, while conversely activating hypoxia-inducible factor 1α to elevate vascular endothelial growth factor expression in lung tissue, thus contributing to deleterious effects [50]. Additionally, this study reveals the harmful impact of SIRT1-mediated deacetylation of FOXP3 in human Tregs under conditions of excess LPG 18:0. Consequently, this study proposes targeting entities upstream of SIRT1 for therapeutic intervention. Enzymes involved in LPG production and their specific receptors emerge as potential therapeutic targets.

The majority of lysophospholipids’ biological effects are mediated through G-protein coupled receptors. However, the precise receptor for LPG has not yet been identified. It is hypothesised that the LPA receptor (LPAR) and GPR55 may act as receptors for LPG [51, 52]. Employing publicly accessible single-cell RNA sequencing data, we assessed the expression of GPR55 and LPAR on Treg cells and others. The findings indicated GPR55 expression on human CD4+ and CD8+ T-cells, irrespective of asthma presence. Further scrutiny revealed that GPR55's expression was primarily located on the surfaces of Tregs, whereas LPAR2 expression encompassed all helper T-cell types (supplementary figure S4) [53]. Our data suggest that the administration of inhibitors targeting LPA receptors (Ki16425) or GPR55 (CID16020046) partially counteracted the suppression by LPG 18:0 on human Treg differentiation and TGF-β expression, underlining the intertwined relationship between LPG, LPAR and GPR55 (supplementary figure S5). Nonetheless, further exploration is required to elucidate the specific receptors for LPG and their cellular expression profiles. Tregs have been demonstrated to confer protective effects in asthma, mitigating inflammation-related tissue damage and forestalling the progression of airway remodelling; thus, they represent significant targets for asthma immunotherapy [54]. The specific influence of LPG 18:0 on Tregs offers a novel approach for the treatment of immune-related diseases, mediated by disruption of metabolite signalling or targeting its specific receptor.

In summary, LPG 18:0 emerges as a promising novel biomarker, with potential applications in asthma diagnosis and assessing treatment response. Quantifying the concentration of LPG 18:0 provides deeper insights into the intricate pathophysiological mechanisms of asthma, enabling tailored therapeutic strategies for individual patients. This exploratory study focuses on the potential of LPG 18:0 as a biomarker in asthma and its impact on CD4+ T-cell dysregulation. LPG 18:0 disrupts mitochondrial function and affects FOXP3 acetylation through the NAD+/SIRT1 pathway in Tregs. However, further research endeavours are imperative to overcome the challenges and constraints encountered during the implementation of LPG18:0, ultimately advancing the refinement and personalisation of asthma management.

Supplementary material

Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.

Supplementary material ERJ-01752-2023.Supplement

Acknowledgements

We express our gratitude to Pingzhang Wang (Dept of Immunology, School of Basic Medical Sciences, Peking University, Beijing, China) for his contributions to the data discussion, and to Chen Huang and Xiaotong Yu (Center of Basic Medical Research, Institute of Medical Innovation and Research, Peking University Third Hospital, Beijing) for their technical assistance. We also acknowledge the valuable editing assistance of Jiawei Ribaudo, a reviewer from the pre-publication support service at the University of Michigan (Ann Arbor, MI, USA).

Data availability

The data supporting the findings of this study are available from the corresponding author upon reasonable request.

References↵Safiri S, Carson-Chahhoud K, Karamzad N, et al. Prevalence, deaths, and disability-adjusted life-years due to asthma and its attributable risk factors in 204 countries and territories, 1990–2019. Chest 2022; 161: 318–329. doi:10.1016/j.chest.2021.09.042↵Maison N, Omony