Activin A–Activated ALK4 Induces Pathogenic Th17‐Involved Endothelial–Mesenchymal Transition in Systemic Lupus Erythematosus–Associated Pulmonary Arterial Hypertension

Abstract

Objective

Autoimmune diseases, such as systemic lupus erythematosus (SLE), are associated with pulmonary arterial hypertension (PAH), a condition that can lead to heart failure. However, whether T cells also contribute to the occurrence of PAH in SLE has not been clarified. The objective of this study was to elucidate the role of Activin A and activated receptor signaling in SLE‐PAH.

Methods

Mass Cytometry (CyTOF) analysis was performed to identify the major affected immune cell population in patients with SLE‐PAH. Serum Activin A and interleukin‐17 (IL‐17) levels in patients with SLE‐PAH, patients with SLE, and healthy donors were determined by enzyme‐linked immunosorbent assay. Cocultures of Th17 cells with pulmonary microvascular endothelial cells (PMECs) and relevant rodent models were used to identify the converged target.

Results

The reduced CD4⁺ T cell number was detected in patients with SLE‐PAH after treatment with immunosuppressant and vasodilator. Increased Th17 cells population and higher serum Activin A and IL‐17 levels were found in patients with SLE‐PAH compared to patients with SLE only or donors. Activin A signals via activin receptor–like kinase 4 (ALK4) in both Th17 cells and PMECs. Overexpressing ALK4 in Th17 cells increased IL‐6 and endothelial–mesenchymal transition (EndoMT) marker levels in cocultured PMECs. We found severe SLE–pulmonary hypertension (PH) in mice by overexpressing ALK4, and alleviated hemodynamic changes in CD4⁺ T cells depletion mice. ALK4 inhibitor vactosertib (TEW‐7197) effectively treated SLE‐PH mice by repressing connective tissue growth factor (CTGF) transcription, which was induced by ALK4 activated pSmad2 and pSTAT3.

Conclusion

Our findings suggest that Activin A activates ALK4 in Th17 cells, thereby inducing IL‐17 secretion. Concurrently, activated ALK4 induces EndoMT in human PMECs (hPMECs) via CTGF up‐regulation. It suggests that ALK4 is a promising therapeutic target for SLE‐PAH.

INTRODUCTION

Autoimmune diseases, including connective tissue disease (CTD), are strongly associated with the development of irreversible end‐stage pulmonary arterial remodeling, leading to pulmonary arterial hypertension (PAH) (defined as a mean pulmonary artery pressure [mPAP] > 20 mm Hg) at rest. ¹ Although CTD‐PAH in White populations primarily relates to systemic sclerosis, systemic lupus erythematosus (SLE)–associated PAH is a more common form of CTD‐PAH in Asian individuals. ¹ Current treatments have yet to achieve complete cure for SLE‐PAH, and the patients continue to face a challenged prognosis. ² , ³ Therefore, identifying molecular targets closely related to the progression of SLE‐PAH is important for the development of interventions.

In patients with SLE, the presence of PAH coincides with significantly higher serum protein levels of interleukin‐17 (IL‐17). ⁴ , ⁵ IL‐17 is primarily secreted by Th17 cells, which are a subtype of CD4⁺ T cells associated with many inflammatory and autoimmune diseases. Transforming growth factor β (TGFβ) and IL‐6 induce CD4⁺ T cells differentiate to Th17 cells and secrete ⁶ , ⁷ , ⁸ IL‐17. Pathogenic Th17 cells were also found to exacerbate local inflammation in an animal model of autoimmune encephalomyelitis. ⁷ However, further investigation is necessary to determine whether the induction of pathogenic Th17 cell differentiation is involved in SLE‐PAH.

Mutations at members of the TGFβ family of receptors, including bone morphogenetic protein type II receptor (BMPR2) and activin receptor–like kinase 1 (ALK1), have been implicated in the development of idiopathic and heritable PAH. ⁹ , ¹⁰ , ¹¹ BMPR2 plays a vital role in maintaining pulmonary vascular endothelial integrity. ¹² , ¹³ In contrast, Activin A, another TGFβ family member, and its receptor ALK4/5 coordinately mediate the abnormal proliferation of pulmonary vascular endothelial cells, which is related to elevated pulmonary artery pressure. ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ Patients with PAH have increased Activin A levels in peripheral blood and elevated levels of proinflammatory cytokines ¹⁸ , ¹⁹ , ²⁰ such as IL‐6. Furthermore, on exposure of CD4⁺ T cells to cytokines, such as IL‐6, IL‐1β, and IL‐23, Activin A was up‐regulated in CD4⁺ T cells, and activated ALK4 in an autocrine manner to induce pathogenic Th17 cells. ⁷ Therefore, we hypothesized that ALK4 activation can induce pathogenic Th17 to produce IL‐17 in SLE‐PAH.

Current treatments for SLE‐PAH include endothelin receptor antagonists, phosphodiesterase type 5 inhibitors, and prostacyclin analogs, which are often administered in combination. In 2023, a fourth potential therapy for PAH, ¹ sotatercept, which is a recombinant human activin receptor type II A (ACTRIIA) fusion protein that acts as a ligand trap for Activin A, was introduced to rebalance growth‐inhibiting BMPR2 signaling. It was reported to significantly decrease pulmonary vascular resistance in patients with PAH and improve their performance on a six‐minute walk test. ²¹ , ²² However, whether this approach is effective for a broader population of patients with CTD, especially those with SLE‐PAH, which has a high regional incidence rate, is still unclear. It suggests specific therapeutic strategies for CTD‐PAH must be considered.

In this study, we used patient‐derived blood samples, cell cocultures, and a rodent model of SLE–pulmonary hypertension (PH) to delineate the mechanism of Activin A–activated signaling in the inflamed pulmonary vasculature and to investigate the role of specific receptors in the pathogenic effects of Th17 cells involved in SLE‐PAH. Our study revealed a significant reduction in the CD4⁺ T cell population in the SLE‐PAH (after treatment compared with before treatment) and confirmed the involvement of CD4⁺ T cells in PH mice. Increased IL‐17 and Activin A levels in the peripheral blood samples of patients with SLE‐PAH promoted the activation of target genes such as connective tissue growth factor (CTGF) in human pulmonary microvascular endothelial cells (PMECs). More severe PH was developed by overexpressing ALK4 in pristane‐induced mice exposed to hypoxia. We treated the mice with a novel ALK4 inhibitor, Vactosertib ( small molecule TEW‐7197, shorten as TEW), and found a significant improvement in hemodynamic indices and vascular remodeling.

MATERIALS AND METHODS

Patient enrollment

Before the study began, all individuals provided written informed permission by the principles established in the Declaration of Helsinki approved by the Ethics Committee of Zhejiang University Medical College (no. 014‐2015). Patients with SLE‐PAH (total n = 45) and patients with SLE without PAH (total n = 46, age‐ and sex‐matched) and patients with primary Sjögren disease (pSjD)–associated PAH (total n = 24) and patients with PAH without pSjD (total n = 27) were enrolled at Peking Union Medical College Hospital between November 2016 and May 2021. PAH was diagnosed based on the 2015 European Society of Cardiology and the European Respiratory Society guidelines by right‐sided heart catheterization, with an mPAP of 25 mm Hg at rest and pulmonary vascular resistance (PVR) of >2 Wood units. ²³ All patients with SLE were diagnosed using the American College of Rheumatology 1997 criteria. ²⁴ All patients with pSjD were diagnosed using the 2016 American College of Rheumatology/EULAR classification criteria for pSjD. ²⁵ , ²⁶ Patients with additional overlapping CTDs, interstitial lung disease, pulmonary thromboembolism, left heart insufficiency, infection, or acute organ dysfunction were excluded.

Animal study design

All animal study protocols were in accordance with the Animals in Research: Reporting In Vivo Experiments 2.0 Guidelines. ²⁷ All animal study protocols were reviewed and approved by Animal Care & Use Committee of Zhejiang University (numbers ZJU20210036 and ZJU20240927, AIRB‐2023‐0317). Animals were randomly assigned to experimental groups. Following the principle of equal opportunity, the random number table method was used to assign each animal. No samples or animals were excluded from the analysis. All quantitative analyses and measurements (gene expression analysis, enzyme‐linked immunosorbent assay [ELISA], right ventricular systolic pressure [RVSP] measurement, Western blot analysis) were performed in a masked manner. All figures are representative of at least three experiments unless otherwise noted. At the time of killing, mice were anesthetized with pentobarbital sodium (1% in 0.9% saline, 0.1 mL/10 g body weight) injected intraperitoneally once. Rats were anesthetized with pentobarbital sodium (2% in 0.9% saline, 0.2 mL/100 g body weight) injected intraperitoneally once. The anesthesia depth was monitored with pedal reflex. At the end of the experiment, all animals were killed by cervical dislocation.

SLE‐PH mouse model establishment

Female wild type (WT) C57BL/6J mice (18–20 g, Charles River, n = 8 mice per group) were randomly grouped according to the experiments. The mice in the SLE and SLE‐PH groups were intraperitoneally injected with 0.5 mL of pristane (MedChemExpress), whereas saline was also intraperitoneally administered to the mice in the normoxia and hypoxia groups. Then, the mice in the hypoxia and SLE‐PH groups were placed in a 10% O₂ chamber and maintained under hypoxic conditions for four weeks to establish the SLE‐PH model.

All animals were housed in a specific pathogen‐free facility (three to four mice per cage) in a room with a temperature of 25°C (SD ± 2°C), were maintained on a 12‐hour light/dark cycle, and were allowed free access to a standard rodent chow diet and water. At the end point of model establishment, the Vevo 2100 system (VisualSonics) was used for transthoracic echocardiography.

Statistical analysis

All the statistical analyses were performed using GraphPad Prism 8 (GraphPad Software). All the data were analyzed without the exclusion of any data points. All experimental data were subjected to a normality (Shapiro‐Wilk) test. Unless otherwise noted, statistical comparisons of samples were conducted using Student's t‐test or one‐way repeated measures analysis of variance with the Dunn–Bonferroni correction for multiple comparisons followed by Dunnett's post hoc test. For nonnormally distributed data, statistical comparisons of samples were conducted using the Mann‐Whitney U‐test. Correlations were evaluated by Spearman's correlation analysis, with the strength of relationships quantified by the correlation coefficient (r_s). P values < 0.05 (two‐tailed) were considered to indicate statistical significance. In the corresponding figures, “*,” “**,” and “***” indicate P < 0.05, P < 0.01, and P < 0.001, respectively. Each figure legend mentions the test used, the criteria for statistical significance, and the number of experimental and biologic replicates.

RESULTS

Th17 cells and IL‐17 levels elevated in patients with SLE‐PAH, with increased Activin A to drive Th17 cell differentiation

To identify the major affected immune cell type in patients with SLE‐PAH, we applied mass cytometry to determine the proportions of immune cells in the peripheral blood samples of patients with SLE‐PAH before and after six months of treatment with vasodilator and immunosuppressant (Figure 1A and Figure S1A); details are listed in Supplemental Table 1. Compared to that in the pretherapy group, the proportion of CD4⁺ T cells was significantly reduced in posttherapy patients, but not for CD8⁺ T cells and other subsets, indicating that CD4⁺ T cells might be the major immune cells involved in the pathogenesis of SLE‐PAH (Figure 1B and Figure S1B and C). To confirm the involvement of Th17 cells in SLE‐PAH, we analyzed Th17 cells proportion in peripheral blood mononuclear cells (PBMCs) from five healthy controls and six patients with SLE‐PAH by flow cytometry (Figure 1C). Quantitative analysis revealed a significant elevation in Th17 cell frequency in patients with SLE‐PAH compared to controls. In addition, we detected elevated IL‐17 levels in SLE‐PAH serum samples compared with control and SLE without PAH serum samples (Figure 1D).

Figure 1.

Th17 cells and IL‐17 were elevated in patients with SLE‐PAH, with increased Activin A to drive Th17 cells differentiation. (A) The percentages of CD4⁺ T cells and CD8⁺ T cells in the peripheral blood of the patients before and after therapy were analyzed by mass cytometry. (B) The quantification of analyzed cells from the four patients. (C) Representative flow cytometry dot plots of PBMCs from five HCs and six patients with SLE‐PAH, and the quantification of Th17 cells is shown on the right. (D) Enzyme‐linked immunosorbent assays were conducted with serum samples from HCs, patients with SLE without PAH (n = 46), and patients with SLE‐PAH (n = 45). (E) Correlation analysis between Activin A level and clinical parameters in patients with SLE‐PAH. (F) The human PBMC‐derived CD4⁺ T cells stimulated by Activin A or Activin A with IL‐6 were collected and examined by Western blot for IL‐17 and RORγt (left) and densitometric quantification was shown in right (n = 3). (G) The mRNA expression of RORC and IL‐17A in human CD4⁺ T cells activated by Activin A or Activin A with IL‐6 (n = 3). Data are shown as means ± SEMs, and P values (Bonferroni corrected) are depicted in the panels throughout all figures as *P < 0.05, **P < 0.01 and ***P < 0.001 using one‐way analysis of variance with repeated measures, followed by Bonferroni correction for multiple comparisons. A, Activin A; CI, cardiac index; HC, healthy control; IL‐6, interleukin‐6; mPAP, mean pulmonary artery pressure; mRNA, messenger RNA; NC, negative control; NT‐proBNP, NT‐pro–brain natriuretic peptide; PAH, pulmonary arterial hypertension; PBMC, peripheral blood mononuclear cell; PVR, pulmonary vascular resistance; RORC, RAR‐related orphan receptor C; RORγt, receptor‐related orphan receptor γ t; SLE, systemic lupus erythematosus.

To investigate the potential involvement of Activin A in CTD‐PAH pathogenesis, we measured the levels of Activin A in serum samples from 33 healthy controls, 34 patients with SLE, 27 patients with pSjD without PAH, 33 patients with SLE‐PAH, and 24 patients with pSjD‐PAH by ELISA (Supplemental Table 2), and detected highest readouts in patients with SLE‐PAH and pSjD‐PAH than healthy controls and non‐PAH groups (Figure 1D and Figure S1D). Then we analyzed the correlation between Activin A levels and key clinical parameters in patients with SLE‐PAH. Pearson correlation analysis revealed that Activin A levels were positively correlation with N‐Termimal‐pro–brain natriuretic peptide (r _s = 0.904), mPAP (r _s = 0.902), and PVR (r _s = 0.896) and negatively correlation with cardiac index (r _s = −0.920) (Figure 1E). Similar correlation was also observed for IL‐17 (Supplemental Figure 1E), suggesting Activin A and IL‐17 may contribute to SLE‐associated PAH progression and serve as potential biomarkers for SLE‐PAH.

Then, to clarify how IL‐17 was produced in the context of increased Activin A in SLE‐PAH, we isolated CD4⁺ T cells from human PBMCs and stimulated with Activin A alone or in combination with IL‐6 for 18 hours. ⁷ We found that retinoic acid receptor‐related orphan receptor γ t and IL‐17, signatures of Th17 cells, ²⁸ , ²⁹ were up‐regulated in Activin A–treated CD4⁺ T cells, with further elevation in the Activin A with IL‐6 group (Figure 1F and Figure S1F). At the same time, extracellular signal–regulated kinase 1 or 2 phosphorylation, a pathogenic Th17 activation signal, ⁷ was significantly increased after Activin A stimulation (Figure S1G). We also detected increased RAR‐related orphan receptor C (RORC), IL‐17A, and granulocyte–macrophage colony‐stimulating factor (GM‐CSF) at the messenger RNA (mRNA) level in Activin A–treated groups (Figure 1G and Figure S1H). To confirm the dependence of Activin A on Th17 differentiation, WT C57BL/6J mouse spleen CD4⁺ T cells were differentiated into Th17 cells. Suppressing Activin A with Follistatin (a natural Activin A antagonist) ¹⁵ , ³⁰ dose‐dependently reduced RORC and IL‐17A mRNA levels (Supplemental Figure 1H). These findings implied that in the context of SLE‐PAH, elevated Activin A might contribute to CD4⁺ T cells differentiation into Th17 cell with IL‐17 expression.

Activin A receptor ALK4 mediating the release of IL‐17 from Th17 cells and induced endothelial–mesenchymal transition in cocultured PMECs

Activin A binds to naïve CD4⁺ T cells to drive them to produce IL‐17 through activation of the type I receptor ALK4 in the context of SLE‐PAH. At the same time, Activin A may also induce endothelial–mesenchymal transition (EndoMT) in endothelial cells. Considering the complexity of the inflammatory conditions in PAH, we investigated the interaction between PMECs and Th17 cells by coculture assays. In these assays, we further examined the role of the possible key player ALK4 by modulating its expression on Th17 cells to confirm its effect on cocultured PMECs. ⁷ We isolated CD4⁺ T cells from WT C57BL/6J mice, treated these cells with ALK4–short hairpin RNA (shRNA), scramble‐shRNA, or ALK4‐Overexpression (OE) lentiviruses with or without Activin A in the presence of IL‐6, and then cocultured differentiated Th17 cells with PMECs isolated from WT C57BL/6J mice (Figure 2A). The knockdown efficiency of ALK4‐shRNA in Th17 cells was evaluated by quantitative polymerase chain reaction (qPCR) and immunoblotting, as the efficiency of ALK4 overexpression (Figure 2B and Figure S2). Further analysis of gene expression in Th17 cells revealed that the levels of IL‐17, GM‐CSF, and RORC were elevated in the ALK4‐OE group, whereas IL‐10 was increased in the ALK4‐shRNA group compared with the ALK4‐OE and scramble‐shRNA groups, respectively (Figure 2C).

Figure 2.

The Activin A receptor ALK4 mediated the release of IL‐17 from Th17 cells and induced endothelial–mesenchymal transition in cocultured pulmonary microvascular ECs (PMECs). (A) Schematic illustration of CD4⁺ T cells interaction with hPMECs. (B) The mRNA expression of ALK4 after treated with ALK4‐shRNA and ALK4‐OE lentiviruses in CD4⁺ T cells (n = 3). (C) The mRNA expression in CD4⁺ T cells treated with ALK4‐shRNA, ALK4‐OE, and scramble‐shRNA lentivirus after 48 hours (n = 3). (D) The mRNA expression in mouse PMECs cocultured with Th17 cells (n = 3). (E) Representative images showing CD31 (green) and SM22α (red) with DAPI (blue) counterstain for nuclei in mouse pulmonary ECs cocultured with Th17 cells assessed by immunofluorescence microscopy from the indicated groups. Scale bar = 50 μm. SM22α was quantified using ImageJ (right) (n = 5). Data are shown as means ± SEMs, and P values (Bonferroni corrected) are depicted in the panels throughout all figures as *P < 0.05, **P < 0.01, and ***P < 0.001 using one‐way analysis of variance with repeated measures, followed by Bonferroni correction for multiple comparisons. 4‐OE, activin receptor–like kinase 4 overexpression; 4‐shRNA, activin receptor–like kinase 4–short hairpin RNA; 6, IL‐6; ALK4, activin receptor–like kinase 4; α‐SMA, α‐smooth muscle actin; Col1α1, Collagen Type I Alpha 1 Chain; CTGF, connective tissue growth factor; EC, endothelial cell; Fn1, Fibronectin 1; GM‐CSF, granulocyte–macrophage colony‐stimulating factor; hPMECS, human PMECs; IL‐17, interleukin‐17; mRNA, messenger RNA; ns, not significant; PAI1, plasminogen activator inhibitor 1; qPCR, quantitative polymerase chain reaction; RORC, RAR‐related orphan receptor C; S‐shRNA, scramble–short hairpin RNA; shRNA, short hairpin RNA; SM22α, smooth muscle 22α; VE, vascular endothelial; WT, wild type. Color figure can be viewed in the online issue, which is available at http://onlinelibrary.wiley.com/doi/10.1002/art.43235/abstract.

To further confirm that the interaction between Th17 cells and PMECs is mediated by ALK4, we cocultured PMECs and Th17 cells for a prolonged period. We found that the expression of genes related to Activin A‐ALK4 signaling, such as CTGF and plasminogen activator inhibitor 1 (PAI1), extracellular matrix (ECM)–related genes Col1α1, and Fn1, inflammatory factor IL‐6 and mesenchymal markers, including α‐smooth muscle actin (α‐SMA) and Vimentin were significantly increased in the ALK4‐OE coculture group but significantly decreased in the ALK4‐shRNA coculture group. Oppositely, the expression of vascular endothelial (VE)‐Cadherin, an endothelial marker, was decreased in the ALK4‐OE coculture group and increased in the ALK4‐shRNA coculture group (Figure 2D). We also examined smooth muscle cell marker 22α (SM22α) expression in mouse pulmonary endothelial cells determined by immunofluorescent staining. We demonstrated that SM22α was highly expressed in ALK4‐OE coculture group and decreased in ALK4‐shRNA coculture group (Figure 2E). These data suggested that Activin A exerted its effects through ALK4 to drive the differentiation of pathogenic Th17 cells to induce inflammation and detrimental EndoMT in cocultured PMECs.

CD4 ⁺ T cell depletion alleviating hemodynamic changes in hypoxia mice

To further investigate the involvement of CD4⁺ T cells in SLE‐PAH, we generated CD4⁺ T cell–depleted mice by using anti‐CD4 antibody GK1.5 in WT C57BL/6J mice and isotype antibody in control mice under chronic hypoxia (10% O₂) exposure for three weeks (Figure 3A). The effectiveness of CD4⁺ T cells depletion was confirmed by a reduction in CD4⁺ T cells from approximately 13% to 5% of total splenic cells examined by flow cytometry (Figure 3B). Compared with the hypoxia group, CD4⁺ T cell–depleted mice exhibited a significant decrease in RVSP, total pulmonary vascular resistance index (TPVRI), and right heart hypotrophy index (Fulton index) (Figure 3C–E). The serum level of IL‐17 was also reduced in CD4⁺ T cell–depleted mice (Figure 3F). However, CD4⁺ T cell depletion had less effect on reversing remodeled small pulmonary arteries with only partially reduced wall thickness (Figure 3G–I). A marked decrease in IL‐17 levels was observed in the CD31⁺ pulmonary vasculature endothelium cells in CD4⁺ T cell–depleted mice (Figure 3I). Furthermore, immunoblotting of homogenized lungs demonstrated that a reduction in IL‐6 and Smad2 phosphorylation, which was accompanied by decreased CTGF in CD4⁺ T cell–depleted mice (Figure 3J). The analysis of genes expression in lungs showed decreased expression levels of IL‐17A, GM‐CSF, Rorc, inhibin beta A (INHBA), CTGF, Col1α1, IL‐6, and Fn1 in CD4⁺ T cell–depleted mice (Figure S3). These findings suggested that CD4 depletion protected against inflammation and hemodynamic changes in hypoxia‐induced PH but exerted less effects on reversing vascular remodeling.

Figure 3.

CD4⁺ T cell depletion alleviates hemodynamic changes in hypoxia mice. (A) The schematic diagram of chronic hypoxia model in CD4⁺ T cell–depleted mice. (B) Representative flow cytometry histogram of CD4⁺ T cells from spleens and quantification of CD4⁺ T cells (n = 3). (C, D, E) Pulmonary hypertension parameters were as follows: RVSP, TPVRI, and Fulton index (n = 6 mice for each group). (F) Enzyme‐linked immunosorbent assay showing IL‐17 level in the serum of different groups. (G) The representative images of pulmonary arteries by immunohistochemistry staining of SM22α of each group indicated in the panel. Scale bar = 50 μm. Blue arrow indicates small pulmonary arteries. (H) Medial wall thickness of small pulmonary arteries. (I) Representative images showing CD31 (green) and IL‐17 (red) with DAPI (blue) counterstain for nuclei in systemic lupus erythematosus pulmonary hypertension mice. White arrow indicates the pulmonary small arterial. Scale bar = 50 μm. (J) The Western blot result showed that pSmad2, CTGF, and IL‐6 expression in lungs of each group (left) and densitometric quantification is shown in the right (n = 3). Data are shown as means ± SEMs, and P values (Bonferroni corrected) are depicted in the panels throughout all figures as *P < 0.05, **P < 0.01, and ***P < 0.001 using one‐way analysis of variance with repeated measures, followed by Bonferroni correction for multiple comparisons. CD4d, CD4‐deleted; CD4dH, CD4‐depleted and hypoxia; CTGF, connective tissue growth factor; FITC, fluorescein Isothiocyanate; IL‐17, interleukin‐17; LV, left ventricle; ns, not significant; RVSP, right ventricular systolic pressure; RV, right ventricle; SM22α, smooth muscle 22α; TPVRI, total pulmonary vascular resistance index; WT, wild type. Color figure can be viewed in the online issue, which is available at http://onlinelibrary.wiley.com/doi/10.1002/art.43235/abstract.

Aberrant activation of ALK4 but not ALK5 inducing proinflammatory factor and EndoMT gene expression in hPMECs

To investigate Activin A signaling associated with SLE‐PAH pathogenesis, we analyzed the downstream mediator in human PMECs (hPMECs) exposed to patients’ serum samples or Activin A. Immunoblotting revealed significant up‐regulation of pSmad2, along with elevated IL‐6 and CTGF expression, known to induce EndoMT, ³¹ following stimulation with SLE without PAH serum, SLE‐PAH serum, or Activin A (Figure 4A and Figure S4A). Notably, SLE‐PAH serum induced greater protein activation than non‐SLE PAH serum. To confirm Activin A dependent effect, we also used Follistatin and found it reduced Smad2 phosphorylation, IL‐6, and CTGF expression in the presence of Activin A.

Figure 4.

The aberrant activation of ALK4 not ALK5 induced pro‐inflammatory factor and endothelial–mesenchymal transition gene expression in h‐pulmonary microvascular endothelial cells (hPMECs). (A) hPMECs stimulated by serum samples from HCs, patients with SLE, and patients with SLE‐PAH, Activin A (25 ng/mL) and Follistatin (100 ng/mL) for 24 hours were collected and examined by western blot. (B) The mRNA expression in hPMECs induced by Activin A (25 ng/mL and 50 ng/mL) for indicated time (n = 3). (C) The mRNA expression in control, ALK4‐OE, ALK4‐truncated, and vehicle vector groups (n = 3). (D) The mRNA expression in hPMECs transfected with ALK4‐OE and ALK5‐OE plasmid. (E) The pSTAT3, pSmad2, and CTGF expression in hPMECs treated with Stattic were examined by Western blot. (F) The densitometric quantification of protein levels of pSmad2, pSTAT3, and CTGF (n = 3). (G) The overlapping binding region of pSmad2 and pSTAT3 in ctgf promoter region was indicated by red line. Data are shown as means ± SEMs, and P values (Bonferroni corrected) are depicted in the panels throughout all figures as *P < 0.05, **P < 0.01, and ***P < 0.001 using one‐way analysis of variance with repeated measures, followed by Bonferroni correction for multiple comparisons. Each data point in the panels represents one independent participant. 4‐OE, activin receptor–like kinase 4 overexpression; ALK4, activin receptor–like kinase 4; AS, Activin A+Stattic; α‐SMA, α‐smooth muscle actin; Col1α1, col1α1, collagen type I alpha 1 chain; CTGF, connective tissue growth factor; F, Follistatin; Fn1, fibronectin 1; hPMECS, human PMECs; HC, healthy control; IL‐6, interleukin‐6; mRNA, messenger RNA; NC, negative control; ns, not significant; PAH, pulmonary arterial hypertension; PAI1, plasminogen activator inhibitor 1; S, Stattic; SLE, systemic lupus erythematosus; ut, untreated by Activin A; VE, vascular endothelial. Color figure can be viewed in the online issue, which is available at http://onlinelibrary.wiley.com/doi/10.1002/art.43235/abstract.

To examine the direct effect of Activin A on hPMECs, hPMECs were subjected to dose and time gradients assay (25–50 ng/mL for 18–24 hours) (Figure 4B). The results revealed up‐regulation of the Activin A target genes (CTGF, PAI1), ³² proinflammatory factors (IL‐6, IL‐8), and ECM‐related genes (Fn1, Col1α1) in a time‐ and dose‐dependent manner. We also detected increased expression of α‐SMA and Vimentin coupled with decreased VE‐Cadherin (Supplemental Figure 4B), paralleled by CTGF expression and Smad2 phosphorylation over time (Supplemental Figure 4C and D).

Given the role of ALK4 in IL‐17 regulation in CD4⁺ T cells, we engineered ALK4 functional domains through lentiviral constructs, including canonical ALK4 full‐length and kinase‐defective truncation ³³ (Figure S5A). Transfection efficiency in hPMECs was validated by qPCR (Figure S5B). ALK4 overexpression led to significant elevation of ECM‐related genes (CTGF, Fn1, Col1α1), PAI1, and proinflammatory factors (IL‐6, IL‐8) when compared to controls. Conversely, kinase‐truncated ALK4 reversed these expression patterns (Figure 4C and Figure S5C). Both ALK4 and ALK5 mediate Activin A/TGFβ signaling through Smad2 phosphorylation, ¹³ and their different roles in Activin A–induced vascular pathology remain undefined. Our studies revealed ALK4 overexpression specifically enhanced ECM genes expression (CTGF, Fn1, Col1α1) and IL‐6 and EndoMT progression (α‐SMA and Vimentin) in hPMECs, concurrent with Bmpr2 down‐regulation and endothelial marker loss (VE‐Cadherin). Notably, these pathologic changes were more pronounced with ALK4 versus ALK5 activation (Figure 4D and Figure S5D), the protein expression of pSmad2, IL‐6, and CTGF by overexpression ALK4 and ALK5 were also confirmed with immunoblotting (Figure S5E).

Considering the key role of CTGF in EndoMT and ECM genes expression, ³⁴ we investigated its transcriptional regulation in Activin A–stimulated hPMECs. The results revealed Stat3 inhibition, via selective antagonist Stattic, ³⁵ attenuated Smad2 phosphorylation and CTGF expression (Figure 4E and F), establishing STAT3 as a synergistical signal molecule for Smad2‐mediated CTGF activation. Mining of chromatin profiling data sets (GSM2635691 and GSM2752894) identified colocalized pSmad2/pSTAT3 binding motifs within the CTGF promoter on chromosome 6 (Figure 4G). These transcription factors exhibited overlapping occupancy at CTGF promoter region, indicating with red rectangles in Figure 4G. These findings suggested ALK4 rather than ALK5 in up‐regulating ECM‐related genes and IL‐6, which activated pSTAT3 to synergize with pSmad2 to induce CTGF expression in hPMECs.

ALK4 overexpression inducing severe PH in SLE mice

To confirm the importance of ALK4 in vivo, we aimed to examine the detrimental role of ALK4 by establishing an SLE‐PH mouse model (Figure 5A). In our previous study, mice with SLE‐PH induced with pristane and exposure to hypoxic conditions (10% O₂) (SLE‐PH mice) exhibited only a mild increase in RVSP without right heart hypotrophy. Therefore, we treated mice weekly with an ALK4‐overexpressing lentivirus via intranasal instillation to reproduce the increased activation of Activin A in the SLE‐PH mouse model and further monitored the mice for five weeks. We observed an increase in the RVSP to approximately 35 mm Hg and a significant increase in the Fulton index (Figure 5B–D).

Figure 5.

ALK4 overexpression induced severe pulmonary hypertension in SLE mice. (A) Schematic illustration of ALK4 overexpression and knockdown in SLE‐PH mice. (B, C, D) Hemodynamic indices including RVSP, TPVRI, and Fulton index were measured in the SLE‐PH mice (n = 6 each). (E) Enzyme‐linked immunosorbent assay showing Activin A, IL‐17, and anti‐dsDNA autoantibody level in the serum of different groups. (F) The representative images of pulmonary arteries (Scale bar = 50 μm) by HE and immunohistochemistry staining of SM22α, ALK4, and IL‐17 in each group. Blue arrows indicated small pulmonary arteries. (G) Representative images showed CD31 (green) and pSmad2 (red) with DAPI (blue) in SLE‐PH mice. White arrows indicated the pulmonary small arteries. Scale bar = 50 μm. (H) The western blot result showed that ALK4, pSmad2, and CTGF expression in the lungs of each group (left) with quantification (right). Data are shown as means ± SEMs, and P values (Bonferroni corrected) are depicted in the panels throughout all figures as *P < 0.05, **P < 0.01 and ***P < 0.001 using one‐way analysis of variance with repeated measures, followed by Bonferroni correction for multiple comparisons. 4‐OE, activin receptor–like kinase 4 overexpression; 4‐shRNA, activin receptor–like kinase 4–short hairpin RNA; ALK4, activin receptor–like kinase 4; anti‐dsDNA, anti–double‐stranded DNA; CTGF, connective tissue growth factor; HE, hematoxylin and eosin; IL‐17, interleukin‐17; LV, left ventricle; ns, not significant; PH, pulmonary hypertension; RV, right ventricle; RVSP, right ventricular systolic pressure; S‐shRNA, scramble–short hairpin RNA; shRNA, short hairpin RNA; SLE, systemic lupus erythematosus; SM22α, smooth muscle 22α; TPVRI, total pulmonary vascular resistance index; WT, wild type. [Correction added on 10 September 2025, after first online publication: Figure 5 has been updated in this version.] Color figure can be viewed in the online issue, which is available at http://onlinelibrary.wiley.com/doi/10.1002/art.43235/abstract.

We also applied ALK4 shRNA to confirm the reversal effects of knocking down ALK4. Compared with that in the control group, ALK4 knockdown resulted in considerable amelioration of the increase in the RVSP, a reduced TPVRI and a decreased Fulton index (Figure 5B–D). The serum levels of Activin A, IL‐17, and mouse anti–double stranded DNA (anti‐dsDNA) antibody, which are markers of SLE in humans, ³⁶ were also decreased (Figure 5E). Alleviation of pulmonary inflammation and decreased perivascular IL‐17 levels were observed in the ALK4 knockdown group, accompanied by reduced infiltration of immune cells, as assessed by hematoxylin and eosin staining and immunohistochemistry staining of IL‐17 (Figure 5F), and the wall thickness of the small pulmonary arteries was markedly decreased (Figure S6A). The decrease in pSmad2 levels was observed in ALK4 knockdown PMECs (Figure 5G). A reduction in ALK4 and Smad2 phosphorylation was accompanied by reduced CTGF protein levels in the ALK4 knockdown group (Figure 5H). Further gene expression analysis revealed significant decreases in the expression levels of IL‐17, Rorc, CTGF, Col1α1, IL‐6, GM‐CSF, and PAI1 in the ALK4 knockdown group, whereas the expression of IL‐10 was increased in this group (Figure S6B). In addition, the expression levels of EndoMT markers (a‐SMA and Vimentin) decreased and the expression of VE‐cadherin was restored in the ALK4 knockdown group. Our results showed that ALK4 had a detrimental effect on pulmonary vascular remodeling in SLE‐PH mice and that decreasing ALK4 expression reversed pulmonary hypertension in vivo.

The potent ALK4 inhibitor TEW reversing PH by suppressing IL‐17 expression in the lungs and EndoMT in hPMECs

To evaluate whether inhibition of ALK4 can suppress inflammation and reverse SLE‐PAH, we used the ALK4 inhibitor TEW, which is in a phase 2 clinical trial for breast cancer therapy with low half maximal inhibitory concentration (IC50) (13 nM) and minimal side effects. ³⁷ , ³⁸ In SLE‐PH mice, TEW treatment with reported dose, ³⁹ , ⁴⁰ reduced lung edema (Figure 6A) and reversed elevated RVSP, TPVRI, and Fulton index (Figure 6B–D). Serum IL‐17, Activin A, and anti‐dsDNA autoantibody levels, elevated in SLE or SLE‐PH groups, were significantly reduced by TEW (Figure 6E). TEW reduced perivascular IL‐17, lung inflammation, and immune cell infiltration (Figure 6F), as well as pulmonary artery wall thickness (Figure 6G). TEW also reversed SLE‐PH‐induced elevations in IL‐17, IL‐6, and ECM genes (CTGF, Col1α1, Fn1), and PAI1 and suppressed EndoMT markers (α‐SMA, Vimentin), restoring VE‐cadherin expression in lung tissues (Figure S6C).

Figure 6.

The potent ALK4 inhibitor TEW reversed pulmonary hypertension. (A) Representative images of the lung of control and treated mice (Scale bars = 1cm). (B, C, D) Hemodynamic indices including RVSP, TPVRI, and Fulton index in each group (n = 6). (E) Enzyme‐linked immunosorbent assay showing IL‐17, anti‐dsDNA autobody, and Activin A level in the serum from different groups. (F) The representative images of pulmonary arteries (scale bar = 50 μm) by HE and IL‐17 and SM22α immunohistochemistry staining of each group. Blue arrows indicated the small pulmonary arteries. (G) Medial wall thickness of small pulmonary arteries. (H) The mRNA expression in h‐pulmonary microvascular endothelial cells (hPMECs) treated with TEW after stimulation with IL‐17 and Activin A (n = 3). (I) IL‐6rα level in the supernatant of hPMECs treated with TEW (n = 3). (J) The protein expressions of pSmad2, Bmpr2, pSmad1/5, CTGF and IL‐6 in hPMECs treated with TEW were examined by western blot. (K) The densitometric quantification of protein levels of pSmad2, CTGF, Bmpr2, and pSmad1/5 (n = 3). (L) The mRNA expression in Th17 cells treated with TEW after stimulation with Activin A (n = 3). Data are shown as means ± SEMs, and P values (Bonferroni corrected) are depicted in the panels as *P < 0.05, **P < 0.01, and ***P < 0.001 using one‐way analysis of variance with repeated measures, followed by Bonferroni correction for multiple comparisons. ALK4, activin receptor–like kinase 4; anti‐dsDNA, anti–double‐stranded DNA; α‐SMA, α‐smooth muscle actin; AT, Activin A with TEW; Bmpr2, bone morphogenetic protein type II receptor; CTGF, connective tissue growth factor; GM‐CSF, granulocyte–macrophage colony‐stimulating factor; HE, hematoxylin and eosin; HPMECS, human PMECs; IA, IL‐17 with Activin A; IL‐17, interleukin‐17; LV, left ventricle; mRNA, messenger RNA; ns, not significant; PH, pulmonary hypertension; RV, right ventricle; RVSP, right ventricular systolic pressure; SLE, systemic lupus erythematosus; SM22α, smooth muscle 22α; T‐20, TEW (20 mg/kg); T‐40, TEW (40 mg/kg); TEW, vactosertib (TEW‐7197, shorten as TEW); TPVRI, total pulmonary vascular resistance index. Color figure can be viewed in the online issue, which is available at http://onlinelibrary.wiley.com/doi/10.1002/art.43235/abstract.

Specifically, TEW inhibits phosphorylation of Smad2. ³⁷ , ⁴¹ , ⁴² Based on the efficient doses reported previously, ⁴⁰ , ⁴³ we applied three different doses for the in vitro assay. A significant increase in CTGF, IL‐6, and α‐SMA expression was observed in hPMECs activated by IL‐17 and Activin A, whereas the expression of the endothelial marker VE‐Cadherin was decreased. Treatment with TEW effectively reversed the IL‐17– and Activin A–induced alterations in gene expression (Figure 6H and Figure S7A) and IL‐6rα secretion (Figure 6I) from hPMECs. We further assessed the effects of TEW on CTGF, IL‐6, and BMPR2 signaling in hPMECs. As shown in Figure 6J and K and Figure S7B, TEW suppressed Activin A–induced pSmad2, CTGF, and IL‐6 expression while restoring BMPR2 levels and pSmad1/5 signaling impaired by Activin A. IL‐17, GM‐CSF, and RORC expression were significantly reduced, whereas IL‐10 expression was recovered in TEW‐treated Th17 cells derived from mice CD4⁺ T cells (Figure 6L and Figure S7C). These results demonstrated that TEW treatment reversed pathologic pulmonary vascular remodeling and the development of SLE‐PH. It effectively suppressed Th17 cells induced inflammation and EndoMT in hPMECs while balancing BMPR2 signaling.

TEW reversing PAH in Monocrotaline (MCT)‐treated rats by rebalancing signaling through the Activin A/ALK4 and BMPR2 pathways

To further confirm that TEW alleviated PAH in an inflammation‐related rat model in which the animals could be catheterized more accurately, we used a rat model of MCT‐induced PAH to reproduce the pathologic characteristics of patients, including the infiltration of inflammatory cells in the pulmonary vasculature. ⁴⁴ Severe lung damages were observed in the MCT group, whereas TEW treatment groups displayed less pulmonary hemorrhage (Figure S8A and B). The survival rates of all groups of rats are demonstrated in Figure S8C. TEW treatment with 40 mg/kg significantly improved the survival rate of MCT‐induced rats.

Our results showed a significant improvement in RVSP, the Fulton index, and the TRVRI in MCT rats treated with TEW (Supplemental Figure 8D–F). We observed that RVSP reduced to nearly 35 mm Hg with lower Fulton index and total pulmonary vascular resistance index (TPVRI) in the Tew‐40 mg/kg group than in the MCT group. Serum IL‐17 and Activin A levels were significantly elevated in the MCT group and decreased in the rats administered TEW (Figure S8G). TEW treatment also significantly reduced lung inflammation and perivascular IL‐17 expression in the pulmonary vascular and significantly decreased the wall thickness of small pulmonary arteries (Figure S9A). Additionally, the gene expression of IL‐17, CTGF, IL‐6, Inhba, Rorc, GM‐CSF, Col1α1, Fn1, Cxcr1, and PAI1 decreased with TEW treatment (Figure S8H and Figure S9B). We also found that TEW alleviated EndoMT in MCT rats. α‐SMA and Vimentin decreased, and VE‐Cadherin increased. Western blot analysis revealed a reduction in pSmad2 protein level and an elevation in Bmpr2 in the TEW‐treated group (Figure S8I). Immunofluorescence staining showed decreased pSmad2 in TEW‐treated rats (Figure S8J).

We also conducted TEW in SU5416‐induced PH with hypoxia in rats, which blocked vascular endothelial growth factor receptor 2 and caused severe pulmonary vascular remodeling (Figure S10A). Images of the lung from each group are shown in Figure S10B. The whole lungs showed poor blood circulation and supply in the Sugen5416 with hypoxia (SuHx) group. Our results showed a significant improvement in hemodynamic indices in the TEW‐treated group, which RVSP decreased from 40 to 35 mm Hg with a lower Fulton index and TPVRI than in the SuHx group (Figure S10C–E). With TEW treatment, serum IL‐17 and Activin A levels were significantly reduced (Figure S10F); moreover, immune cell infiltration and perivascular IL‐17 levels decreased (Figure S10G). The wall thickness of pulmonary small arteries was significantly reduced with TEW treatment (Figure S10H), as was the gene expression of IL‐17, IL‐6, Inhba, Col1α1, and PAI1 decreased with TEW treatment as well as mesenchymal markers (α‐SMA and Vimentin) and VE‐Cadherin increased (Figure S10I). The previously mentioned results confirmed the cell signaling by which TEW inhibits Activin A/ALK4 signaling in immunologically dysfunctional PAH rat models.

DISCUSSION

This study aimed to elucidate the role of Activin A‐ALK4 signaling in SLE‐PAH, with a special focus on the molecular targets that alleviate the initial inflammatory insult on the most vulnerable endothelial layer during the onset of PAH. The reduced CD4⁺ T cell population in patients after treatment and elevated Th17 cells in PBMCs of patients with SLE‐PAH suggested that Th17 cells are involved in SLE‐PAH. We observed significantly higher levels of IL‐17 and Activin A in serum of patients with SLE‐PAH than in those of patients with SLE without PAH. Coculture experiments revealed that ALK4 mediates the interaction between immune cells and endothelial cells. We observed increased expression of CTGF and proinflammatory genes downstream of the Activin A/ALK4 pathway in vitro and in vivo, which suggested the detrimental role of ALK4 in promoting a pathogenic behavior of Th17 cells to induce inflammation and pulmonary vascular remodeling. The specific ALK4 inhibitor TEW effectively suppressed the release of IL‐17 by Th17 cells and prevented EndoMT in hPMECs in an SLE‐PH mouse model and an MCT‐PAH rat model. Finally, we confirmed that CTGF gene transcription is regulated by Activin A/pSmad2 and IL‐6/pSTAT3 in hPMECs. We first addressed the importance of targeting Activin A/ALK4, which also relies on suppressing pathogenic Th17 cells, to strongly contribute to alleviating PAH in the context of inflammation‐related disorders, such as SLE.

We used mass cytometry to observe the dynamic involvement of immune cells before and after treatment. Mass cytometry enables the analysis of immune cells at the single‐cell level, allowing the examination of the intricate immune status of patients with SLE‐PAH. Following immunosuppressant and vasodilator combination therapy, we observed reductions in the populations of CD4⁺ T cells in patients with SLE‐PAH, which allowed us to identify the source of the increase in the IL‐17 level.

The current therapeutic strategy for SLE‐PAH involves combination treatment including immunosuppressants along with conventional vasodilator for PAH, which increases the risk of side effects. ¹ So far, no therapy has been shown to achieve the “one stone kills two birds” effect for SLE‐PAH in the clinic. Immunosuppressive medications, primarily glucocorticoids, cyclophosphamide, and cyclosporine A, are fundamental for SLE treatment. However, these treatments fail to reverse pulmonary vascular remodeling. To improve the prognosis of patients with SLE‐PAH, recent efforts have been made to restore the balance between Activin A and BMPR2 signaling during PH progression. ⁴⁵ Hydroxychloroquine and tacrolimus were not only immunosuppressants but also reported to balance BMPR2 signaling in PAH. ⁴⁶ , ⁴⁷ For patient 1 with SLE‐PAH, tacrolimus was applied to replace cyclophosphamide, and patient 4 with SLE‐PAH reached low‐risk PAH even without vasodilator in this study. A Study of Sotatercept for the Treatment of Pulmonary Arterial Hypertension (STELLAR) trial, ²² , ⁴⁸ suggesting the clinical effectiveness of targeting the Activin A signaling pathway in the context of CTD‐PAH. ²¹ Our findings implied the potentiality of Activin A–BMPR2 balance modulator in addition to the immunosuppressive effect in SLE‐PAH treatment.

Elevated levels of IL‐6, which plays an important role in autoimmune diseases, have also been found in patients with PAH. ¹ , ⁴⁹ IL‐6 is produced by macrophages and T cells through pSTAT3 to facilitate Smad2 phosphorylation upon Activin A activation. Our findings revealed that inhibiting pSTAT3 in hPMECs with Stattic led to decreased pSmad2 and CTGF levels. Thus, it is worth further exploring the interaction between the Activin A/pSmad2/CTGF pathway and IL‐6/pSTAT3 to determine the functional significance of the activation of pSTAT3 in hPMECs. Although we predicted the pSmad2 and pSTAT3 binding sites in the ctgf promoter region via database searches and JASPAR, the precise binding sites of pSmad2 and pSTAT3 in the ctgf promoter on chromosome 6 in hPMECs and the epigenetic modifications of this promoter remain to be confirmed. Although our data suggested a potential association between Activin A signaling and Th17 cell differentiation in SLE‐PAH, the mechanistic interplay between ALK4 and Th17 cells remains to be fully elucidated. The limitations of this study also include the development of targeted therapies, for example, whether the source of Activin A in patients with SLE‐PAH is the same as that in patients with PAH. Further profiling Activin A expression in PBMCs in patients with SLE‐PAH may be able to provide direct evidence.

In conclusion, our research revealed that abnormally activated ALK4 inducing IL‐17 production and secretion in pathogenic Th17 cells in the context of SLE‐PAH. Meanwhile, Activin A binds to ALK4 and promotes the expression of CTGF, α‐SMA, and IL‐6 in hPMECs. Moreover, IL‐6 and pSTAT3 synergistically induce CTGF expression in hPMECs. TEW effectively inhibits ALK4 activity in hPMECs and Th17 cells to alleviate inflammation and EndoMT in the context of SLE‐PAH and PAH associated with other inflammatory conditions. Furthermore, we intend to use Activin A as a biomarker to identify the most suitable patient cohort and explore the potential development of TEW for therapeutic application.

AUTHOR CONTRIBUTIONS

All authors contributed to at least one of the following manuscript preparation roles: conceptualization AND/OR methodology, software, investigation, formal analysis, data curation, visualization, and validation AND drafting or reviewing/editing the final draft. As corresponding author, Drs Li and Yang confirm that all authors have provided the final approval of the version to be published and take responsibility for the affirmations regarding article submission (eg, not under consideration by another journal), the integrity of the data presented, and the statements regarding compliance with institutional review board/Declaration of Helsinki requirements.

Supporting information

Disclosure Form:

Appendix S1. Supporting Information

Data availability statement

The data that support the findings of this study are available on request from the corresponding author. All data relevant to the study are included in the article or uploaded as supplemental information.