Inhibition of immunoglobulin E attenuates pulmonary hypertension

Pulmonary hypertension (PH) is a severe cardiopulmonary disease characterized by pulmonary vascular remodeling. Immunoglobulin E (IgE) is known to participate in aortic vascular remodeling, but whether IgE mediates pulmonary vascular disease is unknown. In the present study, we found serum IgE elevation in pulmonary arterial hypertension (PAH) patients, hypoxia-induced PH mice and monocrotaline-induced PH rats. Neutralizing IgE with an anti-IgE antibody was effective in preventing PH development in mice and rat models. The IgE receptor FcεRIα was also upregulated in PH lung tissues and Fcer1a deficiency prevented the development of PH. Single-cell RNA-sequencing revealed that FcεRIα was mostly expressed in mast cells (MCs) and MC-specific Fcer1a knockout protected against PH in mice. IgE-activated MCs produced interleukin (IL)-6 and IL-13, which subsequently promoted vascular muscularization. Clinically approved IgE antibody omalizumab alleviated the progression of established PH in rats. Using genetic and pharmacological approaches, we have demonstrated that blocking IgE–FcεRIα signaling may hold potential for PAH treatment. Using rat and mouse models of pulmonary hypertension and patients’ data, Shu, Liu, Zhou et al. show that the concomitant increase in immunoglobulin E (IgE) and mast cells expressing the effector receptor FcεRIα has an important role in pulmonary vascular remodeling, and genetic and pharmacological inhibition of the IgE–FcεRIα signaling alleviated the progression of pulmonary hypertension in animal models.


Introduction
Pulmonary hypertension (PH) is a progressive cardiopulmonary disease that results in increased mean pulmonary arterial pressure (mPAP), ultimately leading to failure of the right ventricle and death.
Pulmonary vascular remodeling and in ltration of immune cells are important pathological features in pulmonary arterial hypertension (PAH, WHO Group 1) [1,2]. Chronic in ammation is observed in lung tissues from experimental PH models and PAH patients. Immune cells and pro-in ammatory factors are thought to contribute to vascular remodeling and muscularization. Of note, lymphoid follicles are found in hypoxia-exposed animals and contain T cells, B cells and dendritic cells [3,4]. Moreover, IPAH patients without auto-immune diseases also produce auto-antibodies [5][6][7], suggesting that the adaptive immune response occurs locally to generate antibodies in the lung. Currently available therapies for PAH are unable to reverse the pathological vascular remodeling and do not target the dysregulated immune response in the vessel wall. Elucidating the mechanism of vascular remodeling and in ammation may facilitate the identi cation of novel and more effective therapeutic targets.
Immunoglobulin E (IgE) is an antibody secreted by B cells that undergo IgE class-switch recombination and up-regulate germ-line ε (Glε) transcription [8] in response to Th2 stimulation. B cells and tertiary lymphoid organs are found in arterial wall lesions in atherosclerosis [9]. We previously reported that serum IgE is elevated during the aortic vascular remodeling in cardiovascular diseases such as atherosclerosis [10] and abdominal aortic aneurysm [11]. IgE aggravates perivascular in ammation by binding to FcεRIα, the α-subunit of its high a nity receptor [12]. Mast cells (MCs) and other cell types play a major role in cardiovascular diseases mediated by high serum IgE levels [10]. Moreover, MC in ltration has been identi ed in the lung tissues from patients with idiopathic PAH (IPAH) and rats with monocrotaline (MCT)-induced PH [13,14], while inhibition of MCs by cromolyn, an MC stabilizer, repressed the development of PH in rats [13]. However, the roles of IgE in PH development and the cell types mediating its effects are not known, and whether this effect is amenable to therapeutic targeting remains unclear.
In this study, we identi ed increased serum IgE levels in clinic IPAH patients and experimental PH models. We showed that blocking IgE effectively prevented PH in multiple animal models. We demonstrated that IgE promoted vascular muscularization by binding with FcεRIα, activating MCs and promoting the release of IL6 and IL13. The clinically used anti-IgE antibody Omalizumab showed therapeutic effect in rats with established PH. Informed consents to use lung tissue for research were provided by PAH participants before enrollment.

Single-cell RNA-seq and analysis
Lung tissues from mice under hypoxia exposure for 0, 1, 2 and 4 weeks were harvested and scRNA-seq were performed. A single-cell cDNA library was generated as previously described [15]. More details of the analysis are available in the supplementary material.

Statistical analyses
All statistical analyses were performed using GraphPad Prism 8.0 and SPSS version 23.0. Data are presented as means ± standard error of mean (SEM). The Kolmogorov-Smirnov test was used to assess for normality. An unpaired two-tailed t-test was performed to compare the difference between two groups. A one-way ANOVA with LSD-t test or two-way ANOVA with Bonferroni's post hoc test was performed to evaluate differences between multiple groups. A p-value <0.05 was considered statistically signi cant. Randomization and blinded analyses were used whenever possible.

Additional materials and methods
Additional materials and methods are available in the supplementary material online.

Elevated serum IgE in PAH patients and PH models
We examined IPAH/HPAH patients who did not have allergy, infection or autoimmune diseases (Supplementary Table 1), and found that these PAH patients showed elevated serum IgE levels in comparison with age/sex-matched healthy subjects (Fig.1A). In experimental PH models, elevated serum IgE levels were observed in hypoxia-induced PH mice (Fig.1B) and MCT-induced PH rats (Fig.1C) when compared with control animals.
Adaptive immune response is observed in PH lung [16], therefore we hypothesized that the elevated IgE is produced from lung tissues. To characterize the local immune response in lung tissues and the cells involved, single cell RNA sequencing (scRNA-seq) was performed in lung tissues from mice exposed to hypoxia at different time points (Supplementary Fig.1). The proportion of B cells (identi ed by their characteristic enrichment of Cd19, Fig.1D and E) was increased under hypoxia (Fig.1F). Differentially expressed genes (DEGs, Fig.1G) and their enriched Gene Ontology (GO) terms between consecutive pairs of time points revealed "B cell activation" and "B cell differentiation" at the early stages of PH ( Supplementary Fig.2). GO terms of "regulation of chromosome organization", "regulation of immunoglobin mediated immune response", and "immunoglobulin mediated immune response" (Supplementary Fig.2) suggested that antibody class-switch recombination may have occurred in B cells during PH development. We also found enhanced Th2 response (Supplementary Fig.3A-D), up-regulated transcription of genes in the IL4 pathway in T cells ( Supplementary Fig.3E), as well as enhanced Cd4 + T-B cell interactions ( Supplementary Fig.3F-H), supporting that IgE was produced by activated B cells in PH lung via a classical mechanism. These results were experimentally con rmed by the increased proportion of activated B cells (CD21 + CD23 + ) in hypoxic lung (Fig.1H). Moreover, the upregulation of Glε mRNA expression in both CD19 + B cells and PH lung tissues (Fig.1I) further con rmed that active B cells switched to the IgE isotype. The increased percentage of IgE + cells in PH lungs (Fig.1J) potentially resulted in the elevation of serum IgE levels during PH development.

Blocking serum IgE attenuated the development of PH in mice and rats
To determine the function of IgE in PH development, we used an experimental anti-IgE antibody to neutralize serum IgE in a hypoxic mouse model of PH ( Fig.2A, Supplementary Fig.4A). In mice exposed to hypoxia, the anti-IgE antibody signi cantly attenuated the increase in right ventricular systolic pressure

Fcer1a de ciency attenuated hypoxia-induced PH in mice
To identify the target receptor for IgE in PH development, we assessed the expression of the high a nity receptor FcεRIα in the lung tissues from PAH patients. We observed that FcεRIα mRNA and protein levels were signi cantly up-regulated in the patient lung ( Fig.3A and B). Immunohistochemical staining showed that FcεRIα positive cells were located around pulmonary vessels (Fig.3C), suggesting their potential involvement in pulmonary vascular remodeling. In lung tissues from hypoxia-exposed mice ( Fig.3D and E) and MCT-injected rats ( Fig.3F and G), mRNA and protein levels of FcεRIα were also increased. To investigate the role of FcεRIα in IgE-mediated PH, Fcer1a +/+ (WT) and Fcer1a -/-(KO) mice were exposed to hypoxia to establish PH (Fig.3H). As shown in Fig.3I-K, Fcer1a de ciency protected against hypoxiainduced elevation in RVSP and RV hypertrophy, and decrease in PA AT/ET compared with WT mice.
Histological analyses showed less pulmonary vascular thickening and muscularization in response to hypoxia in Fcer1a knockout mice compared to WT controls (Fig.3L-N). These results indicated that FcεRIα was the receptor involved in IgE-mediated PH development.

FcεRIα expressed in mast cells contributed to PH development in mice
To identify the FcεRIα expressing cells in lung tissues, we analyzed our mouse scRNA-seq data and found that Fcer1a was mainly enriched in MCs (Fig.4A, top). Similarly, the published scRNA-seq data [17] in normal human lungs revealed FCER1A expression in MCs (Fig.4A, bottom). These results were con rmed by the co-localization of FcεRIα + cells and MC markers (c-Kit + cells, Fig.4B). Moreover, we showed an increased percentage of CD45 + c-Kit + FcεRIα + cells (MCs) in hypoxic mice (Fig.4C). This observation is consistent with the previous report of MCs in ltration in lung tissues of PAH patients and PH rats [18].
To evaluate the function of MC-expressed FcεRIα in PH development, MC-speci c Fcer1a knockout mice (Fcer1a MC-/-, MCKO) were generated by crossbreeding Fcer1a ox/ ox mice (WT) mice with Mcpt5-Cre transgenic mice (Fig.4D). Four weeks of hypoxia exposure resulted in PH in WT mice, indicated by increased RVSP and decreased PA AT/ET, which were improved in the MCKO mice ( Fig.4E and G). No signi cant changes were observed in RV hypertrophy, consistent with previous ndings of cromolyn administration in MCT model [13] (Fig.4F). Histological analyses showed decreased wall thickness and vascular muscularization in MCKO compared with WT mice under hypoxia (Fig.4H-J). Together, these results suggested that FcεRIα expressed in MCs contributed to the development of PH.

IgE stimulated MCs to produce IL6 and IL13 in experimental PH models
In order to identify the factors by which IgE-activated MCs mediate PH pathogenesis, MCs were treated with IgE to mimic their response to IgE in lung tissues, followed by RNA-seq to identify IgE-regulated genes. Previous studies indicated that the MCs were recruited from bone marrow to the lung in PH [14]; therefore, bone marrow derived mast cells (BMMCs) were used for this experiment. As shown in Fig.5A, IgE treatment altered the expression of 6,041 genes in BMMCs. GO analyses showed that the IgEregulated genes were primarily associated with in ammation and vascular cell dysfunction (Fig.5B). Importantly, the DEGs enriched in these GO terms included many secreted factors that are known to be PH-related, such as Il6, Il13, Il33, Vegfa, Vegfb, Vegfc, Fgf2, Fgf18, Hbegf, Pdgfa, Ccl2 and Ccl4 (Fig.5A). By combining these 12 genes from the RNA-seq data of IgE-stimulated MCs with our scRNA-seq data from PH animals, we found three candidate genes (Il6, Ccl4 and Il13) that were mainly expressed by MCs and responded to IgE stimulation (Fig.5C). RT-qPCR results con rmed the signi cantly up-regulated expression of Il6 and Il13, but not Ccl4, in CD45 + c-Kit + FcεRIα + MCs sorted from hypoxic lung tissues (Fig.5D). Moreover, increased Il6 and Il13 expression was observed in lung tissues from hypoxic mice (Fig.5E), and their expression was reduced under IgE-FcεRIα blockade, as shown in anti-IgE antibody-treated mice (Fig.5F), Fcer1a knockout mice (Fig.5G), as well as in MC-speci c Fcer1a knockout mice (Fig.5H). Similarly, Il6 and Il13 expression was also increased in MCT-injected rats ( Supplementary  Fig.5A), and was reduced by anti-IgE (Supplementary Fig.5B). Co-culture of MCs and PASMCs was established to study the effects of MC derived IL6 and IL13 (Fig.5I). The conditioned medium of IgEactivated MCs promoted PASMC proliferation, while this pro-proliferative effect of the conditioned medium was reduced with knockdown of Il6 and Il13 in MCs (Fig.5J and K). Through a series of transcriptomic analysis and experimental validation, we have found that IgE-activated MCs produced IL6 and IL13, which potentially contributed to IgE-mediated PH in animal models.

Clinically-used IgE antibody Omalizumab ameliorated established PH in rats
We next evaluated the effect of anti-IgE therapy on the progression of established PH in rats using Omalizumab, a clinically-used recombinant human monoclonal antibody. The treatment was started from 14 days post-MCT injection, when the pulmonary vascular remodeling and mPAP elevation were already present [19] (Fig.6A). After two weeks of treatment, Omalizumab effectively lowered the serum IgE levels ( Fig.6B), reduced mPAP (Fig.6C) and RV hypertrophy (Fig.6D), improved pulmonary artery ejection function (Fig.6E), and attenuated pulmonary vascular wall thickening and muscularization (Fig.6F-H). Anti-IgE therapy by Omalizumab also decreased Il6 and Il13 expression effectively in lung tissues of MCTinduced PH rats (Fig.6I). These results indicate that Omalizumab attenuates the development of established PH, and targeting IgE may serve as an effective therapy for PH patients especially with high serum IgE.

Discussion
In this study, we have discovered that IgE was elevated in PAH patients and animal models of PH, and this promoted the pathogenesis of disease. We have shown that blocking IgE effectively attenuated the development of PH in different models. Using constitutive and cell type-speci c knockout animals, IgE was found to promote PH by stimulating MCs to produce IL6 and IL13, revealing a novel immune-based mechanism in PH ( Supplementary Fig.6). This is the rst report of the function of IgE in pulmonary vascular disease. Our ndings provide pre-clinical evidence supporting the translation of anti-IgE therapies such as Omalizumab in the treatment of PAH.
Elevation of IgE is well recognized to be important in cardiovascular diseases. Our study showed that IgE was also elevated in PH, released from stimulated B cells in lung tissues. Previous studies have reported that auto-antigen exposure in pulmonary arterial vessels could activate the immune system in PH lung tissues [5,7]. These auto-antigens, including lamin A/C, tubulin β-chain in endothelial cells, as well as vimentin and calumenin in broblasts, triggered the production and binding of IgG/IgM isotype antibody. Moreover, in ltration of Th2 cells and production of Th2 cytokines (IL4, IL5, IL13) have been reported in lung tissues from PAH patients or PH animal models [20,21]. In this study, we rst revealed increased Th2 cell-B cell interaction in the lung tissues of PH mice through scRNA-seq. This process may promote antibody class switching into the IgE isotype in B cells, subsequently leading to increased IgE production. Collectively, we propose that the exposure of auto-antigens in damaged pulmonary vessels elicits a local adaptive immune response, resulting in IgE production during PH development.
Exogenous antigens such as mite or pollen may also cause IgE elevation, in the context of asthma or allergic reaction. However, there is no con rmed clinical relationship between asthma and PH. IgE is not necessary for all types of asthma, as IgE elevation and reactivity are not observed in patients with IgEindependent asthma [22]. Besides, The IgE level has an aggravated elevation in patients with asthma at acute attack, which gets back to normal level when patient is recovered [23]. More importantly, IgE derived from exogenous antigens shows component-speci city. Component-speci c IgE may exhibit different disease-causing effects, as illustrated in other types of cardiovascular diseases [24]. In our study, we did not use PH models caused by dust mites or ovalbumin, and the serum and tissue samples were collected from PAH patients without allergy, infection or autoimmune diseases. Hence, IgE was activated by endogenous auto-antigens and was functional in PH development. Transfusing puri ed IgE from PH rats into naïve rat could support the function of endogenous IgE in PH. This experiment is worthwhile further investigation once the puri cation technique is improved.
FcεRIα is the high a nity receptor that interacts with IgE directly. We observed increased FcεRIα expression in the lung tissues of rodent PH models and PAH patients. Perivascular FcεRIα + cells were rare in the lung tissues of non-PAH subjects, which was consistent with a previous report [25]. Remarkably, in PAH lung tissues, we found signi cant increases in FcεRIα + cells around the vascular lesions. The mechanism driving the increase in FcεRIα expression in PH lung is unclear. Importantly, scRNA-seq data of mice, rats and human lung tissues all revealed FcεRIα expression in MCs. Bone-marrow derived mast cells, which in ltrate into the lung tissue and constitute the perivascular MCs, possess high FcεRIα expression [26], but whether these in ltrated MCs contributed to elevating FcεRIα level warrants further investigation. Based on our results, a concomitant increase in IgE production and its effector cells with the high a nity receptor gives rise to a series of downstream responses that may ultimately result in vascular remodeling.
Activation of MCs has been reported in MCT-induced PH rat and human lung tissues from PAH patients [13,14]. Stabilizing MCs through cromolyn in the early stage protects against MCT-induced PH development [13]. However, the mechanism by which MCs are activated remains unknown to date. Our results revealed that increased IgE in PH activated MCs and contributed to the development of PH. Among a range of MC-released downstream mediators, we narrowed down the predominant factors to IL6 and IL13 using transcriptomic and experimental approaches. IL6 is widely recognized to be important in PH, as circulating IL6 was elevated in patients with PAH [27,28]. Moreover, IL6 was reported to be important for IgE generation from B cells, and MC-derived IL6 was shown to promote PH by stimulating B cells [29]. These results suggested that MC-derived IL6 may reciprocally activate B cells to secrete more IgE, forming a vicious cycle in disease pathogenesis. Additionally, IL13 may also participate in PH, as transgenic mice overexpressing IL13 in the lung spontaneously developed PH. IL13 was reported to promote the proliferation of pulmonary arterial smooth muscle cells [30] and migration of pulmonary arterial endothelial cells [31]. Taken together, we have shown that IgE served as an activator of MCs and the subsequent release of cytokines, in particular IL6/IL13, promoted pulmonary vascular remodeling.
Limitations of our study include the species difference in the expression of FcεRIα. In addition to MCs, FcεRIα is also expressed in dendritic cells in human (Fig.4B), whereas its expression is absent in murine counterparts [32]. In ltration of dendritic cells was observed in human PAH lung samples [33]. Previous research has reported con icting roles of FcεRIα expressed in dendritic cells, due to the different research strategies used [34,35]. The observations in humanized mice [35] suggested that IgE activation mediated antigen presentation by dendritic cells. In the pathogenesis of asthma in human [22], IgE activated dendritic cells promoted Th2 cell recruitment and reactivation. Based on these reports and the expression of FcεRIα in human dendritic cells, we speculate that IgE may act on dendritic cells to promote the in ltration of Th2 cells, leading to an uncontrolled Th2 response in PH. Of note, stabilizing MCs by cromolyn is effective in treating PH in rats, but less effective in human [36]. This might be explained by the in ltrated dendritic cells in human lung, which are also activated by IgE but are unresponsive to cromolyn. From a translational perspective, blocking the upstream mediator IgE, instead of a single type of downstream effector cells (MCs or dendritic cells) and factors (IL6 or IL13), may confer better therapeutic e cacy. In order for this concept to be applied in the clinic, further pre-clinical and clinical studies are required, to compare the e cacy of Omalizumab, against and/or in combination with currently available drugs for PAH, in order to determine the most suitable IgE blocker for single or combination therapy. The wealth of data and experiences associated with the use of Omalizumab in other diseases may facilitate its repurposing for the treatment of PAH.
In summary, our study demonstrated that elevated IgE promoted the development of PH, whilst inhibition of IgE attenuated PH in multiple animal models. This pathogenic axis was shown to be activated in PAH patients. We have found that the IgE antagonist Omalizumab ameliorated established PH, providing proof-of-concept supporting this novel therapeutic strategy, especially for PAH patients with high levels of of IgE+ cells in CD19+ cells (n=6). For A, H-J, data are shown as mean ± SEM and the differences between groups were evaluated by unpaired two-tailed t-test. For B and C, data are shown as mean ± SEM and the differences between groups were assessed by one-way ANOVA with LSD-t test. *p<0.05, **p<0.01, ***p<0.001. Proportion of non-muscularized (N), partially muscularized (P), or full muscularized (F) pulmonary vessels of 20-100μm in diameter from indicated rats. All above results of quantitative analysis are shown as mean ± SEM, and differences between multiple groups were evaluated by two-way ANOVA with Bonferroni's post hoc test. *p<0.05, **p<0.01, and ***p<0.001 for anti-IgE versus isotype control-treated animals with hypoxia-or MCT-induced PH, ###p<0.001 for isotype control-treated animals hypoxia vs. normoxia or Saline vs. MCT groups.     IgE stimulated MCs to produce IL6 and IL13 in experimental PH. A. Volcano plot of differentially expressed genes between BMMCs treated with and without IgE, with two samples for each group; P value was adjusted by false discovery rate (FDR). B. Bubble plot of GO pathway enrichment of the differentially expressed genes in response to IgE. C. Dot plot of gene expression in cell clusters. D. Relative mRNA expression of Il6, 1l13 and Ccl4 in sorted CD45+ c-Kit+ FcεRIα+ cells from lung tissues of mice under normoxia and hypoxia. E. Relative mRNA expression of Il6 and Il13 in lung tissues from mice under normoxia and hypoxia. F. Relative mRNA expression of Il6 and Il13 in lung tissues from hypoxic mice injected with IgE-neutralizing antibody or isotype control. G. Relative mRNA expression of Il6 and Il13 in lung tissues from WT or KO mice after hypoxia exposure. H. Relative mRNA expression of Il6 and Il13 in lung tissues from WT or MCKO mice after hypoxia exposure. I. Schematic diagram of the MC and PASMC co-culture experiment. J. Quanti cation of cell number of PASMCs co-cultured with supernatant from IgEtreated MCs transfected with or without shIL6 and shIL13 measured by CCK8 assay. K. Representative western blots and quanti cation of PCNA (proliferating cell nuclear antigen) in the PASMCs. All values are presented as the mean ± SEM. For D, n=6; E, n=4; F-H, n=6 for each group, and differences were evaluated by unpaired two-tailed t-test. For J and K, n=3 for each group, and differences were evaluated by one-way ANOVA with LSD-t test. *p< 0.05, **p< 0.01, and ***p< 0.001.