Leptin Promotes Allergic Airway Inflammation through Targeting the Unfolded Protein Response Pathway

Allergic asthma and obesity are major public health problems in the world. Recent Meta-analysis studies implicated a positive relationship between serum leptin, which is elevated in obese individuals, and the risk of asthma. However, it is not well understood how obesity-associated elevation of leptin increases the risk of asthma. In the current study, we have found that leptin induces the unfolded protein response factor XBP1s in an mTOR- and MAPK-dependent manner in pro-allergic TH2 cells; in vivo, mice fed with high fat diet had increased serum leptin as observed in human obese population and exacerbated asthmatic symptoms, associated with increased XBP1s expression in splenic CD4+ T cells. XBP1s is required for leptin-mediated pro-allergic TH2 cell survival and cytokine production. Our results reveal a previously unappreciated insight that obesity-associated hyperleptinemia contributes to enhanced pro-allergic lymphocyte responses through induction of XBP1s, leading to exacerbation of allergic asthma.

Besides asthma, obesity is another leading health problem worldwide and meta-analysis studies manifested that obesity is a major risk factor for development of asthma [22][23][24] . However, the underlying mechanism whereby obesity increases the risk of asthma has not been well established. Leptin, a classical pro-inflammatory adipokine mainly derived from adipocytes, displays significantly higher amounts in serum and visceral adipose tissue from the obese population than the non-obese population 25,26 , leading to defects in modulating balance between food intake and energy expenditure followed by multiple autonomous complications. Leptin resistance has been described to enhance parasympathetic tone, which leads to bronchoconstriction and obesity-associated asthma 27 . Dysregulation of leptin expression by adipose tissue has been shown to influence lung physiology and mechanics, and associates with asthma development 28 . In addition, hyperleptinemia was proposed to be a crucial factor leading to respiratory failure in leptin-resistant obese subjects 29 . Meta-analysis studies have further implicated a positive relationship between serum leptin and the risk of asthma [30][31][32] . Thus, leptin emerges as a potential mediator driving allergic asthma. However, how leptin participates in the interconnection between adipose tissue and airway inflammation and how leptin-involved defective metabolism influences lung physiology are not well understood.
Our previous studies have shown a pathogenic role of leptin in enhancing allergic responses 33 . We found leptin deficiency leads to attenuated asthma symptoms with decreased eosinophilia, and type 2 cytokine production. Leptin targets the mTOR, MAPK and STAT3 pathways in TH2 cells 33 . However, it is not fully understood how these signals lead to enhanced type 2 cytokine production. In this study, we found that leptin induced XBP1s [spliced form of X-box binding protein 1 (XBP1)], an endoplasmic reticulum (ER) stress-unfolded protein response (UPR) factor important in secretion function of both T and B cells for secretion of cytokines and antibodies, respectively [34][35][36] . We found XBP1s contributed to leptin functions not only in cytokine production but also in cell survival in pro-allergic TH2 cells. In vivo, splenic CD4 + T cells from high fat diet-fed mice (termed HFD mice) with plethoric leptin and exacerbated asthma symptoms, close to obese human subjects, express increased amounts of XBP1s relative to normal chow diet-fed group (termed ND mice). Taken together, our data suggest that the leptin-XBP1s axis upregulates the responsiveness of pro-allergic TH2 cells. These findings provide a novel insight into the potential of obesity-associated elevation of leptin leading to the increased risk of allergic asthma.

Results
Leptin induces an endoplasmic reticulum (ER) stress-UPR factor, XBP1s. We have previously found that leptin deficiency impairs type 2 immune responses and attenuates allergic airway inflammation through attenuation of pro-allergic TH2 cells and ILC2s 33 . ER stress-UPR play a pivotal role in T and B cell secretion function. To further understand how leptin-mediated signals regulate pro-allergic cytokine expression, we examined expression of the ER stress-UPR factor XBP1s in in vitro polarized TH2 cells in the presence or absence of leptin, and found that treatment with leptin induced XBP1s expression compared with no treatment (Fig. 1A). To test the effect of leptin on XBP1s expression in vivo, we generated HFD mice (body weight, 44.02 ± 2.18 g), which expressed increased levels of serum leptin compared with lean mice fed with ND mice (33.92 ± 2.77 g) (Fig. 1B), replicating human studies 37 . We asked whether HFD affected leptin receptor (ObR) expression on various types of immune cells. By using flow cytometry, we found that TH2 cells, ILC2s and TH1 cells from lung draining lymph nodes (LLNs) of HFD and ND mice all expressed comparable levels of ObR (Fig. 1C), indicating that the regulation of ObR is independent of diet or serum leptin levels. Next we examined XBP1s expression in splenic CD4 + T cells isolated from the HFD versus ND mice and found that HFD CD4 + T cells expressed elevated levels of XBP1s protein compared with ND cells (Fig. 1D). These results indicate that leptin induces XBP1s expression, which may contribute to leptin-induced hyper-responsiveness of pro-allergic TH2 cells.
HFD-associated elevation of leptin correlated with aggravated allergic airway inflammation. We have observed above increases in leptin expression in HFD mice compared with their ND controls, and during allergic asthma, HFD T cells expressed higher amounts of XBP1s. We then tested allergic inflammation in the HFD and ND asthmatic mice as previously described 33,38 . We found that during allergic responses, HFD mice had increased levels of leptin in serum relative to ND mice ( Fig. 2A) and elevated expression of XBP1s in splenic CD4 + T cells (Fig. 2B), similar as the steady state (Fig. 1B,D). HFD mice displayed increased infiltrates of eosinophils, lymphocytes and neutrophils in bronchoalveolar lavage fluids (BALFs) compared with ND mice (Fig. 2C). Whereas, PBS challenged HFD and ND mice displayed basal level of BALF immune cell infiltration (Fig. 2D). To evaluate lung inflammation, we stained the lung sections from HFD and ND mice with hematoxylin and eosin. In compliance with BALFs immune cell infiltration, we observed that HFD lungs exhibited increased mononuclear cell infiltration in the peribronchovascular spaces that contained increased numbers of eosinophils ( Fig. 2E). In addition, papain and Ova challenge elicited IgE responses in both HFD and ND mice; Sera and BALFs of HFD mice contained increased levels of Ova-specific IgE compared with ND mice (Fig. 2F, left). IgE expression in PBS challenged HFD and ND mice were only at basal levels ( Fig. 2F, right).
Type 2 immune responses manifest one of the essential hallmarks of allergic asthma. Therefore, we profiled type 2 lymphoid TH2 cells and ILC2s, as well as TH1 cells in asthmatic HFD mice relative to their ND controls, and found that lung associated draining lymph nodes (LLNs) of HFD mice contained increased frequencies and numbers of TH2 cells and ILC2s but the frequencies and numbers of TH1 cells are comparable with ND mice ( Fig. 3A-C). In contrast with asthmatic groups, PBS challenged HFD and ND mice had only a few TH2 cells and ILC2s in LLNs (Fig. 3D); interestingly, both PBS challenged HFD and ND mice contained slightly fewer TH1 cells in LLNs compared with the asthmatic groups (Fig. 3D). Upon ex vivo recall with Ova, HFD LLN cells expressed higher amounts of TH2 cytokines, IL-4, IL-5 and IL-13 and comparable amounts of TH1 cytokine IFN-γ relative to the ND group (Fig. 3E), whereas LLN cells from PBS-challenged HFD and ND mice only expressed basal amounts of these cytokines (data not shown).
Above results showed that HFD mice contained increased TH2 and ILC2 frequencies and expressed more Ova-specific type 2 cytokines versus ND mice. We next asked whether HFD promotes proliferation of these immune cell populations during allergic responses. To address this, we measured the expression of Ki67, a cell proliferation-associated nucleic protein that marks cell at active phases (G1, S, G2 and M) but not the resting phase (G0), in TH2 cells, ILC2s and TH1 cells from the asthmatic mice. We observed that the frequencies of Ki67 + cells were greater in HFD TH2 cells and ILC2s than corresponding ND cells. In contrast, there was no apparent distinction of Ki67 + TH1 cells between HFD and ND mice, which was due to the dominant type 2 immune environment eliciting by a protease type adjuvant papain (Fig. 3F). Taken together, these findings suggest that HFD induces elevated expression of leptin that induces the UPR factor XBP1s and likely renders TH2 cells and ILC2s but not TH1 cells more responsive to antigen stimulation during a type 2 challenge, which subsequently activate and recruit more immune cells, including eosinophils, to the airway and lung, leading to exacerbation of allergic airway inflammation, consistent with our previous observations on leptin deficient mice 33 .
Leptin-XBP1 axis is required for TH2 cell cytokine expression. Since leptin induces XBP1s expression in TH2 cells, we asked whether leptin regulates TH2 cell cytokine secretion via the XBP1s pathway. To explore this, we performed siRNA-mediated Xbp1 gene silencing in differentiated TH2 cells. In vitro differentiated TH2 cells were transfected with siXbp1 or scramble siRNA with or without addition of leptin. We found that expression of XBP1s protein and mRNA was significantly upregulated by leptin treatment and leptin-induced XBP1 expression was downregulated by siXbp1 relative to scramble siRNA treatment (Fig. 4A,B), indicating a successful Xbp1 gene silencing. In the absence of leptin, siXbp1 led to downregulation of XBP1s protein and a strong trend of decreasing its mRNA (Fig. 4A,B). We next examined whether Xbp1 gene silencing affects the induction of cytokine expression in TH2 cells by leptin treatment and found that leptin-induced elevations of TH2 type cytokines (IL-4, IL-5 and IL-13 at both mRNA and protein levels) were reversed by Xbp1 gene silencing (Fig. 4B,C), whereas neither leptin treatment nor Xbp1 knockdown altered the expression of Gata3 mRNA (encoding GATA3, the master transcription factor of TH2 cells). These data indicate that the leptin-XBP1s axis is required for TH2 cell cytokine expressions.

Leptin-XBP1s pathway protects TH2 cells from activation induced cell death. Although XBP1s
is known to enhance cell secretion through regulation of ER function 35,36 , it is not clear whether XBP1s also contributes to the other effects of leptin. We assessed the effect of XBP1 on proliferation of siXbp1 or scramble siRNA transfected TH2 cells for 6 h or overnight culture by CFSE dilution. We found that leptin increased TH2 cell proliferation after overnight but not 6-h culture; addition of Xbp1 gene silencing did not alter the effect of leptin on proliferation (Fig. 5A). In addition to proliferation, we measured activation induced cell death in in vitro differentiated TH2 cells by LIVE/DEAD Green stain and found that leptin-mediated protection on cell death was abolished by Xbp1 gene silencing, indicating an essential role of XBP1 in controlling cell survival (Fig. 5B). Therefore, XBP1s is required for leptin mediated cell survival but not proliferation of pro-allergic TH2 cells.

Leptin induces XBP1s expression in a MEK-and mTOR-dependent manner. Our above results
indicate leptin functions through induction of XBP1s expression (Figs 1A, 4 and 5B). Leptin is known to activate the mTOR and MAPK pathways in TH2 cells 33 . We next asked whether these leptin signals are able to induce XBP1s expression. We found that leptin-induced XBP1s expression could be blocked by addition of either mTOR inhibitor rapamycin or MEK inhibitor PD98059 (Fig. 6A), both of which can block induction of TH2 cell cytokine expression by leptin 33 . XBP1s is known to transactivate genes encoding factors promoting ER function and autophagy 39,40 . We therefore assessed whether leptin signaling affects the expression of XBP1s downstream factors and found that leptin induces Hspa5 (encoding UPR factor BiP), Ddit3 (encoding UPR factor CHOP) and Becn1 (encoding autophagy factor Beclin1), whereas treatment with both mTOR inhibitor rapamycin and MEK inhibitor PD98059 greatly diminished the effects of leptin on induction of these UPR and autophagy factors (Fig. 6B). Thus, leptin regulates XBP1s expression through activation of the mTOR and MEK signal cascades in TH2 cells and induces UPR and autophagy factors which is likely through the induction of XBP1s.

Leptin induces XBP1s expression through activation of IRE1 but not ATF6. After activation, TH2
cells are known to mount massive protein synthesis leads to ER stress-associated UPR, in which activation of IRE1 and/or ATF6 results in increased expression of Xbp1s mRNA and XBP1s protein 35 . To understand how leptin signaling induces XBP1s expression, we examined IRE1 and ATF6 activation by western blot. We found that addition of leptin induces IRE1 phosphorylation (Fig. 7A), whereas it did not alter ATF6 cleavage (Fig. 7B), suggesting that leptin induces XPB1s expression through activation of IRE1 rather than ATF6. We next assessed whether the mTOR and MEK pathways downstream of leptin activate IRE1 and found both mTOR inhibitor rapamycin and MEK inhibitor PD98059 could block leptin induced phosphorylation of IRE1 (Fig. 7C). Therefore, leptin activates the mTOR and MEK pathways that subsequently activate IRE1, leading to increasing XBP1s expression and finally enhancing pro-allergic TH2 cell function.
Taken together, our results demonstrate that obesity-associated elevation of leptin may result in activation of the mTOR/MEK (upstream of MAPK)-IRE1-XBP1 axis in pro-allergic lymphocytes that promote cell survival and cytokine production, therefore exacerbating allergic airway disease (outlined in Fig. 7D).

Discussion
Obesity was identified as a major risk factor in development of allergic asthma by meta-analysis [22][23][24] . Many studies have shown that leptin, an adipokine highly elevated in the obese population, represents a paramount role on affecting lung physiology and mechanics and thus, regulating respiratory function, which is correlated with obesity-associated asthma [27][28][29] . Furthermore, a positive relationship between serum leptin and risk of asthma has been revealed by meta-analysis studies [30][31][32] . However, as a potential mediator driving allergic asthma, leptin is still in lack of being understood on how it participates in defective metabolism-involved lung pathology. Several studies showed that leptin promotes TH1 responses both in vivo and in vitro [41][42][43] , whereas, whether leptin regulates pro-allergic type 2 responses is not yet elucidated. Batra et al. showed leptin promoted TH2 cell development and aggravated TH2-mediated colitis, however, controversial results in vitro showing high concentration of leptin (1 μg ml −1 ) reduced TH2 cell frequency in repeated polarization cultures have been described in the same study 43 . This might be explained by over-dose leptin raised toxicity that could overwhelm its physiological effects. In a human childhood study, the obese asthmatic group had increased plasma IL-4 and IFN-γ, correalted with higher plasma leptin compared with control gorup 42 . ILC2s, another essential component in type 2 immune responses and allergic asthma, produce type 2 cytokines 12-14 , promote TH2 cell differentiation and memory responses in allergic asthma [18][19][20][21] . Furthermore, activated TH2 cells produce IL-2 that promotes ILC2s development. This reciprocal regulation between TH2 cells and ILC2s suggests a feed-forward loop in allergic responses. Our current studies showed that HFD mice with plethoric leptin had increased TH2 and ILC2 proliferation and type 2 cytokine production compared with ND group, which contributes to the exacerbated asthma symptoms.
Leptin modulates immune cell function through activating multiple downstream signaling pathways including JAK2-STAT3, MAPK and PI3K-AKT 44,45 . For instance, leptin enhances TH1/TH17 cell survival through activating ERK1/2 and AKT-mTOR pathways 46 . In consistency with these studies, we have shown that in TH2 cells, leptin also activates STAT3, MEK-MAPK and AKT-mTOR signaling pathways, which contribute to leptin-mediated cell proliferation, survival and cytokine production 33 . It has not been well understood how these leptin signals regulate type 2 cell cytokine production. Upon extracellular stimulation, lymphocytes produce massive amounts of effector cytokines. Accumulation of unfolded (or misfolded) proteins leads to ER stress that activates the UPR pathways 35 . During UPR, ATF6 is cleaved into an active ATF6 fragment (ATF6f) transcription factor, and induces transcription of several genes, including Xbp1; subsequently, endoribonuclease IRE1 undergoes phosphorylation and phosphorylated IRE1 excises a 26-nucleotide fragment from unspliced Xbp1 (Xbp1u) mRNA and forms spliced Xbp1 (Xbp1s) mRNA. XBP1s protein transactivates transcription of many genes that are crucial for secretory function through increasing ER capacity and promoting autophagy 40,47 . These pathways together allow a cell to resolve the endogenous stress of unfolded proteins and maintain intracellular homeostasis. Therefore, XBP1s plays a central role in UPR and cell secretion function. For example, in human necrotizing enterocolitis, XBP1 splicing levels correlate with the severity of mucosal damage that is associated with increased mucosal expression of pro-inflammatory cytokines, IL-6 and IL-8 48 . In our study, we have observed that leptin induces XBP1s expression in TH2 cells and after knock-down of Xbp1, the effect of leptin on TH2 cell cytokine production is diminished (Fig. 4B,C). Previous studies have shown that Xbp1 mRNA splicing can be induced by MAPK signaling in liver cells or PI3K-AKT-mTOR signaling in innate immune cells 49,50 . We have also found that leptin induces XBP1s expression dependent of both MEK-MAPK and mTOR signaling pathways, which leads to activation of IRE1 but not ATF6 (Fig. 7), and XBP1s is required for the anti-apoptotic effects of leptin in TH2 cells (Fig. 5B). In summary, our results suggest leptin as a key risk factor in the development of allergic asthma in obese subjects through induction of the UPR factor XBP1s that promotes survival of pro-allergic lymphocytes and their cytokine expression. These findings may suggest a novel therapeutic approach for treatment of obesity associated allergic asthma.

Materials and Methods
Animals. Six-week-old C57BL/6 mice were fed with either a normal chow diet (ND) or HFD (45 kcal% fat, D12451; Research Diets Inc., New Brunswick, NJ, USA) for 15 weeks. All mice were housed in the specific pathogen-free animal facility at the University of New Mexico Health Sciences Center. All animal experiments were performed with protocols approved by the Institutional Animal Care and Use Committee of the University of New Mexico Health Sciences Center. All methods were performed in accordance with the relevant guidelines and regulations.
Induction of allergic asthma. Age and sex-matched C57BL/6 HFD and ND mice were immunized intranasally with 25 µg papain and 50 µg chicken Ovalbumin (Ova) (or treated with PBS as no asthma controls) for three times on D0, D1 and D14. On D15, Sera, BALFs, LLNs, and left upper lung lobes were collected for analysis of infiltrates and immune responses as described before 33,38 . In vitro TH2 cells differentiation. CD4 + CD25 − CD62L + naïve T cells were sorted from C57BL/6 WT mice and differentiated in a TH2-polarizing condition (5 μg ml −1 anti-IFN-γ and 10 ng ml −1 IL-4) using plate-bound α-CD3/α-CD28 and low serum (3-5% FBS)-containing RPMI medium. Afterwards, the resulting cells were re-stimulated in serum free medium (OpTmizer TM CTS TM T-Cell Expansion SFM, Life Technologies) for intracellular cytokine expression, apoptosis and proliferation assay, and siRNA silencing assay in the presence or absence of leptin (200 ng ml −1 ) as indicated.
Gene silencing. In vitro differentiated WT TH2 cells were starved 24 h in serum free medium on D4. On D5, the TH2 cells were transfected with Xbp1 or scramble siRNA (Santa Cruz Biotechnology, Inc.) and incubated on an anti-CD3 coated plate for 6 h, and afterwards the siRNA transfected TH2 cells were subjected to different treatments as indicated, and were used for cytokine expression, proliferation, cell death and immunoblot assays.
ELISA. Mouse Leptin ELISA kit (#90030, Crystal Chem) was utilized for leptin measurement following the manufacturer's instruction. To measure Ova-specific IgE, plate-bound Ova (100 μg ml −1 ) were used as capture and anti-mouse IgE (23G3, eBioscience) as detection antibody. LLN cells (4 × 10 6 cells ml −1 ) from the asthmatic HFD and ND mice were recalled with various concentrations of Ova for 3 days and the supernatants were collected for measurement of cytokine expression by ELISA using a standard protocol. For in vitro differentiated TH2 cells, the cells were starved for 24 h and transfected with siXbp1 or scramble siRNA. The resulting cells were washed and treated with or without leptin (200 ng ml −1 ) for 6 h, and finally the supernatants were collected and used for measuring cytokines expression by ELISA.
Proliferation assay. In vitro TH2 cell proliferation was assessed by carboxyfluorescein succinimidyl ester (CFSE, C34570, ThermoFisher Scientific) dilution. 1 d after siXbp1 or scramble siRNA transfection, TH2 cells were labeled with CFSE and re-stimulated with plate-bound anti-CD3 in serum free medium with or without leptin (200 ng ml −1 ) for 6 h or overnight incubation. For in-vivo proliferation, single-cell suspensions of LLNs from