**4. Discussion**

Human studies as well as in vivo animal models have reported increased activation of the classical and alternative NF-κB pathways in asthmatic airway tissues and in inflammatory cells [3,33]. As such, NF-κB signaling intermediates are attractive therapeutic targets for airway diseases such as asthma, as the underlying inflammation is independent of stimuli [34] and is mediated at least in part by NF-κB mediated signaling in bronchial fibroblasts (Figure 1). Bcl10 being a critical mediator of NF-κB signaling prompted us to explore its role in fibrotic remodeling in bronchial fibroblasts from severe asthma. To the best of our knowledge, this is the first report providing evidence of elevated protein expression of Bcl10 in the pathogenesis of severe asthma as well as the role of Bcl10-mediated signaling in the LPS-induced pro-fibrotic cytokine expression in bronchial fibroblasts.

NF-κB is a key component of the inflammatory network that controls cytokine production in airway pathology [20]. Overexpression of Bcl10 is an indicator of constitutive

NF-κB activation in tumors including MALT lymphoma [8,35], and persistent NF-κB activation is known to characterize severe uncontrolled asthma [3]. At baseline, the S-As fibroblasts demonstrated differential gene expression of various intermediates of the NF-κB pathway when compared to their healthy counterparts, supporting the notion of activation of NF-κB in severe asthma (Figure 1A). This was further confirmed by the increased protein expression of Bcl10 at basal levels in S-As fibroblasts (Figure 1B). We have also previously reported the increased expression of pro-fibrotic and pro-inflammatory mediators associated with the NF-κB activation, such as IL-6, IL-8, IL-11 and GROα (CXCL1) in these S-As fibroblasts [36]. Thus, the increased Bcl10 expression in S-As fibroblasts appears to signify the activation of the Bcl10-mediated NF-κB pathway in severe asthma. The observation that complete Bcl10 deficiency severely impaired fibroblast function in an immunodeficient individual [37] is testament to its importance in fibroblast response. Numerous lymphoid malignancies are characterized by the constitutive aberrant activation of the NF-κB proinflammatory pathway. Here, we demonstrated a similar pattern of Bcl10-mediated NF-κB activation in airway structural fibroblasts.

Furthermore, Bcl10 was differentially expressed in the sub-epithelial fibroblasts among the varying severities of asthma, ranging from weak expression in control biopsies to moderate expression in mild-to-moderate asthma and strong expression in severe asthma (Figure 2A–F). Just as in the case of MALT lymphoma, high Bcl10 expression in fibrotic airway tissues is paradoxical, considering that Bcl10 is a pro-apoptotic CARD-containing adaptor molecule [38]. However, certain cellular contexts in vivo may influence Bcl10 to behave as an anti-apoptotic molecule. For instance, overexpression of Bcl10 conferred a survival advantage to activated primary B cells even after withdrawal of the activating stimuli [39]. Alternately, the subcellular localization of Bcl10 may be a pre-determining factor in explaining this paradox. Bcl10 was predominantly expressed in the cytoplasm of subepithelial fibroblasts irrespective of the disease severity (Figure 2). However, Bcl10 nuclear expression was detected in the bronchial epithelium of non-asthmatic and asthmatic individuals (Figure 2). The different subcellular localization pattern of Bcl10 between the bronchial fibroblasts and epithelial cells (Figure 3A,B), indicates cell-dependent functional role of Bcl10. Here, it is interesting to note that the NF-κB-independent functions of Bcl10 include actin remodeling. For instance, Bcl10 regulates TCR-induced actin polymerization and cell spreading in T cells, and FcγR-induced actin polymerization and phagocytosis in monocytes/macrophages [40]. Subsequently, the role of Bcl10 in actin dynamics, cytoskeletal and membrane remodeling in macrophages was found to entail phagosome formation [41]. This may perhaps explain the cytoplasmic expression of Bcl10 in S-As fibroblasts indicating Bcl10-dependent actin polymerization in addition to its role in NF-κB activation. Actin dynamics and polymerization are key to the contractile property of fibroblasts [42]. Thus, the presence of Bcl10 in the cytoplasmic compartment of S-As fibroblasts may signify its role in regulating actin dynamics and contraction of bronchial fibroblasts, and by this means, contributing to airway hyperresponsiveness in severe asthma. However, further studies are required to completely understand the physiological relevance of this subcellular localization.

Toll-like receptors (TLRs) such as TLR4 and TLR2 are important for the adaptive Th2 cytokine-driven inflammatory response in asthma [43,44]. Engagement of the TLR initiates the recruitment and activation of several adaptor molecules resulting in the activation of multiple signaling cascades, including NF-κB. TLR activation of the NF-κB pathway regulates the expression of immunomodulatory and inflammatory mediators. For instance, LPS binding to the TLR4 initiates an inflammatory cascade that climaxes in the release of cytokines such as IL-8 and IL-6, and the recruitment of an inflammatory infiltrate of lymphocytes, macrophages and polymorphonuclear leukocytes [45,46]. However, the signal transduction of the TLR4-Bcl10-NF-κB axis in asthmatic fibroblasts is far from understood, and the role of Bcl10 in TLR4-mediated fibroblast function is largely unknown.

Since the basal expression of TLR4, CARMA3, BCL10 and MALT1 was elevated in DHBF when compared to NHBF (Figure 1), we speculated that the TLR4-BCL10-NF-κB axis responds to LPS stimulation in bronchial fibroblasts and stays upregulated in S-As fibroblasts contributing to the activation of NF-κB in these cells. Here, we show that Bcl10 is a mediator of LPS-induced increase in pro-fibrotic cytokine expression (Figure 4). While LPS stimulation for 2 h boosted the expression of Bcl10, IL-6, IL-8 and TGF-β1 in bronchial fibroblasts, IRAK1/4 inhibition reversed the LPS-induced increase in all components, indicating increased Bcl10-mediated signaling contributed to pro-fibrotic cytokine expression upon LPS exposure in bronchial fibroblasts. Although LPS stimulation was previously reported to induce collagen expression in lung fibroblasts [28], there was no increase in COL1A1, COL5A1 and FN1 expression in NHBF and DHBF with LPS stimulation (Figure 4D). This could be attributed to the short exposure to LPS. While 2 h of LPS stimulation was sufficient to induce pro-fibrotic cytokine signaling in these fibroblasts, longer exposure may pave way to increased ECM secretion via activation of phosphoinositide3-kinase-Akt (PI3K-Akt) pathway [28] as well as autocrine signaling from IL-6 and TGF-β1 [47,48].

Since TLR4 signaling activates the canonical NF-κB pathway via the MyD88-dependent as well as the TRIF-dependent pathways [30], we next aimed to verify which of these pathways contributed to pro-fibrotic signaling. While IRAK1/4 inhibition had no effect on the TRIF-dependent pathway, the LPS-induced increase in Bcl10 was abolished and this was accompanied by a corresponding decrease in IL-6, IL-8 and TGF-β1 expression (Figure 5). This indicates Bcl10 signaling directly to its downstream mediators, including IL-6, IL-8 and TGF-β1 in response to LPS in bronchial fibroblasts. Furthermore, since Bcl10 inhibition was not accompanied by alterations in the TRIF-dependent induction of type I interferons, anti-viral signaling and immune defense against viral infections will be sustained in these cells. Considering the activated status of NF-κB pathway in S-As fibroblasts, an exaggerated response to LPS may cause greater extent of inflammatory damage and remodeling changes in severe asthmatic airways.

Increased levels of IL-6 were detected in the sputum of asthmatic subjects when compared to healthy controls, which correlated with impaired lung function in allergic asthma [49]. Further, high circulating levels of IL-6 increased the risk of exacerbations by 10% for each 1-pg/μL increase in baseline IL-6 level [50]. IL-6 is a pleiotropic cytokine with both pro-inflammatory and pro-fibrotic functions [51]. Severe asthma is also characterized by elevated IL-8 levels and associated neutrophilia [52,53]. We have previously reported the upregulation of IL-6 and IL-8 expression at basal levels in S-As fibroblasts when compared to non-asthmatic fibroblasts [36]. The ability of these fibroblasts to produce increased levels of ECM proteins as well as pro-inflammatory and pro-fibrotic cytokines have implicated them in remodeling and inflammation, two key processes involved in the pathogenesis of asthma. IL-8 is known to contribute to the pathogenesis of severe asthma by facilitating various features of airway remodeling, including neutrophil recruitment, epithelial-tomesenchymal transition [54], angiogenesis [55] and proliferation and migration of ASM cells [56]. TGF-β is the central mediator of fibrotic tissue remodeling in asthma [57]. TGF-β gene polymorphisms were recently reported as a risk factor for asthma control [58]. The persistently high levels of TGF-β in severe asthma may contribute to increased collagen secretion from severe asthmatic fibroblasts despite treatment with oral corticosteroids [59]. Here, we showed that Bcl10 mediated the LPS-induced expression of IL-6, IL-8 and TGF-β1 in bronchial fibroblasts, highlighting the pathogenic role of Bcl10-mediated signaling in promoting airway remodeling in severe asthma.

Although extensively studied in immune cells, Bcl10-mediated NF-κB activation is emerging as an important pathway in non-immune cells as well. For instance, angiotensin II promotes liver fibrosis by activating the CBM-3-dependent NF-κB pathway in hepatocytes [16]. Lysophosphatidic acid-induced NF-κB activation and IL-6 production in murine embryonic fibroblasts involves signaling through adapter proteins Bcl10 and Malt1 [60]. In one study, Bcl10 was found to be an essential component of TLR4 response in human primary fibroblasts [37] and Bcl10 deficiency was found to abolish TLR4 signaling in response

to LPS stimuli and subsequent production of IL-6 and IL-8. Our results are consistent with the observations made in this study.

We identified the Bcl10-mediated NF-κB pathway as a mechanism contributing to fibrotic remodeling and inflammation in severe asthma (Figure 6). Further studies are, however, essential to delineate the molecular interactions of Bcl10 to develop a more complete understanding to explain the signal transduction in bronchial fibroblasts. Another interesting option is to explore the kinetics of CBM-3 formation in S-As fibroblasts taking into account the activation of the Bcl10-mediated NF-κB pathway in these fibroblasts.

**Figure 6.** Induction of Bcl10-mediated pro-fibrotic cytokine signaling by LPS in bronchial fibroblasts. (**A**) In bronchial fibroblasts, LPS stimulation of TLR4 receptor activates both the MyD88-dependent and TRIF-dependent pathways. Bcl10-mediated NF-κB activation has downstream effects on airway inflammation and remodeling through the secretion of IL-6, IL-8 and TGF-β1 cytokines. (**B**) Blockade of this Bcl10 signaling cascade may ameliorate both inflammation and fibrosis by impeding the expression of IL-6, IL-8 and TGF-β1 cytokines.

Some of the limitations of our study include the lack of specific Bcl10 inhibition that would demonstrate the direct causal relationship between BCL10 and NF-κB pathway activation and could be addressed by using BCL10 siRNA/CRISPR and NF-κB dual luciferase reporter assay. However, we were able to show that an IRAK1/4 selective inhibitor that inhibits the signaling upstream of Bcl10 abrogated the pro-fibrotic signaling induced by LPS. Since the S-As fibroblasts were derived from patients on medications such as glucocorticosteroids and biologics, the results need to be interpreted with caution, as these medications are known to alter inflammatory cell signaling events in asthmatic airways [61]. Nevertheless, the enhanced signature of key NF-κB genes at baseline in the S-As fibroblasts suggest refractoriness to steroids in these patients. Another shortcoming of our study is the lack of in vivo investigation that we aim to explore in future studies.
