Increased levels of Gab1 and Gab2 adaptor proteins skew interleukin-4 (IL-4) signaling toward M2 macrophage-driven pulmonary fibrosis in mice

M2-polarized macrophages, also known as alternatively activated macrophages, have long been associated with pulmonary fibrosis; however, the mechanism has not been fully defined. Gab1 and Gab2 proteins belong to the Gab family of adaptors and are integral components of the signal specificity in response to various extracellular stimuli. In this report, we found that levels of both Gab1 and Gab2 were elevated in M2-polarized macrophages isolated from bleomycin-induced fibrotic lungs. In vitro Gab1/2 deficiency in bone marrow-derived macrophages abrogated IL-4–mediated M2 polarization. Furthermore, in vivo conditional removal of Gab1 (Gab1MyKO) and germ line knock-out of Gab2 (Gab2−/−) in macrophages prevented a bias toward the M2 phenotype and attenuated bleomycin-induced fibrotic lung remodeling. In support of these observations, Gab1/2 were involved in responses predominated by IL-4 signaling, an essential determinant for macrophage M2 polarization. Further investigation revealed that both Gab1 and -2 are recruited to the IL-4 receptor, synergistically enhancing downstream signal amplification but conferring IL-4 signal preference. Mechanistically, the loss of Gab1 attenuated AKT activation, whereas the absence of Gab2 suppressed STAT6 activation in response to IL-4 stimulation, both of which are commonly attributed to M2-driven pulmonary fibrosis in mice. Taken together, these observations define a non-redundant role of Gab docking proteins in M2 polarization, adding critical insights into the pathogenesis of idiopathic pulmonary fibrosis.

Idiopathic pulmonary fibrosis (IPF) 3 is a chronic interstitial disease of unknown etiology that is characterized by alveolar inflammation, a Th2-cell cytokine milieu, and excess collagen deposition, leading to respiratory failure and death (1)(2)(3). However, the mechanisms underlying this devastating progression remain largely elusive. Macrophages are the predominant innate immune cells in the lung, serving as important regulators of lung injury and repair (4). In response to injury, macrophages demonstrate remarkable plasticity and are capable of acquiring alternatively activated polarization (M2), which is characterized by producing matrix metalloproteases and secreting chemokines, regulating Th2 and fibrotic progression (4 -6). It has been demonstrated that M2 markers associate with pulmonary fibrosis in a mouse model, suggesting an important role for M2 macrophages in the development of pulmonary fibrosis (4,(7)(8)(9). The M2 marker Arg1 (arginase 1) regulates the production of proline, which is required for collagen synthesis, yet recent studies indicate that Arg1 is not essential for regulation of the proline-collagen pathway (10). Additional bleomycin (BLM)-induced mouse experiments indicate that FIZZ1 (found in inflammatory zone 1) promotes the differentiation and survival of myofibroblasts during fibrogenesis. FIZZ1-deficient mice exhibit ameliorated fibrosis in lung tissue (11,12). Recent clinical findings indicate that the expression of Ym1 (chitinase-3-like protein 3) and CD206 (mannose receptor) are also elevated in alveolar macrophages isolated from patients with IPF (13,14). Taken together, these findings highlight the important role of M2 macrophages in the development of pulmonary fibrosis.
M2 polarization is driven by the products of Th2 cells, among which IL-4 is well demonstrated to be an important determinant controlling the process (14 -16). Upon IL-4 stimulation, engagement of the IL-4 receptor (IL-4R) by IL-4 subsequently activates its two major downstream signals, JAK1/STAT6 and PI3K/AKT, in M2-polarized macrophages (17,18). The outcome of this process results in enhanced levels of Arg1, FIZZ1, Ym1, and CD206, which are also clinically featured in IPF (8,14,19). Nevertheless, much work is needed to unveil the signal integration in M2-driven fibrosis.
The Grb2-associated binder (Gab) proteins, which belong to the Gab/DOS family of docking proteins, function as important elements in signal integration during the assembly of downstream signaling complexes (20,21). The Gab proteins, Gab1, Gab2, and Gab3, have been identified in mammals, DOS (daughter of sevenless) has been identified in Drosophila, and Soc1 (suppressor-of-clear) has been identified in Caenorhabditis elegans (22)(23)(24). All Gab docking molecules are characterized by a highly conserved pleckstrin homology domain and a proline-rich domain. Although lacking enzymatic activity itself, Gab contains multiple tyrosine-based activation motifs for binding kinase and phosphatase, through which it confers signal diversity and specificity (20,21,(25)(26)(27).
Gab1 and Gab2 (Gab1/2) are widely expressed in various cell types, including immune cells such as T cells, macrophages, and mast cells (28). In response to multiple extracellular signals, Gab1/2 are rapidly phosphorylated and coordinate the signaling cascades, leading to the altered cellular functions or phenotypes. Gab1/2 have a potential functional redundancy due to their homology and similar expression patterns (29 -33). Recent studies have demonstrated that cardiomyocyte-specific Gab1/2 KO mice show high mortality after birth due to heart failure, suggesting the crucial roles of Gab1/2 in the maintenance of cardiac functions (29). In addition, both Gab1 and Gab2 are reported to participate in EGF-mediated MAPK/ERK signaling (33). Despite their similarity of function, many lines of evidence clearly demonstrate distinct roles for Gab1/2 under certain pathophysiological conditions. Germ line knock-out of Gab1 in mice is embryonic lethal due to defects in the heart, liver, muscle, and placenta (34,35). In contrast, Gab2 KO mice are viable and generally healthy with only mild defects in the allergic response and osteoclastogenesis (36 -38). In addition, both Gab1 and Gab2 promote endothelial cell migration in response to VEGF. However, contrary to Gab1, Gab2 depletion surprisingly results in the increased activation of AKT (39,40). We previously reported that Gab1 enhances allergic asthma by promoting dendritic cell migration (41), whereas Gab2 is involved in asthma by positively regulating goblet cell hyperplasia and mucus secretion in the airway epithelium (42).
In the present study, we observed that the expression of Gab1 and Gab2 are enhanced in alveolar macrophages isolated from BLM-induced fibrotic mice. In vitro and in vivo investigations demonstrated that the deficiency of either Gab1 or Gab2 could result in impaired M2 polarization. Furthermore, myeloid-specific Gab1 KO mice and Gab2 KO mice exhibited significantly attenuated pulmonary fibrosis induced by BLM. However, distinct molecular mechanisms are used by Gab1 and Gab2 to regulate IL-4 -induced macrophage polarization, with Gab1 positively regulating AKT signaling and Gab2 positively regulating STAT6 signaling. Our findings reveal a novel, non-redundant mechanism for Gab1/2 in modulating macrophage function as well as their regulatory roles in mediating the pathogenesis of pulmonary fibrosis.

Gab1/2 are highly expressed in M2-polarized macrophage isolated from fibrotic lungs
To probe the possible role of Gab1/2 in pulmonary fibrosis, we examined the levels of Gab1/2 in alveolar macrophages harvested from BLM-treated mice, which is a widely used animal model of pulmonary fibrosis (43). It was evident that Gab1/2 expression were elevated at both the mRNA and protein levels compared with SAL-treated mice ( Fig. 1 (A and B) and supplemental Fig. 1 (A and B)). These alveolar macrophages were characterized by enriched M2-polarized macrophages, which exhibited the canonical CD206 positivity and a strong up-regulation of M2 markers, including Arg1, FIZZ1, and Ym1 ( Fig. 1,  C and D). Both Gab1 and Gab2 are ubiquitously and relatively abundantly expressed. Nevertheless, Gab1 germ line knock-out is lethal in embryos (34,35). Therefore, our strategy was to develop a conditionally deleted Gab1 system (Gab1 MyKO ) to probe Gab1 function in macrophages. In contrast, complete Gab2 knockouts appear normal and are generally healthy (36 -38). Given the fact that Gab1 expression appears earlier and more abundant than Gab2, it is believed that loss of Gab2 is probably compensated somewhat by Gab1 or other Gab members. To determine whether there is a putative compensation between Gab1 and Gab2 in macrophages, which was previously reported but remains controversial, we detected the levels of Gab1/2 in Gab1 MyKO and Gab2 Ϫ/Ϫ mice. As shown in Fig. 1 (E and F), deletion of Gab1 in bone marrow-derived macrophages (BMDMs) resulted in unaltered Gab2 expression (Fig. 1E), whereas the loss of Gab2 resulted in a slight increase in Gab1 at both the mRNA and protein levels (Fig. 1F). This observation implies the synergistic role of Gab1/2 in macrophages. Overall, these data suggest that Gab1/2 are potentially involved in M2 polarization, which regulates the progression of pulmonary fibrosis.

Loss of Gab1/2 in macrophages impairs IL-4 -induced M2 polarization ex vivo
To further investigate the roles of Gab1 and Gab2 in macrophage M2 polarization, BMDMs were isolated from Gab1 f/f , Gab1 MyKO , Gab2 ϩ/Ϫ , and Gab2 Ϫ/Ϫ mice and then stimulated with IL-4, followed by M2 assessment. As indicated in Fig. 2 (A  and B), both Gab1 and Gab2 deficiency in macrophages impaired M2 activation with a substantial decrease in the expression of Arg1, FIZZ1, and Ym1 compared with the controls. Consistent with this finding, the reduction of Arg1 levels in Gab1/2 deficiency was confirmed using immunoblotting and immunofluorescence assays (Fig. 2 (C and D) and supplemental Fig. 2 (A and B)). We also found that the IL-4-mediated up-regulation of CD206 was attenuated in Gab1 MyKO and Gab2 Ϫ/Ϫ BMDMs (Fig. 2, E and F). Additionally, IL-4R␣ expression was unaltered in Gab1 MyKO and Gab2 Ϫ/Ϫ BMDMs (supplemental Fig. 3, A and B), and the loss of Gab1/2 had no effect on macrophage viability, as evaluated by annexin V-propidium iodide staining (supplemental Fig. 3, C and D). Collectively, these ex vivo observations support a positive role of Gab1/2 in the control of macrophage M2 polarization.

Ablation of Gab1/2 leads to a decrease in chitin-induced M2 polarization in vivo
To further explore the in vivo observations related to Gab1/2 in M2 polarization, Gab1 MyKO and Gab2 Ϫ/Ϫ mice were administered chitin intraperitoneally. Chitin is an important component of fungi and helminths and is essential for inducing mouse M2 activation (44,45). The in vivo levels of M2 markers were assessed at the transcriptional and protein levels in chitin-elic-ited macrophages isolated from the Gab1 MyKO and Gab2 Ϫ/Ϫ mice. We found that the up-regulation of the M2 markers induced by chitin was markedly abrogated in both Gab1-and Gab2-deficient macrophages (Fig. 3, A and B), and Arg1 and FIZZ1 expression were further analyzed using immunoblotting assay, with consistent results (Fig. 3, C and D). Furthermore, the in vivo individual deletion of Gab1/2 resulted in a reduction in CD206 following the chitin challenge compared with the cor- A, the mRNA expression levels of Gab1/2 (relative to ␤-actin) in alveolar macrophages obtained from SAL-or BLM-treated mice (n ϭ 6/group) were assessed by qPCR. B, Gab1/2 protein levels in alveolar macrophages obtained from SAL-or BLM-treated mice were determined by immunoblotting analysis. Each sample represents three mice. C, representative results of the immunostaining of CD206 (red)/DAPI (blue) in lung sections from SAL-and BLM-challenged mice (n ϭ 6/group). Scale bars, 25 m. D, analysis of Arg1, FIZZ1, and Ym1 mRNA expression levels by qPCR in alveolar macrophages from SAL-and BLM-treated mice (n ϭ 6/group). E, protein (top) and mRNA (bottom) levels of Gab2 were measured in BMDMs from Gab1 f/f and Gab1 MyKO mice. F, protein (top) and mRNA (bottom) levels of Gab1 were measured in BMDMs from Gab2 ϩ/Ϫ and Gab2 Ϫ/Ϫ mice. Data from three independent experiments are shown. Error bars, S.D. *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001.

The roles of Gabs in M2 polarization
responding controls (Fig. 3, E and F). Together, these data demonstrate that Gabs are required for M2 polarization in vivo.

Gab1/2 deficiency protects mice from BLM-induced pulmonary fibrosis
M2-polarized macrophages are known to contribute to the pathogenesis of pulmonary fibrosis (15,46), so we then determined whether the loss of Gab1 and Gab2 affects the progression of pulmonary fibrosis. To test this possibility, Gab1 f/f , Gab1 MyKO , Gab2 ϩ/Ϫ , and Gab2 Ϫ/Ϫ mice were challenged with BLM. M2 markers were determined by quantitative PCR (qPCR) on day 21, indicating that the expression of Arg1, FIZZ1, and Ym1 were significantly decreased in purified alveolar macrophages from Gab1 MyKO and Gab2 Ϫ/Ϫ mice compared with the corresponding controls (Fig. 4, A and B). Cells obtained from bronchoalveolar lavage fluid (BALF) were further analyzed by flow cytometry using alveolar macrophages markers (F4/80 ϩ CD11b Ϫ ), revealing that the percentages of

The roles of Gabs in M2 polarization
alveolar macrophages in the lungs were unchanged due to the loss of Gab1/2 (supplemental Fig. 4, A and B). Meanwhile, Gab1/2 deletion efficiency was confirmed by an immunoblotting assay (supplemental Fig. 4, C and D). Additionally, the total number of leukocytes in BALF was determined, and the results showed that the influx of inflammatory cells into the lungs of BLM-treated Gab1 MyKO and Gab2 Ϫ/Ϫ mice remained unaltered compared with the corresponding controls (supplemen-tal Fig. 4, E and F). Fibrotic injury can be evaluated by assessing excess and aberrant collagen deposition in a BLM-induced model of pulmonary fibrosis (43). We thus collected the whole lungs from BLM-challenged transgenic mice. Collagen accumulation was determined by Masson's trichrome staining. Histological analysis indicated a substantial decrease in collagen deposition and less fibrotic thickening in the alveolar walls in the Gab1 MyKO and Gab2 Ϫ/Ϫ mice compared with the corre-

The roles of Gabs in M2 polarization
sponding controls. In addition, hydroxyproline assays also showed reduced collagen expression (Fig. 4, C and D). The attenuated pulmonary fibrosis in Gab1 MyKO and Gab2 Ϫ/Ϫ mice was further supported by the reduced expression of fibrotic markers, such as fibronectin, ␣-SMA, and Col I (Fig. 4, E and F). To investigate whether the protective effect observed in Gab1 MyKO and Gab2 Ϫ/Ϫ mice was dependent on the impaired M2 polarization, BMDMs originated from Gab1 f/f , Gab1 MyKO , Gab2 ϩ/Ϫ , and Gab2 Ϫ/Ϫ mice were treated with IL-4 for the induction of M2 macrophages and administered into clodronate liposome-treated mice by the intratracheal route 7 days after BLM induction. Compared with the relative controls, animals that received Gab1 MyKO BMDMs or Gab2 Ϫ/Ϫ BMDMs exhibited reduced lung fibrosis, as determined by Masson's trichrome staining and hydroproline assays (Fig. 4, G and H). Together, these results indicate that high levels of Gabs in macrophages predispose mice to pulmonary fibrosis and, vice versa, that Gab1/2 deficiency protects mice from M2-associated pulmonary fibrosis.

Gab1/2 function in the distinct signal responsiveness to IL-4 activation
Next, we explored the mechanisms underpinning the defective M2 polarization in Gab1/2-deficient macrophages. Binding of the Th2 cytokine IL-4 signal with IL-4R␣ subsequently recruits a downstream signaling complex to activate tyrosine

The roles of Gabs in M2 polarization
phosphorylation events, which are essential for M2-polarized macrophages. It has been demonstrated that following IL-4R activation, cytoplasmic STAT6 and AKT are simultaneously phosphorylated to diverge IL-4 responsiveness (6,17,47,48). STAT6 is a well-defined transcription factor downstream of IL-4R activation that regulates M2-related genes, including Arg1, FIZZ1, Ym1, and CD206. As indicated in Fig. 5 (A-D), STAT6 phosphorylation was measured in the Gab1/2-deleted BMDMs in response to IL-4, and our results showed that loss of Gab1 had no effect on IL-4-activated STAT6 signaling (Fig. 5A), whereas loss of Gab2 caused a significant attenuation of STAT6 phosphorylation (Fig. 5B). Intriguingly, we then found diminished AKT phosphorylation in Gab1 MyKO BMDMs (Fig. 5C) but only a slight decrease in AKT phosphorylation in Gab2 Ϫ/Ϫ BMDMs (Fig. 5D). In addition, in vitro Gab1/2 were inactivated using shRNA-mediated deletion in RAW264.7 cells as a further assessment of IL-4R responsiveness. As illustrated in supplemental Fig. 5, Gab1 knockdown resulted in a marked reduction in AKT phosphorylation but not in STAT6 (supplemental Fig.  5A); however, Gab2 knockdown led to decreased phosphorylation of STAT6, but AKT remained intact (supplemental Fig.  5B). Our preliminary findings link reduced AKT activation rather than STAT6 activation to impaired M2 polarization in Gab1 MyKO mice. To address the AKT-dependent effects, we took a genetic approach to rescue AKT signaling in Gab1 MyKO BMDMs by adenovirus-mediated myristoylated AKT (myr-AKT; constitutively active AKT) transduction. As illustrated in Fig. 5E, immunoblotting analysis revealed increased levels of basal phospho-AKT in Gab1 MyKO BMDMs after adenoviral transduction of myr-AKT. qPCR analysis indicated that overexpression of active AKT increased the induction of M2 genes (Arg1, FIZZ1, and Ym1) following IL-4 stimulation (Fig. 5F). Finally, we determined whether this signal preference mediated by Gab1/2 is phenotypically correlated with BLM-induced M2 activation. To probe this possibility, we harvested alveolar macrophages from BLM-treated Gab1 MyKO and Gab2 Ϫ/Ϫ mice (Fig. 5, G and H). We observed a considerable reduction in phosphorylated AKT in Gab1 MyKO mice but not in Gab2 Ϫ/Ϫ mice (Fig. 5G); moreover, the activation of STAT6 was more evidently attenuated in Gab2 Ϫ/Ϫ than in Gab1 MyKO mice (Fig. 5H). Although this trend was readily apparent, it depends to a large extent on macrophages and their particular circumstances. Overall, our findings suggest that Gab1/2 function in the distinct signal responsiveness to IL-4 activation, supporting the notion that the docking proteins confer IL-4 signal preference in M2-polarized macrophages.

Gab1/2 selectively favors the formation of the signal complex driven by IL-4
To examine the mechanism by which Gab1/2 regulate the IL-4 signal preference, we conducted co-immunoprecipitation

The roles of Gabs in M2 polarization
assays to compare the signal interactions modulated by Gab1/2 in response to IL-4. We observed that Gab1 but not Gab2 selectively interacted with the p85 subunit of PI3K upon IL-4 stimulation, which may account for a stronger decline in AKT activation in Gab1 deficiency than that in Gab2 deficiency (Fig. 6A). It has been demonstrated that IL-4R enables JAK1 recruitment and then sequentially mediates STAT6 activation, thereby regulating a subset of M2-associated genes (6). As expected, we found that loss of Gab2 in BMDMs suppressed the phosphorylation of JAK1 (Fig. 6B). In addition, IL-4 stimulation promoted Gab2 interaction with IL-4R␣ (Fig. 6C). Further experiments revealed that the interaction of IL-4R␣ with JAK1 induced by IL-4 was substantially abrogated in Gab2 Ϫ/Ϫ BMDMs (Fig. 6D). Together, it is likely that Gab1/2 confer the IL-4 signal preference through selective association with either PI3K/AKT or JAK1/STAT6, respectively. We concluded that this preferred formation of a signal complex is responsible for the IL-4 signal preference observed in Gab1/2-deficient macrophages.

Discussion
It has been extensively reported that alternatively activated (M2) macrophages are involved in the progression of lung fibrosis; however, the precise mechanism underlying this process is still not clear (4,9,49,50). In this study, we defined a novel role of the Gab docking protein family in regulating M2 polarization-mediated lung fibrosis. We found that an increased level of Gab1/2 in alveolar macrophages is coupled with M2 skewing in mice challenged with BLM. Gab1/2 deletion in macrophages suppressed M2 polarization, which was phenotypically correlated with the attenuated fibrosis observed in Gab1 MyKO and Gab2 Ϫ/Ϫ mice. Mechanistically, Gab1/2 are synergistically involved in IL-4-regulated Th2 responses but serve as a non-redundant control for IL-4 activation. Our results further demonstrated that Gab1 regulates AKT activation, whereas Gab2 mediates STAT6 enhancement in response to IL-4, which supports the notion that the Gab1/2 docking proteins confer the IL-4 signal preference in M2-associated lung fibrosis. These findings may help us to better understand the finely tuned role of docking proteins in the pathogenesis of idiopathic pulmonary fibrosis.
Macrophages are mononuclear phagocytes, crucial determinants in orchestrating almost all forms of wound healing and fibrotic repair (5,51). In tissues, monocytes/macrophages respond to various environmental cues and are rapidly activated. This process is highly plastic and characterized by functional polarization into activated macrophages. Conceptually, polarized macrophages have been categorized as M1 (classically activated) and M2 (alternatively activated) (48,52,53). This M1/M2 paradigm has been questioned recently in terms of its accuracy and practicality, given the high heterogeneity of polarized macrophages under certain circumstances (54). Despite that, the functions of M2-polarized macrophage are widely accepted in this phenotypic shift toward fibrosis in the lungs. Emerging evidence has demonstrated that monocytes/ macrophages could be skewed into M2 polarization by a broad array of pro-fibrotic cytokines, leading to the marked elevation of M2 biomarkers, including Arg1, Ym1, FIZZ1, and CD206, in the BLM-induced murine model (15,(55)(56)(57). Analysis of clinical samples from IPF patients has clearly shown the pathological relevance of these observations, implying a promising therapeutic potential (58). Undoubtedly, elucidating this mechanism is a high priority and may facilitate the development of macrophage-directed interventions for this currently untreatable disease.
In this work, we first observed an elevated level of Gab1/2 in M2-polarized macrophages harvested from the lungs of BLMinduced mice, suggesting the potential involvement of Gab1/2 in M2-driven fibrosis (Fig. 1). Further investigation demonstrated an important role of Gab1/2 in promoting M2 polariza- Figure 6. Gab1 binds to p85 and Gab2 promotes the association of JAK1 with IL-4R␣. A, RAW264.7 cells were stimulated with IL-4 (20 ng/ml) for 30 min, and total cell lysates were subjected to immunoprecipitation (IP) with anti-Gab1 antibody (left) or anti-p85 antibody (right) followed by immunoblotting analysis with anti-p85 antibody, anti-Gab1 antibody, or anti-Gab2 antibody. B, BMDMs from Gab2 ϩ/Ϫ and Gab2 Ϫ/Ϫ mice were stimulated with IL-4 (20 ng/ml) for the indicated periods, and whole-cell lysates were measured by immunoblotting analysis using anti-phospho-JAK1 and JAK1 antibodies. C, RAW264.7 cells were treated with IL-4 (20 ng/ml) for 30 min. Total cell lysates were prepared and subjected to immunoprecipitation with anti-Gab2 antibody (left) or anti-IL-4R␣ antibody (right) followed by immunoblotting analysis with anti-IL-4R␣ antibody or anti-Gab2 antibody. D, BMDMs isolated from Gab2 ϩ/Ϫ and Gab2 Ϫ/Ϫ mice were stimulated with IL-4 (20 ng/ml) for 30 min. Whole-cell lysates were immunoprecipitated with anti-JAK1 antibody. Immunoblotting analysis was performed with anti-IL-4R␣ antibody. Data from three independent experiments are shown.

The roles of Gabs in M2 polarization
tion in BMDMs from transgenic mice lacking either Gab1 or Gab2 (Fig. 2). Herein, we employed LysM-Cre-mediated excision to induce a selective deletion in monocytes/macrophages. LysM-driven Cre recombinases are also expressed to some degree in neuronal subsets, neutrophils, and certain other blood cells, although LysM-Cre-mediated deletion in macrophages is still highly efficient (59 -61). However, there is currently no true macrophage-specific Cre deletion technique. Thus, testing the deletion efficiency by immunoblotting assays is essential, as stated by Peter Murray (60). Accordingly, the degree of Gab1 knockdown in purified alveolar macrophages was assessed by immunoblotting analysis. As indicated in supplemental Fig. 4C, Gab1 levels were efficiently lowered in purified alveolar macrophages. Chitin is a structural component of fungi and helminths, and the administration of chitin strongly induced macrophage M2 polarization, manifested by the upregulated expression of M2 markers (44,45). Herein, we used in vivo models of chitin to clarify the functional roles of Gab1/2 in M2 polarization (Fig. 3), suggesting a broad effect of Gab1/2 on M2-polarized macrophages. In this work, our data clearly define the positive regulation of Gab1/2 in M2-polarized macrophages, which contributes to pulmonary fibrosis. Future studies are needed to evaluate additional functions of Gab1/2 in tissue-specific macrophages and M2-associated disorders.
Gab proteins, which serve as docking proteins for integrating phosphorylation, comprise a family of adaptor proteins including mammalian Gab1, Gab2, and Gab3 (22). Gab1 and Gab2 are ubiquitously expressed with a widespread expression pattern (22, 25). Gab3 has limited expression in hematopoietic cells, and Gab3-deficient mice exhibit no obvious defects in normal development and hematopoiesis (27,62). Thus, this work focused on dissecting the functions of Gab1 and Gab2 in macrophages. In fact, M2 polarization can be triggered by a broad array of cytokines and pathogenic agents, so we further investigated a possible mechanism by which Gab1/2 mediate M2 polarization in response to canonical IL-4 stimulation. Although the regulation of M2 by IL-4 is diverse, emerging evidence clearly indicates that upon stimulation, IL-4 binds IL-4R␣ and then recruits IL-2R␥ (type I receptor), leading to the activation of the non-receptor tyrosine kinase JAK1, enabling the phosphorylation, dimerization, and translocation of STAT6 to the nucleus, where it binds to the promoters of IL-4responsive genes (6,17,18,63). In addition, the PI3K/AKT signaling pathway is also engaged by IL-4R␣ in parallel to the JAK1/STAT6 pathway to regulate the expression of M2 markers (47,64). Intriguingly, we found a distinctive role of Gab1/2 in IL-4-mediated downstream elements, although Gab1/2 deficiency in macrophages led to an apparently similar M2-associated fibrosis (Fig. 4). It is evident that in response to IL-4 stimulation, Gab1 preferentially activates PI3K/AKT signaling, whereas Gab2 favors JAK1/STAT6 activation (Fig. 5). Consistent with this, we noticed a preferential interaction of Gab1 with p85, whereas Gab2 facilitates the interaction of IL-4R␣ with JAK1 (Fig. 6), attributable to the M2-associated phenotypic shift and fibrosis observed in the mice. Despite lacking catalytic activity, docking proteins confer signal specificity through their selective preference for protein interactions. In this work, we presented an illustrative example that the Gab1/2 docking pro-teins enable the IL-4 signal preference, commonly contributing to M2 polarization and the development of fibrosis. Future studies may provide additional information regarding the precise contribution of the IL-4 signal preference by Gab1/2 divergence in M2 subpopulations in fibrotic lungs. With such efforts, clarifying the signal convergence and divergence should greatly facilitate our understanding of the pathogenesis of this idiopathic pulmonary disorder.
In conclusion, our findings highlight a non-redundant role of the Gab1/2 docking proteins in M2 polarization, an important determinant of M2-associated pulmonary fibrosis. Interfering with the Gab1/2-mediated IL-4 signal preference could be a more effective and less toxic therapeutic option for macrophage-guided therapy.

Mice
Gab1 flox/flox mice (41) were crossed with LysM cre/ϩ mice to generate conditional Gab1 KO mice (Gab1 MyKO ). Gab2 Ϫ/Ϫ mice were bred with Gab2 ϩ/Ϫ mice to generate Gab2 KO animals (65). All animal protocols were approved by the animal care and use committee of the Zhejiang University School of Medicine.

Antibodies and reagents
Recombinant murine IL-4 and M-CSF were obtained from Peprotech (Rocky Hill, NJ). The following antibodies were used for immunoblotting and immunoprecipitation: rabbit poly-

BLM-induced pulmonary fibrosis
Gab1 f/f , Gab1 MyKO , Gab2 ϩ/Ϫ , and Gab2 Ϫ/Ϫ mice (8 -10 weeks old) were intratracheally instilled with 2.5 units/kg BLM. Mice subjected to the same volume of SAL served as controls, and the mice were sacrificed by injection of pentobarbital 21 days after the challenge for the analysis of lung fibrosis. All experiments in this study were reviewed and approved by the animal care and use committee of Zhejiang University School of Medicine.

Histopathology and immunohistochemical analyses
For histopathology analysis, lungs were inflation-fixed with an intratracheal injection of 4% paraformaldehyde in PBS at 25 cm H 2 O and immersed in the same fixative to preserve the pulmonary architecture, followed by paraffin embedding. The sections (5 m) were stained with Masson's trichrome for assessment of collagen deposition. For immunostaining, the slides were probed with antibodies against CD206 (1:50; BioLegend).

BALF analysis
After 21 days of BLM administration, mice were euthanized by pentobarbital injection. The left lungs were tied at the bronchus, and the right lungs were lavaged with 0.5 ml of cold sterile Ca 2ϩ -and Mg 2ϩ -free PBS. This procedure was repeated three times, and ϳ1.5 ml of BALF was recovered from each animal. The samples were pelleted at 1000 rpm for 5 min at 4°C, and supernatants were stored at Ϫ80°C until analyzed. To isolate and purify alveolar macrophages, the pellets were resuspended in DMEM/F-12 medium (Hyclone, Logan, UT) supplemented with 10% heat-inactivated FBS (Hyclone), 50 units/ml penicillin, and streptomycin (Hyclone) and plated for at least 90 min. Non-adherent cells were discarded, and adherent cells were gathered for FACS analysis. Macrophages were enriched to about 90% after quick adhesion, as described previously (66,67).

Hydroxyproline assays
Lung tissues were weighed and acid-hydrolyzed with 6 N HCl, and hydroxyproline content was determined as a measure of fibrosis with a colorimetric assay, as described previously (68).

myr-AKT transduction
For myr-AKT transduction, BMDMs from Gab1 MyKO mice were plated at 5 ϫ 10 5 cells/well in 12-well tissue culture dishes and cultured overnight. Viral supernatants (empty vector (EV) or myr-AKT (Applied Biological Materials, Richmond, Canada)) were then added, and after 10 h, the medium was replaced, and cells were allowed to grow until further analysis.

Flow cytometry
Cells were washed with cold PBS three times and stained with fluorochrome-conjugated monoclonal antibodies against mouse CD206 (BioLegend) for 30 min at 4°C. After incubation, cells were washed with cold PBS three times and subsequently subjected to FACS analysis using an AECA NovoCyte TM system (ACEA Biosciences, San Diego, CA). Data were analyzed using FlowJo software version 7.6.

Immunoblotting and co-immunoprecipitation assays
Cells were washed with cold PBS and lysed in radioimmune precipitation lysis buffer (150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 50 mM Tris-HCl (pH 8.0), and 0.1% SDS) with EDTA-free protease inhibitor tablets and phosphatase inhibitors (Roche Diagnostics). Equal amounts of protein were separated by SDS-PAGE and transferred onto nitrocellulose membranes (Pall, Port Washington, NY) for immunoblotting with primary antibodies as indicated.

The roles of Gabs in M2 polarization
were washed six times with Nonidet P-40 lysis buffer and resuspended in 1ϫ protein loading buffer. Protein levels were determined by immunoblotting as indicated above. Three independent experiments were performed.

Densitometry analysis
Shorter exposures of films were chosen for densitometry analysis to ensure that the intensities of bands were in the linear range of the film, and the integrated densities of indicated bands were calculated with National Institutes of Health ImageJ Software. The relative abundance of phosphorylated protein was calculated as a ratio of the density of the phosphorylated band divided by the density of the total protein band.

Chitin administration
Chitin (Sigma-Aldrich) was washed with PBS and sonicated for 30 min on ice. About 800 ng of chitin was injected intraperitoneally, and peritoneal macrophages were collected 2 days after administration.

Immunofluorescence
After stimulation with IL-4 for 48 h, BMDMs were washed with PBS, fixed in 4% paraformaldehyde solution for 20 min, and permeabilized with 0.1% Triton X-100 for 20 min. After serum blocking for 1 h at room temperature, BMDMs were then incubated with anti-Arg1 antibody overnight. The next day, BMDMs were washed with PBS, incubated with Alexa Fluor 594 goat anti-rabbit IgG (1:400; Invitrogen) for 1 h at room temperature, washed with PBS, and immersed in DAPI to visualize the nuclei. Staining was imaged with an inverted confocal microscope (Carl Zeiss, Göttingen, Germany).

Statistical analysis
All results were expressed as the mean Ϯ S.D. The p value was calculated using the GraphPad Prism version 5 statistical program and determined by two-tailed Student's t test. A value of p Ͻ 0.05 was considered statistically significant.
Author contributions-X. G., T. L., H. C., X. Z., and Y. K. designed the study; X. G., T. L., Y. X., X. X., and Z. Z. performed experiments, collected, and analyzed data; Y. Z., J. X., and K. X. provided advice on experimental design; X. G., X. Z., and Y. K. wrote the paper. All authors approved the final version of the manuscript.