Hermansky–Pudlak syndrome (HPS) comprises a group of inherited disorders caused by mutations that alter the function of lysosome-related organelles. Pulmonary fibrosis is the major cause of morbidity and mortality in HPS-1 and HPS-4 patients. However, the mechanisms that underlie the exaggerated injury and fibroproliferative repair responses in HPS have not been adequately defined. In particular, although Galectin-3 (Gal-3) is dysregulated in HPS, its roles in the pathogenesis of HPS have not been adequately defined. In addition, although chitinase 3-like 1 (CHI3L1) and its receptors play major roles in the injury and repair responses in HPS, the ability of Gal-3 to interact with or alter the function of these moieties has not been evaluated. In this article, we demonstrate that Gal-3 accumulates in exaggerated quantities in bronchoalveolar lavage fluids, and traffics abnormally and accumulates intracellularly in lung fibroblasts and macrophages from bleomycin-treated pale ear, HPS-1–deficient mice. We also demonstrate that Gal-3 drives epithelial apoptosis when in the extracellular space, and stimulates cell proliferation and myofibroblast differentiation when accumulated in fibroblasts and M2-like differentiation when accumulated in macrophages. Biophysical and signaling evaluations also demonstrated that Gal-3 physically interacts with IL-13Rα2 and CHI3L1, and competes with TMEM219 for IL-13Rα2 binding. By doing so, Gal-3 diminishes the antiapoptotic effects of and the antiapoptotic signaling induced by CHI3L1 in epithelial cells while augmenting macrophage Wnt/β-catenin signaling. Thus, Gal-3 contributes to the exaggerated injury and fibroproliferative repair responses in HPS by altering the antiapoptotic and fibroproliferative effects of CHI3L1 and its receptor complex in a tissue compartment-specific manner.
Hermansky–Pudlak syndrome (HPS) is a group of inherited autosomal recessive disorders that occur worldwide (1–3). Ten genetic subtypes HPS-1 to HPS-10) have been described, with each mutation affecting the function of lysosome-related organelles (LROs) (3, 4). Although pulmonary fibrosis occurs in HPS-2 (5), this fatal complication has been appreciated largely in HPS-1 and HPS-4 patients, whose genetic defects are in biogenesis of LRO complex 3 (BLOC-3), which includes HPS1 and HPS4 proteins (6–11). HPS-1 is particularly common in northwest Puerto Rico, where 1:1800 people are affected and the carrier frequency is 1 in 21 persons (12–14). Due to the untreatable and progressive nature of the pulmonary fibrosis of HPS, this complication is the leading cause of death for those with the disorder (15). However, the mechanism(s) by which LRO-related defects lead to the exaggerated injury and fibroproliferative repair responses, defined in HPS and models of the disorder (7, 10, 16–19), have not been adequately defined.
Galectin-3 (Gal-3) is a β-galactoside–binding lectin that is expressed in the nucleus, cytoplasm, or within the extracellular milieu of cells from multiple organs including the lung (20, 21). It has pleiotropic effector functions including the ability to regulate cell death responses (22, 23) and tissue fibrosis (20, 21, 24–26). Recent studies from our laboratories demonstrated that the concentrations of Gal-3 in bronchoalveolar lavage (BAL) fluids from HPS patients were significantly higher than in samples from patients with idiopathic pulmonary fibrosis or controls, and correlate with HPS disease severity (27). Additionally, studies demonstrated that dermal fibroblasts from HPS subtypes that are associated with pulmonary fibrosis manifest abnormal Gal-3 trafficking and exaggerated intracellular Gal-3 accumulation compared with cells from controls and HPS subtypes that do not have pulmonary fibrosis (27). However, the roles of Gal-3 in the exaggerated injury and fibroproliferative repair responses in HPS have not been defined. In addition, the possibility that Gal-3 will manifest different effects in different HPS tissue compartments has not been addressed.
Chitinase 3-like 1 (CHI3L1) is a pleiotropic glycoprotein that also inhibits cell death and drives fibroproliferative repair. Its effects are mediated by at least two receptors. One is the multimeric chitosome that includes IL-13α2 and TMEM219, and the other is CRTH2 (28–31). Recent studies from our laboratory demonstrated that the levels of circulating CHI3L1 are higher in HPS patients with pulmonary fibrosis compared with those who remain fibrosis-free, and that these levels correlate with disease severity. Using murine models, we also demonstrated that a defect in CHI3L1 inhibition of epithelial apoptosis and exaggerated CHI3L1-driven fibroproliferation play important roles in HPS fibrosis. We showed that BLOC-3 proteins differentially contribute to the trafficking of CHI3L1 receptor components; IL-13Rα2 trafficked abnormally, but CRTH2 trafficked normally in the absence of HPS1. These studies also demonstrated that the abnormal IL-13Rα2 trafficking abrogated the antiapoptotic effects of CHI3L1 and contributed to the enhanced injury responses and sensitivity to apoptosis in BLOC-3 HPS patients and murine models of these disorders (31). In contrast, CHI3L1 drives fibrosis via interactions with CRTH2, which traffics normally in BLOC-3 HPS and HPS-1–deficient mice. When viewed in combination, these studies highlight the importance of the CHI3L1 axis in the exaggerated sensitivity of the epithelium to injury and the augmented fibroproliferative repair response in HPS. They also demonstrate that IL-13Rα2 and Gal-3 have overlapping effector profiles, and that the trafficking of both is dependent on BLOC-3 proteins. However, a relationship between Gal-3 and CHI3L1 and its receptors in HPS has not been defined.
We hypothesized that Gal-3 influences the HPS fibroproliferative repair response in a tissue-specific manner, via interactions with CHI3L1 and or its receptor components. Hence we characterized the levels of Gal-3 in BAL from bleomycin-treated wild-type (WT) and HPS-1–deficient mice, characterized the expression and trafficking of Gal-3 in epithelial cells, fibroblasts, and macrophages from these animals, and defined the effects of extracellular and intracellular Gal-3 on these cell populations. We also used biophysical approaches to define the interactions of Gal-3, CHI3L1, and components of its receptor complex. These studies demonstrate that Gal-3 is upregulated in murine models of HPS-associated pulmonary fibrosis, and that Gal-3 accumulates intracellularly because of defective trafficking in fibroblasts and macrophages from HPS-1–deficient mice. We also show that extracellular Gal-3 is a potent stimulator of epithelial apoptosis, whereas intracellular Gal-3 drives fibroproliferative repair by inhibiting fibroblast apoptosis, increasing fibroblast proliferation and myofibroblast transformation, and increasing M2-like macrophage differentiation. Lastly, these studies demonstrated that Gal-3 physically interacts with CHI3L1 and its IL-13Rα2 and TMEM219 receptor components, which abrogates CHI3L1-induced antiapoptotic signaling, augments Wnt/β-catenin signaling, and contributes to the development of epithelial cell death and tissue fibrosis in the lungs of HPS-1–deficient mice.
Materials and Methods
Knockout and transgenic mice
Pale ear mice (HPS1−/−) and Gal-3 knockout mice were obtained from Jackson Laboratory. Pale ear/Gal-3 double-mutant mice were generated in our laboratory. All mice were congenic on a C57BL/6 background and were genotyped as previously described (31).
Sex-matched, 8-wk-old WT, pale ear, Gal-3−/−, and Pale ear/Gal-3 double-mutant mice (≥4 mice per group per experiment, repeated at least three times) were exposed to a single bleomycin injection (1.25 U/kg; Teva Parenteral Medicines, Irvine, CA) via intratracheal administration. Mice were sacrificed and evaluated at days 7 and 14 to examine apoptosis and fibrosis, respectively.
Total cellular RNA was obtained using TRIzol reagent (Invitrogen), according to the manufacturer’s instructions. mRNA was measured using real-time RT-PCR as described previously (32, 33). The primer sequences for extracellular matrix genes were obtained from PrimerBank (http://pga.mgh.harvard.edu/primerbank/) or the same as previously used (32, 34, 35).
Gal-3 levels in mouse BAL samples, cell lysates, and cell culture supernatant were quantified using an ELISA kit (R&D Systems) following the manufacturer’s instructions.
Double-label immunofluorescence staining was undertaken using paraffin-embedded lungs from WT and pale ear mice to localize the expression of Gal-3. Monoclonal anti–Gal-3 (Santa Cruz), anti-F4/80 (Abcam), CD68 (Abcam), and anti-CHI3L1 (R&D) Abs were used in these evaluations as previously described (29).
Cell proliferation assay
Cells were seeded (4000 cells/well) in 96-well plates. Water Soluble Tetrazolium Salt-1 reagent (Roche, Basel, Switzerland) to detect mitochondrial dehydrogenase enzymes was added directly to cell culture medium between 16 and 24 h after serum starvation. After 1 h of incubation, the plate was read at 450 nm with equipment reference reading at 630 nm.
Mouse lungs were removed en bloc, inflated to 25-cm pressure with PBS containing 0.5% low melting point agarose gel, fixed, embedded in paraffin, sectioned, and stained. H&E and Mallory’s trichrome stains were performed in the Research Histology Laboratory at Brown University. BAL and lung inflammation were assessed as described previously (32).
Quantification of lung collagen
Animals were anesthetized, median sternotomy was performed, and right-heart perfusion was completed with calcium and magnesium-free PBS. The heart and lungs were then removed. The right lung was frozen in liquid nitrogen and stored at −80°C until used. Collagen content was determined by quantifying total soluble collagen using the Sircol Collagen Assay kit (Biocolor; Accurate Chemical & Scientific, Westbury, NY) according to the manufacturer’s instructions.
End labeling of exposed 3′-OH ends of DNA fragments in paraffin-embedded tissue was undertaken with the TUNEL in situ cell death detection kit AP (Roche Diagnostics). After staining, 8–10 random pictures were obtained from each lung or cell culture well, and a minimum of 200 cells was visually evaluated in each section. The labeled cells were expressed as a percentage of total nuclei.
Primary alveolar macrophages, lung fibroblasts, and type II lung epithelial cells from WT or pale ear mice were used in the experiments. MLE12 cells were obtained from ATCC. Treatments include recombinant CHI3L1 (rCHI3L1; 500 ng/ml), recombinant Gal-3 (rGal-3; 100 ng/ml), TGF-β (20 ng/ml), H2O2 (100 μg/ml), Gal-3 small interfering RNA (siRNA), Gal-3 overexpression construct, and bleomycin (100 μg/ml). Cells were cytospun onto slides for TUNEL analysis as previously described. Cellular RNA or protein was extracted for additional analyses.
Coimmunoprecipitation and Western blots
Proteins from the lung lysate of WT mice or IL-13 transgenic (Tg) mice were clarified by centrifugation for 10 min at 4°C. Gal-3, IL-13Rα2, and CHI3L1 were immunoprecipitated (IP) with anti–Gal-3 rabbit polyclonal Ab (sc-53127, against full-length Gal-3; Santa Cruz), anti-CHI3L1 (MAB2649, against the catalytic domain; R&D), or anti–IL-13Rα2 mAb (AF539, against the extracellular domain; R&D), respectively, using Catch and Release V2.0 (Reversible Immunoprecipitation System; EMD Millipore, Darmstadt, Germany). Similarly, recombinant human proteins were IP with anti-Gal-3 rabbit polyclonal Ab (sc-53127, against full length Gal-3; Santa Cruz), anti-CHI3L1 (AF2599, against the catalytic domain; R&D), or anti-IL-13Rα2 mAb (AF146, against the extracellular domain; R&D). The precipitates were subjected to immunoblotting with Abs against Gal-3, IL-13Rα2, CHI3L1, and TMEM219 (AF7556, against the extracellular domain; R&D), respectively. Western blots for signaling were performed as previously described (28, 30).
Equipment and running buffer preparation for surface plasmon resonance
Surface plasmon resonance (SPR) was conducted at Brown University EPSCOR Proteomics facility with the GE Biacore X100 Plus Package (which includes the in-line buffer degasser that removes air bubbles to minimize their disruption to protein–protein interactions) and analyzed via the Biacore Evaluation Software on the Windows 2000 machine in the facility. HBS-EP+ buffer was diluted 1:10 with MilliQ water (final concentration: 10 mM HEPES, 150 mM NaCl, 3 mM EDTA, and 0.005% Surfactant P20; BR100826; GE Healthcare Life Sciences) for use within 24 h and vacuum steri-filtered using Stericup-GP 250 ml Express Plus PES 0.22 μm (SCGPU02RE; EMD Millipore) to remove potential debris or precipitates. This buffer was the primary buffer for any interaction studies for the flow system.
Recombinant proteins for SPR
Proteins involved in these SPR experiments were purchased from ACROBiosystems. Posttranslational glycosylation is critical for modulating binding in some situations. Proteins were reconstituted at the manufacturer-recommended concentration indicated below in HBS-EP running buffer, the same as the assay running buffer to minimize bulk refraction effect, and frozen in 20-μl aliquots at −80°C. Proteins that have been thawed are kept for <1 wk in 4°C.
General protocol of preconcentration optimization and immobilization of ligand onto sensor chip CM5
Before immobilizing ligand onto the CM5 chip, the protein Gal-3 required a preconcentrating pH screen to optimize ionic association of the protein to a nonactivated CM5 chip surface, which intrinsically carries a slightly negative charge. Gal-3 was screened for ideal immobilization buffer sodium acetate (10 mM NaOAc-HCl or NaOAc-HOAc) buffer pH 4–5.5 in 0.5 pH increments. When protein was diluted to 40-10 μg/ml concentration in immobilization buffer, which was in most immobilizations a 1:10 dilution, the protein exhibited in theory a positive charge, because the isoelectric point of Gal-3 is >5.5, creating ionic interactions with the CM5 chip. The amine coupling strategy via the appropriate Amine Coupling Kit (BR100050; GE Healthcare) was employed to covalently link ligand onto the Biacore CM5 Sensor Chip (29149604; GE Healthcare), a glass slide with a gold cover that is covered with a carboxy-dextran matrix. 1-Ethyl-3-(3-dimethylaminopropyl)-carbodiimide (0.4 M) and N-hydroxysuccinimide (0.1 M) were each stored when reconstituted with MilliQ-grade water at −80°C and used within 4 mo. As recommended by Biacore, 1 M ethanolamine-HCl (pH 8.5) was stored at 4°C and used within 4 mo. Upon need to immobilize ligand onto the chip, chip was docked and primed twice with the running buffer. Then the chip was injected with a 1:1 v/v injection of 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide/N-hydroxysuccinimide that had been automatically mixed together seconds ago. This mix was flowed across the chip at 5 μl/s for 7 min. Afterward, ligand (Gal-3, for example) at the specified concentration of either 20 or 10 μg/ml diluted from the 400 μg/ml stock with the appropriate pH NaOAc buffer (in the case of Gal-3 [pH 5]) was injected into the system in which the HBS-EP running buffer was flowing at 5 μl/s for 7 min. Finally, ethanolamine was injected at 5 μl/s for 7 min to inactivate active cross-linking residues on the CM5 chip.
SPR binding kinetics studies
Steri-filtered 100 mM α-lactose in water was used as regeneration solution with the Gal-3 chip. The regeneration consisted of a 30-s injection of the regeneration solution over the Gal-3–coated surface. For kinetic study, samples of analyte (the protein injected into the flow cell to test binding with Gal-3) were serially diluted 2-fold from typically 400 or 100 nM by the running buffer in the flow system immediately before using the Biacore instrument. Typically, 90–120 s analyte injection association was used with between 180 and 240 s dissociation observed for appropriate model fitting. Experiments were done at least three times at various serial dilution patterns to establish that the kinetic constant values were independent of the concentration of the analyte.
Mouse data are expressed as mean ± SEM. As appropriate, groups were compared by ANOVA with the Bonferroni posttest; follow-up comparisons between groups were conducted using a two-tailed Student t test. A p value ≤0.05 was considered to be significant.
Animal experiments were approved by the Institutional Animal Care and Use Committee of Brown University in accordance with federal guidelines.
The levels of Gal-3 are increased in lungs from pale ear mice
The pale ear mouse has a null mutation of the HPS1 gene and shares many aspects of the human HPS-1 disease phenotype (36, 37). To determine whether Gal-3 is dysregulated in pale ear mice, the levels of lung lysate Gal-3 mRNA and BAL Gal-3 protein were assessed in WT and pale ear mice at baseline and after bleomycin challenge. The latter assessments were undertaken during the fibroproliferative repair phase of the bleomycin response as previously defined (29). At baseline, the levels of Gal-3 mRNA and protein were similar in the lungs of pale ear versus WT mice (Fig. 1A, 1B). In WT mice, bleomycin administration caused a fibroproliferative repair response that was associated with a significant increase in tissue and BAL Gal-3 (Fig. 1A, 1B). The levels of tissue and BAL Gal-3 in pale ear mice exceeded the levels in WT mice after bleomycin treatment (Fig. 1A, 1B). Immunohistochemical (IHC) evaluations of the lungs from these mice demonstrate that Gal-3 is expressed by inflammatory cells in areas of injury and fibrosis (Fig. 1C). Immunofluorescence staining also demonstrated that macrophages are a major cellular source of Gal-3 in pale ear and WT animals (Fig. 1D). These studies demonstrate that the expression and accumulation of Gal-3 are increased during bleomycin-induced fibroproliferative repair in pale ear versus control mice.
HPS fibroblasts express and contain more Gal-3 than WT cells and have a Gal-3 trafficking defect
Studies were next undertaken to define the expression and trafficking of Gal-3 in lung fibroblasts from WT and pale ear mice. In these experiments, we characterized the levels of Gal-3 mRNA and extracellular and intracellular Gal-3 using primary cells from WT and pale ear mice before and after stimulation. At baseline, the levels of Gal-3 mRNA in fibroblasts from pale ear mice were greater than in cells from WT animals (Fig. 2A). There were also modest increases of intracellular Gal-3 protein levels in cells from pale ear mice (Fig. 2B). However, the baseline levels of extracellular Gal-3 were not different in cells from the two groups (Fig. 2C). Treatment with TGF-β1 or H2O2 increased the levels of mRNA encoding Gal-3 in fibroblasts from both WT and pale ear mice (Fig. 2A, 2D). This was associated with increased levels of intracellular Gal-3 in pale ear cells compared with WT controls (Fig. 2B, 2E). Despite having higher levels of Gal-3 mRNA and higher levels of intracellular Gal-3 than in WT cells, pale ear cells manifested significantly lower levels of Gal-3 secreted into cell culture media after TGF-β1 or H2O2 stimulation (Fig. 2C, 2F). In combination, these findings demonstrate that HPS lung fibroblasts express and contain more Gal-3 than WT cells and have a Gal-3 trafficking defect that augments the intracellular accumulation of Gal-3 and decreases the secretion of Gal-3 into the extracellular space.
HPS macrophages express and contain more Gal-3 than WT cells and have a Gal-3 trafficking defect
To see whether the findings noted earlier were specific for fibroblasts, similar investigations were undertaken using macrophages. As noted in Fig. 2G, CHI3L1 treatment increased the levels of Gal-3 mRNA in macrophages from both WT and pale ear mice, with the levels in pale ear exceeding those in WT cells. In addition, CHI3L1 treatment increased the levels of intracellular Gal-3 accumulation in pale ear cells; the levels in HPS cells exceeded those of WT controls (Fig. 2H). In contrast, pale ear macrophages manifested significantly lower levels of secreted Gal-3 in cell culture media compared with WT cells (Fig. 2I). Thus, as for the HPS fibroblasts, HPS macrophages expressed and contained more Gal-3 than WT cells and exhibited a Gal-3 trafficking defect that augments the intracellular accumulation of Gal-3 and decreases the secretion of Gal-3 into the extracellular space.
Gal-3 expression and production are not seen in epithelial cells from WT and pale ear mice
To determine whether Gal-3 was expressed in epithelial cells, we isolated alveolar type 2 epithelial cells from lungs from WT and pale ear mice and evaluated the levels of Gal-3 mRNA and intracellular and extracellular Gal-3 protein before and after stimulation. At baseline, Gal-3 mRNA was not readily appreciated in cells from WT or pale ear mice (data not shown). In addition, stimulation with TGF-β1, bleomycin, or H2O2 did not induce levels of intracellular or extracellular Gal-3 that could be appreciated in our evaluations (data not shown). Thus, in contrast with fibroblasts and macrophages, lung epithelial cells did not display induction or production of Gal-3 in WT or pale ear mice.
Endogenous Gal-3 regulates apoptosis and fibroproliferative repair in vivo in pale ear mice
We next compared the bleomycin-induced epithelial injury (apoptosis) and fibroproliferative repair (collagen accumulation) responses in WT mice, Gal-3 null mice, pale ear mice, and pale ear/Gal-3 double null mutant animals. When compared with the WT controls, bleomycin-challenged pale ear mice manifest exaggerated levels of alveolar type II cell apoptosis and collagen accumulation (Fig. 3A, 3B). Importantly, the exaggerated bleomycin-induced epithelial cell death and fibrotic responses in pale ear mice were significantly decreased in the absence of Gal-3 (Fig. 3A, 3B). In accord with the later findings, bleomycin administration increased the levels of mRNA encoding α-smooth muscle actin (α-SMA) in lungs from WT and pale ear mice; this inductive response was exaggerated in lungs from pale ear animals, and this increase was abrogated in pale ear/Gal-3 double-mutant mice (Fig. 3C). Studies by others have demonstrated that alternatively activated (M2) macrophages play an important role in the development of lung fibrosis (38), either via direct effects on matrix or indirectly by promoting fibroblast accumulation and activation (39, 40). Bleomycin administration increased the levels of mRNA encoding the M2 macrophage marker CD206 in lungs from WT and pale ear mice, and this inductive response was exaggerated in lungs from the HPS-1–deficient mice (Fig. 3D). Importantly, the genetic removal of Gal-3 from WT mice or in pale ear mice led to significant decreases in pulmonary CD206 expression (Fig. 3D). This was particularly impressive in pale ear/Gal-3 double-mutant animals whose levels of expression of CD206 after bleomycin treatment were below those in similarly treated WT controls (Fig. 3D). When viewed in combination, these studies demonstrate that bleomycin-treated pale ear mice manifest exaggerated levels of epithelial cell apoptosis, pulmonary fibrosis, and myofibroblast and M2 macrophage accumulation. They also indicate that Gal-3 plays a key role in these events because the exaggerated apoptosis, fibrosis, and myofibroblast and M2 macrophage accumulation in bleomycin-treated pale ear mice were eliminated in pale ear/Gal-3 double-mutant animals.
Extracellular Gal-3 induces apoptosis of primary lung epithelial cells from pale ear mice in vitro
Extracellular Gal-3 has been shown to induce, and intracellular Gal-3 to inhibit, cellular apoptosis (22, 23). Because epithelial cell death plays an important role in the pathogenesis of pulmonary fibrosis (41), studies were undertaken to define the effects of Gal-3 on epithelial cell apoptosis. In these experiments, we harvested primary type II alveolar epithelial cells from pale ear mice and WT mice, and investigated the effects of extracellular rGal-3 in the presence and absence of bleomycin using TUNEL evaluations. These studies demonstrated that rGal-3 augmented TUNEL staining in epithelial cells from WT mice and pale ear mice (Fig. 4A). Moreover, the levels of apoptosis were more pronounced in cells from pale ear animals (Fig. 4A). Bleomycin also increased the TUNEL staining of cells from WT and pale ear mice (Fig. 4B), and rGal-3 augmented the levels of apoptosis in these bleomycin-treated WT and pale ear cells (Fig. 4B). Importantly, pale ear cells had higher levels of baseline apoptosis and were more sensitive to rGal-3–induced apoptosis than WT cells (Fig. 4). These effects were not mediated by alterations in intracellular Gal-3; the levels of apoptosis were not altered by treatment with Gal-3 siRNA and as noted earlier, significant levels of mRNA encoding Gal-3 were not seen in epithelial cells (Supplemental Fig. 1). In combination, these studies highlight the ability of extracellular Gal-3 to induce epithelial cell apoptosis and the enhanced sensitivity of epithelial cells from pale ear mice to the cell death–inducing effects of Gal-3.
Intracellular Gal-3 regulates fibroblast death, proliferation, and differentiation in vitro
We next investigated the effects of extracellular and intracellular Gal-3 on fibroblast cell death, proliferation, and differentiation. In the first experiments, H2O2 was used to induce cellular injury and apoptosis. As can be seen in Fig. 5A, H2O2 increased the number of TUNEL+ fibroblasts from WT and pale ear mice. Importantly, the pale ear fibroblasts were relatively resistant to apoptotic stimuli, manifesting ∼40% fewer TUNEL+ cells than WT cells (Fig. 5A). Interestingly, this protection appeared to be mediated by intracellular Gal-3 because siRNA-mediated knockdown of Gal-3 increased the levels of H2O2-induced apoptosis in pale ear cells to levels that were comparable with those in WT cells (Fig. 5A). In addition, overexpression of Gal-3 protected WT and pale ear fibroblasts from oxidant-induced cell death (Fig. 5B). Gal-3 mRNA levels were quantitated to confirm the effects of Gal-3 knockdown or overexpression (data not shown). In contrast, extracellular rGal-3 treatment did not have a similar effect on fibroblast apoptosis (Supplemental Fig. 2A). These results demonstrate that the heightened levels of Gal-3 that are seen in fibroblasts from pale ear mice protect them from oxidative stress–mediated cell injury.
To assess the role of Gal-3 on fibroblast proliferation, we incubated fibroblasts with and without TGF-β1. In these experiments, TGF-β1 increased the proliferation of fibroblasts from WT and pale ear mice (Fig. 5C, 5D). Importantly, the levels of TGF-β1–induced fibroblast proliferation were higher in cells from pale ear than cells from WT mice (Fig. 5C, 5D). These effects were mediated by intracellular Gal-3 because siRNA-mediated knockdown of Gal-3 decreased the levels of TGF-β1–induced proliferation in pale ear cells to levels that were comparable with those in WT cells (Fig. 5C). In addition, overexpression of Gal-3 exaggerated the TGF-β1–driven proliferation of pale ear cells (Fig. 5D). Gal-3 mRNA levels were quantitated to confirm the effects of Gal-3 knockdown or overexpression (data not shown). In combination, these results demonstrate that the heightened accumulation of Gal-3 in cells from pale ear mice augments their proliferative response to TGF-β1.
Studies were also undertaken to define the role(s) of intracellular Gal-3 in myofibroblast differentiation. When the levels of expression of α-SMA were used as a surrogate for myofibroblast differentiation and accumulation, TGF-β1 was a powerful stimulator of α-SMA in WT and pale ear cells; the pale ear response exceeded that in the WT cells (Fig. 5E). Gal-3 knockdown reduced the levels of mRNA encoding α-SMA in pale ear cells to levels that were comparable with those in WT cells (Fig. 5E). In addition, overexpression of Gal-3 augmented the ability of TGF-β1 to stimulate α-SMA mRNA accumulation in WT and pale ear cells (Fig. 5F). In contrast, extracellular rGal-3 treatment did not augment the ability of TGF-β1 to stimulate α-SMA mRNA accumulation (Supplemental Fig. 2B). Consistent with the findings with α-SMA expression, Gal-3 knockdown reduced the levels of vimentin expression in pale ear and WT cells (Supplemental Fig. 3A), and overexpression of Gal-3 augmented the ability of TGF-β1 to stimulate vimentin expression (Supplemental Fig. 3B). These studies demonstrate that intracellular Gal-3 augments TGF-β1–induced myofibroblast differentiation in WT and HPS1-deficient fibroblasts.
Intracellular Gal-3 regulates macrophage differentiation in vitro
Because of the Gal-3 abnormalities that were seen in macrophages, we next characterized the effects of Gal-3 on macrophage responses. In the first experiments, macrophages were incubated with H2O2, and cell death was assessed using TUNEL assessments. In these experiments, extracellular and intracellular Gal-3 did not appear to play a major role in the induction or regulation of macrophage cell death responses because WT and pale ear macrophages exhibited similar levels of TUNEL+ cells after H2O2 treatment (Fig. 6A), and the siRNA-mediated knockdown or the overexpression of Gal-3 did not reveal differences in the survival of these cell populations (Fig. 6A, 6B). Similarly, exogenous rGal-3 treatment did not have any effect on the survival of these cells (Supplemental Fig. 4A). Thus, Gal-3 does not regulate oxidant-induced macrophage apoptosis.
Previous studies from our laboratory and others have highlighted an association between M2 macrophages and pulmonary fibrosis, and have demonstrated that CHI3L1 is a prominent stimulator of M2 macrophage differentiation (29, 32). To address the role(s) of Gal-3 in these responses, we characterized the ability of CHI3L1 to induce macrophage CD206 expression (31). In keeping with previous findings from our laboratory, rCHI3L1 was a potent stimulator of CD206 expression by macrophages from WT and pale ear mice (Fig. 6C, 6D). The macrophages from the pale ear mice manifested exaggerated M2/CD206 responses when compared with the cells from WT animals (Fig. 6C). These effects were mediated by intracellular Gal-3 because siRNA-mediated knockdown of Gal-3 decreased the levels of CHI3L1-induced CD206 expression in pale ear cells to levels that were comparable to those in WT cells (Fig. 6C). In addition, overexpression of Gal-3 exaggerated the CHI3L1-driven expression of CD206 in WT and pale ear cells (Fig. 6D). Gal-3 mRNA levels were quantitated to confirm the effects of Gal-3 knockdown or overexpression (data not shown). Exogenous rGal-3 treatment did not modulate M2/CD206 responses (Supplemental Fig. 4B). Thus, CHI3L1 promotes M2 macrophage differentiation in a Gal-3–dependent manner.
Gal-3 diminishes the ability of CHI3L1 to activate antiapoptotic signaling pathways
Previous studies from our laboratory demonstrated that CHI3L1 mediates its antiapoptotic effects, at least in part, by activating MAPK kinase and or AKT signaling pathways that decrease apoptosis (28, 30). Because the present studies demonstrate that Gal-3 is a prominent stimulator of epithelial apoptosis, and previous studies demonstrated that CHI3L1 has profound antiapoptotic effects, studies were undertaken to determine whether Gal-3 could abrogate the antiapoptotic effects and signaling of CHI3L1. In these experiments, we treated MLE12 cells with rGal-3 in the presence or absence of rCHI3L1. In accord with our findings using alveolar type 2 cells (described earlier), rGal-3 was a potent stimulator of apoptosis and CHI3L1 was a potent inhibitor of this response (Fig. 7A). To define the intracellular signaling events associated with these effects, we investigated the effects of rGal-3 on CHI3L1-induced activation of ERK1/2 and AKT using MLE12 cells. In Fig. 7B, ERK and AKT were activated by rCHI3L1, and the activation of ERK and Akt were diminished (Fig. 7B) in cells cotreated with rCHI3L1 and rGal-3. These studies demonstrate that Gal-3 inhibits the antiapoptotic effects of CHI3L1 and CHI3L1-induced antiapoptotic signaling.
Gal-3 plays a critical role in CHIL1-induced Wnt/β-catenin signaling in macrophages
Previous studies from our laboratory demonstrated that CHI3L1 uses IL-13Rα2 to activate Wnt/β-catenin signaling (28). Therefore, we used murine peritoneal macrophages from Gal-3 null mutant mice to evaluate the roles of Gal-3 in this signaling response. These experiments demonstrated that CHI3L1 activated Wnt/β-catenin in murine peritoneal macrophages, and that these responses were significantly decreased in the absence of Gal-3 (Fig. 7C). Exogenous rGal-3 treatment did not have a major effect on CHI3L1 activation of Wnt/β-catenin signaling in murine peritoneal macrophages (Supplemental Fig. 4C). Together, these results demonstrate that Gal-3 plays a critical role in CHI3L1-induced Wnt/β-catenin signaling in lung macrophages.
Gal-3 binds to IL-13Rα2 and CHI3L1
Because Gal-3–induced apoptosis was associated with decreased CHI3L1-induced, chitosome-mediated antiapoptotic signaling, studies were undertaken to determine whether Gal-3 altered chitosome structure and/or function. Because antiapoptotic CHI3L1 signaling requires IL-13Rα2 and TMEM219 (30), studies were first undertaken to determine whether Gal-3 bound to either moiety and/or altered the aggregation of the IL-13Rα2–TMEM219 receptor complex. Coimmunoprecipitation (Co-IP) evaluations were undertaken with lung lysates from WT and IL-13 Tg mice because IL-13 stimulates IL-13Rα2 and Gal-3 (28, 42). These studies demonstrated that Gal-3 and IL-13Rα2 physically bind to one another because the immunoprecipitation of one always brought down the other (Fig. 8A). Co-IP assays were also used to evaluate the interactions between CHI3L1 and Gal-3. In these experiments, A549 cells were transfected with plasmids containing full-length cDNA encoding Myc-tagged Gal-3 or CHI3L1, lysates were generated, and Co-IP was undertaken. These studies demonstrated that Gal-3 also physically binds to CHI3L1 (Fig. 8B). Similar results were obtained in Co-IP experiments using lung lysates from WT and IL-13 Tg mice (Fig. 8C). In addition, in Biacore SPR evaluations, Gal-3 binds IL-13Rα2 in a 1:1 model with an ∼1 nM equilibrium dissociation constant (KD) value (data not shown). In addition, Gal-3 binds CHI3L1 with moderate affinity (KD = ∼94 nM) with biphasic curve fitting (data not shown). In accord with these findings, colocalization studies using multiple-label IHC demonstrated that Gal-3 and CHI3L1 were similarly expressed in the cytoplasm and to a lesser degree in the nuclei of the lung cells (Fig. 8D), many of which were CD68+ macrophages (Fig. 8E). In combination, these studies demonstrate that Gal-3 physically interacts with IL-13Rα2 and CHI3L1.
Gal-3 competes with TMEM219 for IL-13Rα2
Previous studies from our laboratory demonstrated that optimal CHI3L1-induced antiapoptotic signaling requires CHI3L1 interaction with IL-13Rα2, which, in turn, recruits TMEM219 to the multimeric complex (30). Thus, studies were undertaken to determine whether Gal-3 altered the binding of IL-13Rα2 and TMEM219. Hence co-IP experiments were undertaken using recombinant proteins. These studies confirmed our previous findings that Gal-3 binds to CHI3L1 (Fig. 9A) and IL-13Rα2 (Fig. 9B). Importantly, Gal-3 competed with TMEM219 for IL-13Rα2 binding; rhGal-3 diminished the binding between TMEM219 and IL-13Rα2 (Fig. 9C). These findings demonstrate that Gal-3 can abrogate the formation of the IL-13Ra2/TMEM219 chitosome complex and, in so doing, diminish chitosome antiapoptotic signaling.
HPS is an autosomal recessive genetic disorder associated with highly penetrant pulmonary fibrosis that develops in the fourth and fifth decades of life and is the leading cause of death in HPS1 and HPS4 patients (2–4, 6–10, 31). HPS patients can also experience many other clinical manifestations including oculocutaneous albinism and a bleeding diathesis (13, 43). Studies of HPS patients have implicated the abnormal trafficking of melanosome in the pigment-related dermal and ocular abnormalities, and abnormal formation of platelet granules in the bleeding diathesis (13, 43–47). However, despite these conceptual advances that highlight the importance of LRO and vesicle trafficking, the pathogenesis of the pulmonary fibrosis in HPS patients is poorly understood. Specifically, the LRO-related mechanisms that underlie the exaggerated injury and fibroproliferative repair responses in HPS have not been defined, and therapeutics have not been identified to alter these survival-limiting responses. Previous studies from our laboratories demonstrated that the levels of Gal-3 are increased in the lungs and BAL fluids from patients with HPS, and that Gal-3 traffics abnormally in dermal fibroblasts from HPS patients prone to lung fibrosis (27). However, the contributions that Gal-3 makes to the pathogenesis of HPS have not been defined.
To address these issues, we characterized the roles of Gal-3 in the pulmonary fibrosis responses of HPS-1–deficient pale ear mice. These studies demonstrate that the levels of Gal-3 are increased in BAL fluids from pale ear mice, and that Gal-3 traffics abnormally and accumulates in exaggerated quantities intracellularly in fibroblasts and macrophages from HPS-1–deficient animals. They also demonstrate that Gal-3 induces epithelial apoptosis when in the extracellular space, inhibits apoptosis in the intracellular space, and stimulates fibroblast proliferation, myofibroblast differentiation and accumulation, and M2-like macrophage differentiation in fibroblast and macrophages, respectively. Lastly, these studies also demonstrate that Gal-3 physically binds to CHI3L1 and IL-13Rα2, and competes with TMEM219 for binding to IL-13Rα2, which abrogates CHI3L1-induced IL-13Rα2–mediated antiapoptotic signaling and increases Wnt/β-catenin signaling. In combination, these studies demonstrate that the dysregulation of Gal-3 plays a key role in the pathogenesis of HPS, contributing in a variety of ways to the exaggerated injury response and exaggerated fibroproliferative repair response that characterize this disorder by interacting with CHI3L1 and its receptors in a compartment-specific manner.
Studies from our laboratory and others have demonstrated that injury and apoptosis are prerequisites for the development of fibrosis and tissue remodeling (29, 31, 41). In keeping with this well-documented paradigm, epithelial apoptosis and enhanced fibroproliferative repair are well-known to coexist in tissues from patients with diseases like idiopathic pulmonary fibrosis and HPS pulmonary fibrosis (19, 36, 37, 48–50). In contrast with fibrotic disorders such as radiation pneumonitis, in which the initiating injury is known, the fibrosis in HPS occurs spontaneously. This is likely the result of the well-documented enhanced sensitivity of cells and tissues from these patients and their murine models to apoptosis-inducing stimuli (31, 37, 48). The present studies contribute to our understanding of the mechanisms that underlie this enhanced sensitivity by highlighting multiple mechanisms that likely contribute to these exaggerated cell death responses. Specifically, they demonstrate that extracellular Gal-3 is a potent inducer of epithelial apoptosis. This is in keeping with prior studies that showed that extracellular Gal-3 induces apoptosis by activating mitochondrial apoptosis pathways (23). Second, they demonstrate that Gal-3 physically binds to the antiapoptotic ligand CHI3L1 and its receptor IL-13Rα2, and blocks TMEM219 binding to IL-13Rα2. The importance of IL-13Rα2 and the binding of TMEM219 to IL-13Rα2 in CHI3L1-induced antiapoptotic effector and signaling responses was previously described by our laboratory (28, 30). These Gal-3 interactions decrease CHI3L1 activation of antiapoptotic signaling pathways such as Erk1/2 and Akt/Protein Kinase B. Prior studies from our laboratory also demonstrated that there is a trafficking defect of IL-13Rα2 in epithelial cells in HPS-1–deficient mice (31). This likely exaggerates the apoptosis-inducing effects of Gal-3 by limiting the amount of IL-13Rα2 that can combine with TMEM219 to form antiapoptotic signaling chitosome complexes in HPS epithelial cells. Interestingly, the present studies also demonstrate that CHI3L1 is also a potent stimulator of Gal-3. One can speculate that injury induces CHI3L1, which, in turn, increases the levels of extracellular Gal-3 that enhances the sensitivity of HPS epithelial cells to apoptotic stimuli directly and by interacting with CHI3L1 and its receptor components.
Tissue fibrosis is characterized by a decrease in epithelial cells, increased levels of epithelial cell apoptosis, and the accumulation of myofibroblasts that show remarkable resistance to apoptotic stimuli (51, 52). Tissue fibrosis is also associated with the accumulation of M2-like macrophages whose contributions to the pathogenesis of tissue fibrosis are well documented (53, 54). The mechanisms by which epithelial cells are induced to undergo apoptosis, whereas nearby myofibroblasts become apoptosis resistant, have long puzzled students of pulmonary fibrosis (55, 56). Nor have the mechanisms by which macrophages undergo M2-like differentiation been adequately defined. The present studies advance our understanding of the mechanisms that likely underlie these responses in HPS and possibly other fibrotic disorders. As noted earlier, they demonstrate that Gal-3 traffics abnormally and accumulates in fibroblasts and macrophages but is not produced in detectable quantities by epithelial cells in HPS-1–deficient mice. They also show that intracellular Gal-3 is a potent inhibitor of fibroblast apoptosis, while stimulating fibroblast proliferation and myofibroblast differentiation. This is in keeping with the known sequence similarities between Gal-3 and the apoptosis inhibitor Bcl-2 (22). Furthermore, the current studies demonstrate that Gal-3 is a potent stimulator of M2 macrophage differentiation in vivo and in vitro. One can easily see how these effects can contribute to the accumulation of myofibroblasts, as well as the M2 macrophages in HPS and the profibrotic effects of Gal-3 that have been reported in the liver (24), kidney (25), lung (20), and in scleroderma (26).
CHI3L1 is a member of the 18 glycosyl hydrolase gene family that binds to but does not degrade chitin polysaccharides. It is present in the circulation of normal individuals and is present in exaggerated quantities in the circulation of patients with a variety of diseases, many of which are characterized by inflammation, injury, repair, and/or tissue remodeling (32, 33, 57–59). Recent studies from our laboratory and others have demonstrated that CHI3L1 plays critical roles in antipathogen, Ag-induced, and oxidant-mediated responses by regulating essential biologic responses including cell death, inflammasome activation, Th1/Th2 cytokine balance, macrophage differentiation and accumulation, and Erk, Akt and Wnt/β-catenin signaling (32, 33, 57, 60–63). In many of these settings, CHI3L1 is likely produced as a protective response based on its ability to simultaneously decrease epithelial apoptosis and stimulate fibroproliferative repair (29, 31). To understand the mechanisms by which CHI3L1 mediates these responses, CHI3L1 receptors/binding partners have been sought and characterized. These studies defined a multimeric receptor called the chitosome, which contains IL-13Rα2 and its coreceptor TMEM219 (28, 30). They also demonstrated that the ability of CHI3L1 to optimally inhibit epithelial cell death and activate Erk1/2 and Akt signaling requires IL-13Rα2 and TMEM219 with CHI3L1 binding to IL-13Rα2, which then attracts TMEM219 to the receptor complex (30). In contrast, CHI3L1 activates Wnt/β-catenin signaling via a pathway that requires IL-13Rα2 but is independent of TMEM219 (28, 30). The present studies demonstrate that Gal-3 binds to CHI3L1 and IL-13Rα2, and blocks the binding of TMEM219 to IL-13Rα2. They also indicate that these interactions decrease Erk and Akt signaling while augmenting Wnt/β-catenin signaling. At present, it is not entirely clear how these physical interactions occur. However, one could speculate that IL-13Rα2 is an α receptor subunit that can interact with two β-subunit coreceptors, TMEM219 and Gal-3. In this construct, CHI3L1-stimulated IL-13Rα2 would interact with TMEM219 to mediate antiapoptotic effector responses, whereas IL-13Rα2 interactions with Gal-3 would augment apoptotic responses and Wnt/β-catenin activation.
The present studies of Gal-3, along with prior studies of CHI3L1 and its receptors, have provided important new insights into the pathogenesis of pulmonary fibrosis in HPS. They also provide the basis upon which therapies for HPS-associated pulmonary fibrosis can be proposed. Specifically, one can easily see how interventions that decrease the production and/or accumulation of Gal-3 can be therapeutically useful in this disorder. One can also envision the development of effective combination therapies because Gal-3–based therapies may very well act in an additive or synergistic manner with interventions that augment membrane expression of IL-13Rα2 or block CRTH2. Additional investigations will be required to assess the utility of each of these approaches.
In summary, the present studies demonstrate that Gal-3 is increased in the extracellular space, traffics abnormally, and accumulates in lung fibroblasts and macrophages in mice that lack HPS-1. These abnormalities contribute to the pathogenesis of HPS pulmonary fibrosis, with the extracellular Gal-3 stimulating epithelial apoptosis and intracellular Gal 3 enhancing fibroblast survival and proliferation and myofibroblast and macrophage differentiation. Finally, these effects are mediated by the ability of Gal-3 to bind to CHI3L1 and IL-13Rα2, and block IL-13Rα2-TMEM219 binding, which decreases antiapoptotic and increases Wnt/β-catenin signaling. Additional investigation of the roles of Gal-3 and its interactions with CHI3L1 and its receptors in HPS and related diseases is warranted.
This work was supported by the American Thoracic Society/Hermansky-Pudlak Syndrome Network (to Y.Z.), National Institutes of Health Grant U54 GM115677 (to Y.Z. and B.S.S.), National Institutes of Health Pilot Grant P20 GM103652 (to Y.Z.), National Institutes of Health Grants R01 HL093017 and U01 HL108638 (to J.A.E.), National Institutes of Health Grant R01 HL115813 (to C.G.L.), and by the Intramural Research Program, National Human Genome Research Institute, National Institutes of Health (to B.R.G. and W.A.G.).
The online version of this article contains supplemental material.
Abbreviations used in this article:
biogenesis of LRO complex 3
chitinase 3-like 1
small interfering RNA
α-smooth muscle actin
surface plasmon resonance
The authors have no financial conflicts of interest.