Pseudomonas aeruginosa provokes a painful, sight-threatening corneal infection. It progresses rapidly and is difficult to treat. In this study, using a mouse model of P. aeruginosa keratitis, we demonstrate the importance of a carbohydrate-binding protein, galectin-8 (Gal-8), in regulation of the innate immune response. First, using two distinct strains of P. aeruginosa, we established that Gal-8−/− mice are resistant to P. aeruginosa keratitis. In contrast, mice deficient in Gal-1, Gal-3, and Gal-9 were fully susceptible. Second, the addition of exogenous rGal-8 to LPS (TLR4 ligand)–stimulated bone marrow-derived macrophages suppressed 1) the activation of the NF-κB pathway, and 2) formation of the MD-2/TLR4 complex. Additionally, the expression level of reactive oxygen species was substantially higher in infected Gal-8−/− bone marrow neutrophils (BMNs) compared with the Gal-8+/+ BMNs, and the P. aeruginosa killing capacity of Gal-8−/− BMNs was considerably higher compared with that of Gal-8+/+ BMNs. In the bacterial killing assays, the addition of exogenous rGal-8 almost completely rescued the phenotype of Gal-8−/− BMNs. Finally, we demonstrate that a subconjunctival injection of a Gal-8 inhibitor markedly reduces the severity of infection in the mouse model of P. aeruginosa keratitis. These data lead us to conclude that Gal-8 downmodulates the innate immune response by suppressing activation of the TLR4 pathway and clearance of P. aeruginosa by neutrophils. These findings have broad implications for developing novel therapeutic strategies for treatment of conditions resulting from the hyperactive immune response both in ocular as well as nonocular tissues.

This article is featured in Top Reads, p.357

Pseudomonas aeruginosa produces a fulminating, highly destructive, sight-threatening corneal infection in humans (1, 2). Contact lens wear is the prime risk factor (3–5). One of the serious consequences of P. aeruginosa corneal infection is blindness resulting from an overactive immune response leading to persistent inflammatory reaction in the stroma, corneal ulceration with subsequent corneal scarring, and, in some cases, perforation. The current therapy includes antibiotic treatment, which reduces the bacterial burden; despite this, tissue damage occurs, often rapidly, due to uncontrolled inflammation. Although the polymorphonuclear leukocyte (PMN)–predominant innate response is clearly vital in achieving bacterial eradication, the persistent dysregulation of innate responses is the primary cause for acute manifestations of P. aeruginosa keratitis. Because P. aeruginosa keratitis progresses rapidly, the development of effective strategies to control the disease at an early stage to prevent the uncontrolled immune response in the cornea is a high priority. The innate immune system is the first line of defense against pathogens and is initiated by pattern recognition receptors, which respond to invading microbes (6–13). The first step in the induction of innate immune response in the context of bacterial infections is generally achieved by TLR-mediated activation of the NF-κB pathway, which results in the induction of inflammatory cytokines that are secreted from the cell to mediate downstream inflammatory effects that clear the infection (14–16). TLR4 is critical in host resistance to P. aeruginosa keratitis, as its deficiency results in increased PMN infiltration and proinflammatory cytokine production, impaired bacterial killing, and a susceptible phenotype (17).

Galectin-8 (Gal-8) belongs to a tandem repeat–type subfamily of carbohydrate-binding proteins, galectins (18, 19). It contains two distinct carbohydrate recognition domains (CRDs), the N-terminal CRD (Gal-8N) and the C-terminal CRD (Gal-8C), linked by a linker domain. Carbohydrate-binding specificity of the two CRDs is distinct. A distinguishing feature of Gal-8N is its strong preference for α2,3-sialylated glycans. Very high affinity to α2,3-sialylated glycans is a unique feature of Gal-8N that sets Gal-8 apart from all other members of the galectin family (20, 21). Gal-8 is widely expressed in many organs and tissues under physiological or pathological conditions. It has been reported that Gal-8 activates antibacterial autophagy to defend cells against bacterial invasion (22). Depending on the context, Gal-8 has been shown to have both proinflammatory as well as anti-inflammatory properties (23). In a number of cell types, including osteoblasts, liver, spleen, and lungs, Gal-8 stimulates the secretion of various chemokines and cytokines, including TNF-α, IL-1β, MCP-1, and IL-6, and is thus proinflammatory (24). Other published studies using experimental models of autoimmune diseases including rheumatoid arthritis (25), uveitis (26), and encephalomyelitis (27) emphasize that Gal-8 exhibits anti-inflammatory effects. In some scenarios, Gal-8 acts as a proinflammatory molecule in selected resting cells of the immune system but displays anti-inflammatory properties when these cells become activated (23). Therefore, at present, whether Gal-8 restrains an exacerbated response or, to the contrary, fuels an ongoing inflammatory response is an open question. In the current study, we demonstrate that 1) Gal-8−/− mice are resistant to P. aeruginosa keratitis, 2) Gal-8 exacerbates the P. aeruginosa keratitis pathology by downmodulating the TLR4 pathway, and that 3) Gal-8 downmodulates the TLR4 pathway by binding to CD14 and inhibiting the formation of the MD-2/TLR4 complex. We further demonstrate that Gal-8–modulated signaling results in repression of reactive oxygen species (ROS) expression and reduction in the P. aeruginosa killing capacity of neutrophils. Finally, we explored the potential of pharmacological targeting of Gal-8 for treatment of P. aeruginosa keratitis and demonstrated that subconjunctival injection of Gal-8N inhibitor reduces the severity of P. aeruginosa keratitis in a mouse model.

Seven- to 8-wk-old female wild-type (WT, C57BL/6) mice were purchased from The Jackson Laboratory (Bar Harbor, ME). The Gal-8−/− mouse strain was created from an embryonic stem cell clone (14305A-F8), obtained from the KOMP Repository (https://www.komp.org) and generated by Regeneron Pharmaceuticals (28) on the C57BL/6N background. The Gal-8 null status of the knockout mice was confirmed by Western blotting (29). The Gal-8−/− mice on the C57BL/6N background were backcrossed at least 10 generations to the C57BL/6J background. Animals were housed in animal facilities approved by the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC) at Tufts University, and all experimental procedures were in complete agreement with the AAALAC guidelines. A cytotoxic strain, P. aeruginosa 6077, and a P. aeruginosa invasive strain, 6294, were used. Bacteria were cultured overnight on tryptic soy agar plates at 37°C before each infection. rGal-8 was produced and purified as previously described (30) with a few modifications. Briefly, lysates of bacteria expressing rGal-8 were chromatographed on a β-lactose–conjugated Sepharose column (EY Laboratories; 1-ml bed volume). The lectin was eluted from the column with a buffer containing 100 mM β-lactose and dialyzed against PBS containing 2% glycerol and 4 mM 2-ME and stored at −80°C. Endotoxin was removed from each preparation of lectin by Detoxin-Gel endotoxin removing gel (Thermo Scientific), and endotoxin levels were detected by a ToxinSensor chromogenic Limulus amebocyte lysate endotoxin assay kit (GenScript). Endotoxin levels of all Gal-8 preparations used in this study were <0.1 endotoxin units/µg. Before use, each preparation of Gal-8 was also tested for carbohydrate-binding activity and specificity by the RBC agglutination assay. To prepare the Gal-8 affinity matrix, 5 mg of rGal-8 was conjugated to 330 mg of Pierce N-hydroxysuccinimide–activated agarose dry resin in accordance with the manufacturer’s instructions (Thermo Scientific, Waltham, MA).

Corneas of WT and Gal-8−/− mice were infected under deep anesthesia induced by i.p. injection of ketamine and xylazine. Central corneas of mice were scarified with three parallel 1-mm incisions using a 26G needle, and a 5-μl drop containing 100 CFU of strain P. aeruginosa 6077 or P. aeruginosa 6294 was applied to the eye. The eyes were examined on day 1 and day 3 postinfection (p.i.) for the development of corneal keratitis and opacity. The degree of opacity was graded from 0 to 4, as previously described by Berk et al. (31): 0, eye macroscopically identical to the uninfected control eye; 1, partial corneal opacity covering the pupil; 2, dense corneal opacity covering the pupil; 3, dense opacity covering the entire anterior segment; and 4, perforation of the cornea, phthisis bulbi (shrinkage of the globe after inflammatory disease). To enumerate bacterial load, infected corneas of WT and Gal-8−/− mice were harvested, and each cornea was homogenized in 250 µl of tryptic soy broth with 0.05% Tween 20. Then, 10 µl of the corneal homogenate was serially diluted in tryptic soy broth with 0.05% Tween 20 and selected dilutions were plated in triplicate on Pseudomonas isolation BBL agar plates (Becton Dickinson, Franklin Lakes, NJ). Plates were incubated overnight at 37°C, and the numbers of viable bacteria were manually counted. Infiltration of PMNs in infected corneas was quantified by a myeloperoxidase assay (16). Briefly, corneas were harvested at day 1 p.i, homogenized in 1.0 ml of 50 mM phosphate buffer (pH 6.0) containing 0.5% hexadecyltrimethylammonium bromide (Sigma-Aldrich, St. Louis, MO), subjected to four freeze-thaw cycles, and sonicated in an ice bath (20 s). After centrifugation, 100 µl of the supernatant was added to 2.9 ml of 50 mM phosphate buffer containing o-dianisidine dihydrochloride (16.7 mg/100 ml, Sigma-Aldrich) and H2O2 (0.0005%). The change in absorbance was monitored at 460 nm (5 min at 30-s intervals), and the numbers of PMNs per cornea were calculated on the standard curve prepared using mouse bone marrow neutrophils (BMNs).

To determine whether Gal-8 influences bacterial growth, varying concentrations of P. aeruginosa were incubated with rGal-8 (0.75 µM) for 3 h at 37°C, and then bacterial counts were enumerated by plating on BBL agar plates as described in the previous section. To determine whether Gal-8 influences the adhesion of P. aeruginosa to the corneal surface, corneas of Gal-8−/− and WT mice were scarified and a 5-µl drop of bacteria (100 CFU, P. aeruginosa 6077) was applied to the cornea; at 2 and 4 h p.i., corneas were harvested and processed for bacterial enumeration.

P. aeruginosa–infected corneas were harvested from WT and Gal-8−/− mice on day 1 p.i., pooled groupwise (20–30 corneas/group), and digested in 200 µl of DMEM containing collagenase and DNase (100 µg/ml each, Roche Diagnostics, Indianapolis, IN) for 45 min at 37°C in a water bath. After incubation, the corneas were disrupted by grinding with a syringe plunger on a cell strainer, and single-cell suspensions were made in complete DMEM medium. For PMN sorting, cells were blocked with an unconjugated anti-CD32/CD16 mAb (clone 2.4G2, 30 min) in staining buffer (2% FBS, 5 mM EDTA in PBS) and then incubated for 30 min at 4°C with a mixture of CD45 (clone 30-F11), CD11b-PerCP (clone M170), and Ly6G (clone 1A8). All Abs were purchased from BD Bioscience (San Jose, CA). Then, cells were washed three times using PBS and incubated with allophycocyanin/fixable viability dye (1:1000, 30 min, 4°C), washed again, and resuspended for FACS sorting on a FACSAria cell sorter (Becton Dickinson). The yield of neutrophils was 1–1.5 × 105 cells/20 infected corneas.

FACS-sorted WT and Gal-8−/− neutrophils of infected corneas (1–1.5 × 105 cells) were lysed in buffer RLT (Qiagen) and were processed for transcriptome analysis by RNA sequencing (RNA-seq). Briefly, total RNA was isolated from cell lysates using a RNeasy mini kit (Qiagen) and quantified using an Agilent 5200 fragment analyzer system (Agilent Technologies, Santa Clara, CA). Yield of RNA was 1 ng/5000 neutrophils. All RNA samples had 28S/18S ratios between 1.8 and 2.5. Purified RNA samples (20–40 ng) were sent to Tufts University Genomics Core for library preparation and sequencing. The Ovation RNA-seq system (NuGEN Technologies) was used to construct cDNA libraries from total RNA, and cDNA libraries were multiplexed with three samples per lane and loaded onto flow cell lanes. Sequencing-by-synthesis of 51-nt length was performed on an Illumina HiSeq 2500 sequencing system (NuGEN Technologies). DESeq2 was used to normalize the expression values of the transcript isoforms. A stringent filtering criterion of raw reads per kilobase of transcript per million mapped reads value of 50 in at least one sample was used to obtain expressed transcripts. ANOVA was then performed on the log-transformed data to generate fold change between samples collected from different experimental conditions and the p value for each transcript. Differentially expressed mRNA isoforms were filtered at a cutoff of >50 reads with a fold change >1.5 and p < 0.05.

Protein extracts of infected corneas of WT and Gal-8−/− mice were prepared in a radioimmunoprecipitation assay (RIPA) buffer supplemented with a protease inhibitor mixture (cOmplete tablets; Roche Applied Science, Mannheim, Germany) and 2% SDS. For preparation of tissue lysates, three to four corneas were pooled and considered one biological replica. Aliquots of lysates containing 10–30 μg of proteins were subjected to electrophoresis in a 4–15% gradient or 12% SDS-PAGE gels (Bio-Rad, Hercules, CA) and transferred to a nitrocellulose membrane. The blots were blocked with Odyssey blocking buffer (OBB; LI-COR Biosciences, Lincoln, NE) and incubated with mouse anti-NLRP3 (clone Cryo-1, 1:1000, AdipoGen, San Diego, CA), anti–caspase-1 (clone Casper-1, 1:1000; AdipoGen), and goat anti–IL-1β (clone AF-401-NA, 1:1000, R&D Systems, Minneapolis, MN) primary Abs in OBB (4°C, overnight). The secondary Abs used were anti-goat IgG IRDye 800CW and anti-mouse IgG IRDye 680LT (LI-COR Biosciences) diluted in OBB (1:10,000, 25°C, 45 min). Blots were then scanned with the Odyssey infrared imaging system using Image Studio v3.0 software (LI-COR Biosciences). After image acquisition, the blots were stripped using the NewBlot nitrocellulose stripping buffer (LI-COR Biosciences) and reprobed with mouse anti–β-actin (clone AC-15, 1:10,000, Santa Cruz Biotechnology, Dallas, TX) as a primary Ab, and anti-mouse IgG IRDye 680LT (LI-COR Biosciences) as a secondary Ab.

Bone marrow–derived macrophages (BMDMs) were prepared by maturing bone marrow cells for 6–7 d in DMEM medium containing 10% FBS, 2 mM l-glutamine, 1% penicillin/streptomycin, 1 mM sodium pyruvate, 50 µM 2-ME (Life Technologies), and 10% L929 conditioned medium as a source of CSF (32). More than 99.5% of the cells prepared by this procedure were CD11b+F4/80+ macrophages as characterized by flow cytometry analysis.

To detect the effect of exogenous rGal-8 on the expression levels of the key components of the TLR4 pathway, mature BMDMs in 100-mm petri dishes were serum starved in 5 ml of DMEM for 2–3 h, rGal-8 (20 µg/ml) was added, and, after a 30-min incubation with Gal-8, cells were stimulated with LPS (100 ng/ml Escherichia coli/LPS, Invitrogen) for 30 min at 37°C. At the end of incubation period, cells were washed, lysed in RIPA buffer, and cell lysates were subjected to Western blot analysis to detect the expression levels of phospho-IκBα using mouse anti–phospho-IκBα (clone 5A5, Cell Signaling Technology) as a primary Ab. To determine whether the effect of Gal-8 on the innate immune response is carbohydrate-dependent, BMDMs were incubated with Gal-8 in the presence of 0.1 M β-lactose, α2,3-sialyl lactose (3′SL), or α2,6-sialyl lactose (6′SL). To detect the effect of exogenous rGal-8 on secretion of IL-6 and TNF-α, mature BMDMs were stimulated with LPS in the presence of rGal-8 as described above except that the cells were stimulated with LPS for 4 h and nigericin (10 µM, Sigma-Aldrich) was present in the media during the last 45 min of incubation. At the end of the incubation period, IL-6 and TNF-α released into culture supernatant were measured in triplicates by using sandwich ELISA kits (R&D Systems, Minneapolis, MN).

To determine whether Gal-8 binds to LPS, wells of 96-well flat-bottom plates (Costar) were coated with rGal-8 (2.5–20 µg/ml PBS, 2 h, room temperature), and nonspecific binding sites were blocked with 1% BSA in PBS (1 h, room temperature). Biotinylated LPS (Invitrogen, 50 µl, 1 µg/ml PBS containing 1% BSA) was added to the galectin-coated wells, and plates were incubated at room temperature for 1 h and then washed with PBS containing 0.05% Tween 20. Streptavidin-HRP (50 µl, 1:40 dilution, R&D Systems) was added to each well, and plates were incubated for 20 min, washed, and air dried. Tetramethylbenzidine (TMB) chromogenic substrate (50 µl/well, R&D Systems) was added to each well for HRP detection, reactions were stopped by 1 N H2SO4 (50 µl/well), and plates were read in a microplate reader (FilterMax F5 multimode, Molecular Devices). Wells coated with mouse CD14 (10 ng/ml, R&D Systems), a well-known LPS-binding protein, served as a positive control. In some experiments, wells were coated with Gal-3 for comparison purposes.

To determine whether CD14 and TLR4 are Gal-8–binding proteins, matured BMDMs in 100-mm petri dishes were stimulated with LPS (E. coli-LPS, Invitrogen, 100 ng/ml DMEM, 2 h), washed with PBS, and then lysed in 0.5 ml of RIPA buffer supplemented with a protease inhibitor mixture. Protein concentrations of the lysates were adjusted to 1 mg/ml, 500 µg of a protein aliquot was precleared by incubation with 50 µl of unconjugated agarose beads (2–4 h at 4°C), and supernatants were collected and incubated overnight with Gal-8–conjugated agarose beads (50 µl at 4°C). Then, beads were washed with PBS three times by centrifugation, bound proteins were eluted by boiling the beads in 30 µl of Laemmli sample buffer for Western blot analysis using anti-TLR4 (1:100 dilution, Santa Cruz Biotechnology), and anti-CD14 (1:1000 dilution, clone H-4, Cell Signaling Technology) as primary Abs. Secondary Abs used were anti-rabbit IgG IRDye 800CW and anti-mouse IgG IRDye 680LT (LI-COR Biosciences) diluted in OBB (1:10,000, 25°C, 45 min). Controls included incubation of lysates with Gal-8 affinity beads in the presence of β-lactose (100 mM, a pan inhibitor of galectins) or sucrose (100 mM, a noncompeting sugar).

Because CD14 is the key LPS-binding protein on BMDMs, to determine the effect of Gal-8 on LPS binding to CD14, serum-starved BMDMs in 100-mm petri dishes were treated with rGal-8 (20 μg/ml in DMEM) for 30 min, FITC-LPS (Sigma-Aldrich, 100 ng/ml in DMEM) was added, and plates were incubated for an additional 1 h at 37°C. Plates were then washed three times with cold PBS to remove any traces of unbound LPS, stained with a viability dye, and then processed for flow cytometry to detect bound FITC-LPS. Cells were acquired by a BD LSR II, and data were analyzed by FlowJo.

BMDMs pretreated with rGal-8 (20 µg/ml in DMEM) for 30 min as described above were incubated with FITC-LPS (Sigma-Aldrich, 100 ng/ml in DMEM, 1 h, 37°C). Cells were then washed twice with cold PBS, sequentially stained with PE/MD-2/TLR4 complex Ab (dilution 1:100, clone MTS510, eBioscience) and a viability dye, and then processed for flow cytometry.

BMNs were isolated as described by Swamydas and Lionakis (33). Briefly, femurs and tibias of WT and Gal-8−/− mice were collected; bone marrow was flushed over a cell strainer (70 μm). Cells were washed and resuspended in 3 ml of DMEM and loaded slowly over two layers of Histopaque 1119 and 1077 (Sigma-Aldrich), respectively, and centrifuged at 2500 rpm (25 min, room temperature). Neutrophils were collected from the interface of the Histopaque 1119 and Histopaque 1077 layers. More than 95% of the cells collected from the interface were CD11b+Ly6G+ neutrophils as characterized by flow cytometry analysis.

To determine whether Gal-8 modulates ROS expression in BMNs, WT and Gal-8−/− BMNs (2 × 105 cells) were incubated in 100 µl of 1× ROS label in ROS assay buffer (BioVision, Milpitas, CA; 30 min, 37°C, protected from light) and were then cocultured with P. aeruginosa 6294 (P. aeruginosa/BMN ratio of 1:1, 1:5, and 1:10). After a 2-h coculture period, fluorescence intensity of ROS was measured by flow cytometry. Cells were acquired by BD LSR II and data analyzed by FlowJo. BMNs incubated with 1× ROS inducer (BioVision, Milpitas, CA) for 1 h before flow cytometry served as a positive control. For rescue experiments, the assays were performed using Gal-8−/− BMNs in the presence of rGal-8 (5–20 µg/ml Gal-8).

To assess the role of Gal-8 in the modulation of bacterial killing capacity of neutrophils, BMNs prepared from WT and Gal-8−/− mice were seeded in triplicates in a 48-well plate (5 × 104 cells/well in 250 µl of DMEM + 10% FBS without antibiotics). Cells were allowed to adhere to wells (1 h at 37°C in a CO2 incubator) and were then cocultured with P. aeruginosa 6294 at a BMN/P. aeruginosa ratio of 1:1, 1:5 and 1:10 (in 250 µl, 4 h). At the end of the incubation period, cells were homogenized in the same media, and homogenates were processed for bacterial enumeration as described above. For rescue experiments, Gal-8−/− BMNs were preincubated with an ascending concentration of exogenous rGal-8 (10–20 µg/ml, 30 min, 37°C) and coculture incubations were carried out in the presence of rGal-8.

To examine the effect of Gal-8 inhibition on the outcome of keratitis, a Gal-8 inhibitor (3,4-dichlorophenyl 3-O-(5-carboxy-1-methyl-[1H]-benzo[d]imidazole-2-ylmethyl)-1-thio-α-d-galactopyranoside), referred to hereafter as 19a, was used (34). This inhibitor is a benzimidazole-derivatized galactoside designed and synthesized based on the crystal structure of Gal-8 that selectively inhibits the Gal-8 N-terminal domain (Kd of 1.8 µM), with a weaker affinity for Gal-3 (Kd of 5 µM) and little affinity for Gal-1, Gal-2, Gal-4, Gal-7, and Gal-9 (34). To examine the effect of Gal-8 inhibition, corneas of two groups of WT mice were infected with P. aeruginosa (five corneas/group). The control group received a subconjunctival injection of 10 μl of vehicle (PBS + 23% DMSO), and the experimental group received a 10-μl injection of a Gal-8 inhibitor, 19a (5 mg/ml in vehicle), immediately prior to P. aeruginosa infection. The eyes were examined on day 1 p.i., the degree of keratitis was recorded, and the corneas were harvested 1) individually for bacterial enumeration by plating the homogenate on agar plates and PMN enumeration using the myeloperoxidase assay as described earlier, or 2) pooled groupwise (five corneas/group) and processed for quantification of IL-1β. Briefly, corneas were homogenized in 500 µl of PBS containing 1% BSA and a protease inhibitor mixture (Roche Diagnostics, Indianapolis, IN) and centrifuged at 12,000 × g for 5 min. A 100-μl aliquot of supernatant was assayed in triplicate by ELISA to quantify IL-1β, using a DuoSet kit (R&D Systems, Minneapolis, MN).

Differences between two groups were compared using two-tailed unpaired t tests using GraphPad Prism software (version 6.05). Data in graphs represent mean ± SEM. A p value of <0.05 was considered significant.

To determine whether Gal-8 has an impact on the course of P. aeruginosa keratitis, central corneas of Gal-8−/− and WT (C57BL/6) mice were infected with P. aeruginosa, and at day 1 p.i., the severity of bacterial keratitis was graded by slit lamp and then corneas were harvested for bacterial and PMN enumeration. As expected, WT mouse corneas developed severe keratitis (Fig. 1A); the degree of keratitis was reduced in Gal-8−/− mice as evident from a substantial decrease in opacity score (Fig. 1A), bacterial load (Fig. 1B), and PMN infiltration (Fig. 1C). Similar results were obtained on day 3 p.i. (Supplemental Fig. 1). Also, similar results were obtained regardless of whether the corneas were infected with the cytotoxic strain P. aeruginosa 6077 (Fig. 1) or an invasive strain P. aeruginosa 6294 (Supplemental Fig. 2) in four independent experiments. Additional studies revealed that rGal-8 did not influence bacterial growth (Supplemental Fig. 3A) or adhesion of bacteria to the abraded corneas of WT and Gal-8−/−mice (Supplemental Fig. 3B). Thus, the resistance of Gal-8−/− mice to P. aeruginosa keratitis is not due to the ability of Gal-8 to facilitate bacterial growth or bacterial adhesion, nor is it likely to be due to the ability of Gal-8 to facilitate bacterial entry into the cells because similar results were obtained regardless of whether the infection was induced with the extracellular cytotoxic strain P. aeruginosa 6077 or an invasive strain P. aeruginosa 6294.

FIGURE 1.

Gal-8 knockout mice are resistant to P. aeruginosa keratitis. Central corneas of Gal-8−/− and WT (C57BL/6) mice were challenged with P. aeruginosa and the severity of bacterial keratitis was graded on day 1 postinfection (p.i.) using a scoring system ranging from 0 to 4 to grade the degree of keratitis as described in Materials and Methods, and then corneas were harvested for bacterial and PMN enumeration. (Ai) Representative eye images comparing the disease severity of WT and Gal-8−/− corneas. (Aii) Corneal opacity score of Gal-8−/− and WT mice on day 1 p.i. (B) Bacterial load and (C) PMN counts in Gal-8−/− and WT mice at day1 p.i. are shown. Data are plotted as mean ± SEM. In (Aii) and (B), combined results of three independent experiments are shown (n = 18). In (C), combined results of two independent experiments are shown (n = 20). Statistical levels of significance were analyzed by a Student t test. ***p < 0.001 versus WT.

FIGURE 1.

Gal-8 knockout mice are resistant to P. aeruginosa keratitis. Central corneas of Gal-8−/− and WT (C57BL/6) mice were challenged with P. aeruginosa and the severity of bacterial keratitis was graded on day 1 postinfection (p.i.) using a scoring system ranging from 0 to 4 to grade the degree of keratitis as described in Materials and Methods, and then corneas were harvested for bacterial and PMN enumeration. (Ai) Representative eye images comparing the disease severity of WT and Gal-8−/− corneas. (Aii) Corneal opacity score of Gal-8−/− and WT mice on day 1 p.i. (B) Bacterial load and (C) PMN counts in Gal-8−/− and WT mice at day1 p.i. are shown. Data are plotted as mean ± SEM. In (Aii) and (B), combined results of three independent experiments are shown (n = 18). In (C), combined results of two independent experiments are shown (n = 20). Statistical levels of significance were analyzed by a Student t test. ***p < 0.001 versus WT.

Close modal

Next, we investigated whether Gal-8 has the capacity to influence the immune response. Because TLR4 plays a critical role in host resistance to P. aeruginosa keratitis (17), we first examined whether the expression of genes related to the TLR4 pathway is altered in Gal-8−/− neutrophils of infected corneas. Briefly, FACS-sorted WT and Gal-8−/− neutrophils of infected corneas were processed for RNA-seq and pathway analysis by Ingenuity Pathway Analysis software (Qiagen). Analysis of the data revealed that many genes of the innate immune pathway including NLRP3, NF-κB, TNF, and CD14 were upregulated in Gal-8−/− neutrophils (Fig. 2A). Next, we performed studies to determine whether the expression levels of key components of the TLR4 pathway are upregulated in infected corneas of Gal-8−/− mice. Western blot analysis of infected corneas of Gal-8−/− and WT mice, harvested on day 1 p.i., revealed that the expression levels of various components known to be upregulated upon activation of the TLR4 pathway including pro–IL-1β, NLRP3, and pro–caspase-1 are substantially higher in infected Gal-8−/− corneas compared with corresponding WT mouse corneas (Supplemental Fig. 4). These data suggest that activation of the TLR pathway is augmented in infected Gal-8−/− corneas and that Gal-8 dampens the innate immune response. Next, to further confirm the role of Gal-8 in downmodulation of the innate immune response, mature BMDMs were stimulated with 100 ng/ml LPS (TLR4 ligand) in the presence and the absence of exogenous rGal-8 (20 µg/ml). We found that exogenous rGal-8 downregulated phosphorylation of IκBα (Fig. 2B), a key readout to assess activation of the NF-κB pathway. A critical outcome of TLR4 pathway activation is the secretion of inflammatory cytokines. Specifically, TNF-α and IL-6 expression are key readouts of activation of the TLR4 pathway. To conclusively establish the suppressive role of Gal-8 on activation of the TLR4 pathway, additional studies were designed to determine whether treatment with exogenous rGal-8 will diminish secretion of TNF-α and IL-6 in BMDMs. In one study, mature BMDMs were serum starved overnight and were then stimulated with 100 ng/ml LPS (TLR4 ligand) in the presence and the absence of exogenous rGal-8 (4 h, 37°C). At the end of the incubation period, levels of key inflammatory cytokines released into culture supernatants were measured by ELISA. Indeed, the addition of exogenous rGal-8 markedly reduced the secretion of TNF-α and IL-6 (Fig. 2C). The inhibitory effect of Gal-8 on LPS-induced activation of the TLR pathway was specifically counteracted by 3′SL, a Gal-8N competing saccharide, but not by 6′SL, a Gal-8 noncompeting saccharide (Fig. 2B, 2C), suggesting that the carbohydrate-dependent function of Gal-8N is directly involved in the inhibitory effect of Gal-8 on activation of the TLR4 pathway.

FIGURE 2.

Gal-8 dampens the innate immune response. (A) The expression levels of genes related to the innate immune pathway are upregulated in Gal-8−/− neutrophils of infected corneas. FACS-sorted WT and Gal-8−/− neutrophils of infected corneas were processed for RNA-seq and pathway analysis by Ingenuity Pathway Analysis software (Qiagen). Histogram depicts upregulated expression of genes related to the innate immune response. Combined results of two independent experiments are shown (20 corneas/group). (B) Gal-8 negatively regulates TLR4 signaling. Mature BMDMs were serum starved overnight and were then stimulated with 100 ng/ml LPS (TLR4 ligand) in the presence and absence of exogenous rGal-8 (20 µg/ml) or rGal-8 + control saccharides for 30 min at 37°C. Electrophoresis blots of cell lysates were processed for Western blot analysis using Abs to phospho-IκBα. Representative blots of three independent experiments done in duplicates are shown. (C) Exogenous rGal-8 diminishes secretion of TNF-α and IL-6 in BMDMs. Mature BMDMs were serum starved overnight and were then stimulated with 100 ng/ml LPS in the presence and absence of exogenous rGal-8 (20 µg/ml). At the end of the incubation period, TNF-α and IL-6 released into culture supernatants were measured by ELISA (n = 4). Data are plotted as mean ± SEM. Statistical levels of significance were analyzed by a Student t test. *p < 0.05, **p < 0.01.

FIGURE 2.

Gal-8 dampens the innate immune response. (A) The expression levels of genes related to the innate immune pathway are upregulated in Gal-8−/− neutrophils of infected corneas. FACS-sorted WT and Gal-8−/− neutrophils of infected corneas were processed for RNA-seq and pathway analysis by Ingenuity Pathway Analysis software (Qiagen). Histogram depicts upregulated expression of genes related to the innate immune response. Combined results of two independent experiments are shown (20 corneas/group). (B) Gal-8 negatively regulates TLR4 signaling. Mature BMDMs were serum starved overnight and were then stimulated with 100 ng/ml LPS (TLR4 ligand) in the presence and absence of exogenous rGal-8 (20 µg/ml) or rGal-8 + control saccharides for 30 min at 37°C. Electrophoresis blots of cell lysates were processed for Western blot analysis using Abs to phospho-IκBα. Representative blots of three independent experiments done in duplicates are shown. (C) Exogenous rGal-8 diminishes secretion of TNF-α and IL-6 in BMDMs. Mature BMDMs were serum starved overnight and were then stimulated with 100 ng/ml LPS in the presence and absence of exogenous rGal-8 (20 µg/ml). At the end of the incubation period, TNF-α and IL-6 released into culture supernatants were measured by ELISA (n = 4). Data are plotted as mean ± SEM. Statistical levels of significance were analyzed by a Student t test. *p < 0.05, **p < 0.01.

Close modal

Taken together, our findings that the TLR4 pathway is augmented in infected Gal-8−/− corneas and that exogenous rGal-8 reduces TLR4 pathway activation in the LPS-treated BMDMs lead us to conclude that Gal-8 downmodulates the TLR4 pathway, and thereby plays a critical role in regulation of the innate immune response.

We next sought to define the mechanism by which Gal-8 represses the TLR4 pathway. The first step in the activation of TLR4 signaling is the delivery of LPS to CD14 that is present on the macrophage cell surface or in soluble form. LPS is then transferred from CD14 to the myeloid differentiation (MD-2) protein. The LPS/MD-2 subsequently interacts with TLR4, resulting in assembly of the TLR4/MD-2/CD14 complex that triggers the signaling pathways including activation of NF-κB pathway and secretion of inflammatory cytokines. To define the mechanism by which Gal-8 downmodulates the innate immune response, we first conducted studies to establish whether Gal-8 binds to LPS, TLR4, or CD14. To determine whether Gal-8 is an LPS-binding protein, Gal-8–coated wells were sequentially incubated with 1) biotinylated LPS, 2) streptavidin-HRP, and 3) TMB chromogenic substrate (R&D Systems) for HRP detection. This study revealed that unlike Gal-3 and CD14, Gal-8 is not an LPS-binding protein (Fig. 3A). To determine whether Gal-8 binds to TLR4 or CD14, BMDM lysates were incubated overnight with Gal-8–conjugated agarose beads, and bound proteins were examined along with total cell lysates (input) by Western blot using anti-TLR4 and anti-CD14. TLR4 did not bind to Gal-8 beads (Fig. 3B). In contrast, CD14 bound to Gal-8 beads (Fig. 3B, lane media). Binding of CD14 to Gal-8 was inhibited by a competing saccharide, lactose, whereas a noncompeting sugar, sucrose, had no effect (Fig. 3B), suggesting that CD14 binds to Gal-8 in a carbohydrate-dependent manner. Subsequently, we conducted studies to determine whether Gal-8 hinders TLR4 signaling by 1) interfering with the uptake of LPS by CD14, or 2) by deterring delivery of LPS from CD14 to MD-2 and formation of the TLR4/MD-2/CD14 complex. To determine whether Gal-8 interferes with the binding of LPS to CD14, BMDMs pretreated with rGal-8 (30 min) were incubated with FITC-LPS (1 h) and were then processed for flow cytometry to detect cell-bound FITC-LPS. These experiments revealed that Gal-8 does not interfere with binding of CD14 to LPS as indicated by complete overlap of the FITC-LPS plot regardless of whether the cells were pretreated with rGal-8 (Fig. 3C). Next, to determine whether Gal-8 prevents formation of the MD-2/TLR4 complex, BMDMs pretreated with rGal-8 were stimulated with FITC-LPS (30 min), stained with MD-2/TLR4 complex Ab, and then processed for flow cytometry. As expected, stimulation with LPS resulted in the formation the MD-2/TLR4 complex (Fig. 3D). Treatment with rGal-8 inhibited formation of the MD-2/TLR4 complex. The inhibitory effect of Gal-8 on formation of the MD-2/TLR4 complex was abrogated by 3′SL, a Gal-8 competing saccharide, but not by 6′SL, a noncompeting saccharide (Fig. 3D). These data lead us to conclude that Gal-8 binds to CD14 and downmodulates the TLR4 pathway by inhibiting formation of the MD-2/TLR4 complex in a carbohydrate-dependent manner.

FIGURE 3.

Gal-8 dampens TLR4 signaling by inhibiting MD-2 complex formation in a carbohydrate-dependent manner. (A) Unlike Gal-3, Gal-8 is not an LPS-binding protein. Biotinylated LPS was added to the Gal-8–coated wells, and the plates were sequentially incubated with streptavidin-HRP and then TMB chromogenic substrate. Wells coated with known LPS-binding proteins, Gal-3 and CD14, served as positive controls. Combined results of three independent experiments in triplicates are shown (n = 9). (B) Gal-8 binds to CD14 but not TLR4. BMDM lysates were incubated overnight with Gal-8–conjugated agarose beads in the presence or absence of lactose or sucrose (100 mM), and bound proteins were examined along with total cell lysates (input, 30 µg of protein) by Western blot using anti-TLR4 and anti-CD14. A representative blot of three independent experiments is shown. (C) Gal-8 does not interfere with the binding of LPS to CD14. BMDMs pretreated with rGal-8 were incubated with FITC-LPS, washed, and then processed for flow cytometry to detect bound FITC-LPS. Data are representative of three independent experiments in duplicate. (D) Gal-8 prevents delivery of LPS from CD14 to MD-2 and thereby prevents formation of the MD-2/TLR4 complex. BMDMs pretreated with rGal-8 (0.75 μM) in the absence and presence of control sugars were stimulated with 100 ng/ml FITC-LPS (30 min), washed, stained with MD-2/TLR4 complex Ab (MTS510, eBioscience), and then processed for flow cytometry. Left, Representative FACS plot of three independent experiments in duplicate is shown (n = 6). Right, Combined results of two to three independent experiments in duplicate are shown (n = 6 for LPS and rGal-8 groups; n = 4 for 3′SL and 6′SL groups). Data are plotted as mean ± SEM. Statistical levels of significance were analyzed by a Student t test. *p < 0.05, **p < 0.01.

FIGURE 3.

Gal-8 dampens TLR4 signaling by inhibiting MD-2 complex formation in a carbohydrate-dependent manner. (A) Unlike Gal-3, Gal-8 is not an LPS-binding protein. Biotinylated LPS was added to the Gal-8–coated wells, and the plates were sequentially incubated with streptavidin-HRP and then TMB chromogenic substrate. Wells coated with known LPS-binding proteins, Gal-3 and CD14, served as positive controls. Combined results of three independent experiments in triplicates are shown (n = 9). (B) Gal-8 binds to CD14 but not TLR4. BMDM lysates were incubated overnight with Gal-8–conjugated agarose beads in the presence or absence of lactose or sucrose (100 mM), and bound proteins were examined along with total cell lysates (input, 30 µg of protein) by Western blot using anti-TLR4 and anti-CD14. A representative blot of three independent experiments is shown. (C) Gal-8 does not interfere with the binding of LPS to CD14. BMDMs pretreated with rGal-8 were incubated with FITC-LPS, washed, and then processed for flow cytometry to detect bound FITC-LPS. Data are representative of three independent experiments in duplicate. (D) Gal-8 prevents delivery of LPS from CD14 to MD-2 and thereby prevents formation of the MD-2/TLR4 complex. BMDMs pretreated with rGal-8 (0.75 μM) in the absence and presence of control sugars were stimulated with 100 ng/ml FITC-LPS (30 min), washed, stained with MD-2/TLR4 complex Ab (MTS510, eBioscience), and then processed for flow cytometry. Left, Representative FACS plot of three independent experiments in duplicate is shown (n = 6). Right, Combined results of two to three independent experiments in duplicate are shown (n = 6 for LPS and rGal-8 groups; n = 4 for 3′SL and 6′SL groups). Data are plotted as mean ± SEM. Statistical levels of significance were analyzed by a Student t test. *p < 0.05, **p < 0.01.

Close modal

As described earlier, innate immune response results in the recruitment of neutrophils at the site of infection. Production of ROS by neutrophils is a vital component of the innate immune response that enables killing and clearance of pathogens. Therefore, we next investigated the role of Gal-8 in modulation of the ROS pathway and the bacterial killing capacity of neutrophils. In this study, we first determined whether expression of genes related to the ROS pathway is altered in Gal-8−/− neutrophils. FACS-sorted WT and Gal-8−/− neutrophils of infected corneas were processed for RNA-seq and pathway analyses by Ingenuity Pathway Analysis software (Qiagen). Analysis of the data revealed that many genes of the ROS pathway were upregulated in Gal-8−/− BMNs (Fig. 4A). To determine whether expression of ROS is upregulated in infected Gal-8−/− BMNs, WT and Gal-8−/− BMNs cocultured with P. aeruginosa 6294 (BMN/P. aeruginosa ratio of 1:10, 2 h) were incubated with ROS label for 30 min, and then fluorescence intensity of ROS was measured by flow cytometry. Infected WT BMNs expressed low levels of ROS (Fig. 4B, red). Expression levels of ROS were substantially higher in corresponding Gal-8−/− BMNs (Fig. 4B, green). To determine whether Gal-8 modulates the bacterial killing capacity of neutrophils, freshly isolated WT or Gal-8−/− BMNs were incubated with an ascending dose of P. aeruginosa (4 h), homogenized, and processed for bacterial enumeration. The numbers of P. aeruginosa colonies were substantially higher in the WT BMNs compared with the Gal-8−/− BMNs (Fig. 5A), indicating that the bacterial killing capacity of Gal-8−/− BMNs is considerably higher compared with that of Gal-8+/+ BMNs. To determine whether the exogenous rGal-8 rescues the altered bacterial killing capacity of Gal-8−/− neutrophils, Gal-8−/− BMNs were preincubated with ascending concentrations of rGal-8 (30 min), cocultured with P. aeruginosa 6294 (2 h), and then processed for bacterial enumeration. In these experiments, exogenous rGal-8 reduced the killing capacity of Gal-8−/− BMNs in a dose-dependent manner, resulting in a complete rescue of the phenotype at an rGal-8 concentration of ≥10 µg/ml or higher (Fig. 5B). To determine whether the bacterial killing capacity of BMNs is totally dependent on ROS, BMNs were preincubated with a ROS inhibitor (diphenyleneiodonium, 2 h) prior to coculture with P. aeruginosa. ROS inhibitor almost completely abolished the bacterial killing capacity of BMNs (Fig. 5C).

FIGURE 4.

Gal-8 downmodulates ROS expression in neutrophils. (A) Genes related to the ROS pathway are upregulated in Gal-8−/− neutrophils of infected corneas. FACS-sorted WT and Gal-8−/− neutrophils of infected corneas were processed for next-generation transcriptomic RNA sequencing (RNA-seq) and pathway analyses. (Ai) A representative heatmap of three independent experiments is shown. All signals are compared with a mean value, and change from the mean is visually represented by a color assignment. (Aii) Combined results of three independent experiments in triplicates are shown (n = 9). (B) ROS expression is upregulated in Gal-8−/− neutrophils. WT and Gal-8−/− BMNs were preincubated with 1× ROS label for 30 min at 37°C prior to coculture with P. aeruginosa 6294 (2 h), and then fluorescence intensity of ROS was quantified by flow cytometry. BMNs incubated with a ROS inducer (BioVision) served as a positive control. (Bi) Representative FACS plot. (Bii) Quantification of ROS expressed as mean fluorescence intensity (MFI) ± SEM. Combined results of two independent experiments in duplicate are shown. *p < 0.05.

FIGURE 4.

Gal-8 downmodulates ROS expression in neutrophils. (A) Genes related to the ROS pathway are upregulated in Gal-8−/− neutrophils of infected corneas. FACS-sorted WT and Gal-8−/− neutrophils of infected corneas were processed for next-generation transcriptomic RNA sequencing (RNA-seq) and pathway analyses. (Ai) A representative heatmap of three independent experiments is shown. All signals are compared with a mean value, and change from the mean is visually represented by a color assignment. (Aii) Combined results of three independent experiments in triplicates are shown (n = 9). (B) ROS expression is upregulated in Gal-8−/− neutrophils. WT and Gal-8−/− BMNs were preincubated with 1× ROS label for 30 min at 37°C prior to coculture with P. aeruginosa 6294 (2 h), and then fluorescence intensity of ROS was quantified by flow cytometry. BMNs incubated with a ROS inducer (BioVision) served as a positive control. (Bi) Representative FACS plot. (Bii) Quantification of ROS expressed as mean fluorescence intensity (MFI) ± SEM. Combined results of two independent experiments in duplicate are shown. *p < 0.05.

Close modal
FIGURE 5.

Gal-8 reduces bacterial killing capacity of neutrophils via ROS pathway. (A) Bacterial killing capacity of Gal-8−/− neutrophils is enhanced. Freshly isolated BMNs of WT or Gal-8−/− mice were incubated with an ascending dose of P. aeruginosa 6294 for 4 h, homogenized, and 10-µl aliquots of homogenates were plated on BBL blood agar overnight for bacterial enumeration. Note that the numbers of P. aeruginosa colonies are substantially higher in the WT BMNs compared with the Gal-8−/− BMNs. (B) Exogenous rGal-8 rescues the altered bacterial killing capacity of Gal-8−/− neutrophils. BMNs were preincubated with ascending concentrations of rGal-8 (5–20 µg/ml) for 30 min at 37°C prior to coculture with P. aeruginosa and then processed for bacterial enumeration. (C) Bacterial killing capacity of neutrophils is dependent on the ROS pathway. BMNs were preincubated with ROS inhibitor (diphenyleneiodonium, 10 µM) prior to coculture with P. aeruginosa and then processed for bacterial enumeration. Combined results of three independent experiments in duplicate are shown (n = 6). Data are plotted as mean ± SEM. Statistical levels of significance were analyzed by a Student t test. *p < 0.05, **p < 0.01, ***p < 0.001. PA, P. aeruginosa

FIGURE 5.

Gal-8 reduces bacterial killing capacity of neutrophils via ROS pathway. (A) Bacterial killing capacity of Gal-8−/− neutrophils is enhanced. Freshly isolated BMNs of WT or Gal-8−/− mice were incubated with an ascending dose of P. aeruginosa 6294 for 4 h, homogenized, and 10-µl aliquots of homogenates were plated on BBL blood agar overnight for bacterial enumeration. Note that the numbers of P. aeruginosa colonies are substantially higher in the WT BMNs compared with the Gal-8−/− BMNs. (B) Exogenous rGal-8 rescues the altered bacterial killing capacity of Gal-8−/− neutrophils. BMNs were preincubated with ascending concentrations of rGal-8 (5–20 µg/ml) for 30 min at 37°C prior to coculture with P. aeruginosa and then processed for bacterial enumeration. (C) Bacterial killing capacity of neutrophils is dependent on the ROS pathway. BMNs were preincubated with ROS inhibitor (diphenyleneiodonium, 10 µM) prior to coculture with P. aeruginosa and then processed for bacterial enumeration. Combined results of three independent experiments in duplicate are shown (n = 6). Data are plotted as mean ± SEM. Statistical levels of significance were analyzed by a Student t test. *p < 0.05, **p < 0.01, ***p < 0.001. PA, P. aeruginosa

Close modal

Our findings that Gal-8−/− mice are resistant to P. aeruginosa keratitis suggest that inhibiting Gal-8 may improve the outcome of the disease. To explore the potential of pharmacological targeting of Gal-8 for the treatment of P. aeruginosa keratitis, two groups of mice were infected with P. aeruginosa. Prior to challenge with P. aeruginosa, one group of mice received a subconjunctival injection of a Gal-8N inhibitor (19a), and the other group of mice was treated with vehicle only. The severity of bacterial keratitis was graded on day 1 p.i. and then corneas were harvested for bacterial and neutrophil enumeration and for quantification of IL-1β. Treatment with Gal-8 inhibitor substantially reduced P. aeruginosa keratitis as detected by reduction in corneal opacity (Fig. 6A), bacterial load (Fig. 6B), PMN infiltration (Fig. 6C), and expression level of IL1β (Fig. 6D). These data lead us to conclude that targeting Gal-8 is an attractive strategy for the development of novel treatment for blinding immunopathology resulting from bacterial keratitis and possibly other disorders resulting from chronic inflammation.

FIGURE 6.

Gal-8 inhibition reduces the severity of P. aeruginosa keratitis. Corneas of WT mice were infected using P. aeruginosa strain 6077. Immediately prior to infection, control group mice received a subconjunctival injection of vehicle (10 µl of PBS + 23% DMSO) and experimental group mice received a 10-μl injection of the Gal-8N inhibitor 19a (5 mg/ml in vehicle). On day 1 p.i., the eyes were examined to assess the severity of corneal keratitis and then corneas were harvested for bacterial and PMN enumeration and quantification of IL-1β. (Ai) Photographs of vehicle- and Gal-8N inhibitor–treated corneas. (Aii) Quantification of corneal opacity score (n = 30). (B) Bacterial load (n = 32). (C) Infiltration of PMNs (n = 19). (D) Quantification of total IL-1β by ELISA (n = 6). Data are plotted as mean ± SEM. Statistical levels of significance were analyzed by a Student t test. *p < 0.05, **p < 0.01, ***p < 0.001.

FIGURE 6.

Gal-8 inhibition reduces the severity of P. aeruginosa keratitis. Corneas of WT mice were infected using P. aeruginosa strain 6077. Immediately prior to infection, control group mice received a subconjunctival injection of vehicle (10 µl of PBS + 23% DMSO) and experimental group mice received a 10-μl injection of the Gal-8N inhibitor 19a (5 mg/ml in vehicle). On day 1 p.i., the eyes were examined to assess the severity of corneal keratitis and then corneas were harvested for bacterial and PMN enumeration and quantification of IL-1β. (Ai) Photographs of vehicle- and Gal-8N inhibitor–treated corneas. (Aii) Quantification of corneal opacity score (n = 30). (B) Bacterial load (n = 32). (C) Infiltration of PMNs (n = 19). (D) Quantification of total IL-1β by ELISA (n = 6). Data are plotted as mean ± SEM. Statistical levels of significance were analyzed by a Student t test. *p < 0.05, **p < 0.01, ***p < 0.001.

Close modal

The present study was designed to investigate the immunomodulatory role of Gal-8 in P. aeruginosa keratitis. We demonstrate in the present study that Gal-8, which is known to be highly upregulated in pathological corneas, including P. aeruginosa–infected corneas (35), plays a critical role in the pathobiology of P. aeruginosa keratitis. First, we established that Gal-8−/− mice are resistant to P. aeruginosa keratitis as demonstrated by a marked reduction in bacterial load and opacity in infected Gal-8−/− corneas compared with the corresponding WT corneas. Clinical isolates of P. aeruginosa fall into one of two categories: invasive strains and cytotoxic strains. Both strain types cause human corneal disease (36), but the disease they cause has been shown to involve different pathological processes in an animal model (37). Our findings that Gal-8−/− mice are resistant to P. aeruginosa keratitis regardless of whether the corneas are infected with the cytotoxic strain P. aeruginosa 6077 or an invasive strain P. aeruginosa 6294 suggest that the observed phenotype of Gal-8−/− mice is unrelated to the invasiveness and cytotoxicity virulence determinants. Interestingly, a published study has shown that Gal-8−/− animals are also partially resistant to the growth and development of primary tumors and metastatic lesions (38). In contrast, in Trypanosoma cruzi chronic infection, Gal-8 protects from the disease as shown by a generalized increase in heart, skeletal muscle, and liver inflammation and fibrosis in Gal-8−/− mice compared with the WT mice (39). Other published studies using experimental models of autoimmune diseases including rheumatoid arthritis (25), uveitis (26), and encephalomyelitis (27) emphasize that Gal-8 exhibits anti-inflammatory effects, mostly by modulating the adaptive immune response. Dual behavior of Gal-8, acting as both a proinflammatory as well as an anti-inflammatory molecule based on the context and microenvironment of tissue can easily be explained by the differential glycosylation profile exhibited by the cells of the innate and adaptive immune system as well as naive versus activated immune cells and the expression of other galectins in the milieu. In this respect, Gal-1 has also been shown to have a range of proinflammatory and anti-inflammatory functions dependent on its expression and cellular localization (40, 41).

Various members of the galectin family have been shown to bind to bacteria and kill the bound bacteria. For example, human Gal-3 binds to and induces death of Candida albicans (42), Gal-2 and Gal-3 kill Helicobacter pylori (43, 44), and fish Gal-8 has been shown to exert bactericidal activity against some Gram-negative bacterial pathogens (45). Also, Stowell et al. (46) have demonstrated that rGal-8 is able to selectively kill bacteria expressing human blood group Ags, both in vivo and in vitro. In addition, a number of studies have shown that that intracellular Gal-8 targets invading microbes for their destruction by autophagy (22). None of these mechanisms of bacterial killing is likely to be directly relevant to the reduced severity of P. aeruginosa keratitis we observed in the Gal-8−/− mice because we found that 1) Gal-8−/− mice were resistant to P. aeruginosa keratitis regardless of whether the corneas were infected with the extracellular cytotoxic strain P. aeruginosa 6077 or an invasive strain P. aeruginosa 6294, and that 2) Gal-8 does not influence P. aeruginosa growth. Regarding the mechanism by which Gal-8−/− mice resist the P. aeruginosa keratitis, we demonstrated that Gal-8 downmodulates the innate immune response. The first step in induction of the innate immune response in the context of bacterial infections is generally achieved by TLR-mediated activation of the NF-κB pathway that results in the induction of 1) proinflammatory cytokines, 2) the Rho pathway, 3) assembly of inflammasomes, and 4) recruitment of neutrophils that ultimately clear the infection. That Gal-8 dampens the innate immune response is demonstrated by our 1) RNA-seq data showing that many genes of the innate immune pathway including NLRP3, NF-κB, TNF, and CD14 are upregulated in Gal-8−/− neutrophils compared with that of Gal-8+/+ neutrophils of P. aeruginosa–infected mouse corneas; 2) Western blot data showing that expression levels of proteins known to be upregulated upon activation of the TLR4 pathway including pro–IL-1β, NLRP3, and pro–caspase-1 were substantially higher in infected Gal-8−/− corneas compared with corresponding WT mouse corneas; and 3) findings that BMDMs stimulated with LPS in the presence of exogenous rGal-8 exhibited reduction in activation of the NF-κB pathway as indicated by the expression of a markedly reduced level of phosphorylated IκBα and proinflammatory cytokines TNF-α and IL-6. Our striking finding that unlike Gal-8−/− mice, mice deficient in several other members of the galectin family, including Gal-1, Gal-3, and Gal-9, were fully susceptible to P. aeruginosa keratitis suggests that Gal-8–mediated resistance to P. aeruginosa keratitis involves the affinity of N-terminal CRD (N-CRD) of Gal-8 for 3′-sialylated galactosides that is unique among animal galectins (20, 21). In support of this notion, specific inhibition of N-CRD of Gal-8 by 3′-SL abrogated the effect of exogenous rGal-8 on suppression of the TLR4 pathway in BMDMs stimulated with LPS. In contrast, control saccharide, 6′-SL, which lacks the affinity for the N-CRD (Gal-8N) of Gal-8 had no effect. Taken together, these data suggest that the carbohydrate-dependent function of Gal-8N is directly involved in the inhibitory effect of Gal-8 on activation of the TLR4 pathway.

That Gal-8 dampens the innate immune response by suppressing formation of the MD-2/TLR4 complex via interacting with CD14 in a carbohydrate-dependent manner is demonstrated by our robust mechanistic studies showing that 1) CD14 is a Gal-8–binding protein; unlike Gal-3, Gal-8 is not an LPS-binding protein; Gal-8 is also not a TLR4 binding; 3) Gal-8 does not interfere with the delivery of LPS to CD14; and 3) Gal-8 prevents delivery of LPS from CD14 to MD-2 and thereby averts formation of the MD-2/TLR4 complex. Again, our finding that the inhibitory effect of Gal-8 on the formation of MD-2/TLR4 complex was abrogated by 3′SL, a Gal-8 competing saccharide, but not by 6′SL, a noncompeting saccharide, suggests that Gal-8–mediated suppression of formation of the TLR4 complex involves the affinity of N-CRD of Gal-8 for 3′-sialylated galactosides.

In most cases, killing and clearance of pathogens is dependent on ROS production by neutrophils at the site of infection. ROS deficiency in humans results in recurrent and severe bacterial infections, whereas their unregulated release leads to pathology from excessive inflammation. That Gal-8 downmodulates the neutrophil-mediated killing of bacteria in a ROS-dependent manner is suggested by our findings that 1) many genes of the ROS pathway were upregulated in Gal-8−/− BMNs, 2) expression levels of ROS are substantially higher in infected Gal-8−/− BMNs compared with Gal-8+/+ neutrophils, and 3) ROS inhibitor almost completely abolished the P. aeruginosa killing capacity of BMNs. Our findings that the bacterial killing capacity of Gal-8−/− BMNs is considerably higher compared with that of Gal-8+/+ BMNs, and that exogenous rGal-8 completely rescues the enhanced bacterial killing phenotype of Gal-8−/− neutrophils, lead us to conclude with a high level of confidence that Gal-8 is a key player in modulation of the bacterial killing capacity of neutrophils.

Regardless of the mechanisms involved, our findings that treatment with a Gal-8 inhibitor markedly reduces the severity of P. aeruginosa keratitis has broad implications for developing novel therapeutic strategies for treatment of conditions resulting from the hyperactive immune response both in ocular as well as nonocular tissues. At present, treatment of P. aeruginosa keratitis is a major clinical problem. As described earlier, P. aeruginosa keratitis is notorious for causing rapidly fulminant disease often associated with corneal melting and permanent vision loss. Although the current therapy including broad-spectrum topical antibiotics ultimately eliminates the bacteria, it fails to salvage vision in ∼30% of patients who develop long-term moderate-to-severe vision loss (1). Unfortunately, strategies to treat P. aeruginosa keratitis have not changed in the last 30 y or so, and there is a tremendous need to develop evidence-based therapy to downmodulate an unwanted, excessive immune response. Implications of developing novel treatment for an unwarranted, out of control immune response extend far beyond the diseases of ocular tissues. Indeed, excessive inflammation due to an unchecked inflammatory cytokine storm is responsible for pathogenesis of many cytokine storm-associated diseases (47, 48), including sepsis, cancers, autoimmune diseases, infections, and other inflammatory diseases, where TLR signaling plays a significant role. Again, targeting Gal-8 may help downmodulate an unwanted immune response in such cases.

The authors have no financial conflicts of interest.

We thank Dr. Albert Tai (Tufts Genomic Core) for expert help and guidance in RNA sequencing analyses. We also thank Drs. Tanveer Zaidi and Pablo Argueso for helpful discussions.

This work was supported by the National Eye Institute Grant R01EY028570, the Massachusetts Lions Eye Research Fund, the New England Corneal Transplant Fund, an unrestricted challenge award from Research to Prevent Blindness, and by Galecto Biotech.

The online version of this article contains supplemental material.

The RNA sequence data presented in this article have been submitted to the ArrayExpress database (https://www.ebi.ac.uk/biostudies/arrayexpress) under the accession number E-MTAB-12376.

Abbreviations used in this article:

19a

3,4-dichlorophenyl 3-O-(5-carboxy-1-methyl-[1H]-benzo[d]imidazole-2-ylmethyl)-1-thio-α-d-galactopyranoside

BMDM

bone marrow–derived macrophage

BMN

bone marrow neutrophil

CRD

carbohydrate recognition domain

Gal-8

galectin-8

Gal-8N

Gal-8 N-terminal CRD

N-CRD

N-terminal CRD

OBB

Odyssey blocking buffer

p.i.

postinfection

PMN

polymorphonuclear leukocyte

RIPA

radioimmunoprecipitation assay

RNA-seq

RNA sequencing

ROS

reactive oxygen species

3′SL

α2,3-sialyl lactose

6′SL

α2,6-sialyl lactose

TMB

tetramethylbenzidine

WT

wild-type

1.
Ung
L.
,
J.
Chodosh
.
2021
.
Foundational concepts in the biology of bacterial keratitis
.
Exp. Eye Res.
209
:
108647
.
2.
Lakhundi
S.
,
R.
Siddiqui
,
N. A.
Khan
.
2017
.
Pathogenesis of microbial keratitis
.
Microb. Pathog.
104
:
97
109
.
3.
Ormerod
L. D.
,
R. E.
Smith
.
1986
.
Contact lens-associated microbial keratitis
.
Arch. Ophthalmol.
104
:
79
83
.
4.
Pachigolla
G.
,
P.
Blomquist
,
H. D.
Cavanagh
.
2007
.
Microbial keratitis pathogens and antibiotic susceptibilities: a 5-year review of cases at an urban county hospital in north Texas
.
Eye Contact Lens
33
:
45
49
.
5.
Schein
O. D.
,
L. D.
Ormerod
,
E.
Barraquer
,
E.
Alfonso
,
K. M.
Egan
,
B. G.
Paton
,
K. R.
Kenyon
.
1989
.
Microbiology of contact lens-related keratitis
.
Cornea
8
:
281
285
.
6.
Akira
S.
,
K.
Takeda
.
2004
.
Toll-like receptor signalling
.
Nat. Rev. Immunol.
4
:
499
511
.
7.
Pandey
R. K.
,
F. S.
Yu
,
A.
Kumar
.
2013
.
Targeting Toll-like receptor signaling as a novel approach to prevent ocular infectious diseases
.
Indian J. Med. Res.
138
:
609
619
.
8.
Pearlman
E.
,
Y.
Sun
,
S.
Roy
,
M.
Karmakar
,
A. G.
Hise
,
L.
Szczotka-Flynn
,
M.
Ghannoum
,
H. R.
Chinnery
,
P. G.
McMenamin
,
A.
Rietsch
.
2013
.
Host defense at the ocular surface
.
Int. Rev. Immunol.
32
:
4
18
.
9.
Rathinam
V. A.
,
K. A.
Fitzgerald
.
2016
.
Inflammasome complexes: emerging mechanisms and effector functions
.
Cell
165
:
792
800
.
10.
Sahoo
M.
,
I.
Ceballos-Olvera
,
L.
del Barrio
,
F.
Re
.
2011
.
Role of the inflammasome, IL-1β, and IL-18 in bacterial infections
.
ScientificWorldJournal
11
:
2037
2050
.
11.
Storek
K. M.
,
D. M.
Monack
.
2015
.
Bacterial recognition pathways that lead to inflammasome activation
.
Immunol. Rev.
265
:
112
129
.
12.
Hobden
J. A.
,
S. A.
Masinick
,
R. P.
Barrett
,
L. D.
Hazlett
.
1997
.
Proinflammatory cytokine deficiency and pathogenesis of Pseudomonas aeruginosa keratitis in aged mice
.
Infect. Immun.
65
:
2754
2758
.
13.
Church
L. D.
,
G. P.
Cook
,
M. F.
McDermott
.
2008
.
Primer: inflammasomes and interleukin 1β in inflammatory disorders
.
Nat. Clin. Pract. Rheumatol.
4
:
34
42
.
14.
Karmakar
M.
,
Y.
Sun
,
A. G.
Hise
,
A.
Rietsch
,
E.
Pearlman
.
2012
.
Cutting edge: IL-1β processing during Pseudomonas aeruginosa infection is mediated by neutrophil serine proteases and is independent of NLRC4 and caspase-1
.
J. Immunol.
189
:
4231
4235
.
15.
Dinarello
C. A.
2009
.
Immunological and inflammatory functions of the interleukin-1 family
.
Annu. Rev. Immunol.
27
:
519
550
.
16.
Davis
B. K.
,
H.
Wen
,
J. P.
Ting
.
2011
.
The inflammasome NLRs in immunity, inflammation, and associated diseases
.
Annu. Rev. Immunol.
29
:
707
735
.
17.
Huang
X.
,
W.
Du
,
S. A.
McClellan
,
R. P.
Barrett
,
L. D.
Hazlett
.
2006
.
TLR4 is required for host resistance in Pseudomonas aeruginosa keratitis
.
Invest. Ophthalmol. Vis. Sci.
47
:
4910
4916
.
18.
Bidon-Wagner
N.
,
J. P.
Le Pennec
.
2002
.
Human galectin-8 isoforms and cancer
.
Glycoconj. J.
19
:
557
563
.
19.
Zick
Y.
,
M.
Eisenstein
,
R. A.
Goren
,
Y. R.
Hadari
,
Y.
Levy
,
D.
Ronen
.
2002
.
Role of galectin-8 as a modulator of cell adhesion and cell growth
.
Glycoconj. J.
19
:
517
526
.
20.
Ideo
H.
,
T.
Matsuzaka
,
T.
Nonaka
,
A.
Seko
,
K.
Yamashita
.
2011
.
Galectin-8-N-domain recognition mechanism for sialylated and sulfated glycans
.
J. Biol. Chem.
286
:
11346
11355
.
21.
Carlsson
S.
,
C. T.
Oberg
,
M. C.
Carlsson
,
A.
Sundin
,
U. J.
Nilsson
,
D.
Smith
,
R. D.
Cummings
,
J.
Almkvist
,
A.
Karlsson
,
H.
Leffler
.
2007
.
Affinity of galectin-8 and its carbohydrate recognition domains for ligands in solution and at the cell surface
.
Glycobiology
17
:
663
676
.
22.
Thurston
T. L.
,
M. P.
Wandel
,
N.
von Muhlinen
,
A.
Foeglein
,
F.
Randow
.
2012
.
Galectin 8 targets damaged vesicles for autophagy to defend cells against bacterial invasion
.
Nature
482
:
414
418
.
23.
Tribulatti
M. V.
,
J.
Carabelli
,
C. A.
Prato
,
O.
Campetella
.
2020
.
Galectin-8 in the onset of the immune response and inflammation
.
Glycobiology
30
:
134
142
.
24.
Zick
Y.
2022
.
Galectin-8, cytokines, and the storm
.
Biochem. Soc. Trans.
50
:
135
149
.
25.
Eshkar Sebban
L.
,
D.
Ronen
,
D.
Levartovsky
,
O.
Elkayam
,
D.
Caspi
,
S.
Aamar
,
H.
Amital
,
A.
Rubinow
,
I.
Golan
,
D.
Naor
, et al
2007
.
The involvement of CD44 and its novel ligand galectin-8 in apoptotic regulation of autoimmune inflammation
.
J. Immunol.
179
:
1225
1235
.
26.
Sampson
J. F.
,
E.
Hasegawa
,
L.
Mulki
,
A.
Suryawanshi
,
S.
Jiang
,
W. S.
Chen
,
G. A.
Rabinovich
,
K. M.
Connor
,
N.
Panjwani
.
2015
.
Galectin-8 ameliorates murine autoimmune ocular pathology and promotes a regulatory T cell response
.
PLoS One
10
:
e0130772
.
27.
Pardo
E.
,
C.
Cárcamo
,
R.
Uribe-San Martín
,
E.
Ciampi
,
F.
Segovia-Miranda
,
C.
Curkovic-Peña
,
F.
Montecino
,
C.
Holmes
,
J. E.
Tichauer
,
E.
Acuña
, et al
2017
.
Galectin-8 as an immunosuppressor in experimental autoimmune encephalomyelitis and a target of human early prognostic antibodies in multiple sclerosis
.
PLoS One
12
:
e0177472
.
28.
Valenzuela
D. M.
,
A. J.
Murphy
,
D.
Frendewey
,
N. W.
Gale
,
A. N.
Economides
,
W.
Auerbach
,
W. T.
Poueymirou
,
N. C.
Adams
,
J.
Rojas
,
J.
Yasenchak
, et al
2003
.
High-throughput engineering of the mouse genome coupled with high-resolution expression analysis. [Published erratum appears in 2003 Nat. Biotechnol. 21: 822.]
Nat. Biotechnol.
21
:
652
659
.
29.
Chen
W. S.
,
Z.
Cao
,
S.
Sugaya
,
M. J.
Lopez
,
V. G.
Sendra
,
N.
Laver
,
H.
Leffler
,
U. J.
Nilsson
,
J.
Fu
,
J.
Song
, et al
2016
.
Pathological lymphangiogenesis is modulated by galectin-8-dependent crosstalk between podoplanin and integrin-associated VEGFR-3. [Published erratum appears in 2016 Nat. Commun. 7: 12063.]
Nat. Commun.
7
:
11302
.
30.
Poland
P. A.
,
C. L.
Kinlough
,
R. P.
Hughey
.
2015
.
Cloning, expression, and purification of galectins for in vitro studies
.
Methods Mol. Biol.
1207
:
37
49
.
31.
Berk
R. S.
,
L. D.
Hazlett
,
K. W.
Beisel
.
1987
.
Genetic studies on resistant and susceptibility genes controlling the mouse cornea to infection with Pseudomonas aeruginosa
.
Antibiot Chemother (1971)
39
:
83
91
.
32.
Broz
P.
,
D. M.
Monack
.
2013
.
Measuring inflammasome activation in response to bacterial infection
.
Methods Mol. Biol.
1040
:
65
84
.
33.
Swamydas
M.
,
M. S.
Lionakis
.
2013
.
Isolation, purification and labeling of mouse bone marrow neutrophils for functional studies and adoptive transfer experiments
.
J. Vis. Exp.
(
77
):
e50586
.
34.
Hassan
M.
,
S.
van Klaveren
,
M.
Håkansson
,
C.
Diehl
,
R.
Kovačič
,
F.
Baussière
,
A. P.
Sundin
,
J.
Dernovšek
,
B.
Walse
,
F.
Zetterberg
, et al
2021
.
Benzimidazole-galactosides bind selectively to the Galectin-8 N-terminal domain: structure-based design and optimisation
.
Eur. J. Med. Chem.
223
:
113664
.
35.
Chen
W. S.
,
Z.
Cao
,
L.
Truong
,
S.
Sugaya
,
N.
Panjwani
.
2015
.
Fingerprinting of galectins in normal, P. aeruginosa-infected, and chemically burned mouse corneas
.
Invest. Ophthalmol. Vis. Sci.
56
:
515
525
.
36.
Fleiszig
S. M.
,
T. S.
Zaidi
,
M. J.
Preston
,
M.
Grout
,
D. J.
Evans
,
G. B.
Pier
.
1996
.
Relationship between cytotoxicity and corneal epithelial cell invasion by clinical isolates of Pseudomonas aeruginosa
.
Infect. Immun.
64
:
2288
2294
.
37.
Cole
N.
,
M. D.
Willcox
,
S. M.
Fleiszig
,
F.
Stapleton
,
B.
Bao
,
S.
Tout
,
A.
Husband
.
1998
.
Different strains of Pseudomonas aeruginosa isolated from ocular infections or inflammation display distinct corneal pathologies in an animal model
.
Curr. Eye Res.
17
:
730
735
.
38.
Shatz-Azoulay
H.
,
Y.
Vinik
,
R.
Isaac
,
U.
Kohler
,
S.
Lev
,
Y.
Zick
.
2020
.
The animal lectin galectin-8 promotes cytokine expression and metastatic tumor growth in mice
.
Sci. Rep.
10
:
7375
.
39.
Bertelli
A.
,
L. M.
Sanmarco
,
C. A.
Pascuale
,
M.
Postan
,
M. P.
Aoki
,
M. S.
Leguizamón
.
2020
.
Anti-inflammatory role of galectin-8 during Trypanosoma cruzi chronic infection
.
Front. Cell. Infect. Microbiol.
10
:
285
.
40.
Sundblad
V.
,
L. G.
Morosi
,
J. R.
Geffner
,
G. A.
Rabinovich
.
2017
.
Galectin-1: a jack-of-all-trades in the resolution of acute and chronic inflammation
.
J. Immunol.
199
:
3721
3730
.
41.
Lei
T.
,
S.
Moos
,
J.
Klug
,
F.
Aslani
,
S.
Bhushan
,
E.
Wahle
,
S.
Fröhlich
,
A.
Meinhardt
,
M.
Fijak
.
2018
.
Galectin-1 enhances TNFα-induced inflammatory responses in Sertoli cells through activation of MAPK signalling
.
Sci. Rep.
8
:
3741
.
42.
Kohatsu
L.
,
D. K.
Hsu
,
A. G.
Jegalian
,
F. T.
Liu
,
L. G.
Baum
.
2006
.
Galectin-3 induces death of Candida species expressing specific β-1,2-linked mannans
.
J. Immunol.
177
:
4718
4726
.
43.
Sasaki
T.
,
R.
Saito
,
M.
Oyama
,
T.
Takeuchi
,
T.
Tanaka
,
H.
Natsume
,
M.
Tamura
,
Y.
Arata
,
T.
Hatanaka
.
2020
.
Galectin-2 has bactericidal effects against Helicobacter pylori in a β-galactoside-dependent manner
.
Int. J. Mol. Sci.
21
:
2697
.
44.
Park
A. M.
,
S.
Hagiwara
,
D. K.
Hsu
,
F. T.
Liu
,
O.
Yoshie
.
2016
.
Galectin-3 plays an important role in innate immunity to gastric infection by Helicobacter pylori
.
Infect. Immun.
84
:
1184
1193
.
45.
Zhang
T.
,
S.
Jiang
,
L.
Sun
.
2020
.
A fish galectin-8 possesses direct bactericidal activity
.
Int. J. Mol. Sci.
22
:
376
.
46.
Stowell
S. R.
,
C. M.
Arthur
,
M.
Dias-Baruffi
,
L. C.
Rodrigues
,
J. P.
Gourdine
,
J.
Heimburg-Molinaro
,
T.
Ju
,
R. J.
Molinaro
,
C.
Rivera-Marrero
,
B.
Xia
, et al
2010
.
Innate immune lectins kill bacteria expressing blood group antigen
.
Nat. Med.
16
:
295
301
.
47.
Chan
L.
,
N.
Karimi
,
S.
Morovati
,
K.
Alizadeh
,
J. E.
Kakish
,
S.
Vanderkamp
,
F.
Fazel
,
C.
Napoleoni
,
K.
Alizadeh
,
Y.
Mehrani
, et al
2021
.
The roles of neutrophils in cytokine storms
.
Viruses
13
:
2318
.
48.
Kumar
V.
2020
.
Toll-like receptors in sepsis-associated cytokine storm and their endogenous negative regulators as future immunomodulatory targets
.
Int. Immunopharmacol.
89
(
Pt B
):
107087
.

Supplementary data