Osteopontin Depletion in Nonhematopoietic Cells Improves Outcomes in Septic Mice by Enhancing Antimicrobial Peptide Production

Abstract

Sepsis is characterized as system-wide dysregulation of inflammation responses following primary bacterial infection [1, 2]. It accounts for one of the leading causes of death in intensive care units and is responsible for an estimated 5 million deaths annually worldwide [1, 3, 4]. So far, the options for sepsis treatment are limited to supportive care and antibiotics, and no targeted therapies have been successfully developed [5]. Therefore, identifying novel therapeutic targets and developing effective therapeutic strategies are urgently required for sepsis treatment.

As a complex ecosystem, gut microbiota consists of trillions of bacteria and is involved in various biological processes, including food digestion, hormone production, and immune system development and maturation [6]. Accumulating evidence has indicated gut microbiota as a pathophysiologic factor of sepsis [7]. Maintaining a diverse and balanced gut microbiota is critical for enhancing host immunity against enteric and systemic pathogens, and dysbiosis of gut microbiota potentially increases host susceptibility to sepsis [8]. Antimicrobial peptides (AMPs) are a group of broad-spectrum innate antibiotics produced by epithelia, certain immune cells, and microbes. They play an important role in shaping and controlling the bacterial loads and composition of gut microbiota to ensure beneficial homeostasis at the intestinal barrier [9]. The mechanisms by which AMPs clear drug-resistant bacteria include biofilm permeation, resensitization, intracellular bacteriostatic function, immune activity regulation, and biofilm inhibition. Among them, host defense peptides neutralize intracellular toxins, thereby enhancing innate immunity and exerting antimicrobial effects. For example, human LL-37 can modulate immune cytokines by stimulating chemokine release and inhibiting lipopolysaccharide (LPS)–mediated infection [10]. AMP production is mainly regulated by NF-κB signaling in response to bacterial infection or FOXO signaling under noninfectious conditions [11, 12]. Dysregulation of AMPs is implicated in the pathogenesis of gastrointestinal diseases, such as Crohn disease and necrotizing enterocolitis [13, 14]. Moreover, recent studies have indicated the involvement of AMPs in sepsis. For example, AMP LL-37 could ameliorate septic death in a mouse sepsis model induced by cecal ligation and puncture (CLP) [15].

Osteopontin (OPN) is a multifunctional extracellular matrix protein that plays important roles in various physiologic and pathologic processes [16]. Upregulation of OPN is observed in many pathologic conditions, including inflammatory disorders and tissue injuries [17–19]. By interacting with cell surface receptors, such as integrins and CD44, OPN has been reported to activate multiple intracellular signaling, such as ERK, PI3 K/AKT, and NF-κB [20]. Previous studies have demonstrated significantly increased plasma OPN in nonsurvival sepsis cases than in survival ones [21], improved septic mortality and inflammation in Opn^−/− mice [22], and ameliorated sepsis-induced acute lung injury upon OPN neutralization [23]. However, the underlying mechanism by which OPN deficiency improves septic outcomes remains unknown.

In the present study, we for the first time demonstrated that OPN depletion in nonhematopoietic cells but not hematopoietic cells contributes to improved sepsis outcomes, and we showed enhanced AMP production in Opn^−/− mice as compared with wild type (WT) mice under noninfectious conditions. By fecal microbiota transplantation assay and transient neutralization of OPN, we found that Opn^−/− mice could reduce the bacterial loads of gut microbiota from WT mice, resulting in improved septic outcomes. Mechanistically, by employing intestinal organoids, we demonstrated that OPN from nonhematopoietic cells could directly regulate AMP production by modulating the AKT-FOXO3a signaling pathway. Therefore, OPN could serve as a therapeutic target for sepsis prevention and treatment.

METHODS

Mice

Animal protocols were approved by the Institutional Animal Care and Use Committee of the Department of Liver Diseases of the Shuguang Hospital.

Microbiome Depletion and Fecal Microbial Transplantation

Mouse luminal bacteria were depleted with the administration of 200 μL of mixed antibiotics solution through gavage for 3 consecutive days. Feces from WT donor mice were suspended in 30% glycerol diluted in phosphate-buffered saline (1 mL, C10010500BT; Gibco) and stocked at −80 °C until use. In total, 200 μL of the fecal mixture was gavaged into recipient mice 4 times over 12 days.

Detailed methods can be found in the Supplementary material.

RESULTS

Elevated OPN Expression in Septic Mice

Consistent with previous studies [24], we found that OPN levels in serum and peritoneal lavage fluid were significantly increased at 24 hours after treatment with LPS (intraperitoneal, 5 mg/kg; Supplementary Figure 1A and 2B) or CLP (Supplementary Figure 1C and 1D).

Significantly Reduced Septic Mortality and Inflammation in Opn^−/− Mice

Genetic Opn depletion could ameliorate septic death and inflammation [24]. To verify this point, LPS- and CLP-induced sepsis was performed in WT and Opn^−/− mice. A significant decrease in septic mortality was observed in Opn^−/− mice as compared with WT mice (Supplementary Figure 2A and 2B). Meanwhile, decreased proinflammatory cytokine induction (Supplementary Figure 2C and 2D), lung infiltration of Ly-6G⁺ cells (Supplementary Figure 2E), and lung wet/dry ratio (Supplementary Figure 2F) were detected in Opn^−/− mice as compared with WT mice at 24 hours after LPS- or CLP-induced sepsis. The higher body temperature was detected in Opn^−/− mice as compared with WT mice at 24 hours after LPS-induced sepsis (Supplementary Figure 2G). In the open-field experiment, Opn^−/− mice were higher in total activity path and central activity time as compared with WT mice (Supplementary Figure 2H). These results confirm that OPN deficiency plays a protective role in septic inflammation and mortality.

OPN Does Not Affect LPS-Induced Macrophage Activation

OPN deficiency improves sepsis outcomes [24], but the underlying mechanism is largely unknown. Given that excessive macrophage activation frequently leads to the surge of proinflammatory cytokines and high mortality in sepsis [2, 25], we then examined whether OPN could affect macrophage activation. When compared with macrophages derived from WT mice, macrophages derived from Opn^−/− mice had higher interleukin 6 (IL-6) and tumor necrosis factor α (TNF-α) induction at 8 hours (mRNA level) and 12 hours (protein level) after LPS treatment (Supplementary Figure 3A and 3B). However, MAPK signaling activation was comparable in both genotypes (Supplementary Figure 3C). OPN is greatly induced during sepsis; therefore, the in vitro LPS treatment without OPN supplement may not be able to sufficiently reflect in vivo macrophage activation in WT mice. To address this concern, we generated bone marrow–derived macrophages from WT mice and treated them with recombinant mouse OPN (rmOPN), LPS, or LPS + rmOPN. LPS treatment for 4 hours drastically induced IL-6 and TNF-α as compared with rmOPN treatment alone (Supplementary Figure 3D), whereas the addition of rmOPN could not further boost LPS-induced production of IL-6 and TNF-α. These findings suggest that OPN does not affect LPS-induced macrophage activation.

Nonhematopoietic Opn Depletion Contributes to the Improved Septic Outcomes

Enhanced bacterial clearance is closely associated with reduced septic inflammation, organ dysfunction, and mortality. To investigate whether Opn^−/− mice have enhanced bacterial clearance and in turn improve septic inflammation and mortality, bacterial loads in blood and peritoneal lavage fluid were examined in both genotypes at 24 hours after CLP treatment. Significantly decreased bacterial burdens were detected in Opn^−/− mice as compared with WT mice (Figure 1A). Given that macrophage is a major phagocytic cell type responsible for antibacterial defense in septic hosts, we then compared the phagocytosis and bacterial-killing abilities of macrophages in both genotypes. As illustrated in Supplementary Figure 4A, peritoneal macrophages derived from WT or Opn^−/− mice displayed comparable phagocytosis and bacterial-killing abilities. In addition to macrophages, other types of bone marrow–derived cells, such as neutrophils, dendritic cells, and mesenchymal cells, either directly or indirectly reduce inflammation and enhance bacterial clearance in sepsis. To investigate whether the protective effects in Opn^−/− mice are attributed to bone marrow–derived cells, we lethally irradiated WT and Opn^−/− mice and then reconstituted them with bone marrow from WT or Opn^−/− mice (Supplementary Figure 4B). These chimeric mice were then subjected to CLP-induced sepsis. Chimeras solely with hematopoietic Opn depletion (Opn^−/− > WT) did not show improved septic mortality as compared with WT chimeras (WT > WT; Figure 1B). Chimeras with nonhematopoietic Opn depletion (WT > Opn^−/−, Opn^−/− > Opn^−/−) displayed less septic death, reduced induction of IL-6 and TNF-α, and decreased bacterial burdens (Figure 1BitalicD). These findings suggest that nonhematopoietic but not hematopoietic Opn depletion contributes to the improved septic outcomes in Opn^−/− mice.

Figure 1.

Nonhematopoietic but not hematopoietic Opn depletion contributes to improved septic outcomes. A, Bacterial numbers in blood and PLF from WT and Opn^−/− mice at 24 hours after CLP treatment. When compared with WT mice, Opn^−/− mice showed significantly reduced bacterial burdens in blood and PLF samples. B, Survival rates of indicated chimeras in CLP-induced sepsis. Chimeras with nonhematopoietic Opn depletion demonstrated improved septic mortality. Data were analyzed by log-rank test. C, Serum levels of IL-6 and TNF-α of indicated chimeras were detected at 24 hours after CLP treatment (n = 4). Chimeras with nonhematopoietic Opn depletion revealed improved septic inflammation. D, Bacterial burdens in blood and PLF of indicated chimeras were detected at 24 hours after CLP treatment (n = 4). Chimeras with nonhematopoietic Opn depletion indicated reduced bacterial burdens in blood and PLF samples. Data were analyzed by Student t test and 1-way analysis of variance with Tukey post hoc test. Data are shown as mean ± SEM. *P < .05. **P < .01. ***P < .001. CFU, colony-forming units; CLP, cecal ligation and puncture; IL-6, interleukin 6; NS, not significant; OPN, osteopontin; PLF, peritoneal lavage fluid; TNF-α, tumor necrosis factor α; WT, wild type.

Open in new tabDownload slide

Decreased Bacterial Loads and Increased Bacterial Diversity in the Gut Microbiota of Opn^−/− Mice

Differences in gut microbiota could lead to distinct responses to sepsis even in genetically identical mice [26]. Since we found that nonhematopoietic Opn depletion contributes to the improved septic outcomes, we hypothesized that Opn depletion may affect gut microbiota and lead to improved sepsis outcomes. To test this hypothesis, we first cohoused WT and Opn^−/− mice for at least 3 weeks to eliminate possible gut microbiota differences caused by different housing conditions. When compared with WT mice, cohoused Opn^−/− mice still displayed significantly reduced induction of IL-6 and TNF-α at 24 hours after CLP treatment (Supplementary Figure 5A and 5B). To examine whether Opn deficiency could affect gut microbiota, the bacterial loads and composition of gut microbiota were respectively measured by quantitative polymerase chain reaction (qPCR) and high-throughput sequencing of the 16S rRNA gene in cohoused WT and Opn^−/− mice. qPCR assays with 3 pairs of universal primers of the 16S rRNA gene [27–29] showed that cohoused Opn^−/− mice had significantly lower bacterial loads in fecal and cecum samples as compared with WT mice (Figure 2A). Fluorescent in situ hybridization of bacteria in cecum with probes targeting the bacterial 16S rRNA gene confirmed the decreased bacterial loads in cohoused Opn^−/− mice (Supplementary Figure 6A). The results of 16S rRNA sequencing was assayed (Supplementary Figure 6B and 6C). To assess the richness and diversity of gut microbiota between WT and Opn^−/− mice, alpha diversity analyses were performed with Ace, Chao, and Shannon indices. These indices were significantly higher in Opn^−/− mice than WT mice (Figure 2B), indicating an increase in the richness and diversity of gut microbiota in Opn^−/− mice. Beta diversity analyses were also performed to compare the gut microbiome profiles between WT and Opn^−/− mice by principal coordinate analysis and principal components analysis. As illustrated in Figure 2C, WT and Opn^−/− mice exhibited different gut microbiota compositions. Differential abundance analysis at the phylum level revealed significant changes in Bacteroidetes, Firmicutes, Tenericutes, Cyanobacteria, and Verrucomicrobia between WT and Opn^−/− mice (Figure 2D). Among them, Bacteroidetes and Firmicutes are 2 dominant microbial phyla representing 90% of gut microbiota [30]. Specifically, the abundance of Bacteroidetes was significantly reduced in Opn^−/− mice as compared with WT mice (71.79% vs 53.65%, P < .05), whereas the abundance of Firmicutes was greatly increased in Opn^−/− mice (17.02% vs 34.33%, P < .05). The changes in Bacteroides and Firmicutes may contribute to the lower septic death in Opn^−/− mice because previous findings have shown that Bacteroides is the most common infectious cause in intra-abdominal sepsis [31], Firmicutes is negatively correlated with septic inflammation [32], and shifts of the Bacteroides-to-Firmicutes ratio were associated with mortality in patients who were critically ill [33]. These findings indicate that Opn deficiency affects gut microbiota, as evidenced by decreased bacterial loads, increased bacterial diversity, and altered phylum abundance, which may together contribute to the improved sepsis outcomes in Opn^−/− mice.

Figure 2.

Opn^−/− mice display reduced bacterial loads and distinct bacterial composition in gut microbiota as compared with WT mice. A, Bacterial loads in stool and cecum of WT and Opn^−/− mice were measured by quantitative polymerase chain reaction of the 16S rRNA gene at a noninfectious condition. When compared with WT mice, Opn^−/− mice showed reduced bacterial loads in stool (n = 6–8) and cecal (n = 4 or 5) samples. Data were analyzed by Student t test. Data are shown as mean ± SEM. B, Alpha diversity analyses with Ace, Chao, and Shannon indices. When compared with WT mice, Opn^−/− mice demonstrated significantly increased bacterial richness and diversity in gut microbiota. C, Beta diversity analyses by principal coordinate analysis (PCoA) and principal components analysis (PCA) revealed that WT and Opn^−/− mice had distinct composition in gut microbiota. D, Differential abundance analysis at the phylum level indicated significant changes in Bacteroidetes, Firmicutes, Tenericutes, Cyanobacteria, and Verrucomicrobia between WT and Opn^−/− mice (n = 8). Data were analyzed by Student t test. Data are shown as mean ± SD. *P < .05. **P < .01. OPN, osteopontin; OTU, operational taxonomic unit; WT, wild type.

Open in new tabDownload slide

Enhanced Production of AMPs in Opn^−/− Mice

As gut microbiota can be directly regulated by bacterial-specific immunoglobulin A (IgA) [34] and AMPs [35], we compared their levels between WT and Opn^−/− mice. The proportions of IgA-producing B cells (IgA⁺CD19⁺) in cecum lamina propria and IgA-coated bacteria in feces were comparable in both genotypes (Supplementary Figure 7A and 7B), suggesting that IgA may not contribute to the gut microbiota differences between WT and Opn^−/− mice. We then examined the levels of 10 AMPs, including Lcn2, S100A9, Reg3γ, DUOX2, BD1, BD2, BD3, and NOS2, in cecum tissues and cecal epithelia. Of those 10 AMPs, 8 in cecal tissues and 6 in cecal epithelia were significantly upregulated in Opn^−/− mice as compared with WT mice (Figure 3A and 3B). Meanwhile, Western blot confirmed the upregulation of LCN2, S100A9, and REG3g in cecum tissues and cecal epithelia of Opn^−/− mice (Figure 3C and 3D). To verify the increased AMP production in Opn^−/− mice, a zone-of-inhibition assay was performed. When compared with protein extracts from the cecal epithelia of WT mice, the cecal epithelial protein extracts from Opn^−/− mice exhibited significantly stronger bactericidal activity (Figure 3E). Moreover, immunohistochemical staining authenticated increased LCN2 and S100A9 expression in the cecum of Opn^−/− mice (Figure 3F). In addition to cecal tissues, AMP upregulation was observed in the intestinal tissues of Opn^−/− mice (Supplementary Figure 7C). These results indicate that Opn^−/− mice produce higher levels of AMPs, which may reshape the gut microbiota in Opn^−/− mice and lead to improved septic outcomes.

Figure 3.

Opn^−/− mice display increased AMP production as compared with WT mice. A and B, Quantitative polymerase chain reaction assays of 10 AMPs in cecum tissues and cecal epithelia (n = 5). Assays showed that Opn^−/− mice produce higher levels of AMPs as compared with WT mice. C and D, Western blot assays of LCN2, Reg3g, and S100A9 in cecum tissues and cecal epithelia. Assays confirmed higher AMP production in Opn^−/− mice as compared with WT mice. E, Zone-of-inhibition assays were performed to compare the bactericidal activity of protein extracts of cecal epithelia isolated from WT or Opn^−/− mice. The protein extracts from Opn^−/− cecal epithelia demonstrated higher bactericidal activity than those from WT cecal epithelia. F, Immunohistochemical staining of LCN2 and S100A9 in cecum. Opn^−/− mice displayed higher cecal LCN2 and S100A9 expression than WT mice. Magnification, 40×. Data were analyzed by Student t test. Data are shown as mean ± SEM. *P < .05. **P < .01. Scale bars, 50 μm. AMP, antibacterial peptide; OPN, osteopontin; RFC, relative fold change; WT, wild type.

Open in new tabDownload slide

Opn^−/− Mice Reshape Gut Microbiota From WT Mice to Improve Septic Outcomes

AMPs could control bacterial loads and shape bacterial composition of gut microbiota to affect sepsis outcomes [9]. Since Opn^−/− mice produced higher levels of AMPs, we then investigated whether Opn^−/− mice could improve sepsis outcomes by reshaping gut microbiota. WT and Opn^−/− mice were first treated with antibiotics for 3 consecutive days (days 1–3) to deplete their autochthonous microbes and then transplanted with the same volumes of fecal slurries (10%, wt/vol) from WT mice at days 4, 7, 10, and 13 (Figure 4A). When compared with vehicle control, antibiotics-mediated depletion of gut microbiota was confirmed in WT or Opn^−/− recipients by measuring the bacterial loads in fecal samples at day 4 (Figure 4B). After fecal microbiota transplantation, the fecal bacterium loads in recipients were quantified by qPCR of the 16S rRNA gene at the indicated time points (Supplementary Figure 8A), and the bacterial composition of gut microbiota in recipients was evaluated by high-throughput sequencing of the 16S rRNA gene at day 15 (Supplementary Figure 8B and 8C). When compared with WT recipients, bacterial loads of Opn^−/− recipients were significantly reduced at days 10, 13, and 15. Alpha diversity analysis with the Shannon index showed a higher diversity of gut microbiota in Opn^−/− recipients as compared with WT ones, but the analyses with the Ace and Chao indices demonstrated no difference. Beta diversity analysis revealed that Opn^−/− and WT recipients had distinct microbiota compositions. In Opn^−/− recipients, differential abundance analysis at the phylum level identified a significant decrease in Bacteroidetes (80.08% vs 62.09%, P < .05) and increase in Proteobacteria (17.73% vs 36.25%, P < .05) as compared with WT recipients (Supplementary Figure 8D). Moreover, CLP-induced inflammation was significantly reduced in Opn^−/− recipients, as evidenced by the decreased levels of IL-6 (Figure 4C), TNF-α (Figure 4D), and pulmonary neutrophil infiltration (Figure 4E). These findings suggest that the elevated AMPs in Opn^−/− mice may reshape the bacterial loads and composition of gut microbiota and therefore improve septic outcomes.

Figure 4.

Opn^−/− mice reshape gut microbiota from WT mice to improve septic inflammation. A, A schematic of gut microbiota depletion by antibiotics administration, followed by fecal microbiota transplantation to transfer WT fecal slurry into WT or Opn^−/− mice. B, The bacterial loads in fecal samples were measured by 16S rRNA quantitative polymerase chain reaction at day 4 to confirm the depletion of gut microbiota. C and D, Enzyme-linked immunosorbent assays showed that Opn^−/− recipients displayed significantly reduced serum and PLF IL-6 and TNF-α levels at 24 hours after CLP treatment. E, Immunohistochemical staining of MPO revealed that Opn^−/− recipients displayed significantly reduced pulmonary neutrophil infiltration at 24 hours after CLP treatment. Magnification, 20×. Data were analyzed by Student t test and 1-way analysis of variance with Tukey post hoc test. Data are shown as mean ± SEM. *P < .05. **P < .01. Scale bars, 50 μm. Abx, antibiotics; CFU, colony-forming units; CLP, cecal ligation and puncture; IL-6, interleukin 6; MPO, myeloperoxidase; OPN, osteopontin; PLF, peritoneal lavage fluid; TNF-α, tumor necrosis factor α; WT, wild type.

Open in new tabDownload slide

Neutralizing OPN Is Sufficient to Enhance AMP Production in WT Mice

OPN can act as a secreted protein and interact with certain cell surface receptors, such as integrin and CD44, to execute its function. Since Opn^−/− mice could produce high levels of AMPs and reshape gut microbiota to improve septic outcomes, we sought to investigate whether neutralization of OPN is sufficient to affect AMP production, change gut microbiota, and improve septic outcomes. When compared with isotype control, administration of neutralization antibody of OPN significantly decreased fecal bacterium loads (Figure 5A) but did not affect bacterial composition (data not shown), and it increased cecal expression of AMPs, such as LCN2, S100A9, Reg3g, Cryptdin1, and Ido1 (Figure 5B and 5C). Zone-of-inhibition assays showed that cecal epithelial protein extracts from OPN-neutralized mice displayed stronger bactericidal activity than those of isotype controls (Figure 5D). Moreover, as compared with isotype controls, OPN neutralization significantly reduced serum levels of IL-6 and TNF-α and decreased the pulmonary infiltration of neutrophils at 24 hours after CLP treatment (Figure 5E and 5F). These findings indicate that OPN neutralization is sufficient to induce AMP production, reduce bacterial loads, and in turn improve septic inflammation.

Figure 5.

Transient OPN neutralization is sufficient to induce AMP production, decrease bacterial loads, and improve CLP-induced septic inflammation. A, Quantitative polymerase chain reaction assays of the 16S rRNA gene showed that OPN neutralization reduced the bacterial loads in fecal samples (n = 5 or 6). B–D, Quantitative polymerase chain reaction (n = 3 or 4), Western blot, and zone-of-inhibition assays demonstrated that OPN neutralization increased AMP production in cecum tissues. E, Enzyme-linked immunosorbent assays revealed that OPN neutralization decreased serum IL-6 and TNF-α at 24 hours after CLP treatment (n = 3). F, Immunohistochemical staining of MPO indicated that OPN neutralization improved pulmonary neutrophil infiltration at 24 hours after CLP treatment. Magnification, 20×. Data were analyzed by Student t test. Data are shown as mean ± SEM. *P < .05. **P < .01. Scale bars, 50 μm. AMP, antibacterial peptide; CFU, colony-forming units; CLP, cecal ligation and puncture; IgG, immunoglobulin G; IL-6, interleukin 6; MPO, myeloperoxidase; OPN, osteopontin; PLF, peritoneal lavage fluid; RFC, relative fold change; TNF-α, tumor necrosis factor α; WT, wild type.

Open in new tabDownload slide

OPN Directly Regulates AMP Production by Modulating the AKT-FOXO3a Pathway

We established organoids from intestinal epithelia. The cultured intestinal organoids showed distinct crypt-villus architecture (Supplementary Figure 9A) and specific expression of intestine markers (Supplementary Figure 9B). To investigate whether OPN regulates AMP production directly, WT organoids were treated with OPN neutralization antibody, and Opn^−/− organoids were treated with rmOPN. OPN neutralization significantly increased the production of Lcn2 and Reg3g in WT organoids, whereas rmOPN treatment significantly reduced their expression in Opn^−/− organoids (Figure 6A and 6B), suggesting that OPN directly regulates AMP production. Previous studies have demonstrated that AMP production is mainly regulated in an NF-κB– or FOXO-dependent manner. Since our findings showed that Opn^−/− mice produce high levels of AMPs under physiologic conditions, we focused on whether OPN affects FOXO-dependent AMP production. Previous studies have indicated that activated AKT could phosphorylate FOXOs (FOXO1, FOXO3, and FOXO4) to inhibit their nuclear accumulation and transactivity [36, 37]. Interestingly, OPN has been reported to interact with presumed downstream effectors, such as integrin α_Vβ₃ [38] and CD44 [36, 37], to activate AKT. Therefore, we hypothesized that OPN may directly regulate AMP production by modulating the AKT-FOXOs signaling. To test this hypothesis, Western blot was performed to detect the phosphorylation changes of AKT and FOXOs in organoids treated with either OPN neutralization antibody or rmOPN. When compared with WT organoids, Opn^−/− organoids showed decreased phosphorylation levels of AKT and FOXO3a; phospho-FOXO3a was the only phosphorylated FOXO protein signal detectable in the lysates of intestinal organoids (Figure 6C). OPN neutralization in WT organoids decreased the phosphorylation levels of AKT and FOXO3a, whereas rmOPN treatment in Opn^−/− organoids increased their phosphorylation levels, suggesting that OPN could negatively regulate FOXO3a transactivity. Moreover, in Opn^−/− organoids, rmOPN-induced increases in AKT and FOXO3a phosphorylation could be partially reversed by treating with OPN neutralization antibody (Figure 6D). Integrin is one of the OPN receptors. Studies have shown that OPN can bind to integrin α₄β₁ to phosphorylate FOXO3a, thereby inhibiting the activity of FOXO3a. Silencing integrin β₁ significantly reduced the expression of AMPs in Opn^−/− organoids (Figure 6E). Meanwhile, rmOPN failed to induce AKT and FOXO3a phosphorylation in integrin β₁–silenced Opn^−/− organoids (Figure 6F). These findings suggest that OPN directly regulates AMP production via modulating the AKT- FOXO3a signaling pathway.

Figure 6.

OPN modulates AMP production via the AKT-FOXO3a pathway. A and B, Quantitative polymerase chain reaction assays showed that OPN neutralization at WT intestinal organoids and OPN treatment at Opn^−/− intestinal organoids respectively enhanced and inhibited the expression of LCN2 and Reg3g (n = 3). C, Western blot assays revealed that OPN neutralization at WT organoids inhibited AKT activation and FOXO3a phosphorylation, whereas OPN treatment of Opn^−/− organoids enhanced AKT activation and FOXO3a phosphorylation (n = 3). D, Western blot assays demonstrated that in Opn^−/− organoids, rmOPN treatment could enhance AKT activation and FOXO3a phosphorylation and that the addition of an OPN neutralization antibody could reverse such rmOPN-induced AKT activation and FOXO3a phosphorylation (n = 4). E, Quantitative polymerase chain reaction assays indicated that knockdown β₁ at Opn^−/− organoids and OPN treatment respectively failed to inhibit the expression of LCN2 and Reg3g (n = 3). F, Western blot assays showed that knockdown β₁ at Opn^−/− organoids and OPN treatment respectively inhibited AKT activation and FOXO3a phosphorylation. This experiment was replicated twice. Data were analyzed by Student t test and 1-way analysis of variance with Tukey post hoc test. Data are shown as mean ± SEM. *P < .05. **P < .01. AMP, antibacterial peptide; OPN, osteopontin; RFC, relative fold change; rmOPN, recombinant mouse osteopontin; WT, wild type.

Open in new tabDownload slide

DISCUSSION

Previous studies have demonstrated that OPN is upregulated in septic mouse models and that genomic depletion or protein neutralization of OPN could improve septic inflammation, organ injury, and mortality [21, 23, 24]. However, the underlying mechanism remains elusive. In this study, we for the first time demonstrated that nonhematopoietic but not hematopoietic depletion of OPN led to the improved septic inflammation, bacterial clearance, and septic outcomes (Figure 1). This finding inspired us to shift our focus from immune cells to gut microbiota to explore the mechanism by which OPN deficiency improves septic outcomes.

Gut microbiota is critical for septic onset and outcomes. Previous studies have demonstrated that prior exposure to antibiotics could disrupt the homeostasis of gut microbiota and lead to the onset of sepsis [39]. We demonstrated that Opn^−/− mice had decreased bacterial loads and distinct bacterial composition of gut microbiota as compared with cohoused WT mice, suggesting that OPN may affect sepsis outcomes by regulating gut microbiota (Figure 2). The decreased bacterial loads in Opn^−/− mice may at least in part contribute to the improved septic outcomes. Regarding bacterial composition, Opn^−/− mice showed decreased Bacteroidetes and increased Firmicutes in their gut microbiota as compared with WT mice (Figure 2D). Given that Bacteroides is the most common infectious cause in intra-abdominal sepsis [31] and Firmicutes is negatively correlated with markers of septic inflammation [32], such alterations may also contribute to the improved sepsis outcomes in Opn^−/− mice. In fact, a recent study suggested that shifts in the Bacteroides-to-Firmicutes ratio were associated with mortality in severe cases [33].

AMPs are a group of evolutionarily conserved endogenous peptides providing immediate responses to various microorganisms or viral pathogens. Because of features such as a broad spectrum of activity, effective bacterial killing, and minimal resistance, AMPs gain special interests to the development of novel agents against multidrug-resistant bacteria. However, clinical application of artificially synthesized AMPs has been greatly limited due to their instability, toxicity, and undesirable bioavailability [40, 41]. OPN depletion could induce AMP production (Figures 3, 5C and 5D, and 6; Supplementary Figure 7C), suggesting that transient regulation of OPN levels is sufficient to affect AMP production. Therefore, OPN may serve as a therapeutic target in AMP-related clinical applications.

The regulatory mechanisms of AMP production have been fundamentally elucidated in fruit fly, Drosophila melanogaster [42]. In Drosophila, bacterium-induced AMP production is mainly regulated by Toll and immune deficiency pathways, which are respectively activated by gram-positive and gram-negative bacterium infection [43]. Becker et al demonstrated that AMPs can be induced by enhanced nuclear FOXO activity in response to starvation and to insulin signaling mutation and inhibition, and this FOXO-mediated AMP production is conserved [12]. Given that Opn^−/− mice displayed enhanced AMP production under noninfectious conditions, we mainly focused on the putative regulatory interaction between OPN and FOXO signaling. By employing cultured intestinal organoids, we demonstrated that OPN could negatively regulate AMP production by activating AKT, a negative regulator of FOXO signaling [36, 37], which in turn phosphorylated FOXO3a and inhibited its transactivity (Figure 6). This regulatory interaction between OPN and FOXO signaling is consistent with a previous finding showing that OPN promoted the survival of activated T cells by inhibiting FOXO3a [44].

CONCLUSION

In the present study, we investigated the underlying mechanism by which OPN deficiency improves septic outcomes. We demonstrated that OPN depletion could induce AMP production, affect the bacterial loads and composition of gut microbiota, and improve septic outcomes. Mechanistically, we found that OPN could activate AKT and subsequently increase FOXO3a phosphorylation and inhibit FOXO3a transactivity, suggesting that OPN may modulate AMP production through the AKT-FOXO3a signaling pathway.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org/). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.

Notes

Author contributions. C. Y., X. K., X. L., and H. W. designed experiments and wrote the manuscript. C. Y. and D. X. performed most in vitro and in vivo experiments and curated data. J. J. and G. L. contributed to sequencing analysis. Y. L. and Y. G. provided methodology and performed fluorescent staining of intestinal organoids. X. S. and F. W. performed flow cytometry assays of IgA-producing B cells and IgA-coated bacteria. X. K., X. L., and H. W. supervised and conceived of the study.

Data availability statement. The data sets used and analyzed during the current study are available from the authors on reasonable request; some have already been included in this article. Sequencing information is available in figshare at https://doi.org/10.6084/m9.figshare.21385656.

Ethics approval and consent to participate. All experiments using animals were conducted according to guidelines and policies of the Institutional Animal Care and Use Committee of the Department of Liver Diseases of Shuguang Hospital (PZSHUTCM200320006).

Financial support. This work was supported by the National Natural Science Foundation of China (82070633 and 81873582 to X. K., 31870905 to H. W., 81874436 to Y. G., 82000616 to D. X.); Program of Shanghai Academic/Technology Research Leader (20XD1403700 to X. K.); and National Key Sci-Tech Special Project of China (2018ZX10725-504 to Y. G.).

References

Author notes