The Modulation of Regulatory T Cells via HMGB1/PTEN/β-Catenin Axis in LPS Induced Acute Lung Injury

Sepsis-induced acute lung injury (ALI)/acute respiratory distress syndrome (ARDS) remains the leading complication for mortality caused by bacterial infection. The regulatory T (Treg) cells appear to be an important modulator in resolving lung injury. Despite the extensive studies, little is known about the role of macrophage HMGB1/PTEN/β-catenin signaling in Treg development during ALI. Objectives: This study was designed to determine the roles and molecular mechanisms of HMGB1/PTEN/β-catenin signaling in mediating CD4+CD25+Foxp3+ Treg development in sepsis-induced lung injury in mice. Setting: University laboratory research of First Affiliated Hospital of Anhui Medical University. Subjects: PTEN/β-catenin Loxp and myeloid-specific knockout mice. Interventions: Groups of PTENloxp/β-cateninloxp and myeloid-specific PTEN/β-catenin knockout (PTENM−KO/β-cateninM−KO) mice were treated with LPS or recombinant HMGB1 (rHMGB1) to induce ALI. The effects of HMGB1-PTEN axis were further analyzed by in vitro co-cultures. Measures and Main Results: In a mouse model of ALI, blocking HMGB1 or myeloid-specific PTEN knockout (PTENM−KO) increased animal survival/body weight, reduced lung damage, increased TGF-β production, inhibited the expression of RORγt and IL-17, while promoting β-catenin signaling and increasing CD4+CD25+Foxp3+ Tregs in LPS- or rHMGB-induced lung injury. Notably, myeloid-specific β-catenin ablation (β-cateninM−KO) resulted in reduced animal survival and increased lung injury, accompanied by reduced CD4+CD25+Foxp3+ Tregs in rHMGB-induced ALI. Furthermore, disruption of macrophage HMGB1/PTEN or activation of β-catenin significantly increased CD4+CD25+Foxp3+ Tregs in vitro. Conclusions: HMGB1/PTEN/β-catenin signaling is a novel pathway that regulates Treg development and provides a potential therapeutic target in sepsis-induced lung injury.


INTRODUCTION
Sepsis is a systemic inflammatory response syndrome which may result in acute lung injury (ALI) and/or acute respiratory distress syndrome (ARDS) (1). ARDS is a type of respiratory failure characterized by rapid onset of widespread inflammation in the lungs, symptoms include shortness of breath, rapid breathing, and bluish skin coloration (2). Despite recent progress in developing many pharmacological interventions for ALI/ARDS, there have been no successful clinical trials for drugs treating these disorders, implying that there are complex molecular mechanisms in sepsis-driven inflammatory responses.
High-mobility group box 1 protein (HMGB1), a highly conserved and ubiquitous DNA binding nuclear protein, is a key mediator during inflammatory responses in sepsis (3). HMGB1, as an innate "danger signal" (alarmin), plays a key role in the initiating innate and adaptive immune response (4)(5)(6). As a late mediator, HMGB1 can be actively released from endotoxin-stimulated macrophages following lipopolysaccharide (LPS) and by TNF-α or IL-1β stimulation. Blockade of HMGB1 via antibody targeting protects against LPS lethality in mice, whereas administration of HMGB1 in mice results in developing endotoxemia and lethality (7). HMGB1 contributes to the endotoxin-induced ALI through activating NF-κB translocation, increasing levels of proinflammatory cytokines, and enhancing lung permeability (8)(9)(10). Extracellular HMGB1 augmented autoimmune response through stimulating dendritic cell maturation and macrophage activation, whereas HMGB1 deficiency resulted in increasing the number of lymph node CD4 + Foxp3 + regulatory T (Treg) cells during inflammatory response (11). Moreover, disruption of HMGB1 promotes the ability to induce Treg and enhances antitumor immunity (12).
Recently, CD4 + CD25 + Foxp3 + Tregs have been shown to be crucial for the resolution of endotoxin-induced lung injury via both TGF-β-dependent and -independent pathways (13). TGF-β induces Treg-mediated suppressive activity and Foxp3 expression (14,15). The development and survival of CD4 + CD25 + Tregs in vivo was depressed by the increased phosphatase and tensin homolog deleted on chromosome ten (PTEN) activity via distinct IL-2 receptor (IL-2R) signaling, which is associated with downstream mediators of PI3K (16). Deficiency of myeloid PTEN increases PI3K signaling and reduces endotoxin-induced inflammatory response and lung injury (17). Indeed, loss of PTEN leads to an increasing nuclear accumulation of β-catenin (18) and promotes PI3K, which P3 and activates downstream PDK1 and Akt (19). Increasing phosphorylation of Akt by PDK1 enhances Akt activity and facilitates Treg induction (20), whereas deletion of PDK1 in T cells results in reducing Treg numbers in vitro and in vivo (21). Thus, the modulation of Treg development might involve in multiple pathways during lung inflammation and injury.
Using a well-established model of lung injury and an in vitro co-culture system, we identified a novel regulatory pathway of HMGB1/PTEN/β-catenin signaling on Treg induction during inflammatory response. We demonstrated that HMGB1 promoted lung inflammation through activating myeloid PTEN-mediated innate immunity. Lacking myeloid PTEN ultimately resulted in promoting β-catenin activation and TGF-β production, which in turn induced CD4 + CD25 + Foxp3 + Tregs and suppressed endotoxin-mediated inflammation in the lung. Our data document that HMGB1/PTEN/β-catenin signaling is critical for development of Tregs in the resolution of sepsis-induced lung injury.

Mice
The floxed β-catenin (β-catenin flox ) mice (The Jackson Laboratory, Bar Harbor, ME), and the mice expressing Cre recombinase under the control of the Lysozyme M (LysM) promoter (LysM-Cre; The Jackson Laboratory) were used to generate myeloid-specific β-catenin knockout (β-catenin M−KO ) mice. In brief, homozygous β-catenin flox mice were interbred with homozygous LysM-Cre mice, and the heterozygous offspring were then backcrossed to the homozygous β-catenin flox mice to generate β-catenin M−KO (LysM-Cre-β-catenin flox ) mice. The C57BL/6 wild-type (WT) and PTEN flox mice were purchased from The Jackson Laboratory (Bar Harbor, ME). The expression of β-catenin was detected in spleen and myloid cells, respectively ( Figure S1). The myeloid-specific PTEN knockout (PTEN M−KO ) mice were generated as described (22). Mouse genotyping was performed by using a standard protocol with primers described in the JAX Genotyping protocols database, and the expression of PTEN was detected as described (22). All animals were housed in animal facility under specific pathogen-free conditions. Animals at 8-10 weeks of age were used in all experiments.

Mice Treatment
To establish the animal model of ALI, mice were anesthetized with i.p. ketamine (150 mg/kg) and acetylpromazine (13.5 mg/kg), and then an incision (1-2 cm) was made on the animal neck to expose the trachea. A 20-gauge catheter was inserted into the lumen of trachea. 50 µl of LPS (Escherichia coli 055:B5; Sigma-Aldrich, 100 µg/mouse), diluted in sterile water was instilled via the catheter. Sterile water was used in the control group (8-10 mice per group) (13). To determine the role of HMGB1 during LPS-induced ALI, mice were instilled with 100 µg/mouse of anti-HMGB1 (Product# 326052233, Shino-TEST Co, Tokyo, Japan) immediately after LPS instillation. Control mice received the same volume of saline solution or control IgG (Sigma-Aldrich). To generate mouse model of endotoxin-induced sepsis, mice were injected with LPS (750 µg/mouse, i.p.) as described (23). In some experiments, mice were administrated with recombinant HMGB1 (rHMGB1, 50 µg/mouse, i.p., product# 4652, Sigma-Aldrich) or vehicle PBS. Since previous reports showed that maximal lung injury and HMGB1 expression occurred between 12 and 48 h after LPS instillation (24), all animal studies were executed at 24 h after LPS, rHMGB1, anti-HMGB1, control IgG or saline treatment.

Analysis of the Permeability Index
The permeability index, reflexing the damage of alveolar epithelial and endothelial permeability, was evaluated by administrating human serum albumin (i.v. 25 µg; Signa-Aldrich, MO) 1 h prior to sacrificing the animal. The blood and BALF were collected at the time of sacrifice. ELISA assay was performed to measure the level of human albumin concentration using a human serum albumin ELISA kit (Cayman Chemical, Ann Arbor, MI). The pulmonary permeability index was defined as the human albumin concentration in BAL fluid/serum ratio.

Analysis of Bronchoalveolar Lavage Fluid (BALF)
The mice were anesthetized before exposure of the trachea. After the catheter was inserted into the lumen of trachea, the lungs were then lavaged 3 times with 0.8 ml of sterile saline. The total collected lavage averaged 1.4-1.7 ml/mouse. BALF was centrifuged at 800 × g for 10 min at 4 • C. The cell-free supernatants were stored at −80 • C for later analysis. The cell pellet was re-suspended in PBS and counted by a hemacytometer. The differential staining was performed with Diff-Quik staining solutions to count enriched alveolar macrophages as described (25).

Myeloperoxidase Activity Assay
The presence of myeloperoxidase (MPO) was used as an index of lung neutrophil accumulation as described (26). The frozen tissue samples were homogenized and separated by centrifugation. Supernatants were analyzed for MPO activity by spectrophotometry at 655 nm, and the change in absorbance was measured. One unit of MPO activity was defined as the quantity of enzyme degrading 1 µmol peroxide/min at 25 • C per gram of tissue.
Frontiers in Immunology | www.frontiersin.org transcription to cDNA was performed by using SuperScript III First Strand Synthesis System (Invitrogen). Quantitative realtime PCR was performed using the DNA Engine with Chromo 4 Detector (MJ Research, Waltham, MA). In a final reaction volume of 25 µl, the following were added: 1 × SuperMix (Platinum SYBR Green qPCR Kit; Invitrogen, San Diego, CA) cDNA and 10 µM of each primer. Amplification conditions were: 50 • C (2 min), 95 • C (5 min), followed by 40 cycles of 95 • C (15 s) and 60 • C (30 s). Primer sequences used for the amplification of TNF-α, TGF-β, IL-17A, IL-23, RORγt, Foxp3, and HPRT are shown in Supplementary Table 1. Target gene expressions were calculated by their ratios to the housekeeping gene HPRT.

In vitro Transfection and Treatments
After 24 h cell culture, 1 × 10 6 macrophages/well were transfected with 100 nM of HMGB1 siRNA or non-specific control siRNA using lipofectamine 2000 reagent (Invitrogen), and incubated for 24 h. Non-specific (NS) siRNA as a control.
In some experiments, cells were pretreated with 10 µg/ml of rHMGB1 or 10 µg/ml of anti-HMGB1 for 24 h, and then were supplemented with 1 µg/ml of LPS for additional 6 h. The HMGB1 siRNA and control siRNA were purchased from Santa Cruz Biotechnologies (Santa Cruz, CA).

Macrophage/T Cell Co-cultures
The HMGB1 siRNA-transfected macrophages or macrophages isolated from WT, PTEN flox , PTEN M−KO , β-catenin flox , and βcatenin M−KO mice were suspended at 5 × 10 5 cells/ml and cultured on 60 mm plates. After the cells were stimulated with LPS (1 µg/ml) for 6 h, spleen T cells were then added into cultures at a macrophage/T cell ratio of 1:10 as described before (27). The co-cultured cells were incubated for 24 h, and then macrophages and spleen T cells were harvested for the Western blots, real-time PCR, and flow cytometry analysis.

Statistical Analysis
All experiments were repeated three times. Data are expressed as mean±SD and analyzed by Permutation t-test and Pearson correlation. Per comparison two-sided p-values <0.05 were considered statistically significant. Multiple group comparisons were performed using one-way ANOVA with the posthoc test. The body weight loss was analyzed by using student's t-test. All analyses were made using SAS/STAT software, version 9.4.

Myeloid Cell-Specific PTEN Is Critical for the HMGB1-Mediated Inflammatory Response in Acute Lung Injury
To determine whether myeloid cell-derived PTEN plays a role in HMGB1-mediated inflammatory response during lung injury, we used myeloid cell-specific PTEN knockout (PTEN M−KO ) mice as described (22). Indeed, increased animal survival was observed in PTEN M−KO mice, but not in PTEN flox control mice (Figure 3A  p < 0.01). These findings suggest that myeloid PTEN is a critical mediator for HMGB1-induced inflammatory response during ALI.

DISCUSSION
In this study, we have demonstrated, for the first time, that the HMGB1/PTEN/β-catenin signaling represents a novel regulatory pathway to induce CD4 + CD25 + Foxp3 + Tregs in sepsis-induced lung injury.
Using the animal model of ALI, we found instillation of LPS triggered systemic inflammatory response and induced ALI, which was accompanied by induction of HMGB1. Though the exacerbated lung damage was shown in LPS instilled lungs, neutralization of HMGB1 with anti-HMGB1 antibody provided significant protection against ALI as evidenced by increasing animal survival and decreasing pulmonary edema. These findings are consistent with previous reports that intratracheal instillation of live bacterial or HMGB1 mediates an acute inflammatory response characterized by the development of pulmonary edema and increased intrapulmonary production of proinflammatory cytokines (8,17,29).
Numerous studies have revealed the ability of CD4 + CD25 + Foxp3 + Tregs to control immune responses in lung injury (13,(30)(31)(32). In a mouse model of LPS-induced ALI, we found that increased HMGB1 levels mitigated the accumulation of CD4 + CD25 + Foxp3 + Tregs leading to exacerbated lung damage. Interestingly, increasing HMGB1 release and protein expression enhanced PTEN activation on alveolar macrophages after LPS instillation. However, neutralization of HMGB1 suppressed PTEN, which was accompanied by increased CD4 + CD25 + Foxp3 + Tregs and reduced IL-17A in LPS-treated mice. Consistent with previous reports that deletion of PTEN enhanced the expansion of CD4 + CD25 + Tregs (33), our results indicate that PTEN might serve as a negative regulator of Treg peripheral homeostasis during lung inflammation.
Further evidence of PTEN-mediated modulation of Tregs in ALI was obtained from myeloid cell-specific PTEN knockout (PTEN M−KO ) mice. We found that, in contrast to the PTEN flox mice, PTEN M−KO mice treated with LPS or rHMGB1 had reduced lung injury, neutrophil accumulation, proinflmmatory mediators, and increased animal survival. Moreover, myeloid PTEN deficiency increased β-catenin expression and phosphorylation of PDK1 and Akt on macrophages, accompanied by increased peripheral Tregs and Foxp3 expression yet decreased RORγt and IL-17A. Since increasing release of HMGB1 induced macrophage PTEN activation, while deleting myeloid PTEN promoted Tregs, we believe that PTEN is a mediator in the modulation of innate and adaptive immunity during lung inflammation. Indeed, alveolar macrophages are essential for the initiation of innate immune response by binding the toll-like receptors (TLRs) (34). In response to TLRs, PTEN activation on macrophages triggers inflammatory response via regulating PI3K signaling (35,36). Notably, our current data demonstrated that myeloid PTEN deficiency promoted β-catenin activation, consistent with our previous report that PTEN-mediated β-catenin signaling regulated Foxo1-TLR4 activation in lung inflammation (37), suggesting the endogenous innate immune signaling most likely contributes to the Treg induction. Indeed, expression of stabilized β-catenin controls Treg development and survival (38). Activation of β-catenin regulates inflammatory response and promotes anti-inflammatory mediator (39). Thus, our findings implicate that disruption of macrophage HMGB1 or PTEN, and activation of β-catenin may be a key pathway in the regulation of Treg development during lung injury.
The mechanisms underlying the macrophage HMGB1/PTEN/β-catenin signaling-mediated Treg induction appear to be complex during ALI. Our data showed that HMGB1 blockade or PTEN loss increased TGF-β release. However, reduced TGF-β release was observed from β-catenin deficient-macrophages in response to rHMGB1 stimulation. This is consistent with previous report that β-catenin was required for the TGF-β production to regulate immunity during inflammatory response (39). Indeed, TGF-β is a potent regulator of the immune and inflammatory system. In vitro stimulation of naïve CD4 + T cells in the presence of TGF-β increased the expression of CD4 + CD25 + Foxp3 + associated with in vivo suppressive activity during lung inflammatory response (40). Disruption of TGF-β impaired the development of Foxp3 + Tregs and may lead to the multifocal inflammatory cell infiltration and multiorgan failure in mice (28,41). Moreover, TGF-β inhibited RORγt activity and Th17 cell differentiation in human CD4 + T cells (42). TGF-β-induced Foxp3 inhibited Th17 cell differentiation by regulating RORγt function (43). TGF-β promoted the development of Treg and expansion Foxp3 + -expressing CD4 + CD25 + Tregs in vivo (44,45). Lung-resident tissue macrophages can generate Foxp3 + Tregs through increasing TGF-β expression (46). Consistent with this notion, we found increased TGF-β expression and secretion by alveolar macrophages were accompanied by increased CD4 + CD25 + Foxp3 + Tregs and reduced RORγt/IL-17A after anti-HMGB1 treatment or myeloid PTEN deletion in our animal models. This implies that TGF-β may be essential for the induction of CD4 + CD25 + Foxp3 + Tregs during HMGB1induced inflammatory response. On the other hand, we found HMGB1 knockdown markedly inhibited macrophage PTEN expression in the co-culture system. This is consistent with deletion of myeloid PTEN, which increased the expression of PDK1, Akt, and β-catenin. Although PTEN deficiency increased the frequency of CD4 + CD25 + Foxp3 + Tregs, ablation of myeloid β-catenin resulted in reduced CD4 + CD25 + Foxp3 + Tregs and increased RORγt/IL-17A. Indeed, our previous study has shown that disruption of PTEN increased β-catenin, which in turn promoted PI3K/Akt signaling to native feedback to regulate TLR4-driven inflammatory response (47). Increased β-catenin activity enhanced TGF-β production on macrophages, whereas β-catenin deficiency lost the ability to produce TGF-β, myeloid cell motility and adhesion leading to impairing tissue repair (48). Hence, the HMGB1/PTEN/β-catenin signaling regulates Treg induction through multiple signaling pathways. Recent works indicated that PDK1, a downstream of PI3K signaling, plays an important role in the regulation of Treg function (21). PDK1 deficiency suppressed Treg accumulation while increasing IL-17-expressing population leading to enhancing inflammatory response (21). Activation of Akt by PDK1 phosphorylation promoted Tregs and enhanced their suppressive capacity to the Th17 cell differentiation (20). Furthermore, increased Akt phosphorylation enhanced β-catenin transcriptional activity (49). Activation of β-catenin is essential for the stimulation of Treg induction while inhibition of inflammatory T cells (39). These data are consistent with our results that activation of PDK1/Akt/β-catenin enhanced Treg induction and suppressed IL-17A transcription regulated by RORγt in vitro and in vivo. Although our current study was based on the primary ALI and it might have some modified signaling pathways with secondary ALI (systemic inflammation), our findings suggest that HMGB1/PTEN/β-catenin signaling is critical to contribute to the induction of CD4 + CD25 + Foxp3 + Tregs in sepsis-induced lung injury.
In the present study, we observed that HMGB1 can be induced in endotoxin-stimulated macrophages during sepsis. HMGB1 induction activates PTEN and inhibits PI3K/PDK1/Akt leading to suppressed β-catenin activity, which then decreases TGF-β release from macrophages, results in diminished Foxp3 + Treg induction. Blockade of HMGB1 or macrophage PTEN deletion activates PI3K/PDK1/Akt and β-catenin signaling, which in turn enhances macrophage TGF-β leading to increased Foxp3 Treg induction while inhibiting Th17 cell differentiation during sepsisinduced lung injury.
In conclusion, the macrophage HMGB1/PTEN/β-catenin signaling displays a distinct capacity to regulate the development of CD4 + CD25 + Foxp3 + Tregs during lung inflammation. Induction of Tregs ultimately alleviated inflammatory response and facilitated resolution of lung injury. By identifying the regulatory pathway of HMGB1/PTEN/β-catenin signaling on Treg induction, our studies provide the rationale for novel therapeutic strategies for treating sepsis-induced lung injury.

DATA AVAILABILITY
The raw data supporting the conclusions of this manuscript is available, without undue reservation, to any qualified researcher.

ETHICS STATEMENT
The animal study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health. The study protocol were approved by the Institutional Animal Care and Use Committee of Anhui Medical University (No: LLSC2013007).

AUTHOR CONTRIBUTIONS
MZ contributed to the experimental design, performed research, analyzed data, and wrote the first draft of manuscript. MD, RT, HL, ZG, and ZJ collected and analyzed the human samples. HF wrote and revised the manuscript. CL performed in vitro experiments. X-LC and BK contributed to the study concept, research design, data analysis, and finalized the manuscript.