BMSCs ameliorate septic coagulopathy by suppressing inflammation in cecal ligation and puncture-induced sepsis

Sepsis is a complex pathological condition that comprises bacterial infection, inflammation, coagulopathy and innate immunity. Despite advances in management, sepsis still represents major healthcare problems worldwide leading to a substantial consumption of health care resources as a result of its high morbidity and mortality. In the past decades, many immune modulatory compounds have been proposed to restore homeostasis in patients with sepsis; however, no single treatment has been established to control sepsis despite extensive efforts. Therefore, there is an urgent need for effective targeted therapies for sepsis (Angus and van der Poll, 2013; Ulloa and Tracey, 2005; Kaukonen et al., 2014).

Inflammation and coagulation play pivotal roles in the pathogenesis of sepsis. Emerging evidence demonstrates the extensive crosstalk between these two pathways. Inflammation induces coagulation, which, in turn, contributes to secondary inflammation (Levi et al., 2004; Chu, 2010; Choi et al., 2006; Russell, 2006). The products and stimuli derived from the infective microbe activate inflammatory cells, including macrophages and neutrophils, causing them to release inflammatory cytokines. These cytokines then induce a series of sequential results, characterized by vascular dysfunction, abnormal endothelial permeability, disseminated intravascular coagulation, and tissue and organ injury (London et al., 2010; Rodriguez-Gaspar et al., 2001). Multiple studies have shown that pro-inflammatory cytokines are the main mediators of inflammation-activated coagulation (Wang et al., 2015; van der Poll et al., 2001). IL-6 initiates the coagulation activation; tumor necrosis factor-α (TNF-α) and IL-1 are involved in the regulation of physiological anticoagulation (van der Poll et al., 1994). In addition, the protein C system, protease-activated receptors (PARs) and the plasminogen–plasmin system, which regulate the procoagulant, also modulate the inflammation (Esmon, 2002; Coughlin, 2000; Szaba and Smiley, 2002).

Bone marrow stromal cells (BMSCs) are multipotent progenitor cells that have the capacity to self-renew and differentiate into osteoblasts, chondrocytes, adipocytes, myoblasts and fibroblasts (Chamberlain et al., 2007). It has been reported that BMSC injection is a promising cell-based therapy in various diseases (Le Blanc and Pittenger, 2005; De Miguel et al., 2012; Das et al., 2013). It is also known that transfusion of BMSCs has protective effects in mouse models of sepsis induced by cecal ligation and puncture (CLP) (Németh et al., 2009; Mei et al., 2010; Islam et al., 2012). However, whether BMSCs affect septic coagulopathy and how BMSCs ameliorate septic coagulation remains unclear and deserves to be investigated.

To assess the effect of BMSCs on CLP-induced sepsis, BMSCs were transplanted to C57BL/6 mice 6 h after CLP when mice had just recovered from the anesthesia. The mice were then euthanized at the moribund stages. The time of euthanasia was defined as the survival time. Statistical analyses revealed that BMSCs significantly elongated the survival time compared with those treated with PBS (Fig. 1A). Since bleeding time is an index for coagulation, we measured the tail bleeding time 24 h after CLP (Fig. 1B). The results showed that the group with the CLP treatment had a shortened bleeding time and that the group with BMSC treatment after CLP had a longer bleeding time compared with the group subjected to CLP alone. The results suggest that the shortened bleeding time in septic mice was reversed by BMSCs.

To determine whether the BMSC treatment alleviated inflammation induced by CLP, ELISA was used to determine the concentration of key inflammation factors in the plasma. The results showed that the plasma levels of TFPI, TNF-α and IL-1β in the CLP group were significantly higher than in the sham-operated control mice, but not in the mice treated with both CLP and BMSC injection (Fig. 1C–E). The results suggest that BMSCs blunt the inflammation induced by the CLP and protect the septic mice from the acute coagulopathy and inflammatory response.

In sepsis, pathological angiogenesis damages the endothelial barrier, which increases the secretion of multiple angiogenic factors. To determine whether injection of BMSCs alleviated endothelial damage, the proteome profiler mouse angiogenesis array was employed to measure angiogenic factors in the plasma. The results showed that the CLP group had increased plasma levels of CXCL1, CXCL10, CXCL16, MCP1 (CCL2), MMP3, MMP8 and PTX3 (Fig. 1F,G). However, in the group with CLP followed by BMSC treatment, the plasma levels of these angiogenic factors were significantly lower than those in the CLP alone group. The results further suggest that BMSCs alleviate the damage induced by CLP.

Lung inflammation and injury leading to pulmonary dysfunction is one of the common symptoms in septic shock. To determine whether BMSCs alleviated CLP-induced lung inflammation and injury in mice, lung sections were subjected to Hematoxylin and Eosin staining for histological analyses (Fig. 2A) and neutrophil elastase (NE; also known as ELANE) immunohistochemistry staining to assess the level of NE⁺ cells in the lung (Fig. 2B,C). The CLP significantly increased lung inflammation and NE+ cells in the lung. Compared with the CLP group, the group with CLP followed by BMSC injection showed a modest pulmonary inflammatory response. Consistently, CLP increased the plasma level of NE and injection of BMSCs attenuated the increase of plasma NE induced by CLP (Fig. 2D). Furthermore, fluorescent staining indicated that Ly6G-labeled neutrophils and tissue factor (TF)-positive coagulation were increased in the CLP group. However, the effect of CLP was reduced by addition of BMSCs (Fig. 2E,F). Quantitative real-time RT-PCR analyses revealed that expression of cytokines CXCL1, CXCL10, MCP1, IL-6 and IL-10, as well as inflammation-related proteins TF, MMP1, PAR1 (F2R), toll-like receptor 4 (TLR4) and intercellular cell adhesion molecule 1 (ICAM1), at the mRNA level was higher in the CLP group than in the sham- or BMSC-treated groups (Fig. 2G). Protein expression of TF, NE and PAR1, which are related to inflammation and coagulation, was increased after CLP and this increase was attenuated by BMSCs. Interestingly, the BMSC injection had no obvious effect on other biomarkers, such as antithrombin-III (ATIII) and tissue factor pathway inhibitor (TFPI) (Fig. 2H). Together, the data suggest that CLP-induced pulmonary inflammation and coagulopathy can be alleviated by BMSCs.

Sepsis increases vascular permeability by disrupting endothelial integrity, which is one of the major problems induced by sepsis. To determine whether BMSCs alleviate HUVEC damage induced by bacterial endotoxin LPS, HUVECs were co-cultured with BMSCs and then treated with LPS. The HRP penetration assay was used to determine the permeability of HUVECs. As expected, LPS treatment increased the permeability (Fig. 3A). However, the increased permeability was compromised by co-culture with BMSCs. Cell counting analyses revealed that HUVEC proliferation was inhibited by LPS and the inhibition was abolished by BMSCs (Fig. 3B). Furthermore, cell migration assay demonstrated that LPS inhibited HUVEC migration and that BMSCs significantly increased the migration activity of HUVECs even under LPS stimulation (Fig. 3C,D), whereas there was no difference between the BMSC group and BMSC+LPS group (data not shown). In addition, the endothelial cell tube formation test indicated that HUVECs formed more tubes under LPS stimulation, which was curbed by co-culture with BMSCs (Fig. 3E). Secretion of 6-keto prostaglandin F1α (6-k-PGE1α) is a normal function of HUVEC cells. Treating HUVECs with LPS reduced the secretion of 6-k-PGF1α. The inhibition was abrogated by co-culture with BMSCs (Fig. 3F). Secretion of TXB2 is increased when the integrity of endothelial cells is disrupted. Similarly, treating HUVECs with LPS increased secretion of TXB2. Co-culture with BMSCs reduced TXB2 secretion induced by LPS (Fig. 3G). Together, the data demonstrate that BMSCs protect LPS-induced damage of HUVECs.

To explore whether BMSCs affect activation of the endothelium coagulation cascade, we detected TF, NFκB-p65 (NFKB3) and JNK expression at the protein level. Results shown in Fig. 3H indicated that BMSC co-culture led to a down-regulation of TF, p65 and p-JNK expression in HUVECs.

Since macrophages promoted inflammation in sepsis, we then investigated whether BMSCs affected the ability of primary peritoneal macrophages to secrete pro-inflammatory factors. The peritoneal macrophages were treated with LPS for 6 h. Total RNA was extracted for quantitative real-time RT-PCR analyses of pro-inflammatory factor expression at the mRNA level. The results showed that treatment with LPS increased gene expression of CXCL1, CXCL10, CXCL16, MCP1, IL-1β, IL-6, IL-10 and TNF-α, as well as adhesion molecules (ICAM1 and VCAM1) and coagulation-related biomarkers (TF, MMP1 and PAR1). However, the increased expression was reduced by co-culture with BMSCs (Fig. 4A). Immunostaining showed that expression of PAR1 and TF was increased when the macrophages were treated with LPS. Co-culture with BMSCs diminished this increase (Fig. 4B). Protein expression of TF, JAK2 and p-Stat3 was enhanced by LPS, but not in the BMSC co-culture group (Fig. 4C). The results suggested that BMSCs also attenuate the inflammatory reaction in macrophages.

Coagulation dysfunction is one of the most notable symptoms of sepsis, which is a cascade of complex processes closely related to inflammation (Levi et al., 2004). Since stromal cells, especially BMSCs, have anti-inflammatory and immune modulation effects, infusion of stromal cells, especially BMSCs, is emerging as a prominent therapy for a range of diseases and pathologies (Le Blanc and Pittenger, 2005; De Miguel et al., 2012; Das et al., 2013). Herein, we reported that BMSCs alleviate the surge of plasma levels of pro-inflammation TNF-α and IL-1β after CLP. In addition, CLP-induced lung injury and overexpression of chemokines and cytokines CXCL1, CXCL10, MCP1, IL-6, IL-10 and ICAM1 in the lung were attenuated by the BMSC treatment. Moreover, co-culture with BMSCs restricted the ability of macrophages to respond to LPS with respect to overexpression of inflammatory factors CXCL1, CXCL10, CXCL16, MCP1, IL-1β, IL-6, IL-10, TNF-α, ICAM1 and VCAM1. Since uncontrolled inflammation activates platelets, induces endothelial disorder, disseminates intravascular coagulation, and leads to tissue damage and organ failure (Fig. 5), the data demonstrate that BMSCs have the potential for treating CLP-induced sepsis.

The vascular dysfunction and coagulation imbalance are symptoms of sepsis pathogenesis (Lee and Slutsky, 2010). It has been reported that BMSCs protect endothelial cells from inflammation attack (Rahbarghazi et al., 2013; dos Santos et al., 2012). We previously also reported that BMSCs ameliorate sepsis-induced damage by regulating expression of thrombomodulin (TM) and endothelial protein C receptor (EPCR) (Tan et al., 2016). The data in this report further demonstrate the protective effect of BMSCs against septic coagulopathy. Angiogenesis is an intricate and well balanced process (Kim et al., 2013). Our results showing that BMSC treatment reduced septic angiogenesis both in vivo and in vitro suggest that the damage of the endothelium in sepsis is reduced by BMSCs. This is further supported by the in vitro data showing that LPS-induced damage in migration, proliferation and permeability of endothelial cells was alleviated by co-culture with BMSCs.

TF is an inflammation-activated membrane receptor that initiates the coagulation pathway and promotes angiogenesis (Mackman, 2004). PARs (PAR1–PAR4) belong to a family of transmembrane G-protein-coupled receptors (Zhao et al., 2014). PAR1 is a thrombin receptor and serves as a receptor for TF factor VIIa complex and factor Xa (Pawlinski, et al., 2004; Reinhardt et al., 2014), and it also participates in mediating activated protein C signaling in endothelial cells (Mosnier and Griffin, 2003). Emerging evidence suggests that MMP1 plays a pivotal role in inflammation and coagulation by activating PAR1 in sepsis (Tressel et al., 2011). Our results showed that expression of TF, MMP1 and PAR1 was regulated by BMSCs both in vivo and in vitro. This further demonstrates that BMSCs alleviate inflammation and coagulation induced by CLP.

Neutrophils actively participate in sepsis pathogenesis and the neutrophil extracellular traps promote thrombin generation (Gould et al., 2014; Caudrillier et al., 2012; Sreeramkumar et al., 2014; Kolaczkowska and Kubes, 2013). Moreover, the NE level has been shown to be a risk factor for the severity of sepsis (Mihara et al., 2013). Our results showing that BMSCs cleared the pulmonary neutrophil infiltration as well as causing a reduction in the boost in plasma NE levels and lung NE expression induced by CLP further demonstrated that BMSCs modulate neutrophil recruitment in septic mice.

Taken together, we show both in vivo and in vitro that BMSCs effectively ameliorate septic coagulopathy, alleviate vascular damage, reduce inflammation, attenuate acute lung injury and improve the survival rate. Our study demonstrates that coagulation protection by BMSC injection is a promising therapy for treating sepsis.

Male C57BL/6 mice (6 to 7 weeks old) were obtained from Shanghai SLAC Laboratory Animal Limited Liability Company (Shanghai, China). The animals were maintained in a pathogen-free facility with ambient temperature of 23±3°C, relative humidity of 55±10%, and 12 h light:12 h dark cycle. All animal procedures were approved by the Institutional Animal Care and Use Committee of the Wenzhou Medical University.

The CLP mouse model was prepared as described (Dejager et al., 2011). A sham group was operated similarly without ligation and puncture. Subsequently, the mice were subcutaneously injected with 1 ml pre-warmed saline. After 6 h, the mice were injected with sterile PBS or BMSCs (1×10⁶/0.3 ml PBS) via tail vein. For the survival experiments, the mice were closely observed every 12 h for 4 days and euthanized at the moribund stage. For other experiments, the mice were sacrificed 24 h after the CLP.

Bleeding assay was performed in mice 24 h after surgery by tail tip amputation, and then the tail was immersed in saline at 37 °C. The bleeding time was recorded (Ye et al., 2014).

The levels of 6-k-PGF1α and thromboxane B2 (TXB2) in the supernatant were examined using ELISA kits (eBioscience, San Diego, CA) following the manufacturer's instructions. IL-1β, TNF-α, NE and TFPI levels in the mice plasma were measured by specific ELISA kits (R&D Systems).

Plasma collected from three groups of mice were probed with the Mouse Angiogenesis Antibody Proteome Profiler Array (R&D Systems) according to the manufacturer's instructions. For in vitro study, HUVECs (15×10⁴/well) were seeded in the lower chamber of 24-well Transwell plates (pore size of 0.4 μm; Corning) which were covered with 120 μl polymerized Matrigel (BD Biosciences) and the same number of BMSCs were seeded in the upper chamber. Samples without BMSCs in the upper Transwell were used as a control. After LPS (1 μg/ml) treatment for 3.5 h, images were captured with an Olympus CKX41 light microscope.

HUVECs were purchased from Chi Scientific (Jiangsu, China) and cultured in RPMI 1640 medium (Gibco, Life Technologies, Germany) supplemented with 10% fetal bovine serum (FBS) (Sigma-Aldrich) and 1% penicillin/streptomycin (Solarbio, Beijing, China) at 37°C in a 5% CO₂ humidified chamber. BMSCs derived from C57BL/6 mice were obtained from Cyagen Biosciences (Guangzhou, China) and cultured in the specialized medium with 10% FBS according to the manufacturer's instructions as described (Tan et al., 2016). Peritoneal macrophages were isolated from 6- to 7-week-old C57BL/6 mice. Briefly, mice were injected peritoneally with 2–3 ml Starch broth (Sigma), and then the peritoneal cavity was washed with RPMI 1640 medium several times to isolate peritoneal exudate cells. The cells were cultured for 2 h and the adhered cells were harvested for experiments. No bacterial were observed under a microscope. No mycoplasma were detected with the mycoplasma detection kit.

For Transwell co-culture, HUVECs (5×10⁵/well) or peritoneal macrophages (5×10⁵/well) were seeded to the lower chamber and an equal number of BMSCs to the upper chamber of 6-well Transwell plates (pore size of 0.4 μm; Corning). HUVECs (5×10⁵/well) or peritoneal macrophages (5×10⁵/well) seeded in 6-well plates without BMSCs were used as a control. After overnight incubation, 1 μg/ml LPS (from Escherichia coli O111:B4, Sigma) was added to the LPS and LPS+BMSCs groups, with PBS (HyClone) as the solvent control. After a 6 h incubation, HUVECs or peritoneal macrophages were collected for subsequent analyses.

HUVECs (5×10⁴/well) were seeded in the upper chamber and BMSCs in the bottom of a 24-well Transwell plate (pore size of 0.4 μm; Corning). After the cells were confluent, LPS (1 μg/ml) was added to the apical medium for 6 h followed by replacement with 200 μl HRP solution (50 ng/ml). The medium of the lower chamber was collected for penetrating HRP detection after incubation for 1 h. Briefly, 20 μl medium was transferred to a 96-well plate followed by addition of 150 μl substrate (400 μg/ml O-phenylenediamine in 0.05 M citric acid and 0.1 M phosphate with 0.012% hydrogen peroxide). After incubation at 37°C for 10 min, the reaction was terminated by adding 50 μl of 2 M sulfuric acid. The OD value was measured using a microplate reader (Molecular Devices, Hercules, CA).

Cell Counting Kit-8 (CCK-8) assay (Dojindo, Japan) was used to determine the viability of cells. BMSCs or HUVECs were seeded in a 96-well plate at 2000 cells per well. After 24 h, the supernatant of BMSCs was transformed into HUVECs plates as BMSC-treated group, LPS (1 μg/ml) was added to the medium for 6 h. Then, the supernatant was replaced with fresh medium in each well and CCK-8 solution (20 µl) was added for incubation for 1 h at 37°C. OD₄₅₀ values were measured using a microplate reader (Molecular Devices).

BMSCs (2×10⁵/well) suspended with 0.5 ml medium were seeded in the lower chamber of 24-well Transwell plates (pore size of 5 μm; Corning), which had the same number of HUVECs suspended with 0.2 ml medium in the upper chamber. Samples with no BMSCs in the lower chambers were used as a control. After overnight incubation, LPS (1 μg/ml) was added to the medium followed by incubation for 6 h. The non-migrating cells were wiped away with a cotton swab. The membranes were fixed with 4% paraformaldehyde for 20 min, stained with 0.1% Crystal Violet (Solarbio, Beijing, China) for 30 min, and washed three times with PBS. Cells were counted under a light microscope (Nikon, Tokyo, Japan).

Lung, HUVECs and peritoneal macrophages were homogenized and lysed in the RIPA buffer (Thermo Scientific) containing 1% PMSF and 1% protein phosphatase inhibitor mixture (P1260, Applygen, Beijing, China). The protein concentrations were determined using the BCA kit (Thermo Scientific). Lysates containing 40 μg proteins were separated on 8–12% SDS-PAGE gels and transferred to polyvinylidene difluoride (PVDF) membranes (EDM Millipore, Billerica, MA). The membranes were blocked with 5% non-fat milk in TBST (1×TBS, 0.1% Tween 20) for 1 h at room temperature. The membranes were then incubated with the following primary antibodies overnight at 4°C. After washing with TBST three times, the membrane were then incubated with goat anti-rabbit HRP-conjugated polyclonal antibody (1:10,000; Bio-Rad, Cat# 170-6515) for 1 h at room temperature. The primary antibodies were: rabbit anti-NE (1:200; Abcam, Cat# ab68672), anti-TF (1:100; Abcam, Cat# ab151748), anti-ATIII (1:500; Santa Cruz, Cat# sc-271987) and anti-TFPI (1:200; Santa Cruz, Cat# sc-365920), and rabbit anti-phospho-cJun N-terminal kinase (p-JNK) (1:1000; Cat# 4668), anti-β-actin (1:2000; Cat# 4970), anti-Janus kinase 2 (JAK2) (1:1000; Cat# 3230), phospho-signal transducer and activator of transcription (p-Stat3) (1:1000; Cat# 9145), p65 (1:1000; Cat# 8242) and phospho-p65 (p-p65) (1:1000; Cat# 3033) NF-κB antibodies, all from Cell Signaling Technology. The ECL Plus Chemiluminescent Reagent (Thermo Scientific) was used to visualize the proteins. The specific bands were quantitated by ImageJ (NIH, USA).

The Trizol reagent (Invitrogen, CA) was used to extract total RNA from the lung and peritoneal macrophages. The GoScript Reverse Transcription System kit (Promega) was used for first-strand cDNA synthesis according to the manufacturer's protocols. The LightCycler (Roche Diagnostics, Mannheim, Germany) and SYBR Green (Roche Diagnostics, Mannheim, Germany) were applied to detect gene expression with the primer sequences listed in Table S1.

Lung tissues were fixed in 4% paraformaldehyde, dehydrated, embedded and sectioned at 4 μm thickness for H&E staining. For histochemical staining, the lung sections were dewaxed, rehydrated and antigen retrieved by autoclave in citrate buffer (0.01 M, pH 6.0) for 10 min. The endogenous peroxidase was quenched by incubation with 3% hydrogen peroxide for 30 min. The sections were blocked with preimmune goat serum for 30 min and incubated with rabbit anti-NE antibody (1:200; Abcam, Cat# ab68672) overnight at 4°C. After washing with PBS, the sections were incubated with horseradish peroxidase (HRP)-conjugated goat anti-rabbit antibody (Zhongshan Golden Bridge Biotechnology, Beijing, China) for 30 min at room temperature. Thereafter, the slides were stained with a DAB kit (Sigma) and visualized under a microscope (Nikon, Tokyo, Japan). The intensity and localization of the positive staining was assessed by the IPP 6.0 True Color Image Analysis System. Three random areas were selected from each section for measurement.

For tissue immunofluorescence staining, the pre-treated lung sections were blocked with 5% BSA in PBS-T (PBS with 0.1% Tween 20) and incubated overnight at 4°C with the primary antibodies against Ly6G (1:100, Santa Cruz, Cat# sc-53515) and TF (1:100; Abcam, Cat# ab151748). The sections were then incubated for 1 h at room temperature (RT) with FITC (1:200; Abcam, Cat# ab6717) and nuclei were stained with DAPI (1:200; Cell Signaling Technology, Cat# 4083) for 5 min.

For cell immunofluorescence staining, macrophages were fixed with 4% paraformaldehyde at room temperature for 20 min, and then washed three times with PBS, permeabilized with 0.2% Triton X-100 for 15 min, blocked with 5% BSA at RT for 30 min, and incubated with primary antibodies against PAR1 (1:100, Santa Cruz, Cat# sc-5605) and TF overnight at 4°C. The slides were incubated with FITC for 1 h and then DAPI was added for 5 min. Images were taken with a laser scanning confocal microscope (LSCM, Leica) in three random areas.

Kaplan–Meier survival was analyzed using the log-rank test by SPSS software version 20.0 (SPSS). Categorized variables were compared by one-way Analysis of Variance (ANOVA) or Student's t-test. P<0.05 was considered statistically significant.