Bone marrow stromal cells attenuate sepsis via prostaglandin E(2)-dependent reprogramming of host macrophages to increase their interleukin-10 production

Abstract

Sepsis causes over 200,000 deaths yearly in the US; better treatments are urgently needed. Administering bone marrow stromal cells (BMSCs -- also known as mesenchymal stem cells) to mice before or shortly after inducing sepsis by cecal ligation and puncture reduced mortality and improved organ function. The beneficial effect of BMSCs was eliminated by macrophage depletion or pretreatment with antibodies specific for interleukin-10 (IL-10) or IL-10 receptor. Monocytes and/or macrophages from septic lungs made more IL-10 when prepared from mice treated with BMSCs versus untreated mice. Lipopolysaccharide (LPS)-stimulated macrophages produced more IL-10 when cultured with BMSCs, but this effect was eliminated if the BMSCs lacked the genes encoding Toll-like receptor 4, myeloid differentiation primary response gene-88, tumor necrosis factor (TNF) receptor-1a or cyclooxygenase-2. Our results suggest that BMSCs (activated by LPS or TNF-alpha) reprogram macrophages by releasing prostaglandin E(2) that acts on the macrophages through the prostaglandin EP2 and EP4 receptors. Because BMSCs have been successfully given to humans and can easily be cultured and might be used without human leukocyte antigen matching, we suggest that cultured, banked human BMSCs may be effective in treating sepsis in high-risk patient groups.

Figures

Figure 1

Effect of intravenous injection of BMSCs on the course of sepsis after CLP. (a) Survival curves of mice after CLP and a variety of treatments using BMSCs from C57/BL6, FVB/NJ and BALB/c mice, as well as C57/BL6-derived fibroblasts. (b) BMSC treatment effects on kidney function, as reflected by serum concentration of creatinine (SCr). The number of mice in all measurements is as follows: sham, n = 5; CLP, n = 13; CLP + BMSC, n = 14. Tubular injury scores are shown at right. (c) Intense PAS staining of hepatocytes is shown after sham operation and BMSC treatment. No staining can be seen in CLP. After treatment (CLP + BMSC), the red staining by PAS in hepatocytes indicates partial glycogen storage capacity. Scale bar, 20 μm. (d) Alanine aminotransferase (ALT) and aspartate aminotransferase (AST) concentrations in the liver after sham and BMSC, CLP or CLP and BMSC treatment. (e) Serum amylase concentrations after sham and BMSC, CLP or CLP and BMSC treatment. (f) DAB staining of caspase-3 cells in untreated spleen sections and BMSC-treated spleen sections. A quantitative comparison between the numbers of apoptotic splenic cells in treated versus untreated mice (right) shows a significant decrease with BMSC treatment. Scale bar, 100 μm. (g) Serum TNF-α and IL-6 concentrations after sham and BMSC, CLP or CLP and BMSC treatment. (h) Serum IL-10 concentrations at 3, 6 and 12 h after CLP. n = 8−11 at each time point. Error bars represent means ± s.e.m.; *P < 0.05; **P < 0.01.

Figure 2

Fate of injected BMSCs and effect of BMSC treatment on survival of normal and immune cell–depleted mice. (a–c) Immunohistochemical staining showing that BMSCs prelabeled with Q-dot (red punctate staining; a) travel to the lung (b) and take up residence in close proximity to macrophages (c). The latter cells were immunostained with an antibody to Iba1 (ionized calcium-binding adaptor molecule-1, a specific marker of the macrophage lineage47) and visualized with Alexa-Fluor-488 conjugated to a secondary antibody (green). Scale bar, 10 μm. (d–f) Summary of the effectiveness of BMSC treatment of mice genetically lacking or depleted of certain subsets of immune cells or soluble mediators. Survival curves show survival percentage of macrophage-depleted mice with or without BMSC treatment (d), survival percentage of BMSC-treated CLP mice and untreated mice after neutralizing IL-10 or blocking the IL-10 receptor (e) and survival percentage of after treatment with BMSCs derived from Il10−/− septic mice (f). *P < 0.05.

Figure 3

Effect of BMSC treatment on leukocyte trafficking. (a,b) The average number of circulating monocytes (a) and the number of circulating neutrophils (b) after BMSC treatment. The cell counts are the average of data from five mice per group, 24 h after the induction of CLP. (c) The average number of macrophages isolated from lungs of CLP mice with no treatment, with BMSC treatment and BMSC treatment after depletion of circulating monocytes using clodronate. The numbers of macrophages are per random microscope visual field. Five mice were studied in each group. The numbers in parentheses are the total number of fields in which cells were counted. Error bars represent means ± s.e.m. **P < 0.01; ***P < 0.001.

Figure 4

Characterization of control and stimulated mononuclear cells in CLP mice in vivo and in vitro. (a) The distribution of monocytes (Mon), macrophages (Mac) and polymorphonuclear cells (PMN) among cells isolated from septic lungs. Error bars represent means ± s.e.m. (b) FACS plots of lung mononuclear cells. The cells were first gated on forward scatter (FSC) and CD11b. The R1 group (high FSC, dim CD11b) represents resident macrophages. R2 (high FSC and bright CD11b) represents a group of myeloid cells. R3 (medium FSC and bright CD11b) is a mixed population of monocytes and PMN cells when it is further gated on a monocyte-lineage marker (F4/80) and a PMN marker (GR1). The FACS is a representative example of the four mice shown on the bar graph in a. Blue color indicates monocytes and red indicates polymorphonuclear cells in both a and b expressed as percentage of total number of cells analyzed. (c) Quantification of IL-10 secretion after ex vivo LPS treatment of lung macrophages. Six hours after the induction of CLP, lung macrophages were isolated from mice with or without BMSC treatment (four mice per group), cultured and treated ex vivo with LPS. IL-10 production of these isolated macrophages is shown 3 and 5 h after LPS stimulation. (d) To test whether the increased IL-10 production could be due to a BMSC-macrophage interaction, macrophages from bone marrow were cocultured in vitro with BMSCs and stimulated with LPS. IL-10 production of macrophages cocultured with BMSCs is shown compared to macrophages that had no contact with BMSCs at 1, 3, 5 and 7 h after LPS stimulation. (e) Lungs of four BMSC-treated and four untreated mice were used for a FACS experiment with CD11b and an intracellular marker for IL-10. The number of IL-10 producing monocytes and macrophages after the treatment is shown. (f) Bacterial counts in the peritoneal space and in the circulation in 12 untreated (red) and 10 BMSC treated (blue) mice. (g) Myeloperoxidase (MPO) abundance in the kidney and the liver after BMSC treatment. Error bars represent means ± s.e.m. *P < 0.05, **P < 0.01 and ***P < 0.001.

Figure 5

Studies of molecular alterations underlying the effect of BMSCs on macrophages. (a) NF-κB abundance in nuclei isolated from triplicate samples of BMSCs 30, 60 and 90 min after addition of LPS to the culture medium. (b) Western blot (WB) analysis of COX2 abundance in LPS-stimulated cocultures. Density measurements of three western blots are quantified in the bar graph. (c) The COX2 activity after LPS treatment of cocultures (triplicate samples). (d) Western blot analysis of COX2 abundance in Tlr4−/− (KO) BMSCs or in cultures treated with antibody to TNF-α (anti–TNFα). (e) Prostaglandin E2 (PGE2) abundance in macrophage cultures or coculture supernatants in a variety of conditions 3 and 5 h after LPS stimulation. Four samples were run for each condition. (f) COX2 enzyme activity after iNOS inhibition and 1, 3 and 5 h after LPS stimulation in triplicate samples. Error bars represent means ± s.e.m. *P < 0.05, ***P < .001.

Figure 6

Summary of studies of the molecular pathways involved in the interaction between BMSC and macrophages. (a) IL-10 concentration changes in supernatants of cocultures in a variety of treatment conditions after LPS stimulation. Colored graphs (except for the blue color that labels the bone marrow macrophages as the source of all consecutive experiments) show treatments that eliminate the effect of BMSCs on macrophages. Black graphs show IL-10 levels after LPS stimulation, whereas open graphs show the control (nonstimulated) values. The experiments where the conditions eliminated the effect are colored. Purple color shows the effect of septic environment, green color shows agents and cellular compartments related to the PGE2 pathway and pink color shows agents related to the nitric oxide pathway. Three separate kinds of macrophages (bone marrow macrophages; peritoneal macrophages and the RAW264.7 cell line) were examined initially. Because they behaved identically in the assay, we used bone marrow derived macrophages (BM MF) for the rest of the experiments. In the box labeled 1, the effect of septic environment on the BMSCs is studied in BMSCs from TLR4-, MyD88-, TNFR1- and TNFR2- deficient mice, or antibody to TNF-α was used to neutralize the effect of TNF. The box labeled with 2 shows the cytokines and agents that have been implicated in the literature in immunomodulation of T cells by BMSCs, including COX1/2 and iNOS inhibitors. The box labeled 3 shows studies of the COX2 pathway, including the prostaglandin receptors EP1–EP4. Finally, in the box labeled with 4, we show studies related to nitric oxide. (b) A summary of our current hypothesis about the mechanisms that underlie the interactions between BMSCs and macrophages in the CLP sepsis model. Bacterial toxins (for example, LPS) and circulating TNF-α act on the TLR4 and TNFR-1 of the BMSCs, respectively. This results in the translocation of NF-κB into the nucleus. This activation process seems to be nitric oxide dependent. Activated NF-κB induces the production of COX2, resulting in increased production and release of PGE2. PGE2 binds to EP2 and EP4 receptors on the macrophage, increasing its IL-10 secretion and reducing inflammation.