Severely burned mice are susceptible to sepsis stemming from Enterococcus faecalis translocation due to the impaired generation of M1 macrophages (M1Mϕs) in local translocation sites. In our previous studies, CCL2 has been characterized as a major effector molecule on the burn-associated generation of M2Mϕs, an inhibitor cell type for resident Mϕ conversion into M1Mϕs. In this study, we tried to protect burned mice orally infected with E. faecalis utilizing CCL2 antisense oligodeoxynucleotides (ODNs). We show that M2Mϕs in mesenteric lymph nodes (MLNs) were not demonstrated in burned mice treated with CCL2 antisense ODNs. M1Mϕs were not induced by heat-killed E. faecalis from resident Mϕs transwell-cultured with mesenteric lymph node macrophages (MLN-Mϕs) from burned mice, while M1Mϕs were induced by the same antigen from resident Mϕs transwell-cultured with Mϕs which were isolated from burned mice treated with CCL2 antisense ODNs. Bacterial growth in MLNs was shown in burned mice orally infected with a lethal dose of E. faecalis. However, after the same infection, sepsis did not develop in burned mice treated with CCL2 antisense ODNs. These results indicate that bacterial translocation and subsequent sepsis are controlled in burned mice orally infected with a lethal dose of E. faecalis by gene therapy utilizing CCL2 antisense ODNs.
Infectious complications are responsible for a high mortality rate of thermally injured patients. Although the concept of bacterial translocation in these patients is not yet completely standardized, bacterial translocation beyond the intestinal lumen is one of the mechanisms that promotes infectious complications 1, 2. The evidence of bacterial translocation are: (i) nosocomial infections have been correlated with indigenous gut bacteria (e.g. Escherichia coli) isolated in blood cultures and (ii) enteric microorganisms have been identified in the blood of cirrhotic patients with spontaneous bacterial peritonitis 3. Antibiotics are effective in diminishing the colonization and multiplication of bacteria which are translocated from the intestine. However, due to defects of the host's antibacterial innate immunities, the very small amounts of bacteria that escape from these treatments are sufficient to spread systemically in thermally injured patients. Excessive antibiotic usage (amounts and duration) leads to the generation of untreatable strains of bacteria. A new paradigm is needed to treat burn patients with bacterial translocation-related infectious complications. Therefore, we attempted to immunologically control infectious complications caused by bacterial translocation through the recovery of damaged host antibacterial defenses in thermally injured patients.
The important roles of macrophages (Mϕs) in antibacterial innate immunity have been described in many papers 4–10. M1Mϕs (IL-12+ IL-23+ IL-10− Mϕs) generated from resident Mϕs by the stimulation with a microbial antigen or cytokines are potent effector cells that kill invaded microorganisms 11–13. In contrast, M2Mϕs (IL-12− IL-23− IL-10+ Mϕs) 14, 15 are shown to be inhibitory on Mϕ conversion from resident Mϕs to M1Mϕs 16. CCL17 and IL-10 released from M2Mϕs are characterized as effector molecules for inhibiting Mϕ conversion from resident Mϕs to M1Mϕs 16. Therefore, M1Mϕs are not generated in hosts where M2Mϕs predominate 7, 17.
CCL2 is a chemokine that attracts and activates mononuclear cells. The necessity of this chemokine for Th2-cell generation has been well demonstrated 18. Thus, CCL2-knockout mice resisted Leishmania major infection 18, while CCL2-overexpressing transgenic mice were susceptible to infections with Listeria monocytogenes or Mycobacterium tuberculosis19. We previously demonstrated that herpes encephalomyelitis 20 and cryptococcal encephalitis 21 are not severely developed in mice depleted of CCL2. Recently, the increased level of CCL2 has been demonstrated in sera of thermally injured patients 22 as well as severely burned mice 23. These mice have already been characterized as mice susceptible to sepsis stemming from Enterococcus faecalis translocation 24. In the subsequent study 25, utilizing CCL2 knockout mice, a role of CCL2 on resident Mϕ conversion into M1Mϕs or M2Mϕs was explored. In contrast to severely burned wild-type mice, M1Mϕs were induced and M2Mϕs were not induced in burned CCL2-knockout mice stimulated with the E. faecalis antigen. SCIDbgMN mice (SCID-beige mice depleted of functional Mϕs and PMNs) inoculated with Mϕs that were isolated from burned CCL2-knockout mice were shown to be resistant against lethal doses of E. faecalis infection, whereas all of the SCIDbgMN mice inoculated with Mϕs from burned WT mice died after the same infection. Also, burned CCL2-knockout mice treated with rCCL2 were shown to be susceptible to E. faecalis infection, and M2Mϕs were isolated from these mice 25.
In the present study, we tried to protect thermally injured mice orally infected with a lethal dose of E. faecalis by gene therapy utilizing phosphorothioate-CCL2 antisense oligodeoxynucleotides (ODNs). Antisense ODNs, ribozymes and small-interfering RNA have been used for cytokine knockdown therapy 26, 27. As compared with alternative technologies for blockage of CCL2, antisense ODNs have a higher specificity and probability of success 28. The advantage of antisense ODNs designed as phosphorothioates specifically to heterogeneous nuclear RNA or mature mRNA sequences is resistance to degradation by RNases 26. Therefore, for the blockage of CCL2 in severely burned mice orally infected with E. faecalis, phosphorothioate-CCL2 antisense ODNs were utilized in this study.
Reduced CCL2 levels in the sera of severely burned mice treated with CCL2 antisense ODNs
In our previous studies 23, 24, CCL2 produced in response to burn injury was shown to play a major role on the M2Mϕ predominance in severely burned mice. Therefore, for the elimination of M2Mϕs; we tried to reduce serum CCL2 levels of severely burned mice by treatment with CCL2 antisense ODNs. Various concentrations of CCL2 antisense ODNs were administered to mice 2 and 12 h after burn injury. Sera, obtained from these mice 24 h after burn injury, were assayed for CCL2 by ELISA. Serum specimens obtained from normal mice treated with saline and severely burned mice treated with scrambled ODNs were utilized as controls. CCL2 was not detected in the sera of normal mice, whereas the sera of severely burned mice treated with scrambled ODNs contained 1.3 ng/mL of CCL2. However, 77–100% of CCL2 was eliminated from the sera of severely burned mice after treatment with 1 μg/mouse (Fig. 1A) or more (10 and 100 μg/mouse, Fig. 1B) of CCL2 antisense ODNs. These results indicate that the gene therapy utilizing CCL2 antisense ODNs is feasible to decrease CCL2 levels in severely burned mice.
Disappearance of M2Mϕs in severely burned mice treated with CCL2 antisense ODNs
The disappearance of MLN-M2Mϕs in severely burned mice treated with CCL2 antisense ODNs was examined. Severely burned mice were treated twice with 10 μg/mouse of CCL2 antisense ODNs 2 and 12 h after burn injury. Mϕs isolated from mesenteric lymph nodes (MLN-Mϕs) of these mice 1–8 days after burn injury were cultured for 24 h without any stimulation. Culture fluids harvested were assayed for CCL17 as a biomarker of M2Mϕs. The amounts of CCL17 detected in the culture fluids were compared with those of CCL17 that were produced by the same MLN-Mϕs derived from controls (burned mice treated with scrambled ODNs). In the results, CCL17-producing Mϕs (M2Mϕs) appeared in burned mice 2–4 days after burn injury. However, such Mϕs were not demonstrated significantly in MLN-Mϕs of severely burned mice treated with CCL2 antisense ODNs (Fig. 2). These results indicate that gene therapy utilizing CCL2 antisense ODNs inhibits MLN-M2Mϕ-generation in severely burned mice.
Induction of M1Mϕs in severely burned mice treated with CCL2 antisense ODNs
We tried to induce M1Mϕs from resident Mϕs transwell-cultured with MLN-Mϕs from severely burned mice treated with CCL2 antisense ODNs. MLN-Mϕs (upper chamber), isolated from severely burned mice treated with 10 μg/mouse CCL2 antisense ODNs, were transwell-cultured with resident Mϕs (lower chamber). Before the cultivation, resident Mϕs were cultured with 105 heat-killed E. faecalis for 6 h and washed with media three times. Twenty-four hours after cultivation, cells in the lower chamber were tested for their abilities to produce CCL5 and IL-12, biomarkers for M1Mϕs. In the results, M1Mϕs were not generated from antigen-stimulated resident Mϕs in transwell cultures performed with MLN-Mϕs from severely burned mice. However, both IL-12 and CCL5 were produced by antigen-stimulated resident Mϕs transwell-cultured with MLN-Mϕs from severely burned mice that were previously treated with CCL2 antisense ODNs (Fig. 3A). These results indicate that M1Mϕs are inducible from resident Mϕs transwell-cultured with MLN-Mϕs that were derived from severely burned mice treated with CCL2 antisense ODNs. On the other hand, the abilities to produce IL-10 and CCL17 were examined for resident Mϕs after transwell cultured with Mϕs (lower chambers) isolated from MLNs of burn mice treated with or without CCL2 antisense ODNs. M2Mϕ properties were demonstrated in resident Mϕs transwell-cultured with Mϕs from MLNs of burn mice. However, resident Mϕs did not change to M2Mϕs after transwell-culture with Mϕs from MLNs of burn mice treated with CCL2 antisense ODNs (Fig. 3B).
Effect of CCL2 antisense ODNs on the susceptibility of burned mice to E. faecalis oral infection
Severely burned mice were treated with CCL2 antisense ODNs once daily for 5 days beginning 2 h after burn injury. At 24 h after burn injury, these mice were infected orally with 107 CFU/mouse of E. faecalis. Survival and bacterial growth in these mice were compared with those of severely burned mice treated with scrambled ODNs. In the results, 100% of normal mice orally infected with E. faecalis survived, while 100% of burned mice treated with scrambled ODNs died within 5 days of infection. At this time, 84% of severely burned mice treated with CCL2 antisense ODNs survived (Fig. 4A). In the next experiments, the growth of bacteria in MLNs of severely burned mice 2 days after E. faecalis oral infection was examined. E. faecalis was not detected in the MLNs of normal mice orally infected with E. faecalis, whereas 1.8×104 CFU/g organ of the pathogen was detected in the MLNs of severely burned mice treated with scrambled ODNs. When CCL2 antisense ODNs were administered to severely burned mice before and after E. faecalis infection (once daily for 5 days beginning 2 h after burn injury), the growth of the pathogen was not demonstrated in MLNs of these mice (Fig. 4B). These results indicate that the sepsis caused by E. faecalis translocation is effectively suppressed in severely burned mice treated with CCL2 antisense ODNs.
M1Mϕs appearing in MLN-M1Mϕs have been identified as a major host's antibacterial effector cell against E. faecalis translocation 24, 25. However, resident Mϕs transwell-cultured with MLN-M2Mϕs from burned mice did not convert into M1Mϕs although they were stimulated with a bacterial antigen. M2Mϕs are inhibitory of the Mϕs conversion from resident Mϕs to M1Mϕs. Recently, M2Mϕs have been classified into three subpopulations: M2aMϕs (IL-10+ CCL17+ FIZZ1+ Mϕs), M2bMϕs (IL-10+ CCL1+ LIGHT+ Mϕs) and M2cMϕs (IL-10+ CXCL13+ FIZZ1+ Mϕs) 9. Except for the chemokine-producing profile, the discrimination of M2aMϕs and M2cMϕs is impossible at this time 9, 29, 30. In our previous study 25, M2aMϕs and M2cMϕs were isolated from MLNs of mice 2–8 days postburn injury, and M2bMϕs were isolated from MLNs of mice 10–28 days postburn injury. In this study, Mϕs were isolated from MLNs of mice 1–8 days after burn injury, and these Mϕs produced CCL17, CXCL13 and IL-10 into their culture fluids (CCL1 was not produced by them). These results indicate that M2Mϕs utilized in this study were a mixture of M2aMϕs and M2cMϕs. Since the appearance of M2aMϕs or M2cMϕs was not demonstrated in CCL2-knockout mice exposed to severe burn injury 25, this indicates that CCL2 is required for the generation of M2aMϕs and M2cMϕs. M2bMϕs were induced in CCL2-knockout mice exposed to severe burn injury 25. Therefore, we hypothesized that MLN-M1Mϕs are inducible at translocation sites of severely burned mice orally infected with E. faecalis if the appearance of MLN-M2aMϕs and M2cMϕs is controlled in mice 1–8 days after severe burn injury.
In the results, normal mice and severely burned mice treated with CCL2 antisense ODNs did not carry M2Mϕs in their MLNs. When antigen-stimulated resident Mϕs were transwell cultured with MLN-Mϕs that were isolated from severely burned mice treated with CCL2 antisense ODNs, M1Mϕs were generated. Bacterial translocation and subsequent sepsis did not develop in normal mice orally infected with 108 CFU/mouse or more of E. faecalis, while all severely burned mice orally infected with 107 CFU/mouse of the pathogen died within 5 days of infection. However, bacterial growth in MLNs of severely burned mice treated with CCL2 antisense ODNs was not demonstrated significantly, and 84% of these mice survived. These results indicate that sepsis stemming from E. faecalis translocation in severely burned mice is controllable by the gene therapy utilizing CCL2 antisense ODNs, through the elimination of MLN-M2aMϕs and M2cMϕs (or induction of MLN-M1Mϕs) at the translocation site. Blockage of IL-10 may influence the functions of all phenotypes of M2Mϕs; however, this intervention may lead to the unregulated systemic inflammation through the inhibition of regulatory T-cell functions. In fact, disadvantages of IL-10 blockage in host antibacterial infections have been documented 31, 32. Therefore, we did not use IL-10 antisense ODNs in this study.
Using SCIDbg mice depleted of Mϕs and PMNs (SCIDbgMN mice), we have preliminarily examined whether orally infected pathogen causes infectious complications. After decontamination, these mice were infected orally with vancomycin-resistant Enterococcus faecium (VRE, ATCC 700221 strain), and the growth of VRE in the liver and MLNs was examined using EF agar containing vancomycin. In these experiments, we confirmed a source of the pathogen for sepsis developed in burn mice orally infected with E. faecium. That is to say, the vancomycin-resistant property of enterococci was used as a biomarker of the pathogen, which was translocated from intestine. When 105 CFU/mouse of VRE was given to SCIDbgMN mice, all of them died within 3–5 days of infection. VRE (105.7–106.2 CFU/g organ) was detected in tissue specimens taken from these mice 2 days after infection. No other bacteria were detected in these tissue samples. In addition, all SCIDbgMN mice exposed to the same dose of heat-killed VRE survived, and no bacteria were detected in tissue specimens from these mice. These results indicate that the development of infectious complications in these mice was caused by VRE given orally.
Various cells such as neutrophils, monocytes/Mϕs, dendritic cells, eosinophils and certain T-cell subpopulations are known to be producers of CCL2 33. So far, we do not know which cells are the major source of CCL2 in burned mice. Certain monocyte/Mϕ populations exposed to stress have been described as producer cells for CCL2 34. These monocytes/Mϕs may play a role on the CCL2 production in burned mice. In our previous studies utilizing severely burned mice 7, neutrophils with the functions to produce CCL2 and IL-10 have been demonstrated, and these neutrophils are designated as PMN-II. PMN-II may be the major cell to produce CCL2 in mice 1–3 days after burn injury. PMN-II were clearly distinguished from normal PMNs and immunopotentiating PMNs (PMN-I) by the ability to express CD11b and CD49d surface antigens and cytokine/chemokine-producing profile 7. Thus, PMN-II (CD11b+CD49d− PMNs) are CCL2 and IL-10-producing cells, whereas PMN-I (CD11bCD49d+ PMNs) are IL-12 and IFN-γ-producing cells. However, neither CCL2 nor IL-10 was produced by neutrophils isolated from burn mice that were previously treated with CCL2 antisense ODNs (Supporting Information Fig. 1). These results indicate that CCL2 production by PMN-II is controllable by CCL2 antisense ODNs gene therapy. Further studies are needed.
Materials and methods
Eight to ten weeks-old male BALB/c mice (The Jackson Laboratory, Bar Harbor, ME, USA) were used in these experiments. Experimental protocols for animal studies were approved by the Institutional Animal Care and Use Committee of the University of Texas Medical Branch at Galveston.
Bacteria, reagents and media
As previously described 24, 25, E. faecalis (49757 strain) grown in brain heart infusion broth for 18 h at 37°C was utilized in this study. Heat-killed E. faecalis (5×107 CFU/mL) were prepared by heating bacteria at 65°C for 30 min. No viability of the bacteria was confirmed by plating an aliquot of the heat-killed bacteria on TSB agar plates. Murine rCCL2 was purchased from BD Biosciences (San Jose, CA, USA), and mAb directed against CCL2 was obtained from BioLegend (San Diego, CA, USA). rCCL5, rCCL17 and mAbs directed against these chemokines were purchased from R&D Systems (Minneapolis, MN, USA). Biotin-conjugated anti-CD3, anti-F4/80 and anti-CD19 mAbs were obtained from eBioscience (San Jose, CA, USA). Phosphorothioated CCL2 antisense ODNs (5′-AAGCGTGACAGAGACCTGCATAGTGGTGG-3′) and scrambled ODNs (5′-CCACCACTATGCAGGTCTCTGTCACGCTT-3′) were purchased from Sigma-Genosys (The Woodlands, TX, USA). RPMI-1640 medium supplemented with 10% heat-inactivated FBS, 2 mM L-glutamine, 100 U/mL penicillin and 100 μg/mL streptomycin was utilized for the cultivation of various Mϕ preparations.
Thermally injured mice were created according to our previously reported protocol 23–25. This procedure consistently produced a third degree burn on approximately 25% of total body surface area (TBSA) for a 26-g mouse. Immediately after thermal injury, physiologic saline (1 mL per mouse, i.p.) was administered for fluid resuscitation. Deaths within 5 days of 25% TBSA flame burn were not demonstrated after our burn procedure. As controls, mice were anesthetized and shaved but were not exposed to the gas flame. They also received physiologic saline (1 mL per mouse, i.p.). Buprenorphine (2 mg/kg) was given s.c. every 12 h during the postburn period. Sham burn animals also received identical regimens of analgesics (buprenorphine) throughout the study period.
Preparation of mesenteric lymph node Mϕs (MLN-Mϕs)
Mϕs (F4/80+ cells) were prepared from MLNs of various groups of mice, as previously described 24, 25. F4/80+ cells with 94% or more purity were consistently obtained using this technique.
Severely burned mice were subjected to CCL2 antisense ODN gene therapy. Thus, burned mice were treated twice with CCL2 antisense ODNs at 2 and 12 h after burn injury. Based on our preliminary studies, CCL2 antisense ODNs were administered s.c. to burned mice at doses ranging from 0.01 to 100 μg/mouse. Weight loss, reduced appetite and abnormal body temperature were not demonstrated in normal mice treated with 100 μg/mouse of CCL2 antisense ODNs twice a day for 7 days. The effect of the gene therapy was confirmed by measuring CCL2 levels in the sera of these mice 24 h after burn injury, because the maximum level of CCL2 in sera of these mice was reached within 24 h of severe burn injury. CCL2 in serum specimens was assayed by ELISA. To determined the efficacy of CCL2 antisense ODNs on the generation of M2Mϕs, MLN-Mϕs were isolated from severely burned mice treated twice with 10 μg/mouse of CCL2 antisense ODNs (2 and 12 h after burn injury) 1–8 days after burn injury. Then, their abilities to produce CCL17 were compared with those of MLN-Mϕs isolated from burned mice treated with scrambled ODNs. Next, M1Mϕs were induced from antigen-stimulated resident Mϕs transwell cultured with MLN-Mϕs that were isolated from burned mice treated with CCL2 antisense ODNs. Transwell cultures were performed with MLN-Mϕs (5×105 cells/mL, upper chamber) and resident Mϕs (1×106 cells/mL, lower chamber) that were previously stimulated for 6 h with 105 heat-killed E. faecalis. Twenty-four hours after cultivation, the upper chamber was removed and Mϕs in the lower chamber were washed with media. Then, Mϕs in the chamber were cultured for an additional 24 h. Culture fluids harvested were assayed for CCL5 and IL-12 (p35/p40 heterodimer) using ELISA. When Mϕs with the abilities to produce CCL5 and IL-12 (but not CCL17) were detected in the lower chamber of transwell cultures, they were considered to be M1Mϕs. When Mϕs with the abilities to produce CCL17 (but not CCL5 and IL-12) were detected in the lower transwell chambers, they were considered to be M2Mϕs.
E. faecalis oral infection
As previously described 24, 25, mice were decontaminated by an antibiotic mixture before E. faecalis oral infection. Then, decontaminated mice were treated orally with lansoprazole (a proton-pump inhibitor, 0.5 mg/mL) to stabilize infection conditions. Four hours after treatment, these mice were exposed to burn injury. The mice were then treated with CCL2 antisense ODNs once daily for 5 days beginning 2 h after burn injury. One day after burn injury, the mice were infected orally with 107 CFU/mouse of E. faecalis. The severity of infectious complications induced by E. faecalis oral infection in these mice was evaluated by (i) the growth of the bacteria in MLNs and (ii) the mortality rates of the test groups in comparison with the controls, as previously described 24, 25.
The results obtained were analyzed statistically using ANOVA test. Survival curves were analyzed using the Kaplan–Meier test. All calculations were performed on a computer using the program Statview 4.5 from Brain Power. A value of p<0.05 was considered significant.
This work was supported by Shriners of North America grant #88400.
Conflict of interest: The authors declare no financial and commercial conflict of interest.