Interleukin-10 Gene Transfer: Prevention of Multiple Organ Injury in a Murine Cecal Ligation and Puncture Model of Sepsis
- First Online:
- Cite this article as:
- Kabay, B., Kocaefe, C., Baykal, A. et al. World J. Surg. (2007) 31: 105. doi:10.1007/s00268-006-0066-9
The aim of this study was to determine the effect of immunoregulatory cytokine interleukin-10 (IL-10) gene therapy on multiple organ injury (MOI) induced by a cecal ligation and puncture (CLP) model of sepsis in mice.
Male Balb/c mice subjected to CLP were treated with either an hIL-10-carrying vector or an empty control vector. We assessed the degree of lung, liver, and kidney tissue destruction biochemically by measuring myeloperoxidase (MPO) and malondialdehyde (MDA) activity. Histologic assessments were based on neutrophil infiltration in lung and liver tissue. IL-10 protein expression was examined immunohistochemically, and ultrastructural changes in the liver were studied by transmission electron microscopy. We analyzed the expression of tumor necrosis factor-α (TNFα) mRNA by reverse transcription polymerase chain reaction 3, 8, and 24 hours after CLP in all organs.
Organ damage was significantly reduced by hIL-10 gene transfer, which was associated at the tissue level with reduced MPO activity in the liver, lung, and kidney and decreased leukocyte sequestration and MDA formation in the lung. The liver MDA was not significantly higher in the hIL-10 gene therapy group than in the controls and seemed not to be affected by hIL-10 gene transfer. The reduced portal tract neutrophilic infiltration and preserved ultrastructure of the hepatocytes also showed that tissue function was not impaired. The lung and kidney TNFα mRNA expression was suppressed markedly in the hIL-10 gene therapy group, but liver TNFα mRNA expression varied over time.
These findings showed that IL-10 gene therapy significantly attenuated sepsis-induced MOI.
Despite remarkable advances in antibiotic therapy and other medical and surgical techniques, postsurgical and posttraumatic sepsis and sepsis syndromes remain a major cause of death in intensive care units. All the clinical research in this field shows that sepsis is a time-dependent and complex problem involving several host factors, basically consisting of immunologic mediators acting for and against the complex mechanism leading toward death.1 The most frequent cause of death in septic patients is acute lung injury, seen in its most severe form as acute respiratory distress syndrome (ARDS). Injury of other organs, such as the liver and kidney, usually follow, resulting in multiple organ dysfunction syndrome (MODS).2
Cytokines are considered important mediators in the development of this lethal multiorgan damage, and the proinflammatory cytokine tumor necrosis factor-α (TNFα) plays a pivotal role in this process.3 Interleukin-10 (IL-10) modulates the expression of cytokines and strongly inhibits the production of IL-1, IL-6, IL-8, IL-10 itself, IL-12, TNFα, and various chemokines by activated monocytes/macrophages.4–7 IL-10 also enhances production of the IL-1 receptor antagonist and the soluble p55 and p75 TNF receptors.8 Thus, IL-10 induces a shift from the production of proinflammatory to antiinflammatory mediators.
Many functional activities of polymorphonuclear neutrophils (PMNs) are also regulated by IL-10.9 A major component of sepsis-induced organ injury is the recruitment and accumulation of activated PMNs. This multistep process involves adhesion of PMNs to the vascular endothelium. This adhesion relies on the expression of adhesion molecules induced by proinflammatory cytokines.10 Thus, we postulated that the modulation of inflammatory responses by the antiinflammatory cytokine IL-10 may inhibit sepsis-induced organ injury.
In the present study, we used gene therapy tools to study the role of IL-10 in a murine model of sepsis-induced organ injury. We hypothesized that IL-10 gene therapy would attenuate tissue oxidative injury: first by reducing neutrophil infiltration and second by modulating sepsis-induced changes in TNFα activity.
MATERIALS AND METHODS
Cecal Ligation and Puncture
Male BALB/c mice weighing 18 to 24 g were maintained under standard laboratory conditions. Mice were not fed the night before surgery. The animals were anesthetized with ketamine (80 mg/kg) and xylazine (16 mg/kg) by subcutaneous (SC) injection. We made a small abdominal midline incision and ligated the cecum just below the ilocecal junction with 3/0 silk, and punctured it with a 22-gauge needle, through and through, producing two holes. Gentle pressure was applied to the ligated cecum to extrude a small amount of stool, and the cecum was placed back in the abdomen. The abdominal incision was then closed with continuous nylon sutures, and 1 ml of saline was administered subcutaneously for fluid resuscitation. All experiments were carried out with approval of the Animal Use Committee of the Hacettepe University.
Groups of Mice
We divided the mice into four groups. Mice that were IL-10-negative (n = 12) or IL-10-positive (n = 12) were subjected to CLP. The IL-10-positive mice received liposomal 100 μg IL-10 carrying plasmid DNA (pcDNAhIL103.1/GS); and the IL-10-negative mice received liposomal 100 μg empty plasmid DNA (pcDNA3.1/GS) intraperitoneally 24 hours before the surgical procedure. Sham-operated mice (n = 9) underwent the previously described procedure without ligation and puncture of the cecum. For reverse transcription polymerase chain reaction (RT-PCR) analysis, control mice (n = 2) were taken from among those animals not subjected to any surgical procedure or plasmid DNA injection.
Cationic Liposome/DNA Preparation
Plasmid DNA was transferred into GeneHogs Escherichia coli encoding for IL-10 (pcDNA3.1hIL10/GS) and empty control vector (pcDNA3.1/GS; Research Genetics, Huntsville, AL, USA). pcDNA3.1hIL10/GS vector contains an expression cassette consisting of the cytomegalovirus (CMV) immediate early promoter and enhancer followed by bovine growth hormone (BGH) and SV40 polyadenylation signal. The IL-10 gene is fused to a C-terminal peptide encoding the V5 epitope for detection with the anti-V5-HRP antibody, which is supplied by the manufacturer. The vector also encodes the bleomycin resistance gene for selection in E. coli. Plasmid DNA was isolated, and endotoxin was removed using the Gene Elute endotoxin-free Plasmid Midiprep Kit (Sigma-Aldrich, St. Louis, MO, US). We used pcDNA3.1/GS as the control vector, which contains an expression cassette consisting of the CMV immediate early promoter and SV40 polyadenylation signal followed by bleomycin resistance gene for selection in E. coli.
We used in vivo Lipozome Transfection Reagent (20 mM DOTAP:cholesterol (1:1 molar ratio) extruded liposomes in 5% glucose; Sigma-Aldrich) to transfer plasmid DNA, as described by the manufacturer. Preliminary calibration mixing tests were done to determine the appropriate amounts of DNA and liposomes that produce complexes suitable for injection. The quality of liposome–DNA complexes was determined by the A400 values of the mixture. A400 of 0.5 to 0.9 for a 1:20 dilution in water indicates that the mixture is suitable for in vivo studies.
IL-10 and TNFα mRNA detection by RT-PCR
Primer pairs utilized to amplify the TNFα (NM_013693) and the GAPDH (M_32599) genes
5′-ACC ACA GTC CAT GCC ATC AC-3′
5′-TCC ACC ACC CTG TTG CTG TA-3′
5′-CCT GTA GCC CAC GTC GTA GC-3′
5′-TTG ACC TCA GCG CTG AGT TG-3′
Histopathology and Detection of IL-10 Protein
For the detection of IL-10 protein on tissue sections, 4-μm cryostat sections of 10% buffered formalin fixed tissues were deparaffinized and rehydrated and then labeled with the primary monoclonal antibody anti-V5-HRP. Enzyme conjugated secondary antibodies, biotinylated anti-mouse immunoglobulin G (IgG) and streptavidin-horseradish peroxidase (HRP) were applied (LSAB®2 System; Dako, Glostrup, Denmark). Following the secondary antibody incubation, sections were quenched for endogenous peroxidase activity and blocked for nonspecific binding interactions for 1 hour using an albumin-based blocking solution. Chromogenic staining was done with 3,3-diaminobenzidine to produce a brown color. Sections were counterstained with Harris’ hematoxylin.
Immediately after the animals were sacrificed, slivers of hepatic parenchyma were immersed in fixative and cut in 1 mm3 cubes. Fixation was carried out for 1 hour in phosphate buffer (pH 7.4, 265 mOsm). We processed specimens using the fast method through water, graded alcohols, and propylene oxide to Araltide CY212. They were preoriented and polymerized in Araltide CY212. Each cube was sectioned and ultrathin silver-gray sections were cut with a diamond knife, and picked up on bare 300 mesh copper grids; they were then stained in a saturated solution of aqueous uranyl acetate for half an hour and counterstained for 1 minute in alkaline lead citrate solution. Electron micrographs were made with the use of a Jeol 1220 Electron Microscope operated at 80 kV.
Assessment of Pulmonary and Liver Neutrophil Sequestration
Liver and lung tissue neutrophil sequestration was determined according to a method described elsewhere.11,12 Tissues were fixed in 10% formalin solution and paraffin-embedded sections (4 μm, H&E stain) were examined by light microscopy. A single pathologist, blinded to the groups, examined at least two sections of each specimen. Neutrophils were identified by their size (10–12 μm) and their segmented nuclei. Alveolar septal wall neutrophil sequestration in the peripheral lung parenchyma and only neutrophils that were present in the hepatic sinusoids or had extravasated into the tissue were counted. Neutrophils in necrotic areas were not counted in the liver. The number was expressed as the mean number of neutrophils per histologic power field (×400).
Tissues were homogenized in 20 mM potassium phosphate buffer (pH 7.4) and the homogenate was centrifuged for 5 minutes at 10,000×g at 4°C. The supernatant was discarded, and the pellet was resuspended in 50 mM potassium phosphate buffer (pH 6.0) containing 0.5% hexadecyltrimethylammonium bromide (HETAB). The suspension was frozen and sonicated for 10 seconds. Following incubation for 2 hours in a waterbath (60°C), the suspension was centrifuged at 10,000×g for 5 minutes. The supernatants were subjected to myeloperoxidase (MPO) assay, and MPO activity was assessed by measuring H2O2-dependent oxidation of o-dianisidin. One unit of enzyme activity was defined as the amount of MPO present that caused a change in the absorbance of 1.0/min at 410 nm and 37°C.
The tissue malondialdehyde (MDA) content, as an index of lipid peroxidation, was determined by thiobarbituric acid (TBA) assay according to Mihara and Uchiyama.13 Briefly, 3 ml 1% phosphoric acid and 1 ml 0.67% TBA solution were added to a 0.5 ml tissue sample. The mixture was heated in boiling water for 1 hour; after cooling, 4 ml of 1-butanol was added and mixed vigorously. After centrifugation, the butanol phase was removed, and absorbance at 532 nm was used to calculate the MDA concentration. The MDA standard was prepared from 1,1,3,3-tetramethoxypropane. The MDA concentration in tissue homogenate was expressed in relation to the protein concentration, which was measured by the method of Lowry et al.14 The results are expressed as nanomoles per milligram of protein.
Quantification of Organ Function and Injury
Blood samples were taken after 18 hours in all animals and centrifuged to separate the plasma. All plasma samples were analyzed within 24 hours using standard laboratory techniques. The following marker enzymes were measured in the plasma as biochemical indicators of MOI or dysfunction: Liver injury was assessed by measuring the increase in plasma concentrations of alanine aminotransferase (ALT) and aspartate aminotransferase (AST); renal dysfunction was assessed by measuring the increase in plasma concentrations of creatinine, and the acid-base balance and blood gas analyses were evaluated by measuring the arterial blood concentrations of pH, PaO2, and bicarbonate anion (HCO3−) using a pH blood gas analyzer as described elsewhere.15
Values are expressed as the median (range). Data were analyzed by statistical software (SPSS for Windows 10.0; SPSS, Chicago, IL, USA). Biochemical data were compared using parametric one-way analysis of variance (ANOVA) with Tukey’s HSD post hoc tests and nonparametric histopathologic scoring data analyzed by the Kruskal-Wallis test. The mean TNFα mRNA expression values are compared to the mean of the sham group by two-way nonparametric ANOVA (Friedman’s test). The time points that satisfy the significance cutoff (P < 0.05) are indicated by the asterisk in Figure 7 (see later).
Expression of IL-10 Protein
Pulmonary and Liver Neutrophil Sequestration; MPO and MDA Analysis
Organ dysfunction measurements in all the treatment groups of mice
CLP + empty vector
CLP + hIL-10 gene
19.58 ± 3.90
49.53 ± 9.30†
29.66 ± 5.02‡
39.53 ± 13.00
99.68 ± 12.91†
45.00 ± 13.47‡
0.23 ± 0.04
1.15 ± 0.17†
0.65 ± 0.13‡
7.38 ± 0.02
7.26 ± 0.01†
7.42 ± 0.02‡
34.18 ± 1.10
30.38 ± 1.34†
33.03 ± 1.70‡
98.33 ± 0.81
87.66 ± 5.08†
95.66 ± 1.63‡
25.33 ± 1.92
17.45 ± 1.19†
23.20 ± 1.16‡
Empty vector transfer + CLP resulted in a significant decrease in the arterial levels of PaO2, PaCO2, pH, and HCO3−, when compared with the sham-operated animals. The hIL-10 gene therapy significantly reduced lung dysfunction caused by CLP.
Histologic Evaluation of Lung and Liver Injury
TNFα RT-PCR mRNA Expression
The CLP-operated animals demonstrated an early increase of TNFα transcription in blood (1.5-fold change) and lung (1.4-fold change) at hour 3, which reached significance (P < 0.05)—changes of 1.8-fold and 1.7-fold, respectively—over 24 hours. TNFα expression in liver increased significantly (P < 0.05)—1.3-fold to 2.3-fold changes—within 8 to 24 hours. Also a significant 25% increase in TNFα expression was observed in the kidney of the CLP-operated group from hour 3 to hour 24 (Fig. 7).
The IL-10 gene transfer group demonstrated a trend of TNFα expression similar to that of the CLP group in blood and lung (expression patterns parallel to those in Fig. 7B,C) but this up-regulation met the significance criteria (P < 0.05) only at the 24-hour time point in lung and at the 8 and 24 hour points in blood. This significant induction reached 1.5-fold and 1.4-fold changes at 24 hours in blood and lung, respectively. The increase in TNFα expression was 1.9-fold change in liver at 8 hours and remained stable up to the 24-hour time point. The expression pattern in kidney did not display significant change (Fig. 7D).
Figure 7 shows the expression pattern of TNFα in all animals. IL-10 gene therapy suppressed TNFα mRNA expression in lung (Fig. 7B) and kidney (Fig. 7C) tissue, but the suppression pattern reversed in the liver (Fig. 7A) tissue over time.
The findings of this study showed that systemic hIL-10 production, induced by a single administration of liposomal IL-10-carrying plasmid DNA before CLP, inhibited multiple organ dysfunction. Previous studies using exogenous recombinant IL-10 (rIL-10) or blocking endogeneous IL-10 have investigated the role of IL-10 in sepsis-related conditions and reported that the site and timing of administration, as well as the animal model, are all factors contributing to outcome.16–18 Thus, we used the technique of in vivo gene transfer described in this study for systemic delivery of hIL-10 as an immunomodulator to inhibit a systemic inflammatory response before CLP.
One important discussion on the use of immunomodulator approaches for sepsis is the activity of the agent. Continuous endogenous synthesis would overwhelm the limitations of recombinant IL-10, which has a half-life less than 4 hours.19–21 In our study, intraperitoneal liposomal gene transfer is a systemic method of gene delivery,20 and de novo IL-10 synthesis is observed primarily in the peritoneum but is detectable throughout the entire body, principally in all endothelial structures. In our study, systemic delivery of IL-10 and endogenous synthesis in the peritoneum ameliorated MOI in the animals. Intraperitoneal delivery of IL-10 is a novel approach that gave us the chance to elucidate the controlled impact of the cytokine in a preconditioned environment. One additional impact of local expression is to control the inflammatory response at its foci. Chun et al.22 reported favoring the liposomal gene transfer method because this type of gene expression was far more persistent than that induced by the viral vector system. Liposomal complexes readily extravasate across vascular endothelial barriers in lung tissue and offer the advantage of transgene expression throughout the tissue.23 Moreover, nonviral gene delivery using complexes of cationic liposomes and plasmid DNA is nontoxic to the animals and can be used as a safe vehicle for gene transfer in a variety of cancer gene therapy clinical trials.24–28 Adenoviral vectors may not be the best vector system to carry an immune modulator gene, as they are long known to provoke a strong immune response and there are limitations with in vivo use.29
We used the more clinically relevant model of CLP and administered plasmid DNA 24 hours beforehand.30,31 Systemic gene delivery before CLP may not mimic real-life sepsis treatment, but it enabled us to observe the effects of IL-10 on remote organ injury before the release of proinflammatory mediators. The reason for administering liposomal complexes 24 hours before CLP is that intraperitoneal injections of plasmid/cationic liposome complexes produce gene expression in tissues 24 hours after injection, and this transgene expression can be detected for 10 days after injection.22 However, the slower onset of transgene expression effectiveness limited the application of this technique in inflammatory conditions, and a practical way of achieving this has yet to be perfected. The use of expression-controlled plasmid DNA vectors, which would activate the transgene after the septic insult, and the regional administration of these vectors into selected tissue compartments or cell type-specific liposomal delivery methods may help create clinically applicable models and define the role of hIL-10 in combating intraabdominal sepsis.
Several experimental reports describe successful results of using gene therapy for sepsis and tissue injury.32–34 Oberholzer et al.33 used a CLP model and observed that the intrathymic administration of adenoviral vectors—but not systemic administration—was protective. This effect was attributed to the prevention of thymocyte apoptosis. We used the same model and observed that systemic administration of liposomal IL-10 gene transfer attenuated oxidative MOI. This was possibly achieved by the reduced infiltration of neutrophils into the tissues, leading to decreased amounts of TNFα mRNA, and by blocking release of oxygen free radicals from activated macrophages, which stopped the progression to multiple organ failure.
Systemic neutrophil activation and infiltration into the tissues plays a central role in the pathogenesis of organ injury in sepsis-related conditions.2,10,35 The lung is one of the most severely affected organs, and the effects are seen early. Several proinflammatory and antiinflammatory mediators are released secondary to neutrophil infiltration. Interplay between proinflammatory and antiinflammatory mediators serves to mediate and regulate these immune and inflammatory responses. IL-10, which plays a complex role in these processes, has multiple properties. It has been shown to inhibit alveolar macrophage production of the proinflammatory mediators involved in ARDS. According to the study of Armstrong and Milla,36 ARDS developed in patients with low circulating and bronchoalveolar lavage levels of IL-10 and increased lavage levels of TNFα, highlighting the potential importance of interplay between proinflammatory and antiinflammatory mediators. Furthermore, the highest concentrations of TNFα are found in bronchoalveolar lavage fluid from patients with sustained ARDS, suggesting that locally elevated cytokine levels are responsible for lung tissue injury. In our study, the highly suppressed neutrophilic infiltration and TNFα mRNA levels in the mice that underwent hIL-10 gene transfer possibly protected their lungs from cytokine-mediated injury. These findings are in accordance with those of Takakuwa et al.,32 who observed markedly reduced TNFα mRNA expression in lung tissue before the induction of sepsis. Takakuwa et al.32 also studied plasma TNFα concentrations and TNFα mRNA expression in the lung and liver tissue after induction of septic insult. They found that plasma TNFα was significantly suppressed by IL-10 gene therapy after induction of sepsis but that it was suppressed further in animals pretreated with IL-10. They observed a similar pattern of TNFα mRNA expression in lung and liver tissue.
Although liver failure is generally thought to be a late complication after pulmonary and renal failure during sepsis, the results of recent studies indicate that hepatocellular function is depressed early after the onset of sepsis and persists during the late, hypodynamic stage.35,37 Hepatocellular dysfunction during sepsis is associated with early increased gene expression as well as circulating levels of TNFα, which remain elevated as sepsis progresses38 Despite the widespread expression of hIL-10 in the liver in our study, TNFα mRNA expression tended to increase beginning at 24 hours after CLP. Similar levels of elevated TNFα mRNA expression were observed in all of the groups other than the controls. The TNFα expression patterns we observed are in contrast with those of other studies, but our experimental design was insufficient to clarify this effect and requires more work. However, increased MPO activity and neutrophil infiltration were well correlated with the morphologic alterations observed histologically in the liver; and hIL-10 gene transfer exerted a protective effect biochemically and structurally. We also observed that there is a significant decrease in neutrophil infiltration and MPO content in the liver, yet there was an increase in MDA levels. Taken together with the elevated TNFα mRNA expression during the late phase of septic insult, a possible explanation might be the late hepatic dysfunction that results from microcirculatory perturbations, causing a mismatch of blood supply to metabolic demand.39 Under physiologic conditions, liver microcirculation is maintained by a balance between vasoconstrictors (endothelin-1) and vasodilators (nitric oxide and carbon monoxide). Under inflammatory conditions, this regulated balance becomes disrupted, favoring vasoconstriction; and IL-10 might have different regulatory functions under septic conditions.40 Hepatocellular damage may be a combination of early toxic and late microperfusion-related hepatocyte injury; and the role of IL-10 needs to be clarified in this context.
Several investigators have studied exogenous administration of IL-10 protein in experimental models of sepsis. IL-10 is completely protective in the lipopolysaccharide model of sepsis, but in the CLP model it can be either protective or harmful depending on the time of intervention.4,17 Remick et al.18 reported that increasing amounts of IL-10 administered to mice before CLP resulted in no survival benefit. Similarly, IL-10 pretreatment had no effect on pulmonary neutrophil sequestration in either the lung or the peritoneum. This may be due to the short biologic half-life of recombinant IL-10 and the limited access of rIL-10 to the tissue interstitium. Giving IL-10 protein exogenously did not seem to maintain high tissue levels of protein.19,36 Thus, continuous administration of IL-10 protein is needed to maintain biologically active concentrations.
We found that overexpression of hIL-10 in septic animals, evoked by systemic liposomal gene transfer, ameliorates MOI, resulting in reduced neutrophil sequestration and down-regulation of local TNFα mRNA expression. These results indicate the therapeutic potential of IL-10 gene transfer as a strategy against multiple organ dysfunction and warrant further studies to elucidate the potential role of IL-10 in septic conditions.