Endothelial dysfunction is a potential contributor to multiple organ failure and mortality in aged mice subjected to septic shock: preclinical studies in a murine model of cecal ligation and puncture

Male mice homozygous for the Nos3^tm1Unc targeted mutation (eNOS^-/-) on a C57BL6 background were obtained from The Jackson Laboratory, Bar Harbor, ME, USA (strain name: B6.129P2-Nos3^tm1Unc/J; stock number: 002684). Male C57BL/6 mice (six to eight weeks of age) were used as wild-type controls (Jackson stock number: 000664). Aged (24 months) C57BL/6 mice were obtained from colonies of the National Institute on Aging. Animals were kept in a 12 h/12 h light/dark cycle at 21 to 23°C with free access to a standard chow diet.

Acute sepsis was induced in mice by cecal ligation and puncture (CLP) as previously described [15] with modifications [16]. Briefly, mice were anesthetized by ketamine/xylazine cocktail (intraperitoneal (i.p.)), the abdomen was shaved, wiped with 70% isopropanol and a midline abdominal incision (1 to 2 cm) was performed. The cecum was exteriorized, ligated with a sterile silk suture 1 cm from the tip and double punctured with a 20-gauge needle. The cecum was squeezed to assure expression of a small amount of fecal material and was returned to the abdominal cavity. The incision was closed with auto-clips and kept clean by povidone-iodine (Betadine). Mice were resuscitated with i.p. injection of 1 ml of lactated Ringer's solution. Sham-operated mice were treated as described above with the exception for the ligation and puncture of the cecum. Pain was prevented by buprenorphine (0.1 mg/kg; subcutaneous (s.c.) 30 minutes before surgery and every 12 hours thereafter). For the survival study, mice were constantly monitored for 48 hours. Mice that survived this period of time were euthanized by cervical dislocation. In a separate set of studies, mice were sacrificed 12 hours following CLP by opening the chest cavity with an incision in the diaphragm. A sample of whole blood was collected for analysis of organ function using a comprehensive metabolic panel. Aortae were isolated and assessed for endothelial function. Finally, major organs were snap frozen and kept at -80°C until use. All animal procedures described in this study have been approved by the University of Texas Medical Branch Institutional Animal Care and Use Committee.

Samples of whole blood were collected from septic mice, placed in lithium-heparin tubes and immediately processed for quantitative determinations of alanine aminotransferase (ALT), albumin (ALB), alkaline phosphatase (ALP), amylase (AMY) total calcium (Ca²⁺), creatinine (CRE), glucose (GLU), phosphorus (PHOS), potassium (K⁺), sodium (Na⁺), total bilirubin (TBIL), total protein (TP), and urea nitrogen (BUN) as described [16] by a VetScan Chemistry Analyzer (Abaxis, Union City, CA, USA) using the VetScan Comprehensive Diagnostic Profile reagent rotor.

Blood from CLP or sham-operated mice was collected in K₂EDTA blood collection tubes and centrifuged at 4°C for 15 minutes at 2,000xg within 30 minutes of collection. Plasma was isolated, aliquoted and stored at -80°C until use. The EMD Millipore's MILLIPLEX™ MAP Mouse CVD Magnetic Bead Panel 1 kit was used for the simultaneous quantification of the following analytes: sE-Selectin, sICAM-1, Pecam-1, sP-Selectin, PAI-1 total, proMMP-9, and thrombomodulin (Merck Millipore, Darmstadt, Germany). Luminex uses proprietary techniques to internally color code microspheres with two fluorescent dyes and to create distinctly colored bead sets of 500 5.6 μm polystyrene microspheres or 80 6.45 μm magnetic microspheres, each of which is coated with a specific capture antibody. After an analyte from a test sample is captured by the bead, a biotinylated detection antibody is introduced. The reaction mixture is then incubated with streptavidin-phycoerythrin (PE) conjugate, the reporter molecule, to complete the reaction on the surface of each microsphere. The Luminex instrument acquires and analyze data using the Luminex xMAP fluorescent detection method and the Luminex xPONENT™ acquisition software (Thermo Fisher Scientific, Waltham, MA, USA).

Thoracic aortae were isolated from mice 12 hours following CLP or sham procedures. Aortic rings (approximately 2 mm in length) were placed in 5 ml organ baths filled with oxygenated (95% O₂ to 5% CO₂) Krebs-Henseleit solution at 37°C, were mounted onto isometric force transducers (Kent Scientific, Minneapolis, MN, USA), and connected to a PowerLab data acquisition system (AD Instruments, Dunedin, New Zealand) for the detection of isometric tension, as described [17]. Aortic rings were allowed to equilibrate for 45 minutes under a resting tension of 1.5 g and subsequently contracted with phenylephrine (PE; 1 μM). Once a plateau was reached, cumulative concentration response curves to acetylcholine (Ach) or to the nitric oxide (NO) donor sodium nitroprusside (SNP) were performed in order to assess the smooth muscle relaxation in response to endothelium-derived or exogenously applied NO. The experiments were repeated in at least four aortic rings, each from a different mouse.

Immunodetection of protein carbonyl groups (index of oxidative stress) was performed in tissue homogenates from thoracic aortas collected from mice undergoing 12 hours of severe sepsis, using the OxyBlot kit (Millipore) as described [18]. Protein carbonyl groups were derivatized with 2, 4-dinitrophenylhydrazine (DNPH) generating a stable dinitrophenylhydrazone (DNP) product immunodetected by specific anti-DNP antibodies. Following derivatization of the aortic tissue homogenates with DNPH, 20 μg of protein extracts were subjected to SDS-polyacrylamide gel electrophoresis (SDS-PAGE) using NuPAGE Novex 4% to 12% Bis-Tris Midi gels (Invitrogen, Thermo Fisher Scientific, Carlsbad, CA, USA). After electrophoresis, the proteins were transferred to a PVDF membrane. Membranes were then blocked in Starting Block T20 (TBS) blocking buffer (Thermo Fisher Scientific) and incubated with a rabbit anti-DNP primary antibody (1:500) provided by the manufacturer. The primary antibody was detected using an OxyBlot supplied horseradish peroxidase (HRP)-conjugated anti-rabbit immunoglobulin (Ig)G secondary antibody. Immunoreactivity was visualized by chemiluminescent detection. The Oxyblot method produces a smear-like blot, indicative of the oxidative modification of multiple proteins of different molecular weights. Blots were scanned and a densitometric analysis was performed on the totality of the modified proteins using NIH Image J software.

Tissue malondialdehyde (MDA) levels, an index of cellular injury/oxidative stress, were detected as described [16] in aorta, liver and mesenteric vascular bed homogenates from mice subjected to 12 hours of CLP-induced sepsis, using a fluorimetric MDA-specific lipid peroxidation assay kit (Enzo Life Sciences, Farmingdale, NY, USA) according to the manufacturer's instructions. The assay is based on the BML-AK171 method in which two molecules of the chromogenic reagent N-methyl-2-phenylindole (NMPI) react with one molecule of MDA at 45°C to yield a stable carbocyanine dye with a maximum absorption at 586 nm.

Myeloperoxidase activity was measured in aorta, liver and mesenteric vascular bed extracts from wild-type, eNOS^-/- and aged mice subjected to 12 hours of CLP-induced severe intra-abdominal sepsis, as described [16] using a commercially available myeloperoxidase (MPO) fluorometric detection kit (Enzo Life Sciences). The assay utilizes a non-fluorescent detection reagent, which is oxidized in the presence of hydrogen peroxide and MPO to produce its fluorescent analog. The fluorescence is measured at excitation wavelength of 530 to 571 nm and emission wavelength of 590 to 600 nm.

The IV lateral liver lobes were harvested from mice subjected to sham operation or CLP at 12 hours. Mitochondria from livers were isolated by differential centrifugation as described [19]. In brief, livers were homogenized on ice-cold mitochondrial isolation buffer (MSHE-BSA, pH 7.2) composed of 210 mm mannitol, 70 mm sucrose, 5 mm HEPES, 1 mm EGTA, and 0.5% (w/v) fatty acid-free BSA using a drill-driven teflon glass homogenizer. Homogenates were centrifuged at 600 g for 10 minutes at (4°C). Following centrifugation, the lipid phase was carefully removed by aspiration and the remaining supernatant was decanted to a separate tube and centrifuged at 10,000 g for 10 minutes (4°C). The supernatant fraction was removed, the pellet was suspended in MSHE-BSA, and the centrifugation procedure was repeated two more times. Finally, the pellet was resuspended in a small volume (approximately 100 μl) of MSHE-BSA. Total protein was determined using the Pierce BCA Protein Assay Reagent (Thermo Fisher Scientific). Mitochondrial preparations were used for bioenergetic analysis within 1 to 3 hours.

The XF24 Extracellular Flux Analyzer (Seahorse Biosciences, North Billerica, MA, USA) was used to measure mitochondrial bioenergetic function, as described [19]. Respiration by the mitochondria (10 μg/well) was sequentially measured in a coupled state with substrate present (basal respiration, State 2), followed by State 3 (phosphorylating respiration, in the presence of adenosine 5′-diphosphate sodium salt (ADP) and substrate), State 4 (non-phosphorylating or resting respiration) following conversion of ADP to adenosine triphosphate (ATP), State 4o, induced with the addition of oligomycin. Next, maximal uncoupler-stimulated respiration (State 3u) was detected by the administration of the uncoupling agent carbonylcyanide-4-trifluorometh-oxyphenylhydrazone (FCCP). At the end of the experiment the Complex III inhibitor, antimycin A was applied to completely shut down the mitochondrial respiration. This `coupling assay’ examines the degree of coupling between the electron transport chain (ETC.), and the oxidative phosphorylation (OXPHOS), and can distinguish between ETC. and OXPHOS with respect to mitochondrial function/dysfunction. In a second set of studies, electron flow experiments were conducted. This method allows the functional assessment of selected mitochondrial complexes together in the same time frame. Mitochondrial electron transport was stimulated by the addition of pyruvate/malate (10 mm/2 mm, respectively, in order to enable the activity of all complexes); with succinate (10 mm, in the presence of the Complex I inhibitor rotenone, 2 μM, in order to direct the electron flow exclusively through complexes II, III and IV) or with the artificial substrates ascorbate/TMPD (10 mm/100 μM, respectively, in the presence of the Complex III inhibitor antimycin at 4 μM, in order to selectively activate Complex IV) (3).

Data are shown as mean ± SEM. One-way and two-way ANOVA with Bonferroni's multiple comparison test were used to detect differences between groups. Statistical calculations were performed using Graphpad Prism 5 analysis software (GraphPad Software, La Jolla, CA, USA).

To determine whether mice exhibited an eNOS deficiency- and/or an age-associated vulnerability to sepsis, we compared mortality rates of eNOS^-/- and aged (24 months) versus young mice (C57BL/6, approximately two months of age) in response to CLP. Mortality was assessed every 12 hours for two days. At the end of the second day, the mice that survived were sacrificed by opening the chest cavity under deep anesthesia. In our model of sepsis, approximately 70% of the eNOS deficient mice died by 24 hours, showing the highest mortality rate among the animal groups studied (Figure 1). At the same time, there was 40% mortality in the aged mice. In contrast, no mortality was detected in the wild-type control group. By 48 hours, the survival rates of the aged mice and the young were 20% and 50% respectively; none of the eNOS^-/- mice used in the study survived to 48 hours.

In a different set of experiments, tissues and blood were harvested from eNOS^-/-, aged and young mice subjected to 12 hours of CLP. A complete metabolic panel for detection of plasma markers of organ injury was performed through the VetScan Chemistry Analyzer (Table 1). The analysis revealed a slight trend of CLP-induced increase of the liver injury markers ALT and ALP and the pancreatic injury marker AMY in wild-type mice. Significantly higher increases were noted in the same parameters, both in the aged and eNOS^-/- mice. Blood urea nitrogen (BUN), phosphorus (PHOS) and creatinine (CRE) levels (index of renal and hepatic failure) increased in the eNOS^-/- and aged mice but remained in the standard range in the young wild-type septic young mice. Analysis of tissue oxidative stress by the MDA assay and tissue mononuclear cell infiltration by the MPO activity assay were carried out in aorta, mesenteric bed (a vascular district known to be exquisitely abundant in endothelium and sensitive to eNOS/NO-dependent effects on the peripheral vascular homeostasis) and liver homogenates, showed an overall increase in sepsis-induced tissue damage both by aging and by eNOS deficiency (Figure 2).

Plasma levels of sE-Selectin, sICAM-1, Pecam-1, sP-Selectin, PAI-1 total, proMMP-9, and thrombomodulin were simultaneously detected in young, aged or eNOS^-/- mice 12 hours following CLP or sham procedure (Figure 3). Low baseline values of the markers of endothelial dysfunction were detected in sham-operated mice, grossly unchanged within the groups. Although 12 hours is a relatively short time to expect increased plasma levels of sE-Selectin, sP-Selectin and Pecam-1 in septic young mice compared to the sham group, their levels did increase significantly in the aged and eNOS-deficient septic mice. The analysis of the data revealed a similar trend of increase for pro-MMP-9 and thrombomodulin. Twelve hours of CLP induced about 15-fold increase in both pro-MMP-9 and thrombomodulin plasma levels in young septic mice compared to young, sham-operated mice. More pronounced increases were detected in the septic eNOS^-/- with the septic aged mice showing the highest levels among the groups. No statistically significant differences were detected between young, aged and eNOS-deficient septic mice for plasma levels of SICAM-1 and PAI-1. Detection of protein carbonyl groups, index of oxidative damage, was performed by OxyBlot analysis of aortae isolated from mice subjected to CLP. As shown in Figure 4A and B, some protein carbonylation was already detectable in wild-type septic mice, but the induction (measured by software-assisted densitometric analysis of the chemiluminescent signal) increased more substantially (by about three- and two-fold, respectively) in aged mice and eNOS-deficient mice. Dysfunction of the endothelium during sepsis was determined by assessing the Ach-induced vasorelaxant responses in vitro in aortic rings precontracted with phenylephrine. As expected, aortic rings from eNOS^-/- mice were completely unable to relax in response to Ach (Figure 4C and D). Aged mice showed some basal level of endothelial dysfunction, as evidenced by a shift to the right of the sigmoidal concentration response curve to Ach in the sham group. CLP-induced sepsis produced a significant degree of endothelial dysfunction, the severity of which was more profound in the aged mice compared to the young mice (mean LogEC₅₀ from -7.8 to -7.5 in young mice, sham versus CLP; mean LogEC₅₀ from -7.3 to -6.8 in aged mice, sham versus CLP). Moreover, both aged and eNOS-deficient vessels showed increased sensitivity to the NO donor SNP (Figure 4E and F).

In the current model of severe sepsis, the survival rate in the young group of mice was approximately 50% at 48 hours. In the next set of studies, we aimed to investigate whether the degree of endothelial dysfunction correlates with the outcome of sepsis. Mice were divided into two groups: those that survived and those that did not survive 48 hours of sepsis. Mice were closely monitored for the 48-hour period of time and thoracic aortae from the non-survivors were harvested at the time of death and processed for assessment of in vitro vascular response to endothelium-derived NO by Ach stimulation. Mice that survived the 48 hours of sepsis were euthanized by opening the chest cavity and the aortic ring assay was performed subsequently. A distinct difference of severity of endothelial dysfunction was detected between the young mice that survived and those that died (Figure 5). In a subsequent study, a similar approach was used for the assessment and comparison of endothelial function in survivor versus non-survivor aged mice. Since the survival rate at 48 hours of CLP is only 80% in the case of the aged mice, we restricted the CLP to 24 hours. Aortic rings were harvested from five non-survivor mice at the time the death was scored (see figure legend for additional methodological details). The survivors were euthanized 24 hours following CLP. These data indicate that endothelial dysfunction precedes mortality and suggest that it may be a predictor of it.

CLP caused a significant reduction of the total mitochondrial function based on different bioenergetic measurements (`coupling’ and `electron flow’ assays) in isolated mouse liver mitochondria ex vivo (Figure 6A and B), indicative of an overall, sepsis-associated impairment of mitochondrial function. A decrease in oxygen consumption rate (OCR) was also noted both in sham-operated aged and eNOS^-/- mice compared to the sham-operated young mice (all in the absence of CLP). CLP, in turn, tended to further worsen mitochondrial function in all three groups (Figure 6C and D).