| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
From the W.M. Keck Center for Transgene Research and the Department of Chemistry and Biochemistry, University of Notre Dame, Notre Dame, Indiana
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
700,000 people annually in the United States, with a mortality of 30%.1,2
This disease is associated with a response to infection causing systemic inflammatory response syndrome, which involves a variety of symptoms including hypothermia and hyperthermia, tachycardia and tachypnea, and anomalous white blood cell counts.3
As a consequence, the inflammatory system is hyperactive and evokes both cellular and humoral responses, such as the release of cytokines and chemokines that serve as mediators for crosstalk between endothelial cells (ECs) and epithelial cells on one hand, and neutrophils and macrophages on the other.4,5
The inflammatory state is closely related to coagulation through several connecting elements, which include: 1) receptors and membrane proteins, eg, endothelial protein C receptor (EPCR), thrombomodulin (Tm), and protease-activated receptors (PAR); 2) hemostasis-related proteins, eg, tissue factor (TF), factor (F) VII, FXI, and FXII; 3) systems that regulate thrombin formation, such as protein C (PC) system components, tissue factor pathway inhibitor, and anti-thrombin-III; 4) inflammatory mediators, eg, tumor necrosis factor (TNF)-
, interleukin (IL)-1, IL-6, and inducible nitric oxide synthase; and 5) cell adhesion mediators for extravasation of inflammatory cells, such as E-selectin, P-selectin, and integrins.6
All of these pathways function at early stages of the infection in attempts to rid the system of the infectious agent, but undergo a loss of control in severe sepsis, leading to unregulated hemostasis. In many cases, disseminated intravascular coagulation (DIC) is established,7-9
with resultant contributions to presentations of hypotension, hypoperfusion, and multiple organ dysfunction syndrome. This advanced stage of the disease normally culminates in multiple organ failure and death. Natural anti-coagulant mechanisms, namely the PC pathway, anti-thrombin-III, and tissue factor pathway inhibitor, as well as a functional fibrinolytic system, are needed to maintain hemostasis. However, after systemic inflammatory response syndrome in sepsis, anti-thrombin-III and PC are consumed, and damage to the endothelium impairs the EC-derived Tm- and EPCR-modulated activation of PC.10 In addition, hypofibrinolysis is a typical outcome during advanced stages of the disease,11 thus exacerbating the potential coagulopathy. These findings have led to many approaches for attempting to control the multiple septic responses, and replacement of natural endogenous anti-coagulants was believed to offer potential benefits. However, although clinical trials in severe sepsis with anti-thrombin-III and tissue factor pathway inhibitor replacement therapies have proved disappointing,12-14 suppression of thrombin generation by recombinant human activated PC (aPC) may indeed correlate with improved survival,15 at least in cases of severe sepsis coupled with DIC.
In principal, the aPC system offers many advantages as a combined anti-thrombotic/profibrinolytic/anti-inflammatory therapeutic agent that is capable of down-regulating thrombin generation by attenuating levels of FV and FVIII,16,17
and enhancing clot dissolution by stimulating fibrinolysis directly via plasminogen activator inhibitor-1 (PAI-1) inactivation.18
In addition, an indirect fibrinolytic effect of aPC is observed through reduction of thrombin levels and consequent activation of thrombin activatable fibrinolytic inhibitor (TAFI).19,20
In vitro experiments suggest that aPC also serves as an inflammatory modulator through several routes, eg, down-regulation of PAR-mediated thrombin-induced proinflammatory signals,21
inhibition of the generation of cytokines, such as TNF-,22
modulation of macrophage migration inhibitory factor (MIF),23
and inhibition of the responses of human mononuclear phagocytes to lipopolysaccharide (LPS), interferon-
, or phorbol ester.22,24
In addition, aPC attenuated ICAM-1 and E-selectin expression,25
and inhibited apoptosis in EC cultures25
through its receptor/co-receptor relationships with EPCR and PAR-1.26
aPC has also been shown to inhibit LPS-induced translocation of nuclear factor (NF)-
B in monocytes,27
to inhibit leukocyte cell adhesion,28
and to attenuate LPS-induced pulmonary vascular injury by inhibiting activated leukocytes,29
all potentially contributing to its anti-inflammatory effectiveness independent of its anti-coagulant activity. In a rat model of endotoxemia, aPC prevented LPS-induced hypotension, an effect that was dependent on its serine protease activity.30
Although administration of aPC appears to be beneficial to a subset of patients with severe septicemia, there is little available evidence to suggest that deficiencies in PC would lead to a more severe host response to the septic state,31 ie, although therapeutically somewhat effective, there is no proof that aPC is on the mechanistic pathway to septic responses. Thus, a new experimental model of PC-deficient mice becomes a valuable tool to study gene associations in this disease, particularly with regard to those genes that might have a reciprocal relationship with the PC pathway. A model of LPS-induced endotoxin shock was recently used in mice heterozygous for a targeted PC deficiency (PC+/–).9 Although the LPS dosages in this case were very high and somewhat mechanistically unrealistic, PC+/– mice showed a higher mortality rate, a more elevated inflammatory response, and a more severe DIC when compared to wild-type (WT) mice.
Various animal models are available for the study of sepsis,32 with LPS bolus injection and cecal ligation and puncture (CLP) among those models in widest use. Each has its advocates and critics. The CLP model has gained wide acceptance in the scientific community because of its polymicrobial nature, and is considered to reproduce the pathophysiology of sepsis closely. In the present study, we used the CLP model to explore the outcome of polymicrobial sepsis in PC+/– mice in a detailed manner. In particular, hemodynamic parameters and gene expression in major organs were analyzed. The current article presents a summary of our findings.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Mice were housed in microisolation cages and given access to food and water ad libitum. The procedures for generation of mice heterozygous for a null PC allele (PC+/–) have been described earlier.33 PC+/– mice were bred to the >F10 generation in the C57BL/6J strain. All animal experiments were performed in accordance with protocols approved by the Institutional Animal Care and Use Committee.
Genotypic Analyses
PC genotyping was performed on ear punch or tail-tip DNA by polymerase chain reaction (PCR) methodology using three different primers: 1) a common forward primer (CGTGATGAGTTTCAGGCAGTGAGAG) for both the WT and null alleles, located 300 bp downstream of the beginning of the first intron of the PC gene; 2) a specific reverse primer (GAGATAAGCAGATCCTGTGGATTGC) for the WT allele situated in intron 1 at the 5' flank of the PC gene at a location 200 bp upstream of the start codon of exon 2; and 3) a reverse primer (ATTCGCAGCGCATCGCCTTCTATC) present in the NEO cassette for the null allele. Amplicons of 1000 bp and 600 bp were obtained for the WT and null alleles, respectively.
The PCR conditions were as follows: samples were initially denatured for 5 minutes at 95°C, followed by 30 cycles of denaturation (95°C, 30 seconds), annealing (64°C, 50 seconds), and extension (72°C, 1 minute and 50 seconds). PCR was concluded with one additional extension period at 72°C for 10 minutes. PCR reactions were performed on a DNA engine Tetrad, PTC-225 (MJ Research, Watertown, MA).
CLP Surgery
For CLP,32,34,35 male WT and PC+/– mice (ages 2 months and 6 months) were sedated with rodent cocktail (9 mg/ml ketamine, 1.8 mg/ml xylazine, 0.3 mg/ml acepromazine in saline) before surgery, which was performed under a surgical microscope. The abdomens of the sedated mice were shaved. A longitudinal 1-cm incision was then made in the peritoneum, the cecum was localized and exposed, and then ligated just below the ileocecal valve with a sterile 5.0 silk ligature. The anti-mesenteric cecal surface was then punctured twice with a 20-gauge needle. The bowel was then replaced into the peritoneal cavity.
Necropsy
Animals were sacrificed at 6 hours or 24 hours after CLP surgery. For experimental purposes, the animals were divided in four different groups. The first group was maintained undisturbed up to 1 week for the survival analysis. The second group was sacrificed by exsanguination via the orbital vein and the blood placed into 3.8% citrate, at a volume ratio of 9:1, blood:citrate. An amount of 25 µl of this sample was diluted immediately in 475 µl of 3% acetic acid (Unopette, Becton Dickinson Vacutainer System; Becton Dickinson, Franklin Lakes, NJ) to lyse red blood cells. White blood cells were then counted on a Neubauer (Fisher Scientific, Pittsburgh, PA) hematocytometer using x100 magnification under a transmission microscope. Another drop of blood was spotted onto histological slides and smeared to perform white blood cell differentials. The resulting blood smear slides were immediately treated with a fixative spray (CH3OH, HOAc) and stained with Giemsa solution. The remaining plasma that was obtained from the mice was retained for further use in plasma assays. At this point, the mice were disinfected by immersion in 70% alcohol for peritoneal lavage. A 20-gauge sterile venous catheter was connected to a 3-ml syringe containing 2 ml of sterile saline and inserted into the peritoneum. Peritoneal lavage fluid (PLF) was collected in the syringe and dispensed in an Eppendorf tube. PLF (25 µl) was diluted into 475 µl of 3% acetic acid (Unopette microcollection system) for leukocyte counts. Differentials were performed on cytospin slides, as above. The third group of mice was first anesthetized with rodent cocktail at the proposed sacrifice time points. Blood was collected through vena cava puncture into 3.8% citrate at a volume ratio of 9:1, citrate:blood, for plasma assays of coagulation factors. The mice were then perfused through the left ventricle with saline until no additional blood emerged from the right auricule. The liver, kidneys, and spleen were removed and divided for histological analyses, myeloperoxidase (MPO) assays, and RNA isolation, for which the latter samples were stored in RNA Later (Ambion, Austin, TX) at –20°C. The fourth group of mice was used for heart rate and blood pressure (BP) monitoring.
Implantation of the BP and Heart Rate Probes
At a time of 7 to 10 days before CLP, mice were sedated with isoflurane using a ventilator and placed under a surgical microscope. A longitudinal 1-cm incision was made on the neck. The salivary glands were separated and the trachea and left carotid artery were exposed. The distal end of the carotid artery was ligated with a 6.0 silk thread to avoid back-flow or flow-through. The pressure sensor (Transoma Medical-DSI, St. Paul, MN) that is encapsulated in a cannula was filled with a gel provided by the manufacturer. A loose loop was made in the proximal end of the carotid artery with a 6.0 silk thread and the cannula was inserted up to the aortic arch. The proximal loop was tied to restrain the cannula. The remainder of the probe (battery and transmitter) external to the cannula was inserted in a pocket made between the skin and the muscle layer covering the ribs on the left side of the mouse. Finally, the skin was sutured with 5.0 ethicon thread (Harvard Apparatus, Holliston, MA). After CLP surgery (performed under isoflurane anesthesia), the heart rates, along with diastolic and systolic pressures, were continually recorded.
Histochemistry and Immunohistochemistry
Lungs, livers, and kidneys were fixed in 4% paraformaldehyde-phosphate-buffered saline (PBS) for 2 hours, embedded in paraffin blocks, and serially sectioned at 3 to 4 µm thickness. Tissue sections were deparaffinated, rehydrated, and stained with hematoxylin and eosin. In addition, immunostains for fibrin(ogen) and leukocytes (CD45) were performed to complete the analysis of the histopathology.
Colony Forming Unit Assays
Fresh PLF (10 µl) samples were serially diluted up to 1:105 in saline and 10 µl of each dilution was plated on LB plates and incubated at 37°C overnight. After this time, the colonies were counted and the results were expressed as colony-forming units/10 µl.
Plasma Assays
All assays were performed at room temperature.
PAI-1 Levels
PAI-1 levels were determined using a two-antibody-sandwich enzyme-linked immunosorbent assay (ELISA) in 96-well microtiter plates coated with a monoclonal mouse anti-mouse PAI-1 antibody (H34G6; Molecular Innovations, Southfield, MI). The wells were blocked with 5% milk. The coated plates were treated with 100 µl of plasma (1:5 dilution in PBS) or standard (0 to 12.5 ng/ml murine PAI-1; Molecular Innovations), and incubated for 2 hours. After four washes with PBS/Tween-20, 100 µl of 4 µg/ml rabbit anti-mouse PAI-1 polyclonal antibody (Molecular Innovations) in PBS was added to the wells and incubated for 2 hours. The plates were then washed and 100 µl of a 1:2000 diluted goat anti-rabbit IgG-alkaline phosphatase conjugate (Bio-Rad, Richmond, CA) was added and further incubated for 2 hours. After washing 4 times, 100 µl of the colorimetric substrate pNPP (Sigma, St. Louis, MO) was added and incubated for 30 minutes. The absorbance was measured at 405 nm in a plate reader.
PC Assays
PC assays were also performed using ELISA to detect its levels in plasma using an in-house polyclonal monospecific rabbit anti-mouse PC to coat the wells. The plasma sample (100 µl of a 1:10 dilution of plasma in PBS) or standard (0 to 5 nmol/L recombinant murine PC, provided by Dr. Angelina Lay of this laboratory) in PBS was added to the plates and incubated for 2 hours. After suitable washing with PBS/Tween-20, an in-house chicken anti-mouse PC polyclonal antibody in PBS was added to the wells and incubated for an additional 2 hours, followed by addition of goat anti-chicken IgG-horseradish peroxidase conjugate (Bio-Rad), and further incubated for 2 hours. The colorimetric substrate TMB (3,3',5,5'-tetramethylbenzidine) was used for development. After incubation for 30 minutes in the dark, the absorbance at 540 nm was determined and subtracted from the absorbance at 450 nm.
Fibrinogen Assays
Clottable fibrinogen levels were measured using the Fibri-Prest Automate coagulometric assay kit (Diagnostica Stago, Parsippany, NJ), which consists of the determination of the clotting time of the plasma sample after the addition of an excess amount of thrombin. Calibrations were accomplished with different concentrations of the human plasma Unicalibrator from the assay kit. The assay was performed in a StART 4 coagulometer (Diagnostica Stago).
Activated Partial Thromboplastin Time (APTT) Assays
Activated partial thromboplastin time (APTT) assays were performed on plasma samples that were incubated with an equal volume of PTA reagent (Diagnostica Stago) for 3 minutes at 37°C. Clotting times were determined by coagulometry after the addition of an equal volume of a 25 mmol/L CaCl2 solution.
Prothrombin Time (PT) Assays
Prothrombin time (PT) assays were accomplished by measuring clotting times by coagulometry after the addition of an equal volume of PT reagent (thromboplastin, Diagnostica Stago).
FXI and FXII Levels
FXI and FXII levels were determined on dilutions of plasma samples that were added to FXI- or FXII-deficient plasmas (Diagnostica Stago) and APTT reagent (Diagnostica Stago) for 3 minutes at 37°C. Clotting times were determined by coagulometry after the addition of 25 mmol/L CaCl2. A calibration curve was constructed using different dilutions of artificial plasma, and the results are expressed relative to the average FXI or FXII mouse levels for the resting WT group.
Cytokine Assays
Cytokine assays were determined on plasma and PLF. TNF-, IL-6, and MIP-2 were determined using Quantikine-M murine kits (R&D Systems, Minneapolis, MN). For the simultaneous detection of other cytokines (IL-1ß, IL-2, IL-4, IL-5, IL-10, IL-12p70, and interferon-) and chemokines (MIP-1, KC, Eotaxin, RANTES, TARC, SDF1ß, JE, and MCP5), a multiplex ELISA assay was used (Pierce Biotechnology, Rockford, IL). The assay is a traditional quantitative sandwich ELISA, and the difference relies on the fact that each well of a microtiter plate is precoated with an antibody microarray, thus allowing for detection of up to 36 different analytes. In this case, the plates were analyzed using a SearchLight CCD camera (Pierce Biotechnology).
Blood Chemistry Analyses
Blood chemistry analyses were performed on plasma samples prepared from blood collected from the vena cava. Determinations were made of alkaline phosphatase, amylase, aspartate aminotransferase, creatinine kinase, creatinine, -glutamyltransferase, glucose, lactate dehydrogenase, lipase, inorganic phosphate, total bilirubin, and urea. The VET-TEST 8008 (IDEXX, Ellisvile, MO) was used for the assays. For 12 tests, 150 µl of plasma was required. Standards were supplied by the manufacturer.
MPO Activities
MPO activities were performed as described earlier36 on samples from livers, kidneys, and spleens from saline-perfused animals, first using a heat step to eliminate spurious peroxidase activity. The MPO assays were performed on lysed perfusate cells in microtiter 96-well plates by adding 50 µl of 0.75 mmol/L H2O2, 80 mmol/L PBS, pH 5.4, 50 µl of sample or standard (in triplicate), and 25 µl of 8 mmol/L TMB in dimethyl sulfoxide. The plates were incubated in the dark for 2 minutes and the reaction was terminated with 50 µl of 1 mol/L HCl. The absorbance at 540 nm was subtracted from the absorbance at 450 nm and the values were converted to units of peroxidase activity using the standard curve that is simultaneously generated during assays of the samples. The standards were prepared using various concentrations (0 to 0.5 U) of horseradish peroxidase (POD; Roche, Indianapolis, IN) in 0.5% CTAB, 50 mmol/L PBS, pH 6.0.
Expression and Quantitation of Hemostasis and Inflammation Transcripts
RNA Isolation
A 15- to 20-mg portion of liver, kidney, or lung was added to 2 ml of DNase/RNase-free screw cap tubes containing 1 g of 1-mm-diameter zirconia beads (Biospec Products, Bartlesville, OK) and 600 µl of RLT buffer (Qiagen, Valencia, CA). The samples were then homogenized in a Fast-Prep FP120 shaker (Thermo Savant, Holbrook, NY). The purification procedure continued using the RNeasy system (Qiagen). Once isolated, the RNA samples were analyzed for purity and quantitation in an Agilent Bioanalyzer 2100 (Agilent, Palo Alto, CA).
Gene Expression Assays
Expression of mRNA of IL-1, IL-6, TNF-, MPO, TF, PAI-1, TM, and EPCR was studied at the different time points in liver, kidney, and lung by quantitative (Q) reverse transcriptase (RT)-PCR. Expression of coagulation FXI and FXII was quantitated only in liver, and the expression of PC was determined in liver and kidney. Q-RT-PCR was performed in one step in a 7700 ABI PRISM TaqMan real-time PCR apparatus (Applied Biosystems, Foster City, CA).
Each reaction contained 100 ng of total RNA, 2 µl of 10x Master Mix (dNTP mixture with dUTP instead of DTTP), 6 mmol/L MgCl2, 50 nmol/L each primer, 100 nmol/L probe, 16 U RNase inhibitor (Promega, Madison, WI), 2 U Multiscribe reverse transcriptase (ABI), and diethyl pyrocarbonate-treated H2O (20 µl). Thermocycling conditions were, sequentially: 48°C for 30 minutes and 95°C for 10 minutes, 40 cycles at 95°C for 15 seconds, 60°C for 1 minute, with fluorescent readings at the end of each cycle. The amount of mRNA detected was quantitated using individual standard curves. All samples were run in triplicate. FAM fluorescence was measured at the end of each 60°C period. The mRNA levels of the gene of interest were expressed as a ratio to the housekeeping gene, RPL19, and the averages for each experimental group were expressed relative to the WT resting levels for a particular gene.
Statistical Treatment of the Data
For survival analysis, the data were analyzed by the Kaplan-Meyer treatment and the comparison of survival between different genotypes was accomplished using the log-rank test with the NCSS software (NCSS, Kaysville, UT). For data analysis and comparison of single pairs, the Student’s t-test was used. For overall group comparisons, Kruskal-Wallis one-way analysis of variance test was performed. Tukey-Kramer tests were conducted to analyze multipairwise comparisons.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
The initial experiments that compared survival of WT versus PC+/– mice after CLP were performed with young mice at 2 months of age. No significant survival differences in these two groups were found by Kaplan-Meier analyses of the data. However, similar experiments conducted with older mice at 6 months of age did indeed show a significant difference in survival between these groups (Figure 1)
. Additionally, survival at 6 months of age after CLP was significantly lower than the survival presented by 2 months of age mice under the same treatment. Thus, it is clear that there is an age dependency involved in the susceptibility of PC+/– mice in this model of polymicrobial sepsis, and the remainder of the studies described herein was performed on mice at 6 months of age.
|
Hypotensive Effects after CLP Surgery
A hallmark feature of the fatal course of sepsis in humans is the severe hypotension and shock that occurs. For this reason, we have measured continual central BP and heart rates of mice after CLP, using a radiotelemetric method after insertion of suitable probes in the central thoracic aortas of WT and PC+/– mice. The probes were placed in the mice (n = 4 in each group) 1 week before the CLP procedure in order that the BP and heart rates would stabilize before the experimental procedure. The experiments began after these values returned to normal. Point-by-point average data of synchronized plots are shown in Figure 2
. In WT mice, heart rates began to decline at a time 5 hours after CLP surgery, and were minimal at 6 to 8 hours after CLP (Figure 2a)
. Significantly, PC+/– mice were more susceptible to this change than WT mice, with heart rates of this latter group decreasing to 50% of their pre-CLP values. The central BP (average of systolic-diastolic pressures) decreased in both groups of mice and paralleled the temporal responses of the heart rates. These values decreased to significantly lower levels in PC+/– mice, relative to the BP of WT mice (Figure 2b)
. Thus, a more severe sepsis-induced shock state accompanied a PC deficiency, a heretofore unrecognized potential function of PC.
|
Leukocyte counts decreased steeply and significantly, from values of 6000 to 7000/µl in WT and PC+/– mice in presurgical blood, to 2000 to 2500/µl in both groups 6 hours after CLP (Figure 3a)
. At this latter time point, there was no significant difference between the decreases in WT and PC+/– mice. However, although circulating leukocytes increase somewhat at 24 hours in WT mice, this does not occur in PC+/– mice, and leads to significantly higher blood leukocyte counts in WT mice as compared to PC+/– mice, both of which remain significantly lower than the values found in presurgical mice. A significantly higher leukocyte count has been found 6 hours after CLP PLF of PC+/– mice, relative to similar WT mice. Large increases in PLF leukocyte counts occur at 24 hours after CLP (Figure 3b)
, but the differences between the groups seen at 6 hours after surgery are not evident at the 24-hour time point. The PLF is enriched in neutrophils in both groups of mice at 24 hours after CLP, displaying counts of (10,000 ± 3000 neutrophils/µl blood in WT mice and 11,000 ± 2000 neutrophils/µl blood in PC+/– mice, as compared to resting mice (2500 ± 500 neutrophils/µl blood in both WT and PC+/– mice). These data demonstrate the increased neutrophil migration into the peritoneal cavity as a major inflammatory response in this model, as is the case with the disease state in humans.
|
Histological examination of the tissues did not show severe pathologies at 24 hours after CLP surgery, not unexpected when taking into account the acute nature of the process that led to shock and sudden death. Particularly, analysis of hematoxylin and eosin-stained slides of WT and PC+/– liver sections presented leukocyte accumulation and necrosis in the parenchyma, a typical sign of peritonitis. Lungs showed edemas in the alveolar spaces and leukocyte accumulation in the peripheries of the pulmonary arterioles, an indication of a mild pneumonia. Kidneys displayed a mild proteinurea in the tubuli and the hearts showed no pathological symptoms. Additionally, fibrin(ogen) immunostaining did not show any signs of fibrin(ogen) deposition in organs, with the exception of the edemas found occasionally in lungs. Further, CD45 staining did not reveal evidence of leukocyte infiltration. Overall, no overt differences in histology were found between WT and PC+/– organs.
Coagulation Factor Assays
Because a strong potential for coagulopathy and DIC are common features of severe sepsis, and are displayed to various degrees in different animal models of sepsis, assays of critical coagulation factors that might produce this state have been performed. APTTs increased from 37 ± 2 (n = 13) seconds and 57 ± 8 seconds (n = 12) in resting WT and PC+/– mice, respectively, to 70 ± 12 seconds (n = 9) and 91 ± 14 seconds (n = 7) in WT and PC+/– plasmas 24 hours after CLP, respectively. PT values increased from 11.8 ± 0.3 seconds (n = 4) and 11.0 ± 0.8 seconds (n = 4) in resting WT and PC+/– mice, respectively, to 23 ± 3 seconds (n = 4) and 20 ± 3 seconds (n = 4) in WT and PC+/– plasmas 24 hours after CLP, respectively. Additionally, no significant differences were found in these parameters between WT and PC+/– mice at any particular time point. This suggested that consumption of coagulation factors has occurred after septic challenge, thus showing that the coagulation system has been activated. Plasma PC levels did not change from their resting values at 6 hours after CLP in both groups of mice, but were significantly attenuated in each at 24 hours after surgery. At all time points, plasma PC in PC+/– mice maintained its ratio of 50% of that of WT plasma (Figure 4a)
. Plasma fibrinogen levels, determined at 6 hours after CLP, were unchanged from those of resting mice, and were not different between WT and PC+/– mice (Figure 4b)
. However, at 24 hours after CLP, the plasma fibrinogen levels were significantly elevated by 2- to 2.5-fold, with a significantly larger increase found in WT mice relative to PC+/– mice. These increases at 24 hours most likely reflect the acute phase response of fibrinogen, which more than counterbalanced any possible consumptive coagulopathy resulting from the inflammatory response. Further and substantially contributing to a possible prothrombotic state were the elevations observed in plasma levels of PAI-1, which increased approximately fivefold at 6 hours and 24 hours after CLP, when compared to these same values in plasmas of resting mice of both genotypes (Figure 4c)
. Lastly, intrinsic pathway components, specifically FXII and FXI, at 24 hours after CLP, declined to 50% of the values of WT and PC+/– mice. No significant differences were observed between the two genotypes. This decrease is, most likely, because of the consumption of the coagulation factors because of the hypercoagulable state evident at this time point.
|
For measurement of the plasma cytokine responses in both groups of mice, the 24 hours after CLP surgery was selected. The choice for this particular time point was justified by previous experimental evidence.37
After CLP surgery, cytokine levels keep increasing at 8 hours after CLP and reach a plateau at 10 hours later. Measurements at earlier times might reflect the early response to the surgery itself (anesthesia and opening of abdomen). Moreover, the 24-hour time point was shown in the past to be effective for these types of measurements after CLP.38-40
Plasma levels of 18 cytokines and chemokines were thus measured in WT and PC+/– mice 24 hours after CLP surgery. Both groups showed low levels (less than 100 pg/ml) of MIP-1, IL-1ß, IL-4, and IL-12, moderate levels (100 to 2000 pg/ml) of TNF-, interferon-, MIP-2, MCP-5, RANTES, TARC, SDF1ß, IL-2, and IL-5, and highly elevated levels (more than 2000 pg/ml) of IL-6, IL-10, KC, JE, and eotaxin at this time point. Although differential trends were observed in many of these groups, individual variations between mice were too large to achieve statistical significance with a reasonable number of mice. By pooling plasma samples, this issue can be artificially controlled, because the large individual variations that are observed in many cases would not be recognized, but we did not feel that this was an effective manner of proceeding.
However, levels of several cytokines, namely IL-1, IL-2, IL-5, IL-6, and IL-12, and also possibly TNF-, were significantly higher in PC+/– mice 24 hours after CLP in comparison to similarly treated WT mice, and these values are listed in Table 1
. These cytokines were undetectable in plasmas of untreated or in sham-surgical mice. Additionally, the differences in plasma levels of IL-4, IL-10, and KC approached statistical significance (P < 0.1), being more elevated in PC+/– mice 24 hours after CLP than in similarly treated WT mice. Although these may not be the only cytokines/chemokines that are differentially elevated in this model, very large numbers of mice would be required to achieve significance for many of these other agents, and we felt that this initial screen was ample for highlighting some critical inflammatory components that are possibly differentially affected in this model.
|
At 24 hours after CLP, blood levels of creatinine kinase (Figure 5a)
and creatinine (Figure 5b)
were elevated in both WT and PC+/– mice, as compared to control mice. For both of these parameters, the values were significantly higher in the case of treated PC+/– mice, as compared to similar WT mice. With regard to blood urea nitrogen values (Figure 5c)
, only the PC+/– mice displayed significantly higher values than the other groups tested. Hepatic injury was observed to a similar extent in both WT and PC+/– mice, as shown by the increase in aspartate aminotransferase at 24 hours after CLP surgery (Figure 5d)
. These results suggest that kidney and heart injury occur to a greater extent in the septic PC+/– mice, at the time point measured, than in WT mice, and this can be a contributing factor to the higher mortality rate of these latter animals after septic challenge.
|
The amounts of mRNA from genes related to the coagulation cascade, such as TF, PC, EPCR, TM, PAI-1, FXI, and FXII, or linked inflammatory components, such as MPO, IL-1ß, IL-6, and TNF-, were determined by RT-PCR from liver, kidney, and lung tissues. The primers and probes used for each assay are listed in Table 2
.
|
. In livers, TF mRNA levels decreased 24 hours after challenge, whereas lung expression of TF increased significantly at the 24-hour time point. In kidneys, TF expression displayed an increase 24 hours after CLP surgery relative to resting levels. No significant differences were observed in hepatic, pulmonary, or renal TF expression between WT and PC+/– mice at the 24-hour time point. EPCR expression showed a tendency to increase in liver and lungs 24 hours after CLP. Importantly, this temporal increase was more remarkable in kidney, also showing higher EPCR levels in PC+/– mice compared to WT individuals at the late time points (Figure 6b)
. TM also displayed increases at 24 hours after CLP in kidney and liver. Although there were no significant differences between resting WT and PC+/– mice, higher TM mRNA levels were observed in PC+/– livers, relative to WT livers, at the 24-hour time point (Figure 6c)
. In resting PC+/– mice, PC gene expression was, 60% of the PC mRNA levels of resting WT mice, in both liver and kidney, where PC expression totaled 33% of the liver expression. PC mRNA levels decreased at 24 hours after CLP, relative to resting mice. Interestingly, whereas WT individuals displayed a 50% decrease in PC expression in liver and kidney at the 24-hour time point, relative to the resting stage, the decrease in PC expression was more dramatic in PC+/– mice (66% in kidney and 75% in liver) at the same time point (Figure 6d)
. These results are consistent with the plasma PC levels determined by ELISA.
|
. At the 24-hour time point, hepatic PAI-1 expression was significantly higher in PC+/– mice, compared to WT individuals.
|
. This finding may support the undetectable MPO activity measured by the functional assay (not shown), and provided additional evidence of the poor leukocyte infiltration seen in the histology (not shown). Levels of TNF- remained low before and after challenge and showed no variation when compared in the different experimental groups (data not presented). Levels of IL-1ß mRNA increased 24 hours after challenge relative to resting mice in kidneys and, less significantly, in livers. In both organs, there was a tendency toward higher expression of IL-1ß in PC+/– mice, compared to WT mice (Figure 7c)
. Finally, expression of IL-6 showed a remarkable difference between resting and challenged mice in kidney (3000-fold). In the liver, this difference was less pronounced (100-fold) and less significant. Also, higher levels of IL-6 mRNA were observed in liver and kidney from PC+/– mice, compared to their WT littermates, 24 hours after challenge (Figure 7d)
. It is believed that the elevated expression of IL-1ß, IL-6, and PAI-1 in PC+/– mice was mainly because of EC expression levels and additionally Kupffer cells in the liver, but not expression by leukocytes, based on the histology and MPO results. |
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Investigators in a previous report9
used a model of gram-negative endotoxin shock in PC+/– and WT mice after LPS administration (50 mg/kg). A comparatively higher mortality rate in PC+/– mice was attributed to the exacerbation of DIC in PC+/– mice as compared to WT mice. The dose of LPS used in this work was approximately fivefold higher than the dose applied in comparable studies. The administration of such a high level of LPS may cause a response because of the toxicity of the endotoxin per se in terms of its direct interaction with organs, instead of a systemic inflammatory response that would result from a lower dose of this agent. In the present study, CLP was used as a model of polymicrobial sepsis and focused on hemodynamic parameters, organ failure, and focal gene expression. It is believed that CLP mimics sepsis symptoms in terms of the kinetics of cytokine expression much more accurately than the LPS model used in the previous study, which showed differences with clinical studies and with the model used in this work.3,45
As examples, serum TNF-37,46,47
and LPS levels48
were not elevated in the same kinetic manners in septic patients or CLP-treated mice, as was found in LPS-injected mice. Moreover, the kinetics and magnitude of cytokine expression described by the LPS models do not reproduce appropriately the cytokine profile of sepsis.37
However, CLP may present other types of complications because of its polymicrobial nature, because it is not feasible to control the type and titer of the gastrointestinal flora in different individuals. Because the intestinal microflora is predominantly composed of gram-negative anaerobic bacteria, eg, Escherichia coli, we performed PLF bacterial counts on LB media appropriate for this type of microorganism. There were no significant differences in the colony-forming units found in WT and PC+/– peritoneal lavages at 24 hours after CLP. This result may indicate that the increased mortality in PC+/– mice is not because of bacterial count differences, which had been suggested earlier.49
Among the original goals of the present study, was to conduct a comparative analysis of sepsis between young mice (8 to 12 weeks old) and older mice (26 to 30 weeks old) to evaluate the prognosis of PC+/– mice with age. There was no difference in the survival of WT and PC+/– mice at the younger age. Age dependence on survival after CLP has been previously studied,50,51 and revealed improved survival in mature mice as compared to young mice. Nonetheless, under our experimental conditions, survival was more favorable in the younger group than in the older group, in which PC+/– mice displayed a higher mortality rate than WT mice. The explanation for these observations may be a function of the increase in the risk of thrombosis with age, because of higher expression of procoagulant proteins such as fibrinogen, FVIII, and FIX,52 among other factors. Moreover, individuals with a heterozygous PC deficiency have a higher rise in the propensity to a coagulopathy with age than healthy individuals,53-56 which could explain the higher susceptibility to sepsis of PC+/– mice in the older age group.
Hemodynamic parameters, namely central BPs and heart rates, were continuously monitored before CLP and during the 24 hours after surgery and closely paralleled the phenomena observed in human sepsis. At 6 hours after challenge, the mice entered septic shock, displaying clear signs of hypotension and bradycardia. Importantly, the drop in both parameters was remarkably more dramatic in PC+/– mice as compared to WT mice, likely the most significant factor in the higher mortality in PC+/– mice. Moreover, poor tissue irrigation and hypoxia because of hypotension explains the renal damage observed by the elevated plasma levels of BUN and creatinine. Heart or muscle damage, as indicated by the elevated plasma creatinine kinase values, was higher in PC+/– mice compared to WT mice and hepatic failure, noted by the increase in plasma aspartate aminotransferase levels, was equivalent in both WT and PC+/– 24 hours after CLP. Hence, a more severe septic shock state in PC+/– mice supports previous evidence that PC contributes to the regulation of BP.30,57
Importantly, although previous studies described the effect of aPC administration on reduction of hypotension independent of its anti-coagulant properties,30
the present study shows an aggravated hypotensive state in mice expressing 50% of normal PC levels. The convergence of the results of these different approaches strongly suggests a mechanistic role of aPC in the prevention of hypotension.
The elevated plasma levels of TNF- and IL-1ß observed throughout the inflammatory state6,58
likely contribute to the up-regulation of tissue factor (TF) that is observed in this report. Furthermore, endothelium damage because of inflammation can activate blood coagulation. The procoagulant state induces the generation of thrombin and thrombin signaling through its PARs on ECs and/or platelets,59
and triggers inducible nitric oxide synthase expression60,61
and NO release. This provides an additional mechanism for hypotension in septic shock. Moreover, this plausible pathway for hypotension in sepsis-induced DIC, and PC signaling in ECs through EPCR and PAR-1 or PAR-2,62
may provide another mechanism for a protection against hypotension. A co-localization of PAR-1 with EPCR can reduce the availability of PAR-1 for thrombin signaling and thus participate in protection against inflammation through PC signaling.63
This protection could be lost in a PC deficiency and the diminution of plasma aPC levels could in this manner be a factor in hypotension. The decrease in the hemodynamic parameters seen after CLP-induced sepsis, which is more dramatic in PC+/– mice compared to WT mice, supports this hypothesis, taking into account that the plasma PC levels in PC+/– mice are approximately one-half of the value in WT mice at the time of shock. Additionally, at later time points, PC+/– mice displayed a more remarkable decrease in plasma PC levels than WT mice (Figure 4a)
. Furthermore, PC gene expression in liver and kidney showed a more striking drop in PC+/– mice compared to WT mice (Figure 6a)
.
A statistically significant decrease in blood leukocytes was observed in both WT and PC+/– mice 24 hours after CLP, the effect being more prominent in the later case. However, the total number of leukocytes measured in PLF did not differ between both genotypes at the 24-hour time point, as observed at 6 hours after CLP, in which the leukocyte count was higher in PC+/– as compared to WT PLF. Additionally, there is no evidence of inflammatory cell infiltration in the tissues, as displayed by CD45 staining (not shown) and MPO functional and gene expression assays at the 24-hour time point. Importantly, cellular debris was observed in the cytological analysis of the PLF of both groups at 24 hours after CLP by Giemsa staining. Based on these results, it appears that more significant leukocyte death in the PLF of PC+/– mice may result in the compensation in the leukocyte count at the present time point. A protective role of PC against apoptosis shown recently,25 justifies our assumption.
Interpretation of the hemostasis analyses presented herein indicates mild signs of consumption of factors from both the extrinsic and intrinsic pathways of blood coagulation 24 hours after CLP, as evidenced by the prolongation of the PTs and APTTs, respectively, along with the decrease in FXI and FXII activities. Further, FXI and FXII mRNA levels in the liver remained invariant after the challenge, implying that the decrease in their respective activities is not transcriptionally regulated (not shown). Overall, these results show that the consumption of coagulation factors is similar in WT and PC+/– mice. On the other hand, there was an 2.5-fold elevation in fibrinogen levels after CLP. This phenomenon reflects the acute phase response of this protein and possibly counterbalances an effect of a consumptive coagulopathy. Moreover, the blood platelet count remained unchanged 24 hours after the challenge, at 500,000 platelets/ml blood in both WT and PC+/– mice. Hemostasis factor differences between WT and PC+/– mice were not evident and did not account for a severe DIC in a PC deficiency. These results are not in accord with the observed aggravation of DIC observed in PC+/– mice after LPS,9
and this likely reflects differences in the models used. Thus, it seems that the PC contribution to amelioration of sepsis does not necessarily depend on a protective role against DIC, although DIC enhances the progress of the disease, but more so as a modulator of hypotension and bradycardia, and on its anti-inflammatory properties.
From the preceding considerations, it is apparent that EPCR-assisted PC activation and signaling are critical for the host defense against septic shock. Expression of EPCR in cell cultures decreased after treatment with mediators of the inflammatory response, eg, TNF-,64
which is in agreement with the impairment in PC activation seen in septic shock patients. However, up-regulation of EPCR gene expression, and increase in the release of a soluble form of EPCR by shedding from the endothelium, were observed in rats as a result of an endotoxin challenge.65
The same study showed that endotoxin-stimulated EPCR up-regulation was inhibited by hirudin, a thrombin inhibitor. Further incubation of rat ECs with thrombin or a PAR-1 agonist resulted in an increase of EPCR expression. These results imply that EPCR expression after endotoxin treatment occurs through PAR-1-dependent thrombin signaling. An increase in EPCR mRNA expression is observed 24 hours after CLP challenge predominantly in lung and kidney. Additionally, the increase in EPCR mRNA expression after CLP-induced sepsis was higher in PC+/– mice than in WT mice, which is supported by the hypothesis that thrombin levels are more elevated because of the PC deficiency. Thus, there may be a higher level of sEPCR in PC+/– mice compared to WT mice, contributing to impairment of the anti-inflammatory and anti-coagulant activities of PC.
Expression of Tm decreases on vascular endothelium in vitro and in vivo66,67
after treatment with endotoxin via TNF- up-regulation and signaling. Increased soluble Tm (sTm) values were observed in septic DIC models,68
which supports our observations. Interestingly, it was shown that Tm expression after endotoxin challenge was increased in macrophages but decreased in ECs.69
These observations suggest the importance of this receptor during sepsis in the regulation of coagulation through PC activation in different organs. Tm may serve anti-coagulant or anti-inflammatory purposes elsewhere, as expressed by other cell types. The significant increase in TM mRNA observed in the liver is not surprising, considering that this organ is rich in Kupffer cells, the hepatic macrophages. Further, transcription of TNF- or other proinflammatory cytokines that are known to down-regulate Tm expression, was not as elevated in liver as compared with the other organs and remained unchanged 24 hours after CLP.
It is believed that transcription factors NF-B and AP-1 intervene in the expression of TF after septic or traumatic shock70,71
in the vital organs via cytokine-mediated induction. Up-regulation of TF was observed in lung, kidney, liver, and vascular ECs in several inflammation and sepsis studies using different animal models,72,73
which is in agreement with our results in kidney and lung. Nonetheless, in the liver, we found a decrease in TF mRNA. The liver is the last organ to receive venous blood and, therefore, highly affected during the hypotensive shock caused by sepsis, as shown by the elevated plasma aspartate aminotransferase. It appears that in this model of sepsis, the liver was protected from a coagulopathy through lower cytokine expression, compared to other organs, leading to a decrease in TF and an increase in Tm expressions by the Kupffer cells. This observation is supported by the fact that fibrin(ogen) deposition was absent in this organ 24 hours after CLP.
Overall, the presented evidence confirms that individuals with a congenital PC deficiency have lesser chances of survival after sepsis because of a more severe hypotensive state, with organ malfunctioning. Remarkably, PC concentrations of 50% of normal are not sufficient to provide protection against systemic inflammation and shock, showing the importance of this protein in vivo. These findings imply that PC therapy would be an excellent choice for the treatment of septic patients with a congenital PC deficiency, in agreement with previous observations.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by grants from the National Institutes of Health (HL19982 and HL73750 to F.J.C.) and the Kleiderer-Pezold endowed professorship (to F.J.C.).
Accepted for publication June 21, 2004.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
, or phorbol ester. J Immunol 1994, 153:3664-3672[Abstract]
). The Tukey-Kramer test showed a significant difference between the 24-hour PC+/– group and the other groups (
