INTRODUCTION
The host response in sepsis includes the activation of the coagulation and complement systems, and the release of inflammatory mediators (1). Central to inflammation is also the presentation of chemokines on the luminal side of the endothelium and the adhesion and extravasation of leukocytes through the blood vessel wall. This process is partly mediated by membrane-bound proteoglycans (PGs) on leukocytes and endothelial cells and is accompanied by increased vascular permeability (2).
The PGs are composed of a core protein with covalently attached glycosaminoglycans (GAGs). The GAGs are unbranched chains of polysaccharides and are in most cases sulfated at various positions, which gives the molecules a highly negative charge and allows subdivision of the different molecules into classes such as chondroitin sulfate (CS) and heparan sulfate (3). Because of the variability of the position of the sulfate groups along the polysaccharide chain, GAGs display considerable sequence heterogeneity. This structural diversity allows GAGs to interact and modify the actions of various molecules, including cell-adhesion molecules, growth factors, cytokines, chemokines, proteases, and antimicrobial peptides (AMPs) (3-6).
On the luminal side of the endothelium, GAGs constitute a negatively charged network, the glycocalyx. The glycocalyx creates a physical and an electrostatic barrier that maintains the fluid balance between blood and tissue (7), and the loss of the glycocalyx has been associated with subendothelial edema in vivo (8).
Experimental data suggest that endothelial GAGs are involved in inflammation. Endothelial cells in cell culture release GAGs when stimulated by proinflammatory substances (9). These results have recently been confirmed in vivo (10). In addition, postcapillary venules in the rat shed the PG syndecan-1 upon stimulation with chemoattractant N-formylmethionyl-leucyl-phenylalanine ex vivo (11), and syndecan-1 has been demonstrated to contribute to mortality of burned and Pseudomonas aeruginosa-infected mice (12). A recent study suggests that GAGs are involved in the development of endotoxin-induced acute lung injury in the rat (13).
In the present study, we wanted to investigate if circulating GAG levels are altered in patients with septic shock, reflecting shedding of PGs. Indeed, we found that the GAG as well as the syndecan-1 levels are increased in the patients, and that the GAG level correlates with mortality. We also found that the GAG at relevant concentrations inhibits the endogenous antibacterial activity of plasma in vitro.
MATERIALS AND METHODS
Patients
All consecutive patients admitted to the intensive care unit (ICU) of Lund University Hospital, Sweden, because of sepsis between February 8, 2005, and April 4, 2006, and who fulfilled the criteria for septic shock (14) were included. The local research ethics committee of this institution approved the project, and informed consent was obtained from the patients or the next of kin of the unconscious patients. Severity of the organ dysfunction was defined using the Sequential Organ Failure Assessment (SOFA) score on the basis of measurements in the first 24 h of admission (15, 16). Patients who had received heparin/heparinlike drugs were excluded. The mortality rate within 90 days after inclusion was registered. The controls consisted of patients scheduled for surgery of an intracranial tumor. Informed consent was obtained. None of the control patients had ongoing steroid treatment or had been given heparin/heparinlike drugs.
Sample collection
Within 24 h after admission to the ICU, blood samples were drawn from an already existing arterial catheter and were collected in EDTA-treated vacuettes. In 9 patients, additional blood samples were obtained 4 days later. Arterial blood samples from the control patients were similarly drawn before induction of anesthesia. The samples were immediately centrifuged for 10 min at 800 g in room temperature. The plasma supernatant was removed and stored at −70°C until analysis.
Measurements of procalcitonin and C-reactive protein
Procalcitonin was measured in the plasma samples using an immunoluminometric assay kit from BRAHMS Aktiengesellshaft (Henningsdorf, Germany, Product no. 109.050; sensitivity, 0.01 ng/mL). Concentrations of C-reactive protein (CRP) in venous blood were measured by latex-enhanced immunoturbidimetry (Roche Diagnostics).
GAG assay
Isolation and detection of sulfated GAGs from plasma was performed using a kit from Euro-Diagnostica (Malmö, Sweden) and as previously described (17). Briefly, 10 μL of plasma sample or standard (CS-A at 1, 2.5, 5, 10, or 25 μg/mL) were added in duplicate to 20 μL of an acidulous buffer and gently agitated for 15 min at room temperature. Two hundred microliters of Alcian blue solution was added, and the mixture was left to precipitate for 1 h. A polyvinylidene fluoride membrane (Amersham) was prewet in ethanol, blocked for 1 h in 1% (vol/vol) Triton X-100, and assembled in a 48-well slot blot apparatus (Bio-Rad). Two hundred microliters prewash-buffer was added to each well, and 100 μL was evacuated. The samples were transferred to the slot blot apparatus and passed through the membrane by vacuum followed by two washes with 300 μL of 50% (vol/vol) ethanol in 0.05 M MgCl2. The membrane was then removed and dried. The membrane was mounted between two plastic sheets and scanned in reflectance-color mode in an Afga scanner. The red channel of the output RGB-matrix was used for the calculations. A least-squares standard curve was applied to the standard. Minimum assay sensitivity was 1 μg/mL GAG, and the blots saturated at 25 μg/mL.
Syndecan-1 enzyme-linked immunosorbent assay
A human syndecan-1 enzyme-linked immunosorbent assay (Diaclone, Täby, Sweden) targeting the soluble syndecan-1 molecule was performed according to the manufacturer's instructions. Briefly, 100 μL of samples or standard and 50 μL of diluted biotinylated antigen were pipetted in duplicate to the precoated wells and were incubated for 1 h. After three washes, 100 μL streptavidin horseradish peroxidase conjugate was added to each well, and was incubated for 30 minutes. After another three washes, 100 μL of substrate solution was added, and the color was allowed to develop for 13 min. The reaction was stopped by adding 100 μL of 1 M H2SO4, and the absorbance was measured at 450 nm. The assay was linear from 8 to 256 ng/mL, and samples above this concentration were diluted.
Radial diffusion assay
To assess the ability of GAGs to interfere with the antibacterial activity of human plasma and specific AMPs, a radial diffusion assay (RDA) was used (18). Blood from the antecubital vein of six healthy donors was drawn into plastic tubes containing EDTA. Plasma was collected after centrifugation for 10 min at 600 g and was stored at −70°C until use. Escherichia coli (strain 37.4) isolates were grown overnight at 37°C in 10 mL of 3% (wt/vol) tryptic soy broth (TSB). To obtain mid logarithmic-phase organisms, 200 μL of this culture was subcultured in 10 mL fresh TSB and grown for an additional 2 h at 37°C to an optical density at 620 nm of approximately 0.4. The bacteria were centrifuged at 900 g for 10 min and washed once in 10 mL Tris buffer (10 mM, pH 7.4), and diluted 100 times in 10 mM Tris buffer (pH 7.4). One percent (wt/vol) of low-electroendosmosis-type agarose (Sigma-Aldrich) in 10 mM Tris buffer (pH 7.4) containing 0.02% (vol/vol) Tween 20 (Sigma-Aldrich) and 0.03% (wt/vol) TSB was brought to the boiling point, cooled to 50°C, and then mixed with bacterial suspension (giving approximately 4 × 106 colony-forming units) and poured into a 20-cm Petri dish. A series of wells (diameter, 4 mm) were punched in the plate after the agarose had solidified. Six microliters of plasma samples or 5 μM LL-37 (Innovagen, Lund, Sweden) or 5 μM bactericidal permeability-increasing protein ([BPI] Wieslab, Lund, Sweden) with or without 0.1, 1, 10, or 100 μg/mL CS-A, CS-B, CS-C, or heparin (all from Sigma-Aldrich) with or without 25 μg/mL diethylaminoethyl (DEAE)-Dextran (Sigma-Aldrich) was applied to the wells, and the plates were incubated for 3 h at 37°C. An overlay agar composed of 6% TSB and 1% (vol/wt) of low-electroendosmosis-type agarose was then poured over, and the plates were incubated upside down for 18 h in 37°C to visualize growth of bacterial colonies. Antibacterial activity was indicated by a clear zone corresponding to the lack of bacterial growth around the wells. The diameter of the clear zone surrounding the wells was measured with a metric scale scribed in 0.1-mm increments.
Statistical analysis
The differences in the plasma level of GAG or syndecan-1 between patients and controls and between survivors and nonsurvivors were assessed with Wilcoxon rank sum test, and the data are expressed as median followed by 25th and 75th percentiles in square brackets when not otherwise indicated. The difference between levels at inclusion compared with 4 days later was assessed with Wilcoxon signed rank test. Correlations to Sequential Organ Failure Assessment (SOFA) score were assessed with Spearman rank test. The differences in antibacterial activity were evaluated with one-way repeated measurement ANOVA followed by post hoc testing using the Bonferroni correction, and the data are expressed as means ± SEM. To compare the recovery of antibacterial activity in the presence of DEAE-Dextran, a paired Student t test was used. The number of patients or independent experiments is indicated by "n." Differences were considered statistically significant when P < 0.05. All data were processed in Matlab 7.0.1, and figures were made with Adobe Illustrator CS.
RESULTS
Patient characteristics
The median age of the septic shock patients was 65 years (range, 28-87 years; n = 18), and there were 6 men and 12 women. Survival of the septic shock patients 90 days postsubmission to the ICU was 55% (10/18 patients). The median cardiovascular SOFA score of the patients was 3 (range, 1-4), and the median total SOFA score was 14 (range, 5-18). The median CRP level was 224 ng/mL (range, 54-542 ng/mL), and the median procalcitonin level was 39 ng/mL (range, 0.5-445 ng/mL). Blood culture was positive for E. coli in four patients, Streptococcus species in three patients, Staphylococcus species in two patients, and Enterococcus species in one patient. In eight of the patients, the blood culture was negative.
The median age of the controls was 64 years (range, 23-85 years; n = 18), and there were 8 men and 10 women.
Plasma GAG and syndecan-1 levels
The median plasma GAG level of septic shock patients (2.7 μg/mL [range, 1.9-4.8 μg/mL], n = 18) was significantly higher than of the controls (1.8 μg/mL [range, 1.7-2.0 μg/mL], n = 18; P < 0.01; Fig. 1A). The GAG level of the septic shock patients 4 days postsubmission had not changed significantly compared with the level at inclusion (2.4 μg/mL [range, 2.0-3.6 μg/mL] vs. 3.2 μg/mL [range, 2.3-4.2 μg/mL], n = 9). In the septic shock group, the median GAG level was higher in the nonsurvivors (4.6 μg/mL [range, 3.1-8.6 μg/mL], n = 8) than in the survivors (2.0 μg/mL [range, 1.6-2.6 μg/mL], n = 10; P < 0.01; Fig. 1B). The results were statistically significant also when the two patients with the GAG levels above 15 μg/mL were excluded, neither of whom survived. The median syndecan-1 plasma level of the septic shock patients (246 ng/mL [range, 180-496 ng/mL], n = 18) was significantly higher than of controls (26 ng/mL [range, 23-31 ng/mL], n = 18; P < 0.001; Fig. 2). In the septic shock group, the plasma level of syndecan-1 did not correlate with mortality (not shown) and was not statistically significantly different at 4 days postsubmission compared with at inclusion (299 ng/mL [range, 104-563 ng/mL] vs. 240 ng/mL [range, 181-612 ng/mL], n = 9). The GAG levels of the septic shock patients correlated neither to the cardiovascular SOFA score nor to the total SOFA score (Fig. 3, A and B). The syndecan-1 levels correlated to the cardiovascular SOFA score (Fig. 3C). We could also detect a weak correlation to the total SOFA score (Fig. 3D). We could not detect any correlation between GAG and syndecan-1 levels, and neither GAG nor syndecan-1 levels correlated to the plasma levels of CRP or procalcitonin (not shown).
;)
Fig. 1: Plasma GAG level in septic shock patients and controls measured with an Alcian blue slot-binding assay. The median GAG level was significantly higher in sepsis patients (▿) compared with controls (○, A). Among the septic shock patients, the median GAG level was significantly higher in the patients who died within 90 days postsubmission (◃) compared with the survivors (▹, B). The median is indicated (horizontal line), and the symbols represent individual patients. Wilcoxon rank sum test.
Fig. 2: Plasma syndecan-1 level in septic shock patients (▿) and controls (○) measured with enzyme-linked immunosorbent assay. The levels were significantly higher in sepsis patients compared with controls. The median is indicated (horizontal line), and the symbols represent individual patients. Wilcoxon rank sum test.
Fig. 3: Correlation between GAG and cardiovascular SOFA score (A) and total SOFA score (B) as well as between syndecan-1 and cardiovascular SOFA score (C) and total SOFA score (D). The correlation coefficient (r) and the level of statistical significance are indicated. Spearman rank correlation test.
Effect of GAGs on the endogenous antibacterial activity of plasma
All GAGs tested were found to inhibit the antibacterial activity of plasma at a concentration of 10 μg/mL (Fig. 4), which corresponds to the upper range of the concentrations measured in the sepsis patients. The decreased antibacterial activity at 10 μg/mL was recovered in the presence of the polycation DEAE-Dextran to bind GAGs (Fig. 5). Glycosaminoglycans or DEAE-Dextran alone, at the concentrations used, did not affect bacterial growth (data not shown). We also found that at concentrations above 1 μg/mL, CS-B, CS-C, and heparin, but not CS-A, inhibited the antibacterial action of isolated antibacterial peptide human cathelicidin LL-37 in aqueous solution (Fig. 6A). In the same concentration range, all GAGs inhibited the antibacterial effect of BPI in aqueous solution (Fig. 6B).
Fig. 4: Antibacterial activity of human plasma against E. coli in the presence or absence of GAG, as assessed by an RDA. Chondroitin sulfate-A (•), CS-B (▴), CS-C (▪), and heparin (▾, 10 and 100 μg/mL) all significantly inhibited the endogenous antibacterial activity of plasma (*). The data were analyzed by one-way repeated measurement ANOVA followed by post hoc testing using the Bonferroni correction (P < 0.05). Values are means ± SEM (n = 6).
Fig. 5: Antibacterial activity against E. coli of human plasma containing GAGs (10 μg/mL) in the presence (filled bars) or absence (open bars) of DEAE-Dextran, as assessed by an RDA. All GAGs significantly inhibited the endogenous antibacterial activity of plasma (*). The antibacterial activity was recovered in the presence of DEAE-Dextran (25 μg/mL, †). The data were analyzed by Student paired t test (P < 0.05). Values are means ± SEM (n = 4).
Fig. 6: Inhibition of the antibacterial activity of the human cathelicidin LL-37 (A) and BPI (B) against E. coli by addition of GAGs, as assessed by an RDA. Chondroitin sulfate-B, CS-C, and heparin significantly inhibited the antibacterial activity of LL37, and all GAGs inhibited the antibacterial activity of BPI (*). The data were analyzed by one-way repeated measurement ANOVA followed by post hoc testing using the Bonferroni correction (P< 0.05). Values are means ± SEM (n=4).
DISCUSSION
The present results demonstrate that circulating GAGs are increased in septic shock patients, which indicate a deranged GAG turnover. Importantly, the GAG levels correlated to mortality. Considering the limited number of patients investigated, it is premature to determine whether GAG could be useful as a prognostic marker in patients with severe sepsis. Previously, endocan-1, an endothelial PG, has been shown to possess properties as a marker for sepsis (19). The findings are supported by the observation that the urine levels of GAG in patients with meningococcal sepsis are related to the severity of the disease (20).
The sources of the GAGs resulting in the increased levels in patients can, however, not be determined on the basis of the present results, but it seems plausible that they at least partly originate from shedding of the endothelial glycocalyx. This hypothesis is supported by our finding that the levels of syndecan-1, an endothelial PG, were 10- to 100-fold increased in the patients. The levels of syndecan-1 correlated to the cardiovascular SOFA score. It is well established that syndecan-1 participates in inflammation, and that shedding from the endothelium promotes leukocyte adherence (11, 21-23). In rat heart capillaries, disruption of glycocalyx is associated with subendothelial edema (8). It could be speculated that shedding of GAGs, including syndecan-1, plays a role in the dysregulation of the endothelial function, resulting in increased permeability in septic shock. Surprisingly, there was no correlation between syndecan-1 and GAG levels. Syndecan-1 is, however, a subgroup among GAGs, and the lack of correlation suggests multiple origins of the plasma GAGs. This is supported by the fact that the levels of GAG were about 10 times higher than the levels of syndecan-1. Sepsis involves disruption of the vascular barriers, and it cannot be excluded that this results in leakage of GAGs from connective tissue and the basement membrane into the circulation. In a recent study, Rehm and colleagues (24) found that GAG as well as syndecan-1 levels rose after ischemia and reperfusion during vascular surgery. This correlated to shedding of glycocalyx as detected by microscopy. Taken together, these and our results suggest that the shedding of endothelial GAGs could be a response to vascular stress from various causes.
We found that GAGs inhibit the antibacterial activity of plasma as well as isolated AMPs in vitro. This suggests that an increased plasma GAG level per se could impair the innate immunity by binding locally released AMPs as well as activated complement fragments as previously demonstrated (25). It seems reasonable to assume that the inhibitory effects exerted by the negatively charged GAGs, at least partly, are caused by electrostatic interactions with the positively charged endogenous AMPs such as LL-37 and BPI. Interestingly, the observation that the polycation DEAE-Dextran reversed the inhibitory effects on plasma antibacterial action suggests that therapies aiming to neutralize or remove excessive GAGs could be successful. Indeed, a recent report using DEAE-cassettes with the aim to remove LPS from the blood of sepsis patients showed promising results (26). Although the study was designed to bind LPS, it is plausible that GAGs, which bind tightly to DEAE groups in plasma (27), were also bound to the cassette. Hence, our results suggest that the use of similar scavengers could be generalized to any form of sepsis, including sepsis caused by gram-positive bacteria.
In conclusion, we have shown that GAG and syndecan-1 levels are increased in the blood of septic shock patients and that the GAG level correlates to mortality. We have also shown that GAG, at concentrations present in the sepsis patients with a poor prognosis, inhibits the plasma antimicrobial activity in vitro. These results motivate further investigations in larger clinical materials.
ACKNOWLEDGMENT
The authors thank Mina Davoudi for expert technical assistance.
REFERENCES
1. Aird WC: The role of the endothelium in severe sepsis and multiple organ dysfunction syndrome.
Blood 101:3765-3777, 2003.
2. Parish CR: The role of heparan sulphate in inflammation.
Nat Rev Immunol 6:633-643, 2006.
3. Kreuger J, Spillmann D, Li JP, Lindahl U: Interactions between heparan sulfate and proteins: the concept of specificity.
J Cell Biol 174:323-327, 2006.
4. Kjellen L, Lindahl U: Proteoglycans: structures and interactions.
Ann Rev Biochem 60:443-475, 1991.
5. Taylor KR, Gallo RL: Glycosaminoglycans and their proteoglycans: host-associated molecular patterns for initiation and modulation of inflammation.
FASEB J 20:9-22, 2006.
6. Baranska-Rybak W, Sonesson A, Nowicki R, Schmidtchen A: Glycosaminoglycans inhibit the antibacterial activity of LL-37 in biological fluids.
J Antimicrob Chemother 57:260-265, 2006.
7. Michel CC, Curry FE: Microvascular permeability.
Physiol Rev 79:703-761, 1999.
8. van den Berg BM, Vink H, Spaan JA: The endothelial glycocalyx protects against myocardial edema.
Circ Res 92:592-594, 2003.
9. Klein NJ, Shennan GI, Heyderman RS, Levin M: Alteration in glycosaminoglycan metabolism and surface charge on human umbilical vein endothelial cells induced by cytokines, endotoxin and neutrophils.
J Cell Sci 102(Pt 4):821-832, 1992.
10. Marechal X, Favory R, Joulin O, Montaigne D, Hassoun S, Decoster B, Zerimech F, Neviere R: Endothelial glycocalyx damage during endotoxemia coincides with microcirculatory dysfunction and vascular oxidative stress.
Shock 29:572-576, 2008.
11. Mulivor AW, Lipowsky HH: Inflammation- and ischemia-induced shedding of venular glycocalyx.
Am J Physiol Heart Circ Physiol 286:1672-1680, 2004.
12. Haynes A, Ruda F, Oliver J, Hamood AN, Griswold JA, Park PW, Rumbaugh KP: Syndecan 1 shedding contributes to
Pseudomonas aeruginosa sepsis.
Infect Immun 73:7914-7921, 2005.
13. Bashenko Y, Ilan N, Krausz MM, Vlodavsky I, Hirsh MI: Heparanase pretreatment attenuates endotoxin-induced acute lung injury in rats.
Shock 28:207-212, 2007.
14. American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference: Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis.
Crit Care Med 20:864-874, 1992.
15. Knaus WA, Draper EA, Wagner DP, Zimmerman JE: APACHE II: a severity of disease classification system.
Crit Care Med 13:818-829, 1985.
16. Moreno R, Vincent JL, Matos R, Mendonca A, Cantraine F, Thijs L, Takala J, Sprung C, Antonelli M, Bruining H, et al.: The use of maximum SOFA score to quantify organ dysfunction/failure in intensive care. Results of a prospective, multicentre study. Working Group on Sepsis related Problems of the ESICM.
Intensive Care Med 25:686-696, 1999.
17. Bjoernsson S: Quantitation of proteoglycans as glycosaminoglycans in biological fluids using an Alcian blue dot blot analysis.
Anal Biochem 256:229-237, 1998.
18. Lehrer RI, Rosenman M, Harwig SSSL, Jackson R, Eisenhauer P: Ultrasensitive assays for endogenous antimicrobial polypeptides.
J Immunol Methods 137:167-173, 1991.
19. Scherpereel A, Depontieu F, Grigoriu B, Cavestri B, Tsicopoulos A, Gentina T, Jourdain M, Pugin J, Tonnel A-B, Lassalle P: Endocan, a new endothelial marker in human sepsis.
Crit Care Med 34:532-537, 2006.
20. Oragui EE, S SN, Kyd P, Levin M: Increased excretion of urinary glycosaminoglycans in meningococcal septicemia and their relationship to proteinuria.
Crit Care Med 28:3002-3008, 2000.
21. Gotte M: Syndecans in inflammation.
FASEB J 17:575-591, 2003.
22. Gotte M, Joussen AM, Klein C, Andre P, Wagner DD, Hinkes MT, Kirchhof B, Adamis AP, Bernfield M: Role of syndecan-1 in leukocyte-endothelial interactions in the ocular vasculature.
Invest Ophthalmol Vis Sci 43:1135-1141, 2002.
23. Mulivor AW, Lipowsky HH: Role of glycocalyx in leukocyte-endothelial cell adhesion.
Am J Physiol Heart Circ Physiol 283:1282-1291, 2002.
24. Rehm M, Bruegger D, Christ F, Conzen P, Thiel M, Jacob M, Chappell D, Stoeckelhuber M, Welsch U, Reichart B, et al.: Shedding of the endothelial glycocalyx in patients undergoing vascular surgery with global and regional ischemia.
Circulation 116:1896-1906, 2007.
25. Nordahl EA, Rydengård V, Nyberg P, Nitsche DP, Mörgelin M, Malmsten M, Björck L, Schmidtchen A: Activation of the complement system generates antibacterial peptides.
Proc Natl Acad Sci U S A 101:16879-16884, 2004.
26. Bengsch S, Boos KS, Nagel D, Seidel D, Inthorn D: Extracorporeal plasma treatment for the removal of endotoxin in patients with sepsis: clinical results of a pilot study.
Shock 23:494-500, 2005.
27. Staprans I, Felts JM: Isolation and characterization of glycosaminoglycans in human plasma.
J Clin Invest 76:1984-1991, 1985.
