Endothelial barrier dysfunction preceding breakdown of microcirculatory flow is a key event in the pathogenesis of sepsis that significantly contributes to organ failure and death in patients (1). Despite great efforts to improve sepsis therapy the mortality rate of septic patients remains high. It has been increasingly recognized that therapeutic approaches to treat microvascular dysfunction in sepsis might help to prevent organ failure and thereby improve therapeutic outcome (1). However, there is still no specific therapy available yet to stabilize the endothelial barrier function in septic patients.
Previous studies have shown that the sphingolipid sphingosine-1-phosphate (S1P) is critically involved in the maintenance of the endothelial barrier (2–4). S1P binds to a subset of five G-protein coupled receptors (S1P1–5) leading to a large variety of cellular events such as cell proliferation, migration, and immune cell trafficking (5). Based on studies in single microvessels it has been demonstrated that a continuous release of S1P from erythrocytes and platelets is required to maintain the endothelial barrier under resting conditions (3, 6).
In a clinical study, Winkler et al. (7) have found significantly decreased S1P levels in close correlation to the disease severity of septic patients indicating a causal relationship between S1P serum levels and loss of microvascular barrier function in sepsis. Supporting these clinical observations different in vivo models have shown that augmented levels of S1P effectively block inflammation-induced increase of microvascular permeability (4, 8–10). The barrier-protective effects of S1P are mediated through binding at the S1P receptor-1 leading to activation of Rho family GTPases, cytoskeletal reorganization, adherens- and tight junction assembly (11–15). However, clinical trials applying the unspecific S1P receptor agonist FTY720 to patients with multiple sclerosis have revealed severe side effects such as bradyarrhythmia, macular edema, and hypertension (16). The development of SEW2871 as a specific agonist for S1P1 presents a promising therapeutic approach to use selectively the barrier enhancing properties of S1P signaling.
Initial studies have shown that SEW2871 improves endothelial barrier function and thereby improves microcirculatory flow (17, 18) suggesting that SEW2871 might be a specific therapeutic agent for microvascular barrier dysfunction and breakdown of microcirculation in sepsis. However, SEW2871 has not been tested yet in a clinically relevant sepsis model that encompasses the complexity of bacterial sepsis.
Therefore, the present study aimed to analyze the therapeutic and potential side effects in long-term treatment of SEW2871 in a clinically relevant rodent model of bacterial abdominal sepsis that enables the evaluation of microvascular endothelial barrier functions, organ failure, and macrohemodynamic parameters (19). In this polymicrobial sepsis model the intravenous application of SEW2871 did not attenuate the breakdown of microvascular barrier functions but caused severe cardiac side effects resulting in increased lethality. The present study provides further insights in the complexity of sepsis treatment with potentially suitable therapeutic agents.
MATERIALS AND METHODS
As indicated below sphingosine-1-phosphate (Cayman chemicals, CatNo 62570, Germany) was used at different concentrations (100 nM, 1 μM, 5 μM) alone or in combination with the specific sphingosine-1-phosphate receptor-1 agonist SEW2871 (Cayman chemicals, CatNo 10006440, Germany) at 100 nM, 1 μM, or 5 μM. TNF-α at 100 ng/mL (Biomol, Hamburg, Germany) was used to induce endothelial barrier disruption (20). All reagents were dissolved and stored according to the manufacturers’ recommendations. SEW2871 was dissolved in ethanol and then further diluted with isotone saline (NaCl 0.9%) according the recommendation found in the datasheet (Cayman Chemical, SEW2871, item no. 10006440). DMSO was not used to dissolve SEW 2871 because anti-inflammatory characteristics of DMSO are known and thus can influence the results. The final stock was filtered through a 0.22 μm filter before application. The protocol of dissolving and dilution was exactly the same for both CASP+SEW2871 and Sham+SEW2871 groups (see below).
We used human dermal microvascular endothelial cells (HDMEC) (PromoCell, Heidelberg, Germany) which were characterized to be a suitable model to study microvascular endothelial barrier regulation in vitro(20). Cells were used from passages two to eight and grown in endothelial growth medium containing supplement mix and were passaged using Detach kit-30 (both PromoCell).
Immunostaining has been described in detail previously (21). In brief, HDMECs were grown to confluence on cover slips for 7 to 10 days. After incubation with mediators as indicated below, fixation with 2% formaldehyde and treatment with 0.1% Triton X-100, we used a polyclonal goat antihuman VE-cadherin antibody (R&D Systems, Wiesbaden, Germany) as primary antibody (1:1,000 in PBS each). As secondary antibodies, we used a Cy3-labeled donkey anti-goat IgG (Dianova, Hamburg, Germany, diluted 1:600 in PBS). Cover slips were mounted on glass slides with 60% glycerol in PBS, containing 1.5% n-propyl gallate (Serva, Heidelberg, Germany). All experiments were performed in triplicates for each condition.
Measurement of transendothelial resistance
ECIS 1600R (Applied BioPhysics Inc, Troy, NY) was used to measure the transendothelial resistance (TER) of HDMEC monolayers to assess endothelial barrier integrity as described previously (21).
After animal care committee approval (laboratory animal care and use committee of the district of Unterfranken, Germany), experiments were performed on male Sprague–Dawley rats (n = 31; aged 8–10 weeks; 300–350 g BW; Janvier (Le Genest St. Isle, France)). Animals were kept under conditions that conformed to the National Institutes of Health “Guide for the Care and Use of Laboratory Animals,” approved by the Government of Unterfranken as well as for those of the US National Institutes of Health. All rats were maintained on a standard diet and water ad libitum at 12 h day and night cycles. Animals were not fasted prior to the procedure.
For the in vivo experiments we utilized a modified model of polymicrobial sepsis that combines the previously described colon ascendens stent peritonitis model (CASP) with specific modifications enabling simultaneous macrohemodynamic and mircohemodynamic monitoring. This sophisticated in vivo model was described in detail previously (19). In brief, after anesthetization of the animals using isoflurane (2.0 vol.%; Forene, Abbott, Wiesbaden, Germany)/nitrous oxide inhalation (Air Liquide, Düsseldorf, Germany) (FiO2 0.28) the right jugular vein and left carotid artery were cannulated for intravenous medication and continuous blood pressure and heart rate measurements (Hewlett Packard Model 88S, Hamburg, Germany), respectively. Furthermore, the collection of arterial blood samples for repeated blood gas analyses was feasible. Anesthesized animals underwent median laparotomy, the caecum was identified and gently exteriorized to explore the proximal colon. 20 mm aboral from the ileocaecal valve, the wall of the ascending colon was pierced with a suture at the antimesenteric side. After the perforation of the colonic wall with a 14 G needle a modified 10 ch plastic tube (tip of a suction catheter, type “Ideal,” B. Braun Melsung, Germany) was inserted and fixed with the prepared suture at the colonic wall. This technique ensured to maintain the perforation of the colon during the duration of the experiments. Stool was milked from the caecum toward the colonic stent using two cotton swabs until stool appeared on the top of the stent. Then the gut was reinserted into the peritoneal cavity and the stent was flushed with 2 mL sodium chloride 0.9% (Fresenius Kabi, Bad Homburg, Germany) to distribute the feces into peritoneal cavern. Afterward the peritoneum and the skin were closed with continuous two layer sutures. Sham-operated animals only received median laparotomy without perforation of the colon.
After the operation procedure and a reconstitution time of 30 min hemodynamic parameters were measured and samples for the first blood gas analyses were taken using ABL505 blood gas analyser (Radiometer, Copenhagen, Denmark). The volume of blood that was taken for blood analysis was substituted by an equal volume of 0.9% sodium chloride (Fresenius Kabi, Bad Homburg, Germany). Analgesia was maintained by continuous intravenous application of fentanyl (2 μg/100 g BW/h fentanyl (Fagron, Barsbüttel, Germany)). All animals had free access to water and food. At the end of the experimental procedures all animals were euthanized by anesthesia overdose.
Randomization of animals
Twelve hours after sepsis induction animals were randomized in the following groups: Sham-operated animals receiving 0.9% sodium chloride 5 mL/12 h which served as the control group (Sham+NaCl, n = 8), Sham-operated animals receiving SEW2871 0.5 mg/kg BW/12 h (total volume 5 mL/12 h) (Sham+SEW2871, n = 5), CASP-operated animals either receiving 0.9% sodium chloride 5 mL/12 h (CASP+NaCl, n = 5) or SEW2871 0.5 mg/kg BW/12 h (total volume 5 mL/12 h) (CASP+SEW2871, n = 13). Before the continuous administration of sodium chloride or SEW2871 started, the NaCl-treated animals received 1 mL 0.9% sodium chloride as a single shot (bolus) and the SEW2871-treated animals received SEW2871 0.5 mg/kg BW diluted in 1 mL 0.9% sodium chloride, respectively.
Animal preparation for monitoring of microvascular leakage
Twenty-four hours after the operation, animals were re-anesthesized as described above and a tracheotomy was performed for mechanical ventilation using a rodent ventilator (Type: 7025, Hugo Sachs Elektronic KG, March-Hugstetten, Germany) with a fixed ventilation regime (FiO2 0.28 and respiratory rate of 75 per minute). With the start of controlled ventilation the anesthesia regime was changed utilizing midazolam (Midazolam-ratiopharm, Ratiopharm, Ulm, Germany) 0.7 mg/100 g BW/h, Fentanyl 7 μg/100 g BW/h and for a sufficient anesthetic depth 0.7 Vol% isoflurane (without N2O) was adapted. Animals’ body temperature was always kept at 37°C using a heating plate.
In the next step relaparotomy was performed, the mesentery was gently exteriorized and placed on a customized device as has been described before (19). Using a modified inverted Zeiss microscope (Axiovert 200, Carl Zeiss, Göttingen, Germany) equipped with different lenses (Achroplan × 10 NA 0.25/ × 20 NA0.4/ × 40 NA 0.6) enabled continuous measurements of microvascular leakage. Images were captured using digital camera ColorSnap CF driven by Metamorph analysis software and digitally recorded for off-line analysis as described below. In the time course of experiments the upper surface of the mesentery was continuously moisturized with 37.5°C crystalloid solution (Sterofundin, B.Braun Melsungen AG, Melsungen, Germany).
Macrohemodynamic monitoring and blood gas analyses
Mean arterial pressure (MAP) and heart rate were continuously measured throughout the experiments and documented after the first operation/sepsis induction (baseline), after 12, 24 h and before the end of the experiment (26 h). Simultaneously, blood gas samples were withdrawn and at the end of the experimental procedures (after intravital microscopy) whole blood samples were taken for laboratory investigation. The blood volume taken for blood analyses was substituted by an equal volume of 0.9% sodium chloride (Fresenius Kabi, Bad Homburg, Germany). Furthermore, fluid resuscitation was conducted with 0.5 mL to 1.0 mL 0.9% sodium chloride (Fresenius Kabi, Bad Homburg, Germany) depending on the MAP with a threshold of 60 mm Hg.
Intravital measurement of capillary endothelial barrier properties
To assess changes of microvascular permeability during the experimental procedures fluorescent digital images were taken after single intravenous injection of fluorescein isothiocyanate (FITC) conjugated albumin 5 mg/100 g BW (Sigma, Deisenhofen, Germany) (19). Fluorescent images were taken using a 100 W mercury lamp and a filter set consisting of a 450 nm to 490 nm excitation and a 520 nm emission filter within the inverted microscope (Zeiss microscope, Axiovert 200, Carl Zeiss, Göttingen, Germany). Microvascular permeability was then estimated by determining the extravasation of FITC-albumin by measurements of integrated optical intensity as described previously (22). Using Image J software six to eight randomly selected intravascular and interstitial areas near the mesenteric venules were selected for the measurements of optical intensity intra- and extravascular. The analysis was carried out by a blinded observer.
Evaluation of blood samples and measurements of S1P serum levels
After intravital microscopy blood samples were taken for the determination of creatinine levels using routine laboratory methods. The serum concentration of S1P was evaluated at the end of the experimental procedures using S1P ELISA kit (Antibody Research, 682861). According to the manufacturer's recommendations wells were precoated with Human S1P antibody and an equal amount of rat serum was used. After incubation and washing S1P detection antibody labeled with biotin was added. Again after incubation and washing Strept-Avidin-HRP was added to form immune complex. Finally, chromogenic HRP enzyme substrate and stop solution was added. The intensity of developed color is proportional to the concentration of S1P present in the sample and read at 450 nm using a plate reader (ELISA Reader, Tecan, Maennedorf, Switzerland).
Histo-pathological analysis of lungs
Lungs were removed after experimental procedures for histo-pathological studies. The tissues were fixed in formaldehyde 3.5% (Otto Fischer, München, Germany) for more than 24 h. Tissues were then embedded in paraffin and subsequently sections were stained with H&E for the analyses of morphological alterations within the tissues. Sections were photographed using a Keyence BZ-9000 microscope. BZ-II analyser software from Keyence (KEYENCE Corporation, Japan) was used to measure thickness of alveolar septa. In each animal four sections were analyzed by the determination of the thickness of three randomly chosen alveolar septa. The analysis was performed by a blinded observer (19).
Values throughout are expressed as mean ± SEM. Statistical analyses were performed using GraphPad Prism 6.0d. For parametric parameters, possible differences were assessed with t test or ANOVA followed by Tukey test. For non-parametric data, Kruskal–Wallis following Dunn post-test or Mann–Whitney U test were used for significant differences. The tests used for respective experiments are indicated in the figure legends. Statistical significance is assumed for P < 0.05.
Sphingosine-1-phosphate and its agonist SEW2871 cause dose-dependent endothelial barrier breakdown
To confirm and to further analyze the previously described effects of sphingosine-1-phosphate and its specific S1P receptor-1 agonist SEW2871 in terms of endothelial barrier properties, we measured the TER within 24 h after incubation of confluent endothelial monolayers (HDMEC) with S1P and SEW2871 at different concentrations. As shown in Figure 1A, the incubation with S1P in different concentrations caused an endothelial barrier breakdown revealed by significantly decreased TER in comparison to control conditions (Fig. 1A). In parallel, it was observed that the incubation of SEW2871 also leads to a dose-dependent alteration of endothelial barrier (Fig. 1B). Only the lowest concentration of 100 nM SEW2871 did not show compromised barrier properties compared with control.
In the next step, we tested whether S1P and SEW2871 were effective to attenuate inflammation-induced endothelial barrier breakdown. Based on the results above, only low concentrations (i.e., 100 nM) of both S1P and SEW2871 were used in the following experiments. To induce an inflammation-induced endothelial barrier breakdown TNF-α at 100 ng/mL known as potent pro-inflammatory cytokine was used as described previously (20). As shown in Figure 1C, 4 h after incubation with TNF-α a significantly decreased TER was detectable (Fig. 1C). The specific S1P receptor-1 agonist SEW2871 100 nM partially attenuated the TNF-α-induced loss of endothelial barrier function. Interestingly, S1P did not show any barrier stabilizing effects under inflammatory conditions but rather caused a further impairment of endothelial barrier function (Fig. 1C).
Immunofluorescence staining of VE-cadherin known as a pivotal junctional protein in the maintenance of endothelial barrier function underlined the functional data presented above. Under control conditions the staining of VE-cadherin showed a linear staining pattern at the cell borders (Fig. 2A). After incubation with S1P and SEW2871 an irregular staining pattern of VE-cadherin was determined that became more pronounced with increased concentrations of S1P and SEW2871 (Fig. 2A). After incubation with TNF-α a loss of VE-cadherin staining at the cell borders became evident next to an abundant irregular staining pattern (Fig. 2B). In contrast to S1P SEW2871 partially attenuated the TNF-α effects as already shown by functional analysis, i.e., TER measurements (Fig. 2B).
Taken together, these experiments demonstrated that the incubation of S1P and SEW2871 cause an endothelial barrier breakdown in a dose-dependent manner as revealed by functional and immunostaining data in vitro. Importantly, loos of endothelial barrier properties under inflammatory conditions were only attenuated by application of SEW2871 in a low concentration.
Severe side effects after intravenous application of SEW2871 in vivo in animals with sepsis
Based on the in vitro results we used SEW2871 in a low dose to test therapeutic and potential side effects in a polymicrobial sepsis model. Twelve hours after sepsis induction the therapeutic regime started with a bolus application of either SEW2871 0.5 mg/kg BW dissolved in 1 mL sodium chloride 0.9%/ethanol as described above or only 1 mL sodium chloride 0.9% followed by continuous application of SEW2871 or sodium chloride, respectively. The application of SEW2871 did not show any side effects in sham-operated animals (n = 5; Sham+SEW2871). However, under septic conditions severe cardiac side effects were determined such as bradyarrhythmia and asystolic arrest leading to death (n = 8 out of 13; CASP+SEW2871) (Fig. 3, A and C). The deceased animals were excluded from the following analysis. In comparison all CASP-operated animals with NaCl-treatment survived and, thus, no animal in this group was excluded (n = 5, CASP and NaCl).
Macrohemodynamic effects of long-term treatment with SEW2871
Macro-hemodynamic values including MAP and heart rate were measured at four different time points as already mentioned above. Within the first 24 h no significant differences in MAP were detected between the different groups (Fig. 3B). Between 24 and 26 h a drop in MAP was observed in all experimental groups. This effect can be explained by systemic reaction due to repeated surgical procedures (tracheotomy, relaparotomy), duration of reanesthesia with mechanical ventilation and intravital microscopy with exteriorized gut (mesenteric traction syndrome) to measure endothelial barrier function. Furthermore, it has to be taken in consideration that a fluid resuscitation has been conducted starting with MAP below 60 mm Hg. As shown in Figure 4 CASP-operated animals (CASP+NaCl: 9.6 ± 1.5 mL; CASP+SEW2871: 5.6 ± 0.93 mL) needed more additional volume (significant for CASP+NaCl) compared with Sham-operated animals (Sham+NaCl: 1.63 ± 0.42 mL; Sham+SEW2871: 2.0 ± 0.84 mL).
CASP-operated animals with SEW2871 (CASP+SEW2871: 398 ± 16/min) treatment showed a decreased heart rate with a significant difference at time point 24 h (Sham+NaCl: 477 ± 12.5/min; Sham+SEW2871: 474 ± 14/min; CASP+NaCl: 430 ± 23/min) (Fig. 3D). At the end of the experiment the differences were still detectable but not on a significant level anymore.
In conclusion, macrohemodynamic measurements throughout the entire time course of the experiment revealed that the long-term treatment with SEW2871 does not have any obvious side effects under non-septic conditions. However, the intravenous application of SEW2871 under septic conditions caused a significantly increased lethality due to severe cardiac side effects. Even animals not showing acute cardiac side effects in the beginning of the treatment were characterized by a decreased heart rate after long-term treatment.
SEW2871 does not improve endothelial barrier properties under septic conditions
To examine a potential effect of SEW2871 in terms of a stabilization of sepsis-induced endothelial barrier breakdown, the efflux of intravenously applied FITC-albumin 4 kDa was measured in mesenteric venules. In contrast to Sham-operated animals both CASP-groups showed an obvious efflux of FITC-albumin from intravasal to extravasal (Fig. 5A) indicating the presence of microvascular leakage (capillary leakage syndrome) as a result of an impaired endothelial barrier function. The quantification of FITC-albumin extravasation within each group confirmed the visual impression of significantly increased efflux of FITC-albumin in CASP-operated animals compared with Sham-operated groups (Fig. 5B). SEW2871 was not capable of improving mesenteric microvascular barrier properties under septic conditions in vivo.
In addition to the examination of the mesenteric microcirculation lung tissue sections were analyzed. CASP-operated animals showed a pulmonary endothelial dysfunction as revealed by significantly increased thickness of alveolar septa and infiltration of neutrophils (Fig. 6, A and B). The application of SEW2871 did not have any protective effect in terms of severity of lung edema under inflammatory conditions.
Effects of SEW2871 on metabolic homeostasis and regulation
Since endothelial dysfunction resulting in capillary leakage syndrome is accompanied with systemic metabolic changes blood samples were taken for blood gas analyses after the first operation, after 24 h, and at the end of the experiment (Table 1). After the first operation pH values of all four experimental groups were in the normal range. In the further course of the experiment pH values shifted toward basic pH values with an increased base excess which was especially pronounced in the Sham-operated animals. Furthermore, at 24 h CASP-operated animals showed a slightly increased respiratory rate (cut-off: 115/min) suggesting a compensation for the increasing lactate levels over the time. Interestingly, only in the CASP+SEW2871 group the respiratory parameters under spontaneous breathing conditions confirmed the severe histo-pathological changes seen in lung sections of these animals. After re-anesthesia and relaparotomy pH-parameters have dropped to acidic values with further increased lactate levels that were most pronounced in CASP-operated animals. Due to the mechanical ventilation regime at this time point no differences in respiratory values were obvious anymore.
Due to the fact that not only the lung with its respiratory exchange but also the kidney plays a pivotal role in the regulation of metabolic homeostasis kidney function was evaluated at the end of the experiment (Fig. 7). The differences in creatinine levels were not significantly different but slightly increased in both CASP groups (Sham+NaCl: 0.32 ± 0.04 mgL/dL; Sham+SEW2871: 0.37 ± 0.05 mg/dL; CASP+NaCl: 0.40 ± 0.09 mg/dL; CASP+SEW2871: 0.48 ± 0.0.07 mg/dL) taking into account that CASP-treated animals received a significantly higher amount of additional volume. This fluid resuscitation therapy has to be considered a kidney protective therapy. The analysis of potassium levels determined no differences between the experimental groups (Sham+NaCl: 5.0 ± 0.20 mol/L; Sham+SEW2871: 4.30 ± 0.28 mol/L; CASP+NaCl: 4.83 ± 0.43 mol/L; CASP+SEW2871: 4.60 ± 0.54 mol/L).
Furthermore, liver enzymes were evaluated since sepsis can also induce a hepatocellular injury. CASP-operated animals show elevated values compared with sham animals but without significance (aspartate transaminase (AST): Sham+NaCl: 258.3 ± 33.83 U/L; Sham+SEW2871: 326.4 ± 64.12 U/L; CASP+NaCl: 709 ± 504.7 U/L; CASP+SEW2871: 426.7 ± 135.7 U/L; alanine transaminase (ALT): Sham+NaCl: 54.81 ± 4.99 U/L; Sham+SEW2871: 58.98 ± 3.96 U/L; CASP+NaCl: 101.2 ± 54.03 U/L; CASP+SEW2871: 75.56 ± 14.72 U/L) (Fig. 7C).
Serum levels of Sphingosine-1-phosphate are reduced in SEW2871 treated animals under inflammatory conditions
Finally, the concentration of S1P in the serum of each animal was measured at the end of the experiments (Fig. 8). The serum levels of S1P in Sham animals with SEW2871 treatment (Sham+SEW2871: 284.6 ± 7.08 ng/mL) were slightly increased compared with the Sham+NaCl group (266.1 ± 7.67 ng/mL). The same increase was determined in CASP+NaCl animals (276.9 ± 7.35 ng/mL). Interestingly, in the CASP+SEW2871 group the serum levels of S1P were markedly decreased (239.7 ± 4.01 ng/mL) which reached significance when compared with Sham+SEW2871 and CASP+NaCl.
In recent past, it has been increasingly recognized that endothelial barrier dysfunction is a pivotal and critical event in the pathogenesis of systemic inflammation response syndrome and sepsis. The impairment of the endothelial barrier due to pro-inflammatory stimuli results in an altered microcirculation that is characterized by decreased tissue perfusion leading to multi-organ dysfunction and death (1, 23). However, specific and effective therapeutic approaches to stabilize the endothelial barrier under septic conditions are not available in the human situation yet. Previous studies have shown that S1P and S1P1 agonists are supposed to stabilize endothelial barrier function in different acute tissue injury models (8–10, 13, 17, 24, 25). Thus, in this study, we examined the effects of the specific S1P1 agonist SEW2871 in polymicrobial sepsis using the rodent colon ascendens stent peritonitis (CASP) model (19, 26). In contrast to the previously published studies mentioned above, the prolonged application of SEW2871 in the CASP model did not attenuate the sepsis-induced endothelial barrier breakdown but was rather associated with severe side effects such as bradyarrhythmia and cardiac arrest resulting in an increased lethality. Interestingly, these severe side effects were only observed under septic conditions since sham-operated animals receiving SEW2871 did not reveal any abnormalities. Septic animals treated with SEW2871 that did not show cardiac side effects were characterized with increased vascular permeability as shown by FITC-albumin extravasation in the mesenteric microcirculation and thickened alveolar septa comparable with CASP-operated animals receiving only sodium chloride. The present study shows for the first time in a clinically relevant animal model that the specific S1P1 agonist SEW2871 cannot improve sepsis-induced endothelial barrier dysfunction and is associated with increased mortality due to cardiac side effects when the intravenous application of SEW2871 occurs after the establishment of septic conditions.
Prolonged exposure to S1P and SEW2871 results in endothelial barrier breakdown in vitro
Sphingosine-1-phosphate and agonists have been previously proposed to enhance endothelial barrier function in vitro by the activation of the G-protein coupled S1P receptor-1 leading to activation of Rho family GTPases, cytoskeletal reorganization, adherens-, and tight junction assembly (11, 13–15). However, in parallel it has also been determined that S1P and S1P1 agonists cause a decrease of S1P1 expression on endothelial cells due to receptor internalization and degradation in a time- and dose-dependent manner and thus can impair endothelial barrier integrity, whereby the specific S1P1 agonist SEW2871 only induces a loss of receptor expression at higher concentrations (>250 nM) (27–29). In contrast to published data showing a barrier-stabilizing effect after a short-term treatment of endothelial monolayers with S1P or S1P agonists we observed an endothelial barrier disruption under steady state conditions after long-term treatment with S1P or SEW2871 in a time- and dose- dependent manner. Based on the observations in the literature mentioned above, our finding of a significant decrease of trans-endothelial electrical resistance, the formation of intercellular gaps, and irregular distribution of VE-cadherin could be explained by a loss of normal S1P-S1P1 signaling after long-term exposure. Since it has been reported that septic patients have decreased serum levels of S1P (7), we further tested whether the application of S1P or SEW2871 improves endothelial barrier function under inflammatory conditions in vitro. The application of low-dose S1P (100 nM) did not restore TNF-α-induced endothelial barrier breakdown but led to a further impairment of endothelial barrier. In contrast, the low-dose application of SEW2871 (100 nM) partially attenuated endothelial barrier properties as shown by increased TER and less pronounced irregular staining pattern of VE-cadherin compared with endothelial monolayers treated with TNF-α alone.
Effects of SEW2871 in a clinically relevant sepsis model
Given the fact that the low-dose application of SEW2871 resulted in an improvement of endothelial barrier function under pro-inflammatory conditions, we examined the effect of long-term SEW2871 application in vivo using a modified CASP model (19). Compared with other in vivo inflammation models such as lipopolysaccharide (LPS)- or cecum ligation and puncture (CLP) model, the CASP model presents a clinically relevant model of polymicrobial abdominal sepsis and enables long-term observations (26). In contrast, the LPS model characterized by a high reproducibility due to the defined application of bacterial lipopolysaccharide is more a hyper-inflammation model without sustained bacterial focus and is associated with an early onset of high mortality. A second in vivo sepsis model is the CLP-model that has been proposed to mimic an abdominal polymicrobial sepsis (30). However, it has been recognized that the CLP model is compared with the CASP model a model of intra-abdominal abscess formation with minor signs of systemic inflammation (31). A disadvantage of the CASP model is the interindividual range of inflammatory response in the animals, even if a diffuse peritonitis was detectable in all animals of the current study. Furthermore, it has to be mentioned that the experimental setup presented here with CASP-model and evaluation of microcirculation at the end of the experiments represents a two-hit model with re-anesthesia and relaparotomy that strongly influences the cardiopulmonary performance. It must also be considered that in this study the animals received a fluid replacement that is an essential part in sepsis therapy and improves the survival rate compared with CASP-operated animals without fluid replacement. This may explain why we have no mortality within time course of our experiments in the CASP+NaCl group and the extent of renal dysfunction is less pronounced than reported previously by our group (32). These limitations should be considered in the analysis of this current study.
Intravenous application of SEW2871 led to severe cardiac side effects under septic conditions
In contrast to previously published studies showing beneficial effects of S1P1 agonists in terms of stabilizing the endothelial barrier under inflammatory conditions, the aim of the current study was to evaluate whether SEW2871 has still a therapeutic effect under established septic conditions in vivo since in a clinical setting the drug application usually occurs at the time point of diagnosis and not in a preventive manner. Therefore, SEW2871 was intravenously administrated 12 h after sepsis induction. It is known from our own experiences (19) and from the literature (26) that at time point 12 h CASP-operated animals reveal signs of systemic inflammation response and loss of microvascular barrier properties. Due to fact that the sepsis focus was not removed in this study, we decided a continuous application of SEW2871 following the first bolus.
Our data show for the first time that under septic conditions the intravenous administration of SEW2871 causes severe acute cardiac side effects such as bradyarrhythmia and cardiac arrest after the initial application resulting in significantly increased lethality. Furthermore CASP-operated animals treated with SEW2871 revealed a decreased heart rate after long-term exposure as seen at the time point 24 h. So far, cardiac side effects have been mainly reported for unspecific S1P receptor agonists such as FTY720 that are able to stimulate S1P3(16, 33, 34). In case of the specific S1P1 agonist SEW2871 there is only one study available that has determined that the application of SEW2871 leads to an increased incidence of irreversible postischemic arrhythmias in an ex vivo heart ischemia and reperfusion model (35). In our study, the observed lethal cardiac side effects mostly occurred with the first application of SEW2871 what is comparable with clinical studies where patients with multiple sclerosis received an unspecific S1P agonist that also caused bradyarrhythmia after the first drug administration (16). Interestingly, the cardiac side effects were only detectable in septic animals treated with SEW2871. We did not observe the severe cardiac side effects in the animals with CASP alone, suggesting that is rather unlikely that these side effects may have occurred spontaneously in septic animals. Furthermore, the fact that we had specific effects on the endothelial barrier in our in vitro experiments following SEW2871 application clearly demonstrates that SEW2871 was dissolved properly and was thus effective in our experiments. Therefore, it can be speculated that the pro-inflammatory milieu plays an important role in changing the S1P1 signaling and/or the affinity of SEW2871 to other S1P receptors. Another possibility that cannot be completely excluded is that microembolisms might be responsible for the severe cardiac side effects directly after drug application, even if macroscopically no ischemic areas in organs or tissues of these animals were detectable in our experiments.
However, the observation of severe cardiac side effects under septic conditions was initially unexpected since a recently published work by Coldewey et al. (2) has shown that SEW2871 even in higher concentrations did not alter the cardiac function under pro-inflammatory conditions induced by LPS/PepG co-administration as revealed by echocardiography 18 h later. In the latter study, SEW2871 was applied 1 h after LPS/PepG injection as a single bolus. It remains actually unclear if SEW2871 did not show any severe side effects at this early time point since septic/inflammatory conditions have not yet been established so strong compared with the time point of SEW2871 administration in the present study (12 h after CASP operation). In the lack of a support for this idea further detailed investigations will be necessary.
Regarding the arterial blood pressure no differences were obvious between the different experimental groups. The drop of MAP in Sham-operated animals at the end of the experiment might be explained by relaparotomy with exteriorized gut and re-anesthesia with combined inhalative and intravenous anesthesia. To explain why there are no differences between CASP- and Sham-operated animals, it has to be considered that CASP animals required an increased additional fluid volume to maintain the MAP in these animals. This difference was statistically only significant for the CASP+NaCl group but also in the CASP+SEW2871 animals a clear trend could be determined.
Long-term treatment with SEW2871 does not attenuate sepsis-induced microvascular leakage
In recent years, there has been an increasing body of evidence showing that S1P and S1P1 agonists stabilize endothelial barrier function in vivo under inflammatory conditions and thus lead to an improved outcome (8–10, 13, 17, 18, 24, 25). However, Shea et al. (29) have reported that the long-term exposure of FTY720 in a LPS- and bleomycin-induced lung injury model increases vascular permeability, fibroproliferative response, and even mortality.
The application of SEW2871 in our CASP model did not protect the animals to develop a capillary leakage syndrome as revealed by significantly increased FITC-albumin extravasation in the mesenteric microcirculation. Further clinical parameters that SEW2871 was not able to restore sepsis-induced endothelial barrier breakdown and organ failure are also the requirement of intravasal volume replacement, elevated creatinine values, and increased thickness of alveolar septa in the SEW2871-treated CASP group comparable to septic animals treated only with NaCl. Our findings indicate that the time point and the duration of SEW2871 administrations seem to play a pivotal role in the ability to abolish endothelial barrier breakdown under inflammatory conditions since in our experimental setup the drug administration started when septic conditions were already completely established and sustained until the end of the experiment. In contrast, in the other studies mentioned above, the application of S1P or its agonists mostly occurred as single shot administration in parallel with induction of inflammation or even as preventive therapy. Secondly, the inflammation was mostly induced by LPS or ischemia/reperfusion. Furthermore, as already emphasized for the in vitro findings it has been shown that the long-term exposure of S1P or its agonists impairs endothelial barrier function due to receptor internalization and degradation. This functional antagonism is presumably also possible in vivo and might also explain why the serum levels of S1P were significantly reduced in septic animals treated with SEW2871. Interestingly, S1P levels were increased in the other CASP-group suggesting that the release of S1P might be a rescue mechanism in the early onset of sepsis. This may also explain the difference to observations in clinical and experimental studies in which S1P was generally reduced in sepsis (2, 7). It can be speculated that levels of S1P wave during septic inflammation depending on the phase of sepsis.
Another explanation for the discrepancy between our data and previous studies includes different drug doses and application ways. In the current study, we used a low dose but it can be assumed that a higher dose of SEW2871 will enhance the severe cardiac side effects. Therefore, we did not carry out any further experiments with higher doses.
In conclusion, our study showed for the first time that the application of the specific S1P1 agonist SEW2871 causes severe cardiac side effects and cannot attenuate the inflammation-induced endothelial barrier breakdown in a clinical relevant in vivo sepsis model. Further preclinical studies are required to study the use of S1P agonist under septic conditions and to understand the mechanisms of potential side effects of SEW2871.
The authors are grateful to Alexia Yarezki, Veronika Heimbach, and Silvia Koch for skilful technical assistance.
1. Opal SM, van der Poll T. Endothelial barrier dysfunction in septic shock. J Intern Med
2015; 277 3:277–293.
2. Coldewey SM, Benetti E, Collino M, Pfeilschifter J, Sponholz C, Bauer M, Huwiler A, Thiemermann C. Elevation of serum sphingosine-1-phosphate attenuates impaired cardiac function in experimental sepsis
. Scientific reports
3. Curry FE, Clark JF, Adamson RH. Erythrocyte-derived sphingosine-1-phosphate stabilizes basal hydraulic conductivity and solute permeability in rat microvessels. Am J Physiol Heart Circ Physiol
2012; 303 7:H825–H834.
4. Camerer E, Regard JB, Cornelissen I, Srinivasan Y, Duong DN, Palmer D, Pham TH, Wong JS, Pappu R, Coughlin SR. Sphingosine-1-phosphate in the plasma compartment regulates basal and inflammation-induced vascular leak in mice. J Clin Investig
2009; 119 7:1871–1879.
5. Anliker B, Chun J. Lysophospholipid G protein-coupled receptors. J Biol Chem
2004; 279 20:20555–20558.
6. Schaphorst KL, Chiang E, Jacobs KN, Zaiman A, Natarajan V, Wigley F, Garcia JG. Role of sphingosine-1 phosphate in the enhancement of endothelial barrier integrity by platelet-released products. Am J Physiol Lung Cell Mol Physiol
2003; 285 1:L258–67.
7. Winkler MS, Nierhaus A, Holzmann M, Mudersbach E, Bauer A, Robbe L, Zahrte C, Geffken M, Peine S, Schwedhelm E, et al. Decreased serum concentrations of sphingosine-1-phosphate in sepsis
. Crit Care
8. McVerry BJ, Peng X, Hassoun PM, Sammani S, Simon BA, Garcia JG. Sphingosine 1-phosphate reduces vascular leak in murine and canine models of acute lung injury. Am J Respir Crit Care Med
2004; 170 9:987–993.
9. Peng X, Hassoun PM, Sammani S, McVerry BJ, Burne MJ, Rabb H, Pearse D, Tuder RM, Garcia JG. Protective effects of sphingosine 1-phosphate in murine endotoxin-induced inflammatory lung injury. Am J Respir Crit Care Med
2004; 169 11:1245–1251.
10. Szczepaniak WS, Zhang Y, Hagerty S, Crow MT, Kesari P, Garcia JG, Choi AM, Simon BA, McVerry BJ. Sphingosine 1-phosphate rescues canine LPS-induced acute lung injury and alters systemic inflammatory cytokine production in vivo. Transl Res
2008; 152 5:213–224.
11. Lee MJ, Thangada S, Claffey KP, Ancellin N, Liu CH, Kluk M, Volpi M, Sha’afi RI, Hla T. Vascular endothelial cell adherens junction assembly and morphogenesis induced by sphingosine-1-phosphate. Cell
1999; 99 3:301–312.
12. Singleton PA, Chatchavalvanich S, Fu P, Xing J, Birukova AA, Fortune JA, Klibanov AM, Garcia JG, Birukov KG. Akt-mediated transactivation of the S1P1 receptor in caveolin-enriched microdomains regulates endothelial barrier enhancement by oxidized phospholipids. Circ Res
2009; 104 8:978–986.
13. Camp SM, Bittman R, Chiang ET, Moreno-Vinasco L, Mirzapoiazova T, Sammani S, Lu X, Sun C, Harbeck M, Roe M, et al. Synthetic analogs of FTY720 [2-amino-2-(2-[4-octylphenyl]ethyl)-1,3-propanediol] differentially regulate pulmonary vascular permeability in vivo and in vitro. J Pharmacol Exp Ther
2009; 331 1:54–64.
14. Garcia JG, Liu F, Verin AD, Birukova A, Dechert MA, Gerthoffer WT, Bamberg JR, English D. Sphingosine 1-phosphate promotes endothelial cell barrier integrity by Edg-dependent cytoskeletal rearrangement. J Clin Investig
2001; 108 5:689–701.
15. Lee JF, Zeng Q, Ozaki H, Wang L, Hand AR, Hla T, Wang E, Lee MJ. Dual roles of tight junction-associated protein, zonula occludens-1, in sphingosine 1-phosphate-mediated endothelial chemotaxis and barrier integrity. J Biol Chem
2006; 281 39:29190–29200.
16. Kappos L, Radue EW, O’Connor P, Polman C, Hohlfeld R, Calabresi P, Selmaj K, Agoropoulou C, Leyk M, Zhang-Auberson L, et al. A placebo-controlled trial of oral fingolimod in relapsing multiple sclerosis. New Engl J Med
2010; 362 5:387–401.
17. Wang Z, Sims CR, Patil NK, Gokden N, Mayeux PR. Pharmacologic targeting of sphingosine-1-phosphate receptor 1 improves the renal microcirculation during sepsis
in the mouse. J Pharmacol Exp Ther
2015; 352 1:61–66.
18. Lee JF, Gordon S, Estrada R, Wang L, Siow DL, Wattenberg BW, Lominadze D, Lee MJ. Balance of S1P1 and S1P2 signaling regulates peripheral microvascular permeability in rat cremaster muscle vasculature. Am J Physiol Heart Circ Physiol
2009; 296 1:H33–H42.
19. Flemming S, Schlegel N, Wunder C, Meir M, Baar W, Wollborn J, Roewer N, Germer CT, Schick MA. Phosphodiesterase-4-inhibition dose-dependently stabilizes microvascular barrier functions and microcirculation in a rodent model of polymicrobial sepsis
2014; 41 6:537–545.
20. Flemming S, Burkard N, Renschler M, Vielmuth F, Meir M, Schick MA, Wunder C, Germer CT, Spindler V, Waschke J, et al. Soluble VE-cadherin is involved in endothelial barrier breakdown in systemic inflammation and sepsis
. Cardiovasc Res
2015; 107 1:32–44.
21. Baumer Y, Drenckhahn D, Waschke J. cAMP induced Rac 1-mediated cytoskeletal reorganization in microvascular endothelium. Histochem Cell Biol
2008; 129 6:765–778.
22. Bekker AY, Ritter AB, Duran WN. Analysis of microvascular permeability to macromolecules by video-image digital processing. Microvasc Res
1989; 38 2:200–216.
23. Goldenberg NM, Steinberg BE, Slutsky AS, Lee WL. Broken barriers: a new take on sepsis
pathogenesis. Sci Transl Med
2011; 3 88:88s25.
24. Hemdan NY, Weigel C, Reimann CM, Graler MH. Modulating sphingosine 1-phosphate signaling with DOP or FTY720 alleviates vascular and immune defects in mouse sepsis
. Eur J Immunol
2016; 46 12:2767–2777.
25. Awad AS, Ye H, Huang L, Li L, Foss FW Jr, Macdonald TL, Lynch KR, Okusa MD. Selective sphingosine 1-phosphate 1 receptor activation reduces ischemia-reperfusion injury in mouse kidney. Am J Physiol Renal Physiol
2006; 290 6:F1516–F1524.
26. Lustig MK, Bac VH, Pavlovic D, Maier S, Grundling M, Grisk O, Wendt M, Heidecke CD, Lehmann C. Colon ascendens stent peritonitis—a model of sepsis
adopted to the rat: physiological, microcirculatory and laboratory changes. Shock
2007; 28 1:59–64.
27. Graler MH, Goetzl EJ. The immunosuppressant FTY720 down-regulates sphingosine 1-phosphate G-protein-coupled receptors. FASEB J
2004; 18 3:551–553.
28. Oo ML, Thangada S, Wu MT, Liu CH, Macdonald TL, Lynch KR, Lin CY, Hla T. Immunosuppressive and anti-angiogenic sphingosine 1-phosphate receptor-1 agonists induce ubiquitinylation and proteasomal degradation of the receptor. J Biol Chem
2007; 282 12:9082–9089.
29. Shea BS, Brooks SF, Fontaine BA, Chun J, Luster AD, Tager AM. Prolonged exposure to sphingosine 1-phosphate receptor-1 agonists exacerbates vascular leak, fibrosis, and mortality after lung injury. Am J Respir Cell Mol Biol
2010; 43 6:662–673.
30. Dejager L, Pinheiro I, Dejonckheere E, Libert C. Cecal ligation and puncture: the gold standard model for polymicrobial sepsis
? Trends Microbiol
2011; 19 4:198–208.
31. Maier S, Traeger T, Entleutner M, Westerholt A, Kleist B, Huser N, Holzmann B, Stier A, Pfeffer K, Heidecke CD. Cecal ligation and puncture versus colon ascendens stent peritonitis: two distinct animal models for polymicrobial sepsis
2004; 21 6:505–511.
32. Schick MA, Baar W, Flemming S, Schlegel N, Wollborn J, Held C, Schneider R, Brock RW, Roewer N, Wunder C. Sepsis
-induced acute kidney injury by standardized colon ascendens stent peritonitis in rats - a simple, reproducible animal model. Intensive Care Med Exp
2014; 2 1:34.
33. Sanna MG, Liao J, Jo E, Alfonso C, Ahn MY, Peterson MS, Webb B, Lefebvre S, Chun J, Gray N, et al. Sphingosine 1-phosphate (S1P
) receptor subtypes S1P1 and S1P3, respectively, regulate lymphocyte recirculation and heart rate. J Biol Chem
2004; 279 14:13839–13848.
34. Forrest M, Sun SY, Hajdu R, Bergstrom J, Card D, Doherty G, Hale J, Keohane C, Meyers C, Milligan J, et al. Immune cell regulation and cardiovascular effects of sphingosine 1-phosphate receptor agonists in rodents are mediated via distinct receptor subtypes. J Pharmacol Exp Ther
2004; 309 2:758–768.
35. Tsukada YT, Sanna MG, Rosen H, Gottlieb RA. S1P1-selective agonist SEW2871
exacerbates reperfusion arrhythmias. J Cardiovasc Pharmacol
2007; 50 6:660–669.