Cardiac surgery is still associated with a high early perioperative in-hospital mortality (1, 2). Surgery related mortality has partially been attributed to the surgery induced acute inflammatory response. In this context, cardiopulmonary bypass (CPB) induces severe inflammation and is associated with higher risk of organ failure such as acute kidney injury (AKI), which increases the risk of not surviving the hospital stay by 2-fold (3–5). Although techniques such as off-pump procedures may reduce postoperative mortality, the acute inflammatory response after cardiac surgery remains challenging due to the massive cardiovascular stress after surgery (6). Mechanisms of surgery induced inflammation and other causes of systemic inflammation are very similar. For example, the clinical presentation of sepsis patients compared to patients with severe postoperative inflammation after cardiac surgery is indistinguishable and hemodynamic instability, volume deficiency and lactate acidosis are leading symptoms (7, 8). Hallmarks are hypotension caused by vasoplegia, edema formation caused by endothelial barrier disruption and uncontrolled cytokine secretion with an increased vulnerability for secondary infections. All of these processes are potentially regulated by the vascular protective G-protein coupled signalling lipid sphingosine-1-phosphate (S1P) (9). S1P supports crucial functions relevant in inflammation and in the control of hyperinflammatory states: I) Regulation of the endothelial cell (EC) barrier: S1P is fundamental for EC barrier stabilization by regulating the EC cytoskeleton (10, 11). In-vivo models have demonstrated a robust stabilizing effects of S1P preventing lung edema and microvascular permeability induced by toxins (endotoxin, lipopolysaccharide) or high-volume ventilation (12, 13). II) Recruitment of lymphocytes and host response: S1P is the only known inducer of lymphocyte migration from lymphatic organs into peripheral blood. Mice with depleted S1P levels are lymphopenic (14–16). This effect of S1P on the immunity is induced by FTY-720 (Fingolimod), which is a standard of care in Multiple sclerosis. Moreover, S1P agonists are effectful inhibitors of viral induced inflammation in mice infected with influenza (17). Finally, S1P signaling may depend on its main carrier, high-density lipoprotein (HDL) or albumin (18).

There is growing evidence that S1P levels are low in patients with sepsis induced systemic inflammation and that low S1P levels correlate with a higher mortality and morbidity (19–22). Moreover, decreased S1P levels are predictive of hyperinflammatory shock with similar precision as the sequential organ failure assessment (SOFA) (20). Manipulating S1P-controlled processes by S1P substitution or S1P-receptor activation are discussed as therapeutic options to attenuate the systemic inflammatory response. These effects are within the range of expectancy due to the well-known activities of S1P on endothelial integrity, the immune response and as a pro-survival factor (20, 23–25). Considering the observations in human systemic inflammation and current experimental data, we aimed to investigate S1P in patients undergoing cardiac surgery induced inflammation.

This report analyses S1P levels in cardiac surgery induced inflammation. Our main findings are that S1P levels are disrupted by heart surgery, changes of S1P levels inversely correlate with acute phase markers for inflammation, and the influence of CPB on S1P levels seems negligible compared to off-pump procedures. Cardiac surgery is an extremely strong inflammatory stimulus induced by tissue trauma, contact activation of the blood with non-endothelial surfaces, coagulation activation, endotoxemia and ischaemia-reperfusion injury (5, 27). This provokes an immediate acute-phase response, which is characterized by alterations of cytokines levels and acute-phase proteins (28, 29). Clinically, the resulting systemic inflammatory response is associated with severe intra- and postoperative complications as myocardial, respiratory and acute kidney injury (AKI), neurological impairment, bleeding and in its most severe form multiorgan failure (27). A recently published systematic review, including data of more than 14,000 patients, analysed which markers were useful in determining the magnitude of surgery induced injury (30). For example, the authors found that peak CRP was 36-times higher in patients after cardiac surgery compared to pre-surgery and that high concentrations of IL-6 and CRP were closely associated with the degree of surgical injury (30). PCT is another well-described marker, but rather suitable for postoperative infectious complications after cardiac surgery (31). Elevated levels of IL-6, CRP, leucocytes and PCT indicate a substantial inflammatory reaction during and after surgery in our cohort. Inflammation in cardiac surgery is often attributed to CPB and there is a still on-going scientific debate on whether operative strategies avoiding CPB may be superior. In our study there was no significant difference in the measured parameters of inflammatory response in both patient groups with and without CPB. Additionally, we did not observe a relevant difference in S1P levels between the two groups and S1P levels are altered in all patients regardless of the use of CPB. In contrast to the clinically established inflammatory markers IL-6, CRP or leukocyte counts, S1P may better indicate a disturbed immune response together with endothelial cell injury. Our data suggests that there is a reliable and prompt S1P drop after the surgical stimulus. The time from the surgical stimulus until the significant S1P drop is very short. Therefore, we believe that, due to the important role of circulating S1P in stabilizing endothelial integrity and immune response, S1P may function as a potential marker of endothelial injury during and after cardiac surgery, which might be predictive for recovery.

A hallmark of systemic inflammation is the damage to the endothelial cell layer and recovery of endothelial cell function determines the outcome. This is most relevant in systemic reactions induced by bacterial pathogens such as in sepsis (32). Cardiac surgery induces a sepsis-like syndrome associated with a high degree of endothelial cell dysfunction. Recent studies report drastically increased postoperative levels of endothelial markers in blood such as vascular endothelial cadherin and endocan, which correlate with the severity of endothelial cell injury after cardiac surgery (33, 34). Von-Willebrand factor regulates adhesion of immune cells to injured endothelium and has been previously studied in cardiac patients (35). In contrast to components released by injured endothelial cells, S1P is a signalling molecule regulating endothelial cell structure via specific G-protein coupled receptors (36). The endothelium is an important source of circulatory S1P and maintains cell function in an auto-protective manner (37). Cardiac surgery may interrupt this balance and injured endothelial cells may produce less S1P, which further promotes endothelial dysfunction and barrier breakdown. An analogue mechanism has been discussed by other researchers for chronic endothelial dysfunction in atherosclerosis (38). Furthermore, circulating S1P levels are not only determined by endothelial cells. Hematopoietic cells, erythrocytes and platelets are a rich source for S1P as well and all contribute to plasma-S1P levels (15). Cardiac surgery induces haemolysis; bleeding complications are common and high platelet turnover and microthrombi are frequent. This could be reason for the observed acute changes in S1P concentration during surgery when hematopoietic S1P sources are compromised. Circulating S1P further depends on its carrier proteins HDL and albumin, which contribute to S1P related effects (39). During recovery after surgical trauma healing mechanisms, such as increased production of anti-inflammatory proteins as HDL or albumin, may be involved in restoring S1P homeostasis and explain our observation.

Another function of S1P is its protective role in hyperinflammatory states. Experimental studies have been performed in various models. For instance, in mice infected with H1N1 influenza, administration of S1P receptor 1 agonists attenuate the cytokine release by pulmonary endothelial cells (17). Another well-studied role for S1P is the regulation of lymphocyte egress from lymphatic organs into the blood (14). It is possible, due to the lack of the immunoregulative function of S1P in the perioperative phase following cardiac surgery, that patients are vulnerable for infectious postoperative complications. Even though it is a comparably small cohort with a short observational period of only 4 days, we have found that patients with S1P levels not recovering to preoperative levels stayed longer on the ICU and had a higher SOFA score on POD1.

S1P levels dropped in every individual. The analysis of S1P kinetics revealed the most relevant decrease before the onset of CPB and before equivalent off-pump-procedures, respectively. The decrease in S1P coincided with invasive surgical measures as sternotomy, luxation of the heart and cannulation for CPB and the administration of unfractionated heparin in a high dose to prepare for upcoming CPB or off-pump procedure. According to the standard institutional protocol, on- and off-pump patients receive an initial i.v. bolus of 300-400 IE/kg heparin and one possibility is that S1P levels are influenced by heparin. Yatomi et al. measured 60% lower S1P levels in non-coagulated blood (191 ± 79 vs. 484 ± 82 pmol/mL) (40). Thus, it is possible that heparin is causing the drop of S1P. The release of S1P by thrombin-activated platelets into the circulation has previously been described (41, 42). In trauma, tissue factor (TF) is exposed and initiates the coagulation cascade. Binding of activated factor (F) VII to TF allows the binding of FX and its conversion to FXa, which than stimulates thrombin generation (43). Consequently, targeting FXa, via heparin, may attenuate platelet activation and subsequently cause S1P levels to drop. The role of platelet aggregation inhibitors in S1P release by platelets has recently been investigated in pre-clinical studies. The release of S1P induced by activation of thrombin receptor (PAR-1) is inhibited by aspirin in-vitro and ex-vivo (44). Moreover, FXa induces the expression of S1P producing kinases and subsequently increases S1P formation (45). Taken together, FXa inhibitors, such as heparin, may decrease circulating S1P levels, which has been lately included in a US patent description (US20170296549A1). Mechanistically, this could be an effect of inhibited thrombin activation of platelets, which is a main source for circulating S1P. Reduced S1P levels in open heart surgery might be affected by injured endothelial cells, reduced hematopoietic sources and carriers or by iatrogenic inhibition of the coagulation system. Interestingly, this is the first report showing that high dose anticoagulation with heparin may influence S1P levels in patients undergoing cardiac surgery. We cannot exclude that this hypothetical pleiotropic effect of heparin might have a larger effect on vascular- and immunomodulatory S1P than we expected. A S1P/heparin interaction on the immune response has been described and therefore cannot be excluded also under high-dose heparin treatment during cardiac surgery (46).

Limitations of this study are that it was carried out at a single center and involved a relatively small number of patients. The results are strictly observational and therefore, the identity of underlying mechanisms remains speculative. Nevertheless, due to the reliable drop of S1P in all patients, we believe that our observations warrant follow-up studies. For instance, confirmatory studies are needed to evaluate a possible relationship of high dose heparin with the drop of S1P levels. The observed drop of S1P is robust, however, we do not know if the observed lower S1P levels will have a functional significance on S1P signalling. Follow-up interventional studies should address the potential of S1P or S1P mimics to benefit patients undergoing cardiac surgery.

In conclusion, in this prospective observational study we report a severe drop in circulatory S1P in patients undergoing cardiac surgery independent of the use of CBP. S1P concentrations might be negatively affected by endothelial injury, loss of S1P sources or other intraoperative events such as anticoagulation by heparin. It is an intriguing possibility to utilize measurements of circulatory S1P to predict the severity of surgery-induced inflammation, ICU length of stay and general clinical outcome. Moreover, our observations encourage future interventional studies to investigate the therapeutical potential of S1P or S1P mimics in cardiac surgery.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Ethics committee - Aerztekammer Hamburg. The patients/participants provided their written informed consent to participate in this study.

Study conception and design: GG and MW. Acquisition of data: GG, EM, HS, AP, AN, ES, and MW. Analysis and interpretation of data: GG, EM, KA, ES, GD, BT, and MW. Drafting manuscript: GG, ES, KA, GD, and MW. Critical revision: GG, AN, SK, HR, CZ, ES, GD, BT, and MW. All authors contributed to the article and approved the submitted version.

The study was financed by internal institutional funding.

SK declares the following competing interests: He received research support by Ambu, E.T.View Ltd, Fisher & Paykel, Pfizer and Xenios; lecture honorarium from Arjo-Huntleigh, Astellas, Astra, Basilea, Bard, Baxter, Biotest, CSL Behring, Cytosorbents, Fresenius, Gilead, MSD, Orion, Pfizer, Philips, Sedana, Sorin, Xenios and Zoll. SK received consultant honorarium from AMOMED, Astellas, Baxter, Bayer, Fresenius, Gilead, MSD, Pfizer and Xenios. MW declares the following competing interests of funding from Sartorius lung research.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.