Title: Decreased serum level of sphingosine‐1‐phosphate: a novel predictor of clinical severity in COVID‐19
Abstract: Article9 December 2020Open Access Transparent process Decreased serum level of sphingosine-1-phosphate: a novel predictor of clinical severity in COVID-19 Giovanni Marfia Corresponding Author Giovanni Marfia [email protected] orcid.org/0000-0002-5849-7000 Laboratory of Experimental Neurosurgery and Cell Therapy, Neurosurgery Unit, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan, Italy Istituto di Medicina Aerospaziale "A. Mosso", Aeronautica Militare, Milan, Italy Aldo Ravelli" Research Center, Milan, ItalyThese authors contributed equally to this work as first authors Search for more papers by this author Stefania Navone Stefania Navone Laboratory of Experimental Neurosurgery and Cell Therapy, Neurosurgery Unit, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan, Italy Aldo Ravelli" Research Center, Milan, ItalyThese authors contributed equally to this work as first authors Search for more papers by this author Laura Guarnaccia Laura Guarnaccia Laboratory of Experimental Neurosurgery and Cell Therapy, Neurosurgery Unit, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan, Italy Department of Clinical Sciences and Community Health, Università degli Studi di Milano, Milan, Italy Search for more papers by this author Rolando Campanella Rolando Campanella Laboratory of Experimental Neurosurgery and Cell Therapy, Neurosurgery Unit, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan, Italy Search for more papers by this author Michele Mondoni Michele Mondoni Respiratory Unit, ASST Santi Paolo e Carlo, Department of Health Sciences, Università degli Studi di Milano, Milan, Italy Search for more papers by this author Marco Locatelli Marco Locatelli Laboratory of Experimental Neurosurgery and Cell Therapy, Neurosurgery Unit, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan, Italy Aldo Ravelli" Research Center, Milan, Italy Department of Medical-Surgical Physiopathology and Transplantation, Università degli Studi di Milano, Milan, Italy Search for more papers by this author Alessandra Barassi Alessandra Barassi Laboratory of Clinical Biochemistry, ASST Santi Paolo e Carlo, Department of Health Sciences, Università degli Studi di Milano, Milan, Italy Search for more papers by this author Laura Fontana Laura Fontana Department of Medical-Surgical Physiopathology and Transplantation, Università degli Studi di Milano, Milan, Italy Search for more papers by this author Fabrizio Palumbo Fabrizio Palumbo Istituto di Medicina Aerospaziale "A. Mosso", Aeronautica Militare, Milan, Italy Search for more papers by this author Emanuele Garzia Emanuele Garzia Istituto di Medicina Aerospaziale "A. Mosso", Aeronautica Militare, Milan, Italy Reproductive Medicine Unit, ASST Santi Paolo e Carlo, Università degli Studi di Milano, Milan, Italy Search for more papers by this author Giuseppe Ciniglio Appiani Giuseppe Ciniglio Appiani Istituto di Medicina Aerospaziale "A. Mosso", Aeronautica Militare, Milan, Italy Search for more papers by this author Davide Chiumello Davide Chiumello SC Anestesia e Rianimazione, ASST Santi Paolo e Carlo, Milan, Italy Search for more papers by this author Monica Miozzo Monica Miozzo Department of Medical-Surgical Physiopathology and Transplantation, Università degli Studi di Milano, Milan, Italy Unit of Research Laboratories Coordination, Fondazione IRCCS Ca' Granda, Ospedale Maggiore Policlinico, Milan, Italy Search for more papers by this author Stefano Centanni Stefano Centanni Respiratory Unit, ASST Santi Paolo e Carlo, Department of Health Sciences, Università degli Studi di Milano, Milan, ItalyThese authors contributed equally to this work as senior authors Search for more papers by this author Laura Riboni Laura Riboni Department of Medical Biotechnology and Translational Medicine, LITA-Segrate, Università degli Studi di Milano, Milan, ItalyThese authors contributed equally to this work as senior authors Search for more papers by this author Giovanni Marfia Corresponding Author Giovanni Marfia [email protected] orcid.org/0000-0002-5849-7000 Laboratory of Experimental Neurosurgery and Cell Therapy, Neurosurgery Unit, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan, Italy Istituto di Medicina Aerospaziale "A. Mosso", Aeronautica Militare, Milan, Italy Aldo Ravelli" Research Center, Milan, ItalyThese authors contributed equally to this work as first authors Search for more papers by this author Stefania Navone Stefania Navone Laboratory of Experimental Neurosurgery and Cell Therapy, Neurosurgery Unit, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan, Italy Aldo Ravelli" Research Center, Milan, ItalyThese authors contributed equally to this work as first authors Search for more papers by this author Laura Guarnaccia Laura Guarnaccia Laboratory of Experimental Neurosurgery and Cell Therapy, Neurosurgery Unit, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan, Italy Department of Clinical Sciences and Community Health, Università degli Studi di Milano, Milan, Italy Search for more papers by this author Rolando Campanella Rolando Campanella Laboratory of Experimental Neurosurgery and Cell Therapy, Neurosurgery Unit, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan, Italy Search for more papers by this author Michele Mondoni Michele Mondoni Respiratory Unit, ASST Santi Paolo e Carlo, Department of Health Sciences, Università degli Studi di Milano, Milan, Italy Search for more papers by this author Marco Locatelli Marco Locatelli Laboratory of Experimental Neurosurgery and Cell Therapy, Neurosurgery Unit, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan, Italy Aldo Ravelli" Research Center, Milan, Italy Department of Medical-Surgical Physiopathology and Transplantation, Università degli Studi di Milano, Milan, Italy Search for more papers by this author Alessandra Barassi Alessandra Barassi Laboratory of Clinical Biochemistry, ASST Santi Paolo e Carlo, Department of Health Sciences, Università degli Studi di Milano, Milan, Italy Search for more papers by this author Laura Fontana Laura Fontana Department of Medical-Surgical Physiopathology and Transplantation, Università degli Studi di Milano, Milan, Italy Search for more papers by this author Fabrizio Palumbo Fabrizio Palumbo Istituto di Medicina Aerospaziale "A. Mosso", Aeronautica Militare, Milan, Italy Search for more papers by this author Emanuele Garzia Emanuele Garzia Istituto di Medicina Aerospaziale "A. Mosso", Aeronautica Militare, Milan, Italy Reproductive Medicine Unit, ASST Santi Paolo e Carlo, Università degli Studi di Milano, Milan, Italy Search for more papers by this author Giuseppe Ciniglio Appiani Giuseppe Ciniglio Appiani Istituto di Medicina Aerospaziale "A. Mosso", Aeronautica Militare, Milan, Italy Search for more papers by this author Davide Chiumello Davide Chiumello SC Anestesia e Rianimazione, ASST Santi Paolo e Carlo, Milan, Italy Search for more papers by this author Monica Miozzo Monica Miozzo Department of Medical-Surgical Physiopathology and Transplantation, Università degli Studi di Milano, Milan, Italy Unit of Research Laboratories Coordination, Fondazione IRCCS Ca' Granda, Ospedale Maggiore Policlinico, Milan, Italy Search for more papers by this author Stefano Centanni Stefano Centanni Respiratory Unit, ASST Santi Paolo e Carlo, Department of Health Sciences, Università degli Studi di Milano, Milan, ItalyThese authors contributed equally to this work as senior authors Search for more papers by this author Laura Riboni Laura Riboni Department of Medical Biotechnology and Translational Medicine, LITA-Segrate, Università degli Studi di Milano, Milan, ItalyThese authors contributed equally to this work as senior authors Search for more papers by this author Author Information Giovanni Marfia *,1,2,3, Stefania Navone1,3, Laura Guarnaccia1,4, Rolando Campanella1, Michele Mondoni5, Marco Locatelli1,3,6, Alessandra Barassi7, Laura Fontana6, Fabrizio Palumbo2, Emanuele Garzia2,8, Giuseppe Ciniglio Appiani2, Davide Chiumello9, Monica Miozzo6,10, Stefano Centanni5 and Laura Riboni11 1Laboratory of Experimental Neurosurgery and Cell Therapy, Neurosurgery Unit, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan, Italy 2Istituto di Medicina Aerospaziale "A. Mosso", Aeronautica Militare, Milan, Italy 3Aldo Ravelli" Research Center, Milan, Italy 4Department of Clinical Sciences and Community Health, Università degli Studi di Milano, Milan, Italy 5Respiratory Unit, ASST Santi Paolo e Carlo, Department of Health Sciences, Università degli Studi di Milano, Milan, Italy 6Department of Medical-Surgical Physiopathology and Transplantation, Università degli Studi di Milano, Milan, Italy 7Laboratory of Clinical Biochemistry, ASST Santi Paolo e Carlo, Department of Health Sciences, Università degli Studi di Milano, Milan, Italy 8Reproductive Medicine Unit, ASST Santi Paolo e Carlo, Università degli Studi di Milano, Milan, Italy 9SC Anestesia e Rianimazione, ASST Santi Paolo e Carlo, Milan, Italy 10Unit of Research Laboratories Coordination, Fondazione IRCCS Ca' Granda, Ospedale Maggiore Policlinico, Milan, Italy 11Department of Medical Biotechnology and Translational Medicine, LITA-Segrate, Università degli Studi di Milano, Milan, Italy *Corresponding author. Tel: +39 0255034268; Fax: +39 0255038821; E-mail: [email protected] EMBO Mol Med (2021)13:e13424https://doi.org/10.15252/emmm.202013424 See also: H Rosen & MBA Oldstone (January 2021) PDFDownload PDF of article text and main figures. Peer ReviewDownload a summary of the editorial decision process including editorial decision letters, reviewer comments and author responses to feedback. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions Figures & Info Abstract The severity of coronavirus disease 2019 (COVID-19) is a crucial problem in patient treatment and outcome. The aim of this study is to evaluate circulating level of sphingosine-1-phosphate (S1P) along with severity markers, in COVID-19 patients. One hundred eleven COVID-19 patients and forty-seven healthy subjects were included. The severity of COVID-19 was found significantly associated with anemia, lymphocytopenia, and significant increase of neutrophil-to-lymphocyte ratio, ferritin, fibrinogen, aminotransferases, lactate dehydrogenase (LDH), C-reactive protein (CRP), and D-dimer. Serum S1P level was inversely associated with COVID-19 severity, being significantly correlated with CRP, LDH, ferritin, and D-dimer. The decrease in S1P was strongly associated with the number of erythrocytes, the major source of plasma S1P, and both apolipoprotein M and albumin, the major transporters of blood S1P. Not last, S1P was found to be a relevant predictor of admission to an intensive care unit, and patient's outcome. Circulating S1P emerged as negative biomarker of severity/mortality of COVID-19 patients. Restoring abnormal S1P levels to a normal range may have the potential to be a therapeutic target in patients with COVID-19. Synopsis The study demonstrates that patients with COVID-19 experience a significant reduction of serum sphingosine-1-phosphate (S1P). The decrease of S1P associates with the number of erythrocytes, a major source of circulating S1P, as well as with the levels of high-density lipoprotein (HDL)/apolipoprotein M (apoM) and albumin, the most important transporters of circulating S1P. The serum levels of S1P, erythrocytes, apoM and albumin are lower in COVID-19 patients admitted to intensive care unit (ICU) than in no-ICU patients. Serum S1P negatively correlates with clinical parameters including Pneumonia Severity Index and days of hospitalization. S1P levels exhibit a strong power in predicting both ICU admission and mortality. The paper explained Problem Since December 2019, COVID-19 has widely spread throughout the world, causing more than one million of deaths. Although clinical characteristics of COVID-19 patients have been widely reported, significative disease-associated biomarkers remain unknown. We evaluated the potential of circulating sphingosine-1-phosphate (S1P) as a prognostic and predictive biomarker in COVID-19. Results We report demographic, clinical, and laboratory findings of 111 patients with COVID-19, compared to 47 healthy subjects. Several blood parameters, including erythrocyte and lymphocyte number, neutrophil-to-lymphocyte ratio, and biochemical variables encompassing albumin, ferritin, D-dimer, and fibrinogen were found different in COVID-19. The novel, major finding of this work was the significant decrease of serum S1P level in COVID-19 patients, which was significantly related to the decrease of erythrocytes, the major cellular source of circulating S1P, as well as of the two key S1P transporters apoM and albumin. Of relevance, the serum levels of S1P, RBCs, apoM, and albumin exhibited the lowest values in patients admitted to intensive care unit (ICU). Multivariate logistic regression analyses revealed that S1P was the only parameter significantly associated with ICU admission, as well as the strongest predictor for both ICU admission and mortality risk. Impact The results of this study establish S1P as a novel circulating biomarker negatively associated with COVID-19 severity and morbidity, and shed light on the pathophysiology of the disease. Interventions to restore S1P abnormal levels should be considered as a potential therapeutic strategy for reducing the hazard of disease progression and death, and for mounting an effective immune response in patients with COVID-19. Introduction In December 2019, a novel enveloped RNA betacoronavirus, named severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), rapidly spread from China throughout the globe, causing the pandemic respiratory disease COVID-19. Lombardy has been the most affected Italian region by the COVID-19 (Sotgiu et al, 2020) with > 175,000 cases, and 17,414 ascertained deaths (as of October 29, 2020, source: Italian Health Ministry). The clinical spectrum of COVID-19 varies from asymptomatic/paucisymptomatic forms to critical conditions requiring support in an intensive care unit (ICU), and characterized by respiratory and multiple organ failure, systemic manifestations as septic shock, and even death (Guan et al, 2020; Gattinoni et al, 2020). The management of COVID-19 patients is constantly evolving, and so far, several medications are available, although without efficacy confirmation by clinical trials in most instances. It is believed that the cytokine storm caused by systemic overproduction of pro-inflammatory cytokines, including interleukin-6 (IL-6), is a relevant cause of disease severity, progression, and death in COVID-19 patients (Song et al, 2020b). However, even though clinical features directly related to the rapid and intense inflammation, the lack of direct evidence makes unclear how the exuberant inflammatory process and organ failure is completed. Despite the importance of the older age and underlying comor-bidities as prognostic factors associated with severity/death of COVID-19 patients, the evolution of SARS-CoV-2 infection is often very heterogeneous and unpredictable, and robust prognostic biomarkers, and potential therapeutic targets are still lacking. Thus, the actual, rapidly expanding knowledge on COVID-19 highlights the need for a better understanding of its pathophysiology, and for developing novel biomarkers with efficacious prognostic value, and potential therapeutic applications. Among multiple mediators of inflammation, sphingosine-1-phosphate (S1P) emerged as a key signal (Obinata & Hla, 2019), involved in the regulation of multiple pathophysiological processes, including vascular physiology, and immunity (Yanagida & Hla, 2017; Bryan & Del Poeta, 2018). This bioactive sphingoid is produced intracellularly by sphingosine kinases, and, after extracellular release, it exerts pleiotropic effects through binding to specific G protein-coupled receptors (S1P1–5; Liu et al, 2012). In physiological conditions, S1P is present at high concentrations in blood, by the contribution of erythrocytes (Hänel et al, 2007), endothelial cells (Venkataraman et al, 2008), and platelets (Yatomi et al, 2000). To the opposite, its levels in other tissues are low, creating a vascular S1P gradient, crucial for S1P to exert its regulatory roles (Yanagida & Hla, 2017). In the systemic circulation, about 60% of plasma S1P is carried by high-density lipoproteins (HDL), bound to apolipoprotein M (apoM), and about 30% by albumin, and the biological properties of S1P attached to ApoM and albumin are different (Kurano & Yatomi, 2018). On these premises and taking into account that S1P is involved in viral infections (Wolf et al, 2019), and in sepsis (Winkler et al, 2019), the objective of this study was to determine the serum levels of S1P, and its transporters apoM and albumin, to evaluate their clinical importance as prognostic/predictive biomarkers in COVID-19. Results Characteristics of the study population Demographic and clinical characteristics of the two groups are shown in Table EV1. HLT and COV groups were very similar for sex and age, as well as for hypertension or diabetes incidence. The HLT population presented all the biochemical parameters within the normal range, whereas COV patients exhibited a pattern of hematological and biochemical abnormalities. In particular, we found significant differences in RBC, HGB, HCT, RDW, WBC, neutrophils, lymphocytes, and monocytes values (Table EV2). Moreover, the calculated neutrophils-to-lymphocytes ratio (NLR) accounted for 1.78 (IQR: 1.47–2.09) in HLT subjects and 7.05 (IQR: 5.34–8.85) in COV patients, with a strong statistically significant difference (P < 0.0001) between the two groups. Among the biochemical analytes, we found multiple statistically significant differences between COV and HLT (Table EV3), including significant decrease in total proteins, albumin, total cholesterol, HDL-C, 25-OH vitamin, and significant increase in fibrinogen, urea, ferritin, GGT, AST, ALT, CRP, LDH, NT-proBNP, and IL-6. Total bilirubin, albeit within the normal range, was significant lower in COV patients than in HLT. Furthermore, the coagulation profile was altered in COV patients with a significant increased level of D-dimer and fibrinogen. To the contrary, coagulation parameters PT and aPTT, and electrolytes, including calcium, sodium, potassium, and chloride were similar in HLT and COV cohorts. ELISA assays provided a mean value of serum S1P concentration in the HLT group of 0.87 µM, which is very similar to that reported by a recent study on 174 healthy blood donors, and obtained by LC/MS/MS (Daum et al, 2020). The measurement of serum levels of S1P and its transporter apoM revealed a highly significant decrease in both molecules in COV patients, compared to HLT subjects (Fig 1A and B). In particular, the median S1P values were 0.87 and 0.69 µM (Fig 1A), and the median apoM values were 39.0 and 24.3 µg/ml (Fig 1B), in HLT and COV, respectively. These results suggest a systemic involvement of S1P/apoM complex in COVID-19. However, the Pearson correlation between S1P and apoM gave no statistical significance, even though a modest positive trend was present (Fig 1C). Interestingly, serum S1P levels significantly correlated with RBC, HGB, and HCT values (Fig 1D–F). Figure 1. Serum levels of S1P and apoM in HLT and COV patients A, B. Serum levels of S1P (A) and apoM (B) in HLT (n = 47) and COV (n = 111) patients. The box plots represent the interquartile range with median value (central line); the whiskers represent the measured range of HLT and COV. Each measurement was run in triplicate, and performed at least twice. Two-tailed Student's t-test was used for statistical analysis. P < 0.0001 is reported. C–F. Pearson correlation between S1P and apoM (C), RBC number (D), HGB concentration (E), HCT value (F). Scatter plots, together with the fitted regression line, are shown. Pearson correlation was performed for statistical analysis. Exact P values or P < 0.0001 are reported. Download figure Download PowerPoint The Pearson correlation demonstrated a significant correlation between S1P and both HDL-C (Fig 2A), and albumin (Fig 2B). In parallel, significant correlations were found between apoM and HDL-C (Fig 2C) and albumin (Fig 2D). To the contrary, no significant correlation was found between PLT and S1P, as well as PLT and apoM. Figure 2. Correlations between S1P/apoM and their transporters A, B. Correlations between S1P and HDL-C (A), and albumin (B). C, D. Correlations between apoM and HDL-C (C) and albumin (D). Data information: Scatter plots and fitted regression line are shown in each figure. Each S1P and apoM measurement was run in triplicate, and performed at least twice in independent assays. Pearson correlation was performed for statistical analysis. Exact P values or P < 0.0001 are reported. Download figure Download PowerPoint Identification of disease's severity markers To investigate the role of S1P and apoM as circulating biomarkers in COVID-19 infection severity, we stratified patients based on whether or not they needed ICU admission, categorizing our COV patient cohort into ICU and noICU groups. As shown in Table EV4, both groups were homogeneous in terms of age and BMI. As expected, ICU patients showed a significant increase in mortality rate, even if their frequency of hypertension and diabetes was significantly lower than that of noICU. Furthermore, the comparison between blood and biochemical parameters revealed that ICU patients, compared to noICU, have a significant decrease in RBC, HGB, HCT, and lymphocytes, concomitant to a significant increase in WBC, neutrophils, NLR, urea, AST, ALT, LDH, CRP, and D-dimer (Table EV5). Intriguingly, patients admitted to ICU had significantly lower values of S1P, apoM, and of albumin and HDL-C, compared to the noICU group (Fig 3A–D), suggesting the potential role of S1P and its blood transporters apoM, albumin, and HDL-C, as circulating biomarkers of disease severity. Figure 3. Serum levels of S1P and its blood transporters in noICU and ICU patients A–D. The serum concentrations of S1P (A), apoM (B), albumin (C), and HDL-C (D) in noICU (n = 89) and ICU (n = 22) patients are shown. The box plots represent the interquartile range with median (central line); the whiskers represent the measured range of noICU and ICU patients. Each S1P and apoM measurement was run in triplicate, and performed at least twice in independent assays. Two-tailed Student's t-test was used for statistical analysis. P < 0.0001 is reported. Download figure Download PowerPoint To determine the possible relation between serum levels of S1P and apoM and the severity of infection in COVID-19 patients, we used PSI, length of hospitalization, and NLR as parameters. Interestingly, a significant negative correlation was found between S1P and PSI, days from hospital admittance, and NLR (Fig 4A–C). Similarly, significant negative correlations were found between apoM, PSI, and NLR (Fig 4D and F), but not between apoM and days from hospital admission (Fig 4E). Figure 4. Correlation between S1P/apoM and COVID-19 infection severity A–F. Pearson correlation between S1P (A–C) or apoM (D–F) and PSI (A, D), days from admission (B, E), and NLR (C, F) in COV (n = 111). Scatter plots and fitted regression line are shown in each figure. Exact P values or P < 0.0001 are reported. Download figure Download PowerPoint In univariable logistic regression analysis, PSI, S1P, apoM, RBC, HGB, HCT, WBC, NLR, CRP, and albumin were significantly associated with ICU admission. Results of multivariable regression analysis showed that NLR and S1P are independent predictors for ICU admission. Noteworthy, S1P, among the considered parameters, is the most important risk factor for ICU admission for COVID-19 patient (OR: 39.45, [95% CI: 1.51–1031.60]; P = 0.027; Table EV6). Further, to estimate the potential prognostic role of S1P, COV patients were stratified by the cutoff value of S1P 0.60 µM, according to Youden index. Cox regression analysis revealed that S1P is a significant predictor of ICU admission, and in-hospital mortality (Fig 5A and B). Of relevance, the mortality rate in COV patients with S1P < 0.60 μM (13.6%) was found significantly higher (P < 0.05) than that with S1P ≥ 0.60 μM (4.5%) In addition and interestingly, patients admitted to ICU with low serum S1P showed a meaningful rise of mortality rate up to 33%. Figure 5. Prognostic value of S1P in COV patients A, B. Cumulative risk for ICU admission (A) and cumulative survival (B) in COV patients (n = 111), grouped for cutoff value of S1P serum level of 0.60 μM. Statistical analysis was performed by Cox regression. Exact P values are reported. Download figure Download PowerPoint Discussion The major findings of this study are that serum S1P levels are significantly lower in COVID-19 patients than HLT and predict both ICU admission and in-hospital mortality. A very recent lipidomic study, published while our paper was in preparation, reported that S1P is significantly reduced in plasma samples of COVID-19, without indicating molar concentrations, which prevents comparison across independent studies, and without addressing to disease severity and patient outcome (Song et al, 2020a). To our knowledge, this is the first study providing evidence for serum S1P association with COVID-19 severity, and suggesting S1P as a novel circulating biomarker of COVID-19 severity and morbidity. Previous studies demonstrated that the cytokine storm plays a key role in COVID-19, and the increase of pro-inflammatory cytokines is associated with disease severity (Song et al, 2020b). Of note, S1P was implicated in both upstream and downstream cytokine production and increased interstitial levels of S1P at the inflammatory sites induce the expression of pro-inflammatory cytokines (Obinata & Hla, 2019), suggesting that local, interstitial S1P may concur to the cytokine storm of COVID-19. Taking into consideration the relevance of the cytokine storm in COVID-19, we here discuss our findings on serum S1P in COVID-19 patients through two main sections: (i) the possible mechanisms underlying decreased level in serum S1P, including cellular contributors of circulating S1P, and its blood transporters; (ii) the functional consequences and clinical implications of low circulating S1P. To facilitate the reader, Fig 6 provides an overview of the proposed mechanisms underlying S1P involvement in COVID-19. Figure 6. Overview of the proposed mechanisms underlying S1P involvement in COVID-19 pathophysiology and severity After SARS-CoV-2 infection, a local inflammation occurs, with increased pro-inflammatory cytokines (1). This promotes an interstitial increase of S1P, which in turn potentiates cytokine secretion by different cells (2). The exuberant local cytokine levels result in a systemic cytokine storm (3). This gives on to different alterations (4), including anemia (with impaired S1P synthesis/export) and acute-phase response in the liver (with decrease in the negative acute-phase proteins albumin and apoM, which act as S1P transporters). These alterations lead to a progressive drop of circulating S1P, with decrease in both apoM/S1P and albumin/S1P (5). The reduction of S1P in the systemic circulation correlates with COVID-19 severity and patient outcome (6). Download figure Download PowerPoint Mechanisms underlying the reduction of serum S1P Due to their unique S1P metabolism and processing, erythrocytes are a major source of S1P in blood plasma, their count positively associates with plasma S1P levels (Hänel et al, 2007). We found that COVID-19 patients have a significantly reduced RBC count, and that this count, together with total hemoglobin and hematocrit values were directly correlated with serum S1P. These findings, and the reports on low blood S1P in anemic patients (Ohkawa et al, 2008), strongly suggest that RBC decrease may contribute to low serum S1P in COVID-19 patients. Although the mechanism-underlying anemia in COVID-19 is unknown, a possible explanation resides in the finding that pro-inflammatory cytokines inhibit erythropoietin-induced erythropoiesis (Pierce & Larson, 2005). This process was associated with RDW increase (Pierce et al 2005), which we found in our COVID-19 cohort. Furthermore, as a very recent preprint notice reported a significant sphingolipid decrease in RBC from COVID-19 patients (Thomas et al, 2020), we cannot exclude that a decrease in S1P synthesis/export by erythrocytes might also occur. In addition to erythrocytes, the vascular endothelium efficiently contributes to plasma S1P levels through the plasma membrane S1P-transporter Spns2 (Nagahashi et al, 2013). As pro-inflammatory cytokines down-regulate Spns2 in endothelial cells (Jeya Paul et al, 2020), and thus their S1P export, it is possible that the cytokine storm in COVID-19 patients contributes to an endothelial-induced S1P drop. In agreement, patients with sepsis have decreased serum S1P levels (Winkler et al, 2019). A further source of blood S1P are platelets, which efficiently store S1P, and mainly release it into the circulation during clotting, leading to higher S1P levels in serum than plasma (Yatomi et al, 2000). Our COVID-19 cohort had platelet numbers and parameters in the normal range, most probably excluding platelets as responsible for their low serum S1P. Besides possible alterations of S1P release into blood, our data reveal that decreased levels of apoM and a