CytoSorb hemoperfusion markedly attenuates circulating cytokine concentrations during systemic inflammation in humans in vivo

The study protocol was approved by the local ethics committee (CMO Arnhem-Nijmegen; reference nos. NL71293.091 and CMO2019-5730). Twenty-four healthy male volunteers between 18 and 35 years of age were recruited. All subjects provided written informed consent and were included after medical history, physical examination, routine laboratory tests and a 12-lead electrocardiogram revealed no abnormalities. Smoking, use of any medication, previous participation in experimental human endotoxemia, allergies to any of the investigational drugs or signs of acute illness within two weeks prior to the start of the study were considered exclusion criteria. The study was registered at ClinicalTrials.gov (NCT04643639). All study procedures were performed in accordance with the Declaration of Helsinki and its most recent revisions.

We performed an open label randomized controlled repeated experimental human endotoxemia study, the design of which is depicted in Fig. 1. After inclusion, all subjects were intravenously challenged twice with the same dose of LPS (details on the experimental human endotoxemia procedure are provided below). The first LPS challenge (on day 0) served to quantify the extent of the primary cytokine response and to induce endotoxin tolerance. The second LPS challenge (on day 7) served to assess the degree of endotoxin tolerance, as reflected by a markedly attenuated cytokine response compared to the first challenge. Subjects were randomized into one of two groups (CytoSorb or control). During the first LPS challenge only, subjects in the CytoSorb group were treated with CytoSorb hemoperfusion for six consecutive hours.

All endotoxemia-related procedures were performed according to our standard continuous LPS-infusion protocol (12). In short, subjects were admitted to the research unit situated on the Medium Care department of the Radboud university medical center for approximately ten hours. Subjects had to refrain from consumption of alcohol and caffeine (24 h), and food and drinks (12 h) prior to LPS administration. Upon admission on the first LPS challenge day, two antebrachial venous cannulas were placed to allow administration of fluids, study medication and LPS. A radial artery catheter (BD Infusion Therapy Systems, Sandy, Utah, USA) was placed under ultrasound guidance to allow serial blood sampling and continuous hemodynamic monitoring. To reduce the risk of vasovagal responses (14), hydration fluids (2.5% glucose/0.45% sodium chloride) were administered as an initial prehydration bolus of 1.5L in the 45 min prior to LPS administration, and at a rate of 150 mL/hr for the remainder of the experiment. After prehydration, a bodyweight adjusted bolus dose of 1 ng/kg LPS (E. coli type O:113, lot no. 94332B1; List Biological Laboratories, Campbell, USA) was administered intravenously, directly followed by continuous infusion of LPS at a rate of 0.5 ng/kg/hr for three hours. Blood samples were serially obtained to construct time-concentration curves of circulating cytokines. Heart rate and intra-arterial blood pressure were measured using a 4-lead electrocardiogram (M50 Monitor, Philips, Eindhoven, the Netherlands) and a radial artery pressure transducer (Edwards Lifesciences, Irvine, California, USA). Hemodynamic parameters were recorded every 5 s using in-house developed software. Every 30 min, body temperature was measured using a tympanic thermometer (FirstTemp Genius 2, Covidien, Dublin, Ireland). Flu-like symptoms (headache, nausea, cold shivers, myalgia and back pain) were scored every 30 min using a numeric six-point Likert scale (0 = no symptoms, 5 = worst ever experienced). A composite score was then calculated as the sum of all previous elements, resulting in a total symptom score ranging from 0 to 25. Furthermore, general malaise was scored on an eleven-point scale (0 = no symptoms, 10 = worst ever experienced).

The CytoSorb 300 mL hemoadsorption device (CytoSorbents Corporation, Princeton, New Jersey, USA) was placed into an extracorporeal circuit on a hemodialysis machine with a dedicated hemoperfusion blood line set (Gambro Prismaflex System and Gambro HP-X line set, Baxter, Deerfield, Illinois, USA). Prior to use, the CytoSorb adsorber was primed with two liters of sodium chloride 0.9% solution as per the manufacturer’s instructions. Next, a hemodialysis catheter (High Flow Double Lumen 13Fr 250 mm, Joline, Hechingen, Germany) was placed in the right femoral vein under ultrasound guidance and local anesthesia, only in subjects of the CytoSorb group. Initial experiments with heparin or bivalirudin anticoagulation monotherapy were terminated prematurely due to clot formation in the extracorporeal circuit, likely related to LPS-induced platelet activation (15). Therefore, 250 mg of acetylsalicylic acid (ASA) in addition to 80 IU/kg of unfractionated heparin was administered intravenously after placement of the catheters. Directly thereafter, heparin was continuously administered until one hour prior to the end of hemoperfusion. Activated clotting time (ACT) was measured every 30–60 min using a point of care analyzer (Hemochron Signature Elite, Werfen, Bedford, USA), and heparin infusion rates were adjusted throughout the experiment to maintain an ACT-ratio between 2 and 2.5 times the baseline value. As ASA and heparin may both have immunomodulatory effects (16, 17), the same anticoagulant regimen was used in the control group to prevent potential confounding related to the administration of these drugs. The goal of this proof-of-principle study was to investigate whether the CytoSorb device is capable of attenuating plasma cytokine concentrations after induction of systemic inflammation. Previous endotoxemia studies have demonstrated that several cytokines already start to increase within 30 min after LPS administration (6). Therefore, CytoSorb hemoperfusion was initiated fifteen minutes after the initial LPS bolus to maximize the chances of demonstrating a treatment effect. To this end, subjects in the CytoSorb group were treated with CytoSorb hemoperfusion for six consecutive hours at a constant blood flow rate of 250 mL/min. After six hours, the hemoperfusion treatment was terminated, the remaining blood in the extracorporeal circuit was returned and subjects were disconnected from the device.

A common feature of systemic inflammation is reduced vascular reactivity, as reflected by an attenuated increase in blood pressure upon norepinephrine administration (18). To assess whether treatment with CytoSorb hemoperfusion affects vascular reactivity, a norepinephrine sensitivity test was performed at baseline and four hours after the LPS bolus administration on the first LPS challenge day. At both timepoints, norepinephrine was continuously administered in increasing dosages (0.05, 0.1 and 0.2 mcg/kg/min) for five minutes per dose. Blood pressure was measured continuously via the radial artery catheter to construct dose–response curves. High-frequency hemodynamic data were captured using specialized software (ICM + , Cambridge Enterprise, Cambridge, UK). Data were synchronized on start-times of norepinephrine infusion, and noise was removed from the high-frequency signal by calculating simple moving averages. Differences in vascular reactivity between and within subjects over time were analyzed by comparing the baseline-corrected area under the curve (AUC) of the blood pressure–time curve.

To assess changes in single pass cytokine clearance and elimination over time, paired blood samples from the inlet and outlet ports of the CytoSorb adsorber were serially obtained. Cytokine clearance was calculated according to the following formula:where Cl = clearance (in mL/min), Ci = inlet plasma cytokine concentration (in pg/mL), Co = outlet plasma cytokine concentration in (pg/mL), and F = plasma flow through the adsorber (in mL/min, calculated as blood flow × (1-hematocrit). The cytokine elimination rate (E, in pg/min) was calculated as \(E = {\text{Cl}} \times {\text{Ci}}\).

Monocyte HLA-DR (mHLA-DR) expression was assessed in ethylenediaminetetraacetic-acid (EDTA)-anticoagulated whole blood using the Anti-HLA-DR/Anti-Monocyte Quantibrite assay (BD Biosciences, San Jose, California, USA) on a CytoFLEX flow cytometer (Beckman Coulter, Indianapolis, Indiana, USA). Flow data were analyzed using Kaluza Software (Beckman Coulter, Indianapolis, Indiana, USA). The number of antibodies bound per cell was calculated by standardizing the geometric mean of monocyte HLA-DR fluorescence intensity (MFI) to BD Quantibrite phycoerythrin (PE) beads (BD Biosciences, San Jose, California, USA). mHLA-DR expression was assessed on both LPS challenge days one hour before, as well as three and six hours after LPS administration.

To determine plasma cytokine concentrations, EDTA-anticoagulated blood was centrifuged (10 min, 2000 g, 4 °C) directly after withdrawal, and plasma was stored at − 80 °C until analysis. Concentrations of tumor necrosis factor (TNF), interleukin (IL)-6, IL-8, IL-10, macrophage inflammatory protein (MIP)-1α, monocyte chemoattractant protein (MCP)-1, granulocyte colony stimulating factor (G-CSF) and interferon-γ-induced protein (IP)-10 were determined batch-wise using a simultaneous Luminex assay (Milliplex, Millipore, Billerica, USA) as per the manufacturer’s instructions. Hemocytometry parameters were measured in EDTA-anticoagulated whole blood using a fully automated hematology analyzer (Sysmex XN-1000, Sysmex Corporation, Kobe, Japan).

Distribution of data was assessed for normality based on histogram plots and Shapiro–Wilk’s test, and data with a non-Gaussian distribution were log-transformed prior to statistical testing if a nonparametric equivalent test was not available. Normally distributed data are presented as mean ± standard error of the mean, whereas nonparametric data are presented as median and interquartile range. The AUC of the plasma cytokine concentration–time curve was calculated as an integral measure of cytokine responses over time for each LPS challenge day. Within-group differences in cytokine concentrations, body temperature, symptom score, hemodynamic, and hemocytometry parameters over time were analyzed using one-way repeated measures analysis of variance (RM-ANOVA), whereas between-group differences over time were analyzed using two-way RM-ANOVA (time × group interaction term). All data analysis and visualization were performed using R version 4.1.3 (The R Foundation for Statistical Computing, Vienna, Austria).

Intravenous administration of LPS resulted in a profound but transient increase in concentrations of all measured cytokines on both LPS challenge days (Fig. 2 and Additional file 1: Fig. S1). In subjects of the CytoSorb group, the increase in plasma concentrations of cytokines upon the first LPS challenge was significantly attenuated compared to the control group: TNF (median AUC − 58%, p < 0.0001), IL-6 ( − 71%, p = 0.003), IL-8 ( − 48%, p = 0.02), IL-10 ( − 26%, p = 0.03), MCP-1 ( − 34%, p = 0.02), and MIP-1α ( − 39%, p = 0.006), while a trend toward less pronounced G-CSF concentrations was observed ( − 47%, p = 0.054, Fig. 2 and Additional file 1: Fig. S1). Interestingly, plasma concentrations of the chemokine IP-10 were not significantly affected during the first challenge (+ 25%, p = 0.15, Additional file 1: Fig. S1A). Upon the second LPS challenge seven days later, a severely blunted response was observed in both groups, as reflected by significantly attenuated plasma concentrations of all cytokines as compared to the first challenge (all p < 0.0001), indicative of profound endotoxin tolerance (Fig. 2 and Additional file 1: Fig. S1). However, no significant between-group differences in plasma concentrations of any of the measured cytokines were observed during the second challenge (Fig. 2 and Additional file 1: Fig. S1). Furthermore, no significant reduction in mHLA-DR expression, a well-established marker of immunological tolerance, was observed during the first LPS challenge in either group, whereas the decrease in mHLA-DR expression during the second LPS challenge was similar in both groups (Additional file 2: Fig. S2).

Cytokine clearance and elimination rates are displayed in Fig. 3 and Additional file 3: Fig. S3. Although all measured cytokines were removed from the circulation by the adsorber to some degree, some were cleared more effectively than others with median clearance rates throughout the experiment varying from 26 (21–28) mL/min for TNF to 7 (2–21) mL/min for G-CSF (Fig. 3 and Additional file 3: Fig. S3). Clearance rates showed a declining trend over time, possibly due to saturation of the adsorber. For example, between one and six hours after the LPS bolus administration, median clearance rates of IP-10 attenuated by 83%, whereas IL-6 clearance rates decreased by 43%. For some cytokines, such as IL-8 and MIP-1α, clearance and elimination rates became negative after several hours, indicating that the adsorber releases cytokines back into the circulation when fully saturated, although the absolute amounts were small and the impact on plasma concentrations therefore negligible (Fig. 3 and Additional file 3: Fig. S3).

Administration of LPS resulted in a decrease in blood pressure and a compensatory increase in heart rate in both groups during both LPS challenge days, albeit to a less extent upon the second challenge (Fig. 4AB). In subjects treated with CytoSorb during the first LPS challenge, these effects occurred earlier, but were less sustained compared to the control group (both p < 0.0001). Treatment with CytoSorb during the first LPS challenge did not influence the decrease in blood pressure observed during the second LPS challenge (Fig. 4AB). The increase in blood pressure induced by intravenous administration of increasing dosages of norepinephrine at baseline (just prior to the first LPS administration) did not differ significantly from the response four hours later in either of the two groups (Additional file 4: Fig. S4A). This indicates that experimental human endotoxemia did not significantly attenuate norepinephrine sensitivity. Although subjects in the CytoSorb group demonstrated a significantly more pronounced increase in blood pressure than subjects of the control group upon administration of norepinephrine prior to the LPS challenge, the magnitude of this difference did not change upon administration of norepinephrine four hours after the LPS challenge (p < 0.01 at both timepoints, Additional file 4: Fig. S4B).

LPS administration induced a substantial increase in body temperature in both groups during both LPS challenge days (Fig. 4C). Prior to LPS administration on the first LPS challenge day, body temperature was higher in the CytoSorb group. As this measurement was obtained shortly after placement of the dialysis catheter, this difference is probably caused by the fact that subjects in the CytoSorb group were covered by a sterile cloth for 20–30 min during the cannulation procedure, whereas subjects in the control group (in whom no dialysis catheter was placed) were not covered. Nevertheless, the LPS-induced increase in body temperature during the first challenge was less pronounced in subjects treated with CytoSorb (p = 0.006), whereas no significant group differences in body temperature were observed upon the second LPS challenge (p = 0.77, Fig. 4C). Although the onset of the LPS-induced increase in the composite score of flu-like symptoms appeared to be earlier in subjects in the CytoSorb group, this difference did not reach statistical significance (p = 0.08, Fig. 4D). Further exploration of individual symptoms, displayed in Additional file 5: Fig. S5, revealed that the sensation of shivering was experienced earlier and more pronounced in the CytoSorb group (p < 0.0001, Additional file 5: Fig. S5E). In line, the sensation of general malaise was also experienced earlier and more severely in subjects treated with CytoSorb during the first LPS challenge (p = 0.004, Additional file 5: Fig. S5F). Treatment with CytoSorb did not influence symptom score severity during the second LPS challenge, with exception of a slightly more pronounced sensation of myalgia in the CytoSorb group (p = 0.004, Additional file 5: Fig. S5C).

Hemocytometry data during both LPS challenges are displayed in Additional file 6: Fig. S6. Administration of LPS induced a minor decrease in platelet counts, and this effect was slightly more pronounced in subjects treated with CytoSorb hemoperfusion during the first LPS challenge (p < 0.001, Additional file 6: Fig. S6B). Similarly, the LPS-induced increase in total leukocyte and neutrophil counts was also more apparent (p = 0.006 and p = 0.04, respectively, Additional file 6: Fig. S6C, D) in the CytoSorb group during the first challenge, whereas the reduction in lymphocyte counts was less pronounced (p < 0.0001, Additional file 6: Fig. S6E). Although treatment with CytoSorb did not affect the disappearance of monocytes from the circulation within 1–2 h after LPS administration, the number of circulating monocytes returned to baseline concentrations significantly earlier in the CytoSorb group during the first LPS challenge (p = 0.002, Additional file 6: Fig. S6F). During the second LPS challenge, no differences in any of the measured hemocytometry parameters were observed between the two groups.