Forty-seven patients were enrolled in this prospective study. Study approval was obtained from the Ethics Review Board of the University of Düsseldorf (Düsseldorf, Germany). Patients with blunt or penetrating multiple injuries who were admitted to our Level I Trauma Center with an Injury Severity Score (ISS) >16, intensive care unit (ICU) stay >3 days and ages 18 years and older were enrolled in this study. Written, informed consent was obtained from all participants or their legal representatives if the patients were unconscious. Exclusion criteria were death of the patient on the day of admission or within the first 2 days on the ICU and withdrawal of patient consent. In addition, patients with known preexisting immunological disorders or systemic immunosuppressive medication were excluded. The severity of injury was assessed by using the ISS, which is based on the Abbreviated Injury Scale (AIS) [²²], on admission to the emergency room. SIRS and sepsis were defined using the criteria outlined in 2005 by the International Sepsis Forum [²³]. SIRS was considered to be present when patients' conditions fulfilled more than one SIRS criterion. Patients were determined as septic if they fulfilled criteria for SIRS and had a proven source of infection. To evaluate organ dysfunction and/or failure, the Sequential Organ Failure Assessment (SOFA) and Multiple Organ Dysfunction (MOD) scores [²⁴] were determined. Severe sepsis referred to sepsis complicated by organ dysfunction. Organ dysfunction has been defined using the definition by the SOFA score with >2 points for at least one system (respiratory, coagulation, liver, cardiovascular, central nervous or renal system). Septic shock was defined as sepsis with acute persistent circulatory failure unexplained by other causes (>2 points in SOFA score for the cardiovascular system).

The patients included in this study did not receive low-dose hydrocortisone therapy as routine adjuvant treatment for septic shock. Seventeen healthy volunteers served as the control group.

Blood was collected from healthy volunteers and daily from patients from the day of admission until day 9. Heparinized blood was immediately used after collection for neutrophil isolation. In parallel, sera and plasma were harvested by centrifugation and stored at -80°C until further processing.

sFas (detection limit <47 pg/mL), sFasL (detection limit <12 pg/mL) (both evaluated by Hoelzel Diagnostika, Cologne, Germany) and interleukin (IL)-6 (detection limit <0.70 pg/mL) (evaluated by R&D Systems, Wiesbaden-Nordenstadt, Germany) were measured in serum and PMNE (detection limit 3 ng/mL) (evaluated by Milenia Biotec, Gießen, Germany) in plasma samples by using commercially available enzyme-linked immunosorbent assay (ELISA) kits according to the manufacturer's instructions.

Human neutrophils were isolated by discontinuous density gradient centrifugation using Percoll medium (Biochrom, Berlin, Germany) as previously described [²⁵]. After hypotonic lysis to remove contaminating erythrocytes, cells were suspended in phosphate-buffered saline (PBS). Purity and viability were routinely >95% as assessed by forward and side scatter characteristics of FACScan (BD Biosciences, Heidelberg, Germany) and Trypan blue exclusion, respectively.

The monoclonal anti-Fas antibody clone ZB4 (2 μg; Millipore, Schwalbach, Germany) was mixed with 40 μL of Protein G Plus/Protein A-Agarose beads (Calbiochem, Darmstadt, Germany) and incubated for 3 hours with gentle shaking. Then pooled serum from four septic patients was added and incubated for an additional 17 hours at 4°C with gentle shaking. Bound immune complexes were spun down, and the supernatant was stored at -80°C until use.

To neutralize the apoptotic activity of agonistic anti-Fas immunoglobulin (Ig) M antibody (clone CH-11; MBL, Woburn, MA, USA), antibodies (50 ng/mL) were first incubated with recombinant human sFas (R&D Systems, Wiesbaden-Nordenstadt, Germany) for 1 hour and then added to freshly isolated neutrophils (1 × 10⁶/mL) from healthy controls. Cells were further cultured with anti-Fas antibodies in the presence of sFas for 18 hours in RPMI 1640 medium containing 2 mM glutamine (Biochrom, Berlin, Germany) and supplemented with 5% fetal calf serum (FCS) (PAA Laboratories, Coelbe, Germany), 100 U/mL penicillin and 100 μg/mL streptomycin (Invitrogen, Karlsruhe, Germany) at 37°C in a humidified atmosphere containing 5% CO₂ before being assessed for apoptosis.

Additionally, pooled patient serum and sera immunoprecipitated with ZB4 were used to block the activity of agonistic anti-Fas antibodies (clone CH-11; 200 ng/mL). After 1 hour of incubation, patient serum (10%) containing CH-11 antibodies was added to freshly isolated control neutrophils (1 × 10⁶/mL). Cells were further cultured overnight in RPMI 1640 medium containing 2 mM glutamine (Biochrom, Berlin, Germany) and supplemented with 100 U/mL penicillin and 100 μg/mL streptomycin (Invitrogen, Karlsruhe, Germany) at 37°C in a humidified atmosphere containing 5% CO₂.

Neutrophil apoptosis was measured by flow cytometry as the percentage of cells with fragmented DNA using the method described by Nicolleti et al. [²⁶]. Briefly, cell suspensions of freshly isolated neutrophils or those incubated overnight were centrifuged at 450 × g for 5 minutes, and then cells were suspended in 300 μl of hypotonic fluorochrome solution (50 μg/mL propidium iodide in 0.1% sodium citrate plus 0.1% Triton X-100). Cell suspensions were stored in the dark at 4°C for at least 3 hours before they were analyzed by flow cytometry (BD Biosciences, Heidelberg, Germany). A minimum of 10,000 events were counted per sample. Results are represented as the percentage of hypodiploid DNA (sub-G1; percentage apoptosis) corresponding to fragmented DNA characteristics for apoptotic cells.

To evaluate differences between the study groups, a Kruskal-Wallis test with Dunn's post hoc test was performed. Correlation between numerical values was evaluated by using Spearman's rank-correlation coefficient (r). Nonparametric receiver operating characteristics (ROC) curves were generated in which the value for sensitivity (true positive rate) was plotted against the false-positive rate (1 - the value of specificity). Analyses were performed using GraphPad Prism software (version 5; GraphPad Software, San Diego, CA, USA). Comparison of ROC curves was performed with MedCalc software (version 11.1.1, MedCalc Software, Mariakerke, Belgium) using the method described by Delong et al. [²⁷]. Data were considered to be statistically significant at P < 0.05.

The 47 patients (31 male, 16 female) enrolled in this study had a mean ISS of 32.9 ± 1.7 (range, 16 to 57). The patients' mean age was 45.9 ± 2.9 years (age range, 20 to 96 years). Among all patients, 18 (38.3%) developed sepsis within 6.1 ± 0.3 days (range, 4 to 9 days) after admission. Among the septic patients, nine patients met the criteria for severe sepsis and four patients met the criteria for septic shock. The infection site of sepsis and microbiological pathogens for each patient are given in Table 1. Five patients died posttraumatically after 30.7 ± 12.3 days (range, 16 to 55 days) as a consequence of multiple organ failure (MOF). The mean ICU stay was 18.1 ± 2.6 days (range, 3 to 74 days). The mean age of the 18 patients (3 female, 15 male) who subsequently developed sepsis (sepsis group) was 53.5 ± 4.6 (range, 20 to 78 years). The mean ISS in this patient group was 36.7 ± 2.8 (range, 16 to 50). Further patient characteristics as well as injury severity and outcomes are shown in Table 2.

Levels of sFas and sFasL were determined in the serum of healthy volunteers (control group) and patients within 24 hours after admission (day 1), at day 5 and at day 9 after major trauma (Figure 1). Patients were divided in two groups: those who subsequently developed sepsis and those with uneventful recovery after major trauma.

Within the first day after admission, sFas values of patients who subsequently developed sepsis, but not the sFas values of those with uneventful outcomes (median, 101.6; interquartile range (IQR), 66.62 to 156.9), were significantly increased (median, 122.5; IQR, 92.84 to 230.7; P < 0.05) compared with the healthy control group (median, 70.29; IQR, 42.9 to 93.29) (Figure 1a). Furthermore, sFas levels in these patients remarkably increased within the next days and peaked at day 5 after trauma (median, 230; IQR, 145.2 to 291.2), whereas the values for patients without development of sepsis normalized at this time point (median, 86.46; IQR, 62.95 to 114.2). sFas values in the sepsis group remained enhanced until day 9 after trauma (median, 187.1; IQR, 80.22 to 297.4) compared with values in the nonsepsis group at the same time (median, 68.95; IQR, 52.44 to 128.8). Significant intergroup differences were detectable between patients with sepsis development and healthy volunteers at day 1 (P < 0.05), day 5 (P < 0.001) and day 9 (P < 0.01). Additionally, sFas levels increased significantly in sepsis patients at day 5 (P < 0.01) and day 9 (P < 0.05) compared with the nonsepsis patients.

In contrast, for both groups (with or without sepsis), sFasL values were on an equivalent level compared with that of healthy controls throughout the entire observation period (P > 0.05) (Figure 1b) and did not show any intergroup differences.

It is well established that sFas may bind to membrane-bound FasL, thus blocking binding of the ligand to the Fas receptor and preventing apoptosis induction in the target cell. Therefore, we assumed that elevated serum levels of sFas may inhibit apoptosis in circulating neutrophils and promote prolonged cellular activity. Neutrophil apoptosis in both patient groups, those who developed sepsis subsequently and those with an uneventful recovery, was significantly reduced within the first day after trauma and continued to be reduced for the entire period until day 9 after trauma. The sepsis patients had a lower rate of neutrophil apoptosis at day 5 (median, 0.94; IQR, 0.6 to 1.67; vs. median, 2.08; IQR 0.64 to 3.36; in the nonsepsis group) and day 9 (median, 0.38; IQR, 0.28 to 0.94; vs. median, 0.47; IQR 0.26 to 2.5; in the nonsepsis group), although this difference did not reach the level of significance (Figure 2a).

We therefore speculated that sFas prevents the activation of Fas on trauma neutrophils, leading to strong inhibition of neutrophil extrinsic apoptosis in sepsis. To prove this hypothesis of sFas-mediated apoptosis inhibition, neutrophils from healthy donors were incubated with an agonistic anti-Fas antibody (CH-11) in the presence of serial dilutions of recombinant human sFas, which has been shown to inhibit FasL-induced apoptosis of Jurkat cells [¹³]. As depicted in Figure 2b, we found that sFas blocks apoptosis in a concentration-dependent manner.

We further investigated whether sFas in the sera of patients with sepsis development might also inhibit CH-11-mediated neutrophil apoptosis. Patient serum contains a broad range of cytokines, especially high levels of granulocyte macrophage colony-stimulating factor (GM-CSF), which is known to reduce the neutrophil apoptosis rate during inflammation by inhibiting the intrinsic apoptosis pathway [²⁸]. Because serum containing high or moderate levels of sFas might also differ in the concentrations of the cytokines mentioned above, we pooled sera from four sepsis patients before immunoprecipitation of sFas by anti-Fas antibodies (ZB4). Then sera were further used to block the proapoptotic activity of CH-11 monoclonal antibodies. As depicted in Figures 2c and 2d, neutrophils incubated with agonistic CH-11 antibodies and sera from sepsis patients immunoprecipitated with ZB4 (low sFas levels) displayed a twofold increased apoptosis rate when compared with cells cultured in the presence of CH-11 antibodies and pooled serum samples (control; high sFas levels).

As shown in Figure 3a, leukocyte counts were found to be significantly increased in septic patients at day 9 after trauma (median, 12.7; IQR, 9.4 to 17.75) compared with the number of leukocytes determined in the nonsepsis group at day 1 (median, 7.7; IQR 6.05 to 9.85; P < 0.01) and day 5 (median, 7; IQR, 6.3 to 10.7; P < 0.05). Neutrophil degranulation was further examined by assessing the levels of PMNE in patients' plasma (Figure 3b). PMNE showed peak levels at day 5 in patients who developed sepsis (median, 301.4; IQR, 217.5 to 474) compared with controls (median, 165.9; IQR, 123.1 to 184.4) and in patients with uneventful recovery (median, 162.8; IQR, 111.4 to 268.9; P < 0.05). Interestingly, PMNE values as well as leukocyte counts were found to correlate with serum sFas concentrations in the sepsis group at day 5 after trauma (r = 0.49; P < 0.05 for both).

IL-6 is a widely accepted inflammatory parameter in response to major trauma and sepsis. Therefore, IL-6 values in patient serum were determined and correlated to the sFas values. As depicted in Figure 3c, IL-6 values of both groups were elevated at day 1 compared with control values, but decreased simultaneously on the following days. Differences were significant at day 5 and day 9 between the sepsis group (day 5: median, 191.7; IQR, 57.37 to 282.2; day 9: median, 54.94; IQR 29.51 to 191.4) and the nonsepsis group (day 5: median, 41.82; range, 22.74 to 69.43; P < 0.05; day 9: median, 19; IQR, 4.11 to 26.75; P < 0.05). In all patients, IL-6 showed a positive correlation with SOFA score at all time points as well as with the MOD score at days 5 and 9 (Table 3). Furthermore, a strong correlation was determined between IL-6 and sFas at day 5 (r = 0.42; P < 0.01) and day 9 (r = 0.4; P < 0.05), but not at day 1 after major trauma. No correlation between sFas and IL-6 values was found in patients with sepsis development (sepsis group).

To investigate the predictive potential of sFas for the development of sepsis after major trauma, sFas values were additionally correlated to SOFA and MOD scores (Table 3). Elevated sFas concentrations determined in patients with sepsis development after severe trauma strongly correlated with patients' SOFA scores from day 1 until day 9 after trauma. In this patient cohort, sFas values at day 5 were also significantly correlated to the MOD score and were positively associated with the development of multiple organ dysfunction (Table 3). However, sFas did not correlate with SOFA and MOD scores of patients with uneventful recovery.

To verify the prognostic potential of sFas in relation to the established prognostic marker IL-6 for sepsis development after major trauma, we established ROC curves for both parameters at each time point. Figure 4 shows ROC curves of sFas and IL-6 at day 1 and day 5 after trauma. Pairwise comparison of the ROC curves displayed no statistical difference between the area under the curve (AUC) for the sFas and IL-6 values at the depicted time points (day 1, P = 0.694; day 5, P = 0.911).

^aISS, injury severity score; ICU, intensive care unit length of stay; Max SOFA, maximal Sequential Organ Failure Assessment score; ^bP < 0.05 between sepsis and nonsepsis groups.

^aSOFA, Sequential Organ Failure Assessment score; MODS, Multiple Organ Dysfunction Score; sFas, soluble Fas; IL-6, interleukin-6; NS, not significant; ^bP < 0.05; ^cP < 0.01; ^dP < 0.001.