S100A8/A9 and sRAGE kinetic after polytrauma; an explorative observational study

Traumatic injury is the leading global cause of mortality for individuals between 15 and 44 years of age [1, 2]. Patients surviving their initial resuscitation following severe trauma are at risk of subsequent complications such as sepsis and multiple organ dysfunction syndrome (MOD), which are responsible for late morbidity and mortality [3‐5].

Tissue injury secondary to trauma results in the release of endogenous molecules, that have been called damage-associated molecular patterns (DAMPs). These endogenous molecules, also known as alarmins, are released passively from necrotic cells, or actively by alternative secretion pathways, and contribute to tissue damage [6, 7].

Alarmins activate immune responses and stimulate the migration of immune cells to the site of injury (reviewed in [8]). They bind to several receptors, including the receptor for advanced glycation end-products (RAGE), Toll-like receptor 4, and the formyl peptide receptors [9].

The most extensively characterized alarmin released in human trauma is the high mobility group box nuclear protein 1 (HMGB1) which is a trigger of organ dysfunction after tissue injury [10, 11]. The release of other alarmins, such as DNA (cfDNA) and mitochondrial DNA has also been documented following trauma [12, 13].

RAGE, a pattern-recognition receptor triggering non-infectious pro-inflammatory processes in response to danger is the target of various DAMPS, including advanced glycation endproducts, members of the S100/calgranulin family, HMGB1 or amyloid-β peptide [14‐16]. Its basal expression is the highest in the lung and it is a known marker of lung injury [16‐18] RAGE can be cleaved by proteolysis, resulting in the circulation of its soluble component. Soluble RAGE (sRAGE) acts as a decoy receptor, which can be rapidly released after various diseases or surgeries and has been implicated in pro-inflammatory or protective phenomenon, a role that remains controversial or disease dependent [19, 20]. Cohen et al. measured sRAGE level in the emergency room (ER) very early after trauma (< 24 h) and showed an association with the severity of injury, coagulopathy and endothelial activation, as well as with late organ dysfunction, particularly in the kidney [21].

One other interesting well-characterized ligand of RAGE is S100A8/A9, also known as calprotectin, but not investigated in the condition of trauma. S100A8/A9 is a dimer of two small calcium-binding proteins, S100A8 and S100A9. It is highly expressed by neutrophils and monocytes, and by activated epithelial and endothelial cells [22, 23]. High persistent levels of S100A8/A9 have been documented in acute and chronic inflammation, tissue remodeling, and infection [24‐27]. S100A8 and S100A9 regulate inflammatory responses, while S100A8/A9 is involved in bacterial clearance and inhibits microbial growth by chelating divalent cations [28, 29].

Our focused interest for the S100A8/A9 complexes derived from the context of known sterile inflammation, tissue repair and infectious risks within the trauma population. Besides, the knowledge that S100A8/A9 can amplify the endotoxin-triggered inflammatory responses of phagocytes [30], coupled to our observation of endotoxin circulation within the same cohort [31], made it attractive to explore.

The objective of this study was to characterize the kinetics of the S100A8/S100A9 complex, as well as its soluble receptor sRAGE, over time after trauma, as potential early biomarkers of subsequent organ dysfunction.

The demographic characteristics of the study population are presented in Table 1. Blood samples were available for analysis in 38 (100%) patients on day 0, in 36 (95%) on day 1, in 33 (87%) on day 3, and in all four sampling time points for 30 (79%) of the enrolled subjects.

The median duration of ventilation was 10 (IQR: 6–16) days, ICU length of stay (LOS) was 9 (IQR: 4–13) days and hospital LOS was 20 (IQR:11–35) days. The mortality was 21%, with 3 patients out of 8, who died within the sampling time (before day 5). An infection was adjudicated to be present in 13 (34%) of patients during the 10 days of observation; the median day of infection was Day 4 (range: 2–9).

The S100A8/A9 complex increased significantly over time (p = 0.001; Friedman test), whereas the sRAGE circulation levels decreased significantly (p < 0.0001; Friedman test), as shown in Fig. 1.

Our study is showing that S100A8/9 levels increased over time, but is lower by Day 5 in patients with the highest ISS at admission. Even more intriguing, our findings reveal an association with infectious complication in patients with the least increase of S100A8/9 circulation over time.

We also complete the data of Cohen et al. [21] and show that the sRAGE levels persist following multiple trauma. We do not find any association with ISS or thoracic injury, but report that its decline over time is in relation with shock and markers of ischemic injury (lactate, need for transfusion).

We also confirm the observation regarding the association of sRAGE at admission and the development of renal failure. Finally, we find significantly higher sRAGE levels on day 0 and day 1 in non-survivors. Our small cohort does not allow us to adjust for other characteristics.

Cohen et al. reported that sRAGE plasma levels, measured at one time-point, increased early after trauma admission in the ER and correlated with the severity of traumatic injury and organ dysfunction particularly renal [1, 2, 21]. We describe a declining sRAGE kinetic after trauma, which question the reason of this phenomenon. It is in contradiction with the theory of renal accumulation suggested by Cohen, but reflects probably the initial insult that will lead to further deterioration of the organ function [21].

The alternative explanations of “shedding” exhaustion versus the consumption of this decoy receptor could be hypothesized, but cannot be elucidated with our work. Indeed, a variety of alarmins have the potential to target a circulating receptor, including HMGB1, which has been showed by Peltz et al. to be persistently elevated up to 136 h after trauma [32]. The further evaluation of this phenomenon would be of interest.

Since, we could assume that the principal source of sRAGE come from the lung where its expression is the highest by far [18], the association with respiratory failure would make sense. We find no association with respiratory failure (MODS or P/F ratio). However, an inverse correlation of sRAGE at day 3 with the ventilation duration could be the reflect of higher alarmin’s release, if we adhere to the sRAGE consumption theory.

The S100A8/9 circulating complex that we studied is also called calprotectin and actively secreted during phagocytosis. It is considered an endogenous DAMP and activator of TLR4 [33]. It is involved in inflammation, enhancement of neutrophils phagocytosis and bacterial clearance [29, 34]. In the setting of trauma cellular injury and ischemia, we were not surprised to measure an increasing level of circulation. Interestingly, the more severely injured reached lower maximum circulating levels, particularly after 5 days.

The patients where the S100A8/9 circulating complex did not increase significantly between D0 and D5 were more prone to infection. Could this be an explanation for the higher risk of infection in severe trauma [3‐5, 35]? It could be speculated that the lack of increase of S100A8/9 over time could be the indirect sign of immune cells exhaustion and decreased immune defense against infections. However, the persistence of systemic inflammation renders the prediction of infection challenging after trauma, as illustrated for example with the study of C-reactive protein [36]. Further studies are needed to validate the role of S100A8/A9 as a biomarker to predict infection after trauma and test its utility for antibiotic early prescription.

As the literature reports that S100A amplify the endotoxin-induced inflammation, we have tested their association (data not shown) [30]. We found none, probably due to our small number of patients and other cofounders.

This study has some limitations. First, it is a relatively small cohort of patients to study the potential of molecular biomarkers to predict outcome post trauma and prognostic inferences are limited by our univariate comparisons; a larger cohort is needed in order to confirm the predicted value of our biomarkers in a multivariate model. Second, the physiological evaluation was limited to general parameters and not determined by more targeted measurement. Finally, although the infectious outcome was not protocolized, it was nevertheless established by adjudication.

In summary, our study is the first to describe the circulation of the S100A8/9 complex in an acute setting and shows its association with infectious risks. We also report for the first time the kinetics of sRAGE in the early phase after multiple trauma. This work illustrates the importance of following the circulating level of biomarker overtime after trauma in order to better define their potential to predict post trauma complications. The utilization of S1008/9 as a tool to stratify infection risk and trigger early interventions need to be validated prospectively.