Release of tissue-damaging factors from infectious agents (such as bacteria, viruses or fungi) or from cells (for example, endothelial or epithelial) after hypoxia, hemorrhagic and septic shock, polytrauma, and other conditions induces release of proinflammatory mediators, the outcome of which is often tissue and organ damage1. Molecules released from infectious agents have been described as pathogen-associated molecular patterns (PAMPs), whereas danger-associated molecular patterns (DAMPs) are those released from damaged tissues in the absence of infectious pathogens, also called 'sterile inflammation'2,3,4,5. PAMPs and DAMPs seem to function via interactions with Toll-like receptor 2 (TLR2) and TLR4 (refs. 3,5), as well as with NOD-like receptors. To date, the only therapies available to patients with hemorrhagic or septic shock are supportive and include fluid resuscitation, pressor drugs for stabilization of blood pressure, replacement of red cells when appropriate, and ventilator support that is not damaging to lungs.

In this issue of Nature Medicine, Qiang et al.6 identify a new molecule, cold-inducible RNA-binding protein (CIRP), which appeared in blood from patients and in animal models after hemorrhagic or septic shock. CIRP mediated inflammation and tissue injury, joining the list of DAMPs involved in the pathogenesis of shock, and may therefore be a target for therapeutic blockade in related clinical conditions in humans.

Although CIRP was originally described in cells exposed to cold temperatures as a suppressor of mitosis and a promoter of cell differentiation7, other conditions (hypoxia, ultraviolet radiation and mild hypothermia) can also induce CIRP expression. Qiang et al.6 found CIRP in the sera from ten patients suffering from hemorrhagic and septic shock in a surgical intensive care unit, but not in sera from healthy donors. Although the authors also showed release of CIRP into the circulation in rats and mice after the onset of hemorrhagic shock, future investigation will have to include clinical studies involving more patients with sepsis, severe sepsis and septic shock, according to an earlier classification7, to confirm the role of CIRP in these cases.

To investigate how CIRP is released into the bloodstream and whether it is mediating an inflammatory response related to shock, the authors used a mouse macrophage-like cell line, peripheral human blood mononuclear cells, the THP-1 cell line (human monocytes) and peritoneal mouse macrophages6, given the crucial role of macrophages in inflammation. As expected, under conditions of normoxia, cells contained nuclear CIRP, whereas hypoxia followed by reoxygenation induced a relocation of CIRP to the cytosol after several hours. CIRP was found in serum and in liver homogenates 4 to 6 h after hemorrhagic shock, suggesting the liver as a possible source of CIRP.

Macrophages exposed in vitro to purified CIRP expressed both tumor necrosis factor-α (TNF-α) and high-mobility group protein B1 (HMGB1), the former a strong proinflammatory mediator and the latter a DAMP that causes proinflammatory cytokines to be released from macrophages, and this effect was recapitulated upon administration of purified CIRP to healthy rats. However, in most cases, expression of TNF-α was extremely low (<5 ng/ml), raising the question of what the inflammatory role of TNF-α is under such conditions. Qiang et al.6 also found that induction of expression of these inflammatory mediators by CIRP signaling required TLR4 expression in macrophages, which is a common but not exclusive receptor of PAMPs and DAMPs3.

In vivo evidence bolstered the role of CIRP in hemorrhagic and septic shock in mice and rats. Survival rates were substantially improved after CIRP inhibition with a neutralizing antibody to CIRP6, and mice lacking CIRP showed similar improved survival after hemorrhagic shock. A probable sequence of events in this scenario may be transcriptional upregulation of CIRP during hemorrhagic shock and other conditions, after which it is then secreted into the extracellular spaces (Fig. 1). Interaction of CIRP with the TLR4-MD2 complex in macrophages results in release of HMGB1, although secretion can also occur from monocytes and dendritic cells, whereas TNF-α release occurs in macrophages and monocytes and a large number of other cell types. TNF-α is an early mediator in adverse events of sepsis, whereas HMGB1 is a late-functioning mediator that initiates release of a series of proinflammatory mediators, mainly from macrophages. The addition of HMGB1 and CIRP to a macrophage cell line enhanced the release of TNF-α6, suggesting that HMGB1 and CIRP may collaborate in septic or hemorrhagic shock. The composite result is a hyperinflammatory response (cytokine storm), development of multiorgan failure and, potentially, death.

Figure 1: Circulating and tissue-resident macrophages and/or monocytes release CIRP in response to shock.
figure 1

Katie Vicari

During hemorrhagic or septic shock, CIRP signals via the TLR2-MD2 complex, resulting in release of TNF-a and HMGB1, the latter causing release of other proinflammatory mediators from macrophages. Collectively, this results in multiorgan failure and lethality.

Full size image

Clinical trials over the last few years have enrolled human subjects with severe sepsis who were treated with one of two TLR4-blocking drugs (eritoran and TAK 242)8,9. Both trials were terminated early because of lack of drug efficacy (improved survival). These failed clinical trials in sepsis are especially disappointing in view of the plethora of DAMPs, PAMPs and other products that are reactive, but not exclusively, with TLR4 (ref. 3), which seems to be a common node. It is possible that CIRP binds to other receptors aside from TLR4. If one were to target CIRP in the appropriate clinical setting, neutralizing antibody to CIRP might be preferable to targeting TLR4, at least on the basis of the failed results of the trials with TLR4-blocking drugs. Two very recent reports in endotoxemic shock suggest that lipopolysaccharide, once it enters the cytosol of macrophages, triggers caspase-11 activation, which is independent of TLR4 and may lead to lethal endotoxemia10,11. This suggests caspase-11 could be a drug target in humans with sepsis.

Hemorrhagic shock is thought to be an example of sterile sepsis, so considerable caution and additional basic studies are needed to complement clinical observations before one can consider translational applications to humans with hemorrhagic shock. Also, although the data suggest CIRP is a candidate target to treat septic humans, we are a long way from such a conclusion. We simply do not have a sufficient number of septic patients who fit the various classifications of sepsis or a reliable and sensitive ELISA system to precisely measure CIRP and thus understand the role of CIRP in conditions driven by DAMPs and PAMPs. Nevertheless, the current study reinforces the literature showing that DAMPs, once released, have the potential to unleash numerous effects that can be harmful or lethal in a variety of clinical settings.