Severe sepsis is the cause of 9% to 22% intensive care unit admissions and is associated with a mortality rate up to 50% [1]. Bacterial antigens trigger the initial cytokine response to infection, which is necessary for the clearance of invading pathogens, but overwhelming activation of immune cells, with excessive production of pro-inflammatory cytokines such as tumor necrosis factor (TNF)-α and interleukin (IL)-6, is thought to be responsible for the clinical manifestation of septic shock [2, 3].

Recognition of pathogen associated molecular pattern (PAMP) by the innate immune system is mediated by pattern recognition receptors (PRR) on leucocytes and epithelial cells.

TLR4, together with CD14 and the MD2 adapter molecule, serves as the main receptor for components from gram negative bacteria such as lipopolysaccharide (LPS) [4], whereas TLR2 is crucial to the propagation of the inflammatory response to components mainly from gram-positive organisms, yeast and mycobacteria including lipoteichonic acid (LTA) and lipoarabinomannan [5–7]. In addition, TLR2 is activated by bacterial peptidoglycan, bacterial lipoproteins and lipopeptides, cell wall structures expressed on virtually all clinically relevant gram positive and gram negative bacterial pathogens [8]. TLR2 might also be involved in LPS induced cell activation since anti-TLR2 antibody partially inhibits IL-12 production of human dendritic cells [9]. TLR activation causes nulear factor kB translocation and mitogen-activated protein kinase phosphorylation, resulting in an enhanced production of inflammatory cytokines such as TNF, IL-1 and IL-6 [4]. In murine macrophages TLR4 expression correlates with the inducibility of the proinflammatory response to LPS [10], whereas stimulation with TLR2 agonist has been associated with the rapid release of IL-10 [11].

The central role of TLR2 and TLR4 in microbial responses suggests that they may be implicated in the pathophysiology and the outcome of human sepsis [12]. Beside initial systemic release of proinflammatory cytokines, prolonged cellular hyporesponsiveness to bacterial components with reduced cytokine response is thought to be a key factor in sepsis with limitation of subsequent ability to mount an appropriate inflammatory defense to secondary infections. TLR regulation might be implicated in hyporesponsiveness, since prolonged LPS-stimulation of human macrophages caused downregulation of TLR2 [13]. In addition tolerance in vitro to bacterial lipoprotein, a compound of gram-positive and gram-negative bacteria, is associated with reduced TLR2 expression [14].

In contrast to data describing TLR downregulation after LPS in vitro, murine sepsis is associated with an increased TLR4 protein expression in tissue [15]. In human sepsis increased protein expression and mRNA of TLR2 and TLR4 on blood neutrophils and monocytes are found compared to healthy individuals [16, 17]. Thus, experimental models in animals and septic human patients display significantly upregulated TLR expression. But, in contrast to the resting state and nonseptic situation peripheral blood monocytes from septic patients secrete reduced quantities of proinflammatory cytokines regardless of their up regulated TLR-expression [18–21], possibly due to intracellulary inhibitory processes [22].

In addition to the questionable functional activity of TLR receptors during sepsis, expression of TLR during human sepsis was not associated with clinical outcome.

We hypothesized that TLR and CD14 expression is increased in sepsis, but inadequately less increased in severe sepsis compared to nonsevere sepsis, possibly associated to monocyte hyporesponsiveness.

We therefore compared the TLR2, TLR 4 and CD14 expression on blood monocytes of patients with sepsis and healthy controls. In patients with sepsis the TLR expression on monocytes was measured sequentially (admission, day 1, day 3, day 7) and correlated with sepsis severity and mortality. To evaluate the functional relevance of the receptor expression, we correlated the TLR and CD14 expression with the cytokine release after stimulation with LPS in a whole blood assay.

As shown by other investigators before, CD14, TLR2 and TLR4 protein expression on monocytes is elevated during sepsis. Interestingly we could show in our study a relatively lower receptor expression (CD14, TLR2 and as a trend in TLR4), in patients with sepsis that died, compared to patients that survived (Table 2). In addition we were able to demonstrate a correlation between receptor expression and cytokine production after LPS stimulation, leading to the assumption that the receptors on monocytes are functionally active during sepsis. These data underline the pathophysiologic role of TLR2 and CD14 for the activation of the immune system and the monocyte hyporesponsiveness in severe human sepsis.

As microorganisms trigger the release of cytokines via TLR's and CD14, these receptors are thought to play a central role in the pathophysiology of sepsis. The ability of the host to sense invasion of pathogenic organism and respond appropriately to control infection is critical to survival. Sepsis can induce monocyte hyporesponsiveness, limiting the antigen induced cytokine release causing immunosuppression. In vitro LPS stimulation is able to cause a downregulation of TLR2 and TLR4 gene expression and surface expression on mouse macrophages inducing monocyte hyporesponsiveness [25–27]. In contrast to downregulation of TLR gene expression after LPS exposure in cell culture, sepsis in mice [15] and humans sepsis causes an upregulation of TLR2, TLR4 and CD14 expression on blood monocytes (table 1). Similar data have been shown by other authors before [28, 29]. Continuous microbial stimulation during bacterial sepsis with a number of different antigenic structures might result in cell activation inducing receptor upregulation [11]. Causative factors might be the release of cytokines during sepsis like IL-6, which has been shown to upregulate TLR4 on human monocytes [21]. In addition neutrophil products like elastase induces TLR4 expression on human monocytes [16].

In our study patients that died had lower cytokine inducibility at all timepoints and therefore more intensive hyporesponsiveness than patients that survived (n.s.). Correlating with this result, lower CD14 and TLR2 expression was found in patients that died (Table 2). In addition, at day 3 cytokine in vitro inducibility (24 hours whole blood LPS-stimulation) correlated with the TLR2 and CD14 protein expression on blood monocytes. Our data indicate that during the course of sepsis the receptors gain again functional activity, with higher cytokine release after in vitro antigen stimulation in cells with higher receptor expression. The data with lower receptor expression in patients that died during sepsis, confirm recent data, showing reduced gene expression for TLR receptors in monocytes of patients according to their state of sepsis: lowest TLR gene expression in septic shock, followed by severe sepsis and sepsis [30, 31]. In addition to Salomao's gene results, we were able to associate TLR protein expression to sepsis severity. In conclusion, inadequate receptor activation might be deleterious for patients with sepsis.

A few years ago, TLR2 was thought to be a specific receptor for antigens from gram positive and the TLR2/CD14 complex was thought to be specific for gram negative antigenic structures. Several recent studies have shown that these receptors are not as specific as thought. Lipoteichonic acid (LTA) from S. pneumonia is able to activate TLR2 and CD14 [32]. In contrast LPS from gram negative bacteria is able to activate also CD14 and TLR2. The contribution of these receptors for host response during infection has been studied in animal knock out models. For example, TLR2 knock out causes reduced inflammatory response to pneumococcal LTA [32]. A strong pro-inflammatory reaction is necessary for local bacterial clearance. As shown in our study, an inadequate downregulation of TLR2 and CD14 (and consecutive reduced cytokine release) during sepsis might have a negative influence on outcome. In addition to reduced bacterial clearance, monocyte hyporesponsiveness has been associated with increased number of nosocomial infections in septic patients. Beside innate immune response, the TLR2 induced cytokine response might also be involved in the physiologic endocrine response to sepsis, since TLR2 deficient mice have an impaired cytokine mediated corticosteroid adrenal stress response [33]. Taken together, these data suggest that a reduced TLR2 and CD14 expression during sepsis might have a negative outcome in sepsis.

Therapies targeting TLR-pathways are currently being investigated for several diseases. Whether TLR blocking [34] or TLR agonists [35] are needed in sepsis is under discussion. Broad knowledge about the regulation of TLR and CD14 induced signalling pathways are needed to anticipate the clinical effect. Our data suggest that additional TLR blocking in patient with sepsis might be dangerous.

In conclusion septic patients are characterized by an increased TLR2, TLR4 and CD14 expression. Death due to sepsis is associated with TLR2 and CD14 downregulation. The precise role of TLR and CD14 regulation in sepsis should be evaluated in further clinical and experimental studies, especially before TLR blocking agents are clinically investigated.